Introduction: Error Theory, What?

Error theories have been proposed and defended in several different areas of philosophy. In addition to ethics, there are error theories about numbers, color, free will, and personal identity, just to mention a few examples. Error theory about some area of thought and discourse, D, is commonly defined as the view that D involves systematically false beliefs and that, as a consequence, all D-judgments, or some significant subset thereof, are false. That is also how we shall understand the term “error theory” in this chapter. (For a discussion of some non-standard versions of error theory, see Olson Olson: 8–11.) We shall thus take moral error theory to be the view that moral thought and discourse involve systematically false beliefs and that, as a consequence, all moral judgments, or some significant subset thereof, are false.

Moral error theories differ in scope. Theories at one end of the spectrum take normative judgments in general—of which moral judgments are a subclass—to be uniformly false, whereas theories at the other end of the spectrum take only a subclass of moral judgments—e.g., those concerning duty and obligation, but not those concerning virtue and vice—to be uniformly false (Anscombe 1958; Williams 1985). Positions between these two extremes are also possible.

In recent debate, there has been a consensus that prominent arguments in favor of moral error theory—such as the argument from queerness, to be discussed in the section “The Argument from Queerness and the Queerness Arguments”—generalize beyond the narrowly moral to the broadly normative domain. Consequently, critics, as well as proponents of the theory, have suggested that a plausible moral error theory must take the form of an error theory about the normative, or more particularly, about the irreducibly normative (Olson 2014; Streumer forthcoming). We shall say more about irreducible normativity in the section “What’s Queer? On Irreducible Normativity.” While the chapter focuses mainly on moral error theory, all the points made and concerns raised apply, mutatis mutandis, to error theories about the broadly normative.

A pertinent question for moral error theorists is what to do with moral thought and discourse. Error theorists who are also abolitionists argue that, since moral judgments are uniformly false, we ought to abolish moral thought and discourse (Garner 2007). An important question here that relates to the aforementioned issue about the scope of moral error theory concerns the meaning of “ought” in the claim that we ought to abolish moral thought and discourse. If moral judgments are uniformly false, it is not true that we ought morally to abolish moral thought and discourse, but it might be true that we ought to do so in some normative, yet non-moral, sense of “ought.” Another possibility is that the “ought” is not to be understood as genuinely normative, but rather as a means–end recommendation about how to achieve certain ends, such as peace and mutual trust in a community. Abolitionists could thus argue that abolishing moral thought and discourse would tend to promote peace and mutual trust better than preserving moral thought and discourse would. But abolitionism is not the only option for moral error theorists. Fictionalists recommend that we take up fictional attitudes to moral thought and discourse (Joyce 2001: chapter 8). Conservationists recommend that we continue to engage in moral thought and use moral talk, although systematically error-ridden (Olson 2014: chapter 9). There are other alternatives as well, but since the moral error theorists’ issue about what to do with moral thought and discourse is logically independent of the truth or falsity of moral error theory we will not consider it further here (see Richard Joyce’s chapter “Fictionalism in Metaethics”).

Moral error theorists typically join forces with non-naturalist realists, against naturalism and non-cognitivism. Against non-cognitivism, they hold that sincere utterances of moral sentences, like “Torture is wrong” and “Happiness is good,” are assertions and not primarily expressions of non-cognitive attitudes. In the remainder of this chapter, we shall say no more in the way of critique of non-cognitivism (see, e.g., Elisabeth Camp’s chapter “Metaethical Expressivism;” Jack Woods’ chapter “The Frege-Geach Problem;” and Matthew S. Bedke’s chapter “Cognitivism and Non-Cognitivism”). Against naturalism, they hold that, according to our ordinary concepts of moral properties and facts, moral properties and facts are irreducibly normative and therefore not reducible to, or wholly constituted by, natural properties and facts. It is initially convenient and useful to view error theory and non-naturalistic realism as jointly opposing naturalism concerning the metaphysical commitments of ordinary moral thought and discourse (Mackie 1977: 31–32). But we shall see in the section “What’s Queer? On Irreducible Normativity” that the issue of naturalism need not be crucial to the debate about moral error theory after all. (On naturalism and non-naturalism, see David Enoch’s chapter “Non-Naturalistic Realism in Metaethics;” Peter Railton’s chapter “Naturalistic Realism in Metaethics;” and Daniel Nolan’s chapter “Methodological Naturalism in Metaethics.”)

