From Scepticism to Anti‐Realism

http://www.diva-portal.org Preprint This is the submitted version of a paper published in Dialectica. Citation for the original published paper (version of record): Tersman, F. (2019) From Scepticism to Anti-Realism Dialectica https://doi.org/10.1111/1746-8361.12276 Access to the published version may require subscription. N.B. When citing this work, cite the original published paper. Permanent link to this version: http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-388933 This is a preliminary version of the article. It will be published in final form in dialectica (DOI: 10.1111/1746-8361.12276). From Scepticism to Anti-Realism by TERSMAN, Folke Uppsala University (P.O. Box 627, 751 26 Uppsala, SWEDEN) Email: [email protected] ABSTRACT A common anti-realist strategy is to argue that moral realism (or at least the non-naturalist form of it) should be abandoned because it cannot adequately make room for moral knowledge and justified moral belief, for example in view of an evolutionary account of the origins of moral beliefs or of the existence of radical moral disagreement. Why is that (alleged) fact supposed to undermine realism? I examine and discuss three possible answers to this question. According to the answer that I think holds most promise, it undermines realism because it renders realism “epistemically incoherent” (in a sense explicated in the paper), and a central aim of the paper is to elaborate and defend that suggestion against certain objections. I end by briefly commenting on the more general significance of the discussion, by considering some other areas (epistemology and vagueness) where similar questions might be raised. 1. Introduction 1 Some arguments against moral realism are classified as being ‘epistemological’. Such challenges are aimed at showing that realism should be abandoned because it cannot account for the existence of moral knowledge. Evolutionary debunking arguments are often said to belong to this category. Darwinian accounts of the origins of our moral beliefs generate the verdict that realists cannot account for moral knowledge, it is held, by committing realists to the conclusion that any correlation that may hold between our moral beliefs and the moral facts is purely coincidental.3 However, an alternative strategy among epistemological challengers is to point to the occurrence of radical disagreement over moral issues, which is taken to illustrate that we cannot gain knowledge of the moral facts that realists posit even when we are in ideal circumstances.4 Yet a further strategy focuses especially on (‘nonnaturalist’) forms of realism that take moral facts and properties to be causally inert. The idea is that this assumption by itself rules out the existence of moral knowledge, independently of complex empirical issues about the causal background of our moral beliefs or about the nature of the existing disagreement.5 What the different epistemological challenges have in common is the shared premise that if it could be shown that realism leads to the sceptical implications the challengers have in mind, then that is a reason to abandon it. The idea is typically not that this would conclusively refute the realist’s position. Yet, it is seen as a significant theoretical cost. Russ ShaferLandau writes as follows apropos the evolutionary debunking strategy: 3 For some prominent examples of this strategy see Street 2006 and Joyce 2006a. 4 See Bennigson 1996, McGrath 2008, Tolhurst 1987, and Wright 1992. See Tersman 2006 for an extensive discussion of this strategy, and Tersman 2017 for the possibility of combining it with the evolutionary debunking one. 5 This strategy is related to the so-called ‘Benacerraf challenge’ against Platonism in mathematics. See Benacerraf 1973. 2 The debunkers claim that if moral realism is true, and if selective pressures have heavily influenced the development of our moral faculties, then we can have no moral knowledge. This by itself does not refute moral realism, but it leaves realists in the deeply unappealing position of being saddled with a thoroughgoing moral scepticism—a logically coherent position that contains about zero appeal. (Shafer-Landau 2012, 1) When confronted with epistemological challenges, realists usually respond by arguing that they can, after all, accommodate the existence of moral knowledge, even in the face of moral disagreement or a Darwinian account of ethics. Alternatively, they argue that competing positions (such as moral constructivism) have similar implications,6 or incur other costs (that are higher). What they typically do not do, as Shafer-Landau’s quote illustrates, is to question the bit according to which the pertinent sceptical implications are a cost. It is that element, however, that provides the focus of this paper. Why would it undermine the realist’s position if it could be shown that it generates the implications in question? Is that question worth asking? It might appear not, as positing objective moral facts may seem utterly pointless if one must concede that we cannot determine which they are. People turn to morality for guidance, and if they hear that, although there are objective truths about how they should behave, it is impossible to find out which truths those are, they are likely to be disappointed. Some may also have been drawn to realism because they hope to secure a sense of meaning and because they fear that a world without objective moral facts is a world where nothing matters. For them too, the finding that moral facts, if they exist, transcend our cognitive abilities is likely to be discouraging. What I want to do in this paper, however, is to explore the possibility of saying something more in support of the idea that the plausibility of realism depends on whether it has the sceptical implications the epistemological challengers have in mind. That is a worthwhile project quite independent of whether realists agree with 6 See Tropman 2014 for an argument to that effect. 