The Argument from Queerness: A Debate on the Ontological Status of Moral Facts, Study notes of Ethics

Download The Argument from Queerness: A Debate on the Ontological Status of Moral Facts and more Study notes Ethics in PDF only on Docsity!

Evaluative Realism and the Argument from

Queerness

Luke Townend

PhD

University of York

Philosophy

September 2020

Abstract

Purely Evaluative Moral Realism is the view that there are moral facts, but they are always and only concerned with goodness and badness, without being con- cerned with reasons or obligations. In this book, I argue that such a view has the virtue of avoiding the Argument from Queerness for Moral Error Theory. Firstly, I lay out and examine the Argument from Queerness to show how it is a threat to moral realism. Next, I isolate and argue against one key premise of the Argu- ment from Queerness; that all moral facts are specially normative, which is to say concerned with a distinctive type of reason for action. Denying this claim creates conceptual space for Purely Evaluative Moral Realism. I then describe this view and demonstrate how it avoids the Argument from Queerness. In addition to all this, I defend Purely Evaluative Moral Realism from objections to show that it does not collapse into any sort of incoherence and is a legitimate rival to more familiar metaethical views. In doing so, I argue that Purely Evaluative Moral Re- alism is compatible with a surprisingly wide range of theories within substantive ethics, and that it allows us to account for many of our intuitions about moral language and moral reasons, so long as we accept that moral reasons may be more like our other reasons than is usually thought, and that they lack the features distinct to special normativity.

• Abstract
• Contents
• Preface
• Acknowledgements
• Author’s Declaration
• 1 The Argument from Queerness
  • 1.1 Moral Error Theory
  • 1.2 The Structure of the Argument from Queerness
  • 1.3 Debunking Accounts
  • 1.4 Queerness Arguments
  • 1.5 The Queerness Argument from Moral Knowledge
  • 1.6 The Queerness Argument from Motivation
  • 1.7 The Queerness Argument from Supervenience
  • 1.8 The Queerness Argument from Special Normativity
• 2 Normativism and Non-Normativism
  • 2.1 The Normativity of Moral Facts
  • 2.2 The Connection between Norms and Values
    • 2.2.1 Moorean Consequentialism as a Conceptual Truth
    • 2.2.2 Values to Be Respected
    • 2.2.3 From Personal Goodness to Moral Norms
• CONTENTS - 2.2.4 The Buck-Passing Account of Value
  • 2.3 The Case for and against the Buck-Passing Account
    • 2.3.1 The Ontological Parsimony Argument
    • 2.3.2 The Heterogeneity Argument
    • 2.3.3 The Open-Feel Argument
    • 2.3.4 The Demystification Argument
    • 2.3.5 The Redundancy Argument
    • 2.3.6 The Similarity Argument
• 3 Purely Evaluative Moral Realism
  • 3.1 What Purely Evaluative Moral Realism Is and What It Isn’t
    • 3.1.1 Commitments of Purely Evaluative Moral Realism
      • Views 3.1.2 Contrasting Purely Evaluative Moral Realism with Similar
  • 3.2 Purely Evaluative Moral Realism and the Argument from Queerness - the Queerness of Special Normativity 3.2.1 Purely Evaluative Moral Realism and the Argument from - the Queerness of Supervenience 3.2.2 Purely Evaluative Moral Realism and the Argument from
  • 3.3 Objections to Purely Evaluative Moral Realism
    • 3.3.1 Restriction on Substantive Theories
    • 3.3.2 Queerness Arguments
    • 3.3.3 Purely Evaluative Moral Realism and Action
• 4 Moral Facts and the Guiding of Action
  • 4.1 Purely Evaluative Moral Realism and Substantive Ethics
    • 4.1.1 Scalar Utilitarianism
    • 4.1.2 Virtue Ethics
    • 4.1.3 Deontology by Way of Honouring Values
    • 4.1.4 Divine Command Theory
  • 4.2 Realism about Non-Special Moral Norms
  • 4.3 Objections to Realism about Non-Special Moral Norms
• CONTENTS - Domains 4.3.1 Problems with the Analogy between the Moral and Epistemic - 4.3.2 Denial of the Self-Consciously Moral Motive
• Conclusions and Summary
• A Virtues of Purely Evaluative Moral Realism
  • A.1 Dissolving the Is-Ought Problem
  • A.2 Free Will and Moral Responsibility
• Bibliography

