The Journal of Value Inquiry 35: 455–476, 2001. REPLIES TO THREE CRITICS © 2001 Kluwer Academic Publishers. Printed in the Netherlands.
455
Replies to Three Critics BERNARD GERT Department of Philosophy, Dartmouth College, Hanover, NH 03755, USA
1. Reply to James Sterba James Sterba has problems with the change in my position from The Moral Rules to Morality. He is aware that I have never accepted the view that rationality requires acting morally, but he thinks that in The Moral Rules I did at least try to show that “all rational agents are still rationally required to publicly endorse morality even if they are not rationally required to abide by its requirements.”1 He quotes the following passage from The Moral Rules: to refuse to advocate the attitude toward the moral rules that they be obeyed by all rational men with regard to all is to increase his chances of suffering evil at their hands. Since he increases his chances of suffering by not advocating this attitude, and has no reason for not advocating it, I conclude that it would be irrational for him not to advocate it.2 However, the above argument does not work; it does not show that it is irrational not to publicly advocate the moral attitude toward the moral rules. The argument does not work because it depends on a person restricting his beliefs to beliefs that are shared by all other rational persons, what I call “rationally required beliefs.” It was only with this restriction that the argument works, and even then, it works only if “publicly advocating” means advocating out loud or to someone. Even with the restriction to rationally required beliefs it does not follow that all rational persons must actually support morality. All that follows is that if all they know about the persons to whom they are talking is that they have all of the rationally required beliefs and desires, it would be irrational to refuse to endorse the attitude toward the moral rules outlined in the quoted passage. This doubly hypothetical argument is correct. Using only rationally required beliefs and trying to justify your attitude toward the moral rules to persons about whom you know nothing except that they also have all of the rationally required beliefs and desires does result in your endorsing the moral attitude toward the moral rules, which is that everyone obey them impartially with regard to everyone. But since both of these conditions need to be satisfied
456
BERNARD GERT
and a person can be rational without satisfying either of them, it is misleading to claim that it is irrational not to endorse morality. No rational person is actually limited to rationally required beliefs and rational persons need not be interested in justifying their attitude toward the rules to others about whom they know nothing except that they have all of the rationally required beliefs and desires. Thus it is misleading to make the categorical assertion that all rational persons are rationally required to endorse morality, even hypocritically. As Sterba points out, when people have rationally allowed beliefs about their characteristics, what groups they belong to, and to whom they are talking, it is clearly not rationally required to even hypocritically endorse the moral attitude toward that moral rules. In the 1998 edition of Morality, I explicitly list the conditions that are necessary in order to reach the conclusion that all rational persons would endorse the moral attitude toward the moral rules. The first condition is the restriction to rationally required beliefs, but as Sterba notes, a further condition is also necessary. Three different conditions can serve this purpose: the additional condition can be either that the person is proposing an attitude to which all rational persons can agree; that the person is impartial with regard to all moral agents with respect to obeying these rules; or that the person regards the ten rules as requirements of an informal public system applying to all rational persons. I summarize this last condition by saying that the rules be considered as moral rules.3 A fourth condition, trying to justify your attitude toward the rules to someone about whom you know nothing except that he is a moral agent, is derived from some comments by Ernst Tugendhat when he compared my justification of morality in The Moral Rules to that in Morality and, like Sterba, preferred the former to the latter. Because Sterba mistakenly regards all the additional conditions as impartiality constraints, he claims that the new conditions cannot really be used to justify morality, because they presuppose it. Moreover he holds that once any of the other conditions are introduced, there is no need for the restriction to rationally required beliefs. This is also mistaken. Even being impartial with regard to all moral agents with respect to obeying the rules, does not require taking the moral attitude toward them, if the person has some religious belief concerning how God wants people to behave. The primary use of the restriction to rationally required beliefs was always to guarantee that the moral system could be understood by all rational persons. Sterba agrees that some restriction like this is needed so that morality or the common moral system, which applies only to those who are not justifiably ignorant of what it prohibits, requires, discourages, encourages, and allows, applies to all normal adult persons. The restriction to rationally required beliefs applies only to the moral system itself, and was intended to eliminate beliefs of which some moral agents were justifiably ignorant. It is primarily “an accessibility restraint,” not an attempt to eliminate biased beliefs, as Sterba claims.4
REPLIES TO THREE CRITICS
457
This restriction does have some strong consequences, for it entails that sophisticated scientific beliefs, of a kind that are not shared by all rational persons, cannot be used to establish the moral system. This means that all attempts to base morality on biology, sociobiology, or any other science or pseudo-science, are ruled out, not because of the falsity of such beliefs, but simply because they are not shared by all those to whom morality applies. Similarly, morality cannot be based on any religion, not because what a particular religion would have us hold is false or biased, but because no religion is known by all rational persons about whom it is appropriate to make moral judgments. Because Sterba does not distinguish between the moral system itself, and the application of that system in making moral decisions and judgments, he does not realize the importance of the limitation to rationally required beliefs. It is clear that Sterba needs to have this same kind of restriction to rationally required beliefs if he regards morality as applying to all rational persons. If an egoist has certain beliefs about God, such as that God punishes egoistic acts with eternal damnation and rewards altruistic acts with eternal bliss, that egoist and an altruist will not end up with Sterba’s compromise; both will favor pure altruism. Furthermore, far from my other conditions begging the question in favor of morality, the first of my conditions, trying to propose an attitude to which all rational persons can agree, is a non-question begging condition with regard to more people than is Sterba’s non-question begging compromise between a pure egoist and a pure altruist. Proposing an attitude to which all rational persons can agree is a condition that guarantees that a person is not begging the question with regard to any rational person. Sterba should simply adopt this condition as his own. Such a condition, when conjoined with the restriction to rationally required beliefs does in fact yield an informal public system that applies to all fallible, rational, vulnerable persons and which is virtually identical to the moral system that we all accept as our common morality. This is in stark contrast to the guide that Sterba derives from his non-question begging compromise between a pure egoist and a pure altruist. That guide seriously deviates from our common morality in that it classifies acting on a low ranking altruistic reason rather than a high ranking egoistic reason as immoral. Indeed, although Sterba calls his compromise between an egoist and an altruist “morality as compromise,” he never shows that the compromise yields morality rather than simply a compromise between egoism and altruism. He admits that he never appeals to “morality as compromise” to resolve any moral problem. The completely counter-intuitive conclusion that it is immoral for a person to act on a low ranking altruistic reason rather than a high ranking egoistic reason, is sufficient to show that whatever Sterba does derive from his non-question begging compromise between a pure egoist and a pure altruist, it is not what we would regard as common morality. Sterba claims that