3 their critics about that element of the challenges, or so I think. One reason is that it may help us to determine what the epistemological challenges can achieve in other contexts. After all, there are intense meta-debates also about epistemology and aesthetics, debates in which realists, expressivists, and so on quarrel about similar issues. The plan of the paper is as follows. Whether the alleged fact that realism generates sceptical conclusions is a reason to abandon realism is likely to depend on the nature, scope and strength of those conclusions. In the next section, I make some distinctions about the types of scepticism the challengers have in mind and address some other preliminary questions. In Section 3, I identify three distinct answers to the question of why realism would be undermined if it could be shown to lead to scepticism. According to the third of those answers, it does so through rendering the target position “epistemically incoherent” (in a sense to be explicated). This is the answer I find most persuasive, and I defend it, and elaborate it further, in Section 4. In the final section, 5, I make some comments about the relevance of the discussion to other areas besides ethics. 2. Realism and Scepticism Moral realism entails that to judge that an action is morally right or wrong, or that a state of affairs is good or just and so on, is to ascribe a real property and thus to have a true or false belief about it. It also entails that some such ascriptions of specific moral properties to specific objects (some ‘positive’ moral claims) are in fact true, and moreover that their truth is independent of us in relevant ways. Different realists have different views about the nature of the properties our moral beliefs allegedly ascribe. Naturalists take those properties to be reducible to natural properties whereas non-naturalists do not. Some non-naturalists also insist 4 that they are causally inert.7 Non-naturalism is a more common target of the type of challenge I focus on than naturalism, but Sharon Street, for example, argues that her version of the debunking argument applies to naturalist versions of realism as well (Street, 2006). In what follows, I assume that the above claims exhaust the content of realism. Note in particular that I do not take the position to involve any epistemological elements. That is, one could be a moral realist without committing oneself to the existence or possibility of moral knowledge or to the truth of any particular moral claim. Admittedly, some realists define their position, unlike me, so that it does have epistemological components. For example, Richard Boyd takes realism to imply that our ‘[o]rdinary canons of moral reasoning … constitute … a reliable method for obtaining and improving (approximate) moral knowledge’ (Boyd 1988, 182). On such a view, it is obviously so that we undermine realism by showing that it rules out the existence of moral knowledge. However, what I want to explore is if it is undermined even if no such epistemological component is taken to be an integral part of the position. As for the nature of the epistemological objections, it should be noted that they rely on the idea that the assumptions realists make about the nature of moral facts rule out the existence of moral knowledge even granted that such facts exist and that some of our moral beliefs are in fact true.8 The assumptions are supposed to entail (either by themselves or in conjunction with further claims) that our moral beliefs violate some other condition for knowledge besides truth, namely some condition having to do with the justification or grounds we have for holding those beliefs. For example, Richard Joyce suggests that an evolutionary account of morality generates sceptical implications because it ‘forces the 7 Well-known naturalists include Richard Boyd, David Brink, and Michael Smith. See Brink 1989, Boyd 1988, and Smith 1994. In the non-naturalist camp we find David Enoch 2011 and Russ Shafer-Landau 2003. Unlike Enoch, Shafer-Landau does not take his position to exclude that moral properties have causal powers. 8 By ‘our moral beliefs’, I mean those of our beliefs whose contents consist of some positive moral claim; i.e., a claim that actually ascribe some moral property. 5 recognition that we have no grounds one way or the other for maintaining these beliefs’ (2006b, 135, see also Kahane 2011). However, claims to the effect that our beliefs in an area are unjustified or ungrounded can differ in a number of respects, such as scope. This raises questions about which of those claims that are relevant in the present context. Some evolutionary debunkers target only a subset of our moral beliefs. For example, Peter Singer appeals to an evolutionary account in an attempt to specifically undermine deontological intuitions and beliefs (Singer 2005). In the discussion about moral realism, by contrast, the evolutionary theory is invoked to establish that realism generates a more far-reaching or indeed global form of moral skepticism. The idea is that it implies that none of our moral beliefs constitutes knowledge. Skeptical claims can also differ with respect to their modal strength. Thus, one can insist that we in fact lack moral knowledge and that no moral belief is actually justified. But one can also hold that it is impossible to gain such knowledge, either just for us humans, given the limitations we are subjected to, or, more strongly, that this is impossible somehow in principle. In the case of the epistemological objections that appeal to the evolutionary theory, the skeptical conclusions the challengers try to associate with realism belongs to the weaker end of the modal spectrum. This makes sense as the influence Darwinian processes have had over our moral thinking presumably is a contingent matter. Other epistemological challengers, however, try to associate realism with modally stronger forms of skepticism. For example, Crispin Wright argues that moral realists, in order to accommodate the possibility of moral disagreements that cannot be attributed to inferential error, bias, ignorance of relevant nonmoral facts or some similar cognitive deficiency, must assume that the facts they posit ‘transcend all possibility of human knowledge’ and ‘elude the appreciation even of the most fortunately situated judge’ (Wright 1992, 151 and 158, respectively). 