PREFACE 7

The findings of this thesis have a direct bearing on the realm of substantive ethics; if Moral Error Theorists are right, there is nothing for substantive ethicists to study. The claim that I will defend in this thesis is that Purely Evaluative Moral Realism is immune to the Argument from Queerness. Purely Evaluative Moral Realism is realism about a limited class of moral facts, while the Argument from Queerness is perhaps the most influential argument for Moral Error Theory. Purely Evaluative Moral Realism limits the range of possible moral facts where a more conventional unqualified moral realism would not, because it entails that nothing is right or wrong. But it preserves claims about goodness and badness, so it does not eliminate the domain completely as Moral Error Theory does. More will be said about what exactly the Argument from Queerness and Purely Evaluative Moral Realism involve in Chapters 1 and 3 respectively, but hopefully the relevance of the one to the other is already clear, even if the details remain obscure at this juncture. If Purely Evaluative Moral Realism can avoid the Ar- gument from Queerness, it will mean that this argument does not lead directly to Moral Error Theory as many have supposed (though it would still warrant the rejection of a great many first-order moral claims). In order for a view to avoid the whole argument, it must avoid each version thereof that poses some genuine threat to realism, so we can only answer in the affirmative if Purely Evaluative Moral Realism allows for further commitments that meet the conditions to avoid each incarnation of the Argument from Queerness without internal contradiction. One other thing to note is that we can only claim that a view can achieve any- thing in particular if that view passes the test of basic coherence and plausibility. As such, I will offer a full defence of Purely Evaluative Moral Realism. What I will not do, however, is argue that it is the correct view of the metaphysics of morality. I have plenty to say about its virtues and implications, but the task of

or wrongness. As noted in Olson 2014, pp. 11-5, Moral Error Theorists cannot simply say that all moral claims are false without qualification because under their view, claims that something does not have some moral property (negative moral claims) are universally true. Moral Error Theorists can make this move so long as they deny the implication from ‘φ is not wrong’ to ‘φ is right or permissible’ as suggested by Pigden (2007, p. 453), and this seems reasonable since they take all of these categories to be empty.

PREFACE 8

comparing it to alternative views is largely ancillary to our discussion. So long as it is not obviously inferior to alternatives, it will remain a live possibility, which will be sufficient to undermine the Argument from Queerness as a direct route to Moral Error Theory. The first chapter will be devoted to describing the Argument from Queerness and its various incarnations, and establishing the strongest forms thereof to more fully motivate the search for a form of realism that does not fall prey to those argu- ments. The Argument from Queerness concludes that there are no moral facts (or, more precisely, that we should not posit or believe in moral facts) because they do not figure into our best explanations of the world, and specifically of human moral thoughts, feelings, and behaviour. What makes explanations invoking moral facts worse than the alternatives is the queerness of moral facts, which is a particular form of strangeness sufficient to motivate ontological suspicion of those entities that possess it. Thus to accept the Argument from Queerness we must accept one or more queerness arguments, which nominate some feature of moral facts as the source of their putative queerness. I will conclude that two of these succeed, namely the ones concerning moral facts’ normativity and their supervenience on natural facts, meaning that there are two successful versions of the Argument from Queerness to overcome. The second chapter will examine a key premise from one of those successful versions. I mean to reject this premise, which I will label Strong Normativism, as it is directly incompatible with Purely Evaluative Moral Realism. It is the claim that the particular and unique normativity so often attributed to moral facts is conceptually necessary to moral facts, meaning that all such facts would be reason- giving in this queer way. Here, I will posit a distinction between normative facts and evaluative facts, the latter of which are not concerned with reasons or duties but with the degree of goodness or badness of a given thing. If evaluative facts can exist in the absence of normative facts, then Strong Normativism is false. I will go on to examine views that imply a necessary connection between normative facts and evaluative facts, foremost among which is the buck-passing account of value, and I will rebut arguments for these views. This opens the door for a form of moral realism that rejects Strong Normativism. That form of realism will be the subject of the third chapter; Purely Evaluative

PREFACE 10

metaethical puzzles that, in one way or another, are owed to the normativity of moral facts. The Argument from Queerness became the focus of this project because it is, in my estimation, the most well-developed of these, and the most legitimate threat to moral realism.

