I. Introduction
H. A. Prichard is known as the author of a paper with one of the best titles in the history of philosophy, “Does Moral Philosophy Rest on a Mistake?” (1912)Footnote 1—and, alongside G. E. Moore, W. D. Ross, and others, as one of the non-naturalistsFootnote 2 whose views dominated metaethics in the early twentieth century. But this common picture of Prichard underestimates his place in the history of ethics, which I believe is central. This is not because he defended completely distinctive ideas; his most important views were shared by other philosophers of his period, from Henry Sidgwick to A. C. Ewing. But it was often Prichard who stated those views most forcefully and defended them best.
These views can be summarized in a slogan Prichard himself did not use: “Duty is underivative.” But this slogan can be applied at three different levels. The first concerns the normative realm as a whole; here it expresses the non-naturalist view that normative truths are sui generis, neither reducible to nor derivable from non-normative truths such as those of science. Duty is underivative in the sense that truths about how we ought to act, in the broadest sense of “ought,” are self-standing. The second level focuses more narrowly on moral judgments, as one kind of normative judgment. Here the claim is that truths about how we ought morally to act are also underivative, not only from non-normative truths but also from any other normative truths; there are no nonmoral “oughts” or values from which moral “oughts” derive. The final level is that of specific deontological duties such as duties to keep promises, not harm others, and so on. For Prichard, these duties do not derive from a more general consequentialist duty to promote good consequences. The main reason we ought to keep our promises or not harm others is just that we ought to; those duties, like the normative realm as a whole and moral duty in general, are self-standing.
Prichard accepted all three of these claims, though he did not distinguish the first two from each other, as many present-day philosophers do. My main interest is in his defense of the second claim, about moral duty in general; this defense appears in his famous argument that it is a mistake to ask “Why ought I to do what I morally ought to do?” because the only possible answer is “Because you morally ought to.” But I will begin by examining his defense of the third claim, about deontological duties, because it sheds light on his methodology in discussing the other two. This is the subject of Section II; later sections will address claims one and two.
II. Deontological Duty Is Underivative
The substantive moral view in “Does Moral Philosophy Rest on a Mistake?” (hereafter “Mistake”) is close to that later defended by Ross, who probably took it over from Prichard. On this view, there is not just one basic moral duty, as consequentialism and also Kant say, but several—some (perhaps) to bring about certain consequences, but others largely independent of consequences, such as duties to keep promises, not harm others, and so on, where the latter duties make the view deontological. The various duties can conflict, but when they do there are no rules for deciding between them: we can only make a direct intuitive judgment about which duty is stronger. In expounding this pluralist view, Prichard did not use the language Ross would introduce of “prima facie duties,” though Prichard at one point suggested the equivalent term “claim.”Footnote 3 But the two philosophers' views were close: there is a plurality of (sometimes) deontological duties, with no single unifying ground for them.
Prichard's defense of this view, and especially his critique of consequentialism, emphasized one of two distinct arguments a deontologist can make. Sixty years after “Mistake,” Robert Nozick said that utilitarian attempts “to derive (approximations of) usual precepts of justice … do not yield the particular result desired, and they produce the wrong reasons for the sort of result they try to get.”Footnote 4 Prichard sometimes made the first of these objections to consequentialism, saying, for example, that it can favor outcomes that are unjust and therefore wrong to produce.Footnote 5 But his more common objection was the second one, about explanation: that even when consequentialism yields the right conclusion about how we ought to act, it gives the wrong reason for it. In “Mistake,” Prichard wrote:
Suppose we ask ourselves whether our sense that we ought to pay our debts or to tell the truth arises from our recognition that in doing so we should be originating something good, e.g. material comfort in A or true belief in B, i.e. suppose we ask ourselves whether it is this aspect of the action which leads to our recognition that we ought to do it. We at once and without hesitation answer “No.”Footnote 6
According to Prichard, we ought to pay our debt because we incurred it, and not because (or only because) of any good that will result. He made a similar objection to Kant's attempt to unify the moral duties under the first formulation of his categorical imperative:
No one could suppose that the reason why an act ought to be done consists in the fact that everyone could do it. Even Kant could not have supposed this. The difficulty escaped him because it didn't occur to him that his criterion of moral rules must express what, on his view, is their reason.Footnote 7
Or, as the same point has recently been put, even if Kant's universal law formulas do flag actions from certain motives or maxims as wrong, they “do not adequately explain why acting on those maxims is wrong. What is wrong with slavery, for example, is not adequately explained by saying that it is impossible for everyone to act [on] the maxim of a would-be slave-owner.”Footnote 8
Prichard's interest in the explanation of particular duties was shared by many other philosophers of the late nineteenth and early twentieth centuries. The British Idealists allowed that utilitarianism mostly yields correct verdicts about which acts are right, but insisted that the reason why those acts are right is not that they promote pleasure; it is that they promote perfectionist goods.Footnote 9 In Principia Ethica, Moore said that we ought most of the time to obey those rules obedience to which will best preserve society, and that these rules will be the same given any plausible theory of what is good. But he still called the question of what is in fact good the “primary ethical question,” because it concerns the reason why those rules are best.Footnote 10 And Prichard's emphasis on explaining particular duties was very much shared by Ross. Ross too sometimes argued that consequentialism yields the wrong conclusions,Footnote 11 but he titled his defense of deontology “What Makes Right Acts Right?” and argued that even when consequentialism is right about which acts are right, it is wrong about why they are right. If we think we ought to keep a promise, he insisted, the reason is not that this will have good consequences; it is simply that we promised.Footnote 12
Prichard's emphasis on explanation also fits a moral epistemology he shared with Ross and C. D. Broad. Unlike Sidgwick, Prichard did not believe we can come to know abstract moral principles by reflecting on them just as abstract principles; moral intuitions are elicited only in particular situations. But what we intuit in a particular situation is that an act's having some nonmoral property tends to make it right or wrong—for example, that its being the return of a favor tends to make it right: “We recognize that this performance of a service to X, who has done us a service, just in virtue of its being the performance of a service to one who has rendered a service to the would-be agent, ought to be done by us.”Footnote 13 Though prompted by a particular situation, the intuition is implicitly general, since it implies, as Prichard emphasized, that any act of returning a favor is, other things equal, right. And this implication, while not equivalent to the principle that we ought, other things equal, to return favors, is sufficiently close to it that our grasp of the principle follows by a small step.Footnote 14
This epistemology contradicts the not uncommon view of Prichard as a moral particularist, who denied the truth of general principles and held that moral knowledge concerns only particular acts as particular. On the contrary, he often emphasized the importance of moral principles, applauding Kant, for example, for recognizing that “any particular right action involves a principle binding on every one always.”Footnote 15 Prichard's epistemology also connects with his main line of objection to consequentialism, since it makes the primary moral intuition one about explanation: that an act's having a nonmoral property like being the return of a favor explains why it is, other things equal, right. But if our primary intuitions are explanatory, then to assess a moral theory for its claims about explanation is to assess it against the most secure moral knowledge we have.
