CHAPTER II - THE ANARCHIST CHALLENGE

Consequentialism and the Argument from Rationality

It has been claimed that only consequentialism can be consistent with the demands of practical rationality. The purpose of this paper is to show why that might not be the case. It is argued that the non-consequentialist might endorse the same conception of rationality that the consequentialist does. In other words, it would not be true that a commitment to a plausible conception of practical rationality entails the rejection of deontological constraints. It is acknowledged that saving deontological constraints from the charge of irrationality might entail endorsing an unfamiliar conception of value. But it is claimed that such a conception would seem to be more plausible that the counterintuitive judgments that follows from the non-existence of deontological constraints.

I. The ‘Paradox of Deontology’

A deontological theory is a moral theory affirming the existence of deontological constraints. The characteristic feature of such moral restrictions is that we are forbidden to violate them even when our violation would be necessary to prevent a greater number of violations. So deontological theories hold that it is impermissible to commit one act of an objectionable type in order to minimize acts of the very same type. But how can the minimization of morally objectionable conduct be morally unacceptable? Robert Nozick discovered this puzzling feature presenting his own theory:

If nonviolation of C is so important, shouldn't that be the goal? How can a concern for the nonviolation of C lead to the refusal to violate C even when this would prevent other more extensive violations of C? What is the rationale for placing the nonviolation of rights as a side constraint upon action instead of including it solely as a goal of one's action?[1]

Some years later, in his widely acclaimed book, The Rejection of Consequentialism, Samuel Scheffler raised this question again.[2] Nozick had claimed that deontological constraints reflect the Kantian principle that “individuals are ends and not merely means; they may not be sacrificed or used for the achieving of other ends without their consent”[3]. Scheffler says that it is natural to interpret Nozick's contention as a suggestion that violations of deontological constraints involve violating the victimized individuals themselves, treating them as means instead of ends.[4] But then, Scheffler would reject the plausibility of Nozick's answer. The reason is simple: regardless of how high the disvalue of a violation of constraint C is, Scheffler reminds us that a greater number of equally weighty violations might take place, and hence at least as much disvalue will ensue if the constraint is not violated.[5] Scheffler also claims that it makes no difference which feature of a violation is singled out as having high disvalue. For nothing one can say about that feature will be capable of explaining a moral rule denying that it is permissible to minimize the occurrence of the very same feature.[6] In The Limits of Morality, Shelly Kagan would seem to share Scheffler’s diagnosis. After his examination of possible deontological answers, Kagan concluded: “not only has the [deontologist] failed to give an adequate explanation; he has failed to give even the beginnings of one.”[7]

Consequentialism is usually understood as an account of right action. The right action is the one that produces the highest-ranked state of affairs in terms of a given conception of the good. On the contrary, deontology could be understood as maintaining that given any impersonal principle for ranking overall states of affairs from best to worst, there will be circumstances in which one is not permitted to produce the best available state of affairs. Deontology could then be understood as a system of constraints on the maximization of moral goodness. For deontology, there are limitations on the conduct of the individual agent which take priority over calculations of overall impersonal value. The weight given to this priority is such that, at least sometimes, it is impermissible to violate these constraints even when such a violation would prevent more numerous violations of the very same constraint or of others equally important.

Consequentialists tend to believe that deontology contradicts a familiar conception of rationality. In Scheffler’s words, such a conception states that “if one has a choice between two options, one of which is certain to accomplish the goal better than the other, then it is, ceteris paribus, rational to choose the former over the latter”.[8] According to Philip Pettit, the key to the main argument for consequentialism is “that every moral theory invokes values such that it can make sense to recommend in consequentialist fashion that they be promoted or in non-consequentialist that they be honored.”[9] And since “the non-consequentialist has the embarrassment of having to defend a position on what certain values require which is without the analogue in the non-moral area of practical rationality,”[10] consequentialism ought to be preferred to deontology.