Acknowledgements

This work was supported by the Arts & Humanities Research Council (grant num- ber AH/L503848/1) through the White Rose College of the Arts & Humanities. I would like to extend my gratitude to my supervisors, Johan Gustafsson and Christian Piller, whose insight and advice have been invaluable to me in assembling this thesis. I would also like to thank Mary Leng, Paul Noordhof, Tom Stoneham, and Richard Yetter Chappell for their valuable commentary on in-progress versions of the text, as well as to Jonas Olson and Charles Pigden for their remarks on some of the core ideas presented herein, and to the membership of my initial examination panel, whose remarks have shaped this revised version of the project. Additional thanks are owed to all the participants in the University of York’s postgraduate work-in-progress seminars, especially Celia Coll, Britt Harrison, and Bridger Landle. Finally, I am immensely grateful for the kindness and support of my mother and father, my brothers, and all of those in York’s Department of Philosophy who have extended their friendship to me over the course of this project.

Chapter 1

The Argument from Queerness

The Argument from Queerness expresses a general worry about moral realism: there is something strange about the idea of moral truth, which doesn’t quite cohere with the rest of the world as we understand it. In this chapter, I will clarify this argument’s assumptions, intended conclusions, and overall structure. The Argument from Queerness consists of two subordinate arguments for conclusions which, taken together, are meant to lead to moral error theory. One of these subordinate conclusions is that we can explain or account for our moral thoughts, feelings, and actions without invoking moral facts. The other is that if there were any moral facts, they would be metaphysically queer. It is this latter conclusion which is of greatest interest for our discussion, so the latter half of this chapter will be devoted to examining the many arguments for it.

1.1 Moral Error Theory

The Argument from Queerness is generally posed as an argument for the onto- logical aspect of Moral Error Theory. Moral Error Theory consists of two core claims: a semantic claim, and an ontological claim. The semantic claim is that moral language is truth-apt, in the sense that when we make moral claims we are trying to express facts about the world, and those claims are capable of being true or false. The Argument from Queerness takes this former claim as given (though Moral Error Theorists usually have independent arguments for it, which I discuss

13

CHAPTER 1. THE ARGUMENT FROM QUEERNESS 14

briefly below). The ontological claim, which is the conclusion of the Argument from Queerness, is that there are no moral facts or properties.^3 I should clarify here what I mean by ‘facts’ in general and ‘moral facts’ in particular. A fact is a state of affairs that obtains, and a fact exists whenever a particular entity has a property or stands in a relation.4,5^ Facts are truthmakers in the sense that the truth values of propositions depends on what facts there are. Insofar as we can assume that truth involves a world-to-content relation, facts fall on the side of the world.^6 All facts are about the distribution of properties (and relations, but I will dis- cuss properties for brevity). Moral facts, then, are about the distribution of moral properties.^7 In other words, they are about whether actions (and sometimes things closely associated with them, such as motives, character traits, and consequences among other things) are good or bad, or right or wrong in a particular way. The particular way in question—the moral—is distinct from other standards by which actions might be assessed, such as practical rationality, but is nevertheless such

(^3) Error theories more generally are views such that all the claims within some domain of

discourse are false. Moral Error Theory fits this schema in that it commits its followers to the falsehood of all positive moral claims, which follows from the ontological claim that there are no moral facts. There are error theories about other domains, to which occasional reference will be made in what follows, but this text is primarily concerned with Moral Error Theory so wherever not otherwise stated it is to this variety that I refer. (^4) Armstrong 1997, p. 1; Finlay 2014, p. 7. Facts might also usefully be described as actual

distributions of properties and relations. While I have borrowed Armstrong’s definition of facts, I should make clear that I do not mean to commit myself to any part of the broader factualist ontology that he develops. (^5) This definition might be simplified if we can account for relations as relational properties,

but that is not a question that we need to settle for the purposes of this project. (^6) Armstrong 1997, p. 13. (^7) Taking this in conjunction with the definition of facts given above, it follows that the claim

that there are no moral facts is equivalent with the claim that there are no moral properties, which is why I spoke of them together when I first laid out the ontological claim of Moral Error Theory.

