I

Most of us agree that morality imposes a set of duties that have at least pro tanto force. W. D. Ross suggests duties of fidelity, reparation, non-maleficence, beneficence, gratitude, justice and self-improvement.Footnote ¹ These duties specify factors that reasonable persons recognize as relevant to the rightness of actions, despite our many disagreements about particular cases, questions of relative weight and theoretical matters.

In light of this modest consensus, many philosophers have wondered whether these duties comprise an ‘unconnected heap’,Footnote ² or whether it is possible to tie the duties together – to offer a unifying account of our moral obligations. Should we try to offer such an account if we can? Why might we want to do so? What form, if any, should a unifying account take?

Philosophers remain deeply divided on these central questions. Some believe that unification is both possible and desirable, while others deny its possibility or desirability. Principle monists believe that the list of duties can be derived from, or justified in terms of, one ultimate moral principle. Principle pluralists disagree.Footnote ³

Kant was a principle monist. So are some contemporary deontologists,Footnote ⁴ along with contractualists,Footnote ⁵ ethical egoists, and others.Footnote ⁶ The standard definition of ‘consequentialism’ entails that consequentialists, too, are principle monists. According to the conventional definition, consequentialist theories evaluate their respective evaluands (acts, rules, motives, etc.) exclusively in terms of the tendency of the evaluand to promote good states of affairs. This single criterion precludes a plurality of ultimate principles.

Consequentialists have been derided for their principle monism since at least the days of Ross and G. E. Moore. A familiar objection to monism challenges the very possibility of unifying the list of pro tanto duties in terms of a single principle. Many pluralists claim that no single principle is consistent with the full list of duties to which we are pretheoretically committed.Footnote ⁷ In a series of papers and his recent book, Ideal Code, Real World, Brad Hooker challenges this pessimism.Footnote ⁸ Employing the method of reflective equilibrium, he tries to unify the set of basic moral duties in terms of a distinctive rule-consequentialist principle.

Hooker's monism has drawn powerful criticisms from pluralists such as Berys Gaut, Phillip Montague, Philip Stratton-Lake and Alan Thomas.Footnote ⁹ Interestingly, however, most of Hooker's critics do not insist that his ultimate principle actually conflicts with the common list of pro tanto duties. Rather, they claim that his principle is otiose. Hooker's principle, they contend, does no justificatory or explanatory work once we have arrived at our list of pro tanto duties.

The charge that Hooker's consequentialism is otiose complements another charge often made against consequentialism – that consequentialists fetishize simplicity. Many philosophers, some of them consequentialists, have noted that the relative simplicity of consequentialism accounts for much of its initial appeal.Footnote ¹⁰ Contemporary pluralists often attempt to discredit monists’ preference for simplicity by psychologizing the preference – by emphasizing the monist's presumed psychological motivations. Susan Wolf observes that ‘the deep human longing for simplicity, completeness and, most of all, uniqueness, moves us, wrongly, to accept [monistic theories] as the whole moral truth.’Footnote ¹¹ Alan Thomas, whose critique of Hooker figures prominently in this article, describes simplicity as a ‘quasi-aesthetic preference’, one which, he implies, should carry little weight in metaethics.Footnote ¹² Other pluralists suggest that consequentialists suffer from a jealous desire to mimic the natural sciences or an excessive commitment to abstract canons of theory choice derived from scientific epistemology. Joel J. Kupperman writes that: ‘The inspiration for act utilitarianism of course is science. Act utilitarianism. . . is designed to achieve (or parody) the extreme generality and simplicity of scientific theory at the highest level: to provide for ethics what a unified field theory would provide for physics.’Footnote ¹³

Surprisingly, some monists do not even try to resist these psychologizing moves. Rather than trying to justify their preference for simplicity they simply attribute it to reasons of ‘personal temperament’.Footnote ¹⁴ This is a substantial concession to the pluralists that no monist should make without a fight.

In this article, I shall defend principle monists against these pluralist objections, using Hooker as my model. I make two related points. First, I argue that the objections rely upon assumptions that conflict with premises to which pluralists themselves must appeal in order to defend their own positions against challenges from moral particularists, such as Jonathan Dancy. Second, I argue that abandoning these assumptions allows us to appreciate certain advantages that monism has over pluralism, virtues that are analogous to the advantages that generalism has over particularism. These include pragmatic virtues that derive from the intensional content of a moral principle. I call these intensional-pragmatic virtues. Pluralists rarely recognize these virtues. My arguments go beyond anything Hooker has offered in his own defense.Footnote ¹⁵

II

I shall highlight some features of Hooker's approach before considering the critiques. Hooker's position typifies broader trends in consequentialist theorizing. Since Bentham, consequentialists have been softening their opposition to axiological pluralism.Footnote ¹⁶ Hooker, himself, leans toward a pluralistic, objective-list theory of well-being.Footnote ¹⁷ But the issue of axiological pluralism is orthogonal to the controversy over unifying the list of pro tanto duties. One can be an axiological pluralist and a principle monist, or an axiological monist and a principle pluralist, or what have you. As a consequentialist, Hooker remains a principle monist, by definition.