In any case, moral error theorists part company with non-naturalist realists in holding that there are no moral facts. Whether moral error theorists hold that there are no moral properties is a slightly more delicate issue. Depending on their general views on the metaphysics of properties, particularly concerning the possibility of uninstantiated properties, moral error theorists may hold either that there (necessarily) are no moral properties, or that there are moral properties that are (necessarily) uninstantiated.

Many different arguments can and have been used to support moral error theory (Joyce Joyce; Olson 2014: 15–17). For example, it is a common view that moral realism requires convergence in moral judgments among fully rational individuals (Smith 1994: chapter 6). If no such convergence is forthcoming, and if ordinary moral thought and talk involve commitment to moral realism, error theory looms. Another possible argument for moral error theory is that moral judgments presuppose a kind of free will that human agents cannot have, in which case moral error theory also looms large.

By far the most discussed argument for moral error theory, however, is the argument from queerness (Mackie 1977: chapter 1). In the next section, we shall look at the structure of the argument, and in the sections that follow, we shall consider in some detail one pertinent version of it. We shall also see what can be said for and against this version. But before we come to that, we shall take notice of a difficulty concerning the exact formulation of moral error theory. (The remainder of this section draws on Olson 2014: 11–15.)

According to the standard version of moral error theory, all moral judgments are false. A moral judgment, let us say, is a judgment that entails that some agent morally ought to do or not to do some action; that there are moral reasons for some agent to do or not to do some action; that some action is morally permissible; that some institution, character trait, or what have you, is morally good or bad; and the like. This raises the question of what to say about the truth-values of negated moral judgments, such as the judgment that torture is not wrong. This latter judgment appears to be a moral judgment, and since it is the negation of a judgment that moral error theory deems false, it appears that the theory should deem the negated judgment true. Yet we know that according to the standard version of moral error theory, moral judgments are uniformly false.

This leads to two worries: First, is moral error theory a coherent theory? Second, can it be maintained that moral error theory lacks moral implications? It is immediately obvious that the standard version of moral error theory has implications for moral theory since it implies that moral judgments are uniformly false. But the worry we shall now address is whether moral error theory has implications that are themselves moral.

Mackie insisted that his error theory is purely a “second-order,” or metaethical, view (Mackie 1977: 15–17; on the distinction between metaethics and ethics, see Mark Schroeder’s chapter “Normative Ethics and Metaethics”). But this can be doubted. If it is false that torture is wrong, it seems to follow, by the law of excluded middle, that it is true that torture is not wrong, which seems to be a first-order moral view. That torture is not morally wrong seems to imply that torture is morally permissible. More generally, then, the apparent upshot is that contrary to Mackie’s contention, moral error theory does have first-order moral implications. And rather vulgar ones at that—if moral error theory is true, any action turns out to be morally permissible!

Things get in one way worse, for it seems that we can also derive an opposite conclusion. According to the standard version of moral error theory, “Torture is permissible” is false. By the law of excluded middle, it follows that torture is not morally permissible, which seems to entail that torture is morally impermissible. More generally, then, the apparent upshot is that any action is morally impermissible! This may not be a vulgar first-order moral implication, but it is surely absurd. It also transpires that moral error theory involves a straightforward logical contradiction since we have derived that it is true that, e.g., torture is morally permissible (since any action is morally permissible) and that it is false that torture is morally permissible (since any action is morally ­impermissible).

One possible way out of the predicament is to restrict the scope of moral error theory and hold that not all moral judgments are false, but only a significant subset thereof, namely positive moral judgments (Sinnott-Armstrong 2006: 34–36). A positive moral judgment is a judgment about what some agent morally ought to do or not to do, what there is moral reason for some agent to do or not to do, and so on and so forth; or what would be morally good or bad, or morally desirable or undesirable, and so on. Importantly, it says nothing about mere permissibility.

Restricting the scope of moral error theory to positive moral judgments in this way saves moral error theory from incoherence and from the absurd implication that anything is morally impermissible. But one may object that it remains a platitude that any action that is not wrong is permissible. In other words, moral error theory would still imply vulgar nihilism about wrongness, according to which anything is permissible. (“Nihilism” can mean different things in different contexts. In this chapter, it simply means the rejection of something, such as wrongness or moral facts more generally.)

