Consequentialism, Agent-Neutrality, and Mahāyāna Ethics

Several Indian Mahāyāna texts express an ethical perspective that has many features in common with Western forms of universalist consequentialism. Śāntideva, in particular, endorses a strong version of agent-neutrality, claims that compassionate agents should violate Buddhist moral commitments when doing so would produce good results, praises radical altruism, uses a critique of the self to support his ethical views, and even offers a reasonably clear general formulation of what we call act-consequentialism. Meanwhile, Asaṅga’s discussions of the motivation behind rules of moral discipline and the permissible reasons for breaking those rules suggests an interesting and complex version of rule-consequentialism. Evidence for features of consequentialism can be found in several Mahāyāna sūtras as well. In reading these sources, inter1 Philosophy Department, Binghamton University. E-mail: [email protected]. I am grateful to Dan Arnold, Christopher Knapp, Steven Scalet, Lisa Tessman, Melissa Zinkin, and multiple anonymous reviewers for their helpful advice and comments. This paper was first published in Philosophy East and West 58, no. 1 (2008), 17-35, and is reprinted by permission. 613 Journal of Buddhist Ethics pretations that draw on virtue ethics may not be as helpful as those that understand the texts as committed to various versions of consequentialism. During the past three decades, Western scholars have begun to study Buddhist ethics in a serious way. This development may soon make possible a fruitful dialogue between Buddhist and Western traditions of ethical reflection, in which each tradition might be enriched by the ideas of the other. However, such dialogue will be very difficult unless we Westerners can find some way of understanding, in our terms, what kind of ethical theory Buddhism might involve. Damien Keown’s influential work on this topic has convinced many scholars that Buddhist ethical views are not very similar to utilitarianism. Keown holds that these views should be understood through analogies either with Aristotelian virtue ethics, or perhaps, in the case of the Mahāyāna, with a theory known as Situation Ethics. However, if we draw on current debates in Western ethics to clarify the range of theories available for comparison, Keown’s position becomes less plausible. Although a case could be made that no form of Buddhist ethics is as similar to Aristotelianism as Keown claims, I focus here on Mahāyāna Buddhism. There are several Indian Mahāyāna texts that express an ethical perspective that has many features in common with the various versions of universalist consequentialism. Of these versions, the one that seems most promising as a basis of comparison with Mahāyāna views has sometimes been called “ideal utilitarianism” or “perfectionist consequentialism.” Consequentialism is a broad and diverse family of ethical theories. Each of these theories gives consequences some kind of decisive role in deciding how we should behave. Thus, an act-consequentialist ethical theory says that actions should be evaluated with reference only to their consequences. Meanwhile, a rule-consequentialist theory says Goodman, Consequentialism, Agent-Neutrality, and Mahāyāna Ethics 614 that we should follow a set of rules whose acceptance would, under some specified circumstances, bring about the best consequences. Various rule-consequentialist theories involve endorsing rules that would produce the best consequences if everybody followed them; or merely if everybody tried to follow them; or perhaps if enough people tried to follow them. And there are additional forms of consequentialism that trace the rightness of actions to consequences in some other indirect way, such as through the motives of the agent. The consequentialist theories that I want to put forward as models of Mahāyāna ethics are universalist: they take into account consequences for all sentient beings over the entire future history of the universe. I will be ignoring those consequentialist theories that are not welfarist, in that they assign intrinsic value to states of affairs that have nothing to do with the goodness of the lives of sentient beings. I shall take it as obvious that the beauty of rock formations and the balance of ecosystems, for example, are not the sorts of matters with which Indian Buddhist ethical thinkers were primarily concerned. The most important consequentialist theory for the history of Western ethics has been classical utilitarianism, which is the conjunction of universalist act-consequentialism with hedonism. Hedonism is the thesis that a being’s welfare depends only on how much happiness or suffering it experiences. Thus, in any situation of moral choice, classical utilitarianism tells us to choose that action which will produce the greatest total excess of happiness over suffering. Hedonism has had some defenders, but many people have found it hard to swallow. Is it really true that being happy is the only thing that 2 Thus some ethicists would classify all the theories I will consider as versions of utilitarianism, using that latter term in a broad sense to mean welfarist, universalist consequentialism. 615 Journal of Buddhist Ethics can make your life go better for you? Many philosophers have thought that it is good to know the truth, to be in love, and to appreciate objects of beauty, for example, and that the goodness of these states does not wholly depend on the happiness they produce. A consequentialist can accommodate these concerns by switching from hedonism to some other theory of welfare or interest. One option would be the preferencesatisfaction theory, on which your life goes well if you get what you want. The other important option is the Objective List theory. On this view, there is a list of features of your life that are intrinsically good or intrinsically bad. The more of the good features you have, the better your life goes; the more bad features it has, the worse it goes. All the things we regard as having genuine, non-derivative significance for wellbeing can go on the list. From now on I will be examining the hypothesis that Buddhism involves some kind of universalist consequentialism. The negative Buddhist view of desire strongly militates against attributing a preferencesatisfaction view to them. I will propose, instead, that the most plausible consequentialist interpretation of Buddhist ethics would employ an Objective List theory of welfare. The main alternative to this consequentialist interpretation draws on the tradition of virtue ethics. However, as many writers in the field agree, the concept of “virtue ethics” is slippery; it is not easy to make clear distinctions between virtue ethics and other forms of ethical thought. One such distinction is especially difficult to draw. Consider a universalist consequentialist view based on an Objective List theory, where the list of intrinsic goods includes certain character traits. Thus, one of the things which a follower of this theory