Hooker also agrees with Ross, John Rawls and many others that a moral theory should aim to fit our considered moral convictions. He accepts, more or less, Ross's list of pro tanto duties, which reflect these convictions.Footnote ¹⁸ Whereas Ross believes that moral theory cannot offer unification beyond what Ross himself proposes, without coming into conflict with our convictions, Hooker believes further unification to be both desirable and possible within reflective equilibrium, without introducing such conflict.Footnote ¹⁹ He offers an original rule-consequentialist principle as the unifying norm:

This principle justifies the pro tanto duties, according to Hooker. It offers the theoretical advantages of greater unity, coherence and integration. It also offers the practical advantage of helping us resolve moral controversies more effectively than the list of duties does on its own, he claims.

III

I now consider some criticisms of Hooker. Stratton-Lake argues that Hooker misunderstands Ross's concept of a prima facie duty (what I am calling a ‘pro tanto duty’). According to Stratton-Lake, Hooker neglects Philippa Foot's distinction between verdictive and evidential moral considerations.Footnote ²¹ Verdictive considerations are general moral judgments, such as ‘Breaking promises is wrong’. Evidential considerations, by contrast, are what Stratton-Lake characterizes as ‘principles of moral salience’. These identify basic reason-giving considerations, or ‘right-making characteristics’, such as the fact that a promise was made. According to Stratton-Lake, Hooker understands pro tanto duties as general verdictive considerations, whereas Ross intends them as evidential considerations.Footnote ²² Stratton-Lake claims that Hooker misunderstands Ross and that this misunderstanding leads Hooker to conclude that the list of Rossian duties is not normatively basic. According to Stratton-Lake, Hooker thinks it both desirable and possible to justify this list in terms of something else and Hooker argues that his rule-consequentialist principle can serve this purpose. However, Stratton-Lake objects, Hooker's principle cannot justify the Rossian duties without begging the question. Hooker's rule-consequentialist principle itself ‘presupposes that certain evidential moral considerations are salient to deliberation and action, namely, considerations of well-being and fairness’.Footnote ²³ Verdictive considerations are not morally basic, whereas evidential considerations are, according to Stratton-Lake. That is why Hooker's rule-consequentialist principle must rest on the latter, rather than justifying them, he concludes.

In a brief reply to Stratton-Lake, Hooker observes that ‘[t]o internalise a rule is, among other things, to come to take certain considerations as salient. . . So rule-consequentialism, in selecting rules, selects a set of evidential considerations.’Footnote ²⁴ So Hooker's position is that rule-consequentialism can justify not just moral verdicts, but evidential considerations, though it justifies the latter indirectly.

It seems to me that Hooker and Stratton-Lake talk past one another. Of course, a unifying principle that takes the promotion of welfare as a non-derivative goal must, in some sense, appeal to a principle of moral salience which identifies the fact that welfare is increased as a non-derivative right-making characteristic. But it is a further question whether that unifying principle will direct moral deliberators to internalize rules such that, once they have internalized them, they will themselves treat facts about welfare as salient in deliberation. The same goes for facts about fidelity or reparation or self-improvement, to cite some other Rossian duties. Hooker argues that his rule-consequentialist principle has just this effect, among others. This is not a question-begging derivation, though it may be hard to see why without considering unrealistic counterfactuals, the full relevance of which will emerge later in this article. For now, we can begin to see the import of unrealistic counterfactuals by considering a possible world in which it would not maximize welfare for moral deliberators to internalize rules such that they themselves treated welfare facts as morally salient. Consider a world in which most human beings are poor judges of human interests. They mistakenly believe that all other human beings enjoy having ice water unexpectedly poured on their heads. We cannot convince them otherwise, try as we might. If such people internalize rules that lead them to treat welfare as morally salient, they will often seize the common opportunity to toss ice water on one another, mistakenly believing this to be an act of kindness. If the population suffers from enough misconceptions of this sort, then at some point it would not promote the general welfare for them to internalize such rules, rather than rules that do not lead them to treat welfare as morally salient. People can still survive if they ignore the welfare of one another, albeit less happily than if they practice beneficence properly. So, at some point, the advantages of having people treat welfare as morally salient are outweighed by the disadvantages, in the imagined world.Footnote ²⁵

If the world I have described is physically possible, then it is but a contingent fact that, in the actual world, with human nature what it is, it promotes the good for people to internalize rules such that they treat welfare as morally salient. In the actual world, human beings happen to be sufficiently accurate judges of one another's welfare that internalizing such rules has a positive net effect. This is a happy contingency, of course, but the fact that it is a contingency does not detract from Hooker's claim that his rule-consequentialist principle supports a list of Rossian duties in the actual world. It does not beg the question, as Stratton-Lake concludes, because it is not a matter of conceptual or even physical necessity that a principle that ultimately appeals to welfare facts as non-derivative right-making characteristics will support rules such that, when we internalize them, we tend to treat welfare facts as evidential moral considerations.

Indeed, Hooker's project depends, more fundamentally than has been recognized, on its culminating in a contingent defense of the correct unifying moral principle, a point that becomes relevant again in section V.Footnote ²⁶ Only because he defends rule-consequentialism as contingently, rather than necessarily, correct can Hooker do what his critics consider impossible. At first appearance, Hooker's strategy does look question-begging. He offers a unifying account that simultaneously supports the Rossian duties while depending on the same evidential considerations which those duties reflect. Hooker's account depends on these evidential considerations as considered convictions in reflective equilibrium. An account that presented both the unifying principle and the Rossian duties themselves as necessarily valid would, indeed, beg its questions. I shall explain this point briefly, switching to the formal mode for the sake of clarity. If it is necessarily true that promise-keeping has deliberative moral salience and also necessarily true that rule-consequentialism is the correct unifying principle, as Hooker intends this, then it is conceptually true that the former follows from the latter and vice versa. In that case, one could add nothing to the epistemic status of either a Rossian duty or the rule-consequentialist principle by placing them into (what would then be a trivial) ‘reflective equilibrium’ with one another.