A better way out is to deny that the implications from “not wrong” to “permissible” and from “not permissible” to “wrong” are conceptual, and maintain instead that they are instances of a generalized conversational implicature (Grice 1989). To illustrate, “not wrong” conversationally implicates “permissible” because normally when we claim that something is not wrong we speak from within a system of moral norms, or moral standards for short. According to most moral standards, any action that is not wrong according to that standard is permissible according to that standard. General compliance with Gricean maxims that bid us to make our statements relevant and not overly informative ensures that we do not normally state explicitly that we speak from within some moral standard when we claim that something is not wrong (Grice 1989: 26ff.). But the implicature from “not wrong” to “permissible” is cancellable. The error theorist can declare that torture is not wrong and go on to signal that she is not speaking from within a moral standard. She might say something like the following: “Torture is not wrong. But neither is it permissible. There are no moral facts and, consequently, no action has moral status.” This cancels the implicature from “not wrong” to “permissible.” (Analogous reasoning, of course, demonstrates why the error theorist’s claim that torture is not morally permissible does not commit her to the view that torture is morally impermissible and hence morally wrong.) On this view, error theory has neither the vulgar implication that anything is permissible nor the absurd implication that anything is impermissible.

But one might object that a fatal problem remains. The law of excluded middle entails that if “Torture is wrong” is false, then “Torture is not wrong” is true. If the latter sentence expresses a moral judgment, then moral error theory after all has first-order moral implications, i.e., implications that by its own lights are false.

In response, recall that according to the definition above, moral judgments are judgments that entail that some agent morally ought to do or not to do some action; that some action is morally permissible or impermissible; that some institution, character trait, or what have you, is morally good or bad; and so on. Now, according to the view on offer, a negated judgment like the one expressed by the sentence “Torture is not wrong” does not entail that torture is permissible; it merely conversationally implicates that it is, since the implicature from “not wrong” to “permissible” is cancellable. Likewise, the judgment expressed by “Torture is not morally permissible” does not entail that torture is impermissible and hence wrong; it merely conversationally implicates that torture is impermissible and hence wrong. Thus negated atomic judgments involving moral terms are not strictly speaking moral judgments, but some such judgments conversationally implicate moral judgments. Since, on this view, sentences like “Torture is not wrong” express judgments that are true, we cannot derive that their negations (such as “Torture is wrong”) are true. This saves moral error theory from the threat of incoherence and from implausible first-order moral implications, and it enables moral error theorists to maintain that all moral judgments, and not just a subset thereof, are false. Let us now consider arguments in favor of error theory and the alleged queerness of moral facts.

The Argument from Queerness and the Queerness Arguments

The argument from queerness is profitably seen as having a bipartite structure. The first part of the argument seeks to identify ways in which moral facts would be metaphysically queer, and hence to establish a presumption against the exist ence of moral facts. Such arguments we can call “the queerness arguments.” The second part seeks to explain why we tend to think and speak as if there are moral facts, although there are none. Such explanations typically appeal to projectivist accounts of moral judgment and belief, according to which we mistake affective attitudes (such as approval and disapproval) for perceptions of mind-independent moral properties and facts (Hume 1998: Appendix 1; Mackie 1977), and to debunking explanations, according to which moral judgment and belief originate and evolve because of their social and evolutionary advantageousness (Joyce 2006). The point is to explain how ordinary moral judgment and belief are products of processes that do not track moral truth.

The second part of the argument from queerness is needed because the queerness arguments, if successful, establish only a presumption against moral facts. There are some prima facie queer entities, whose queerness does not rule them out of existence. Some examples are dark matter, neutrinos, tardigrades, and Impressionist paintings (Platts 1980: 87). Once inquiry reveals how such entities fit into the natural order of things, we typically no longer view them as queer, or we may continue to view them as queer in the sense of appearing utterly different from most other things we encounter. But inquiry and reflection can help us realize that such entities are actually parts of the best explanations of some of our observations and beliefs, and at that point, they no longer seem ontologically mysterious. By contrast, moral facts do not in this way fit into the natural order of things, and they are not parts of the best explanations of our observations and beliefs. Moral facts are, according to error theory, both metaphysically queer and explanatorily redundant.

If successful, the argument from queerness shows that moral error theory is, in the end, more plausible than realism since it establishes a presumption against moral facts and explains our common-sense moral beliefs in ways that do not require or presuppose that they are or can be true.

