II. THE THREAT OF COLLAPSE

The focus throughout will be on particular cases in which adhering to a generally optimizing rule will produce suboptimal results. A standard example will help to discipline the discussion. Suppose that by executing one individual a judge knows to be innocent, s/he can avert a riot which will lead to the deaths of a significant number (call this number N) of persons. Circumstances will differ in each case, but it is true that for some value for N, viable consequentialist theories will require that the judge knowingly execute the innocent person, and will maintain that it is wrong not to do so. What does a rule-consequentialist say about this sort of example?

To consider this question, we must consider what rule is applicable to the case and whether a supporter of RC would endorse that rule. It is plausible to suppose that a rule (call it R) which proscribes killing innocent persons would be adopted since it is part of the possible code of rules ‘whose inculcation in the overwhelming majority of the next generation could reasonably be expected to result in as good consequences as could reasonably be expected to result from any other identifiable code’.Footnote ⁵ If this is the case, then R would be part of the moral code selected by rule-consequentialism, and acts would be permissible/right if and only if they are allowed by this preferred code and wrong otherwise. When the value of N is sufficiently high in the case just described, following R would be suboptimal. This is potentially problematic for the supporter of RC since in such a case, the consequentialist spirit of the view would seem to require that the judge knowingly condemn the innocent person.Footnote ⁶ One option available to the rule-consequentialist is to incorporate an exception into the rule. This move is unattractive since it seems that if this is done in every such case, then RC and act-consequentialism (AC)Footnote ⁷ will have the same implications for action. This point is typically used to motivate part (1) of the dilemma, the claim regarding extensional equivalency, yet HookerFootnote ⁸ responds to this potential problem.

Rule-consequentialists avoid the ‘collapse’ of RC into AC by arguing that RC will not allow exceptions to rules in the way a supporter of AC does. They provide two reasons for this. First, incorporating exceptions into the rules would undermine mutual trust and confidence in other persons, since it would be generally known that departures from the rules were allowed and even encouraged in some cases. Trust and confidence possess great consequential value, and rule-consequentialists claim that it is an advantage of their view over AC that it can in fact preserve these things. Second, creating exceptions to the rules would either result in rules which were too complicated to be understood and internalized, or if the exceptions were codified in newly created rules, would create a virtually endless set of rules. Either option would be exceptionally costly in rule-consequentialist terms, given the high internalization costs of such moral codes.Footnote ⁹ The point is that RC would endorse a set of rules that are optimal in terms of complexity and number. This set would generate prescriptions for action which differ from that of AC, and therefore RC does not collapse into AC. For the sake of argument, I will grant that this is the case. I will suppose that supporters of RC will not generally incorporate exceptions into the rules, and that the ‘collapse’ is avoided since RC and AC do in fact have different implications. The focus will be on part (2) of the dilemma for the remainder of the article.

III. THE THREAT OF INCONSISTENCY

If the rule-consequentialist does not allow such exceptions to the rules, then the view appears to be inconsistent as a form of consequentialism. The reason supporting this claim is that the rule-consequentialist appears to be dedicated to maximizing the good, yet she accepts that one should follow a rule that is part of the preferred moral code even if in a particular case she is absolutely certain this would not maximize the good. Hooker's reply is that this problem is merely an appearance, since rule-consequentialists do not in fact have an overarching commitment to maximizing the good.

Hooker argues first that the rule-consequentialist agent does not have the overarching commitment as part of her moral psychology.Footnote ¹⁰ He points out that the rule-consequentialist's moral psychology could be as follows: (1) her fundamental moral motivation is to do what is impartially defensible, (2) she believes that acting on impartially justified rules is impartially defensible, and (3) she believes that RC constitutes the best system of impartially justified rules. Even given these commitments, the rule-consequentialist need not have any overarching commitment to maximizing the good as part of her moral psychology. Hooker concludes, ‘For such agents, there is nothing incoherent about following rules when this won't maximize the good.’Footnote ¹¹

I agree. An analogous point about moral psychology appears in discussions of Peter Railton'sFootnote ¹² sophisticated act-consequentialist view. A sophisticated consequentialist is one who employs the main principle of consequentialism only as a criterion for determining rightness and does not necessarily use it as a decision-procedure. This allows sophisticated act-consequentialists to be motivated (e.g.) on the basis of stable character traits and dispositions; they need not directly appeal to the consequentialist principle when deciding how to act. These dispositions are subject to what RailtonFootnote ¹³ calls a ‘counterfactual condition’: if the disposition does not produce optimal results over a significant period of time, then (assuming this is possible) the agent will take steps to change that disposition.Footnote ¹⁴ It is generally agreed upon, even by critics of the view,Footnote ¹⁵ that the overarching commitment to maximizing the good embodied in the counterfactual condition need not be part of sophisticated consequentialist agents’ moral psychology. Agents will periodically evaluate their lives from an external perspective, yet in their daily attitudes towards others will guide their lives on the basis of personalized commitments to (e.g.) career, family and friends.