Some of Hooker's critics explicitly avow these modal assumptions. Montague, for example, assumes that our considered convictions must apply, if not to all possible worlds, then at least to certain possible worlds that are extremely different from the actual world. Montague also assumes that our considered convictions apply to the Rossian duties, not just to concrete situations.Footnote ²⁷ We have considered moral convictions at varying levels of abstraction.Footnote ²⁸ Some of these concern concrete situations, while others concern general moral concepts. On Montague's conception of reflective equilibrium, an acceptable theory must match our considered convictions about not only concrete situations (e.g., ‘the My Lai massacre was barbaric’) but also the Rossian duties, themselves (e.g., ‘promise-keeping constitutes an evidential moral consideration’). As Montague writes: ‘moral theories are tested against the Rossian rules,. . . the Rossian rules are more basic than any theory, and. . . if a choice must be made between a Rossian rule and some theory (rule consequentialism, say), then the theory should be jettisoned’.Footnote ²⁹

For example, on Montague's view, if our considered convictions in the actual world indicate that promise-keeping is a Rossian duty, then an adequate moral theory must confirm that promise-keeping is a duty in any (or at least many) possible worlds.

Hooker replies to Montague persuasively and I shall not add to his reply here.Footnote ³⁰ I summarize Montague's critique only to reinforce my claim that Hooker's approach looks question-begging to critics only because they assume that our convictions apply rigidly across a very wide range of possible worlds and apply to general moral duties, not just to concrete situations.

IV

I turn, now, to what I consider the most sophisticated objection yet made against Hooker's monism. Alan Thomas argues that Hooker's rule-consequentialist principle is an ‘idle wheel’ once we commit ourselves to the Rossian list of duties. Thomas applauds Hooker for abandoning what Thomas calls epistemological realism: ‘the view that our ethical beliefs fall into epistemic classes. . . or that within a class our beliefs stand in determinate relations of epistemic priority or subordination’.Footnote ³¹ He proceeds to argue that abandoning epistemological realism creates three problems for Hooker, two of which I shall now discuss. The third I shall postpone until section VI.

First, Thomas argues that Hooker's disavowal of epistemological realism prevents Hooker from explaining the entailment of the Rossian list by the rule-consequentialist principle. This is because ‘the underlying consequentialist assumption stands in a symmetrical relation of mutual support to the principles of the list and thus both its normative content and its explanatory power must co-vary with that of the initial list.’Footnote ³²

Second, Thomas argues that Hooker's rejection of epistemological realism creates ‘an internal tension between the claim that Hooker's position is equally explanatory to that of the pluralist and [the claim] that the underlying consequentialist principle offers normative grounding for the list’.Footnote ³³ Thomas asks us to imagine that ‘the [Rossian] list is revised in light of a change in moral phenomenology, to preserve its claim of explanatory adequacy’. If Hooker's rule-consequentialist principle changes as well, Thomas insists, then ‘the principle is the underlying normative ground of the list in a sense which makes it epistemically prior to it’.Footnote ³⁴

This second argument of Thomas's is highly compressed, but I shall attempt to reconstruct it. He appears to assume that if we observed the rule-consequentialist principle to change in tandem with changes in the Rossian list, reason would oblige us to infer that the Rossian list was actually changing because of the change in the rule-consequentialist principle, ‘tracking’ the latter.Footnote ³⁵ In other words, reason would require us to conclude that the rule-consequentialist principle fell into a prior epistemic class from the list, and did so in virtue of its content, as opposed to our degree of confidence in it. Presumably, the principle would enjoy this epistemic privilege on account of the fact that its content is more abstract and general than are the Rossian duties. Given these relations of epistemic priority, the principle would function as the ‘normative ground’ of the list. This is what Thomas understands Hooker to deny when Hooker embraces reflective equilibrium. That method precludes the claim that more general and abstract principles, as such, enjoy epistemic priority over the more specific and concrete.

Thomas concludes his second objection as follows:

Thomas's objections demand a response from Hooker. Consider, first, Thomas's claim that Hooker cannot explain the entailment of the Rossian list by his rule-consequentialist principle. It is unclear what sort of explanation Thomas thinks one should desire in this case. Consider an analogy involving the epistemic relation between our commitment to a Rossian duty and our considered convictions concerning concrete scenarios. Our commitment to the Rossian duty of promise-keeping, for example, stands in a symmetrical epistemic relation to our conviction that it was wrong for Tony Blair to break a promise he made to his wife on Christmas Day, 2005. Each supports the other and their relation cannot be further explained. Yet our inability to explain the relation never prevents the pluralist from stating the Rossian duties in general terms. I suggest that the analogous impossibility of ‘explaining’ the entailment of the Rossian list by the rule-consequentialist principle should not dissuade Hooker from espousing the latter, especially in light of the virtues of that principle which I shall discuss shortly.

Second, I think that one could explain why the rule-consequentialist principle would co-vary with moral phenomenology without making the epistemological realist's mistaken assumption that the rule-consequentialist principle is epistemically prior to the Rossian list in virtue of the principle's greater abstraction and generality. The reason the rule-consequentialist principle would be different if our considered convictions were different does, indeed, have to do with the content of the principle. But the explanation of this co-variation is not that the principle is epistemically prior to the Rossian list. The principle co-varies with the Rossian list because the two are coextensive. Hooker maintains that his principle requires and prohibits exactly the actions required and prohibited, respectively, by the Rossian list.Footnote ³⁷ Thomas does not challenge Hooker on this score, nor do most of Hooker's other critics.