will try to do is to create good states of character. This is the view that Thomas Hurka has called “perfectionist consequentialism.” It also has many similarities to the view P. J. Ivanhoe refers to as “character consequentialism,” but difGoodman, Consequentialism, Agent-Neutrality, and Mahāyāna Ethics 616 fers from that position in a few respects. Now, of course, a follower of virtue ethics takes as her central ethical goal the cultivation of good states of character. So what is the real difference between virtue ethics and perfectionist consequentialism? Indeed, if Buddhist ethics has strong similarities to each of these two positions, how will we ever tell which position it more closely resembles? Analytic ethicists have been thinking about these positions for some time, and they have discovered certain differences between these ethical perspectives. Of course, we need to keep in mind that there are several different forms of virtue ethics, and that different advocates of that perspective articulate their positions in quite different ways. But it is still possible to identify fundamental differences between virtue ethics and any universalist form of consequentialism. Act-consequentialist views will have practical consequences that virtue ethicists are committed to rejecting. Indirect consequentialists, such as rule-utilitarians, may have fewer practical disagreements with virtue ethicists, but their positions are still very different at the level of theoretical foundations. The basis of Aristotelian virtue ethics is a view called eudaimonism. According to eudaimonism, an action or trait is morally praiseworthy if it constitutes or contributes to the agent’s flourishing (eudaimonia.) This assertion must be understood in the context of Aristotle’s social conception of the self. According to Aristotle in the Nicomachean Ethics, my flourishing does not depend solely on what happens to me; there will be a small group of people, including my relatives and close friends, whom I care about and whose welfare directly contributes to my own (1101a20). My actions should be evaluated in terms of their contribution to the flourishing of myself and of those I care about. 3 See Ivanhoe. If the definition in note 1 is used, we could even call the theory “perfectionist utilitarianism.” 617 Journal of Buddhist Ethics On Aristotle’s view, certain virtues, such as courage, temperance and wisdom, help to constitute my flourishing. These virtues include liberality, or generosity; and Aristotle specifies that a generous person will experience pleasure in giving to others (1120a30). Thus, Aristotle’s virtuous agent must be willing to think about the welfare of others, and not just himself and those he cares about. But at the level of the theory’s foundations, the reason why a certain action of giving is morally praiseworthy is not the benefit to the recipient. Rather, the foundational justification for the value of giving is that doing so is the activity of the soul in accordance with the virtue of generosity, and thus helps to constitute the flourishing of the generous agent. Universalist consequentialists, by contrast, are not eudaimonists. According to consequentialists, one of my actions can be right even though it is harmful to my flourishing, so long as its consequences are sufficiently beneficial to others, including others who are in no important way related to me. Moreover, to a consequentialist, the value of generous acts derives from the benefits they confer on all those involved: the contributions to the welfare of both the giver and the recipient matter, and matter equally, in deciding the value of the action. We have, therefore, identified two important differences between Aristotle’s virtue ethics and all forms of consequentialism. Aristotle holds eudaimonism, whereas consequentialists would reject it; and, for Aristotle, the foundational justification for virtuous acts is their contribution to the flourishing of the agent, whereas for universalist consequentialists, their justification depends on their consequences for all sentient beings. Another important difference between the two views is closely related to these first two. A practitioner of virtue ethics takes her own virtue as her central ethical goal: she is to develop the skills, habits and 4 I am appealing to the distinction between factors and foundations that is discussed in Kagan. Goodman, Consequentialism, Agent-Neutrality, and Mahāyāna Ethics 618 attitudes of mind necessary to be the best agent she can be. But a practitioner of perfectionist consequentialism will have a very different goal: to bring about as much virtue among all sentient beings as possible. That means that a perfectionist consequentialist would be willing to make himself worse in order to make others better, so long as the total amount of virtue in the universe increased. This is a third difference between the two views: that of whose virtue is to be promoted. The differences I have just discussed can be summarized in the following way. Aristotelian virtue ethics is an agent-relative theory, meaning that it gives different aims to different agents. The view gives each agent the aim of that agent’s own flourishing, where the flourishing of each agent involves the flourishing of the small group of people that the agent cares about. But all versions of universalist consequentialism are agent-neutral: they give one common aim to all agents. This common aim is that the lives of all sentient beings go as well as possible. Agentneutrality is a very powerful assumption, which can have quite striking consequences. Because consequentialist theories are agent-neutral, they do not allow agents to assign any moral significance to the distinction between their own welfare and the welfare of others. Everyone’s happiness, and everyone’s virtue, must be taken into account equally. When this agentneutrality is expressed in a theory that directly evaluates actions, the resulting theory is extremely demanding: it can often call for acts of heroic self-sacrifice to benefit others. Moreover, because of its agentneutrality, act-consequentialism could often require an agent to neglect those people she cares about the most in order to benefit people she may never even have met. For example, Peter Singer argues that a genuinely moral agent living in a rich country under modern conditions should 5 The distinction between agent-relative and agent-neutral theories is found at Parfit 55. 