This analogous point made in discussions of act-consequentialism is enlightening, since sophisticated act-consequentialists do possess an overarching commitment to maximizing the good. They acknowledge the familiar point that in order to maximize the good in the long run, one may have to perform some wrong actions in particular instances,Footnote ¹⁶ and they realize that adopting stable dispositions will likely lead them to do precisely this. But the fundamental reason they adopt and maintain these dispositions is that they expect them to maximize the good in the long run. If this strategy fails to realize the most good overall, the dispositions should be abandoned. All this can be true, yet these facts may be kept out of mind at most times. This suggests that a consequentialist's moral psychology and a consequentialist's commitment to maximizing the good can be distinct. Hence, even if rule-consequentialist agents do not have an overarching disposition to maximize the good as part of their moral psychology, we cannot conclude that they do not possess this commitment in some other way. The overarching commitment to maximizing the good could be manifested in the theory itself, for instance, in the principles employed to morally assess rules and actions.

Does rule-consequentialism embody an overarching commitment to maximizing the good in this other sense? Hooker anticipates this possibility and denies it:

Hooker's suggestion appears to be that since the commitment regarding what could reasonably be expected to maximize the good operates on the level of rules, which are then appealed to in evaluating actions, an overarching commitment to maximize the good is not truly part of the theory itself. This is puzzling. This suggestion does nothing to establish that the theory itself does not contain the overarching commitment, since part of the theory consists of its rules, and on RC rules are selected simply on the basis of a commitment to maximizing the good. As I see it, there are two questions: on RC, does a commitment to maximizing the good exist at the level of rule selection? Hooker clearly acknowledges that this is true above. Second, on RC, does a commitment to maximizing the good exist at the level of act judgments? Hooker thinks this is not the case, and therefore that RC does not possess the overarching commitment since (I take it) for this to be true a commitment to maximizing the good must exist at both levels. In the passage above, Hooker sets up a veritable ‘iron curtain’ between the levels of rule selection and act evaluation. However, if points of contact exist between these two levels, then the commitment to maximizing the good at the level of rules will bleed over into the principle for the evaluation of actions.

Close consideration of Hooker's theory reveals several points of contact between these two levels, and hence, the existence of an overarching commitment to maximizing the good within the theory itself. If there was no contact between these two levels, then the Ideal Code would remain static.Footnote ¹⁸ Even if acting on the basis of an Ideal Rule turned out not to maximize expected value in the long run, we cannot change it, for that would take down the ‘iron curtain’ between the levels of rule selection and act judgment. On Hooker's theory, acts should conform to ideal rules. But ideal rules must be such that their inculcation is reasonably expected to maximize the good. In order for persons acting on the basis of the Ideal Code truly to fulfill the purpose of the rule selection process, they must learn something about the actual results of internalizing and following the code. If the theory never takes stock of the acts performed on the basis of the Ideal Rules, then the theory can never improve its educated guesses regarding expected value at the level of rule selection. This, however, would make sustained moral progress impossible; the Ideal Code would only change via a new calculation of the expected value of rules to be performed each generation, as suggested by Hooker.Footnote ¹⁹ This is a troublesome result. Yet HookerFootnote ²⁰ rightly endorses the importance of moral progress: ‘A moral theory should offer us hope of improvement, and not merely improvement in the level of conformity with established norms, but improvement in the norms themselves. Rule-consequentialism, properly conceived, does this. . .’ If RC does not attend to the consequences of following the Ideal Code in the real world, it lacks adequate resources for moral progress. Yet attending to actual consequences at the level of actions in order to facilitate this improvement implies that RC has some form of commitment to maximizing the good at both the level of rules and acts, and hence has an overarching commitment to maximizing the good.