CHAPTER 1. THE ARGUMENT FROM QUEERNESS 16

As I have already stated above, the Argument from Queerness is not strictly an argument for the whole of Moral Error Theory. Rather it is an argument for the ontological thesis, and so against a particular variety of moral realism—non- naturalist realism. When J. L. Mackie advanced it in his book Ethics: Inventing Right and Wrong, it was alongside other arguments ruling out alternative versions of realism, making it the last of a series of arguments which, taken together, were meant to support Moral Error Theory.10, Moral Error Theorists and non-naturalist moral realists share some key com- mitments about the nature of moral facts, which I will grant for the purposes of this chapter. These commitments are cognitivism and non-naturalism. Cog- nitivists hold that moral claims express a belief held by the speaker, in contrast to non-cognitivists who hold that moral language elliptically expresses emotions, prescriptions, or some other non-cognitive (read: non-belief) attitude. This is ef- fectively the semantic claim of Moral Error Theory, as it makes the content of moral claims propositional and truth-apt; they are statements capable of being true or false. The standard argument used by Moral Error Theorists (although not by them alone) to justify their commitment to cognitivism starts with the claim that the meaning of moral terms is determined by the intentions of ordinary speakers when they use them.^12 These intentions are revealed through ordinary moral discourse, which proceeds as though moral claims express exactly the kind of truth evalu- able proposition that cognitivists hold them to express. This is what allows for full-blooded disagreements over questions of morality, and for moral claims to be embedded in formulae of propositional logic.^13 I will not pretend that this alone

(^10) Mackie 1977, pp. 23-4. (^11) While Mackie’s Argument from Queerness, and many that have followed after it, target

non-naturalist realism, some versions of it may also target naturalist realism (Olson, 2017, pp. 64-5). This is true of those versions which rest on a queerness argument (see below) whose targeted feature of moral facts could be naturalized without undermining the claim that said feature is queer (see §1.4). (^12) Mackie 1977, pp. 32-3. (^13) If moral claims were not actually propositional in content, we would run into the Frege-

CHAPTER 1. THE ARGUMENT FROM QUEERNESS 17

does away with non-cognitivism, even if it is deals a significant blow to simpler forms thereof.^14 I will, however, note that this project is foremost about what the Argument from Queerness proves. As such, it is the dialectic between a robust realism and Moral Error Theory which is of most interest to us. In addition, I will try to meet Moral Error Theorists on their own terms with regard to these questions. I will, therefore, treat cognitivism as given from this point. Non-naturalists deny moral naturalism, which is the claim that moral facts are either identical to, reducible to, or constituted by natural facts.^15 The issue of what is and is not natural is a knotty one, but it will come up often enough over the course of my project to make untangling it here worthwhile. For philosophers, the domain of the natural is that of the familiar temporal world, and we can follow many others in stipulating that our guide to what counts as natural is that which is posited by our best current science.^16 General or metaphysical naturalism (as opposed to moral naturalism) is the view that the natural facts are exhaustive, that there are no facts beyond those countenanced by science.^17 Moral Error Theorists and moral realists alike most often appeal to the normativity of moral facts (see §1.8) to justify their commitment to non-naturalism; there is, they claim, simply no way to account for moral facts’ reason-giving properties in a strictly naturalistic framework.^18 (I am not so strongly married to this position because of

Geach or embedding problem (discussed in Geach 1958; 1965.) (^14) In particular, moral quasi-realism as developed in Blackburn 1984; 1993 is notable in that

it is meant to have the advantage of reflecting or vindicating ordinary moral language while still falling under the non-cognitivist banner. See also Hare 1952 and Gibbard 1990. (^15) This is a gloss, but a good one, which I have borrowed from Brink (1989, pp. 156-7). For a

moral fact to be constituted by one or more natural facts is for it to be ontologically dependent thereupon in the sense that the moral fact is nothing over and above the natural fact or facts (as with reduction) while also being multiply realizable in the sense that the same moral fact or property have been grounded in different natural facts or properties. (^16) See Sturgeon 2006, p. 92, Kitcher 2011, pp. 3-4, and Olson 2014, p. 88. This is also one of

the definitions discussed in Hampton 1998, p. 35. ‘Best’ here means best by science’s own lights. (^17) This does not necessarily imply physicalism or materialism; see Chalmers 1996. (^18) See Mackie 1977, p. 32, Dancy 2006, p. 132, and Parfit 2011a, §134.