619 Journal of Buddhist Ethics give almost all of her resources to alleviate the suffering of famine victims, until her material situation, and that of her family, is almost as bad as that of the victims themselves. Virtue ethicists, by contrast, reject such extreme demands. Defenders of virtue ethics try to delimit a sort of personal moral space in which each individual can legitimately promote his own welfare and the welfare of those he cares about, without reference to what might be good for the world as a whole. Though virtue ethicists agree that benevolence is sometimes morally required, they would place commonsense limits on this requirement, and argue that other moral considerations, such as family obligations, can often override its demands. Some consequentialists, especially those who are not act-consequentialists, have also defended ethical theories that are not as demanding as Singer’s view. Thus if we find a thinker presenting an ethical position that is extremely demanding, that is evidence that we are dealing with a form of consequentialism; but if a view is not extremely demanding, it could fall into either of these two categories of ethical theories, or perhaps some other. Consequentialism may call on us to do more than sacrifice ourselves: it may ask us to sacrifice others. A commonly noted feature of the act-consequentialist view is that it can require us to do things that would seem, prima facie, to be wrong, when the consequences of not doing them would be sufficiently terrible. An ideal act-consequentialist agent would have to lie, break promises, or even kill innocent people, when doing so would be of benefit to many sentient beings. By contrast, virtue ethics 6 See, for example, Christine Swanton’s statement that “we (at times) withhold the label ‘benevolent’ in our description of an act which . . . promotes the good of strangers but egregiously fails to express bonds of love to near and dear” (Swanton 4). Another important example of this kind of attitude is found in Wolf. 7 See, for example, Railton. Goodman, Consequentialism, Agent-Neutrality, and Mahāyāna Ethics 620 does not necessarily have this feature. Virtue ethicists may be in a position to side with our natural revulsion against performing certain terrible actions, even when these actions are necessary to avert great evils. Another difference between virtue ethics and consequentialism has come to light only since the work of Derek Parfit, especially in his book Reasons and Persons. Because of its commitment to agent-neutrality, consequentialism regards the divisions between the lives of different individuals as no more significant than differences between different periods of a particular individual’s life. Actions which benefit the agent, or the agent’s family and friends, at the expense of sentient beings in general, are just as irrational as actions which benefit me in the short run but do much greater harm to my long-term interests. These ethical claims could draw support from metaphysical theses that undermine the significance, or even the existence, of the unity of an individual human life, and thereby of fundamental distinctions between persons. Thus, Parfit uses destructive criticisms of the notion of personal identity to undermine egoism and support his own consequentialist views about ethics (307-347). Virtue ethics, by contrast, could receive no support from this kind of metaphysics; it is much more at home with the view of substance found in Aristotle’s metaphysical writings. A significant difference between perfectionist consequentialism and virtue ethics, then, involves the metaphysical bases that could be used to support each of the theories. Thus, even though perfectionist consequentialism is, in some respects, quite similar to virtue ethics, there are various ways to distinguish these views. Once we turn to the interpretation of Buddhist ethics, these differences can be used as tests. If we want to determine which kind of theory can most appropriately be attributed to Buddhists, it makes sense to look for passages that respond to the issues I have just raised. 621 Journal of Buddhist Ethics Of all the productions of the Indian Buddhist tradition, the texts that come closest to a worked-out ethical theory are the two works of Śāntideva: the Bodhicaryāvatāra, or Introduction to the Bodhisattva’s Way of Life, and the Śikṣāsamuccaya, or Training Anthology. In many cases, Śāntideva draws on earlier scriptural sources; but in synthesizing these sources, he creates a system of substantially greater theoretical coherence. The sophistication, generality, and power of Śāntideva’s arguments give him a legitimate claim to be the greatest of all Buddhist ethicists. The heart of Śāntideva’s ethical perspective is expressed in this passage from the Training Anthology: Through actions of body, speech, and mind, the Bodhisattva sincerely makes a continuous effort to stop all present and future suffering and depression, and to produce present and future happiness and gladness, for all beings. But if he does not seek the collection of the conditions for this, and does not strive for what will prevent the obstacles to this, or he does not cause small suffering and depression to arise as a way of preventing great suffering and depression, or does not abandon a small benefit in order to achieve a greater benefit, if he neglects to do these things even for a moment, he is at fault. None of the distinctive characteristics of classical actutilitarianism are missing from this passage. The focus on actions; the 8 In Sanskrit: “bodhisattvaḥ sarva-sattvānāṃ vartamāna-anāgata-sarva-duḥkha-daurmanasyopaśamāya vartamāna-anāgata-sukha-saumanasy-otpādāya ca niḥśāhyataḥ kāya-vāṅ-manaḥparākramaiḥ prayatnaṃ karoti / yadi tu tat-pratyaya-samāgrīṃ na-anveṣate, tad-antarāyapratikārāya na ghaṭate, alpa-duḥkha-daurmanasyaṃ bahu-duḥkha-daurmanasya-pratikārabhūtaṃ n-otpādayati, mahā-artha-siddhy-arthaṃ ca-alpa-artha-hāniṃ na karoti, kṣaṇam-apyupekṣate, sāpattiko bhavati” (Vaidya 12); translation by the author. Bendall and Rouse’s translation, at 16, obscures some crucial philosophical points. Goodman, Consequentialism, Agent-Neutrality, and Mahāyāna Ethics 622 central moral importance of happy and unhappy states of mind; the extension of scope to all beings; the extreme demands; the absence of any room for personal moral space; the balancing of costs and benefits; the pursuit of maximization: every one of these crucial features of utilitarianism is present. Notice also that the passage does not say anything about who receives the benefits or burdens that we are to balance against each other. If we go by this passage alone, we will conclude that Śāntideva is prepared to allow the balancing of the interests of some against the interests of others. But this passage is not all we have to go on; nor does it exhaust the available evidence for an act-consequentialist reading of Śāntideva’s ethical philosophy. Śāntideva offers us not only a statement of consequentialism, but a powerful rhetorical exploration of its demanding nobility. Chapter three of the Introduction is full of poetic expressions of radical altruism and total, self-sacrificing compassion: III. 8: May I avert the pain of hunger and thirst with showers of food and drink. May I become both drink and food in the intermediate aeons of famine. 9: May I be an inexhaustible treasure for impoverished beings. May I wait upon them with various forms of offering. 