Thomas could, however, object that my initial responses miss the force of his objections. Thomas argues that, if Hooker's principle is coextensive with the Rossian list, then the principle is an idle wheel. Nothing I have said so far addresses this claim.

I shall not dispute the validity of Thomas’ idle-wheel argument. Instead, I shall concentrate on its major premise, which is that one moral theory cannot be superior to another, overall, if the second theory is at least as extensionally adequate as the first. In other words, the idle-wheel objection presupposes that the first theory cannot be superior to the second if the second theory dictates results that coincide with our considered moral convictions about particular cases to the same degree as do the results dictated by the first theory. Call this premise extensionality.

If extensionality is true, then monism cannot claim superiority over pluralism without claiming extensional superiority. I have two strategies for challenging extensionality, one direct and one indirect. My indirect argument concludes that pluralists are in no position to espouse extensionality in their debate with monists because they will want to deny extensionality in order to prevail in their debate with moral particularists. My direct argument identifies certain virtues that can give a principle an advantage over coextensive norms that lack these virtues. These are the aforementioned intensional-pragmatic virtues. I offer some reasons to think that a monistic principle, such as Hooker's, possesses greater intensional-pragmatic virtues than the Rossian duties possess.

V

I begin with my indirect challenge to extensionality. Pluralists are generalists, not particularists. They endorse certain general, pro tanto moral principles and hold that the wrongness of an act is a function of its deviation from one or more of these principles. For example, Rossian generalists accept the principle ‘One ought to keep one's promises’. According to Rossians, the fact that an action constitutes a promise-breaking always counts against it, morally.Footnote ³⁸

Generalists face challenges from particularists such as Jonathan Dancy, John McDowell and David McNaughton. Particularists argue that moral judgment does not depend on general moral principles. Generalists and particularists also offer divergent accounts of moral judgment. Generalists hold that we derive our judgment, in a particular case, from our knowledge of general moral norms. Particularists hold that we can scrutinize the details of each case and arrive at a quasi-perceptual judgment about its moral properties, without ever appealing to general norms.Footnote ³⁹

How can generalists respond to particularism? I suggest that generalists need not insist that it is conceptually impossible to reach moral judgments without appealing to general moral principles. They need not even insist that moral judgments derived from general principles are more accurate than those derived by particularist methods. Instead of claiming these epistemological advantages, generalists can claim that their method has pragmatic advantages over particularism. They can claim that it is easier, in the long run, in terms of the cognitive burdens we bear as finite deliberators, to categorize an act in terms of a Rossian duty, such as promise-breaking, than it is to decide whether the act is wrong without employing any such mediating categorization. In practice, if we employ general norms, then we will identify (and avoid or prevent or sanction) a greater number of genuinely wrongful acts than we otherwise would. This is not because general principles are conceptually necessary to moral judgment, or provide greater accuracy. Those are much harder issues to resolve. It is easier to show that general principles are pragmatically useful, as deliberative heuristics.

To illustrate, imagine a society the conventional morality of which includes nothing like the Rossian norm, ‘Breaking one's promise is wrong, pro tanto’. In this society, no one treats promise-breaking, per se, as presumptively wrong. Perhaps they do not separate out the category of promises from a broader category to which they belong, such as ‘predictive representations concerning one's own future conduct’. The latter category includes promises, but also non-promissory predictions upon which no one could reasonably rely, such as ‘I expect I’ll visit Paris within the decade’.

The inhabitants of this society can still draw moral distinctions between broken promises and other unfulfilled, asserted predictions of a non-promissory character. Suppose one such inhabitant, Jill, promises her colleague, Jack, that she will cover his classes the following week. She never shows up. How might Jack proceed to evaluate Jill's conduct? Ex hypothesi, he cannot appeal to a Rossian presumption against promise-breaking, as such. Instead of beginning with this useful presumption, he must proceed on a ‘case-by-case’ basis. He must try to identify certain other facts about Jill's conduct as evidential considerations supporting a verdict that she has acted wrongly.

In principle, Jack can still arrive at the correct conclusion – that Jill has acted wrongly – without relying on the Rossian duties. But arriving at this conclusion, the generalist claims, proves much more difficult for Jack than it does for us. Jack must reconstruct the presumption ‘from scratch’ every time, by surveying previous cases for similarities and differences as compared to Jill's case. This process expends Jack's limited cognitive resources. It is also unnecessary. The various cases do, in fact, have something in common that is non-trivial and codifiable: they are all cases of promising. There is no point in denying this commonality, and much to be gained by codifying it and using the codification to classify new cases as they arise.

Also note that, with all that data to interpret, it might be easier for Jack to get confused, or for Jill to persuade Jack that he had become confused. If, by contrast, Jack can simply appeal to the Rossian norm, ‘Breaking one's promise is wrong, pro tanto’, he might prove a more accurate, in addition to a more efficient, moral evaluator in the long run.

Generalism also has its pragmatic costs, of course. The generalist makes specific mistakes which he would avoid if he scrutinized every case as deeply as he would need to do if he declined to rely upon general principles. In the aggregate, the particularist probably arrives at the correct answer more often than does the generalist, but she expends more cognitive resources than the generalist expends. Conversely, the generalist probably arrives at the correct answer less often, but more efficiently, than does the particularist.