The locus classicus of the argument from queerness is a dense passage from Mackie’s Ethics (1977: 38–42). In my view, four distinct queerness arguments can be discerned in Mackie’s discussion. They concern motivation, epistemology, supervenience, and irreducible normativity (Olson 2014: chapters 5–6). According to the queerness argument concerning epistemology, our ways of knowing about moral facts would have to be queer. This argument can be set aside for our purposes since it concerns the attainability of moral knowledge rather than the existence of moral facts. If successful, the argument supports moral skepticism rather than moral error theory (see Matt Lutz and Jacob Ross’ chapter “Moral Skepticism”). One could argue that realism about moral facts requires a plausible epistemology of moral facts, and that failure to provide one would count against realism. But we won’t pursue that line here.

The queerness argument concerning motivation has as its key premise that moral facts are such that necessarily, if one is aware of them (at least by first-hand acquaintance), one is thereby motivated to act in accordance with them. There is now a general consensus that the weakness of this argument is that its key premise is highly dubious. It seems not implausible that the connection between awareness of moral facts and motivation to act is contingent (see David Faraci and Tristram McPherson’s chapter “Ethical Judgment and Motivation”).

The queerness argument concerning supervenience is much more contested. Its starting point is that moral facts depend on natural facts; there can be no change or difference in moral facts without a change or difference in natural facts that somehow explains or determines the difference in moral facts. The challenge for realists is to account for this seemingly necessary dependence relation. Naturalist realists can respond that since moral facts are natural there is no dependence relation between metaphysically distinct kinds of facts to account for (Brink 1984). But the difficulty naturalists face is to give a naturalistic account of moral facts that accommodates their irreducible normativity. We will return to this in the next section. Non-naturalist realists face the difficulty of accounting for the dependence relation in a way that is consistent with the metaphysical distinctness of moral and natural facts (see Pekka Väyrynen’s chapter “The Supervenience Challenge to Non-Naturalism”). Some error theorists who are impressed by this argument argue that there is no such plausible account to offer, and they conclude that there are no moral facts (Streumer Streumer; forthcoming).

One way for non-naturalist realists to respond is to argue that some natural properties have the further property of making actions right, wrong, obligatory, etc. (Olson 2014: 98–99; Wielenberg 2014: 23–24, 35–36). For example, a utilitarian may argue that an action’s natural property of maximizing happiness makes that action right. The property of being a right-making (wrong-making, ought-making, etc.) property is itself an irreducibly normative property and fundamental truths concerning which natural properties make actions right, wrong, obligatory, etc., may well be necessary truths that we can grasp only by a priori reflection or intuition. The notion of right-making (wrong-making, ought-making, etc.) properties, therefore, does not seem to give rise to additional metaphysical or epistemological puzzles for non-naturalistic realism.

Note that what is made right by a right-making property (such as the property of maximizing happiness) is an action. The relation between the fact that an action has certain natural properties and the fact that it has certain moral properties is plausibly seen as an instance of a kind of generic making-relation, which is not itself irreducibly normative but whose instances may obtain between natural and irreducibly normative relata (e.g., between the fact that an action has the natural property of maximizing happiness and the fact that it has the irreducibly normative property of being right).

Were error theorists to object to the idea of a generic making-relation, or to the idea that fundamental moral truths are necessary and a priori knowable, the target of their objections would be very wide and would not obviously succeed in pointing to something uniquely objectionable about moral or normative facts and properties. It seems, then, that the moral error theorist’s objection will have to focus on the metaphysical queerness of irreducible normativity. In order to see whether that objection can be substantiated, we need to get a better understanding of what irreducible normativity is supposed to be.

What’s Queer? On Irreducible Normativity

It is not very easy to say what irreducible normativity is. Several contemporary non-­naturalist realists take the notion of an irreducibly normative reason to be primitive (Parfit 2011a: 31; Scanlon 2014: 2). In order to illuminate irreducible normativity, it is therefore useful to say what it is not and to contrast it with adjacent but different notions. Following Derek Parfit, we can contrast normativity in the “rule-implying” sense with normativity in the “reason-implying” sense (2011b: 308–10; 326–27). Examples of normative facts of the former kind are facts about what is legal or illegal, grammatical or ungrammatical; and about what accords with rules of etiquette or chess. There is no metaphysical mystery how there can be such facts, for facts about the law and grammar, and about rules of etiquette or chess, are all facts about human conventions. It might, of course, be difficult to say exactly how and why certain conventions originate and evolve, but such difficulties invite no metaphysical mysteries (Mackie 1977: 25–27; Joyce 2001: 34–37; Olson 2014: 118–26). Moreover, for any fact that is normative in the rule-implying sense, e.g., that it is a rule of etiquette that one does not eat peas with a spoon, we can always ask whether we have reason to—or whether we ought to, or are required to—comply with that rule. Those are instances of what Christine Korsgaard and others have called “the normative question” (Korsgaard 1996; Broome 2007).