10. See, I give up without regret my bodies, my pleasures, and my good acquired in all three times, to accomplish good for every being. Clearly, Śāntideva is much more similar to certain actconsequentialist writers such as Peter Singer, who insist on the supreme moral significance of altruistic self-sacrifice, than he is to the advocates of virtue ethics and of other versions of consequentialism who want to allow the individual some moral space to act in ways not dictated by 623 Journal of Buddhist Ethics universalist moral considerations. The ethics of the Introduction, like some forms of consequentialism, is extremely demanding. Another Mahāyāna text, the Sūtra on Upāsaka Precepts, expresses similar views. The practice of generosity advocated in this sūtra extends to such an extreme that “even if a wise person was in the situation whereby he would live if he ate the last handful of food but would die if he gave it away, he should still give it away” (114). Thus, as in actconsequentialism, the bodhisattva is not allowed to make distinctions between his own welfare and the welfare of others, except when these distinctions oppose the natural human tendency and favor others over self. The text seems sometimes to bend over backwards in opposition to the selfish tendencies of human nature: “When friends and foes are suffering, he first benefits his foes” (59). By counteracting natural tendencies to partiality, the bodhisattva can move closer to the ideal of impartial great compassion for all beings. This ideal of impartiality finds expression in many passages in the text: “he benefits both foes and friends without discrimination” (91); his goal is “to be compassionate to all regardless of their relationship to oneself” (85); he “sees all foes as dear friends” (105). This emphasis on impartiality is strong evidence that the ethical view of this sūtra, like consequentialism, is an agent-neutral theory. Of course, by practicing altruism and non-violence, the bodhisattva can achieve various good things for himself: he cultivates the roots of good, develops knowledge of religious truth, and so on. One might wonder whether the ultimate justification for the bodhisattva’s practice is these benefits to himself. In fact, the Precepts Sūtra contains what looks like an explicit endorsement of eudaimonism: “to benefit others is to benefit oneself” (49). Moreover, even some of the bodhisattva’s most impressively altruistic actions, such as being reborn as an animal or in one of the hells for the benefit of others, may not be as costly Goodman, Consequentialism, Agent-Neutrality, and Mahāyāna Ethics 624 as they look: “If this person dwells in the three evil realms, he will not have to suffer as other beings do” (21-22). Once the bodhisattva attains an advanced stage of enlightenment, but before Buddhahood has been achieved, no degree of damage to the physical body will cause the bodhisattva to suffer at all; so being born in unfavorable circumstances is not really a great sacrifice. As regards our first difference between virtue ethics and consequentialism, the question of whether doing the right thing always benefits the agent, there is some reason to suppose that Mahāyāna ethics would agree with virtue ethics. This impression can be shown to be misleading by studying the ritual of the “dedication of merit” (puṇya-pariṇāmanā). This ritual is the subject of chapter ten of the Introduction, and the Precepts Sūtra gives a brief summary of its purpose: “He always transfers his merits and virtues to others” (98). Through his religious activities, the bodhisattva constantly accumulates merit (puṇya). This merit, if he retained it, would cause him to be both happier and more virtuous. But rather than promote his own well-being, or even his own virtue, the bodhisattva gives away this merit in order to make other beings happy and virtuous. In certain early Mahāyāna texts, “dedication of merit” seems to mean, not giving away one’s merit to others, but simply changing the nature of the fruition to be expected from it: instead of ripening as worldly pleasures, it instead contributes to the achievement of Buddhahood. For example, Jan Nattier has discussed this issue in her analysis of the Inquiry of Ugra (Ugra-paripṛcchā), at 114-115. But there are many passages in the Introduction that cannot be read in terms of this early understanding of the dedication of merit, such as these verses from chapter ten: X.2. Through my merit may all those in any of the directions suffering distress in body or mind find oceans of happiness and delight. 625 Journal of Buddhist Ethics X.31. By this merit of mine may all beings without exception desist from every evil deed and always act skillfully. Here Śāntideva certainly seems to be wishing that his own merit will be transferred to others, benefiting them. Verse III.6 also seems to be most naturally interpreted as involving a transfer: III.6. With the good acquired by doing all this as described, may I allay all the suffering of every living being. Chapter ten reaches its climax with a verse like a sledgehammer: X.56. Whatever suffering is in store for the world, may it all ripen in me. May the world find happiness through the pure deeds of the Bodhisattvas. I do not think the prospects of a eudaimonist interpretation of this verse are very good. Nor can it be read in terms of the dedication of merit as a change in the nature of the fruition from that merit. If this interpretation of the purpose of the dedication of merit is correct, then Mahāyāna Buddhism both allows actions to be moral that don’t promote the well-being of the agent, and places the virtue of all beings above the virtue of the individual agent. According to the Introduction and the Precepts Sūtra, as in perfectionist consequentialism, the goal of each agent should be to promote virtue in general, not just the virtue of that agent. One of the easiest tests to apply from the list I have given above is the last: whether reductionist views of personal identity can be used to 9 In Sanskrit: “evaṃ sarvam-idaṃ kṛtvā yan-mayā-sāditaṃ śubham tena syāṃ sarvasattvānāṃ sarva-duḥkha-praśāntiṛt.” Note that it seems possible to translate with a passive: “may all the suffering of all sentient beings be allayed.” There is nothing in the verse that forces us to supply the subject: “may I allay.” Goodman, Consequentialism, Agent-Neutrality, and Mahāyāna Ethics 626 support the moral theory in question. It is now quite well known that in the Introduction, Śāntideva offers a justification of his ethical views that appeals to the Buddhist doctrine of non-self. According to this doctrine, what we take to be the fundamental, and fundamentally significant, distinction between ourselves and the rest of the universe is, in fact, an illusion. In reality, there are no such things as souls, selves, or even human bodies. Reality is a vast and complex process, consisting of innumerable tiny, momentary entities called “dharmas,” which, as I have argued in previous work, a contemporary analytic philosopher would classify as tropes. The process of singling out some of these tropes as constituting “me” and “mine” is profoundly deluded. It leads to attachment, egoism, pride, to greed and hatred, and ultimately to suffering. Once we recognize the nonexistence of the self, however, egoism, along with all forms of practical reasoning that depend on a distinction between self and other, are exposed as irrational: VIII.97: If I give them no protection because their suffering does not afflict me, why do I protect my body against future suffering when it does not afflict me? 99: If you think it is for the person who has the pain to guard against it, a pain in the foot is not of the hand, so why is the one protected by the other? 