II. WHEN IS A RULE-CONSEQUENTIALIST NOT A CONSEQUENTIALIST?

We can define a method of evaluation as consequentialist if it involves appraising evaluands based on how well they promote the good, that is, how good their consequences would be. A theory for the evaluation of actions that appraises each individual action based exclusively on how well it would promote the good can be described as directly consequentialist. A theory for the evaluation of actions that appraises actions by reference to some other standard, where that other standard is in turn determined via consequentialist evaluation, can then be described as indirectly consequentialist.

In the consequentialism literature, ‘the good’ typically refers to the totality of those things with (agent-neutral) intrinsic value; something promotes the good to the extent that it brings more of these things into existence.Footnote ⁵ And this is the source of the problem. Hooker's moral theory is obviously not directly consequentialist, but at first glance it would appear to be indirectly consequentialist, in virtue of determining the moral standing of actions by reference to the code of moral rules whose widespread internalization would have the best consequences. His ‘theory of the good’ would appear to be a mildly prioritarian welfarism.

If Hooker were to adopt HPA, however, then these appearances would turn out to be deceiving. His theory of the good would then include the virtues, but his moral theory would take no account of how much the widespread internalization of different moral codes would contribute (or be expected to contribute) to people's acquisition of the virtues when evaluating moral codes. It would consider only expected effects on aggregate well-being. Hence it would evaluate moral codes based on how well they promoted part of the good but not the good per se, i.e. to what extent they brought into being some but not all of the things with intrinsic value. Thus it would not be consequentialist, indirectly or otherwise. While it is possible in principle that the code whose widespread internalization would have the best consequences with respect to well-being could be the same code whose widespread internalization would have the best consequences all things considered, this would be entirely coincidental.

Why would it matter if Hooker's theory turned out not to be consequentialist? One reason that it might matter is that the principle that we are obligated to obey, ‘the code whose collective internalization has the best consequences’, as Hooker summarizes his rule-consequentialism,Footnote ⁶ is itself an attractive general idea about morality. If it turns out that Hooker's moral theory does not really embody this idea then this may detract from its plausibility. I will call this the ‘loss of support’ problem. Another reason is that there is no readily apparent justification for HPA's taking the intrinsic value of well-being (and possibly of some property of its distribution) into account when evaluating moral codes but not the intrinsic value of virtue. To consider the value of the one but not the value of the other appears arbitrary. If there is an underlying principle of selection at work here then Hooker does not make it explicit, let alone present a rationale for it. I will call this the ‘arbitrariness’ problem.

It may appear that Hooker could address both of these problems by making a comparatively minor amendment to part (b) of HPA, merely revising it to say that prospective moral codes are to be evaluated on the basis of how well they promote the good as a whole, including the cultivation of the virtues. This would indeed remove the worry about Hooker's moral theory not being consequentialist if he adopted HPA. However, revising HPA in this way could lead to rules’ being ‘bootstrapped’ into the ideal code. When calculating how good the consequences of a code's widespread internalization would be, we would need to assume that the dispositions essential to accepting it are virtues. But if the virtues are intrinsically valuable and adding rules to a code increases the number of dispositions that are essential to accepting it, then the widespread internalization of lengthier codes would better promote the good, ceteris paribus, since a world in which most people had internalized these codes would contain more virtues. Rules not useful enough to have made it into the ideal code on the original formulation of (b) would now make the cut simply because adding them would make the code longer. (Of course, Hooker cannot cite the fact that taking the virtues’ value into account when settling on the ideal code leads to this bootstrapping problem as his only reason for not taking it into account; that would be ad hoc.)

Admittedly, even with the addition of HPA Hooker's moral theory would still identify the ideal code with the code whose widespread internalization would have the consequences that were best in a particular respect. Some might argue that his theory would still deserve to be called consequentialist and that my definition of ‘consequentialism’ is therefore too restrictive.Footnote ⁷ However, the loss of support and arbitrariness problems really both stem from the fact that if HPA were to be incorporated into Hooker's moral theory then the theory would no longer evaluate moral codes based on how well they promote the good in its entirety, i.e. how good the consequences of their widespread internalization would be all things considered. Saying that the theory would no longer be consequentialist is no more than a pithier way of saying this. I could restate my argument in this section without using the term ‘consequentialist’ at all, so even if my definition of ‘consequentialism’ is too narrow, the arguments would still stand.