Another point of contact between the two levels is revealed in Hooker'sFootnote ²¹ commitment to a ‘disaster exception’. There is some commitment to maximizing the good at the level of acts since HookerFootnote ²² thinks that persons should perform acts that override a rule that is part of the Ideal Code in order to prevent a disaster. This is a favorable part of Hooker's theory, since it allows RC to escape a blind absolutism that requires (e.g.) telling a murderer the truth about where his prey is hiding. On this matter, HookerFootnote ²³ says, ‘Rule-consequentialism therefore holds that we should break the promise or tell a lie when necessary to prevent a disaster.’ Yet this concedes the importance of attending to the particular consequences of one's actions, at least in certain cases. Notice that I am not claiming that this move itself defeats RC or makes RC collapse into a form of AC. My point is simply that this signifies a commitment to maximizing the good at the level of acts. Since this commitment exists at the level of both rules and acts, an overarching commitment to maximizing the good exists in the theory itself.

Finally, Hooker's theory pays attention at the level of acts to the extent of others’ compliance with ideal rules.Footnote ²⁴ This again signifies the existence of an overarching commitment within the theory itself. An example will clarify why this is the case. In a discussion of rule-consequentialism and doing good, HookerFootnote ²⁵ argues that RC does better than AC as well as other theories with respect to its obligations to provide aid. Hooker presents a case in which he and nine others see ten persons fall from a ship's deck into the water. Hooker imagines a scenario in which he throws one lifejacket to the people in the water but the other observers do nothing, even though each could easily assist. HookerFootnote ²⁶ concludes: ‘I should continue throwing life-jackets overboard until everyone in the water has one. Here, because others aren't helping, I should do more.’ Notice that Hooker's moral assessment of the act in question depends not on the rule (a member of the Ideal Code) which, if everyone felt free to follow it, would produce the maximum expected value, but instead on the reasonably expected consequences of that particular action itself. I submit that if any of these points show that a commitment to maximize the good exists at the level of act evaluation on RC, then this establishes a point of contact between both levels and hence Hooker's theory itself possesses an overarching commitment to maximize the good.

One might object to the latter points above by stating that the RC approach to disasters and others’ non-compliance is rule-governed, and therefore fits perfectly well within Hooker's theory. Perhaps a rule such as ‘Prevent disasters’ is part of any ideal moral code and trumps all other rules with which it may conflict, or perhaps it is the case that all rules in the ideal code will have added to them ‘unless following this rule this will cause a disaster’. This strategy is troublesome for the defender of RC, since it is now fair for critics to question whether RC and AC will have different practical implications and whether RC will allow exceptions to rules in the way that AC does. In short, in addressing the problem in this way, the defender of RC may impale himself on the first horn of the dilemma. If Hooker's action in the life-jacket case is justified by the ‘disaster exception’, then this raises the question, ‘What is a disaster?’ Is it the death of nine persons? Eight? As the reader will no doubt see, this line of questioning raises the possibility that the continual application of a ‘disaster exception’ makes RC a very demanding moral theory, perhaps nearly as demanding as AC, contrary to Hooker'sFootnote ²⁷ claims. One might think Hooker's action of throwing more life-jackets is to be guided by the RC rule concerning giving assistance to others and making sacrifices. This rule is provided by HookerFootnote ²⁸ and states as follows: ‘Over time agents should help those in greater need, especially the worst off, even if the personal sacrifices involved in helping them add up to a significant cost to the agents. The cost to the agents is to be assessed aggregatively, not iteratively.’ If it is true that Hooker should continue to throw lifejackets until each individual in the water is helped, then it seems that moral obligation and cost is being assessed iteratively in this case. After all, Hooker may have already done his part in the aggregate. The point is that Hooker's (correct) judgment to continue helping is driven by the particular good consequences of the act, not the rule such that if all felt free to follow it, it would have the best expected consequences.

At the end of Ideal Code, Real World, HookerFootnote ²⁹ says the following regarding the charge of inconsistency: ‘My account of the rule-consequentialist agent's psychology answers the objection. Someone with this psychology would neither behave like an act-consequentialist nor have an overarching aim to maximize the good.’ As I have argued, this move does not escape the internal inconsistency criticism as it applies to the theory itself. The crucial insight is that the problem stems from the RC theory's assessment of the suboptimal action performed in such cases. This assessment, I will now argue, illustrates the real difficulty and commits the RC theory to a non-consequentialist form of ultimate moral value.

