1. Introduction
Ethical realists hold that our ethical concepts, thoughts and claims are in the business of representing ethical reality, by representing evaluative or normative properties and facts as aspects of reality; and they hold that such representations are at least sometimes accurate. To judge, for example, that intense suffering is bad, in a way that provides impartial reasons to mitigate it, is to represent the evaluative and normative status of such suffering as part of how the world is. Or again, in correctly judging that it is morally wrong to undermine health care access for the poor and sick in order to provide tax cuts for the wealthy, we are representing the evaluative fact that such behavior possesses the evaluative property of wrongness. Some ethical realists add the further claim that ethical properties and facts are ultimately non-natural. As I am understanding ethical realism, such non-naturalist realists maintain that these non-natural properties and facts are nonetheless worldly, i.e. they are as much features of the world or aspects of reality as natural properties and facts.Footnote 1
My aim here is threefold: to elucidate the sort of representation that is involved in ethical evaluation on plausible realist views, to clarify what exactly is represented and how non-naturalism might come into the picture, and to defend worldly non-naturalism against some objections. The first main question, taken up in Section 2, is how we should model evaluation on any realist view, which in turn should guide the identification of which properties and facts are credibly regarded as ‘evaluative’ ones – as opposed to others that are merely related in some way to them but are not themselves the evaluative properties and facts whose representation constitutes evaluation. Then the question is how the non-naturalist element enters the picture for those of us who make this move in the case of ethical evaluation. What role do non-natural properties and facts play, and how do they figure into what is represented in ethical evaluation? This issue is taken up in Section 3.
Finally, once that is clear, we will be in a position to answer certain well-known objections to non-naturalist realism from Frank Jackson (Reference Jackson1998) and Allan Gibbard (Reference Gibbard2003, Reference Gibbard, Horgan and Timmons2006), as well as more recent objections from Matt Bedke (Reference Bedke2014) and Dreier (Reference Dreier, Bjornsson, Strandberg, Olinder, Eriksson and Bjorklund2015). I will argue in Section 4 that the objections all mischaracterize the role played by non-natural properties and facts.
In particular, non-naturalism does not claim that the natural properties we cite in support of evaluations ‘don’t really matter’ and that some mysterious, unrelated non-natural properties have to be introduced to do the real justifying work (Gibbard Reference Gibbard2003, Reference Gibbard, Horgan and Timmons2006; Jackson Reference Jackson1998). Nor does non-naturalism construe evaluative judgment as belief about non-natural properties in any sense that would imply that we should be expected to change or abandon our first-order moral beliefs if we were to lose our metaphysical belief in non-natural properties. Neither is there any implication that it should be possible to settle first-order moral questions – perhaps in surprising ways – through some sort of direct metaphysical discovery (independent of ethical methodologies) of certain non-natural properties in the world (Bedke Reference Bedke2014). Finally, non-naturalist realism needn’t create any special mystery about why rationality requires us to be motivated by our normative judgments (Dreier Reference Dreier, Bjornsson, Strandberg, Olinder, Eriksson and Bjorklund2015).
2. Evaluation, representation, and evaluative properties and facts
We may begin with a simple axiom, which I’ll call the Evaluation-Based Condition on identifying evaluative properties:
Evaluation-Based Condition: if there is anything truly worth calling an ‘evaluative property’ then this must be nothing more or less than the very property we are attributing to something in evaluating it (e.g. in judging that it is good or right), where the attribution of this property is what constitutes evaluating the thing in question.
So in seeking to identify evaluative properties in the world, we need to start by thinking about what we are doing in evaluating things. There are various properties relevant to evaluation, such as good-making properties, but if our concern is with the evaluative property of goodness itself, for example, then what we are after is the property whose representation constitutes evaluation, such that by attributing it to something we are thereby evaluating the thing.
Consider a simple, non-ethical example. Sharpness may be a good-making feature of a knife, but simply attributing sharpness to a knife is not by itself evaluating it. Someone could attribute sharpness to a knife while not evaluating it if she doesn’t regard sharpness as good-making here, perhaps because she mistakes the knife for a bookmark, say, or because she is ignorant of a knife's function, or because she fails to understand that sharpness makes things well-disposed toward cutting. We evaluate something when we judge that it is good, which is clearly something more than judging that it is sharp (or attributing any other such property to it). Of course, in the case of a knife we will cite its sharpness as a good-making feature in support of our evaluation of it as a good knife – just as we’ll cite a bookmark's sharpness as a bad-making feature in support of a negative evaluation of it as a bad bookmark. But the evaluation is more than just saying that something is sharp or not, and the property whose attribution constitutes evaluation will therefore be something more than sharpness. Similar points can be made about evaluative facts: anything worth calling an ‘evaluative fact’ must be nothing more or less than the very fact we aim to state in evaluating something, the stating of which constitutes evaluating it. So we need to understand evaluation.
Moving now to ethics, suppose we evaluate a certain activity or experience as being good. Again, we are not here simply judging that it is pleasant, for example, even if that is part of what we’re doing, and even if we think (though I don’t) that all and only pleasant things are good, taking pleasantness to be the sole good-making feature of activities or experiences. To judge that something is pleasant is no more to evaluate it than judging that something is sharp is to evaluate it. Someone could agree that an experience would be pleasant and yet still sensibly wonder whether it would be good to pursue, so that she clearly hasn’t yet settled on an evaluation of the experience. Indeed, she might even deny outright that it would be good, explicitly rejecting the positive evaluation despite judging the experience to be pleasant.
