I hope the name ‘Robust Realism’ does not catch on, it being, in my opinion, an unnecessary addition – and a bland, committee-like one at that – to the plethora of names for the metaethical view previously known as ‘Non-naturalism’, ‘Platonism’ or ‘Intuitionism’. What certainly deserves to be called robust, though, is David Enoch's defence of this view. His book is full of vigorous and original arguments, which is impressive considering how old and crowded this part of the landscape is.
The robustness Enoch has in mind is metaphysical; in contrast with Parfit, Nagel, Scanlon and others, he thinks non-naturalism should be defended as a frankly ontological theory, committed to a major addendum to the naturalistic world-view. The first half of the book presents two positive arguments for Robust Realism, while the second half responds to standard worries people have about this kind of view, concerning supervenience, epistemology, moral disagreement and moral motivation.
Most philosophers agree that explanatory indispensability is a sufficient reason for ontological commitment; we should believe in the existence of the properties that figure in our best explanations of experience. In chapter 3, Enoch argues that the underlying rationale for this transition also applies to deliberative indispensability; to be consistent we should also accept the existence of the properties we need to believe in in order to deliberate about what to do and want.
His argument for this parity claim begins by noting that a number of belief-forming methods, such as perception, memory, modus ponens and Inference to the Best Explanation (IBE), cannot plausibly be epistemically justified in a non-circular way; you cannot defend the reliability of perception, for example, without at some point appealing to premises you have from perception. A common and plausible response to this observation is that we must acknowledge that some basic belief-forming methods are simply default reasonable: reasonable in the absence of a case either for or against them.
But it would surely be unsatisfactory, Enoch continues, to base our epistemology on a brute list of default reasonable belief-forming methods. Perception, memory and IBE are in; divination and Inference to the Worst Explanation are out, but why? Since they are epistemically basic, we will not find any epistemological justification for the favoured methods, but we should still try to find some principled account of default reasonable belief formation that sorts cases in the right way and presents the good cases in some kind of positive light. The account Enoch proposes is that a belief-forming method is default reasonable if it is indispensable to a rationally non-optional project. Perception, memory, IBE and the like are indispensable to the project of explaining experience; this project is not one we can rationally put aside, hence their default reasonableness.
Sliding, it seems to me, a bit from talk of methods to talk of individual beliefs, Enoch concludes that since their indispensability to the rationally non-optional project of explaining experience is what justifies belief in electrons, so indispensability to the rationally non-optional project of deliberating about what to do will justify belief in whatever we need to believe in in order to deliberate. At any rate, he challenges those who would resist this parity claim to present a principled reason to treat explanatory indispensability and deliberative indispensability differently.
One worry about this is that Enoch neglects to consider an obvious alternative to the indispensability-to-a-rationally-non-optional-project account of default epistemic justification. The alternative I have in mind is to say that a creature is default justified in using whatever basic belief-forming methods nature has equipped it with. True, this could let in some unsavoury methods, like wishful thinking or racial stereotyping, but we are only talking about default justification. And, we could say, the default justification of these methods has been defeated long ago by our evolving total theory of the world, which we have arrived at by using all our belief-forming methods, and which implies that these ones are worthless. On this picture, then, it is reasonable to start out, as it were, with some faith in all our native belief-forming methods, but quickly update this faith in the light of our developing science. For example, we should start out with some default faith in memory, but quickly adjust this faith as we learn about how memory works and when it is reliable. Perhaps there are problems with this view, but it is odd that Enoch does not discuss it.
Even if the claim of parity between explanatory and deliberative indispensability were to fail, the next step of Enoch's argument, namely that belief in robustly realist normative properties is indispensable to deliberation, would remain of interest. For if it works, Enoch could still say to his opponents: ‘Either give up deliberation (which you rationally can't), or face up to the fact that you believe in robustly realist normative properties.’ His basic argument that such belief is indispensable for deliberation is that we cannot consistently ask and deliberate about what we should do without believing that the question has an answer, and indeed an answer which is not of our making. You cannot deliberate between doing A and B unless you believe there are (or at least may be) reasons to make one choice over the other. In the believed absence of reasons, you cannot deliberate, you can only pick. The balance of the argument is then to show that the kinds of belief described by competing metaethical views cannot play this normative role. This part of the argument is somewhat frustratingly (as Enoch admits) outsourced to other chapters, and even to other parts of his work not in the book.
One worry about this is whether normative belief of any kind is necessary for deliberation. Cannot the deliberation of a business tycoon, for example, terminate in a conclusion about which course of action will be most profitable, his pecuniary desire working in the background to make this belief action-guiding? Enoch's answer to this objection is unsatisfactory, because he thinks the idea is that deliberation can terminate ‘by reference to a desire’ (p. 75), which commits the thinker to the normative belief that he has reason to pursue his desire. But the more difficult objection is that it seems possible for deliberation to terminate in conclusions about how we will get things we want (de re), without mention of the fact that we want them.