Prichard's epistemology also widened the range of intuitions a successful moral theory must accommodate. Sidgwick is often taken to have argued that utilitarianism fits not only our intuitions about abstract principles but also our judgments about particular cases.Footnote 16 But, for him, these judgments tended to concern only what is in fact good or right, so it does not matter if utilitarianism gives instrumental explanations for the goodness of things that common sense thinks are good in themselves. Moore criticized Sidgwick sharply on this score, saying that even if hedonism yields the right conclusions about which pleasures are best, by considering their effects on future pleasures, it does not capture our convictions about why they are best, which concern what they are pleasures in now.Footnote 17 Prichard similarly expanded the judgments to be captured by a theory of the right when he included, and in fact made central, judgments about why right acts are right rather than just those about which acts they are.
Prichard expressed his objection to consequentialism in a radical way. He thought the proper description of a moral duty must always mention its explanatory ground: if the reason we ought to perform some act is that it has property F, then our duty in the situation is really to perform the act that has F. Footnote 18 But then, in trying to derive the duty to keep promises from a duty to promote the good, consequentialism turns the duty to keep promises into a quite different duty to promote value—if that is what explains the duty, that is what the duty is.Footnote 19 And this means that consequentialism turns the duty to keep promises into something it is not, and thereby distorts the moral phenomena: in trying to explain the duty to keep promises, consequentialism destroys it.
This idea of distorting the moral phenomena was central to Prichard's argument that moral duty in general is underivative, and, in particular, the ideal was central to the third and most distinctive of three stages in that argument. Before we address that, however, there is a question about how far-reaching his critique of consequentialism was.
Ross thought consequentialism is at least partly right, since for him one important duty is to promote good consequences, and whenever there is a duty to promote some state X, it is in part because X is good. When Prichard in “Mistake” set himself “in opposition to the view that what is right is derived from what is good,”Footnote 20 was he agreeing with the view Ross would later defend, or did he mean to reject consequentialism in some more comprehensive way?
We can distinguish three views about the duty to promote a state of affairs X: (1) The duty to promote X is entirely underivative, and in particular does not depend on any claim that X is good. (2) The duty to promote X depends partly on the claim that X is good and partly on the independent claim that we have a duty to promote whatever is good; X's goodness supplies some but not all of the ground of the duty. (This was Ross's view.) (3) The duty to promote X is grounded wholly in the claim that X is good, because claims about the good directly entail claims about what we ought to do. (This was, notoriously, Moore's view in Principia Ethica.) Prichard clearly rejected view (3), but was his target in “Mistake” just (3), or did he mean also to reject (2)? Jonathan Dancy has recently defended the second reading, arguing that while for Prichard there may be acts whose rightness depends on their consequences (for example, on the fact that they promote pleasure), their rightness never depends on those consequences' being good.Footnote 21
But this reading is hard to square with the text of “Mistake.” First, Prichard's target in that paper was the view that “what is right is derived from what is good,” and that is not the view of (2), which treats the duty to promote whatever is good as underivative. Only view (3) derives the right from the good. Second, many of Prichard's arguments were relevant only to (3). He insisted that “[a]n ‘ought’ … can only be derived from another ‘ought’,”Footnote 22 but (2) agrees with him on this point if it treats the duty to promote whatever is good as underived. More specifically, he argued that deriving the right from the good requires that the good be understood as what “ought to be,”Footnote 23 and this is again not true of (2), which can understand “good” in whatever way it likes so long as it adds an independent duty to promote value as so understood. It is only view (3)'s attempt to derive the right entirely from the good, combined with the principle that an “ought” can only follow from another “ought,” that requires an “ought” within the meaning of “good.” Finally, Prichard located his target in “Mistake” in a passage from Hastings Rashdall's Theory of Good and Evil that argued, just as Moore had, that “right” means “productive of the most good.”Footnote 24 And that is view (3), not (2).Footnote 25
Thus, Prichard's anticonsequentialist argument in “Mistake” was consistent with recognizing a duty to promote the good, as Ross did. Nonetheless, Prichard's substantive view, especially later in his career, was in two respects further removed from consequentialism than Ross's.
First, in his early writings, Prichard held that it is a necessary condition of an act's being right that it produce something good: “unless the effect of an action were in some way good, there would be no obligation to produce it.”Footnote 26 And this view was shared by Ross and E. F. Carritt.Footnote 27 Of course, none of these philosophers were consequentialists; they did not think right acts always maximize the good. But that is because they thought we can stand in certain relations to another person, such as having promised him something or caused him some harm, that make it right for us to produce less good for him rather than more good for someone else: the relation shifts the target of our good-promoting activity from one person to another. But it is still necessary, for an act to be right, that it have some good effect.