102: Without exception, no sufferings belong to anyone. They must be warded off simply because they are suffering. Why is any limitation put on this? 10 His arguments are discussed at length in Williams. These arguments are also a central topic of the exchange between Williams and Siderits in Philosophy East and West 50, no. 3 (2000), 412-459. 627 Journal of Buddhist Ethics These arguments can be seen to be strikingly similar to those offered over a thousand years later by Derek Parfit in Reasons and Persons. In both cases, the author starts with a view that is a version of the nonself doctrine, and attempts to use that view to defend an ethics of selfsacrifice for the good of all sentient beings. Thus, on the last of the criteria mentioned above, Śāntideva’s ethical views closely resemble consequentialism. One difference between act-consequentialism and virtue ethics which remains to be examined is that act-consequentialism could occasionally require morally terrifying actions. As it turns out, there are a number of Mahāyāna texts that argue that, sometimes, a bodhisattva ought to perform actions that would otherwise be considered wrong in order to benefit large numbers of sentient beings. Perhaps the most important text of this kind is Asaṅga’s “Chapter on Ethics,” a part of his larger work The Bodhisattva Stages (Bodhisattva-bhūmi). Several writers, including both Harvey and Keown, have discussed the fact that this text argues for the permissibility of lying, stealing, sexual misconduct, and killing, when these actions are motivated by a compassionate wish to benefit all beings. In this respect, Asaṅga’s views are quite different from those found in most non-Mahāyāna Buddhist texts. Careful examination of Asaṅga’s position reveals a very interesting view about when the precepts may be broken. Here is Asaṅga’s general account of when a bodhisattva can break the precepts: If the bodhisattva sees that some caustic means, some use of severity would be of benefit to sentient beings, and does not employ it in order to guard against unhappiness, he is possessed of fault, possessed of contradiction; there is fault that is not defiled. If little benefit would result for 11 Which is itself part of a larger work, the Stages of Religious Practice (Yogācāra-bhūmi). Goodman, Consequentialism, Agent-Neutrality, and Mahāyāna Ethics 628 the present, and great unhappiness on that basis, there is no fault. (74) Now there is an ambiguity in this passage. Does Asaṅga mean that the precepts can be broken if doing so would be of benefit to sentient beings collectively, in a sense that allows aggregation? Or does he mean that the act of breaking the precepts must benefit sentient beings distributively, so that every being affected must benefit, or at least not be harmed? If we examine the examples that Asaṅga offers of permissible precept-breaking, they turn out all to fit the second, distributive pattern. This fact may be surprising, since Asaṅga includes killing as one example of permissible precept-breaking. Here is the only example he supplies of permissible killing: Accordingly, the bodhisattva may behold a robber or thief engaged in committing a great many deeds of immediate retribution, being about to murder many hundreds of magnificent living beings—auditors, independent buddhas, and bodhisattvas—for the sake of a few material goods. Seeing it, he forms this thought in his mind: “If I take the life of this sentient being, I myself may be reborn as one of the creatures of hell. Better that I be reborn a creature of hell than that this living being, having committed a deed of immediate retribution, should go straight to hell.” With such an attitude the bodhisattva ascertains that the thought is virtuous or indeterminate and then, feeling constrained, with only a thought of mercy for the 12 This ambiguity is not an artifact of the English translation; it also exists in the Sanskrit for the passage, which reads: “bodhisattvo yena kaṭuka-prayogeṇa tīkṣṇa-prayogeṇa sattvānām-arthaṃ paśyati taṃ prayogaṃ daurmanasya-ārakṣayā na samudācarati / sāpattiko bhavati akliṣṭām-āpattimāpadyate / anāpattir-yat parīttam-arthaṃ dṛṣṭa-dhārmikaṃ paśyet prabhūtaś-ca tan-nidānaṃ daurmanasyam” (Dutt 116-117). 629 Journal of Buddhist Ethics consequence, he takes the life of that living being. There is no fault, but a spread of much merit. (70-71) In this case, the being who is killed is actually better off dead, since he was about to condemn himself to an immensely long (though finite) period of horrible suffering. By killing him, the bodhisattva rescues the robber from a fate much worse than mere death. The same pattern is also seen in Asaṅga’s examples of permissible violations of the second precept, which forbids “taking what is not given.” The bodhisattva is allowed to overthrow “violent and pitiless” kings, taking their power from them without consent, but preventing them from incurring further negative karma through their oppressive rule. If robbers steal property from religious communities and shrines, the bodhisattva may steal it back, thus protecting them from the very grave karmic consequences of consuming such stolen items. A bodhisattva may also remove corrupt or incompetent storekeepers and custodians from office, to prevent them from incurring serious karmic misfortune from their own waste and embezzlement (71). Asaṅga never explicitly denies that balancing is permissible. He simply does not address situations in which some beings must be harmed to benefit others. But he is enumerating exceptions to the generally valid rules of morality, rules that he takes very seriously indeed. It seems plausible to assume that he intends these to be the only kind of exceptions that are allowed; if he thought there were more classes of exceptions, he would probably have indicated them. He also makes it clear that the rules can be violated only for the benefit of others. It is never permissible for a bodhisattva to break the precepts for his own benefit, even in a small way. Goodman, Consequentialism, Agent-Neutrality, and Mahāyāna Ethics 630 Though Asaṅga does not give us anything like a worked-out view of balancing between individuals, there is one passage where he countenances something like balancing: When something of body or speech done to someone else would result in pain and unhappiness for a third party, whereas neither party would be moved from an unwholesome to a wholesome situation, the bodhisattva will reflect upon it and reject that act of