458
BERNARD GERT
“morality can be seen to be a non-arbitrary compromise between self-interested and altruistic reasons, and the ‘moral reasons’ that constitute that compromise can be seen as having an absolute priority over the self-interested or altruistic reasons that conflict with them.”5 Nonetheless, contrary to Sterba’s claim, what he labels as moral reasons are not moral reasons at all, but simply reasons that result from a compromise between pure egoism and pure altruism. Sterba has provided no argument showing that they should be taken as moral reasons. Indeed, on any normal understanding of high-ranking and low-ranking, Sterba’s guide misclassifies many cases. For example, if I had a computer scheme for transferring all the fractions of a cent in the bank accounts of other people to my own, it would seem that I had a high-ranking egoistic reason for doing so and that there was only a low-ranking altruistic reason for not doing so. The same seems to be true for many cases of cheating where no one else is hurt much, if at all, by a single act of cheating on an exam, but I may avoid the significant harm of flunking. In order to give any plausibility to his view that a compromise between a pure egoist and a pure altruist yields morality, Sterba must incorporate clearly moral values like fairness as high ranking altruistic reasons. Showing that in a debate between a pure egoist and a pure altruist, the only non-question begging reasons are reasons which involve a compromise between pure egoism and pure altruism is a relatively simple and almost tautologous conclusion. What Sterba never even tries to show is that the compromise reasons he advocates are moral reasons. He simply labels them as such. Furthermore, he never even tries to show that the debate should be between pure egoists and pure altruists. Indeed, on the correct account of our ordinary concept of normative rationality, a pure altruist, who chooses to act on weak altruistic reasons rather than strong egoistic reasons, so that he suffers great harms himself in order to gain small benefits for others, acts irrationally. He should not be a participant in any debate about morality, and any compromise between his irrational view and a rational view is likely to contain some irrational elements. But a justified morality should not have any irrational elements. It is not surprising that the guide that Sterba derives from his compromise between the irrational position of pure altruism and pure egoism has only a faint resemblance to common morality. Limiting the participants in the debate about morality to a pure egoist, who most closely resembles what is known in psychiatry as a sociopath, and a pure altruist, who often acts irrationally, is not likely to result in a guide to behavior that anyone should take very seriously, let alone a plausible description of the guide to behavior that is provided by common morality. Far more plausible representatives of anyone seeking to compromise are egocentric persons and impartial persons. Persons of both types holding rational positions might actually reach a compromise
REPLIES TO THREE CRITICS
459
of somewhat greater complexity, and of far more value, than the compromise reached by egoists and altruists advanced by Sterba. If an egocentric person, who acts only on reasons that benefit himself and those for whom he cares, and an impartialist, who acts only on reasons that benefit people impartially, taking no account of whether they are friends or strangers, were to compromise, they might reach the following conclusion. If an action is a violation of a moral rule, then a person must consider all persons impartially, not basing his violation at all on who is to be benefited and who will be harmed. However, if an action is not a violation of a moral rule, then a person may act on reasons that benefit those for whom he cares. This provides an impartialist with impartial protection from egocentrically motivated violations of the moral rules, but allows an egocentric person to act egocentrically in all situations other than violations of moral rules. Unlike the simple non-question begging compromise between a pure egoist and a pure altruist, the non-question begging compromise between an egocentric person and an impartialist such as a pure act consequentialist, actually could yield our common morality. This more creative non-question begging compromise between two rational views, together with a restriction to rationally required beliefs and recognition that morality only applies to vulnerable and fallible persons, actually yields the complex structure that is common morality. Sterba’s simple compromise between views that no sane person actually holds, completely ignores the internal structure of morality. Part of the problem is that Sterba, like most philosophers, drastically oversimplifies our common morality. He does not realize that morality is a public system that has many complex features. The guide that Sterba derives from his compromise bears only a loose relationship to the guide provided by common morality, it is even farther from morality than the guide provided by extreme act consequentialists. However, Sterba makes some of the same mistakes as extreme act consequentialists. Sterba and extreme act consequentialists regard it as immoral to ever spend any money on luxuries for yourself, a low ranking egoistic reason, rather than donating that money to charities that provide necessities to others, a high ranking altruistic reason. Neither common morality, nor the guide provided by Kant, would have us hold this extreme view, although both would have us regard it as morally good to donate money to worthy charities. Furthermore, according to common morality and almost all philosophical accounts of morality except Sterba’s and extreme act consequentialism, it is immoral to cheat even if I have a great deal to gain by cheating and no one will lose anything by the particular act of cheating. Sterba is mistaken in thinking that there is a unique right answer to the question how people should live, egocentrically or impartially. If any fully informed rational person who is concerned with someone could advocate that she live either egocentrically or impartially, then it is appropriate for such a
460
BERNARD GERT