Carritt and Ross seem to have retained this view, but Prichard implicitly abandoned it. Whereas his early papers recognized a plurality of goods, including pleasure and knowledge,Footnote 28 a 1928 letter from him to Ross said there is only one intrinsic good: virtuous disposition. Pleasure and knowledge are not good, as Ross thought, nor is the active exercise of virtue better than its mere possession; the only good is being virtuously disposed.Footnote 29 But if Prichard believed this after 1928 and still affirmed duties to promote pleasure and knowledge, as he seems to have done, he could not ground these duties even partly in the claim that the states in question are good; the duties would have to be entirely underivative, as on view (1). Nor would it be true for him, as for Ross, that whenever there is a duty to promote X, X must be good.
Second, Ross thought that for states that are good in the strict sense, such as virtue and knowledge, there is a single undifferentiated duty to promote them in all people everywhere; the relations that help ground deontological duties are irrelevant. But in an early paper Prichard wrote that any moral principle must mention two things: “(a) a good thing which the action will produce, (b) a definite relation in which the agent stands either to another or to himself.”Footnote 30 And the same view appeared in his much later paper “Moral Obligation” (published posthumously in 1949). Prichard was there challenging the consequentialist idea that what makes an act right is just its causing something good, and taking as his example of something “indubitably good” an increase in someone's patience, as fit his later view that only virtue is good. He objected that the consequentialist idea ignores the difference between increasing one's own patience and increasing another person's, where the different relations in which we stand to the person whose patience is at issue make these different duties: “We shall be thinking of the acts as duties of different sorts just because the one will be making ourselves better and the other making someone else better.”Footnote 31 Prichard therefore extended the grounding role Ross thought relations play only for deontological duties to all duties whatever, and on that basis Prichard held that our duties to produce the same good in different people can differ in strength. Ross thought that if we ought to spend more time promoting our own virtue than other people's, this is only because we can do so more effectively.Footnote 32 But Prichard held that, effectiveness aside, we have a stronger duty to promote our own virtue because it is ours, and may also have a stronger duty to promote our children's virtue.Footnote 33 Even when a duty is just to promote some good, it matters to the character of the duty whose good it is.
If this was Prichard's view, however, it should be read into his deontological critique of consequentialism. Even when consequentialism yields the right verdict about which act is right, he held, it oversimplifies the explanation of the act's rightness by omitting the crucial fact of relationship: the fact that the person we are acting toward stands in a specified relation to us, even that of being a stranger or being ourself. That he stands in this relation is a vital part of our intuitive understanding of what makes the act right, and in ignoring it consequentialism distorts the moral phenomena. A similar charge about distortion figured in Prichard's argument that moral duty in general is underivative, and in particular in the third stage of a three-stage argument he gave for that claim, as we will see in Section V.
III. Conceptual Minimalism
Prichard's second claim, that moral duty in general is underivative, can be combined with any view about what the content of that duty is. It says that the basic moral duties, whatever they are, are not grounded in any more fundamental considerations, either normative or non-normative. A deontologist like Prichard can say that the duty to keep promises is underivative, but a consequentialist can say the same about his preferred fundamental duty. Asked why we ought to promote the good, a consequentialist can say there is no other answer than that we morally ought to. And that is what the leading consequentialists of Prichard's day—Sidgwick, Rashdall, and Moore—did say. Though differing from Prichard about what the basic moral duties were, they agreed that they were self-standing.
As I said at the start of this essay, Prichard did not distinguish the second claim, about specifically moral duty, from the claim that normative considerations in general are underivative. This is because he held that there are no “oughts” other than the moral “ought,” so there are no normative considerations other than moral ones from which a moral duty could derive. This was an aspect of a conceptual minimalism that he again shared with other philosophers of his period and that grounded a first stage in his argument for underivative duty.
These philosophers thought that all normative claims could be expressed using just a few basic concepts. For some, there was only one basic normative concept: for Sidgwick, that of what one “ought” to do; for Moore in Principia Ethica, that of what is “good”; and for Broad and the middle Ewing, that of what is “fitting.” Others such as Prichard, the later Moore, and Ross recognized two basic concepts, often “good” and “ought.” But they all denied that there were more than a very few irreducible normative concepts.
Faced with a proposed additional normative concept, a minimalist of this kind can respond in either of two ways: by reductively analyzing the concept in question using one of his favored basic ones, or by denying that it is really normative. Both approaches were used by Prichard and his contemporaries. Thus, many of them held that “moral goodness” is just ordinary intrinsic goodness when had by a specific type of object (for example, an attitude to an object with some other moral property such as rightness or goodness). This illustrates the reductive option. Sidgwick, Rashdall, and Moore took a similar line about the concept of “my good,” which they understood as that portion of what is good (however “good” is understood) that is located in my life, or is a state of me. But Prichard used the alternative approach for the related concept of what is a “good to” a person, or, in present-day language, what is “good for” her or a constituent of her “welfare.” This, he claimed, is not a normative concept at all but makes only the descriptive claim that something will satisfy a person's desires or, more accurately, give her pleasure; something's being a “good to” a person, therefore, does not give her or anyone else reason to promote it.Footnote 34 Finally, the same two options present themselves for any proposed nonmoral “ought,” such as the instrumental “ought” represented by Kant's hypothetical imperative. One possibility is to analyze this as a categorical imperative with a distinctive content, as on what is now called the “wide scope” reading of the hypothetical imperative. This treats this imperative as a command to make a certain conditional true, namely, to make it the case that if you desire or intend a certain end, you take (what you believe are) the means necessary to achieve it; this reading is suggested at least once by Sidgwick.Footnote 35 The other option is to deny that the instrumental “ought” is normative, and this was again the view of Prichard and Ross. Prichard argued that Kant's hypothetical imperative is not really an imperative at all, since it makes only the descriptive claim that a certain action is necessary for the achievement of a given goal. When we use “ought” instrumentally, “what we really mean … has to be expressed by the hypothetical statement: ‘If I do not do so and so, my purpose will not be realized’.”Footnote 36 For Ross, the hypothetical imperative says only “that a man who desires certain ends can hope to get them only if he adopts certain means.”Footnote 37 This is why, for Prichard and Ross, there are no nonmoral “oughts.” The only genuine “oughts” are categorical, and since (following Kant) all categorical “oughts” are moral, there are only moral “oughts” or moral requirements.