III. HOOKER'S USE OF REFLECTIVE EQUILIBRIUM

As it happens, however, it is not clear that either the loss of support problem or the arbitrariness problem gives Hooker himself a decisive reason not to adopt HPA, even though they might do so for other rule-consequentialists. To understand why, it is necessary to know something about Hooker's strategy for arguing for rule-consequentialism.

Hooker ‘endorses a reflective equilibrium account of theory selection in ethics’.Footnote ⁸ He acknowledges that he is aptly described as an intuitionist in so far as he is looking for a moral theory ‘that matches and makes sense of our moral convictions (or “intuitions” in some metaphysically and epistemologically neutral sense of the term)’.Footnote ⁹ As this passage suggests, Hooker's case for his rule-consequentialism consists mainly of arguments for why it fits our widely shared pre-theoretical moral convictions, i.e. our initial ‘considered moral judgements’, more closely than do competing theories. The one exception to this rule concerns the theory that Hooker calls ‘Ross-style pluralism’, which says that there are a plurality of moral principles that generate prima facie duties. Hooker acknowledges that Ross-style pluralism fits our considered moral judgements as closely as his rule-consequentialism does, but in this instance he contends that his theory is superior on formal grounds, e.g. it offers a unifying explanation for why we have the obligations that we do and it gives us moral guidance on a wider range of issues, including ones on which we have no widely shared considered moral judgements.Footnote ¹⁰ In defending his theory, in short, Hooker restricts himself to arguing that it does the best job of ‘economically systematizing’ our pre-theoretical moral convictions.Footnote ¹¹

There are two versions of the method of reflective equilibrium: the ‘narrow’ and the ‘wide’. Both direct us to choose between theories of morality based on how closely they cohere with other of our commitments and on the extent to which they possess formal virtues like unity and robustness. Both also tell us that we should sometimes be willing to revise some of our other commitments when this would allow us to settle on an otherwise superior moral theory. But when following the narrow method we are only concerned with how closely our moral theory coheres with our considered moral judgements, and this makes the method conservative with respect to these judgements. We may make limited revisions to our initial considered moral judgements or pre-theoretical moral convictions when we are following this method in order to arrive at a moral theory that is more attractive from a formal standpoint, but these judgements will remain largely intact. When we follow the wide method, in contrast, we aim to arrive at a moral theory that coheres with both our considered moral judgements and our ‘background theories’, which include any philosophical or empirical theories that might bear on the selection of a moral theory. This method presupposes that background theories ought to offer positive support for the moral theory (and perhaps vice versa). In the process of settling on a moral theory we may make selective revisions to both our initial considered moral judgements and our initial background theories. The greater our confidence in our background theories, the more willing we will be to revise our pre-theoretical moral convictions. The wide method can therefore be expected to require more extensive revisions to our initial considered moral judgements than the narrow method.Footnote ¹²

The foregoing description of Hooker's arguments may make it sound as if he favours the method of narrow reflective equilibrium, but in response to criticism on this point he has clarified that in principle his allegiance is to the wide method.Footnote ¹³ He denies that there is any difference between these methods in practice, however, at least in terms of which moral theory we will settle on if we follow them. Terming what I have been calling background theories ‘background conditions’, Hooker claims that in practice these two methods come to the same thing.

This claim is much stronger than it may initially appear. Hooker is not asserting only that the subject areas that the background theories address are at least as controversial as morality and that every plausible moral theory coheres closely with some set of plausible background theories. Rather, he is claiming that every plausible moral theory coheres equally well with every plausible background theory.Footnote ¹⁵ This is a bold claim, given that most of the history of moral philosophy consists of attempts to show that particular background theories favour or disfavour particular moral theories. (It is telling that in the last passage I quoted Hooker equates the coherence that the wide method requires between our background theories and our moral theory with mere consistency.Footnote ¹⁶ No doubt the leading moral theories are all logically consistent with all of the plausible background theories, since their subject matters are different, but coherence is a stronger relation than consistency.Footnote ¹⁷ Someone who fails to notice Hooker's conflation of these notions might fail to be struck by just how strong his claim is.Footnote ¹⁸)

In the previous section I argued that if Hooker were to adopt HPA then strictly speaking his moral theory would no longer be a consequentialist one. I suggested that this could present him with the loss of support problem and the arbitrariness problem. I also showed that attempting to address these problems by modifying part (b) of HPA so that it would take the value of the virtues into account when evaluating moral codes could lead to rules’ being bootstrapped into the ideal code.