Now this is so far just a point about evaluation and the concept of goodness employed in it. We cannot, with Moore (Reference Moore1903), move directly from this sort of ‘open-question’ point to claims about the nature of the property of goodness. Moore famously relied on faulty assumptions about semantics and metaphysics to draw quick conclusions about the impossibility of goodness being identical to pleasantness, or more generally about the falsity of ethical naturalism.Footnote 2 We know now that despite the distinctness of the concepts of goodness and pleasantness, and of judgments about goodness and judgments about pleasantness, it is still possible, as far as that goes, for goodness just to be pleasantness – just as water can be H2O despite the distinctness of the concepts and thoughts about them. Moorean ‘open questions’ don’t lead straightaway to metaphysical conclusions: the fact that we can grant pleasantness but still wonder about goodness doesn’t show that the properties are in fact distinct.
While that is true, however, none of this lends any plausibility to the thought that we can actually understand evaluation and therefore evaluative properties on the model of reductive definitions in the sciences. Consider, for example, the simplistic proposal that when we evaluate activities or experiences as good we are just attributing pleasantness to them using a different concept, goodness, much as we have all along just been talking about H2O using the concept water, and have now discovered the actual nature of the referent. This has no plausibility because while it is true that one possible outcome of our evaluative activity might be the hedonistic conclusion that all and only pleasant experiences are good and worth pursuing, the evaluative activity itself is plainly not one in which we were just talking about pleasantness all along, attributing it to things using a different concept. That is something people could do, of course, but it has nothing to do with the evaluative concerns we have as agents, where the point is not to talk about which things have or lack a property like pleasantness while using a different concept, but is instead to inquire into what kinds of things – including but not limited to pleasantness – have an evaluative status such that they should weigh positively in our deliberations, having normative significance for us (Darwall Reference Darwall2003).
This is why we can raise evaluative questions about all kinds of things in ethical reflection, including experiences already acknowledged to be pleasant, and it is why our deliberative questions could never be settled for us by some revelation that ‘goodness’ happens just to refer to something like pleasantness (on the water/H2O model). Our evaluative practices reveal that our evaluative words and concepts don’t work that way. If they did, they would be useless to us as rational agents occupying the deliberative perspective and trying to determine whether various pleasant things, for example, are truly worthy of pursuit. As Moore would point out, for example, we are not concerned in that capacity with asking, obtusely, whether this pleasant experience is pleasant, for example, whatever concepts we are using. Our concerns are different.Footnote 3
This brings out a basic condition of adequacy for any account of evaluative properties:
Condition of Adequacy: Any account of evaluative properties must make adequate sense of how attributions of evaluative properties like goodness amount to evaluation, understood in a way that properly captures the role evaluation plays in the practical lives of rational agents.
And this is a rich matter when it comes to ethical evaluation. We face the world from a practical, deliberative standpoint as creatures who must reflectively decide what to do. From this standpoint, we ask questions and seek to arrive at informed judgments about what is worthy of pursuit, promotion, and respect as we make choices and act. These are questions broadly about value or goodness: goods to be sought, good ends to be pursued, and values to be respected in our interactions with the world, whether by promoting certain things or by constraining our actions in appropriate ways. The most important sorts of evaluation we engage in take place in this context, from the perspective of deliberating agents seeking to act well, all things considered.
My claim so far, then, is this: any account of the ethical properties and facts represented in ethical evaluation must make sense of how the representation of those properties and facts constitutes this rich form of evaluative activity, playing this role in practical life. This clearly rules out the simplistic suggestion from before. A property like pleasantness may be a good-making property, but it is not goodness and attributing it to something is not in itself to evaluate the thing. I have argued elsewhere that the same point holds even for more sophisticated views that might seek to identify goodness with pleasantness via functional role accounts, for example, even though their accounts of evaluation are more plausible than that of the simplistic naturalistic view above: they still involve a similar misidentification of evaluative or normative properties (FitzPatrick 2011, 2014). I will not further pursue that point here, however, but will focus instead on the positive implications of the claims made so far, if they are correct.