II. Typical Deontologist Answers

Deontologists have presented two major types of answers to the charge of irrationality. Some authors have denied the existence of an impersonal point of view from which the occurrence of a greater number of violations of a particular constraint is worse than the occurrence of a lesser number. Eric Mack, for example, has said that it is surely worse for some individuals, but it is better for others: for those who are not the victims of the greater number of violations.[11] In the same line, John Taurek has argued that numbers should not have moral significance. Because five individuals each losing his life, for example, does not add up to anyone's experiencing a loss five times greater than the loss suffered by any one of the five.[12] And C. S. Lewis had said before: “There is no such thing as a sum of suffering, for no one suffers it.”[13] The implication of this idea is clear. We could not claim that deontology contradicts rationality if it were the case that deontology does not establish the desirability of those state of affairs that we are forbidden to bring about.

There might be something attractive about this line of thought, but its radical implications are too obvious to make it an uncontroversial solution even among deontologists. The reason is that we would like to draw a distinction between the role played by an impersonal point of view in cases of benevolence and cases of justice. For example, we would like to say that a firefighter facing the option of saving one life or saving twenty from some kind of disaster, he must choose saving the twenty. But if the very plausibility of the impersonal point of view is denied, we might not be able to justify those types of judgements.

The second deontological answer is based on a natural temptation. If we have to comply with deontological constraints even if the outcome of our compliance is more extensive violations, their justification could be found in some feature of our own violation. There is a well-known history in the Kantian tradition of stressing that moral agents must strive to keep their own hands clean. It is not surprising then to find this type of consideration in trying to avoid the charge of irrationality. Examples of this strategy abound. Thomas Nagel believes that deontological reasons have their full force against your doing something --not just against its happening.[14] Stephen Darwall also points out that a key feature of morality is the “duty not to compromise one's moral integrity”.[15] Judith Andre says that one is always responsible for one's own actions, but only sometimes for those of other people.[16]

Shelly Kagan has rightly explained, however, that no argument based on the loss of moral integrity can succeed in justifying the existence of deontological constraints unless their violations do actually involve a sacrifice of the agent's integrity. But if deontological constraints are not justified, then it is no sacrifice of moral integrity for the agent to violate them.[17] As Mack points out, a duty such as the one commanding one not to compromise his own moral integrity is a second-order duty. “If one's specific first-order duties all involve maximizing the good or minimizing the bad, then such maximization or minimization will be essential to one's moral integrity.”[18] The same argument can be rephrased in terms of moral responsibility.

III. Why Deontology Might Not Be Paradoxical

The charge of irrationality towards deontology may be put in the following way. First, it is claimed that deontology is committed to a particular ranking of states of affairs. Let us say that this is such a ranking, ordered from best to worst:

1. n violations of constraint C

2. n + 1 violations of constraint C

Then, it is argued that for deontology, there is some possible situation in which an agent is forbidden to violate constraint C n times even though this is necessary to prevent n + 1 violations of constraint C. From this kind of recommendation, the argument goes, we may deduce not the previous but the following ranking:

1. n + 1 violations of constraint C

2. n violations of constraint C

So now it is concluded that this derivation shows why deontology is committed to an implausible conception of rationality. It is possible, however, to challenge that conclusion. When an agent is forbidden to violate constraint C n times even though this is necessary to prevent n + 1 violations of constraint C, we might also infer the following ranking:

1. n + 1 violations of constraint C

2. prevention of n + 1 violations of constraint C by n violations of constraint C

This ranking does not show any inconsistency with the first ranking. Without violating the requirement of transitivity, we may perfectly accommodate both orderings in the following single ranking:

1. n violations of constraint C

2. n + 1 violations of constraint C

3. prevention of n + 1 violations of constraint C by n violations of constraint C

Thus, it might not be true that rationality requires rejecting the existence of deontological constraints. For we might deny that the violation of constraint C with the goal of preventing a greater number of violations of constraint C, is the same type of action involved in the violation of constraint C. And then, we might claim that n non-preventive violations of constraint C are morally better than n + 1 preventive violations of constraint C.