rational person to say that she ought to act in that way.6 In “Gert, Sidgwick, and Hybrid Theories of Rationality,” David Phillips points out that Sidgwick holds this exact position.7 When acting on moral ideals and no moral rules are being violated, many rational people, including John Stuart Mill and Phillips, favor being concerned more with friends than with all moral agents, even though they disagree about how much more. When considering whether to violate a moral rule, however, there is general agreement that a person must be impartial with regard to all moral agents. Sterba never argues that there is a unique right answer to the question of how people should live. Like most philosophers, Sterba holds that egoism or self-interest is the primary alternative to morality. That is why philosophers are generally not taken seriously. Any non-biased look at the world shows that far more harm is caused by immoral actions due to various egocentric views such as nationalism, racism, and religious fanaticism, than by pure self-interest. An egoist is mistaken if he denies that beliefs related to the interests of others are reasons. It is a gross distortion of our concept of rationality to deny that beliefs related to the interests of others are reasons or to hold that people who act on such beliefs when it involves some sacrifice of their own interests are acting irrationally. To accept the pure egoist view of reasons would result in irrationality not being the basic normative concept, for that concept has the feature that no person who is held responsible for his action ever favors himself or anyone for whom he cares acting irrationally. To accept the view of a pure altruist has the same result. “Morality as compromise” seems better characterized as rationality as compromise. Indeed, Sterba does not clearly distinguish between rationality and morality, which is why he thinks he has shown that rationality requires acting morally. Unless irrationality is such that no person held responsible for his actions ever favors himself or his friends acting irrationally, it is of little significance to show that rationality supports morality. It is not the desire to be rational, but concern for others that provides the fundamental motive to act morally. Sterba claims to show that “it is rational for us to accept at the deepest level – in our heart of hearts” that moral reasons are the best reasons.8 However, the standard criticism of immoral people is not that they are not rational, but that they lack the appropriate concern for others. The beliefs that are the reasons for endorsing morality as an informal public system that applies to all rational persons are not identical to the beliefs that are the motives that lead most people to act morally. By trying to make them identical, Sterba distorts our ordinary view of immoral people, for he regards them as not being sufficiently rational when he should regard them as not caring widely or deeply enough. My account of common morality distinguishes between the motives that rational persons may have for acting on the moral rules and moral ideals and the motives that all of them would have for endorsing common morality as
REPLIES TO THREE CRITICS
461
an informal public system that applies to all rational persons, if they accept the two conditions discussed earlier. Unfortunately, their own self-interest, and the interests of their friends, family, country, race, or religion, sometimes provide motives that lead them not to follow the moral guide. It is pointless and counter-productive to regard all those who act contrary to the moral guide because of these motives as acting irrationally. Rationality does support morality, but not in the simple way that philosophers such as Sterba want it to. It is understandable why many philosophers are not satisfied with that level of support, but a careful examination of the concepts of morality and rationality shows that to provide a higher level of support distorts one or the other of these concepts. Sterba credits my account of morality with providing “a very explicit and useful account of the content of morality, but claims that his account of morality shows “that morality is rationally preferable to egoism.”9 In the important sense of “rationally preferable,” this is not true. Although Sterba’s conclusion that each of our views is superior to the other in some ways demonstrates his fairness and willingness to compromise, his conclusion should not be accepted. 2. Reply to David Phillips David Phillips agrees that a hybrid account of rationality allows “for a plausible account of the relationship between morality and reason.”10 However, he holds that his improved version of a hybrid theory is better than mine. Few, if any philosophers are pleased by the practical consequences of a hybrid account. Even Sidgwick, whom Phillips interprets as putting forward a “dualism of practical reason” is troubled by this dualism and would prefer that it not be true.11 I am not troubled by the consequences of a hybrid view and prefer the freedom it provides to the stronger guidance provided by non-hybrid views. My hybrid account correctly describes the fundamental normative concept of rationality in which no person who is responsible for his behavior would ever favor acting irrationally. On this account, when morality and self-interest conflict, it is rationally allowed to act in either way. A non-hybrid view such as rational egoism or the view that rationality requires acting morally, entails that in cases of conflict reason does not allow acting in both ways. Even the view that rationality requires acting on the best reasons does not allow acting in both ways, unless the reasons for acting both ways have equivalent strength. Neither Phillips’s interpretation of Sidgwick’s views nor his criticisms of them are my concern. For reasons other than those put forward by Phillips, Sidgwick’s hybrid theory of rationality is far too restrictive. Sidgwick provides no argument for his view that rationality allows only two points of view, either an egoistic view or the point of view of the universe. Contrary
462
BERNARD GERT
to Sidgwick’s view and the views of most philosophers, it is not irrational to take the interests of your children, country, race, or religion as more important than either self-interest or an undifferentiated interest in everyone. Those who sacrifice their lives, let alone their fortunes, for the benefit of their children, country, race, or religion are not normally regarded as acting irrationally, even when such actions are immoral. On my account of rationality, which is the ordinary concept of rationality when it is taken as the basic normative concept, no person who is responsible for his actions would ever advocate that anyone for whom he cares, including himself, act irrationally. But many such persons would advocate acting to benefit their children, country, race, or religion even when this is both immoral and contrary to their own self-interest. Accounts of rationality by philosophers are too simple. They do not reflect the complexity that is part of the ordinary concept. Philosophical accounts of rationality have so little influence because they assume that self-interest and morality provide the only rational motives for acting. Neither Sidgwick nor Philips explain what is involved in taking the point of view of the universe, nor is it clear that they could explain it in any satisfactory way. It is not possible to replace this way of talking with talk about taking an impartial stance, for the concept of impartiality requires reference to some group with regard to whose members impartiality is required and also the respect in which impartiality is required with regard to this group.12 Talk about taking the point of view of the universe allows them to ignore the problems involved in specifying the group with regard to which morality requires impartiality, and the respect in which it requires that impartiality. Phillips claims that on my view, “egocentric rationality has a special default status: non-egocentric reasons are less basic.13 But these claims are not quite right. It is true that the concept of a reason depends on the concept of an irrational action, and it is also true that an irrational action is defined egocentrically, but there is no concept of egocentric rationality. Both the concept of reasons and the concept of an irrational action are necessary to provide an account of the concept of rationality, so that rationality is neither egocentric nor non-egocentric. Phillips correctly notes that there are two categories of reasons, reasons of self-interest, and reasons related to the interests of others. Although it is never rationally required to act on reasons related to the interests of others, it is not true that non-egocentric reasons are less basic than egocentric reasons in justifying an otherwise irrational action. My definition of reasons is: “Reasons for acting are conscious rational beliefs that can make some otherwise irrational action rational.”14 This definition does not mention, entail, or use any distinction between egocentric and non-egocentric reasons. Furthermore, it is explicitly stated that a “mere change of persons does not change the strength of a reason.”15
REPLIES TO THREE CRITICS
463