What sort of property, then, are we representing in genuine evaluation, as by attributing goodness to something, which is distinct from merely attributing features to it that happen to be good-making? In particular, what is goodness as represented in ethical evaluation, keeping in mind the above reflections on the nature and role of ethical evaluation in the practical deliberations of rational agents? My answer is that goodness (in the ethical sphere) is a property that amounts to an evaluative status with normative implications, which is of concern to rational agents engaging in ethical evaluation. I will focus here on the goodness involved in common non-basic claims of goodness, where goodness is attributed to something by virtue of certain good-making properties (e.g. a policy is good by virtue of enhancing fair equality of opportunity, or an experience is good by virtue of being pleasant), as opposed to the basic goodness of good-makers themselves (such as the goodness of pleasure, which is not itself attributed on the basis of further good-making features). The claim, then, with respect to such non-basic goodness, which is our concern in ordinary forms of ethical judgment, is this:
G: Goodness is an evaluative status things can have, by virtue of possessing certain good-making features (i.e. features that make the things that have them count as satisfying relevant evaluative standards), where things that have this evaluative status are worthy of pursuit, promotion, or respect, such that there are genuine reasons for such pursuit, promotion, or respect.Footnote 4
It is this sort of status that is the property whose attribution (or representation as being instantiated) constitutes evaluating something in the ethical domain, which is what makes this property warrant being called an ‘evaluative property’. It is not pleasantness, even if hedonism turns out to be true, but an evaluative status things have by virtue of being pleasant, such that they merit certain responses on our part, for which there are genuine reasons. Even a hedonist, then, should not claim that goodness just is pleasantness.Footnote 5 The claim should be just that pleasure is good and is the only thing that is (intrinsically) good, so that an activity's or experience's pleasantness is what makes it have the normatively significant evaluative status of being good and worthy of pursuit. The hedonist will then still owe an account of this alleged fact about the unconditional and exclusive good-makingness of pleasantness, or equivalently, a defense of the claim that the relevant standards of goodness – the ones of proper concern to us as deliberative beings – are such as to have that implication. And if she is a naturalist, she will have to provide an account of all this while operating entirely within a naturalistic metaphysics, not smuggling in any irreducibly evaluative or normative assumptions or claims.
Similar points apply to evaluations of actions as right or wrong, and parallel temptations toward overly simplistic identifications of properties. Evaluating an act as right is not just a matter of attributing to it a property like being happiness maximizing. That may well be relevant to rightness as a right-making property, and some may even hold that it is the only one. But saying that an act is happiness maximizing is not in itself to evaluate it as right, any more than attributing pleasantness to an experience is to evaluate it as good. To evaluate an act as right is to attribute to it the normatively significant evaluative status an act has when it has sufficient right-making properties. Rightness is then precisely that normatively significant evaluative status, the representation of which (as instantiated by the act) constitutes evaluating the act as being right, where, as before, the evaluation in question is the sort of ethical evaluation we engage in as rational agents occupying the deliberative perspective. That is,
R: Rightness is an evaluative status acts can have, by virtue of possessing certain right-making features, where acts that have this evaluative status are worthy of choice by a rational agent occupying the deliberative standpoint, and worthy of moral approval or admiration by others, such that there are genuine reasons for these things.
Even, then, if consequentialism turns out to be correct, the claim should not be that rightness just is being happiness maximizing, say, but rather just that all and only happiness maximizing acts are right, and that an act's being happiness maximizing is what makes it have the normatively significant evaluative status of being right and worthy of pursuit and admiration. The consequentialist will still then owe an account of this alleged fact about the right-makingness of happiness maximization, or equivalently, a defense of the claim that the relevant standards of rightness of action – the ones of proper concern to us as rational deliberators – are such as to have that implication. And if she is a naturalist, she will have to provide an account of all this while operating entirely within a naturalistic metaphysics. That is the task she faces.
3. Enter non-naturalism
Nothing so far poses any explicit threat to ethical naturalism: the point has been simply to flag certain temptations we should avoid, to indicate the more complex sorts of properties represented in evaluation as such, and to highlight the rather demanding task naturalists therefore face in accounting for those properties naturalistically. This is, however, where non-naturalism comes in, because the reason why some of us are non-naturalists is precisely that we are skeptical that the task faced by the naturalist can be satisfactorily accomplished.
The worry is that the naturalistic constraint will force undesirable compromises in our understanding of the objectivity and normative force of ethical facts and properties. I have argued elsewhere that it is hard to see how we can capture everything we might want in these respects without positing some irreducibly evaluative or normative properties and facts – where these are things that can be captured (even in principle) only through evaluative or normative concepts, due to the evaluative or normative nature of the property or fact itself (FitzPatrick Reference FitzPatrick and Shafer-Landau2008, Reference FitzPatrick and Brady2011, Reference FitzPatrick2014, Reference FitzPatrick and Sinclairforthcoming).Footnote 6 It is not hard, for example, to come up with a naturalistic account of rightness as an evaluative status related to an act's satisfying certain stipulated standards cashed out in a purely naturalistic way. But if this is to capture a notion of rightness that is genuinely objective and has categorical normative force for us as rational agents, then we also need to posit such facts as that the set of standards in question is in fact the appropriate one for rational human beings to build our practical lives around, having genuine authority for us and appropriately settling the normative questions we face from the deliberative perspective. Yet it is hard to see how any such facts, if they exist, could themselves be understood as just more natural ones.Footnote 7
It is not my purpose here to rehearse or defend those arguments. The point is just that if we do have doubts about cashing out these background evaluative or normative claims about standards or good- or right-makingness in a purely naturalistic way, then we have two choices if we wish to maintain a realist approach: either (1) we stick with naturalism and drop the ambition of capturing such facts, giving up on robust objectivity or categoricity and settling for less, or (2) we continue to posit such facts and maintain that they are irreducibly evaluative or normative: they’re grounded ultimately in basic values or reasons (such as human dignity, or the badness of suffering) that cannot be cashed out fully by any appeal solely to non-evaluative or non-normative properties and facts. It is this second move that brings us to non-naturalism, since it is typically thought that the natural world doesn’t contain any (in principle) irreducibly evaluative or normative properties and facts, at least of the robust sort relevant to ethics.Footnote 8 So for those of us who are led to ethical non-naturalism in this way, this is how the non-naturalism comes in: we are skeptical about capturing everything we want without relying on some irreducibly evaluative or normative facts about standards or good- or right-makingness; so we posit such apparently ‘non-natural’ facts and properties at the bottom of all this.