body-speech on the grounds that it would not comply with the inclinations of the third party. If, on the other hand, he sees that either party, or both would be moved from an unwholesome to a wholesome situation, the bodhisattva will reflect upon it adopting nothing but a thought of mercy, and perform the action . . . . (56-57) In this case, the bodhisattva is permitted to inflict suffering on one party in order to increase the virtue of a second. The permissibility of this form of balancing is easy to explain if we regard Asaṅga as holding that there are two kinds of value, happiness and virtue, with the second being much more important than the first. I shall argue for this interpretation below. Thus we are led to interpret Asaṅga as a kind of rule-utilitarian. On the interpretation suggested by these passages, a Mahāyāna practitioner should normally follow the precepts. These are moral rules that are justified by appeal to the good consequences of following them; they are “lived for the benefit and pleasure of all sentient beings” (48). But 13 Note that Asaṅga’s view is ambiguous between several possible forms of ruleconsequentialism. Asaṅga says nothing about whether the rules are justified by the consequences of their being strictly followed by everyone; or the consequences of their being accepted by everyone, but not necessarily always obeyed; or perhaps by the con631 Journal of Buddhist Ethics these rules may permissibly be broken if, and only if, doing so significantly benefits at least one sentient being other than the agent, and leaves no one worse off than if the rules were followed. In the writings of Śāntideva, we can find at least one explicit example of balancing. Discussing in the Introduction whether and when to sacrifice one’s body for others, he writes: V.86. The body serves the True Dharma. One should not harm it for some inferior reason. For it is the only way that one can quickly fulfill the hopes of living beings. 87. Therefore one should not relinquish one’s life for someone whose disposition to compassion is not as pure. But for someone whose disposition is comparable, one should relinquish it. That way, there is no overall loss. This is an example of balancing between self and other, not between two others. Since we have seen that Mahāyāna texts sometimes treat the difference between the agent and others as morally relevant, the passage does not conclusively show that Śāntideva would countenance harming some innocent persons to benefit a larger number of others. But it seems that he is committed to allowing such an action by his strategy for justifying his ethical views. Śāntideva explicitly says that we are rationally required to accept a lesser amount of suffering in order to prevent a greater. Once we bring to bear on ethics the teaching that there are no metaphysically important differences between different sentient beings, it cannot ultimately matter whether harms are compensated by benefits to the same beings or to others; nor can it ultimately sequences of their being accepted by a large enough number of people; or, perhaps, by the long-term consequences of their being accepted by me in particular. 14 At 69: “All doctors use painful treatments to restore health. It follows that to put an end to many sufferings, a slight one must be endured.” Goodman, Consequentialism, Agent-Neutrality, and Mahāyāna Ethics 632 matter who it is that carries out a harmful action. This strategy inevitably leads to an ethical view that allows balancing. In fact, Śāntideva seems to adopt a more wholeheartedly, or at least more explicitly, consequentialist ethical position than Asaṅga does. Such a position would clearly be entailed by a literal reading of Introduction V.84: “Even what is proscribed is permitted for a compassionate person who sees it will be of benefit.” We should be careful to note that Mahāyāna ethicists, including Śāntideva, do not extend the permission to break moral rules to just anyone. In this passage, Śāntideva requires that the person carrying out the proscribed action must be “compassionate”; moreover, she must actually see—not just theorize—that the action will be beneficial. Ordinary people, who neither understand the way things really are nor have the proper kind of motivation, are seen in this system in much the way that everyone regards children and insane persons: they must strictly follow rules of conduct that have been designed for their benefit. Those with greater compassion and insight, though, may ignore these rules when doing so would be of benefit. A universal permission to disregard moral rules when doing so would be beneficial, such as we seem to find in Introduction V.84, directly implies act-consequentialism. In any given case, any rule we try to apply will either endorse the action that produces the best consequences, in which case following the rule coincides with act-consequentialism, or the rule endorses some other action, in which case the verse allows us to break the rule and do what act-consequentialism tells us to. Thus, Śāntideva’s view must coincide with act-consequentialism, at least as it applies to those who are truly compassionate. The flexibility which is endorsed at the level of theory in Introduction V.84 is reflected in the discussion of at least two sets of practical ethical issues in the Training Anthology. First, Śāntideva has a broader con633 Journal of Buddhist Ethics ception than Asaṅga expresses of the circumstances under which a bodhisattva can make use of the property of others. He writes: If you consume what you believe to be the property of others for your own benefit, you undergo the downfall of stealing. If the value of the goods exceeds the stated threshold, in the Vows of Individual Liberation, you are defeated. But if, as a servant of sentient beings, you simply protect your body with the property of your masters, sentient beings, there is no problem. For it is not always the case that a servant doing work for his master owns the property with which he works. And the Sūtra on Chanting the Dharma Together says, “A bodhisattva should be like a servant, doing whatever needs to be done for all sentient beings.” And if a servant who is totally focused on benefiting the master, but is afflicted by illness and so on, should eat even without informing the master, there is no problem. (79) This ingenious argument seems to be offered by Śāntideva as a way of showing that, if a bodhisattva is genuinely motivated by compassion, he or she may ignore property rights, the topic of the Second Precept of Buddhism, whenever they interfere with working for the benefit of all beings. Śāntideva is also prepared to allow the Third Precept, which forbids sexual misconduct, to be overridden by the more fundamental significance of the welfare of all beings: Also in the world, when a mother and father see their son being impaled on a stake, attachment to enjoyment is not observed in them, due to the power of innate compassion; at such a time, there would be no secret sexual misconduct with [any