Phillips recognizes that my account of rationality “is relatively unrestrictive” but does not appreciate that this is because I take irrationality to be more fundamental than rationality.16 As Philips indicates by quoting my definition of an irrational action early on, it is an irrational act that is defined in a positive way. There is no feature that all rational acts have in common except that they are not irrational. There are even two different categories of rational actions, those that are rationally allowed and those that are rationally required. Philips is correct in claiming I am extremely permissive about what actions are rationally allowed and could note that some rationally allowed actions have no reasons at all. Philips adopts the traditional view of providing a positive definition of a rational action. He adopts a Pareto optimal form of maximizing. This is apparent from his description of what he calls the improved hybrid view. Proponents of this view “will say that actions are rationally justifiable just in case they have the best outcomes when a person’s own interests are given some amount of special weight while otherwise everyone’s interests are weighed equally.”17 Pareto optimality is one of the more plausible versions of a maximizing view of rationality, but Philips’s arguments for limiting special weight to a person’s own interests are not persuasive. Philips’s adoption of a Pareto optimal view is persuasive if a person is proposing some policy to be adopted by contesting parties. In such a situation, it would be irrational not to propose Pareto optimal views because it would be irrational for any contesting party to accept a non-Pareto optimal solution with regard to her own interests. However, from the irrationality of a non-Pareto optimal solution to a policy question when the goal is to promote the interests of everyone as much as possible, nothing follows about the irrationality of a non-Pareto optimal decision when there is no such goal. Nor is it rationally required to adopt such a goal. If rationality required always making Pareto optimal decisions, then all acts of casual vandalism, where a person simply destroys the property of others when he can do so without any fear of harm to himself, would be irrational. The claim that the vandal gains pleasure from his vandalism, so that refraining from vandalism would not be Pareto optimal, is not true in all cases. Just as sometimes a person simply feels like picking a flower, or stepping on an ant, so some people sometimes simply feel like destroying the property of others, or perhaps worse. If there is no chance of being caught, such actions are no more irrational than picking a flower or stepping on an ant. All three may simply be done on a whim and without any thought of pleasure or gain. On Philips’s view, if a person risks harm to himself to protect a colleague even though such a cover-up is immoral, he is acting irrationally. Also, if a person is prepared to sacrifice her own interests for the greater interests of others, but does not act to promote their interests as much as possible, she is acting irrationally. When no examples are provided of the choices being made,
464
BERNARD GERT
it does seem unlikely that a person who is prepared to sacrifice her own interests for the greater interests of others would not seek to promote their interests as much as possible, as long as there were no additional costs to herself. But, it is not irrational to be prepared to sacrifice self-interest to prevent others from suffering serious harm, but still be indifferent to their gaining any additional benefits. That such an attitude may be unusual does not mean that it is irrational, at least not if being irrational is to carry the basic normative force that makes it the fundamental concept. Phillips is not satisfied with a Pareto modification of my account, for he requires that, except for the agent’s interests, everyone’s interests be weighed equally. He thinks that if a person sacrifices the same amount of her own good to gain lesser benefits for one person rather than greater benefits for another she is acting irrationally. For Phillips, it is irrational to flip a coin to decide which person will get the benefits if the benefits to one of them would be greater than the benefits to the other. But if the coin flipper is indifferent between the good of the two persons, why is it irrational not to care who gets the benefits or how much either gets? Philips simply uncritically accepts a maximizing point of view without realizing that this results in almost all people acting irrationally almost all of the time. Part of the plausibility Phillips’s view depends on his uncritical acceptance of the phrase “the point of view of the universe” as expressing some clear point of view. He offers as a fundamental intuition upon which the improved hybrid rests, the following. “It is rationally permissible to take the point of view of the universe, to weigh everyone’s interests equally.”18 Philips does not tell us who is referred to by “everyone.” Does he mean that it is rationally permissible to weigh the interests of some bacteria equally with the interests of moral agents? Perhaps Phillips holds that bacteria do not have interests, that only sentient beings have interests. Then, consider the most primitive form of sentient being and ask whether the intuition that a person should weigh the interests of that being equally with that of moral agents is intuitively plausible. Phillips has no answer to the question: whose interests should a person weigh equally with the interests of moral agents? There is no rationally required answer to this question. In my discussion of impartiality in Morality: Its Nature and Justification, I show that there is no unique answer to the question, “Toward what group does morality require impartiality?” that would be accepted by all rational persons.19 Phillips claims that “the special status of an individual’s own interests stands in need of justification for Gert,” but he does not explain why this is so.20 Strongly justifying a decision involves showing that all rational persons would make the same decision; weakly justifying a decision involves showing that all rational persons could make the same decision. But the only decisions that “need justification” are those that would be irrational if there were no justification. Phillips does not even try to show that it is irrational for people to give
REPLIES TO THREE CRITICS
465
their own interests special status. Instead, he claims that he “needs to show why giving special weight to an individual’s own interest is not arbitrary.”21 But why care if it is arbitrary for an individual to give her own interests special status, unless Phillips, without arguing for the point, holds that being arbitrary is irrational. But that is a completely implausible view. Decisions can be made arbitrarily without being irrational. Indeed, there seems to be almost no relationship between acting arbitrarily, if this means acting without a reason, and acting irrationally. If there is no harm or danger to myself, I can decide to go for a walk for no reason at all, and yet not be acting irrationally. Phillips’s claim that “giving special weight to the interests of her friends, conationals, or conspecifics is arbitrary,” is also completely irrelevant to the discussion of rationality.22 Phillips seems to be claiming that, unless my own interests are involved, it is irrational to be willing to sacrifice my life for my children, but not be willing to sacrifice my life for the children of other people. He offers no argument for this. It is a position that conflicts with the ordinary view of rationality that allows for almost unlimited arbitrariness. Love may always be arbitrary, but it need never be irrational. Phillips limits rational motivation to two categories, one is the same as one of the categories advanced by James Sterba, egoism, but the other is the more plausible impartialist view, weighing everyone’s interests equally, rather than altruism. Phillips thinks that the two categories he advances are the only nonarbitrary categories, and perhaps he is right about that, but he provides no clear criteria for arbitrariness. However, since