It is clear that RC will judge the action of not killing the one to save N other persons (for the requisite value of N) as morally right in some instances in which acting this way produces suboptimal results.Footnote ³⁰ This action is judged to be right on RC given that it is in accord with a rule that is part of the best possible set, as defined earlier. As argued above, RC as a theory itself possesses an overarching commitment to maximizing the good. Such suboptimal acts should not be assessed as morally right, but RC does require that they be assessed this way. In order for this judgment to be acceptable, some other form of moral value must be attributed to these suboptimal actions which accord with the rule in order for them to be properly labeled ‘morally right’.

What is this other form of value? All possible sources of consequentialist value on RC have been taken into account in the judgment regarding the action of not killing the one – its rightness derives from the fact that it is an action in accord with a rule that is a member of the preferred moral code. Yet we know ex hypothesi that it is not the most valuable action in consequentialist terms; it does not promote well-being or whatever else is held to be of intrinsic value to the greatest possible extent. This suggests that in order to make the judgment that the suboptimal act of not killing the one person to save N other persons (for the requisite value of N) is morally right, rule-consequentialists must appeal to an ultimate form of non-consequentialist value. There must be some independent value in adhering to rules, apart from its contribution to pleasure, desire-satisfaction, or other intrinsic goods valued by the consequentialist.

This may shed light on why Hooker discusses the role of intuitions in selecting the preferred moral code. HookerFootnote ³¹ makes clear that he seeks a consequentialist theory that ‘matches and makes sense of our considered convictions (or “intuitions” in some metaphysically and epistemologically neutral sense of the term)’. This desideratum serves on Hooker's version of RC at least to break ties between moral codes that it is reasonably believed would produce equally good results,Footnote ³² and perhaps the closer the fit with conventional morality, the greater the amount of expected value ascribed to the code of rules. HookerFootnote ³³ notes at one point that RC ‘does a better job than its rivals of matching and tying together our moral convictions, as well as offering us help with our moral disagreements and uncertainties’. A defender of RC could argue that the fit with moral intuitions serves as an additional form of value from which rule-consequentialists can draw in order to make the judgment that the suboptimal action of not killing the one person to save N other persons (for the requisite value of N) is morally right. While adherence to common-sense morality may confer moral value upon an act (though I doubt it), the important point for my purposes is that it cannot confer consequentialist value upon this particular suboptimal act. HookerFootnote ³⁴ makes clear that he accepts a methodology involving the notion of ‘reflective equilibrium’Footnote ³⁵ to assess moral theories, and he argues that RC does quite well on this method of assessment. While adherence to conventional moral beliefs may constitute a perfectly fine motivation for being drawn to rule-consequentialism over other moral views,Footnote ³⁶ there is no independent consequentialist value in adhering to intuitions. The rule-consequentialist's judgment of such suboptimal acts commits her to acknowledging some ultimate form of non-consequentialist moral value, if she is to say that a particular suboptimal act of the type under discussion is morally right, since all forms of consequentialist value have been exhausted. In his defense of consequentialism, Hooker offers no argument which establishes the independent value of adhering to rules, following intuitions, etc. Therefore, as a form of consequentialism, RC is properly branded with the charge of internal inconsistency.Footnote ³⁷

I have argued that Hooker's RC is committed to a non-consequentialist form of ultimate moral value in general, given its assessment of suboptimal actions as discussed earlier. Hooker's theory itself has three specific theoretical commitments that tie it to the existence of a non-consequentialist form of ultimate moral value. First, HookerFootnote ³⁸ supports the idea that if two or more moral codes are equal in terms of expected value, then the one closest to conventional morality should determine which acts are wrong and right. As he puts it at one point, ‘from the set of codes with unsurpassed expected value, the one we should follow is whichever is closest to the conventionally accepted code – that is, whatever is closest to the one already established in society’.Footnote ³⁹ This facet of the theory is quite unusual, since the standard consequentialist position would mandate indifference between such codes. (Perhaps a randomized procedure would be used to decide between the relevant codes.) The terms of the discussion rule out any difference in consequentialist value between the codes in question. Hence, Hooker's position on using such a ‘tie-breaker’ signifies his commitment to ascribing ultimate non-consequentialist moral value to conventional morality. While it might be the case that this rarely comes into play, Hooker's view on the ‘tie-breaker’ supports (at least in principle) the notion that (e.g.) homophobic, sexist or racist norms can have moral value in deciding between alternative ideal codes.

