Why Can’t Moral Naturalism and Error Theory Be Friends?

bike thief

Moral realism is the view that moral facts exist. Error theory is the view that moral facts do not exist. Accordingly, one might feel compelled to consider these two views inherently antithetical to one another. I wish to argue the opposite. More specifically, I wish to argue that either (1) error theory is compatible with certain varieties of moral naturalism or (2) error theory is mistaken in regards to what constitutes a moral fact.

But first, what is error theory? In “The Myth of Morality,” philosopher and avowed error theorist Richard Joyce provides us with a very general definition. An error theory, he explains, is any “position that holds that a discourse typically is used in an assertoric manner, but those assertions by and large fail to state truths.” When used in a specifically metaethical context, the term refers to the family of views making these two claims: (1) that moral statements make assertions and (2) that these assertions are false.

According to Joyce, the error theorist arrives at these claims by means of two steps. The first is to determine the semantic status of moral language. Not only does the error theorist think that moral statements are assertoric, but she also thinks that these statements are assertoric in a very special way. Take Joyce, for instance: he considers moral statements to be of the form “if x morally ought to φ, then x ought to φ regardless of whether he cares to, regardless of whether φing satisfies any of his desires or furthers his interests.”

Similarly, John Mackie—another prominent error theoriststates that “ordinary moral judgments include a claim to objectivity.” What both Joyce and Mackie are saying here essentially boils down to the same thing: that at the heart of any moral statement is a categorical imperative. Next, they claim that if a moral statement asserts a categorical imperative, then its truth-hood relies on the fact that “there are objective values or intrinsically prescriptive entities or features of some kind.” Thus, we arrive at the second step of error theory: to deny that any such entities exist.

At this junction, we may restate the two fundamental claims of error theory as follows: (1) that moral statements assert categorical imperatives and (2) that qua assertions of categorical imperatives, they are false.

Still, one can interpret the claim that moral statements assert categorical imperatives in at least one of two ways. It could mean that moral statements necessarily assert categorical imperatives. Or alternatively, it could mean that moral statements sometimes or even usually assert categorical imperatives. While error theory is theoretically compatible with either interpretation, Joyce and Mackie seem to lean towards the former one. For instance, although Joyce admits that possibly “we could get along adequately without such imperatives,” he dismisses such a possibility as “equivalent to saying that we could get along adequately without morality.”

Mackie, for his part, is slightly harder to pin down. As we have already seen, he thinks that “ordinary moral judgments include a claim to objectivity.” At least on the face of it, the qualifier “ordinary” appears to support the latter interpretation. But he also argues moral naturalism fails to capture the “apparent authority of ethics,” to the extent that such a view rejects its “categorically imperative aspect.”

I wish to demonstrate that the first interpretation is untenable. Here, I am arguing for two mutually exclusive possibilities. The first is that error theory is compatible with certain varieties of moral naturalism—by “certain varieties of moral naturalism,” I mean any view that asserts that morality is comprised entirely of hypothetical imperatives. If this possibility is true, then it obviously cannot be the case that moral statements necessarily assert categorical imperatives. The second possibility is that error theory is mistaken in regards to what constitutes a moral fact. As I have already explained, the error theorist understands moral facts to be the truth-makers for categorical imperatives. Thus, if there are no moral facts, then all categorical imperatives are false.

Of course, the error theorist also claims that most—if not all—moral statements assert categorical imperatives. It follows that the second possibility of my thesis may be reduced to the following claim: that the error theorist mistakenly holds that moral statements necessarily assert categorical imperatives. Accordingly, my thesis is that either (1) error theory does not hold that all moral statements necessarily assert categorical imperatives, in which case the view is compatible with certain varieties of moral naturalism, or (2) error theory mistakenly holds that all moral statements necessarily assert categorical imperatives, in which case the view is not compatible with certain varieties of moral naturalism.

Mackie, for his part, does not think that hypothetical imperatives can capture the prima facie authority of ethics. Joyce expresses a similar worry. He writes, for instance, that

[t]he manner in which we condemn Nazis, ignoring any usual desires or interests that they have, is not a peripheral element of moral discourse; it represents a kind of reprehensions that is central. A system of values in which there was no place for condemning Nazi actions simply would not count as a moral system.

In other words, certain actions seem to merit moral repudiation irrespective of the agent’s particular desires. Yet hypothetical imperatives tend to come in the form of conditional statements, such as “if x desires φ, then x ought to do ψ,” The worry, then, is that hypothetical imperatives simply aren’t enough to form a solid foundation for morality.

Suppose you come across a man who is in the process of stealing a bicycle. An avid cyclist yourself, you tell him that he should not steal the bike—or perhaps more forcefully, you might even say that stealing bikes is wrong. If moral imperatives are hypothetical, however, then the soundness of your judgment presumes certain facts about the bike thief’s desires, in which case it remains a very real possibility that these presumptions are simply unwarranted. If the man is a sociopath, say, it might be the case that the antecedent condition of your hypothetical imperative is not satisfied. In such a scenario, your moral judgment would be false.