arbitrariness is not related in any significant way to rationality, Phillips seems to have changed the topic. Unless an arbitrary act is irrational or immoral it needs no justification. The only action that needs rational justification is harming yourself, and that justification is provided by reasons related either to your own interests or to the interests of others. It is irrational not to be motivated by your own interests; it is not irrational not to be motivated by the interests of others. People who are not motivated by anyone’s interest but their own, are what philosophers call psychological egoists, and psychiatrists call psychopaths or sociopaths. But people who are not motivated by the interests of anyone outside of their group, especially if the group is large, such as members of your family, country, race, or religion, are not only not regarded as irrational by most people, they are most people. Nonetheless, most people realize that it is not irrational to sacrifice your interests for the greater interests of those outside of your group, even though such behavior is relatively rare. Phillips claims to be holding a hybrid view, but he does not hold the most important feature of my view, namely, that rational actions share no common features except not being irrational. A rational action is simply an action that is not irrational. This is not a mere tautology, for although everyone agrees that all rationally allowed actions are not irrational, most philosophers, in-
466
BERNARD GERT
cluding Phillips, hold that all rationally allowed actions do share some feature besides being not irrational. Many share Phillips’ view that all rationally allowed actions must be maximizing. It is only by limiting irrational actions to actions in which people harm themselves and by allowing reasons to be beliefs about the benefits to anyone, that the category of rationally allowed actions encompasses the wide variety of actions that are ordinarily regarded as rational. Only this very permissive view of rationality guarantees that no person who is responsible for her actions ever favors herself or those for whom she cares acting irrationally. Phillips never provides his own account of a reason because starting with a positive account of a rational action, talk about reasons is equivalent to talk about rational actions. A rational action is defined simply as that action which a person has the best reasons to do. Thus persons need reasons for all of their actions, even taking a walk. The contrast with defining a reason as a belief that can make some otherwise irrational action rational, is startling. On the account that begins with irrational actions, a person needs a reason for acting only if his action would be irrational otherwise. Although Phillips’s improved hybrid view is somewhat more complex than the standard maximizing view, it is essentially the same; the only difference is that the individual decides how much special weight to give to his own interests. It is the failure to recognize that most reasons have the special role of justifying otherwise irrational actions that leads Phillips to ask, “How can a certain consideration, say that a possible action will save another’s life, have a justifying force both powerful enough to justify seriously harming ourselves and weak enough to ignore completely?23 Since he does not deny that it is rationally allowed to sacrifice your life to save another person’s life, he must deny that it is rationally allowed not to be motivated to save the lives of those who are not members of your group. Is it also irrational not to be motivated to save the lives of non-human animals? How much would you have to be motivated not to be irrational? Philips simply assumes that the justifying strength of a reason and its requiring strength are always the same. Philips’s concept of irrationality fails the test that any adequate concept of irrationality as the basic normative concept must pass, that no person who is responsible for her actions ever favors herself or those for whom she cares acting irrationally. Phillips accepts only a very limited category of rationally allowed actions, only those actions which maximize either your own interests or the interests of all. Phillips may be conflating the concepts of reasons and motives. The strength of reasons is an objective matter, and if a reason is strong enough to justify seriously harming yourself, it is a strong reason. It is unclear what he means by asking how that same reason can be “weak enough to ignore completely.”24 Whether people ignore a reason or not, is a matter of what beliefs motivate them. If people are not motivated at all by beliefs that they will avoid harm to
REPLIES TO THREE CRITICS
467
themselves, then they are irrational, but if they are not motivated at all by beliefs that they will avoid causing harm to some others, they are morally bad persons, though still rational. The strength of a reason is determined by what actions it can justify, not by how strongly it motivates a rational person.25 Some rational persons who are morally bad are not motivated to avoid causing harm to others. As long as they are motivated to avoid harm to themselves, they are considered rational and are not excused from responsibility for their actions. If Phillips accepts the category of moral reasons, perhaps the following analogy will be persuasive. Suppose that a person has made a promise to meet someone for dinner. An ordinary dinner, not one at which momentous decisions will be made. On the way to the dinner, he passes a house on fire, with many people watching a child trapped inside. He has no better training or skills than the people who are watching and it is unlikely that he will be able to affect a rescue, but he decides to try. Everyone would agree that trying to rescue the child morally justifies his breaking the promise to meet someone for dinner. However, everyone would also agree that he is not morally required to try to rescue the child. This is a case where trying to rescue a child is a strong enough reason to morally justify breaking a promise, but the person is not morally required to try to rescue the child even if there were no conflicting moral requirement. Because moral requirements and moral justifications are not the same, something can provide a strong moral justification in those cases where a moral justification is needed, but still not provide a moral requirement. There is a strong analogy between moral requirements and moral justifications and rational requirements and rational justifications. Preventing great harm to others can rationally justify an otherwise seriously irrational act, but preventing this harm is still not a rational requirement.26 My hybrid account of rationality has an egocentric account of irrationality and a non-egocentric account of reasons. This account creates rationally allowed actions with no common feature, except not being irrational. However, on Phillips’s hybrid theory all rationally allowed actions are supported by the best reasons. He calls his account a hybrid account of rationality because in a conflict between your own interests and the interests of all others, it is sometimes rationally allowed to act in either way. My hybrid account limits irrational actions to those involving harm to yourself, while allowing reasons involving preventing harm to others to make harming yourself rationally allowed. Only such a hybrid account provides an account of rationality such that every person who is subject to moral judgment always favors acting rationally, in the sense of never favoring acting irrationally. It is also the only one that allows us to accurately describe the coherent use of that concept. It is only by limiting irrational actions to actions involving harm to self, recognizing that preventing harm to anyone counts as a reason, and defining a rational action as any intentional action that is not irrational that the appro-
468
BERNARD GERT