HPA's claim in part (d) that virtue has intrinsic value is the root of these problems. This claim is too central to HPA for dropping it to be considered a mere amendment; rejecting the claim would be tantamount to rejecting the account. Still, in so far as Hooker argues for rule-consequentialism strictly on the grounds that it does the best job of economically systematizing our considered moral judgements, it is possible that neither of these problems gives him a decisive reason to reject the claim that virtue has intrinsic value. This is especially clear where the arbitrariness problem is concerned. If a moral code whose widespread internalization would produce the best consequences relative to a part but not the whole of the good is the one that best fits our considered moral judgements, then from Hooker's perspective it is not arbitrary to regard that code as the code by which the moral standing of our actions is to be judged.

Things are slightly more complicated where the loss of support problem is concerned. Hooker notes that his moral theory ‘invokes the familiar idea that we should try to live by the moral code whose communal acceptance would, as far as we can tell, have the best consequences, impartially considered’, and he does adduce the attractiveness of this idea as a factor that counts in the theory's favour.Footnote ¹⁹ All of the leading moral theories, he observes, ‘invoke’ or ‘develop’ similarly familiar and attractive general ideas about morality, and if two theories were tied in terms of how well they satisfied all other desiderata for moral theories, ‘but only one of these theories offered an account that starts with familiar and attractive ideas about morality, this theory would seem markedly superior’.Footnote ²⁰ Hooker is not entirely explicit about how familiar and attractive general ideas about morality factor into his use of reflective equilibrium, but he seems to regard such ideas not as background theories but rather as a special sort of considered moral judgement. If Hooker were to adopt HPA, then his moral theory would not invoke or develop the same familiar and attractive idea, since the ideal code will not necessarily be the code whose widespread internalization would have the best consequences. Thus the theory would lose whatever support this factor lends it. Still, Hooker might say that since his moral theory is superior to its rivals in other respects, i.e. in terms of its degree of fit with our considered moral judgements and the extent to which it possesses the formal virtues of moral theories, it does not need to start from any especially attractive general ideas about morality in order to emerge as the best-justified theory. He might also, of course, appeal to some other familiar and attractive general idea about morality that the theory still could be said to develop or invoke.

So the argument in the prior section may give rule-consequentialists who offer more traditional philosophical arguments for the theory decisive reasons to repudiate HPA without giving such reasons to Hooker himself. Of course, if he committed himself to this account then it would be inappropriate for him to continue to describe his theory as an example of rule-consequentialism, at least on one common definition of ‘consequentialism’. And his doing so would mean that he would owe us an argument for the claim that the moral code whose collective internalization would in fact produce the best consequences so far as its effects on well-being are concerned would in fact better systematize our considered moral judgements than the one whose collective internalization has the best consequences when the intrinsic value of the virtues is also taken into account, for he never makes this comparison in Ideal Code, Real World. But these might be prices that he is willing to pay.

IV. INCONSISTENT INTUITIONISM

There is another consideration, however, that I do take to provide Hooker himself with a decisive reason to reject HPA. This is that embracing it would commit him to inconsistent views about the justification of ethical theories, a term that I will use to encompass both deontic theories of morality (i.e. theories of right and wrong action) and theories of virtue. Hooker's distinctive use of the method of reflective equilibrium does not allow him to avoid the force of this argument; on the contrary, this is precisely what gives the argument its force. Given his approach to the justification of moral theories, we might expect Hooker to say that the best theory of virtue is the one that does the best job of economically systematizing our pre-theoretical convictions about virtue. Yet HPA gives our ‘considered aretaic judgements’ no role whatsoever in justifying a theory of virtue. According to this account the best theory of the virtue is derived from Hooker's rule-consequentialism, and he settles on rule-consequentialism because it best systematizes our considered moral judgements. So according to HPA our considered moral judgements drive even aretaic theory selection.