4. Responses to objections
With these clarifications in place, we are now in a position to provide responses to a number of objections.
4.1. Jackson and Gibbard on the superfluousness of non-natural properties and facts
Frank Jackson suggests that non-naturalists have (for some inexplicable reason) decided that the natural features of actions we ordinarily cite as wrong-making features don’t really matter after all and cannot on their own justify avoiding actions that have those features, but need to be supplemented with an additional feature that has the honor of being non-natural: that's what really matters and does the justifying work. He imagines the non-naturalist's thinking going like this:
I see that this action will kill many and save no one, but that's not enough to justify my not doing it. What really matters is that the action has an extra [non-natural] property such that only ethical terms are suited to pick out. (Jackson Reference Jackson1998, 127)
That is, however, not at all what we think. Non-naturalists don’t claim that natural properties like causing suffering and death don’t ‘really matter’ because they’re merely natural, or that they’re ‘not enough to justify’ avoiding such actions, requiring supplementation with additional wrong-making features that are non-natural. Obviously the natural features matter, and of course it would be obtuse to suppose that ethical justification would require adding mysterious non-natural wrong-making features to the list of justifiers – as if the fact that the action gratuitously kills is not enough to make it wrong or to be avoided, and we have to suppose that it also has some obscure non-natural feature (Gibbard calls it ‘exnat’) in order to show that it is to be avoided.Footnote 9
The non-naturalist's claim is instead just that we cannot fully account for the fact that the natural wrong-making features do obviously matter and do justify certain choices without appealing at some point to irreducibly evaluative or normative properties and facts. For example, we might think that we cannot fully account for the wrong-makingness of these natural features, as specified by a certain set of ethical standards S, without appealing to the claim that S is the correct and normatively authoritative set of standards for us to be guided by as rational human beings; and defending that claim may require positing basic facts about the objective dignity of persons or the badness of suffering as grounds for the correctness and authority of S. If those facts in turn seem to be irreducibly evaluative and so seem to resist accommodation within a purely naturalistic metaphysics, then we have thereby been led to non-naturalism. What is non-natural are not extraneous and obscure additional features (like ‘exnat’) added to the list of wrong-makers, but just irreducibly evaluative or normative properties and facts to which we appeal in accounting for the evaluative or normative significance of the familiar natural facts we cite in ordinary ethical justifications. All we deny is that such metafacts about the normative significance of such natural facts (such as their wrong-makingness) are themselves just more natural facts (Dancy Reference Dancy and Copp2006; FitzPatrick Reference FitzPatrick and Shafer-Landau2008, Reference FitzPatrick and Brady2011, Reference FitzPatrick2014); Parfit Reference Parfit2011; Scanlon Reference Scanlon2014, 21, 33).Footnote 10
We are not, then, claiming that some sort of special ethical work is done by trotting out some obscure property and announcing that it is non-natural – as if non-naturalness in itself were some kind of ethical magic that allows some random property to accomplish something that natural properties cannot. The only properties and facts we are talking about in connection with non-naturalness are plainly relevant ethical properties and facts – not some ‘exnat’, but just rightness or goodness, or the badness of suffering or the dignity of persons, or the fact that these things objectively ground certain ethical standards we have reason to live by. And there is no mystery why we should be concerned with these things in our ethical thinking, if they exist. The claim that they are non-natural is not meant to explain why we should be concerned with them, but is instead just a theoretical claim about the metaethical status that these already plainly relevant facts and properties have to possess if our theories are to capture a robust sort of ethical objectivity and normativity: they will have to be non-natural facts and properties because they are irreducibly evaluative or normative.
4.2. Bedke's argument from responses to loss of belief in non-natural properties
If non-naturalist ethical realism is correct then our ethical judgments represent a part of reality, specifically an ethical reality with irreducibly evaluative or normative elements, thus resisting full naturalization. There is a sense, then, in which according to the non-naturalist our ethical judgments are beliefs about non-natural properties: (1) they are beliefs about evaluative or normative properties, such as goodness, and (2) these properties are irreducibly evaluative or normative and therefore non-natural. But Matt Bedke (Reference Bedke2014) has recently argued that ethical beliefs have the wrong sort of ‘epistemic profile’ to be about non-natural properties. He offers two related arguments to this effect, which I will consider in this sub-section and the next. The first has to do with expected reactions to the loss of belief in non-natural properties, and goes as follows.