woman], married or unmarried, whether protected by her family, the Teachings, or the flag. [But,] Goodman, Consequentialism, Agent-Neutrality, and Mahāyāna Ethics 634 where there is benefit to beings [in such conduct,] and no harm to beings, having ascertained the consequences, there is no problem. (93) 15 The idea of this passage seems to be the following: a father whose son is being gruesomely executed would be so shocked and dismayed by the situation that he would not even think about engaging in sex, neither legitimately with his wife nor illegitimately with any other woman. Since bodhisattvas regard all beings as their own family members, their awareness of the terrifying sufferings of cyclic existence should drive all thought of attachment to lustful pleasures from their minds. But under exceptional circumstances, when for some reason engaging in socially forbidden sex would have good consequences and would not lead to any harm, compassion itself requires them to ignore the Third Precept. Asaṅga might agree; he is simply not fully explicit on the range of cases 15 Translation by the author. See the somewhat different translation at Bendall and Rouse 163: “And in the world when a son is impaled in view of mother and father, they do not think of attachment to their own welfare by reason of their natural pity. Secret relations with wives or maids who are protected by the family or religion or the royal standard, would not be forbidden love. If there is here good for people, or no detriment to them, there is no sin when one understands the motive.” Bendall and Rouse’s translation here is questionable; they may be breaking what is semantically one sentence into two. The Sanskrit of the passage reads like this: “loke’pi putre śūlam-āropyamāṇe paśyator-mātā-pitror-na saukhya-saṅgo dṛṣṭaḥ sva-anurūpa-kṛpā-vaśāt prachannas tarhi sasvāmikāsu niḥsvāmikāsu vā kula-dharma-dhvaja-rakṣitāsu kāma-mithyācāro na syāt / sati sattva-arthe sattva-anupaghāte ca-anubandhaṃ nirūpya-adośaḥ.” It seems that we should read the whole text before the daṇḍa as one sentence. If we do, the passage should be translated as I have rendered it above. This proposed translation has the same philosophical upshot as the translation of Bendall and Rouse, except that they read the two conditions—benefit for people, no detriment to them—as disjoined rather than conjoined. But we have ca, “and,” not vā, “or”; both must be satisfied for the action to be permissible. 16 Such sexual relations, on Śāntideva’s view, are still unequivocally forbidden for monks. Only laymen may engage in them. 635 Journal of Buddhist Ethics in which a bodhisattva can break this precept, giving us only a single example in which such a violation would be permissible. If we interpret Śāntideva as an act-consequentialist, we will have to account for passages such as III.14: “Let there never be harm to anyone on account of me.” An act-consequentialist may sometimes have to inflict harms on some in order to prevent greater harms to others. Perhaps Śāntideva is expressing a wish or hope that he never be placed in such a situation. Alternatively, he is aspiring to attain a degree of moral skill so great that he can find actions that benefit all those involved in any situation, without having to harm anyone. It appears we have some reason to regard Śāntideva as holding a somewhat different ethical position from Asaṅga, and one that is substantially closer to modern act-consequentialism. Asaṅga, meanwhile, seems to hold a complex and unusual version of rule-consequentialism, whose implications might well repay further study. If I am right to affirm that these Mahāyāna ethicists hold versions of universalist consequentialism, we are left with a crucial interpretive question: what theory of well-being do they accept? On my view, it makes sense to attribute to Mahāyāna thinkers a view on which happiness and the absence of suffering, as well as virtues and the absence of vices, are elements on an Objective List that defines well-being. In chapter eight, we find a number of verses that indicate that Śāntideva is explicitly interested in alleviating the suffering and promoting the happiness of all beings. Here are just a few: VIII.95. When happiness is liked by me and others equally, what is so special about me that I strive for happiness only for myself? Goodman, Consequentialism, Agent-Neutrality, and Mahāyāna Ethics 636 96. When fear and suffering are disliked by me and others equally, what is so special about me that I protect myself and not the other? 103. If one asks why suffering should be prevented, no one disputes that! If it must be prevented, then all of it must be. If not, then this goes for oneself as for everyone. It seems that, for Śāntideva, happiness must be good and suffering bad. Asaṅga, meanwhile, writes that To undertake and proceed to train oneself in the essence of ethics endowed with these four qualities, should be understood as “wholesome,” because of benefit for oneself, benefit for others, benefit for many people, pleasure for many people, mercy for the world, and welfare, benefit, and pleasure for divine and human beings. (48) Asaṅga too, it seems, regards pleasure as a genuinely valuable consequence of the bodhisattva’s efforts. Scriptural sources from the Mahāyāna tradition also indicate that pleasures, including worldly pleasures, do have some value greater than zero. We read in the Sūtra of Golden Light that if people expound the Sūtra, the benefits will include . . . that the whole of Jambudvīpa will become plentiful, happy, and full of many people and men, that the beings in the whole of Jambudvīpa will be blessed, will experience various pleasures, that beings during numerous hundreds of thousands of millions of aeons will experience inconceivable, most exalted blessings, will have meetings with the Lord Buddhas, in future time will be 637 Journal of Buddhist Ethics fully enlightened in supreme and perfect enlightenment. (40) This fairly typical text certainly regards happiness and pleasure as goods. A great deal of the ethical discourse of the Mahāyāna is specifically concerned with happiness and the absence of suffering. Thus an interpretation of Mahāyāna ethics as a form of classical utilitarianism would have something to be said for it, and would be much closer to the truth than any interpretation in terms of Aristotelian virtue ethics. Indeed, I cannot decisively refute this classical utilitarian interpretation. But I think there is some textual evidence to suggest that, for Mahāyāna Buddhists, pleasure is not the only intrinsic good. Consider, for a first example, the passage above from the Sūtra of Golden Light, which I introduced to suggest that the Mahāyāna tradition attaches some positive value to pleasure. Here, the experience of pleasure is not by any means the only good to be achieved. In the Mahāyāna tradition generally, temporary pleasure is seen as being good, but not very important, while religious and spiritual values are assigned great importance. The Sūtra of Golden Light emphasizes the