Facts that are normative in the reason-implying sense are irreducibly normative, and they are very different. Such facts do not reduce to—and are not wholly constituted by—facts about human conventions or about agents’ motivational states or desires. In the words of the eighteenth-century moral rationalist Richard Price, irreducibly normative facts “have a real obligatory power antecedently to all positive laws, and independently of all will” (Price 1948: 105).

Moral facts are prime examples of irreducibly normative facts, but there may be others. This connects to the question concerning the scope of error theory, which was mentioned in the section “Introduction: Error Theory, What?”. It has been argued, for example, that epistemic facts are irreducibly normative (Cuneo 2007; Bedke Bedke; Olson 2014: chapter 8). But here we shall concentrate on moral facts. Many philosophers hold that according to the ordinary concept of moral facts, they are irreducibly normative, and that ordinary moral judgments are judgments about such facts. (This is of course not to say that ordinary speakers articulate the thought that moral facts are irreducibly normative.) Therefore, moral judgments do not invite the normative question. If we know or believe, e.g., that we ought morally to eat less meat, it makes no sense to ask whether we have irreducibly normative reason to eat less meat; that question will already have been answered. (We could, of course, ask different questions, such as whether eating less meat would be comme-il-faut, or whether there is anything that motivates us to eat less meat, or whether doing so would be conducive to fulfillment of our desires.) This explains why it would be odd, and perhaps conceptually confused, to accept that one ought morally to eat less meat and at the same time hold that one has no reason to do so because social conventions do not recommend or require that one eat less meat, or because one is not motivated to do so, or because one lacks the relevant desire.

Many non-naturalist realists and error theorists maintain that it is impossible to give a plausible naturalistic account of moral facts, precisely because they are irreducibly normative; moral naturalism therefore falls prey to the “normativity objection” (Parfit 2011b: 324–27; Olson 2014: 82–83). But moral naturalism is a broad church and it is perhaps premature to conclude that there can be no plausible naturalistic account of irreducible normativity. If there is such an account, moral naturalists too can agree that ordinary moral thought and talk involve commitment to irreducibly normative facts without thereby yielding to error theory. Note also that it is not uncontroversial that ordinary moral thought and talk do involve such commitment (Foot 1972; Finlay 2008). This indicates, as was suggested in the section “Introduction: Error Theory, What?”, that the divide between naturalism and non-naturalism is not central to the debate on moral error theory after all. Instead, the crucial lines of division are between those who think that ordinary moral thought and discourse involve commitment to irreducibly normative facts (a group which includes non-naturalists, error theorists, and possibly some naturalists) and those who think not, and between those who think that there are irreducibly normative facts and those who think not.

For reasons of space, I shall not say more than I already have said above in defense of the view that ordinary moral thought and discourse involve commitment to irreducibly normative facts (see Olson 2014: 126–35). The pertinent question to consider now is the second of the two mentioned in the previous paragraph, i.e., whether there are any irreducibly normative facts. Here is where error theorists and other critics of irreducible normativity see a mystery: How can there be facts that have a real obligatory power, prior to human conventions and independently of agents’ motivational states or desires? As was just noted, it seems no metaphysical mystery how there can be facts that require or favor certain courses of behavior, where these requiring or favoring relations reduce to facts about conventions or desires. But how can there be facts that require or favor certain courses of behavior, where the relation of requiring or favoring cannot be explained or demystified in terms of facts about human conventions and desires? Irreducibly normative facts seem in this way to involve “a peculiar combination of objectivity and prescriptivity” (Garner 2006: 101; see also Mackie 1977: 40–42; Olson 2014: 116–26). In comparison with facts that are normative in the rule-implying sense, such facts seem queer. Their queerness establishes a presumption against the view that there are irreducibly normative facts and in favor of views according to which there are none, such as the error theorist’s view. To committed realists, this “sheer queerness worry” (Enoch 2011: 134–36) may appear unsophisticated. But it seems not very different from the grounds on which philosophers have worried about or objected to other theories.