If ethical judgments are beliefs, as the cognitive non-naturalist holds, then they should have a mind-to-world direction of fit, as with other beliefs. This means that ethical judgments should tend to drop away if one were to come to think that there is conclusive evidence that their contents are false, as with other beliefs. Suppose, then, that Amy happens upon some teenagers tormenting a cat and judges it to be wrong, but then she also comes to accept evidence against the existence of non-natural properties. In doing so, she would be accepting evidence against the existence of non-natural wrongness in particular, and so against the boys’ action's having that property. So if ethical judgments were really beliefs about things like non-natural wrongness, as the cognitive non-naturalist holds, then we should expect Amy at this point to drop her ethical judgment that the action was wrong. Yet she wouldn’t, and shouldn’t. So, the objection goes, cognitive non-naturalism is false: ethical judgments are not beliefs about non-natural properties.
In response, however, note first that the non-naturalist does not claim that ordinary ethical judgments are de dicto beliefs concerning non-natural properties – as if in believing that the boys’ behavior is wrong Amy is believing that some non-natural property is instantiated. Unless Amy is a philosopher, she won’t even have considered such matters. The claim is only that ethical beliefs, such as the belief that an act was wrong, are beliefs about properties, such as wrongness, of which it is true (de re) that they are non-natural. Amy believes that the boys’ behavior was wrong, but she needn’t have any de dicto belief at all about non-natural properties, even if it is in fact true of wrongness (which she does have beliefs about) that it is a non-natural property. Nor need Amy have any metaethical belief to the effect that ethical judgments are beliefs about non-natural properties. So there is no reason why we should expect Amy to drop her moral belief about the tormenting just because she now reads a book on metaphysics and concludes that there are no non-natural properties: she needn’t so far think that has any bearing at all on wrongness. So cognitive non-naturalism is not saddled with any implausible implications about how ordinary people like Amy should be expected to respond morally to accepting that there is evidence against non-natural properties: it makes exactly the right predictions.
In order to raise the sort of worry Bedke has in mind we need to imagine a different sort of case where, for example, Ben is a moral philosopher who subscribes to cognitive non-naturalist realism. Unlike Amy, Ben takes his ethical judgment to be about a non-natural property, since he takes wrongness to be a non-natural property. So now suppose that Ben subsequently has a crisis of ‘metaethical faith’: after reading the same metaphysics book he too concludes that there are after all no non-natural properties. Here, Bedke argues, there would seem to be an expectation that Ben's ethical judgment should drop away. After all, if his ethical belief was explicitly a belief that a non-natural property, wrongness, was instantiated by the boys’ behavior, then in now coming to accept that there is conclusive evidence that there are no non-natural properties, he has likewise come to accept that there is conclusive evidence that the content of his ethical judgment is false. Just as a belief in angels (where this is understood to be a belief about supernatural entities) should drop away once one accepts that there is conclusive evidence that there are no supernatural entities, so too Ben's ethical belief about the wrongness of tormenting the cat should drop away – even while holding fixed all his beliefs about the natural features of the case he had cited as his reasons for making that judgment (Bedke Reference Bedke2014, 190–191). Yet, Bedke argues, this is not what would or should actually happen. The ‘metaphysical discovery’ that there are no non-natural properties would and should make no difference to first-order ethical judgments, even in cases like Ben’s.
Now this last claim is of course correct, but the problem is that nothing forces the non-naturalist realist to endorse such abandonment of first-order ethical judgments if one should come to think that there are no non-natural properties after all. There are in fact two ways of going forward from such a ‘metaethical crisis’. One would indeed be to retain the metaethical belief that ethical judgments are about non-natural properties (cognitive non-naturalism), in which case the loss of belief in non-natural properties would indeed lead one to drop the ethical judgment as well. This is precisely what we would expect for belief in angels, for example. It is a given that angels, if they exist at all, are supernatural beings (otherwise we’re not really talking literally about angels), so beliefs about angels are beliefs about supernatural beings in a sense that involves believing that supernatural beings exist; and if someone who once believed in angels comes to think that there are no supernatural beings, then she will (typically) abandon her belief in angels. Similarly, then, the non-naturalist could take that path with regard to ethical beliefs after ceasing to believe in non-natural properties. But there is no reason to think that he should, and there is in fact a far more plausible alternative available.
This second, better alternative upon losing one's belief in non-natural properties is instead just to drop the earlier metaethical belief that ethical judgments are beliefs about non-natural properties, and look for a better theory while retaining one's first-order ethical judgments. There are several options here. The first, and least revisionary, would be to aim to retain not only our first-order ethical judgments but also the objectivity and categorical normative force we were after, but to seek now to capture all this somehow within a naturalistic framework (in the hope that some of the considerations that originally drove us to non-naturalism were mistaken). One possibility would be to try to develop a more expansive naturalism that might naturalistically accommodate the kinds of irreducibly evaluative or normative properties and facts that seem to be required for realist forms of objectivity and categoricity (FitzPatrick Reference FitzPatrick2016). Another would be to attempt to capture objectivity and categoricity without positing irreducibly evaluative or normative properties and facts after all.