It has been argued that the consequentialist conception of the right attempts to regulate the conducts of agents by reference to the features we would like to see in the world. But it is not clear why deontology, understood as the theory affirming the existence of deontological constraints, must be opposed to this idea. Given the initial conditions, it is our refraining from violating a deontological constraint that brings about the best state of affairs. Deontology might then be understood as requiring what its critics have been insistently demanding: that each agent in all cases act in such a way as to produce the highest-ranked state of affairs that he is in a position to produce. Thus, deontology might not need to express a commitment to a conception of rationality that is different from the one embraced by consequentialism.[19]

IV. Two Possible Objections

It might be claimed that the distinction between preventive and non-preventive violations is an arbitrary distinction. It might be argued that both descriptions are, after all, descriptions of the same type of state of affairs, for example, the killing of one innocent person. How then could deontology have a preference for one over the other?

This worry is not well founded. The concept of a preventive violation may pick out the same type of state of affairs as the concept of a non-preventive violation; for example, the killing of one innocent person. But this is so only relative to a particular description, one that precludes any reference to causal histories. And there seems to be nothing either irrational or uncommon with a preference ordering characterized by such pattern of individuation of states of affairs. In my ranking of possible states of affairs regarding the outcome of my future tennis match, winning ranks first, but winning out of pity ranks below that of loosing the match. In my ranking of possible states of affairs regarding my future marital status, ending up married to the woman I love ranks above ending up single. But ending up married to the woman I love because of her interest in my money ranks below that of ending up single. The common feature of these rankings is that the value of a certain state of affairs—winning the tennis match or ending up married to the woman I love—is regarded as conditional on a particular characterization of its causal history. This could be how deontology regards the instantiation of certain states of affairs.

As John Broome has shown, there must be such things as rational principles of indifference.[20] Since a device of fine individuation as the one presented here in defense of deontology is available to anyone who has apparently inconsistent preferences, such principles must determine which differences are not enough to justify a preference.[21] Otherwise, as Broome explains, reason would have no role in guiding people through life.[22] Those rational principles of indifference, in general, establish that it is not rational to have a preference between two alternatives unless they differ in some relevant respect. Is the deontological individuation of states of affairs violating any such principles? It is hard to see how it could. The distinction between a particular violation of constraint C from a violation of the same constraint with a different causal history does not seem to be an arbitrary distinction. Many of our ordinary preferences are based on the same type of distinction, and it is implausible to claim that it would make no sense to act upon them.

It might also be claimed that under the understanding of deontology presented here the paradox has been merely pushed up one level. If it is true that, given two states of affairs A and B, A is morally better than B if the number of violations of a moral constraint in A is smaller than the number of violations in B, then we should also say that A is morally better than B if the number of violations of a constraint on preventive violations in A is smaller than the number of violations in B. Thus, in order to achieve the best possible state of affairs, we should be allowed to perform a preventive violation if doing so is necessary to prevent five other preventive violations. Since deontology forbids us to do this, the charge of irrationality has not been answered.

But the prohibition on preventive violations is best understood as a prohibition on a general type of action. The deontological principle for ranking states of affairs places a state of affairs defined by the occurrence of such a general type of action lower than one defined by its absence. So it does not matter whether a constraint C could be a constraint on preventive violations. Because if that is the case, the possibility of a new preventive violation arises, one that is relative to that particular constraint. By violating constraint C in order to prevent a greater number of violations of constraint C, we are simply performing the type of action that will bring about a state of affairs that is ranked lower than the one resulting by our refusal to violate it. In other words, it is true that A is morally better than B if the number of preventive violations in A is smaller than in B. But, again, it is not true that the avoiding of B by means of A must be better than B.