Consider, for example, the reaction that Bertrand Russell reported that he once had to Leibniz’s theory of monads, the monadology. It seemed to him “a kind of fantastic fairy tale, coherent perhaps, but wholly arbitrary” (Russell 1900: xiii). The error theorist’s response to the realist’s ontology of irreducibly normative facts is similar to Russell’s reaction to the monadology. Error theorists may hold that the realist’s theory of irreducibly normative facts is “fantastic” in the sense of unrealistic, odd, or incredible. They may also hold that while the theory is coherent, it is arbitrary, since moral facts are explanatorily superfluous; in particular, moral facts are not required as parts of the explanations of why we make moral judgments and form moral beliefs. We will return to this point in the next section.

Responses to the Argument from Queerness

According to the queerness argument—the first part of the bipartite argument from ­queerness—that we have explored, moral facts are irreducibly normative, and the ­metaphysical queerness of irreducibly normative f acts establishes a presumption against them. There are various ways for moral realists to respond to the argument. As mentioned in the previous section, one line of response rejects the premise that ordinary moral thought and discourse involve commitment to irreducible normativity (Finlay 2008). As we also mentioned in the previous section, however, many realists worry that any account that denies that ordinary moral thought and discourse carry such commitment is thereby vulnerable to the normativity objection. A stubborn but natural response from realists who maintain that ordinary moral thought and discourse involve commitment to irreducible normativity is simply to deny that irreducibly normative facts are queer. Another natural but less stubborn response is to grant that irreducibly normative facts are queer, and that their queerness establishes a presumption against them, but maintain that the presumption can be conclusively rebutted.

As a simple illustration of how the less stubborn response may be developed, consider the proposition that torturing children for fun is morally wrong. Most people believe that this proposition is true, and many are strongly convinced that it is true. Now consider the proposition that no action is morally wrong, which is what moral error theory implies. These two propositions are obviously mutually inconsistent. Presumably, most people who are not already committed moral nihilists are much more confident that the first is true than that the second is true. Presumably, then, for most people, it seems at least initially much more plausible that moral error theory is false than that it is true; insofar as one is confident that, e.g., torturing children for fun is morally wrong, it is rational to reject the error theorist’s claim that moral facts are queer and that no action is morally wrong (Huemer 2005: 116–17; Enoch 2011: 118–21). From this point of view, it would seem more queer or odd for it to be the case that torturing children for fun is not morally wrong than for it to be the case that there are irreducibly normative facts.

In order to respond to this attempted rebuttal, the moral error theorist needs to invoke the second part of the argument from queerness. As noted in the section “The Argument from Queerness and the Queerness Arguments”, the point of the second part is to explain why we make moral judgments and hold moral beliefs, sometimes with great confidence, although they are uniformly false. As indicated, such an explanation typically involves a projectivist account of moral judgment and an account of the social and evolutionary advantageousness of moral judgment. Here, we shall only consider briefly what such accounts may look like (Olson 2014: 141–18).

According to one popular view, the psychological origins of moral norms are our emotions or affective attitudes (Nichols 2004). For example, norms against harming others originate in the intense distress most people feel when witnessing others suffer. Such feelings of distress explain, or at least partly explain, why people are typically strongly motivated to enforce and comply with norms against harming innocents, such as children and animals. It may be that we systematically mistake certain affective responses, such as the distress we typically feel when witnessing others suffer, for perceptions of mind-independent, irreducibly normative, properties, e.g., moral wrongness. In this way, we may come to experience moral wrongness as mind-independent and as independent of human conventions, and come to believe that, e.g., torturing children for fun, has the irreducibly normative property of being wrong. To use Hume’s famous metaphor, the idea is that emotions and affective responses are “gilding or staining all natural objects with the colours, borrowed from internal sentiment [which] raises, in a manner, a new creation” (1998: 163).