This conceptual minimalism contrasts starkly with the approach of much present-day moral philosophy, which operates with a larger number of irreducible normative concepts: the “good for” a person alongside what is simply “good,” as well as prudential, aesthetic, and other “oughts” alongside the moral one. But in my view the minimalism of Prichard and his contemporaries is a far sounder view. Consider the conflict Sidgwick wrestled with, between a principle saying we ought to promote only our own pleasure and one telling us to promote the pleasure of all. If these principles are genuinely to conflict, they must use or at least entail claims using a common concept of “ought”; otherwise, they will talk past each other. But then resolving their conflict is a substantive matter, which involves deciding which of two claims using the common “ought” is stronger. If we like, we can label the first claim “prudential” or “rational” and the second “moral,” but this will not alter the core issue, which is between two substantive claims about what we ought (in the common sense) to do. And the labeling only invites confusion, by suggesting that the issue turns somehow on conceptual questions about what “prudence” or “rationality” and “morality” in the abstract involve. The great virtue of minimalism is that it avoids this confusion by presenting all normative issues as what they really are, namely, substantive issues about which claims using a common concept are true.
It may be objected that this common “ought” need not be labeled “moral”; it could instead be a generic practical “ought,” with moral uses distinguished by, say, their subject matter, such as a reference to others' good. But if this objection grants that there is a single “ought,” it raises only a verbal issue about what that “ought” should be called, and nothing substantive can turn on that. Moreover, I have suggested two arguments Prichard could give for calling this “ought” moral: it treats substantive questions as clearly substantive, and it fits a Kantian tradition whereby all categorical imperatives are moral. Whether or not these arguments are persuasive, it was certainly Prichard's view that the only “ought” is moral.
Given this view, his minimalism combined with his non-naturalism to yield a first stage in his argument that moral duty is underivative, or that it is a “mistake” to ask why we ought to do what we morally ought to do. If there is no “ought” other than the moral one, then there is no “ought” in terms of which this question can be framed; if moral truths are underivative from non-normative ones, as non-naturalism claims, they are completely underivative.
This again was a common view in Prichard's time. Moore said that the question “Why should I do my duty?” is “puzzling,” because it reduces to “‘Why is duty duty?’ or ‘Why is good good?’,” while Carritt wrote, “If any one ask us, ‘Why ought I to do these acts you call my duty?’ the only answer is, ‘Because they are your duty.’”Footnote 38 They too assumed, with Prichard, that the only normative concepts are moral. Sidgwick may seem an exception to this consensus, since he said that we ask “Why should I do what I see to be right?” when we do not ask “Why should I believe what I see to be true?” But his explanation was that we are torn between competing substantive views about what is right, and we express our uncertainty by asking the question; this implies that if it were settled what is right, there would be nothing to ask.Footnote 39 Thus, all these philosophers agreed with Prichard that the question “Why ought I to do what I morally ought?” does not arise, because the first “ought” cannot differ from the second. And that follows from one aspect of their shared conceptual minimalism.
IV. The Epistemological Argument
Prichard's minimalism, therefore, gave him an initial argument for the underivativeness of moral duty, but, on its own, this argument could not establish any substantive conclusions. It showed that whatever normative considerations are overriding should be called “moral,” but Prichard wanted to do more. He wanted to show that the duties in his specific deontological view are underivative, and a consequentialist will want to show that the duty to promote the good is underivative. No mere conceptual claim about “ought” can show that; from conceptual premises only conceptual conclusions follow.
To see this, imagine how an instrumentalist in the style of Thomas Hobbes might respond to Prichard's minimalism. The instrumentalist may initially have said that a supreme principle of rationality tells us to do whatever will most satisfy our desires, and that moral “oughts” about keeping promises or promoting others' pleasure have normative force only when acting on them will be rational in this sense. If the instrumentalist accepts minimalism, he can no longer speak of a rational “ought,” but that only requires him to reformulate his view. He can now say that the supreme moral principle tells us to do whatever will most satisfy our desires, and that more specific “oughts” about promises or promoting pleasure have force only when acting on them will satisfy this principle. His position will be substantively unchanged, and he will still deny that the duties Prichard and the consequentialist want to affirm have force.
Prichard implicitly recognized this possibility, and the second and third stages of his argument were directed against it. The second stage was an epistemological argument that addressed skeptical challenges to the conventional moral duties and paralleled Moore's well-known reply to skepticism about the external world. Moore held that we start out assuming we have direct awareness and therefore knowledge of external objects. A skeptic like David Hume may then propose a theory of knowledge according to which this assumption is false: we are directly aware only of our experiences, and we form our beliefs about external objects by inference from these experiences. But since all such inferences are unwarranted—premises about experiences license only conclusions about further experiences—our beliefs about external objects are unjustified and therefore not knowledge. Moore replied that, whatever theory of knowledge underwrites a skeptical argument like Hume's, we cannot have as much reason to believe it as we do to retain our initial belief that we know external objects. That common-sense belief is the foundation from which philosophy starts, and far from being refuted by a theory like Hume's, it shows that theory to be false.Footnote 40
The second stage of Prichard's argument took a similar line. We start out believing we have an obligation to, say, keep promises. A skeptic like Glaucon in Plato's Republic may ask whether we really do have such an obligation, given that fulfilling it will often work against our interests. The skeptic assumes something like Hobbes's instrumental principle: we have reason to do something only when it will further our desire-satisfaction, and we therefore have no reason to keep a promise when it will not. Prichard replied that we have less reason to believe the instrumental principle than to retain our initial conviction about the duty to keep promises, because that conviction involves direct apprehension of a self-evident truth and therefore is knowledge. That we have the specific duty is the better-grounded conviction, and its being so shows the instrumental principle to be false.