What is more, while we do not know much about the theory of virtue that HPA would yield, we know enough to say that it would conflict rather sharply with some common pre-theoretical convictions about virtue. Rule-consequentialists commonly understand what it is to accept or internalize rules in terms of having one's conscience ‘enforce’ those rules.Footnote ²¹ A person who has internalized certain rules will feel guilty about having violated them, and in so far as her conscience is active while she is deliberating prior to acting she will have to break through what Mill describes as ‘a mass of feeling’ in order to choose an action contrary to those rules.Footnote ²² She may also blame others for violating those rules and feel resentment towards them. This seems to be how Hooker in particular understands the notion of internalizing rules, at least; he writes, for example, that ‘The collective internalization of a code amounts to the establishment of a shared conscience.’Footnote ²³ For Hooker, therefore, being virtuous is having, or at least it is part of having, an ideal conscience.Footnote ²⁴

Yet not only do philosophers who work in the field of ‘virtue ethics’ tend not to understand the virtues as dispositions that are necessary for an ideal conscience, neither does this understanding typically fit the aretaic convictions of people who are not already in the grip of a philosophical theory of virtue. We can see this from the case of Huckleberry Finn. Huck is widely esteemed for his virtue, but not because his conscience was ideal. Indeed, it was so far from ideal as to hinder his doing the morally right thing, although he did the right thing nonetheless by dint of other elements of his character.Footnote ²⁵ We can also see it from the fact that when people do things that we do consider morally obligatory, we sometimes judge them to be less virtuous if their actions were motivated by conscience. One well-known example of this comes from Michael Stocker's paper ‘The Schizophrenia of Modern Ethical Theories’, where a hospital patient is initially pleased to receive a visitor but then learns with dismay that the visitor is motivated exclusively by a feeling of obligation.Footnote ²⁶ Surely this ‘cold-hearted benefactor’ is no paragon of virtue.Footnote ²⁷

So while Hooker rests moral theory selection almost entirely upon closeness of fit with our pre-theoretical convictions about morality, HPA gives our pre-theoretical convictions about virtue no role in aretaic theory selection and even yields a theory of virtue certain to be thoroughly at odds with some of these convictions. Thus if Hooker were to adopt HPA there would be a considerable disparity between what he takes to justify these two types of ethical theories, and in the absence of some rationale for this disparity we would have to regard it as an inconsistency. He offers no rationale for it, however. And what rationale could he offer? There is no denying that we have pre-theoretical convictions about virtue, and the same reasoning that Hooker gives for the claim that their ability to economically systematize our pre-theoretical convictions is the only consideration that can help us choose between moral theories gives just as much support to the claim that this same ability is the only consideration that can help us choose between theories of virtue. Nor is it any help for Hooker to say that there is another consideration that might help us choose between theories of virtue, namely closeness of fit with the most strongly justified moral theory. One could just as easily say that closeness of fit with the most strongly justified theory of virtue is a further consideration that can be used to help us choose between moral theories.

A relatively minor modification to HPA, weakening part (c) to say that the virtues include the dispositions that are essential parts of internalizing the ideal code, as opposed to being identical with those dispositions, would go some distance towards resolving the outright conflicts that I noted between Hooker's theory of virtue and widely shared pre-theoretical convictions about virtue, since these conflicts stem from the fact that the conception of virtue that Hooker advances in Ideal Code, Real World is so straitened. This move will not satisfy someone who maintains that dispositions that are essential parts of internalizing a moral code can never count as virtues, i.e. that to call such a disposition a virtue is always a category error, but it should be enough to satisfy someone who holds only that some of the virtues are something other than dispositions involved in internalizing a moral code.

Making this move would leave Hooker in need of some way to determine what the other virtues are. The solution to this problem, though, is near to hand. While weakening (c) in the way that I described above would render Hooker's account more consistent with our pre-theoretical considered aretaic judgements, the only way for him to eliminate the inconsistency between his approach to the justification of aretaic theories and his approach to the justification of moral theories is to relent and assign our pre-theoretical convictions about the virtues some weight in aretaic theory selection. Again, this runs so counter to the spirit of HPA that it cannot be considered a mere revision to that account; making this move is tantamount to rejecting the account. It would, though, solve two problems at once. First, it would eliminate the current inconsistency in Hooker's approaches to justifying different types of ethical theories. Second, he could use coherence with these convictions as a basis for generating a more comprehensive list of the virtues, one that goes beyond the dispositions involved in internalizing a moral code.

Notice that Hooker's taking this step might necessitate a major shift in his methodology. If we have grounds for favouring one theory of virtue over another that are independent of their degree of fit with a moral theory, such as their degree of fit with our considered aretaic judgements, then we would have to include the theory of virtue that we take to be most strongly justified on these grounds among our background theories when using the wide method of reflective equilibrium in moral theory selection. In that case, even Hooker would have to concede that there is a practical difference between the narrow and wide methods of reflective equilibrium, since a given theory of virtue would almost certainly cohere more closely with some moral theories than others.