What if the thief were to inform you of this fact? If he informs you that he lacks the relevant antecedent condition, would you then retract your judgment? I think Joyce hits upon our ordinary inclinations when he writes, “When we morally condemn a criminal we do not first ascertain the state of his desires.” But while I agree that these concerns reveal an important aspect of common moral thinking, I do not consider them to be fatal to moral naturalism.

Philippa Foot, in “Morality as a System of Hypothetical Imperatives,” asks us to consider a parallel scenario in the domain of etiquette. Imagine a gentlemen’s club, in which a particular club rule states that no one is to bring a woman into the smoking-room. One day, a certain member disregards the rule and enters the room with his mistress. When the secretary informs him of his infraction, he explains that he is quitting the club and cares nothing for his reputation. Will the secretary retract her statement? Not likely.

Foot concludes that if a hypothetical use of “should” gave rise to a hypothetical imperative, while a non-hypothetical use of “should” gave rise to a categorical one, then “should” statements based on rules of etiquette, or rules of a club, would all be categorical imperatives. Yet unless one truly wishes to argue that the rule that one should not bring a woman into the smoking-room is a categorical imperative, the upshot is that hypothetical imperatives may occasionally feel like categorical ones.

The word “feel” here is operative. As Foot explains,

just as one may feel as if one is falling without believing that one is moving downward, so one may feel as if one has to do what is morally required without believing oneself to be under physical or psychological compulsion, or about to incur a penalty if one does not comply.

Even so, we still have not properly responded to Mackie’s claim that the categorically imperative aspect of morality is what imbues it with its apparent authority. Why does morality nonetheless appear to hold a greater authority over us than, say, the rules of etiquette? According to Foot, it is only a matter of degree. To explain the apparent difference, she tentatively advances the theory that “there is nothing behind the idea that moral judgments are categorical imperatives but the relative stringency of our moral teaching.” We may desire a more sophisticated account than this, but I am sympathetic to Foot’s central claim: namely, that the force of moral statements may be explained without an appeal to categorical imperatives.

A more nuanced explanation of moral normativity has been advanced by Peter Singer, who argues that many of our more forceful moral intuitions may in fact stem from our evolutionary heritage. In “Ethics and Intuitions,” he describes an experiment in which subjects were asked to respond to a scenario where two siblings—Julie and Mark—engage in an act of sexual intercourse. Although the evolutionary reasons for condemning such an act are obvious, the scenario as carefully designed to circumvent each of them. For instance, it was a one-time affair, the siblings suffer no physical or psychological damage, they used birth-control and so forth.

As Singer tells us, “most people [were] quick to say that what Julie and Mark did was wrong.” Yet when pressed to justify their judgment, the subjects would appeal to those very reasons the scenario intentionally avoided. Singer explains the phenomena as follows:

Evidently, it is the intuitive response that is responsible for the judgment these people reach, not the reasons they offer, for they stick to their immediate, intuitive judgment, even after they have withdrawn the reasons they initially offered for that judgment, and are unable to find better ones.

Even after exhausting all of his original reasons, the subject would often hold on to his initial judgment. “I can’t explain it,” he would say, “I just know it’s wrong.” In other words, the subject would ultimately appeal to what they felt was a categorical imperative. The problem, however, is that their judgment can be adequately explained by means of evolutionary theory; and as Singer reminds us, “more evolved does not mean better.” He can therefore ask, was the subject’s judgment truly morally binding? Or did it merely feel as such? The most ontologically parsimonious answer, of course, is the latter one.

What both Foot and Singer demonstrate is that we can explain the “apparent authority of morality” without appealing to categorical imperatives. If this is correct, then it follows that the claim that moral statements must necessarily assert categorical imperatives is false—at least provided we can find a tenable theory of moral naturalism. So far, I have shown why the error theorist’s belief that naturalist accounts fail to capture the apparent authority of morality is unfounded. Next, I will show that a naturalist theory of morality can, in fact, be plausible.

The particular view I have in mind is that of Peter Railton, as advanced in his essay “Moral Realism.” A rough sketch of his theory is as follows. Take any agent A. At any point in time, A possesses  a set of “subjective interests”—i.e., her particular “wants or desires, conscious or unconscious.” Now, posit a hypothetical entity A+, such that A+ is identical to A but with the addition of full knowledge and unlimited cognitive capabilities. According to Railton, this hypothetical entity allows us to define A’s “objective interests”: to wit, that which A+ would want A to want.