Prichard himself drew the parallel with the epistemological case. There, he said, we start out believing we have knowledge of, say, mathematics. A skeptic like Descartes can make us doubt whether this is so, but the correct response is to reflect on our original mathematical convictions and see that they were in fact knowledge, so that whatever theory generated Descartes's skepticism must be false.Footnote 41 The stage of being moved by such skepticism is not pointless; it is an essential part of philosophical reflection. But its end-result should be a return to our original convictions, and so it is with moral duty. Skeptics demand a proof that we have a duty to keep promises, but on Prichard's view this involves
the mistake of supposing the possibility of proving what can only be apprehended directly by an act of moral thinking. Nevertheless the demand, though illegitimate, is inevitable until we have carried through the process of reflection far enough to apprehend the self-evidence of our obligations, i.e. the immediacy of our apprehension of them.Footnote 42
Of course, not all philosophers are persuaded by Moore's reply to external-world skepticism, and not all will be persuaded by Prichard on moral duty. This is especially so since Prichard's response turns on an appeal to intuition: specifically, an intuition of the reality of a duty such as the duty to keep promises. But his argument is strengthened by his conceptual minimalism. Proponents of the instrumental view often call it a view about “rationality,” implying that it has priority over merely “moral” claims about promise-keeping or promoting others' pleasure. But if the instrumental “ought” is the same as the moral “ought,” the instrumentalist's claim is conceptually on a par with those other claims and must be weighed substantively against them—it has no epistemic advantage. Consider a case where a person's overriding desire is to become rich and he knows that the most effective way to do so is to kill his cousin and acquire her inheritance. Which is the true “ought” statement applying to him: that he ought simply to kill his cousin, as the instrumental view implies, or that he ought simply not to kill her, as Prichard believed? Surely most people will say the latter, and it is easier for them to do so if both statements use the same “ought.” It likewise strengthens Prichard's reply if we distinguish the skeptic's reading of the instrumental or hypothetical imperative from ones that do not conflict with moral duty and therefore cannot undermine it, such as Prichard's non-normative reading and the wide-scope reading described above.Footnote 43 That the principle is true on one of these other readings does not imply that it is true on the skeptic's reading, which does conflict with duty. And once this distinction is made, the instrumentalist principle read in the skeptic's way looks decidedly unattractive. Surely most people faced with the foregoing example will find it more compelling that the person ought not to kill his cousin than that the supreme normative principle tells us to do whatever will satisfy our desires.
It may be objected that the instrumental view Prichard opposed is not the only one that allows us to ask, “Why ought I to do what I morally ought?” The instrumental view proposes a substantive criterion of rationality and says that moral “oughts” are binding only when they satisfy this criterion. But a different view allows that moral “oughts” have independent force, while holding that they must be weighed against other self-standing “oughts” such as prudential or aesthetic ones, to determine what a person ought to do all things considered. “Why ought I to do what I morally ought?” then becomes the question whether, in a particular case or generally, what I morally ought to do is the same as what I ought to do all things considered. And given the plurality of competing “oughts,” proponents of this view think the answer is sometimes no.
To illustrate this objection, imagine that a person can produce either 9 units of pleasure for herself or slightly more, say, 10 units, for other people. Utilitarians will say she must produce the 10 units, but many will disagree. They will say that if she had a choice between 9 units of pleasure for herself and 1,000 for others, she would have a duty all things considered to produce the 1,000. But when the difference is smaller, as between 9 and 10 units, she does not have that duty but is permitted to prefer her own lesser pleasure. She may also be permitted to prefer the 10 units for others, but the point is that she is not required to do so. And this, the objector says, shows that the moral “ought” does not always uniquely determine the all-things-considered “ought.” The moral “ought” has to be weighed against a prudential “ought” concerning a person's own pleasure, and in that conflict will sometimes lose.
As it happens, Prichard rejected a key premise of this objection. He believed, as Ross also did, that while we have a moral duty to promote other people's pleasure, we have no duty of any kind to promote our own: “we do not think it a duty to aim at—i.e. to act from the desire of—our happiness; nor do we even in fact think that it is a duty to do … those acts which we think will lead to our happiness.”Footnote 44 If the basic normative considerations are all duties, this view implies that in the example above the person is not permitted to produce the 9 units of pleasure for herself; she must produce the 10 units for others. The view therefore underwrites a quick reply to the objection, but it does so at a heavy price. As Ewing and Michael Stocker have pointed out, it implies that the person ought also to prefer 9 units of pleasure for others to 10 for herself, and even 9 units for others to 1,000 for herself. In the latter case, she has some duty to produce the 9 but no duty to produce the 1,000, so she must prefer the lesser pleasure of others to a vastly greater pleasure of her own. And that is absurd.Footnote 45
But Prichard need not have taken this line. Instead, he could have agreed that a person is permitted to prefer 9 units of her own pleasure to 10 units for others, but he could have generated that permission within morality itself, as a moral permission. More specifically, he could have supplemented the underivative moral duties in his and Ross's substantive view with an equally underivative moral permission to pursue one's own pleasure, a permission that, like the duties, is only prima facie and must be weighed against other considerations to determine what one is, all things considered, permitted or required to do. Prichard, Ross, and their contemporaries did not think much about underivative permissions; they tended to assume that the basic moral elements are all duties. But it is entirely consistent with their general approach to recognize such permissions, and if Prichard did, he could say that when a person has a choice between her own lesser pleasure and the vastly greater pleasure of others (say, 9 for herself and 1,000 for others), her other-regarding duty outweighs her self-regarding permission and she ought, all things considered, to produce the greater pleasure. But when the gap between the pleasures is smaller, as in the 9 versus 10 case, her permission outweighs her duty and she may prefer her lesser pleasure. The objection's claim about the 9 versus 10 case is therefore correct, but it does not show that a moral duty can be outweighed by something nonmoral; it only shows that it can be outweighed by a moral permission.