If such moves fail, then a more revisionary option – which still leaves our first-order ethical judgments intact – would be to dial back the initial ambitions concerning objectivity and categoricity. Those of us who are currently non-naturalists must face the fact that at the end of the day we might wind up having to settle for what we can get within a purely naturalistic metaphysics, if objections to non-naturalism ultimately prove decisive. We might then fall back on something along the lines of David Copp's (Reference Copp1995) society-centered, standard-based moral naturalism, for example, as the best we can do in the way of ethical realism.Footnote 11 While this may not give us everything we originally hoped for, it might still be far more attractive than embracing moral nihilism and denying the wrongness of tormenting cats, for example. (Indeed, even if one gave up on ethical realism altogether upon jettisoning belief in non-natural properties, there would still be the option of embracing expressivism, which would allow for retaining even a kind of objectivity and categoricity as well as allowing us to preserve our first-order ethical beliefs.Footnote 12)
The underlying point here is that belief in wrongness is ultimately not like belief in angels: we are not simply changing the subject or ceasing to speak literally if we drop our non-naturalist metaethics and continue to speak of wrongness without positing non-natural properties, as we would be changing the subject or ceasing to speak literally if we continued to speak of angels without positing supernatural entities, as by calling very nice people ‘angels’. Angels are understood by believers to be supernatural entities if they exist at all, so that in believing in angels believers take themselves to be believing in supernatural entities, which explains why losing belief in the supernatural can be expected to result in dropping belief in angels. By contrast, it is not part of ordinary belief about wrongness to take oneself to be believing in a non-natural property, as already noted in connection with Amy. This means that even where someone did take himself to believe in a non-natural property, as Ben did, if he then loses that conviction he still has something to fall back on: he can still have beliefs about wrongness, though stripped of any non-naturalness – a move that is not available to the former believer in angels, who cannot go on (literally) believing in angels that are just stripped of their supernatural nature.
These points are reinforced by recalling the earlier reflections on what ethical belief is. I have emphasized that to have an ethical belief that something is good, say, is not simply to believe that it instantiates some property such as pleasantness. Such a model portrays evaluating something as good as being similar to any other property attribution, such as judging that a piece of metal is magnetic, which in turn encourages the thought that evaluation for the non-naturalist is akin to judging that a crystal has occult healing powers. This in turn encourages Bedke's idea that if the non-naturalist ceases to believe in non-natural properties he should drop his evaluative beliefs, just as one would drop beliefs about the occult healing powers of crystals if one ceased to believe in the existence of occult powers. But my earlier point was precisely that this is an inapt model of evaluation, and so of ethical belief and of the properties represented by it. While it is formally and trivially true (at least on cognitivist assumptions) that evaluation involves attributing a property to something (e.g. attributing goodness to it), this is not like attributing magnetism to a piece of metal, or pleasantness to an experience, or occult healing powers to a crystal. It is instead attributing an evaluative status to something by virtue of its having certain other features (in non-basic cases) together with background facts about how those features make the thing stack up with respect to appropriate evaluative standards; and the evaluative properties represented in evaluation, as instantiated by the evaluated object, are the sorts of properties described in G and R (and their negative correlates, B and W, for badness and wrongness, which we needn’t spell out here).
With this model of evaluation and evaluative properties in place, we can better see how the non-naturalist should respond to a loss of metaphysical belief in non-natural properties. When we evaluate the boys’ cruel behavior as wrong, we are attributing to it a certain normatively significant evaluative status, which it has by virtue of having various features we take to be wrong-making features, i.e. features that make it violate appropriate standards for human action. Now again, some of us are led to non-naturalism by the conviction that we cannot give a fully adequate account of these ethical properties and facts without appealing at some point to irreducibly evaluative or normative properties and facts. We therefore add such elements to the picture, which then counts as non-naturalist. But it is not as if we have thereby switched the objects of our ethical belief to obscure new non-natural properties like ‘exnat’ that have nothing to do with the commonsense understanding of what makes tormenting cats wrong. If that's what we were doing, on the model of attributing new occult healing powers to crystals, then we would indeed have to give up our ethical beliefs if we came to reject the existence of non-natural properties. But we haven’t done anything like that. Everything in the commonsense evaluative picture remains intact: what we are attributing to the act is simply wrongness, and the tormenting is wrong for exactly the reasons common sense tells us: it is a deliberate infliction of pain for amusement. We have just added the claim that at the foundation of the picture lie certain irreducibly evaluative or normative – and therefore non-natural – properties and facts, such as the objective, intrinsic, reason-giving badness of suffering.
If we now come to believe that there are really no non-natural properties and facts in the world, and so no irreducibly evaluative or normative ones, nothing forces us to abandon the entire structure (as if wrongness were just an arbitrary posit, like exnat, proposed as a placeholder for some non-natural property). All that is affected is our understanding of its foundation: we would have to concede that it doesn’t rest on irreducibly evaluative or normative properties and facts after all (unless these turn out to fit into an expansive naturalism after all), but has to be understood in a way that doesn’t rely on such things. For example, perhaps the standards are defensible as the appropriate ones given certain very basic commitments and ends which are at least widely shared, or would be if people underwent improvements in empirical knowledge and deliberated in rich and procedurally rational ways. Again, this may not give us everything we had wanted when we turned to non-naturalism to try to account for more, but if non-naturalism were taken off the table then we would just have to accept that this is the best we can do. Such a view would remain fairly robust and needn’t in any case lead us to dismiss the content of our first-order ethical beliefs as being false. What would lead us to regard the content of our ethical belief as false would be coming to accept that there is conclusive evidence that the features we took to be wrong-making aren’t after all wrong-making, which is to say that we were mistaken about the standards for action and there is, for example, nothing wrong after all with deliberate infliction of suffering for amusement. But dropping the non-naturalist construal of the foundation of the standards doesn’t in any way lead us to such a (morally implausible) conclusion.