Second, Hooker is committed to a non-consequentialist form of ultimate moral value given his position on the publicity of moral rules. Hooker addresses the issue, famously discussed by Henry Sidgwick,Footnote ⁴⁰ of whether the consequentialist basis of morality (assuming it is true) should be kept secret from at least some part of the population. Since Sidgwick's The Methods of Ethics, consequentialists have acknowledged the possibility that it may be better in consequentialist terms if certain persons, and perhaps even the overwhelming majority of persons, did not believe that consequentialism was the correct moral theory. The idea is that the less intelligent group would do worse by attempting to apply the consequentialist test in practice, while the more intelligent class would do better by knowing the consequentialist foundation of morality and using this to decide when to override a conventional rule. This is sometimes discussed as the requirement that consequentialism become a self-effacing or esoteric morality.Footnote ⁴¹ HookerFootnote ⁴² strongly condemns the possibility of a non-public morality; as he says: ‘However, rule-consequentialism, at least as I have formulated it, rules out the objectionable elitism and duplicity from the start. The theory insists that there are the same rules for internalization by the overwhelming majority of everyone.’ Hooker quite clearly judges the question of whether rules are to be public independently of consequentialist considerations. His theory assigns ultimate non-consequentialist value to the publicity of moral rules.

Finally, Hooker's interesting discussion of conflicting rules commits him to a non-consequentialist account of ultimate moral value. HookerFootnote ⁴³ addresses the standard problem of conflicts between rules by following famed rule-consequentialist Richard Brandt's suggestion that in such cases, agents should do what morally well-trained persons would find least unsatisfactory. As Hooker rightly notes, there is a strong deference to virtue ethical reasoning in this suggestion. HookerFootnote ⁴⁴ concludes that in cases of conflict between rules, ‘moral right and wrong cannot be determined except by reference to the dispositions of a virtuous person. The best version of rule-consequentialism accepts that this is true in some cases where different moral considerations pull against each other.’ Hooker's position is not simply one on which virtuous dispositions serve as merely instrumentally valuable indicators. Instead, as his discussion of the similarity between Brandt's suggestion and virtue ethics makes clear, in cases of conflicting rules the dispositions of virtuous persons determine what the best course of action is. This position commits Hooker to assigning ultimate non-consequentialist value to virtuous dispositions; these (not the code of rules whose general internalization has maximum expected value) are used to determine moral right and wrong in such cases.Footnote ⁴⁵

Ultimately, Hooker maintains that rules simply must be impartially defensible and must tie together our moral convictions.Footnote ⁴⁶ If this is the goal of RC, then it is quite mysterious why rules are to be selected by reference to whether their inculcation is expected to maximize the good. Instead, rules should be selected solely according to whether they express our considered moral convictions. HookerFootnote ⁴⁷ thinks that a large part of the appeal of RC is the fact that its implications ‘match quite closely our commonsense deontological intuitions’. This is hardly an uncontroversial claim, even among rule-consequentialists. BrandtFootnote ⁴⁸ explicitly expresses his distrust of intuitions as a basis for morality, and Hooker provides no argument to reconcile RC's commitment to selecting rules whose inculcation will maximize the good with the commitment to rules which embody our moral intuitions. It is unclear how this is a consequentialist view in any substantive sense.

IV. RC AND CONSEQUENTIALISM

At this point, one may reasonably ask: is RC a form of consequentialism? The standard way to draw the distinction between consequentialist teleology and deontology, deriving from Rawls,Footnote ⁴⁹ states that the former defines the right in terms of the good, while the latter does not. One could maintain that RC defines the right independently of the good, but as I have argued, Hooker does not provide arguments to ground the independent value of (e.g.) adhering to rules or following conventions. HookerFootnote ⁵⁰ explicitly states that on RC the ideal code of rules is selected by whether its inculcation can reasonably be expected to maximize the good. Rightness is assessed in terms of conformity with this moral code. The right is defined, at least indirectly, ultimately in terms of the good, and hence Hooker's RC appears to be a form of consequentialism.