6 3. Scepticism and Anti-Realism I shall now consider three distinct arguments to the effect that the plausibility of realism depends on whether it generates skeptical implications. The first is given by Wright. Wright supposes that we are entitled to assume that the truths in an area are epistemically inaccessible (or ‘transcendent’, as he puts it) only if we have something credible to say about why they are thus inaccessible; i.e., only if we can plausibly explain what it is about the subject matter of the area that ‘makes it so’ that the truths it deals with are transcendent (1992, 152). The problem with moral realism, he thinks, is that although there may be compelling such accounts in the case of some areas no plausible explanation applies to ethics. This is the idea that allows him to infer that realism should be rejected from the claim that realists are committed to conceding that the truths they posit are inaccessible One thing to note about this proposal is that it does not seem to work so well for challenges that merely seek to show that realism leads to modally weaker forms of scepticism. For example, consider the evolutionary debunking challenge which is supposed to show that our moral beliefs are unjustified by appealing to the (alleged) fact that they are shaped by evolutionary processes. Arguably, if that fact entails that the beliefs are unjustified, it also explains why that is so. So, the complaint the debunkers have against realism can hardly be that, by rendering our moral beliefs unjustified, realism posits a fact that is inexplicable. There are also other reasons for thinking that Wright’s proposal fails to capture the concerns of at least some challengers. One apparent option for realists who want to meet Wright’s explanatory requirement is to adopt (non-naturalist) assumptions about the nature of moral facts to the effect that they are not only independent of us in semantical and metaphysical senses, as moral realists in general stress, but also unconnected to us causally. This conception of the moral realm seems to offer a potentially successful explanation of its inaccessibility, given the importance causal processes have for our ability to acquire 7 knowledge. Nevertheless, some challengers will insist that this move aggravates the worries realists face instead of helping them to escape them.9 According to the second argument that I shall consider, realism is undermined by the fact that it has skeptical implications, not because it means that realism posits inexplicable facts, but because there are (other) grounds for denying those implications. This second argument is used in the debate about an epistemological challenge that specifically target forms of realism that imply that moral facts are causally inert. Its source of inspiration is the so-called Benacerraf challenge against Platonism in mathematics. Platonists think that our mathematical beliefs posit entities that are mind-independent and bear no causal or spatio-temporal relations to us. In Hartry Field’s well-known reconstruction of Benacerraf’s challenge, those assumptions lead to sceptical worries because they rule out, even in principle, the possibility of a viable explanation of the reliability of our mathematical beliefs. The phenomenon that is to be explained is the supposed fact that we have managed to arrive at mathematical beliefs that are mostly true, or, as Field puts it, that there is a strong “correlation” between our beliefs in that area and the relevant facts.10 What Platonism is taken to exclude is a credible explanation of this correlation which does not just ascribe it to a sheer fluke or a huge coincidence. Field suggests that, if we find that we are in principle unable to provide such an explanation of the correlation that may hold between our beliefs in an area and the relevant facts, then we have a reason to doubt that there is such a correlation in the first place and to suspend judgment about the truth of the beliefs.11 Suspension of judgment is therefore the advice Platonism generates in the case of our mathematical beliefs. 9 For further discusion of Wright's argument, see Tersman 1998. 10 Field 1989, 26. Note that Field focuses on the beliefs held by mathematicians and not those that are held by people in general (‘us’), but that makes no difference here. I shall continue to write about ‘our’ mathematical beliefs. 11 Thus, Field writes: ‘[A] principled inability to […] explain the reliability of certain beliefs tends to undermine the justification of those beliefs.’ (1996, 377). To illustrate this 8 Why is this conclusion in turn supposed to undermine Platonism? According to a thought that motivates the second argument, it does so because the view that our mathematical beliefs are unjustified is simply so counterintuitive. After all, if there are any claims that deserve our trust and whose truth we may justifiably rely upon then they surely include simple mathematical theorems, or so it may seem. So if Platonism generates the opposite verdict, we should adopt some competing and more permissive conception of the nature of mathematical facts, such as, perhaps, a constructivist one. Platonists are vulnerable to the Benacerraf challenge because they hold that mathematical entities are causally inert. Since some non-naturalists make the same claim about moral facts they are taken to face an analogous challenge. A theorist who has considered this analogous challenge and who interprets it along the just indicated lines is David Enoch. Enoch takes the debate between moral realism and its meta-ethical competitors to be a matter of ‘scoring points in an explanatory game’ (Enoch, 2010, 427-429). To accommodate the existence of moral knowledge and justified moral beliefs is one way of earning such points. A theory may fail to do so and still be superior to its competitors in virtue of scoring other points; i.e., by providing explanations of other relevant considerations that are superior to those given by its competitors. Nevertheless, the failure to leave room for moral knowledge lessens the plausibility of a theory, according to Enoch, and provides at least at prima facie reason to look for some other theory. In Enoch’s view, then, the claim that we have moral knowledge and justified moral beliefs may be seen as bit of evidence that meta-ethical theories are to be tested against. The reasonableness of that view could be questioned on the ground that the thesis that we have such knowledge is contested and can be denied for example on expressivist grounds. What I point, he uses his famous example about a remote village in Nepal. Suppose that we happen to believe various things about this village but find that there is reason to think that there is no possible explanation of the reliability of those beliefs. In this case, it would be wise of us to drop our beliefs (see, e.g., Field 2009, 289). 