In a footnote, Enoch quotes Nagel: ‘Once I see myself as the subject of certain desires . . . I still have to decide what to do, and that will include deciding what justificatory weight to give to those desires’ (p. 76). But even if Nagel is right, it is not clear that this will help Enoch, because it is not clear that the kind of deliberation Nagel discusses here is rationally non-optional. Distinguish between Everyday Deliberation, which is simply humdrum thinking about what to do (‘Should I keep working or go home?’), and Nagelian Deliberation, which asks deeper questions about what is worth pursuing (‘Should I care more about my children's happiness than my own?’). My worry is that Everyday Deliberation may be rationally non-optional but possible without normative belief, while Nagelian Deliberation may require normative belief but be rationally optional.
Enoch's other main argument for Robust Realism is in my opinion one of the most important contributions to the metaethical debate of recent years. In chapter 2 he argues that subjectivist, response-dependence and expressivist theories have implausible implications about how to deal with moral disagreement. His first premise is a moral principle, impartiality, which says that in cases of disagreement about what to do together that flows from the parties having different preferences, it is morally required (of each) to try to find some kind of compromise that gives equal weight to the preferences of each, because that reflects the fact that they, the people involved, have equal moral status.
Suppose, for example, that A and B have agreed to spend the afternoon together, and A wants to see a movie, but B wants to play tennis. There is not time for both. In such a case, impartiality implies that they should seek some kind of arrangement that reflects the fact that the preferences of each are equally important. They could flip a coin, or agree to take turns deciding what to do, or some such. But it would be morally wrong of B, say, simply to insist on playing tennis, without making any kind of concession to A to show that B recognizes A's preferences as equally important.
The second premise is also a moral one, namely that in cases of disagreement about what to do together that flow from moral disagreement, the party who is right, and knows it, on the moral issue, is not morally required to accommodate the other's moral belief in a parallel way. We can call this premise ‘no spinelessness’. For example, suppose A and B are to make a joint purchase, option 1 being in other respects superior but involving unnecessary pain to an animal, option 2 involving no such pain. A wants to go with option 1 because she erroneously (we stipulate) thinks animal pain doesn't matter, while B, knowing that it does matter, wants to go with option 2. Here, the intuition is that it is not wrong of B just to insist on option 2, without making any kind of concession to A's erroneous moral view.
The strategy of Enoch's argument is to squeeze theories that analyse moral judgements in terms of preferences (or moral facts as projections of preferences) between these two moral premises. For if moral judgements (facts) are (projections of) preferences of some kind, then by impartiality, we would be morally required to lend equal weight to other people's erroneous moral views, when deciding joint courses of action or designing public policies. But, by no spinelessness, we are not so required. Therefore, moral judgements are not preferences, and moral facts are not projections of preferences.
One worry I have about this argument is the way in which it sets considerations from different areas of inquiry against each other. Enoch thinks of it as an argument that attacks a family of metaethical views on moral grounds. This leads to a lengthy discussion of whether metaethics is ‘morally neutral’. But it seems to me that, in thinking about what is happening in this argument, thinking of subjectivism and expressivism under the rubric ‘metaethical theories’ obscures the real issue. On their face, subjectivism and expressivism are psychological and semantical claims: they say what kind of mental state moral judgements are, and what kind of meaning moral language has. And Enoch usually describes response-dependence views likewise, as claims about the psychology and meaning of normative judgments (pp. 29f.). So the odd and intriguing thing about this argument, it seems to me, is that it moves from moral premises to psychological and semantical conclusions. Why is that defensible? Is it really an argument that expressivism, say, is false, or is it an argument that we shouldn't speak the kind of language expressivists describe going forward, regardless of whether we do so now?
However that may be, it seems to me that Enoch at the very least has thrown down a pressing challenge to metaethical views that understand moral facts, in one way or another, as ‘the children of our sentiments’. Moral spinelessness is not a viable option: we cannot lend weight to any old moral view, when deciding joint action or public policy, just because someone holds it. So defenders of such views owe us an explanation of how I can let the children of my sentiments trump the children of your sentiments, without any kind of compromise, without implying that you are less important than me.
Having discussed Enoch's two main arguments at some length, I hope I have managed to convey the great originality and interest of this book. The more defensive second half, responding to standard worries about non-natural realist views, is littered with smaller instances of the same originality. This book is never boring. It is not suitable for undergraduates, both because it presupposes a fair amount of familiarity with the literature, and because Enoch's arguments are highly networked (argumentative debts are constantly taken out and repaid chapters later; views are excluded in chapter 4 by pointing out that they cannot simultaneously survive the arguments of chapters 2 and 3, and so forth). But for people working in metaethics, it is mandatory reading.