I conclude that philosophers should avoid their common tendency to focus on deliberation in a single case, rather than a series of cases. Imagine an agent with extensive cognitive resources and time to expend on a single moral question. The less generalist is the theory she uses, the more likely will she be to make an accurate decision, if she faces only one question. However, as we give her more questions to answer, holding constant her cognitive resources and time, she approaches a point at which using a more generalist theory would improve her aggregate level of accuracy. Less generalist approaches only look more appealing, across the board, if we disregard considerations of relative efficiency and accuracy, in the aggregate.

I recognize that some particularists deny that their method is, in fact, less efficient than the generalist's. They claim that ‘moral vision’ enables us to perceive the moral truth with greater accuracy and greater efficiency than does generalism.Footnote ⁴⁰ Some particularists also contend that introducing moral principles actually complicates deliberation and generates social conflict.Footnote ⁴¹ But my point is not that generalists can decisively rebut particularism, although I think they have a strong case. Rather, my point is that generalists, including pluralists, must believe that generalism enjoys certain advantages over particularism or they would not be generalists in the first place. I think this generalist commitment places a dialectical burden on pluralists to persuade us that no version of monism could enjoy analogous advantages over pluralism (assuming a version of monism that also captures our considered convictions, of course). At an abstract level, pluralism stands to monism as particularism stands to generalism. The pluralist is in the awkward position of making a generalist case against particularists, while needing to make a ‘particularist’ case against monists.

The pluralist's need to defend generalism against particularism weakens her ability to advance idle-wheel arguments against monists. If extensionality were true, then monism could not, indeed, claim superiority over pluralism without claiming extensional superiority. But if extensionality were true then the generalism/particularism debate would quickly reduce to a stalemate, as well. Generalists have not shown their theory to be extensionally superior to particularism. At most, generalists can claim a range of other advantages over particularism, such as the intensional-pragmatic virtues I have discussed. In doing so, generalists, pluralists included, clear a path for the monist to claim analogous virtues over pluralism, a path which I shall now take.

VI

Earlier we saw how using moral norms that generalize over intensional categories, such as ‘promising’, offers advantages over particularism. I suggest that monism enjoys analogous advantages over pluralism. There are, however, definite differences between the advantage that generalism enjoys over particularism and the advantage that I claim monism enjoys over pluralism. These differences are worth discussing.

First, let us distinguish between first-order and second-order moral judgment. First-order judgment includes, for example, determining the deontic status of specific actions (e.g., permissible). Second-order judgment, by contrast, involves the evaluation of general moral duties, as when one asks ‘Do we have a general moral duty to rescue imperiled strangers?’

The advantages of generalism over particularism emerge at the level of first-order moral deliberation about easy cases, such as the case of Jack and Jill, discussed earlier. These are first-order cases in that they call for moral judgment concerning particular actions. They are easy cases in that the Rossian duties render clear and unequivocal results with respect to them.

By contrast, monism probably offers no advantage over pluralism in first-order deliberation about easy cases. At least, monists need not claim any such advantage. On the contrary, in easy cases, monists deliberate just as pluralists do. (This is not to suggest that agents should derive particular moral conclusions directly from a master principle. This would be prohibitively inefficient as a first-order deliberation procedure, even in easy cases. One cannot be a rule-consequentialist, as opposed to an act-consequentialist, if one expects agents to appeal directly to first principles even in easy cases.)

However, when we consider hard cases, the pragmatic advantages of monism over pluralism begin to emerge. Specifically, these advantages appear in cases of second-order deliberation, in which one must decide whether the Rossian duties support a certain derivative duty. Imagine individuals or groups who accept the same Rossian duties as we do, but who believe that one or more of the duties supports or entails a derivative duty that is more controversial in our society than are the Rossian duties themselves. Controversial derivative duties include, for example, a duty to make a special and substantial effort to help meet the needs of one's biological parents, as such. Another example is a duty to ensure that distant future generations will have something comparable to our present supply of natural resources.

How should one evaluate the claim that one of these controversial derivative duties is supported or entailed by the Rossian duties? Perhaps one could simply focus one's attention on the Rossian duties, along with the proposed derivative duty and the relevant non-moral facts, and wait for the truth to emerge, in a kind of revelation or perceptual insight. Some monists will immediately deny that such revelations could yield moral conclusions, but they need not be so skeptical. Monists need not insist that the process of making the case for controversial duties requires one, as a matter of conceptual necessity, to invoke a monistic principle. The monist's position does not require her to make this claim any more than the generalist's position requires him to insist that it is conceptually impossible to pass first-order moral judgment on a particular individual's actions without appealing to the Rossian duties. Deep reflection on the nature of beneficence, gratitude or other duties could, perhaps, lead one to a definite conclusion as to whether the Rossian duties support certain controversial derivative duties.

An equally important question, however, is how much time, investigation and cognitive resources it takes to reach answers to such second-order questions. Pluralists should not claim to know, a priori, that their method is the most efficient, or even acceptably efficient. If a method existed that was comparably reliable, but also more efficient, then we would have reason to use that method, rather than the pluralist's. Appealing to a smaller set of master principles might be just such a method. Consider, first, the issue of efficiency. It is trivial to formulate an extremely efficient master principle. ‘Flip a coin’ or ‘defer to the opinions of the oldest person present’ are highly efficient rules. But, of course, efficiency is not the only criterion. For the method to be a contender, it must also be roughly as accurate as the pluralist's. Coin-flipping holds no promise of accuracy. Blind deference to one's elders is little better. Now consider Hooker's rule-consequentialist principle. It is less efficient than coin-flipping, to be sure. But the important question is whether Hooker's principle is still more efficient than pluralism, while offering at least approximately as much accuracy as the latter method. If so, then we have reason to prefer Hooker's monism.