9 want to stress in the present context, however, is merely that, in so far as the second argument relies on Enoch’s view, it is not available to those who in fact do deny that we have such knowledge. Those challengers therefore need an alternative answer to the question of why realism would be undermined if it could be shown to lead to scepticism. In the search of such an answer, one may try to argue that the (alleged) fact that realism has such implications instead reveals a problematic internal tension in the realists’ position. Moral realism is defined in this paper so that the thesis that we have moral knowledge, or the ability to gain such knowledge, is not an essential part of it. So, by showing that realists are committed to skepticism we have not shown that their position is self-contradictory. However, there is another sense in which a position may be relevantly incoherent, namely through the fact that some of its elements exclude the justification (rather than truth) of its other elements, perhaps by being inconsistent with the grounds that can be, or have been, offered in support of the theory. That moral realism has this feature is the upshot of the third argument. The third argument accords with another way of interpreting Field’s version of the Benacerraf challenge, which is more faithful to his intentions.12 Unlike what the first interpretation suggests, Field does not think that the fact that Platonism entails that we should drop our mathematical beliefs undermines Platonism because those beliefs should not be dropped. On the contrary, he agrees that we should drop them (by ceasing to take their contents to be literally true). That’s the central component of his ‘fictionalist’ view on the nature of mathematics. Field’s point is rather that, by dropping them, we are no longer Platonists. To be a Platonist is to hold that the entities our mathematical beliefs posit are causally inert, mind-independent, and so on. But it also involves believing that such entities exist and, according to Field, to have mathematical beliefs (and thus to be committed to 12 Note that Field is a fictionalist about mathematics and believes that we should drop our mathematical beliefs (by ceasing to take their contents to be literally true). It is therefore uncharitable to attribute to him the view that the fact that Platonism renders our mathematical beliefs is a problem because they aren’t unjustified. 10 positing a correlation between our mathematical beliefs and the facts). Now, Field assumes that unless it is possible to explain the reliability of our beliefs in an area in a plausible way, they are epistemically impermissible. So, if the assumptions Platonists make about the nature of mathematical entities rule that possibility out in the case of our mathematical beliefs, it means that one part of the position entails that another part (and therefore the position as a whole) cannot justifiably be retained.13 This is why the principled inability of Platonists to explain a correlation between the target facts and our beliefs in the area is a reason to abandon Platonism, either by adopting a competing view about the nature of mathematical facts (perhaps a constructivist one) or (as Field recommends) by dropping them. Let us say that, if a position is such that its truth excludes our being justified in accepting it, then it is ‘epistemically incoherent’. As stressed earlier, the fact that a theory has this property does not make it self-contradictory. Nor does it seem to provide any evidence to the effect that it is false. However, providing such evidence is not the only way to undermine a position, as we may do so also by showing that the positive justification we may have for it is insufficient. Surely, if we lack good grounds or evidence in support of a philosophical theory such as moral realism then we should not accept it. Strictly speaking, of course, the fact that a position is inconsistent with the supposition that it is justified does not imply that it is unjustified. There is a non-trivial step from the claim that a position is unjustified by its own lights and the conclusion that it is unjustified tout court. However, if one is forced to concede that a position one accepts is justified only if it is false, then one has surely ended up in an awkward place. 13 To put the idea slightly differently: Platonists ask us to believe in a claim (the claim that posits the correlation) whose truth cannot be explained on Platonist grounds. This is a problem because the claim in question represents a fact that cannot, unlike perhaps some other facts, plausibly be seen as ‘brute’, and is instead such that we are justified in thinking that it obtains only if we are capable of explaning it. (See, e.g., Enoch 2010 for this way of phrasing the idea.) So, the target position (Platonism) implies (in conjunction with reasonable epistemic principles) that it is unwarranted. 11 According to the third argument, then, the (alleged) fact that realism has sceptical implications is a reason to abandon realism because it renders the position epistemically incoherent. In my view, this is the most promising of the proposals surveyed in this section, as it does not, for example, rest on the contentious claim that we have moral knowledge. Nevertheless, it is not obvious that the third argument really is available to a critic of moral realism. This has in part to do with the way in which I have defined realism. According to the reconstruction of Field’s argument given above, it shows that Platonism is epistemically incoherent because Platonists are supposed to posit an actual correlation between the mathematical facts and our mathematical beliefs. In the case of moral realism, however, things are different. For, on the definition of realism that I use, to be a moral realist is simply to hold that there are (independent) moral truths, and one could have that position and still consistently remain agnostic both about the truth and justification of our actual (positive) moral beliefs. Thus, one can be a realist, on the definition in question, without having to posit an actual correlation between our moral beliefs and the facts (or to an optimistic assessment of one’s ability to explain such a correlation). The question is whether it still could be argued that realism is rendered epistemically incoherent by the supposition that it excludes that our moral beliefs are justified.14 That is the topic of the next section. 