value of virtues, specifically, by comparing them with jewels: “The Buddha’s virtues are like the ocean, a mine of numerous jewels” (41). This analogy strikes me as some evidence that this text is implicitly committed to the intrinsic value of virtuous qualities. Indeed, Mahāyāna writers frequently allude to a number of normatively charged characteristics that human beings can have. Happiness and suffering are just some of these characteristics. The most important good qualities a human can have, from the Mahāyāna point of view, seem to be the “roots of good”: non-greed (arāga,) non-hatred (adveṣa) and non-delusion (amoha.) Though negatively stated, these roots of good Goodman, Consequentialism, Agent-Neutrality, and Mahāyāna Ethics 638 include such obvious virtues as generosity, compassion, and insight. Buddhist texts from various schools often classify these qualities as good (kuśala). Absent some strong argument to the contrary, it would seem to make sense to include them, along with happiness and the absence of suffering, as part of a Buddhist theory of well-being. And doing so would result in the view I have been calling perfectionist consequentialism. Damien Keown has noticed that one might try to make an analogy between Buddhist ethics and Ivanhoe’s proposed theory of character consequentialism. But the analogy strikes him as unhelpful, largely because he doubts the coherence of character consequentialism as an ethical theory. Keown’s attempt to refute character consequentialism, it turns out, involves an anecdote that can help clarify further the nature of Buddhist ethics. He argues as follows: . . . certain choices may corrupt moral character yet have beneficial consequences for society. For example, an emperor may choose to deceive his people in order to promote social harmony. If he succeeds in his aim and takes the knowledge of his lies to the grave, will he have done right or wrong by the standards of character consequentialism? If he is considered to have acted rightly, then character is always subordinate to social good and has no intrinsic value. If he is considered to have acted wrongly, then consequences have no priority in moral judgments, and all reference to ‘consequentialism’ can simply be dropped. (“Karma” 347) This example is valuable largely because we know exactly how Mahāyāna Buddhists would respond to it. It closely parallels the story of King Anala from the Avataṃsaka Sūtra, as cited in Gampopa 348-350. 639 Journal of Buddhist Ethics In this story, the aspiring bodhisattva, Sudhana, is advised by one of his teachers to go visit King Anala and ask him for instruction. But when Sudhana arrives in the city, he discovers that the King is surrounded by frightening, wrathful demons who are constantly engaged in meting out severe punishments to those who violate the city’s laws. Sudhana reacts with horror to this grotesque violation of Buddhist injunctions to be nonviolent and lenient to criminals. But King Anala reveals to him that both the wrathful guardians and their criminal victims are actually illusions created by Anala’s magical powers. As a result of the punishments meted out to these wholly illusory malefactors, the real citizens are terrified into acting rightly. King Anala explains his intentions in this way: It is by this method that I encourage people not to involve themselves in any of the ten nonvirtues, but to demonstrate the path toward the ten virtues. I make this effort to end the suffering of the people in my country and establish them in the path to the omniscient state. (350) By Keown’s own test, King Anala, and the texts that repeat his story with approval, must be consequentialist. But there is a further interesting lesson in the story. One of the things he is trying to bring about in deceiving his people is virtue. He is prepared to tell lies, thereby, perhaps, impairing his own virtue, in order to bring about a much greater total amount of virtue among the populace—including the virtue of not telling lies. We have some reason, then, to interpret King Anala as holding, perhaps tacitly, a view with certain similarities to character consequentialism and perfectionist consequentialism. I have presented evidence that at least some Indian Mahāyāna texts hold an ethical view that is quite similar to Western consequentialism, and perfectionist consequentialism in particular. To what extent Goodman, Consequentialism, Agent-Neutrality, and Mahāyāna Ethics 640 can we generalize from these texts, and attribute a consequentialist view to the Mahāyāna tradition as a whole? It would not be safe to extend this interpretation to all Asian societies that have accepted the Mahāyāna. In China, and in the regions influenced by Chinese culture, Buddhist ethics has been transformed by the influence of Confucianism. But even if we restrict our attention to India, the Mahāyāna tradition is enormously complex and diverse. In particular, the esoteric Buddhism of the Tantras certainly can be interpreted as holding a quite different ethical position from other forms of Buddhism, though this interpretation is not necessarily forced on us. Our examination of the views of Śāntideva and Asaṅga has shown that there may be subtle differences between their positions. Asaṅga seems to hold a kind of rule-consequentialism that allows the rules to be broken when doing so would benefit others and would harm no one; he also seems to have a theory of well-being that accepts both happiness and virtue as good, and assigns somewhat greater significance to virtue. Śāntideva’s views are closely related, but seem to focus more on the direct evaluation of actions, placing less emphasis on rules. Śāntideva is thus somewhat more comfortable with allowing exceptions to generally valid principles. There is considerable, though perhaps not conclusive, evidence for regarding Śāntideva as an act-consequentialist. The analogy with virtue ethics, I have argued, does not represent a very valuable interpretive strategy when it comes to Indian Mahāyāna ethical thought. There seem to be at least some texts from this tradition that could more fruitfully be interpreted in terms of universalist, perfectionist consequentialism. That interpretation might help explain why Mahāyāna texts require such extreme manifestations of generosity; why they allow ethical rules to be broken, if doing so would benefit others; why they advise bodhisattvas to give away their merits and virtues; and why they place such strong and repeated emphasis on altruism and on 641 Journal of Buddhist Ethics benefiting all sentient beings. If this interpretation is accepted, we can hope that discussions of Buddhist ethics will begin to address the many difficulties with, and varieties of, consequentialism that have appeared in Western discussions. It is even possible to hope that the views of Western consequentialist philosophers will be enriched and strengthened by ideas from a version of Buddhism whose goal of promoting universal compassion and benevolence is, in important respects, similar to their own.