Recall the generalist position that I suggested pluralists may need to take in opposition to particularism. Generalists can insist, plausibly, that the process of arriving at first-order moral judgments about an action is both acceptably accurate and more efficient if one appeals to the Rossian duties than if one proceeds as particularists recommend. Monists can make an analogous statement about the process of arriving at a second-order moral judgment concerning the claim that a controversial derivative duty is supported or entailed by the Rossian duties. Monists can claim that the process is acceptably accurate and more efficient if one appeals to a monistic principle than if one limits oneself to contemplating the Rossian duties and the proposed derivative duty.

Consider the aforementioned duty to make a special and substantial effort to help meet the needs of one's biological parents, as such. Do the Rossian duties, individually or jointly, support this duty? Begin with beneficence. It is not uncontroversial in our society that beneficence supports special duties to biological parents, as such. Nor does reflection on the concept of beneficence lead one to such a conclusion. The same holds for the Rossian duty of gratitude. It is not uncontroversial in our society that gratitude supports duties to biological parents, as such. True, biological parents ‘give life’ to their children. But it is not obvious whether this ‘gift’ by itself imposes any duty to make a special and substantial effort to assist one's biological parents in meeting their needs. We must dwell on cases of individuals whose biological parents played no part in raising them, or whose parents were abusive. Although a reasonable case can be made for the conclusion that the gift of life imposes a special duty on offspring, deriving from the Rossian duty of gratitude, a reasonable case can be made for the contrary conclusion, as well. I think the Rossian duties of fidelity and justice (and the non-Rossian duty of loyalty) are also equivocal on this question.

A monistic principle, such as Hooker's, could help one reach conclusions more efficiently than the process of meditating ever more deeply on the Rossian duties. Compared to the Rossian list, Hooker's principle is simpler and more general. These attributes promise greater efficiency of application. Consider a code of rules that imposes the aforementioned duty to one's biological parents. Hooker's principle directs us to ask whether the internalization of that code by the overwhelming majority of everyone everywhere in each new generation has maximum expected value in terms of well-being (with some priority for the worst off).

Now, answering that question may involve difficult computations. Using a smaller set of principles, or even a single principle, does not make the process trivial, or mechanical, or easy. On the contrary, in interesting cases the process remains difficult and factually intensive. I only suggest that using a single principle makes the process at least somewhat more efficient than the potentially interminable process of reflecting on the Rossian duties. Monism at least gives one some idea of which questions to ask.

Imagine two agents, Felix and Oscar. Felix follows the pluralist approach, Oscar the monistic. They consider, independently, a long series of cases that require second-order deliberation, such as the case of the proposed duty to one's biological parents. Suppose that, over time, Felix and Oscar tend to reach the same results by different means. Agreements vastly outnumber disagreements. In light of this pattern, we have reason to favor whichever method, if either, is more efficient. Suppose, further, that Oscar's monistic method proves more efficient. Under the conditions described, we would have reason to favor monism over pluralism as an approach to hard, second-order cases, just as we have reason to favor generalism over particularism as an approach to easy, first-order cases.

Of course, this is all rather speculative. I have not demonstrated, empirically, that a monist and a pluralist would, in fact, often arrive at the same answers in hard, second-order cases. Nor have I proven that, if this were the case, the monist's method would boast greater efficiency. The monist's method may require gathering more non-moral facts, if it uses a master principle such as Hooker's. There are complex empirical questions, worth investigating, about the relative degrees of consensus and levels of efficiency that characterize the two methods. My objective has been simply to respond to those pluralists, such as Thomas, who deny, on philosophical grounds, that a unifying principle such as Hooker's could offer anything beyond what is offered by a set of pro tanto duties that is extensionally adequate to our considered convictions. I suggest that monism offers intensional-pragmatic advantages over pluralism, but only when we ascend to the second-order level. Such ascent is primarily the occupation of theorists and ethics experts, and even they do not spend most of their professional time at the second-order level. To this extent, the advantages of monism, while still pragmatic in character, are easier to overlook because they are mostly advantages that accrue to theorists and ethics experts, theorizing in certain of their professional capacities. By contrast, the pragmatic advantages of generalism over particularism appear even in easy, first-order cases. These are far more common in our everyday lives. To this extent, the advantages of monism over pluralism are not as dramatic or widely relevant as are the advantages of generalism over particularism. But they need not be in order to vindicate monism. Where the advantages exist, they exist.

We are finally in a position to introduce and rebut Thomas's third argument against Hooker. Thomas argues that, without epistemological realism, Hooker cannot support his claim that his rule-consequentialist principle supplies better guidance on contentious moral questions than does the Rossian list, standing alone. But Thomas neglects the possibility that monism offers advantages over pluralism at the second-order level. I have argued that a monistic principle can, indeed, supply better guidance than the Rossian list for the purposes of second-order deliberation. Thomas's conclusion is too quick.