4. Varieties of Scepticism Moral realism is epistemically incoherent to the extent that it generates the sceptical conclusion that we have no good grounds for belief in the existence of moral facts. However, the thesis that the epistemological challengers primarily seek to establish is that realists are committed to a different sceptical claim, namely the claim that no (positive) moral belief is 14 Of course, if Platonism in mathematics is defined similarly, then the same question can be raised in the context of the Benacerraf-Field challenge. Thanks to an anonymous reviewer for stressing this. 12 justified. Whether the challenges provide the basis for a persuasive incoherence charge thus depends on how the two sceptical claims are related to the other. This observation gives rise to a possible objection to the idea that the epistemological challenges (if successful) undermine realism by rendering it epistemically incoherent. On the objection in question, although the claim that no (positive) moral belief is justified rules out our having reasons for thinking of any particular (positive) moral claim that it is a truth, it does not rule out our having reasons for thinking that there are such truths in the first place. That is, the idea is that one may plausibly retain both of the following claims: 1. We are not justified in believing of any (positive) moral claim that it is a truth. 2. We are justified in believing that there are (positive) moral truths. And the point is that, if one can plausibly retain both 1 and 2, then a realist may concede that the epistemological challenges commit her to the claim that our (positive) moral beliefs are unjustified and still deny that this makes her position incoherent in the relevant sense. Is the objection sound? Let us say that, if we think that some (positive) moral claim is true, such as the claim that it is right to give to charity, then we have a ‘first-order’ moral belief. The belief that there are moral facts, by contrast, represents a ‘second-order’ belief about morality.15 We may accordingly call the claim that we are not justified in believing of any (positive) moral claim that it is true ‘first-order moral scepticism’ and the claim that we have no good grounds for thinking that there are moral truths ‘second-order moral scepticism’. What needs to be determined, to assess the objection, is if one can plausibly combine first-order moral scepticism with the denial of second-order moral scepticism. 15 The distinction is stressed for example in Olson 2011. 13 One response to the objection is to argue that it is trivially so that there is no space for the relevant combination by invoking a wide conception of what counts as a (positive) moral claim. One could insist, for example, that the judgment ‘some things are good or bad or right or wrong or...” (where the “...” covers the remaining syntactically basic moral predicates) counts as such a claim. Since that judgment appears equivalent with the thesis that there are moral facts, the conception in question seems to imply that the denial of second-order moral skepticism is outright inconsistent with first-order moral skepticism.16 In what follows, however, I shall pursue the discussion on the basis of the narrower conception of a (positive) moral claim indicated in section 2 (according to which a judgment counts as such a claim only if it ascribes a specific moral property to a specific object). Given the narrower conception, there is no strict inconsistency between first-order moral scepticism and the denial of second-order moral scepticism. This leaves realists with a room for defending the combination of those view by appealing to the fact that there are other cases where we are justified in believing of a consistent sets of positive claims that it contains truths even if we are not justified in thinking of any of those claims that it is a truth. Take a lottery. We can be warranted in thinking that there is a winning ticket without being able to pick it, as we may have evidence to the effect that there is a winner that is not at the same time evidence to the effect that some particular ticket is the winner (evidence about the intentions of the designer of the lottery, perhaps). Does something similar hold in the moral case? That is, are there compelling arguments for the existence of moral truths that are neutral towards firstorder moral scepticism? If the answer to this question is “yes”, then that would refute the idea that realism is epistemically incoherent just because it yields first-order scepticism. One way to try to determine whether there are arguments for the existence of moral facts of the relevant kind is to examine the arguments to that effect that have actually been offered. 16 I thank an anonymous reviewer for pointing out this possibility. 14 This is the strategy I shall use. Note that an argument must meet two constraints to illustrate how one can combine first-order moral scepticism with a denial of second-order moral scepticism. Besides being consistent with first-order skepticism, it must also be sound. A familiar way to justify controversial ontological commitments in any area is to argue that they are implied by, or indispensable for, the best explanation of some set of observational data (indispensable in the sense that their deletion would make the explanation inferior). If they play such a role, then they and the other assumptions the explanation consists in obtain support from the data in question. This is the upshot of the Quine-inspired argument for mathematical realism that appeal to the observation that mathematics, including references to, say, functions and numbers, play a crucial role in modern physics.17 Whether a compelling ‘indispensability argument’ of this kind could be developed for the existence of objective moral facts is highly controversial. That is the central issue in the debate about Gilbert Harman’s well-known challenge against moral realism (according to which moral facts are never posited by the best explanation of anything observable).18 However, the point I presently want to make is that, even if there is room for such a defence, it does not fit the bill. What we are after is evidence for the existence of moral truths that is not at the same time evidence for the truth of any particular moral claim. That is not what we get from an indispensability argument of the kind Quine has suggested in support of mathematical realism. Our observations are supposed to justify mathematical realism via the fact that particular mathematical theorems play an indispensable role in physics. Clearly, this argument works only to the extent that the theorems themselves obtains support from having that role. Thus, it is not consistent with first-order scepticism about mathematics. Similarly, if claims about, say, what is right and wrong figure in the best explanation of some 17 For a competent defense of the argument, see Colyvan 2001. 