This alternative response to the objection has several merits. It explains why, while a person may prefer her own 9 units of pleasure to 10 units for others, she may also prefer the others' 10 units. What she has concerning her own pleasure is only a permission, and if she declines to exercise that permission she does nothing wrong.Footnote 46 The response also yields a version of Prichard's and Ross's view that we have no duty to pursue our own pleasure; if we do not do so, we are again merely not exercising a permission. Finally, the response retains the central virtue of Prichard's minimalism. The issue raised by the 9 versus 10 case is how normative considerations concerning one's own pleasure weigh against considerations concerning the pleasure of others. But if the two sets of considerations are to conflict, their claims about “ought” and “may” must operate in the same field, and if they do, the issue between them is substantive rather than conceptual. Calling both sets of considerations “moral” recognizes this fact, while labeling one “prudential” and the other “moral” suggests falsely that the issue turns somehow on conceptual questions. Making the permission to prefer one's own lesser pleasure a moral permission rightly treats a substantive question as substantive, and similar treatments of aesthetic and other supposed nonmoral “oughts” will do the same.Footnote 47 At the same time, adopting the response would only slightly modify Prichard's epistemological argument. That argument supplements his conceptual claim that the only “oughts” are moral with the substantive claim that certain specific “oughts,” such as those in his deontological view, are better grounded than any contrary principle a skeptic may appeal to. And that claim still stands if, alongside the moral “oughts,” there are some moral permissions.
The history of twentieth-century British philosophy has usually been told in a Cambridge-centered way: there was a late nineteenth-century Idealist movement centered in Oxford that was exploded around the start of the new century by the Cambridge realism of Moore and Bertrand Russell. But there was a simultaneous realist movement in Oxford, initiated by John Cook Wilson and with Prichard and Ross as leading members. Prichard's writings are shot through with what Russell called a “robust sense of reality,” as when Prichard said the ground of a duty to perform an action X cannot be properties X would have if it were performed, because nothing that does not exist can have properties.Footnote 48 It should be no surprise that part of his defense of moral duty mirrors a well-known realist response to skepticism about the external world, insisting that our knowledge of duty is more secure than our belief in any theory that could call that knowledge into question.
V. Distorting Moral Duty
The second, epistemological stage of Prichard's argument was directed mainly against a skeptical view that proposes a substantive criterion for a duty's having force and then says moral duties do not satisfy it. But this stage also challenged anti-skeptical views that say moral duties do satisfy some such criterion and therefore do have force; such a demonstration, it argued, is unnecessary. And Prichard had a further argument against these views: in trying to justify moral duties on some independent ground, they distort the moral phenomena. This was the third stage in his argument for underivativeness, and it echoed his critique of consequentialism.
Prichard's main target here was, again, an instrumentalist view that assumes moral duties ground genuine “oughts” only when fulfilling them will further the agent's interests, say, by maximizing his pleasure. Prichard sometimes argued that this view's key empirical claim is false: it is not true that fulfilling our moral duties always maximizes our pleasure.Footnote 49 But his more common argument was that, even if this empirical claim is true, the view gives the wrong explanation of our moral duties and thereby turns them into something they are not. The version of this argument in “Mistake” was very brief:
The [instrumentalist] answer is, of course, not an answer, for it fails to convince us that we ought to keep our engagements; even if successful on its own lines, it only makes us want to keep them. And Kant was really only pointing out this fact when he distinguished hypothetical and categorical imperatives.Footnote 50
But Prichard developed the argument more fully in later writings such as “Duty and Interest” (1929). There he argued that if what gives a moral duty normative force is the fact that fulfilling it promotes the agent's advantage, then the duty is really one to promote the agent's advantage: as in the critique of consequentialism, what explains the duty also gives its content. But this, he claimed, distorts the moral phenomena, since what we believe is that we have duties to keep promises, relieve others' pain, and so on just as such, or with promise-keeping and pain-relieving as their content. So the attempt to justify the duties instrumentally in fact destroys them.Footnote 51 More generally, Prichard argued that the instrumental justification of moral duties resolves the moral “ought” into the nonmoral one, in effect denying that there is a moral “ought” at all. It “covertly resolve[s] moral obligation into something else which is not moral obligation”; in response, the “apparent tautology ‘moral obligation is moral obligation’” reminds us that “moral obligation … is sui generis.”Footnote 52 Or, what the instrumental justification is doing
is denying by implication that there really are such things as right actions and wrong actions at all in the moral sense of these terms, and maintaining that all that exists instead is right actions and wrong actions in the purely non-moral sense of actions which are, and of actions which are not, conducive to our purpose.Footnote 53
Given Prichard's conceptual minimalism, the distortion here is radical. If the nonmoral “ought” is not an “ought” at all, but makes only the non-normative claim that an act will promote the agent's interest, then the instrumental view does not replace one normative concept with another; it denies normativity altogether. In attempting to justify the moral “ought,” it in effect denies that “oughts” exist.
Prichard's presentation of this argument assumed his radical view that a proper description of a moral duty must mention its explanatory ground, but the argument can be separated from that view. Then it says only that the instrumental view gives the wrong explanation of our moral duties. It says that we ought to keep our promises or promote others' pleasure because this will promote our own pleasure, and even if this does not destroy the duties, it is not the right explanation of them. We ought to do these things just as such.