CHAPTER 1. THE ARGUMENT FROM QUEERNESS 19

reference to any special class of moral facts. The second subordinate argument is at least one queerness argument, which is to say an argument to the effect that moral facts, if they existed, would be ‘queer’. This notion of queerness is a little obscure, and I examine it more thoroughly in §1.4, but for now suffice it to say that queer entities are ontologically or metaphys- ically suspicious in such a way that we shouldn’t admit them to our picture of the world without very good reasons. In other words, they are a cost or burden to any ontology that includes them. Typically queerness arguments nominate some feature of moral facts, and then argue that said feature is in some way problematic and so, therefore, are moral facts.^19 So it seems that moral realism is under attack from two fronts; its commitments appear redundant in the face of debunking accounts, and inherently problematic thanks to queerness arguments. Neither of these considerations is a knockdown ar- gument against realism on its own, but the threat they pose is significantly greater when they are given in combination. This is not just because of the mounting vol- ume of considerations against it—though this is a kind of evidence in itself—but because each consideration undermines arguments against the other as evidence for the antirealist conclusion.^20 Consider debunking accounts. Taken on its own, a debunking account allows us to make something along the lines of Gilbert Harman’s redundancy argument against moral realism, where we reject moral facts on the basis of their redundancy alone.^21 This involves an appeal to a simple quantitative version of Ockham’s razor, whereby we should not posit more entities (in this case facts) than are

(^19) Olson 2014, p. 84. (^20) The idea that the sheer volume of considerations for or against a view is evidence in itself will

be particularly salient if we take seriously the view that little in philosophy is settled by decisive argument, and that we are better tallying the number and quality of arguments or evidence for a position as a metric of its plausibility. See Rawls 1971, p. 20; Lewis 1986, p. viii; and Enoch 2011, §1.4. (^21) Harman 1977, ch. 1. What debunking accounts actually show is that moral facts would be

explanatorily redundant in the first place, and concluding directly from this that there are no moral facts is a further move we are not forced to make.

CHAPTER 1. THE ARGUMENT FROM QUEERNESS 20

explanatorily required.^22 The realist may wish to interject that all things are not equal because, they take it, the debunking account fails to completely explain moral phenomena, even if it tells a convincing story about the elements that it does explain. Such arguments often turn on the claim that whatever the debunking explanation misses can only be explained by a set of facts with some particular quality for which a naturalistic worldview has no room. However, when called to say what this quality is, realists run the risk of nominating the subject of one of the queerness arguments, in which case the debunking account seems all the more plausible for its lack of such troubling commitments. Of course, I cannot prove that every such suggestion would fall into this trap—it would be question-begging to assume that all non-natural properties must be queer—but the most popular and plausible candidate is the apparent normativity of moral facts, which is the subject of the queerness argument discussed in §1.8. On the other side, one objection that could be fielded against the claim that we should reject moral facts simply because they are queer is that the ‘sheer queerness’ of moral facts does not seem to be enough to warrant rejecting them on its own. While it has come to be a philosophical term of art, when Mackie first used the word ‘queer’ it was simply a synonym for ‘strange’. As Platts points out, there are many strange things in the world such as aardvarks and neutrinos, yet we would look askance at anyone suggesting that because these things are so odd, we had best conclude that they don’t actually exist.^23 But even though these things might initially seem queer, when we place them into context as part of the world as a whole, they form part of our best explanations of things.^24 For example, though aardvarks exhibit many unusual traits, our sensory experiences of aardvarks are better explained by their actual presence than by hallucinations or some manner

(^22) In fact, at least some philosophers do not see the multiplication of particular entities as a

theoretical vice (for example, see Lewis 1973, p. 87). There are arguments to the contrary as well (see Nolan 1997) but I am inclined to agree that more facts is not, in and of itself, much of a problem and where I appeal to Ockham’s razor myself it will be to the more standard qualitative version. (^23) Platts 1980, p. 72. (^24) Olson 2014, p. 87.