18 See Harman 1977 and Sturgeon 1988 for two classic contributions. 15 observations, then those observations provide (indirect) evidence for the existence of moral facts. However, it also provides evidence for the truth of the claims themselves. For example, consider Nicholas Sturgeon’s well-known suggestion that the best explanation of social revolutions may attribute them to injustices in the relevant societies (Sturgeon 1988). Clearly, that reasoning entitles us to conclude that there are moral facts only to the extent that it also entitles us to conclude that those injustices have in fact occurred. Another potential argument for the existence of objective truths in a certain domain (which is a special case of the first) appeals to the fact that, over time, inquirers tend to converge on certain answers to the questions that are addressed in the domain. The best explanation of the tendency towards consensus is, it is held, that the answers on which the inquirers converge are true (or ‘closer’ to the truth). This type of argument has been offered for example in support of scientific realism. (See Devitt 1984 for a discussion.) Whether the second strategy might offer any help to a moral realist is also controversial. Moral realists sometimes appeal to the fact that some ethical issues that once were contested (such as the permissibility of slavery) have now been put to rest, at least in some communities (see, e.g., Smith 1994 for this suggestion). Those examples, however, must be weighed against the occurrence also of widespread disagreement, even among seemingly competent inquirers. Besides, there are alternative, sociological explanations of the convergence that do not lend support to a realist view. Be that as it may, for it seems in any case clear that the strategy is not available to someone who wants to combine first-order scepticism with belief in the existence of moral truths. After all, if the convergence gives us reason to posit moral facts, it also gives us reason to believe in the answers on which inquirers converge. The same holds for the very straightforward way of supporting belief in moral facts that consists in stressing that it just seems so obvious that there are true moral judgments, including the truth that it is wrong to torture people for fun. Of course, realism is not the only 16 meta-ethical position that is consistent with the wrongness of such torture, as it also can be accommodated by relativists, constructivists, and so on. However, when the choice stands between realism and some versions of anti-realism, such as Mackie’s moral error theory, it might be seen as decisive. That does not matter in the present context, however, as the indicated argument is a paradigmatic instance of a defence of moral realism that cannot be reconciled with first-order scepticism. Regardless of whether it provides any support for realism, the justification it provides is owed entirely to the justification we might have in believing that torture for fun is wrong or in embracing other equally obvious claims In my view, the most promising candidate of an exception to the just described pattern is an argument that comes from David Enoch (2011). However, Enoch’s argument appears at best reconcilable with some forms of first-order moral scepticism, and its relevance in the present context can be questioned also on other grounds. Enoch thinks that belief in the existence of moral facts (or normative facts more generally) can be justified with reference to their indispensability for another project than the explanatory one that figures in the Quinean-type argument mentioned above. This is the project we engage in when trying ‘to make the decision it makes most sense for one to make’. Normative facts are indispensable to this project, Enoch stresses, since their non-existence would render the deliberations pointless. By engaging in the pertinent project while denying the existence of moral (or normative) facts we would therefore display a kind of irrationality. Given the crucial role the deliberative project plays in our lives, this gives us a reason, Enoch suggests, to believe in the existence of normative facts (including moral facts). The first thing to note about Enoch’s argument is that its relevance in the present context is dubious. The belief in the existence of normative facts is supposed to be justified through being instrumental to a practial project which is taken to be inevitable in the sense that disengaging from it “is not a rationally acceptable option” (2011, 70). It therefore essentially 17 involves a pragmatic element (2011, 63-65). What matters in the present context, by contrast, is whether there are epistemic reasons for positing moral facts in the form of evidence to the effect that there are such facts. It is not clear, in my view, how the argument is supposed to establish that there are reasons of that kind, even granted that it is successful.19 As for the question of the neutrality of Enoch’s argument relative to first-order moral scepticism, this depends on the strength of the scepticism in question. For example, it is hard to combine it with the form of scepticism that Wright suggests realists are committed to, namely the view that we cannot achieve knowledge of the truths the realist posits. Enoch’s defence assumes that the deliberative project is rendered pointless by the non-existence of normative facts, but the problem is that it seems equally pointless on the assumption that normative facts, if any there are, are unknowable. The non-existence of normative facts makes our practical deliberations pointless by ruling out any viable sense in which we can be said to make progress. Thus, if one concedes that there are no normative facts, one will have a hard time explaining why one way of forming normative judgements is somehow more reasonable than another and why it is important to form such judgements in the first place. However, this argument applies equally well to the thesis that normative facts, if any there are, are unknowable. If they are unknowable, progress seems equally excluded. I therefore conclude that, as far as Wright’s version of the epistemological challenge is concerned, Enoch’s defence does not illustrate that a realist can grant the challenger that she is committed to the relevant type of first-order scepticism and still deny that this makes her position epistemically incoherent. This result squares with Enoch’s own acknowledgement that it is urgent for realists to be able to show that the facts they posit are not thus inaccessible (2009, 22). What about the evolutionary debunking version of the challenge, and the type of scepticism that is relevant in that context? Here, things are less clear-cut. On the evolutionary 19 See McPherson and Plunkett 2015 for a helpful discussion of the nature and plausibility of Enoch’s argument. 