priate category of rationally allowed actions can be accommodated. This hybrid account of rationality conflicts with all of the standard philosophical accounts of rationality as acting on the best reasons, that Philips and other philosophers seem to take for granted. However, it is the only account that allows us to accurately describe the ordinary concept of rationality when it serves as the fundamental normative concept. Phillips simply accepts many of the standard philosophical intuitions about rationality without argument, and shows that my account of rationality conflicts with these intuitions. His improved hybrid view of rationality is a maximizing non-arbitrary hybrid account, but it does not allow us to accurately describe the ordinary fundamental normative concept of rationality. Phillips’s account may be congenial to philosophers, but neither it nor any other philosophical theory of maximizing rationality passes the crucial test that no person who is held responsible for his actions would ever advocate that he or anyone for whom he cares act irrationally. If their accounts do not pass this test then they are not defining the fundamental normative concept of rationality. Phillips does not even try to show that my hybrid theory does not pass this test, or that his does, or that this test is not an appropriate test. 3. Reply to Patrick Yarnell Patrick Yarnell has fundamental objections to my account of rationality. He denies my claim that basic reasons must be beliefs about the present or the future, and claims that beliefs about the past can also be reasons. He also objects that what I claim to be paradigm cases of evils, such as sadness and guilt can sometimes be classified as intrinsically desirable or good. These two objections are meant to support his main objection, that a person can sometimes act in a way that he knows will result in his suffering emotional pain or loss of pleasure, without having any of my reasons and yet be acting rationally. Yarnell’s objections involve two mistakes. The first is that he takes all reasons to be basic reasons and the second is that he sometimes uses “reasons” to refer to what I call motives. The first objection is based upon his failing to distinguish basic reasons from derived reasons. In Morality: Its Nature and Justification, I say “beliefs about the past count as reasons only when they are related to basic reasons in the appropriate way.”27 The following discussion of punishment contains the acknowledgment that the fact that someone committed a crime counts as a derived reason for punishing him but makes clear that this is only because punishment for that crime is part of the institution of punishment which is justified by basic reasons related to the present and future. The discussion of gratitude also shows that beliefs about the past are not basic reasons.28
REPLIES TO THREE CRITICS
469
Yarnell’s main objection, that it is sometimes not irrational to want to continue to feel sadness or guilt, is a problem also mentioned in Morality: Its Nature and Justification: “Desires to feel sad, anxious, or displeased are irrational. It is also irrational to desire to feel the various emotions, such as guilt, shame, and remorse, that involve these unpleasant feelings. Often one has reasons for wanting to have these feelings and emotions.”29 However, none of these reasons are provided. I note that “the popularity of tragedies, horror films, and works of art that cause people to feel outrage,” must also be explained.30 No explanation is given in Morality but I say “I hope to discuss [these matters] more fully later in a book on human nature.”31 Yarnell regards these problems as far more serious than I think they are and claims that they require a fundamental revision of my account of rationality. Yarnell provides three excellent examples to support his primary point that someone can refuse to do that which will eliminate experiencing unpleasant feelings, but not have any reasons, in my sense of “reasons,” for such refusals, without being regarded as acting irrationally. Two of his examples involve someone who refuses to take a happy pill which it is claimed will enable the person to cease having unpleasant feelings or to continue having pleasant feelings, without any harmful side effects. The immediate point of these examples it to show that my claim that reasons are limited to beliefs about the present or the future is mistaken, and that beliefs about the past can serve as reasons. Though he does not explicitly limit himself to reasons related to self-interest, most of the reasons that Yarnell considers and rejects as plausible reasons for not taking a happy pill are reasons related to self-interest. He is correct that there do not seem to be any conscious beliefs about self-interest that would justify refusing to take a happy pill to relieve suffering grief at the death of a close friend, or feeling guilt because of sadistic pleasures. He is also correct that we do not regard such refusals as irrational. If Yarnell is right about his first and third cases, both of which involve refusing to take the happy pill, not only is the account of reasons as beliefs about the present and future incorrect, the entire account of rationality is fundamentally flawed. Yarnell says that if we accept his account of his examples, then “under some circumstances, it is rational to intrinsically desire evils such as emotional anguish and loss of pleasure.”32 He even goes further, saying that “Pain and the loss of pleasure, then, are, at least in certain situations, intrinsically desirable. If what is intrinsically desirable is intrinsically good, pain, anguish, and loss of pleasure are, in some cases, intrinsically good.”33 Yarnell is attacking the foundation of the account of rationality in which the list containing death, pain, disability, loss of freedom and loss of pleasure, is used to determine both what counts as an irrational action and what counts as a basic evil or harm. If Yarnell is correct, then not only is my account of rationality fundamentally flawed, but all other accounts of rationality such as
470
BERNARD GERT
classical utilitarianism and all other views which would have us regard pleasure as intrinsically good and pain as intrinsically bad are also fundamentally flawed. I do not use the phrase “intrinsically desirable” or intrinsically good,” for I find these phrases unclear, but Yarnell could make his point without using them. The list of personal goods is a list of things that no person insofar as he is acting rationally would avoid without a reason, and the list of personal evils is a list of things that all such persons would avoid unless they had a reason. Yarnell can be taken as claiming that some pain, particularly grief at the loss of a friend and guilt at taking pleasure in the suffering of others, are not only not personal evils, but personal goods. This conclusion is extremely troubling, and not merely because it is incompatible with my account of rationality. My way of putting the point is different from what Yarnell says. He writes about sadness and guilt being intrinsically desirable or personal goods in certain situations, whereas I have changed that to certain kinds of grief or guilt being intrinsically desirable or personal goods. The idea of something being intrinsically desirable or a personal good in some situations but not in other situations is a misleading way to describe these cases. A better description of such cases is to describe these different situations as providing adequate reasons for desiring a personal evil. In fact, that is the way that Yarnell initially and primarily presents his examples. He claims that these situations show that beliefs about the past can serve as adequate reasons for wanting to continue suffering some unpleasant feelings. Yarnell’s examples can be taken in two ways, both of them incompatible with my account of rational action. One way is to take the past as providing reasons for suffering unpleasant feelings. The second way is to take unpleasant feelings in certain situations as being intrinsically desirable, good, or a personal good. This second way resembles G. E. Moore’s principle of organic unities. Moore allows that although pleasure is normally taken as an intrinsic good, pleasure in the undeserved pain of another person is an intrinsic evil. If my reply to this second way of interpreting his objection is adequate, it will also be an adequate reply to Yarnell’s first formulation. An immediate objection to Yarnell’s first and third cases is that they involve taking a happy pill. Someone claims that the pill allows a person to cease feeling grief for the death of his friend but with no untoward side effects. The person who offers the pill even claims that “it won’t compromise the capacity for sorrow in reaction to other events.”34 But there is no reason whatsoever to believe what she says. The claim of no untoward side effects must include the claim that the pill will not affect your memory. That means that you will not only remember the good times that you had with your friend, you will also remember that he was killed in an auto accident. When you remember this, does the pill make you indifferent to that event? But how could it possibly do that without having all sorts of other effects? If a psychiatrist