Moral error theorists like J. L. Mackie and Richard Joyce have argued that this ­projective error is beneficial rather than pernicious in that it enables moral thought and talk to function well as devices for solving inter- and intrapersonal coordination problems. The thought is, in brief, that natural selection has tended to favor certain patterns of human behavior, such as reciprocating favors; sticking to agreements; punishing perpetrators; parents looking out for their kin; and so on. These selection processes have played a part in shaping our current systems of norms; they account for why we tend to believe, e.g., that there are reasons to return favors, keep promises, punish perpetrators for their misdeeds, and for parents to look after their kin. Human beings will of course sometimes be tempted to violate some of these norms. Breaking promises and omitting to return favors often make sense from a narrowly egoistic perspective, and punishing perpetrators can be costly for the punisher. Moral thought—along with its projective error and experience of moral properties and facts as irreducibly normative—and moral talk—along with its attributions of such properties and references to such facts—enter the picture as social devices that serve to enforce compliance with these norms. We judge that those who fail to return favors and keep their promises act morally wrongly; they are liable to moral blame, i.e., to attitudes of resentment and dislike.

The projective error is also what makes most people perceive a kind of authority in moral norms, which makes them feel bound to act in accordance with them. In this way, moral thought and talk function both intrapersonally “as a bulwark against weakness of will [and] as an interpersonal commitment device” (Joyce 2006: 208). In short, there is such a thing as moral thought and talk partly because “[w]e need [it] to regulate interpersonal relations, to control some of the ways in which people behave towards one another, often in opposition to contrary inclinations” (Mackie 1977: 43).

It is important to be clear about what projectivist debunking accounts of moral judgments and beliefs—like the one just sketched—achieve and do not achieve, if they are successful. First, they do not establish the ontological thesis that there are no moral facts. If successful, they establish at most that moral facts play no explanatory role in the formation of moral judgments and beliefs, and in moral practice more generally. Second, projectivist debunking accounts of moral judgments and beliefs do not establish the epistemological thesis that we cannot attain knowledge about necessary moral truths solely by a priori reflection or intuition. Even if it is true that ordinary moral judgments and beliefs originate in emotions and social conventions and hence are products of processes that do not track moral truth, it does not follow that we cannot attain knowledge a priori about necessary moral truths. It does not even follow that most of our current moral judgments and beliefs are erroneous. It might be that many moral judgments and beliefs that, according to realists, are true (e.g., that happiness is good; that justice should be honored; that cheaters should be punished; etc.) have been evolutionarily advantageous. If so, ordinary moral judgments and beliefs may correlate with moral truths, even though they are products of processes that do not themselves track moral truths. That could explain how it can be that many ordinary and widely held moral judgments and beliefs are true, even if projectivist debunking accounts of them are successful (Enoch 2011: 167–76; Olson 2014: 146; Wielenberg 2014: chapter 4).

It is thus a mistake to infer a nihilist view like error theory directly from the premise that projectivist debunking accounts of moral judgments and beliefs are successful. Moral realism is compatible with such accounts. This should not be surprising, since projectivist debunking accounts explain the psychological and social origins of moral judgments and beliefs, and as such, they have no direct implications for the ontology of moral properties and facts or the attainability of moral knowledge. The point that moral facts play no explanatory roles in the formation of moral judgment and beliefs is no embarrassment to those realists who take moral properties and facts to be causally inefficacious (e.g., Enoch 2011: 7, 162, 177).

What, then, is the point of projectivist debunking accounts of moral judgment and belief if realists can, after all, accept them? Their dialectical point in the error theorist’s critique of realism is twofold. First, they explain why we tend to make moral judgments and hold moral beliefs even though they are uniformly false. Projectivist debunking accounts thereby offer a response to the realist’s argument that moral error theory can be rationally rejected because we believe some moral propositions with comparatively very high confidence. They also defeat the evidence provided by our great confidence in some moral propositions. For if projectivist debunking accounts of moral judgments and beliefs are successful, error theorists can offer them as explanations of why we tend to believe certain moral propositions with great confidence, and these explanations do not require that the propositions are true.

A point worth making in this context is that the hypothesis that moral judgments and beliefs stem partly from affective attitudes explains why moral error theory is emotionally difficult to accept. When one considers the numerous atrocities committed in the past century, it may feel sickening to maintain that none of them were in fact morally wrong, even to proponents of moral error theory. Error theory about other matters, such as numbers and color, may be intellectually difficult to accept, but they do not face the same kind of emotional resistance. The emotional origin of moral judgments goes some way to explaining why reactions to error theory about morality are sometimes rather fierce, while reactions to error theories in other domains typically are less so (Olson 2014: 143–44).