VII

At this point, Thomas might object that my defense of monism presupposes the epistemological realism that Hooker disavows. Once we reject epistemological realism, Thomas might object, we cannot understand the fact that Hooker's principle is simpler, more general or more abstract as giving it greater authority than our concrete judgments. Thomas could suggest that, if Hooker's principle dictates a univocal resolution to a controversy with respect to which our considered convictions remain conflicted, then perhaps it is Hooker's principle, and not our convictions, that ought to be revised, so as to reflect this conflict.

However, I think Hooker can treat the greater generality of his principle as giving it greater authority than more concrete judgments without lapsing into epistemological realism. He need not claim that his principle enjoys greater authority than do our concrete convictions, collectively. All he must claim is that the principle enjoys greater authority than does any one of our convictions, or any small group of them. If a general principle captures our firmly and widely held convictions, and dictates a clear result in a concrete second-order case about which our less confident opinions conflict, then I think we have reason to defer to the general principle. The pluralists with whom I am arguing have accepted, arguendo, that Hooker's principle captures our firmly held and widely shared convictions as well as does the Rossian list. If this is so, then I think we have reason to trust the implications of Hooker's principle in cases in which our convictions are not firmly held and/or widely shared – hard cases at the second-order level.Footnote ⁴²

Pluralists, such as Thomas, admit that Hooker's principle is coextensive with the pro tanto duties. But they want to deny that his principle adds anything non-trivial to the list of duties. I have argued that pluralists cannot consistently conjoin these claims. This is because pluralists, being generalists, are already committed to the idea that, if a principle is coextensive with a set of concrete judgments and is simpler and more general than the latter, then the principle adds something non-trivial, just in virtue of those attributes.

So, at the very least, I think the burden is on the pluralist to explain why explicit formulation and employment of a smaller set of principles does not conserve cognitive resources at the second-order level, as it does at the first-order level. At the second-order level, do we suddenly become better at seeing and responding to similarities such that an appeal to formulated principles becomes superfluous? If not, then what justifies Thomas's confidence that using fewer principles at the second-order level provides no deliberative advantages? Pluralists must answer these questions.

VIII

I have argued that two norms can be extensionally equivalent, but generalize differently over intensional terms in ways that make one the more useful to us, as cognitively limited moral reasoners. This observation invites us to re-examine common assumptions about monists’ preference for simplicity. Monists can do more than they have done to defend that preference against the psychologizing critiques mentioned in the first section. Once we understand simplicity as an intensional concept, the psychologizing critiques lose much, if not all, of their force. Favoring simplicity at the level of the Rossian duties does not constitute a merely ‘quasi-aesthetic preference’, as Thomas claims.Footnote ⁴³ Nor is a preference for simplicity just a matter of personal temperament, nor a basic human longing, nor the symptom of a desire to make ethics more like science, although it may be all of these. Rather, simplicity at the level of the pro tanto duties has tremendous practical value. Because of the simplicity of these duties, we do not have to think as long or as hard when facing a particular question. I suggest that simplicity is a virtue at the second-order level, as well, when the theorist chooses between one ultimate principle and several.

We can now see, also, why Hooker need not insist that smaller, more unified sets of principles, as such, have epistemic priority over larger sets. All he must claim is that, insofar as a smaller set of ultimate principles is easier to apply, and coextensive with a larger set, the smaller set has pragmatic priority. So Hooker can insist that his principle tracks changes in moral phenomenology without insisting that it has epistemic priority over the pluralist principles in virtue of being more unified, as an epistemological realist might claim. His point is not that he is more confident in rule-consequentialism than he is in the pluralist list. Consider again my analogy at the level of the pro tanto moral duties. Compare the proposition ‘Breaking one's promise is wrong, pro tanto’ with the proposition formed by the conjunction of ‘It is wrong (pro tanto) for Joe to break his promise to Sara’; ‘It is wrong (pro tanto) for Jane to break her promise to Alan’; and all similar judgments (those to which we are committed) concerning the wrongness of every particular instance of promise-breaking. The general proposition entails every conclusion entailed by the conjunction: the general proposition is coextensive with the conjunction. I am no more confident that the general proposition is true than I am that the conjunction is true. Yet I still endorse and bear in mind the general proposition rather than the unwieldy conjunction. This makes sense because the general proposition enjoys pragmatic priority, not epistemic priority. Monism may be similarly more efficient than pluralism, in second-order cases. If monism is also comparably accurate, in the aggregate, then it is superior.

IX

The reader may have noticed that my argument in favor of greater unification in moral theory itself appeals to the value of conserving cognitive resources. I claim that more unified theories are superior, ceteris paribus, because they lessen the cognitive burdens of moral deliberation and enable us to employ our cognitive resources for other purposes that serve humanity. My reliance on these values may raise the worry that I have assumed a form of consequentialism at the outset. Indeed, the deep problem here is that the degree of unity we need in our moral principles is itself a normatively loaded question.

I shall now explain, however, that my main argument does not appeal to the value of promoting the good to any greater extent than does the pluralist herself. Consider that the pluralist must argue that pluralism constitutes a stable resting place in between particularism and monism – that there are reasons why a list of seven or so non-derivative duties gives us a pragmatic advantage over particularism, and that no such advantage accrues to Hooker's additional unification.