V. Conclusion

Under the usual understanding of the problem, what is challenged is the deontologist conception of the right. The problem, it is usually claimed, is not about value but about our attitude towards it. “You tell me what the good is, and I will simply argue that we must maximize it”—the consequentialist says. So the question is: how could any restriction on what a person may do to promote the best state of affairs be justified? If the answer presented here is correct, deontology might not be committed to the existence of any restrictions on what a person may do to promote the best state of affairs. The difference between this understanding of deontology and its critics would be a simple difference in the corresponding rankings of moral goodness.[23]

But now it might be claimed that the deontological account of the right has been made plausible by using an implausible account of the good. Even granting that the distinction between preventive and non-preventive violations is not an arbitrary distinction, the critic might still want to ask: How could it be plausible to assign more moral weight to a preventive violation than to a non-preventive violation? To claim that n preventive violations of a given moral constraint are morally worse than n + 1 non-preventive violation of the very same constraint might indeed be counterintuitive. But, certainly, the consequentialist moral theorist would seem to be in no position to argue against deontology on such grounds. If we are then indeed forced to rethink our moral conceptions, endorsing an unfamiliar conception of value would seem to constitute a less dramatic revision than the rejection of deontological constraints that the consequentialist endorses.

[1] Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), p. 30.

[2] Samuel Scheffler, The Rejection of Consequentialism (Oxford: Clarendon Press, 1982).

[3] Nozick, Anarchy, State, and Utopia, p. 31.

[4] Scheffler, The Rejection of Consequentialism, p. 88.

[5] Ibid, p. 87.

[6] Ibid, p. 100.

[7] Shelly Kagan, The Limits of Morality (Oxford: Clarendon Press, 1989), p. 33.

[8] Samuel Scheffler, “Agent-Centered Constraints, Rationality, and the Virtues,” Mind 94 (1985): 409-419, p. 414.

[9] Philip Pettit, “Consequentialism,” in Peter Singer (ed.), A Companion to Ethics (Oxford: Blackwell Publishers, 1997), pp. 230-239, p. 19.

[10] Ibid., p. 18.

[11] Eric Mack, “Moral Individualism: Agent Relativity and Deontic Restraints,” Social Philosophy and Policy 7 (1989): 81-111, p.110.

[12] John M. Taurek, “Should the Numbers Count?” Philosophy and Public Affairs 6 (1977): 203-316, p. 307. For a critique of Taurek's arguments, see Derek Parfit, “Innumerate Ethics,” Philosophy and Public Affairs 7 (1978): 285-301.

[13] C. S. Lewis, The Problem of Pain (London: 1957), p. 104.

[14] Thomas Nagel, The View from Nowhere (Oxford: Oxford University Press, 1986), p. 176.

[15] Stephen L. Darwall, “Agent-Centered Restrictions from the Inside Out,” Philosophical Studies 50 (1986): 291-319, p. 79.

[16] Andre, Judith, “The Demands of Deontology are not so Paradoxical,” Journal of Philosophical Research 16 (1990-91): 408-410, p. 408.

[17] Kagan, The Limits of Morality, p. 31.

[18] Eric Mack, “Deontic Restrictions are not Agent-Relative Restrictions,” Social Philosophy & Policy 15 (1998): 61-83, p. 79.

[19] In other words, if consequentialism is defined only by reference to such a conception, consequentialism is entirely compatible with the existence of deontological constraints.

[20] John Broome, “Can a Humean be Moderate?” in his Ethics out of Economics (Cambridge: Cambridge University Press, 1999), pp. 68-87.

[21] The availability of such a device is the reason why Broome concludes that a Humean cannot be moderate. By this, he means that it is not possible to claim that preferences could be irrational only based on consistency constraints. Ibid.

[22] Ibid., p. 69.

[23] In our personal rankings, an analogous pattern of that found in the deontological ranking does not strike us as problematic. If there is nothing wrong with our own personal rankings involving valuations of states of affairs conditional on their causal histories, the consequentialist should not see any problem in a deontological ranking. After all, consequentialism seems to share the conception of rationality with individual decision theory. They are distinct only insofar as they have different maximandum; the latter deals with individual satisfaction, the former with impersonal good.