Frances Howard-Snyder provides a resourceful way to block the inconsistency objection by arguing that RC is not in fact a form of consequentialism at all. According to Howard-Snyder,Footnote ⁵¹ ‘a view is consequentialist if it does not include any agent-centred element’. On her interpretation, RC gives persons agent-relative goals such as producing states of affairs in which ‘ “Her promises are kept” or “Her children flourish as far as possible” ’.Footnote ⁵² On this reading, RC is a form of deontology since it possesses agent-relative (or agent-centered) elements. If RC were an agent-relative theory, and we suppose that all forms of consequentialism subscribe solely to an agent-neutral account of foundational moral value,Footnote ⁵³ then the fact that no forms of ultimate consequentialist value can be found to support RC's assessment of the suboptimal actions under discussion ceases to be troublesome. On this interpretation, there is nothing irrational or incoherent about acting in accordance with a rule from the preferred code in a case where this is known to produce suboptimal results. On an agent-relative theory, the person you are, your projects and their importance to you can affect the rationality and permissibility of an action.Footnote ⁵⁴ What essentially distinguishes a consequentialist from a non-consequentialist or agent-relative theory, according to Howard-Snyder,Footnote ⁵⁵ is that on the former, valued states of affairs can be described without essential reference to the agent, while this is not true of the latter. Although a full defense of this way of drawing the consequentialist/non-consequentialist distinction is outside the scope of this article, Howard-Snyder's account of this distinction is defensible and widely held. While I think it is troublesome to rule out by definition forms of consequentialism that include agent-centered elements, this account (as we will see) is also accepted by Hooker himself and therefore works for the purposes of the argument in this article. The upshot of this discussion is that Howard-Snyder provides a promising response to the objection that RC is internally inconsistent, although this response requires viewing RC as a variety of non-consequentialism. Howard-SnyderFootnote ⁵⁶ reflects upon the prospects of RC, stating that ‘regarded as a form of deontology rather than a form of consequentialism, it becomes less vulnerable to attack’.

While it may be the case that RC avoids inconsistency when reformulated in the way presented by Howard-Snyder, RC is not yet out of the woods, since this victory is purchased at the price of making the theory desperately unmotivated and alienating the theory from any appeal derived from its consequentialist roots. If RC is best understood as a form of deontology, then it loses its critical edge as a reformer's morality. I will utilize an example discussed by BrandtFootnote ⁵⁷ to clarify the point, although I should not be understood as suggesting that Brandt himself sees RC as a deontological theory. Suppose you live in South Africa during Apartheid, and it is the case that the preferred RC code would include a rule banning race discrimination. One may publicly condemn such race discrimination, join forces with others to put education programs in place, etc. BrandtFootnote ⁵⁸ suggests that in this case the optimal course of action on RC is a sort of compromise which involves ‘avoiding the most offensive things such as interracial sexual contacts and speaking in favor of a completely non-discriminatory system’. Examples such as these in which there exists only minority adherence to a rule of the preferred code illustrate RC's fatal attraction to the status quo. While act-consequentialists acknowledge the incremental nature of social change, it is puzzling that a knowledgeable defender of RC such as Brandt would suggest that interracial sexual contacts per se should be avoided on moral grounds. To do so is racially discriminatory, and if the RC code contains a rule banning such discrimination, then interracial sexual contacts per se should be considered at least morally permissible. Also, RC is subject to the objection to deontology,Footnote ⁵⁹ powerfully presented by Samuel Scheffler,Footnote ⁶⁰ that there is an air of paradox surrounding a theory that holds some acts to be so morally objectionable that they ought not to be performed, even if this results in a state of affairs in which more equally weighty acts of the very same kind occur. ‘Rule-consequentialism’ becomes a misnomer, thereby making its identity and family membership unduly mysterious.Footnote ⁶¹

Howard-Snyder's argument does not help out Hooker's view, since he accepts that all forms of impartial consequentialism, including RC, share a commitment to fundamental agent-neutral value. HookerFootnote ⁶² calls a theory consequentialist ‘if and only if it assesses acts and/or rules (or motives, social codes, virtues, or ways of life) in terms solely of the production of agent-neutral value’. He then goes on to say that ‘rule-consequentialism claims that an act's rightness is determined solely by whether the general acceptance of a system of rules allowing that act would promote agent-neutral value’.Footnote ⁶³ These passages make clear both that Hooker considers RC a form of consequentialism and that RC is committed to assessment solely on the basis of fundamental agent-neutral value. HookerFootnote ⁶⁴ explicitly rejects Howard-Snyder's reformulation of RC, yet his theory is unable to escape the dilemma regarding incoherence. Hooker's theory is deprived of a compelling argument for why adherence to rules is rational in cases when it is known that this adherence will produce suboptimal results.