REPLIES TO THREE CRITICS
471
knows you and thinks that your grief is excessive, she might prescribe an antidepressant, and if your grief was extreme or interfering too much with your life, it might be irrational to refuse to take the anti-depressant. Of course, being depressed often involves wanting to stay depressed, but that is because being depressed is a mental disorder and often causes a person to act or want to act irrationally. Stoics used to claim that people ought to train themselves not to feel sorrow or grief over things not in their control, like the death of a friend, but most people do not think that it is possible to have a friend and not feel sorrow or grief over his untimely death. Yarnell himself says: “To be a friend to someone is, among other things, to intrinsically care for the well-being of the other individual. Intrinsically caring for another essentially involves a certain disposition, the disposition to experience sorrow at the other’s serious misfortune.”35 Since Yarnell claims that being a friend involves having the kind of disposition that friends have, if there are basic reasons for being a friend, then there are reasons for not taking a happy pill. Being a friend is clearly supported by several basic reasons, such as the pleasure of friendship as well as the increased opportunities to give and receive pleasure. Thus, on my account of rationality, a person is not acting irrationally in refusing to take a pill that would weaken the disposition that is essential to being a friend, though, it would also not be irrational to take the pill either. That there are basic reasons for being a friend does not mean that the motives for being a friend are pleasure and increased opportunities for pleasure. Someone who had these as his only conscious motives for being a friend would not have the kind of disposition that friends are supposed to have. Such a person would certainly take the happy pill if it were offered to him, which shows that it is not irrational to view taking the pill as weakening your disposition as a friend. Taking the pill could count as a kind of betrayal of the friendship. But regardless of the motive for being a friend, friendship is supported by the basic reasons. This requires a revision in my account of a reason. The claim that a reason is a conscious rational belief about gaining benefits and preventing harms can be taken as entailing that if a person does not have such a conscious rational belief then he does not have a reason. This would mean that having a derived reason would require having some conscious belief about the relationship between that derived reason and basic reasons. Then the belief that taking a happy pill would weaken your maintaining the disposition of a friend would not count as a reason unless you had some conscious belief about the relationship between being a friend and basic reasons. Although nothing counts as a derived reason unless it is actually supported by basic reasons, a derived reason can be a reason even if a person has no conscious beliefs about the basic reasons which support it. Indeed, I now hold that the facts that correspond to the appropriate rational conscious beliefs to be basic reasons. Eve-
472
BERNARD GERT
ryone knows that being a friend not only benefits yourself, it confers benefits on those who are your friends. Thus being a friend serves as a derived reason, and it is rational to refuse to take a happy pill because taking it is incompatible with being a friend. In Morality I discuss what I call the virtuous answer to the question “Why should I be moral?”.36 I point out that a person may aspire to a good character. Having all the virtues, includes not only acting in the appropriate ways, but also having the appropriate motives and feelings. Similarly, being a friend requires not only acting in certain ways, but also involves having the corresponding motives and feelings. It is not only not irrational to want to be a friend, it is not irrational to be willing to suffer some significant harms in order to achieve or maintain being a friend. It is not irrational to think that taking a happy pill will adversely affect being a friend. Taking the pill will result in knowing that if something terrible happens to another friend you can simply take another pill and avoid the grief. But that makes it more likely that the intensity of your friendship will suffer; it suggests that you are not willing to suffer for your friends. I agree with Yarnell that friendship essentially involves “the disposition to experience sorrow at the other’s serious misfortune.”37 Given this account of friendship, it is not irrational to refuse to take the happy pill to relieve normal grief at the death of a friend. All of this involves having certain views about human nature. If human beings were like Mr. Spock of Star Trek, so that intense friendship is coupled with complete indifference once the friend died, it might be irrational to suffer grief at the death of a friend when this serves absolutely no purpose. But, as Yarnell correctly notes, human beings are not like that. To take a happy pill is to compromise your view of yourself as a friend. That provides you with a reason not to take the pill, for being a friend is supported by basic reasons. Thus, you do have a reason for refusing to take the happy pill; it is not a basic reason, but it is supported by basic reasons. To be a friend makes it more likely that you will benefit others, as well as making it more likely that you will benefit yourself. To see the correctness of this answer, let us add some details to Yarnell’s science fiction example of a happy pill. Suppose that the person who offers you the pill has developed a fantastic ability to predict how various events will affect you. Accepting and building on Yarnell’s science fiction examples, you not only believe what the person says about the pill, you also believe his predictions about what will happen if you take the pill and if you do not. He predicts that if you take the pill, you will retain the close and intense friendships that you now have, but that if you do not take it, you will become bitter and end all of your friendships. You will be so affected by your grief that you will refuse to let yourself become involved with anyone in any intense way at all. Given this information, it would not still be rational to refuse to take the pill. This example shows that it is the present or future effects of your taking
REPLIES TO THREE CRITICS
473