18 debunking version, the relevant form of scepticism is weaker, as the idea is just that realists are committed to the conclusion that we actually lack justified moral beliefs and that our moral beliefs are in fact not appropriately sensitive to their truth. Now, presumably, the meaningfulness of the deliberative project requires just that we could determine the relevant normative facts. It may, therefore, seem that Enoch’s defence of moral realism is unthreatened by the type of scepticism evolutionary debunkers try to saddle realism with. This reasoning is not wholly convincing, however. For the meaningfulness of the deliberative project does not seem to require merely that we could in some abstract way make progress in the search for normative truths, in the sense that there is some world in which our moral beliefs are relevantly sensitive to their truth. It also requires that we have some viable idea about how to proceed in the actual world to make such progress. The point is that the type of sceptical conclusion realism generates, according to the evolutionary debunkers, hardly leaves room for such an idea. According to the conclusion debunkers think that realists are committed to, the correspondence that may hold between our moral beliefs or intuitions (if any such correspondence exists) is the result of a sheer fluke. By realizing this we seem forced to realize that we are in a situation similar to people who find themselves in some desolate landscape and are told that there is no reason whatsoever to trust the only map that is available to them. That they could in some sense have had another and more reliable map is little comfort. What we thus must do is to ponder the prospects of constructing a new such map entirely from scratch. In the ethical case, this would amount to searching for the truth about moral issues without consulting our moral sensibilities; that is, by jumping our Neurathian ship in search for a new one. The unpromising prospects of that project cast doubt on the idea that Enoch’s defence works under the assumption that moral realism has the sceptical implications evolutionary debunkers try to saddle it with. 19 5. Concluding Remarks Why assume, as epistemological challengers do, that the (alleged) fact that moral realism leads to first-order moral scepticism undermines realism? According to the main proposal in this paper, it undermines realism through making it epistemically incoherent. One could question this proposal by providing grounds for belief in the existence of moral facts that are neutral relative to first-order scepticism. In section 4, I argued that the prospects of that strategy are bleak, by showing that some potential arguments for realism are not thus neutral. Admittedly, however, the list of arguments that I considered is incomplete. For example, reconsider the strategy of defending ontological commitments by showing that they play an indispensable explanatory role. I argued in section 4 that such arguments do not have the required neutrality if they proceed via the idea that it is the truth of particular claims in the domain that does the explanatory work (the Quinean argument for mathematical realism is a case in point). However, one could try to develop an indispensability argument for moral realism that does not rely on that idea. That is, one could argue that the thesis that there are moral facts itself helps to explain certain relevant considerations, such as, perhaps, the seeming objectivity of moral discourse, or the fact that moral terms refer, regardless of whether any particular moral claims play an explanatory role.20 When assessing whether that strategy could be used to generate counter-examples against the paper’s main proposal one should bear in mind (as stressed in section 4) that in order for an argument to constitute such a counter-example it must not only be consistent with firstorder moral scepticism. It must also have some likelihood of success. Its likelihood of success depends in turn on how the phenomena in question are to be characterized. For example, if the 20 I thank an anonymous reviewer for pressing me to consider this strategy. 20 proposed explanandum, such as the (alleged) fact that moral terms refer, is taken to entail that there are objective moral facts then an argument for realism that appeals to its capability of explaining that phenomenon begs the question. However, if the fact to be explained is rather, say, that it seems that moral terms refer or that we behave in ways that suggest that we believe so (for example by producing reasons for our moral rulings, or by using indicative sentences to express them), then we may offer non-realist explanations that potentially are superior to those given by realists (because they are more parsimonious). These are possible grounds for questioning the plausibility of arguments of the pertinent kind. However, a complete assessment of the strategy requires consideration of the details of the arguments it might generate, and my conclusions must therefore remain somewhat tentative. I indicated in the beginning that the discussion in the paper might have a broader relevance. One area besides ethics that it might be interesting to look at is epistemology. Epistemic realism is the thesis that there are (independently existing) facts also about what we are justified, or unjustified, in believing. It is sometimes claimed that the disagreement about such issues is just as radical as that which exists in ethics (see, e.g., Cuneo, 2007). So, an antirealist may hope to show that epistemic realists are committed to similar sceptical conclusions as those to which moral realists allegedly are tied.21 If so, that allows her to argue more straightforwardly that epistemic realism is epistemically incoherent. A moral realist can try to reconcile the assumption that moral realism leads to first-order moral scepticism with the denial of the epistemic incoherence of her position by offering evidence for the existence of moral truths that is not also evidence for the truth of any particular moral claim. There is no such loophole for the epistemic realist. For if epistemic realism entails that we are unjustified 21 One might object (as was pointed out by an anonymous referee) to thinking that this argument is available to an epistemic anti-realist on the ground that she cannot consistently claim that the evidence for the position is insufficient (as she denies the existence of objective epistemic facts). Whether epistemic anti-realism does rule out such claims, however, is a controversial issue. See Olson 2014 and Streumer 2017 for interesting discussions. 