However it is stated, Prichard's explanatory objection to instrumentalism, like his objection to consequentialism, turns on an intuition: in this case, it is the intuition that we have some moral duties, whether to keep promises or just to promote the good, that are binding on us as such and apart from any connection to our pleasure. And while an objector may deny this intuition, Prichard again seems to be on strong ground. Surely most of us think there are some things we ought to do, all things considered, even though they will not maximize our own pleasure. And the contrary instrumentalist view, if it is indeed normative, likewise appeals to an intuition, and thus has no epistemic advantage.
Alternatively, an objector may say that Prichard's argument tells only against an implausible instrumental justification of moral duty and does not generalize to all such justifications. This brings us to one of Prichard's distinguishing characteristics as a philosopher: that alongside his profundity he had a strong streak of perversity. This showed itself partly in the issues he discussed; he could worry at great length about a point that seems not to merit the attention, such as whether a not-yet-performed act can have properties. But it is especially evident in his interpretations of other philosophers, which could border on the absurd. He often assumed that there were just two possible views on a topic, his own and some very implausible alternative, and then ascribed the alternative to anyone who rejected his view.
This occurs often in his moral philosophy, where he thought the main alternative to his non-naturalist deontology was a view combining the instrumental principle, read as non-normative, with psychological egoism, the claim that everyone desires only his own good or, more specifically, pleasure. Given these assumptions, the justification of moral duty has to be that fulfilling it will maximize the agent's pleasure. Prichard ascribed this view to some philosophers who did indeed hold it, such as Hobbes, as well as to some who at times suggested it, such as John Stuart Mill and T. H. Green. But he also ascribed it to philosophers for whom it is at least an uncharitable reading, such as Plato and Aristotle, and to some for whom it is ridiculous. He argued that Sidgwick's defense of his axiom of Rational Benevolence claims that acting on the axiom will maximize the agent's satisfaction,Footnote 54 and even interpreted Joseph Butler (Butler!) as a psychological egoist.Footnote 55 These misinterpretations can tempt one to reject Prichard's critiques of the best-known justifications of moral duty as simply missing their target, and that reaction has been common. But in my view it is a mistake, because the core of Prichard's critique still applies to these justifications when they are more charitably understood. Let me illustrate by examining his critique of ancient ethics, and especially of Aristotle.
VI. Prichard on Aristotle
Prichard's view of Aristotle may have changed over time. In “Mistake,” Prichard noted “the extreme sense of dissatisfaction produced by a close reading of Aristotle's Ethics,”Footnote 56 and in so doing, he upset J. O. Urmson, who complained in his introduction to the 1968 edition of Prichard's Moral Obligation how “imperceptive” this remark was about what he thought the greatest ethics book ever written.Footnote 57 But Urmson had not bothered to read the next paragraph! For there Prichard explained that his remark was not intended to criticize Aristotle. Our dissatisfaction with the latter's Ethics arises from the fact that it does not do what we want it to, namely, explain why we ought to do what we think we morally ought to do. And Prichard's main point, as he then repeated,Footnote 58 was that this demand is illegitimate. If anything, “Mistake” praises Aristotle for correctly seeing that duty is underivative.
But this was not at all Prichard's later view. His notorious essay “The Meaning of Agathon in the Ethics of Aristotle”Footnote 59 ascribed to Aristotle the familiar combination of a merely instrumental “ought” and psychological egoism, and did so by arguing that when Aristotle talked of “good” (agathon) he always meant “conducive to the agent's pleasure.” On this reading, Aristotle's argument that virtue is good was really the argument that virtuous action is pleasurable for the virtuous agent,Footnote 60 which Prichard said distorts the moral phenomena just as Hobbes's argument does. Prichard considered, as an alternative, that by agathon Aristotle meant (simply) “good,” but perhaps assuming with Moore that “good” is necessarily agent-neutral, Prichard said this would commit Aristotle to holding that each person desires other people's virtue as much as his own, which Prichard said he did not: Aristotle thought each person cared only about his own virtue.Footnote 61 So Aristotle was another philosopher who reduced the moral to the nonmoral “ought.”Footnote 62
This was, to say the least, an uncharitable reading, and most scholars think it was decisively refuted by J. L. Austin.Footnote 63 They therefore reject Prichard's critique of Aristotle as based on a gross misinterpretation. But the criticism can be separated from the interpretation and, when it is, remains powerful.
Consider the following, more standard interpretation of Aristotle's Ethics. By “good” or agathon Aristotle meant what is “good for” a person, in the sense of contributing to his eudaimonia, well-being, or living a good life. This is not Moore's concept of the (simply) good, nor Prichard's non-normative concept of what will give a person pleasure. It is a normative concept but an independent one, and it grounds reasons that are only agent-relative, so that a thing's being “good for” a person gives only him ultimate reason to pursue it. Thus, when Aristotle said that virtue is “good,” he meant that it is a constituent of the life that is best for a person, and his explanation of why we ought to perform the acts conventionally regarded as duties is that, if done from the right motives, they will express virtue on our part, which is an essential constituent of the good for us.