An obvious justification for using a list of seven or so duties appeals to the fact that, for various contingent reasons, a society that formulates these duties explicitly and internalizes them will enjoy greater welfare than one that formulates more or fewer such duties. The fact that the list contains seven duties and not two, nor two hundred, reflects contingencies of human psychology, such as our cognitive limitations and our degree of innate altruism. Suppose we tried to arrive at roughly the same judgments as the Rossian, but without employing all the Rossian categories. This would prove to be an unnecessary cognitive burden. Avoidable misfortune would result from this use of our limited cognitive resources. Therefore, limiting ourselves to the seven Rossian duties promotes the good.

This rationale for the limited set of Rossian duties plays right into Hooker's hands. He can argue that his rule-consequentialist principle promotes the good, and hence constitutes a rationale for the Rossian duties that simply extends the argument that favors pluralism over particularism. The pluralist, of course, will want to resist this justification. To do so she might appeal to our considered convictions, rather than a rule-consequentialist principle. A society that did not recognize the Rossian duties would not just be a society with less than the optimal level of flourishing, it would contain a tremendous quantity of preventable suffering and a very low absolute level of welfare for many individuals. Our considered convictions condemn that level of suffering, the pluralist could insist. Therefore, our convictions themselves indirectly support the recognition of the Rossian categories. Whereas, we have no such conviction about the level of suffering that would be engendered by widespread failure to use a monistic principle such as Hooker's.

This argument, however, relies on an unsubstantiated premise. It only works if something like Ross's list of duties is really the minimal set necessary to support what our considered convictions deem a morally tolerable world. Each additional duty that the pluralist adds to her list represents an additional degree of unification, beyond what her moral theory otherwise would have exhibited. Each such degree of unification must be justified, on the pluralist's current approach, by showing that our considered convictions condemn the society that would result from the internalization of any less unified moral theory.

The problem for the pluralist is that if, indeed, additional unification can only be justified in terms of avoiding conflict with our considered convictions, then her list of duties appears to be already larger than she can justify. We have no reason to believe that the Rossian list represents the smallest list that remains compatible with our considered convictions. This is especially apparent if we impose the following restriction on ourselves, as I think we must. As we endeavor to justify each successive increase in unity we must limit ourselves, at that stage of unification, to those convictions that are more confidently held and widely shared than are those that directly favor the additional degree of unification under consideration. These convictions, moreover, must be independent, that is, not reflected in the proposed duty itself.Footnote ⁴⁴ Otherwise, the pluralist's argument becomes question-begging. To illustrate, it would be no surprise to find that a conviction against failing to develop one's talents condemns a society that results from the internalization of a moral code that fails to prescribe self-development.

The pluralist will find it difficult to justify the full list of Rossian duties while respecting these constraints because, in the vast majority of cases, simple beneficence mandates actions that are coextensive with those mandated by the other Rossian duties. In order to justify adding a new, discrete duty, the pluralist must argue that our considered convictions condemn any society that results from the failure to add that duty explicitly to its moral code. Yet the extensional difference only manifests itself in atypical cases. There are, for example, cases in which the Rossian duty of non-maleficence condemns physical assault, even though a widely internalized norm permitting physical assault under those conditions would maximize welfare. But such cases are rare. A society the members of which internalize only the beneficence principle and not the duty of non-maleficence will usually refrain from maleficent conduct, anyway, just not under that intensional description. They will not assault their neighbors at will, though they might do so in order to save lives in an unusual emergency. I doubt we have any other considered convictions, ones not concerning non-maleficence per se, that condemn such a society. It may not be an ideal society, but I fail to see that it would have enough other objectionable features to conflict with our considered convictions.

The preceding argument does not lodge the particularist's familiar complaint that the pluralist's list of duties is already too small. Rather, my point is that the pluralist cannot coherently insist that the only legitimate justification for additional degrees of unity involves appealing to our considered convictions. Instead, in order to justify her own preference for moderate unity she must appeal either to epistemic values or (more promisingly) to pragmatic considerations, as I have advocated in this article. She must appeal, in other words, to the very considerations which, if consistently applied, favor monism over pluralism.

X

My argument in favor of principle monism is designed to operate in conjunction with the method of reflective equilibrium. This entails that we cannot purchase greater unity, whatever its advantages, at the cost of compromising our considered convictions. The pluralist may now wish to take advantage of this fact by arguing that monism does, in fact, yield conflicts with our convictions, and hence stands revealed as inferior to pluralism, despite my arguments.Footnote ⁴⁵

This is a legitimate concern which must be addressed on its merits. Nothing in my article strengthens Hooker's case that his version of rule-consequentialism is, in fact, consistent with our considered convictions. But the concern about conflict is separate from the charge that Hooker's rule-consequentialist principle is superfluous. The concern about conflict is also distinct from the charge that Hooker betrays his commitment to reflective equilibrium when he uses his principle to resolve conflicts among the Rossian duties. It is important to keep these several charges distinct. If a monistic principle, such as rule-consequentialism, will not, in fact, lead to conflict with our convictions, then there is a reason to favor it over pluralism, just as Hooker claims.

XI

From the pluralist's standpoint, monists such as Hooker seem myopic in their commitment to a single master principle, given the many different duties and values that glitter on the surface of common-sense morality. My arguments suggest that the monist's choice is not as myopic or arbitrary as it might appear. Monism need not derive from a quasi-aesthetic preference for simplicity, or a desire to mimic natural science, or any personal temperament. Rather, it can be defended in terms of reasonable intensional-pragmatic considerations which anyone can appreciate. These are considerations to which pluralists themselves may want to appeal. Monism is but a natural extension of the same pragmatic impetus toward simplicity and unity in virtue of which generalism succeeds particularism.Footnote ⁴⁶