For instance, let’s say that A ought not to eat that cheeseburger. But what does this mean exactly? Under Railton’s theory, the proposition that eating the cheeseburger is not in A’s objective interest simply means that A+ does not want A to eat it. Furthermore, Railton explains that A’s personal well-being is a function of the congruity between her wants and her objective interests. He writes,

[g]enerally, we can expect that what A+ would want to want were [she] in A’s place will correlate well with what would permit A to experience physical or psychological well-being or to escape physical or psychological ill-being.

If A eats that cheeseburger, she will presumably suffer certain negative consequences. And provided that she is rational, she will thus be deterred from consuming cheeseburgers in the future. Hence, a feedback mechanism exists between the wants of an agent and her objective interests. Although this is not yet a theory of morality but rather an account of hypothetical imperatives of the form “x non-morally ought to do ψ,” once we have this explanation in place, the step up to a theory of morality will not be too difficult.

Railton defines morality as “the assessment of conduct or character where the interests of more than one individual are at stake.” Yet not only can we posit an idealized version of an individual agent, but we can also posit an idealized version of a particular society. For any society S, we may ascertain S’s objective interest by “considering what would be rationally approved of were the interests of all potentially affected counted equally under circumstances of full and vivid information.” Therefore, when we say that “A morally ought to have done ψ,” we are making an appeal to what S+ as a whole would approve or disapprove of.

Even so, it is not necessary that the objective interests of an individual coincide with the objective interests of the society in which she lives. It follows that an individual can always—at least in theory—”opt out” of the objective interests set by S+. Unless we resort to categorical imperatives, however, such a possibility cannot be helped.

Although I have made but a crude sketch of Railton’s view, hopefully it is enough to demonstrate the plausibility of moral naturalism. Such a view also manages to avoid what Mackie has called “the arguments from relativity and from queerness.” The argument from relativity holds that the “radical differences between first order moral judgments make it difficult to treat those judgments as apprehensions of objective truths.” Yet as we have seen, Railton is not interested in moral judgments of the categorical kind, and therefore his view manages to be both relative and objective.

The argument from queerness, moreover, holds that the existence of objective values requires facts “of a very strange sort.” Of course, Railton’s view necessitates no such ontology. Ultimately, then, it appears as if the only real objection the error theorist has against a view such as Railton’s is that it fails to capture the apparent authority of ethics. But as I have already argued, this worry is unfounded.

Let us return to bike thief. “You shouldn’t steal bikes,” you tell him, wagging your finger in disapproval. Asserted as a Railton-style hypothetical imperative, your statement may or may not correspond to any moral facts. One possibility, which we have already entertained, is that the thief is a sociopath, in which case your statement would not map onto any ontological features. Quite simply, it would be false. Alternatively, if the thief subscribes to the objective interests of the society of which you are both members, then your statement does pick out an objective fact—insofar as it correlates to the objective interests of that society—and thereby expresses a truth.

But what happens if you assert not a hypothetical imperative, but a categorical one? Let us say that you inform the thief that “Stealing bikes is (categorically) wrong!” Here, both the error theorist and the Railton-style naturalist are in agreement: there are no moral facts that can satisfy moral assertions of this sort. That said, most moral naturalists would not explicitly acknowledge the conclusion that any statement purporting to assert a categorical imperative is false. But unless a naturalist theory is able to account for the legitimacy of categorical imperatives, I do not see how this conclusion is avoidable.

As we have seen, moral statements do not necessarily assert categorical imperatives. And furthermore, imperatives of this sort are not required for the formation of a plausible theory of morality. Yet the second fundamental claim of error theory does not apply to moral statements that merely assert hypothetical imperatives. Thus, we are faced with the two possibilities I mentioned earlier. If the error theorist admits to such a stipulation, then his view is compatible with certain varieties of moral naturalism. If not, then the error theorist is fundamentally mistaken in regards to what constitutes a moral statement and thereby in regards to what constitutes a moral fact.


Works Cited:

  • Foot, Philippa. “Morality as a System of Hypothetical Imperatives.” Foundations of Ethics. Ed. Russ Shafer-Landau and Terrence Cuneo. Carlton: Blackwell, 2007. Print.
  • Joyce, Richard. “The Myth of Morality.” Foundations of Ethics. Ed. Russ Shafer-Landau and Terrence Cuneo. Carlton: Blackwell, 2007. Print.
  • Mackie, John. “The Subjectivity of Values.” Foundations of Ethics. Ed. Russ Shafer-Landau and Terrence Cuneo. Carlton: Blackwell, 2007. Print.
  • Railton, Peter. “Moral Realism.” Foundations of Ethics. Ed. Russ Shafer-Landau and Terrence Cuneo. Carlton: Blackwell, 2007. Print.
  • Singer, Peter. “Ethics and Intuitions.” The Journal of Ethics, 9.3/4 (2005): 331-352. Web. 25 March 2013.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s