This interpretation does not involve psychological hedonism and need not be conceptually minimalist: it can ground moral duties in claims about eudaimonia that are normative but not moral. Even so, the view it ascribes to Aristotle is open to the Prichard-style objection that it distorts the moral phenomena. Consider a paradigmatically other-regarding duty such as the duty to relieve another's pain. What, Prichard would ask, is the ultimate explanation of why we ought to fulfill this duty? Is it that doing so will make our own lives better, so that the duty is at bottom self-regarding? Or is it that doing so will make the other person's life better? On the standard reading, Aristotle's answer is that it will make our own lives better, by contributing, given the right motives, to our own eudaimonia. But that, the objection says, is not the right explanation. The right explanation is that relieving the other's pain will make her life better, so the duty is not just superficially but fundamentally other-regarding. The view standardly ascribed to Aristotle therefore gives the wrong explanation for the duty to relieve the other's pain, turning what we know intuitively is an other-regarding duty into a self-regarding one. The view is not as crudely egoistic as a hedonistic one, since it does not equate the agent's good with his pleasure. But it is egoistic at a deeper level, because it relates all his reasons for action to his own eudaimonia, which, whatever its content, must be a state of him. It is his good, not anyone else's, that is his ultimate goal, and any view that makes that true violates our intuitive sense that the ultimate point of other-regarding acts is to benefit others.
The philosophers of Prichard's period, though familiar with ancient ethics, thought its approach to moral questions was fundamentally misguided because it involved confusions about the basic normative concepts. Sidgwick identified the key issue when he said that the ancient philosophers did not distinguish between the questions “What ought I to do, all things considered?” and “What will make my life go best?”—or assumed without argument that the answer to these questions must always be the same.Footnote 64 This forced them into implausible claims: for example, that if it is, all things considered, right for me to sacrifice my life in battle, the value that courageous act will contribute to my life must be greater than the value of the hundreds of other virtuous acts I could perform if I lived, so my act involves no real sacrifice on my part.Footnote 65 But the ancient philosophers' approach also gave their theories a disturbingly egoistic cast, not only at the explanatory level Prichard emphasized but also in their more specific claims. Sidgwick said that Aristotle did not have the concept of benevolence, since the virtue on his list closest to it, liberality, is shown just as much in tasteful expenditure on a fine house for oneself as in spending on other people.Footnote 66 Others were repelled by Aristotle's account of the megalopsychos or “proud” man, who enjoys being more virtuous than other people—he is competitive about virtue—and will not give others small benefits, since that is beneath his dignity; only when great things are at stake will he deign to act. Rashdall noted “Aristotle's revolting picture of the high-souled man (megalopsychos),” adding “[o]f course I am aware of the explanations by which all superior people are accustomed to defend the Aristotelian ideal.”Footnote 67 And Ross generalized the criticism when he said that Aristotle's description of the megalopsychos “betrays somewhat nakedly the self-absorption which is the bad side of Aristotle's ethics.”Footnote 68
There are, in fact, many points where Aristotle's ideal agent seems unattractively self-concerned. Aristotle held that a virtuous agent takes pleasure in his virtuous acts; he may also have thought that virtuous acts have their full value only when their goal is successfully achieved, as when an attempt to relieve another's pain does relieve it. So his virtuous agent may take pleasure in objects that include, as an effect of his act, some benefit or good for another.Footnote 69 But surely a virtuous person will be equally or almost equally pleased when another's pain is relieved independently of his own action, as when someone else like a doctor relieves the pain or it goes away by itself. Yet nowhere in his main discussion of virtue in Books II through IV of the Nicomachean Ethics did Aristotle say this; he did not suggest that a virtuous person takes pleasure in states of others that are unconnected to his own acts.Footnote 70 And it is hard to see how he could say this, when only a person's own acts and their effects are appropriately connected to his eudaimonia and can therefore from his point of view be good.
Or consider Aristotle's claim that a virtuous person will sacrifice for a friend and even let the friend perform noble acts rather than do so himself. These acts sound altruistic, but look at Aristotle's justification for them. When a virtuous person gives his friend wealth, he himself “achieves nobility” and therefore assigns “the greater good to himself.” He may let his friend rescue a person from drowning, but the reason is that “it may be nobler to become the cause of his friend's acting than to act himself,” so he still assigns “to himself the greater share in what is noble.” However much his acts benefit others, the reason why he ought to perform them, and even his motive for doing so, is that they will increase his own eudaimonia. It follows that if two friends understand their situation, they will engage in an unseemly competition to act more virtuously. After all, if it is nobler for A to let B save the drowning person, it is also nobler for B to let A do the saving. So we can imagine an Alphonse-and-Gaston routine in which each tries to let the other do the saving—“You do it,” “No, you do it,” “No, you do it”—so he can have “the greater share in what is noble.”Footnote 71 Egoism like this again seems inevitable in a view that grounds all normative demands in an agent's own eudaimonia; and that that is the wrong grounding was most emphasized by Prichard.
Again, however, some may argue that Prichard's objection applies only to eudaimonist theories and not to others that try to justify moral duties on some deeper ground. But in fact the objection is more widely applicable. Though I cannot argue the point here, many of the most influential recent attempts to justify moral duties—R. M. Hare's universalization argument for utilitarianism, John Rawls's veil-of-ignorance argument for his principles of justice, and Christine Korsgaard's appeal to “practical identities”—start with a picture of the moral agent as essentially self-concerned. The agent cares about the fulfillment of his preferences because they are his, or is concerned that social institutions allow him to promote his conception of the good whether or not it is true, or cares about acting from his practical identity. And no justification of moral duty with that kind of starting-point can do other than distort the moral phenomena. The objection must in each case be tailored to the details of the proposed justification, but its core goes back to Prichard.
VII. Conclusion
Many philosophers in the history of ethics and today have wanted to justify moral duties by deriving them from something other than moral duty, either from non-normative facts or from a normative consideration that is not itself moral. A sequence of British philosophers of the late nineteenth and early twentieth centuries rejected this program, holding both that normative truths in general are underivative from non-normative ones, and that more narrowly moral duties are also underivative. This is an important position in the history of the subject and a challenge to some influential projects in philosophical ethics today. But the philosopher who defended the position most fully was Prichard, with his three-stage argument, and he was in particular most explicit in arguing that the attempt to justify moral duties on some other ground not only fails to yield the desired results but distorts the moral phenomena. That makes him, on this issue, a central figure in the history of the subject.