Nothing in the non-naturalist's position, therefore, implies any problematic commitment to dropping plausible ethical judgments such as the condemnation of the boys’ behavior, or to embracing moral nihilism, if it turns out that there are no non-natural properties. The claim is just that as long as non-natural properties remain on the table as possibilities, as we currently believe they do, the most attractive account of ethical properties and facts includes some appeal to such non-natural properties in the form of irreducibly evaluative or normative elements of reality. Obviously if some metaphysical argument convinces us that they are no longer on the table, then we will have to reassess our metaethical picture. But there is no expectation that this should also force a reassessment or abandonment of our first-order ethical views.
4.3. Bedke's argument from responses to direct discovery of non-natural properties
The second argument Bedke offers is also meant to show the irrelevance of beliefs about non-natural properties to our first-order ethical beliefs, in support of the conclusion that ethical judgments are not beliefs about non-natural properties. Return to the case of the boys tormenting the cat. Bedke takes the non-naturalist to hold that ‘there is a possible non-natural property that is possibly exhibited by the boys’ actions, and discovery that this property is so exhibited would dispose one to judge that tormenting the cat is right, ’ even ‘while holding fixed one's beliefs about the natural properties of the case and one's non-cognitive attitudes’ (Bedke Reference Bedke2014, 197). Yet, the argument goes, it is absurd to suppose that there is any such non-natural property you could ‘discover’ through ‘direct cognition’ that ‘would convince you that an action that causes intense pain for fun, and toward which you have strong negative conative attitudes, is actually right.’ To put it another way, Bedke asks: ‘Is it because they are causing it pain that the action normatively matters in the way it does, or because there is some non-natural property or relation at play? Surely the former.’ Again, ‘the direct discovery of any non-natural property or relation (were this possible) would be as idle to the action's judged normative status as a man's height is to the judgment that he won the lottery’ (Bedke Reference Bedke2014, 197).
There is, however, a straightforward response to this challenge. First, as we have seen in connection with Jackson's objection, the non-naturalist is not committed to the existence of any obscure non-natural property, such as Gibbard's ‘exnat’, conceived as a feature possessed by an action alongside its various natural features (such as causing pain, deliberately, for amusement), and capable somehow of canceling out their wrong-makingness. That is, we are in no way committed to a picture according to which it could be the case that if some such mysterious exnat were added to the act's list of features, holding the rest fixed, a hideous act would be made right and we could in principle come to see its rightness by peering at the act through ‘magic spectacles that reveal all non-natural properties’ (Bedke Reference Bedke2014, 198) and thus directly cognizing this exnat or the resulting rightness. Non-naturalists posit no such properties and no such intelligible possibility, metaphysical or epistemic. The only non-natural property we posit that is relevant to this case is wrongness itself (and any irreducibly evaluative or normative properties on which wrongness here depends, such as the intrinsic, reason-giving badness of suffering). And as we have seen, evaluative properties such as rightness and wrongness are not directly cognized, on the model of some direct cognition of occult healing powers, but are rather attributed as resultant properties through an act of evaluation that applies relevant standards to an object in light of its natural features.
There is, then, no possible exnat (to be directly cognized, even in principle) that could change the evaluative status of the action, for the latter status depends simply on the natural wrong-making features and their wrong-makingness. And there is no intelligible possibility of a direct cognition of (a surprising) rightness here, as if rightness were just some distinct occult quality that could be arbitrarily thrown on top of any set of natural properties and so could, for all we know, just show up anywhere. What we have is instead an act with an evaluative status of wrongness, by virtue of the wrong-makingness of its natural features (as implied by the appropriate set of standards for human action), together with the claim that a proper and full account of the ethical properties and facts here ultimately bottoms out in certain appeals to irreducibly evaluative or normative properties and facts, which are for that reason (and no other) best thought of as non-natural.
None of this implies or even suggests a commitment to the intelligible possibility that there is some non-natural property that, while holding everything else fixed, ‘could dispose you to think that walking down the street is wrong, or that & slavery is just’ (Bedke Reference Bedke2014, 197). The non-naturalism I have described gives no such potential role to non-natural properties in evaluation, and so does not, as Bedke suggests, imply any implausible ‘power to change our normative beliefs [while] holding fixed our beliefs about natural properties and holding fixed our conative attitudes’ (Bedke Reference Bedke2014, 197). The only intelligible way for us appropriately to come to change our minds about ethical matters is by way of ordinary ethical methodologies, reflecting on the normative significance of relevant natural features of actions and on the related standards of action appropriate to human beings. And the non-naturalist is no more vulnerable to a surprising and implausible change of mind about the wrongness of tormenting cats, via such ethical reflection on appropriate ethical standards, than anyone else is.