the pill, including its effects on your character, that determines whether it is irrational for you to refuse to take it. There is no need to go into similar detail about the third case. It is clear that getting pleasure from “the spectacle of people starving without any moral hang-ups” is not compatible with wanting to have the moral virtues, at least not in normal human beings.38 Furthermore, in the real world, no one knows what the future will bring. People have limited information and are fallible. Fallibility is, in fact, one of the essential characteristics of human nature which accounts for common morality having the features that it has. Even if you cannot now do anything to help starving people, having the appropriate dispositions makes it more likely that if such occasions arise when you can help, you will be disposed to help. It is so clear that having the moral virtues is supported by basic reasons, that there is no need to provide any further arguments for this. Part of the force of Yarnell’s examples stems from conflating reasons with motives. He states: “The reason that you should mourn the death of friend is that the friend has died. The reason why you should feel guilt or be ashamed is that you have done something immoral or shameful.”39 It is true that your motive for mourning the death of a friend should be the belief that the friend has died and that your motive for feeling guilt should be the belief that you have done something immoral. But the question remains, why is it rational to have these kinds of motives in these kinds of situations? My answer is very similar to that provided by Yarnell. To have these kinds of motives and feelings in these kinds of situations “sometimes constitute who we are and what we value.”40 What makes it rational to be motivated by these kinds of beliefs is that they are essential constituents of our having the kind of character traits that are supported by basic reasons. To have the kind of character traits that are supported by basic reasons requires being motivated by beliefs about the past in the way that Yarnell has described. The case of sadness due to listening to poignant music is the kind of straightforward aesthetic example mentioned earlier. Berlioz “is willing to endure the sadness inspired by poignant music and has no conscious beliefs that continuing to listen will result in greater pleasure.”41 A possible answer is that being the kind of person who is affected in the appropriate way by poignant music, may be supported by basic reasons. Berlioz could believe that he is a better, more sensitive person for being affected in this way by poignant music. However, this kind of answer does not work quite as well in the purely aesthetic case as it does in the two cases where a person’s moral character is involved. I do not yet have a completely satisfactory answer to the aesthetic case, not only the Berlioz case, but also the examples of tragedies, horror films, and works of art that cause people to feel outrage. It may be helpful to consider some analysis of pleasure. It is clear that analyzing pleasure as the satisfaction of desire is not correct. It is far too com-
474
BERNARD GERT
mon for people not to feel pleasure when they have satisfied some desire for that analysis to have any plausibility at all. It is, however, plausible, to think of pleasure as being essentially related to the unmotivated desire to have some activity, feeling, or sensation, continue. To take pleasure in something is to desire for it to continue independent of any beliefs about benefits to be gained or harms to be avoided from its continuing. In normal cases, to have an unmotivated desire for an activity such as having your back scratched to continue, is to get pleasure from it, and to no longer intrinsically want it to continue is to no longer get pleasure from it. This preliminary account of pleasure, derived from Hobbes, is consistent with almost everything that Aristotle says about it, such as, how there can be greater and lesser pleasures, and how a greater pleasure can lead a person away from a lesser one. However, an account of pleasure simply in terms of an unmotivated desire for that which provides the pleasure to continue is not adequate. I have already pointed out that depressed persons sometimes have an unmotivated desire to remain depressed. Although we sometimes talk in an extended sense of a person enjoying his suffering, which supports this account, it is not plausible to regard any unmotivated desire to continue, including compulsive desires, as the sole defining criterian of pleasure.42 The defining criterian of pleasure must include not merely the intentional behavior that is the criterian of desires, but also facial and other expressions such as smiling and laughing. Indeed, that smiling and laughing are part of the criterian of pleasure is what causes the problem in dealing with some of the aesthetic cases. Although the intentional behavior that is the criterian of the desire to continue, is present, the facial and other expressions of pleasure that are part of the criterian of pleasure are not. What is even more troubling is that sometimes the facial and behavioral expressions that are part of the criterian of anxiety, displeasure, or sadness are present. However, it may not be necessary for the appropriate expressions of pleasure to be present at the time of desiring the activity to continue. All that may be required is that such expressions and related bodily behavior be evinced at some later time. It is clear that someone who wants to continue riding a roller coaster, but who never displays any of the appropriate facial expressions of pleasure during the ride, and even displays the appropriate facial expressions of anxiety, will still be said to have enjoyed it when she displays the criterian of pleasure when talking about it later. This does not provide a complete explanation of the rationality of experiencing unpleasant feelings in aesthetic situations, but together with some view about the aesthetic experience making a person more sensitive, it may explain most cases. As I indicated in Morality, a complete answer to this question requires a detailed account of human nature. I do not yet have that account. Patrick Yarnell has forced me to take more seriously than I did, the problems involved in explaining the rationality of wanting to feel sad about the
REPLIES TO THREE CRITICS
475
death of a friend. Recognizing that he uses “reasons” to refer to motives, enables me to endorse the first part of his final claim that “reasons [motives] flow from our deepest commitments.”43 Recognizing that being motivated to suffer at the death of a friend or feeling guilty for having done something immoral are essential constituents of having the kind of character traits we want, allows me to agree with the second part of his final claim that “we will sometimes have non-instrumental reason [motives] to suffer.”44 Yarnell has made it clear that for my account of rationality to be adequate, basic reasons must be supplemented by derived reasons which are related to basic reasons. He has also made clear that an important class of these derived reasons involve having the moral virtues and other desirable character traits. Notes 1. James P. Sterba, “Gert and the Defense of Morality,” Journal of Value Inquiry, this issue. 2. Bernard Gert, The Moral Rules (New York: Harper and Row, 1970) p. 88. 3. Bernard Gert, Morality: Its Nature and Justification (New York: Oxford University Press, 1998), pp. 167–171. 4. Sterba, op. cit. 5. Ibid. 6. See Gert, op. cit. pp. 330–337. 7. See David Philips, “Gert, Sidgwick, and Hybrid Theories of Rationality,” Journal of Value Inquiry, this issue. 8. Sterba, op. cit. 9. Ibid. 10. Phillips, op. cit. 11. Ibid. 12. See Gert. op. cit., ch 6. 13. Phillips, op. cit. 14. Gert, op. cit., p. 56. 15. Gert, op. cit., p. 78. 16. Phillips, op. cit. 17. Ibid. 18. Ibid. 19. Gert, op. cit., pp. 137–146. 20. Phillips, op. cit. 21. Ibid. 22. Ibid. 23. Ibid. 24. Ibid. 25. See Joshua Gert, “Skepticism about Practical Reasons Internalism,” Southern Journal of Philosophy 39 (2001). 26. See Joshua Gert, “Practical Rationality, Morality, and Purely Justifying Reasons,” American Philosophical Quarterly 37 (2000). 27. Bernard Gert, op. cit., p. 62. 28. Bernard Gert, op. cit., pp. 74–75.
476 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.
43. 44.
BERNARD GERT
Bernard Gert, op. cit., p. 47. Ibid. Ibid. Patrick Yarnell, “The Intrinsic Goodness of Pain, Anguish, and the Loss of Pleasure,” Journal of Value Inquiry, this issue. Ibid. Ibid. Ibid. Bernard Gert, op. cit., pp. 347–348. Yarnell, op. cit. Ibid. Ibid. Ibid. Ibid. See Bernard Gert, “Criterian and Human Nature,” in Rudolf Haller and Johannes Brandl, eds., Wittgenstein: Eine Neubewertung; Toward a Re-Evaluation, Vol. II (Wien: VerlagHölder-Pichler-Tempsky, 1990), pp. 106–114. Yarnell, op. cit. Ibid.