21 in believing of any epistemic claim that it is true, then (since the claim that we are unjustified in believing of any epistemic claim that it is true is itself an epistemic claim) it entails that we are not justified in believing in the truth of one of its implications. Another position that may be considered here is epistemicism about vagueness (see, e.g., Williamson, 1994). According to epistemicism, there is a fact of the matter as to whether (i.e., it is true or false that) a vague predicate also applies to a borderline case. It is just that we can never know whether it does. Does the latter component make their theory vulnerable to the objection that it is epistemically incoherent? It does not. If it were to imply that some propositions that ascribe vague predicates to borderline cases are true, then it would be difficult to combine the idea that there is sufficient evidence for it with the claim that all ascriptions of vague predicates to borderline cases are unwarranted. However, it does not in fact imply that, as it merely entails that such propositions are either true or false. Presumably, it is easier to provide evidence for that thesis that is not at the same time sufficient evidence for the ascription of a vague predicate to some borderline case than for the thesis that there are true such ascriptions. So, moral realists and epistemicists are not in the same boat.* REFERENCES BENACERRAF, P. 1973, “Mathematical Truth”, Journal of Philosophy 70, 661-79. BENNIGSON, T. 1996, “Irresolvable Disagreement and the Case Against Moral Realism”, Southern Journal of Philosophy, 411–37. BOYD, R. 1988, “How to be a Moral Realist”. In G. Sayre-McCord (Ed.), Essays on Moral Realism. Ithaca: Cornell University Press, 181–228. * I wish to thank Don Loeb, Olle Risberg, and the anonymous reviewers for dialectica who were involved in the review process and whose comments have significantly improved the paper. 22 BRINK, D.O. 1989, Moral Realism and the Foundations of Ethics. Cambridge: Cambridge University Press. BROSNAN, K. 2011, “Do the Evolutionary Origins of Our Moral Beliefs Undermine Moral Knowledge?”, Biology and Philosophy 26, 51-64. COLYVAN, M. 2001, The Indispensability of Mathematics, Oxford: Oxford University Press. CROW, D. 2015, “Causal Impotence and Evolutionary Influence: Epistemological Challenges for Non-Naturalism”, Ethical Theory & Moral Practice 19 (2), 379-395. CUNEO, T. 2007, The Normative Web, Oxford: Oxford University Press. CUNEO, T. & SHAFER-LANDAU, R. 2014, “The Moral Fixed Points: New Directions for Moral Nonnaturalism” Philosophical Studies, 171, 399-443. DEVITT, M. 1984, Realism and Truth, Princeton, NJ: Princeton University Press. DUMMETT, M. 1978, Truth and Other Enigmas, London: Duckworth. ENOCH, D. 2010, “The Epistemological Challenge to Metanormative Realism: How Best to Understand It, and How to Cope with It”, Philosophical Studies, 148, 413–438. ENOCH, D. 2011, Taking Morality Seriously: A Defense of Robust Realism. Oxford: Oxford University Press. FIELD, H. 1989, Realism, Mathematics, and Modality. Oxford: Blackwell. FIELD, H. 2005, “Recent Debates About the A Priori”, in T. Gendler and J. Hawthorne (Eds.), Oxford Studies in Epistemology. Oxford: Oxford University Press, 69–88. FIELD, H. 2009, “Epistemology Without Metaphysics”, Philosophical Studies 143, 249-290. HARMAN, G. 1977, The Nature of Morality, New York: Oxford University Press. HUEMER, M. 2007, “Compassionate Phenomenal Conservatism”, Philosophy and Phenomenological Research 74, 30-55. JOYCE, R. 2006a, The Evolution of Morality, Cambridge, Mass.: MIT Press. Joyce, R. 2006b, “Metaethics and the Empirical Sciences”, Philosophical Explorations 9, 23 133- 48. JOYCE, R. 2013, “The Evolutionary Debunking of Morality”, in J. Feinberg and R. ShaferLandau (Eds.), Reason and Responsibility, 15th edition, Cengage. JOYCE, R. 2016, “Confessions of a Modest Debunker”, in U. Leibowitz, and N. Sinclair (Eds.), Explanation in Mathematics and Ethics, Oxford: Oxford University Press. KAHANE G. 2011, “Evolutionary Debunking Arguments”, Noûs 45, 103–25. MCGRATH, S. 2008, “Moral Disagreement and Moral Expertise”. In R. Shafer-Landau (ed.), Oxford Studies in Metaethics: Volume 4. Oxford: Oxford University Press, 87-108. MCPHERSON, T. & PLUNKETT, D. 2015, “Deliberative Indispensability and Epistemic Justification”. In R. Shafer-Landau (ed.), Oxford Studies in Metaethics: Volume 10. Oxford: Oxford University Press, 104-33. OLSON, J. 2011, “In Defense of Moral Error Theory”, in M. Brady (ed.), New Waves in Philosophy, New York: Palgrave Macmillan, 62-84. OLSON J. 2014, Moral Error Theory: History, Critique, Defence, Oxford: Oxford University Press. RUSE, M. 1986, Taking Darwin Seriously, Oxford: Blackwell. SHAFER-LANDAU, R. 2003, Moral Realism: A Defence. Oxford: Oxford University Press. SHAFER-LANDAU, R. 2012, “Evolutionary Debunking, Moral Realism and Moral Knowledge”, Journal of Ethics & Social Philosophy 7, 1-37. SINGER, P. 2005, “Ethics and Intuitions”, The Journal of Ethics 9, 331–352. SMITH, M. 1994, The Moral Problem. Oxford: Blackwell. STREET, S. 2006, “A Darwinian Dilemma for Realist Theories of Value”, Philosophical Studies 127, 109-166. STREUMER, B. 2017, Unbelievable Errors, Oxford: Oxford University Press. 24 STURGEON, N. 1988, “Moral Explanations”, in G. Sayre-McCord (Ed.), Essays on Moral Realism. Ithaca: Cornell University Press, 229-55. TERSMAN, F. 1998, “Crispin Wright on Moral Disagreement”, Philosophical Quarterly 48, 359-65. TERSMAN, F. 2006, Moral Disagreement, New York: Cambridge University Press. TERSMAN, F. 2017, “Debunking and Disagreement”, Noûs 51, 754-774. TOLHURST, W. 1987, “The Argument from Moral Disagreement”, Ethics 97, 610–21. TROPMAN, E. 2014, “Evolutionary Debunking Arguments: Moral Realism, Constructivism, and Explaining Moral Knowledge”, Philosophical Explorations 17, 126-40. WILLIAMSON, T. 1994, Vagueness, London: Routledge. WRIGHT, C. 1992, Truth and Objectivity, Cambridge, Mass.: Harvard University Press. 25