4.4. Dreier on non-naturalist realism and motivational requirements
Finally, the above elucidations of non-naturalism and responses to Bedke's objections also provide a straightforward answer to worries raised by Dreier (Reference Dreier, Bjornsson, Strandberg, Olinder, Eriksson and Bjorklund2015) in connection with explaining the irrationality of failing to be motivated by one's own normative judgments. Dreier first of all characterizes non-naturalist realists as holding that ‘ought judgments are simply beliefs about which things have a certain ineffable non-natural property’ (Dreier Reference Dreier, Bjornsson, Strandberg, Olinder, Eriksson and Bjorklund2015, 180). He then imagines beings who ‘are disposed to be motivated by their judgments about which actions and states have some other non-natural property (the shmought property, perhaps).’ What, he wonders, would make it the case that it is rational to be motivated by beliefs about one ineffable non-natural property (the ought property) rather than by beliefs about the other (the shmought property, or exnat, or whatever)? Realists, he claims, have nothing to say about this crucial matter, and so the view is ‘unacceptably mysterian’: for ‘the idea that what we have had in mind all along when we talked about rationality was the disposition to be motivated by which actions we believe to have a certain ineffable property seems pretty strange’ (Dreier Reference Dreier, Bjornsson, Strandberg, Olinder, Eriksson and Bjorklund2015, 180).
Pretty strange indeed, but fortunately not an idea belonging to any sensible non-naturalist realism. First of all, it is worth emphasizing again that non-naturalists do not claim that ordinary ethical judgments are beliefs about non-natural properties in the sense that they are beliefs that certain actions ‘have a certain ineffable property’; nor do we hold that what people have ‘had in mind’ in talk of rationality was something about an ineffable property. In fact, since the whole point of the move to non-naturalism is the insistence that ethical judgments are irreducibly evaluative or normative, non-naturalists are committed to rejecting any reduction of ethical beliefs to de dicto metaphysical beliefs about the instantiation of non-natural properties, such as the reduction of the belief that tormenting cats is wrong to the belief that tormenting cats ‘has a certain ineffable property’ (the latter plainly not being an evaluative or normative judgment). The belief that tormenting cats is wrong is the belief that tormenting cats is wrong, not a belief that some unnamed ineffable property is instantiated (which would be a non-evaluative, metaphysical belief). What the non-naturalist is claiming is just that it is true of wrongness that it is irreducibly evaluative and therefore a non-natural property – though not a particularly ‘ineffable’ one, it's worth noting.
To return to Dreier's objection, then: what the non-naturalist thinks about ought judgments is (1) that they are irreducibly normative judgments (not reducible to metaphysical judgments that some strange properties are instantiated), which play a certain role in practical deliberation about reasons, and (2) that normative truths about reasons or about what ought to be done can be fully accounted for only within a non-naturalistic framework, since the account will ultimately bottom out in some irreducibly evaluative or normative elements. Now one could obviously make the connection to rationality look mysterious by ignoring 1 above and just asking why rationality should be concerned with some non-natural property, X, rather than with another, Y, being told nothing more about the actual natures of X and Y – using a dummy term such as ‘exnat’ or ‘shmought’ merely to indicate some obscure non-natural property. Put that way there is of course nothing to say and it is indeed utterly mysterious why rationality should be concerned with such things. But none of this is a problem for non-naturalists because we do not simply posit generic, ‘ineffable’ non-natural properties and claim that rationality is concerned with them, as if it were the non-naturalness of the properties that does some magical ethical work by itself, regardless of the specific nature of the properties in question.
What rationality is concerned with is essentially evaluative or normative properties such as goodness, rightness, or the property of being a reason for acting, as such, and there is nothing mysterious about this connection when we understand these properties in the context of their role in ethical evaluation as described earlier – rather than just thinking of them as obscure non-natural properties on the model of occult healing powers, which are for some inexplicable reason supposed to be of interest to rational agents. It is no mystery why a rational agent occupying the deliberative perspective, trying to figure out how to live, should be concerned with what there is genuine reason to do (or ought to be done) and with what is good to pursue and what will constitute acting well (or rightly) for her in her circumstances: this is precisely what rational deliberation is about. In claiming that goodness, rightness and reasons are ultimately to be understood as non-natural we are not thereby positing some ‘menagerie of curiosities’ (as Bedke puts it, Reference Bedke2014, 190, 198–99) featuring ‘exnat’ or ‘the shmought property’ – properties it would indeed be bizarre to care about, since they are just meaningless posits divorced from any plausible role in agency and evaluation, thus lacking any intelligible claim to our attention.
Bedke is right, then, that learning that such curious and free-floating properties are instantiated would be as irrelevant ‘to our first-order normative attitudes as learning about spiral nebulae or waterspouts’ (Bedke Reference Bedke2014, 197). But this irrelevance in no way carries over to goodness, rightness, or ‘oughtness’. The non-naturalist's claim about the metaphysical status of these latter properties comes in only after their essential roles in agency and evaluation have been established, simply insofar as it appears that their nature is irreducibly evaluative or normative, making them not purely natural aspects of the world (at least as ‘the natural’ is typically understood). Adding that last claim about irreducibility and therefore non-naturalness does nothing to obscure the relevance of these properties to practical rationality, understood as the virtue pertaining to the sphere of ethical deliberation and evaluation.Footnote 13
Disclosure statement
No potential conflict of interest was reported by the author.