Secondly and relatedly, projectivist debunking accounts help establish that moral error theory is a more parsimonious theory than realism about irreducible normativity. Error theory involves fewer ontologically fundamental and unexplained facts than does realism, without loss in explanatory power concerning how and why we make moral judgments and form moral beliefs (Olson 2014: 87–88, 147–48). Realists about irreducible normativity are of course likely to object that error theorists buy parsimony at the expense of plausibility, since error theory implies that many propositions that seem highly plausible are not true. But that only takes us back to the first dialectical point of the error theorist’s employment of projectivist debunking accounts of moral judgments and beliefs.

However, while its greater ontological parsimony renders error theory in one respect preferable to realism, it is important to observe the limitations of ontological parsimony. Considerations of ontological parsimony do not constitute evidence of the absence of moral facts. At most, they establish the absence of such evidence. In other words, assuming again that the error theorist’s proffered projectivist debunking account of moral judgments and beliefs is successful, that leaves an absence of evidence that there are moral facts; it does not provide evidence of absence of moral facts. (This point is forcefully made in Morton & Sampson 2014.)

This reinforces the importance of the bipartite structure of the argument from queerness. As I have indicated repeatedly, projectivist debunking accounts of moral judgments and beliefs do not on their own establish a presumption against moral facts. That is why the first part of the argument from queerness, what I have called a queerness argument, is needed. As before, realists can and probably will dispute the queerness of moral facts, and of irreducible normativity in particular. But this will not refute the argument from queerness, for it cannot be an adequacy condition of an argument that it suffices to convince the opponents. Consider again the criticism of Leibniz’s monadology, that it seems a fantastic and wholly arbitrary theory. Imagine that the monadology enthusiast responds to this criticism in a way that parallels the realist’s response to the error theorist’s claim that irreducibly normative facts are queer, i.e., by simply refusing to admit that the monadology is fantastic and wholly arbitrary. Such a response notwithstanding, most philosophers presumably find the monadology fantastic and arbitrary, and the entities it concerns ontologically mysterious. And most philosophers presumably take such worries to constitute a presumption against monads and the truth of the monadology. Error theorists take the same views of realist theories that vindicate irreducibly normative facts, and of the alleged facts these theories are about.

At this point non-naturalist realists may try a somewhat different tack and argue that the claim that irreducibly normative facts are intolerably queer is implicitly or explicitly premised on a prejudicial commitment to a kind of philosophical naturalism, according to which only properties and facts that are in some relevant sense continuous with natural science exist (e.g., Scanlon 2014: 17–18). This kind of philosophical naturalism is not only philosophically contentious; it is also unclear what conclusions to draw from it, since it is notoriously unclear which kinds of properties and facts are in the relevant sense continuous with contemporary natural science, and also what it is for properties and facts to be relevantly continuous with contemporary natural science (see Daniel Nolan’s chapter “Methodological Naturalism in Metaethics”).

While it is true that some philosophers who find moral facts queer are or seem inclined toward some kind of philosophical naturalism (e.g., Mackie 1946: 78–80; cf. Enoch 2011: 135), queerness arguments need not be premised on any kind of philosophical naturalism. Someone who finds irreducible normativity queer need not thereby hold that Cartesian souls, abstract objects, God, or whatever other kind of entities commonly taken to conflict with philosophical naturalism—even Leibnizian monads—are queer. This confirms the suspicion that the divide between naturalism and non-naturalism need not be seen as a crucial line of conflict in the debate on moral error theory. The queerness argument against moral facts that we have explored here is not premised on a prejudicial commitment to a contentious kind of philosophical naturalism, but on the sheer queerness of irreducibly normative facts.

Related Topics

Chapter 1, “Non-Naturalistic Realism in Metaethics;” Chapter 2, “Naturalistic Realism in Metaethics;” Chapter 4, “Fictionalism in Metaethics;” Chapter 5, “Metaethical Expressivism;” Chapter 8, “Realism and Objectivity;” Chapter 9, “Metaphysical Relations in Metaethics;” Chapter 10, “The Supervenience Challenge to Non-Naturalism;” Chapter 14, “The Frege-Geach Problem;” Chapter 18, “Cognitivism and Non-Cognitivism;” Chapter 19, “Ethical Judgment and Motivation;” Chapter 31, “Moral Skepticism;” Chapter 36, “The Varieties of Normativity;” Chapter 41, “Metaethical Quietism;” Chapter 42, “Methodological Naturalism in Metaethics;” Chapter 43, “Normative Ethics and Metaethics.”