, Volume 57, Issue 4, pp 591–609 | Cite as

Moral Agency and the Paradox of Self-Interested Concern for the Future in Vasubandhu’s Abhidharmakośabhāṣya

  • Oren HannerEmail author


It is a common view in modern scholarship on Buddhist ethics that attachment to the self constitutes a hindrance to ethics, whereas rejecting this type of attachment is a necessary condition for acting morally. The present article argues that in Vasubandhu’s theory of agency, as formulated in the Abhidharmakośabhāṣya (Treasury of Metaphysics with Self-Commentary), a cognitive and psychological identification with a conventional, persisting self is a requisite for exercising moral agency. As such, this identification is essential for embracing the ethics of Buddhism and its way of life. The article delineates the method that Vasubandhu employs to account for the notion of a selfless moral agent, with particular emphasis on his strategies for dealing with one central aspect of agency, self-interested concern for the future.


Agency Self Personal identity Buddhism Vasubandhu Abhidharmakośabhāṣya 

The Buddhist critique of the self has provoked a long-lasting metaphysical debate between Buddhist and non-Buddhist philosophers concerning the nature of personal identity. Alongside its metaphysical implications, for many classical Buddhist thinkers, this principle primarily had a deep ethical and soteriological significance. In his Treasury of Metaphysics with Self-Commentary (Sanskrit Abhidharmakośabhāṣya, henceforth AKBh),1 for example, Vasubandhu (fourth to fifth centuries CE) maintains that it is only through realizing that there is no permanent self (anātman) that freedom from the cycle of births and deaths (saṃsāra) is possible.2 For the most part, modern scholarship on Buddhist ethics has indeed leaned towards this particular interpretation of the relation between the principle of no-self and ethics. The common conception has been that the cognitive reification of the self, and the accompanying emotional clinging, constitutes an obstacle to acting morally (Goodman 2009a, pp. 111, 213; Gross 1997, pp. 338–339; Harvey 2000, p. 36; Ives 1992, pp. 117–120; King 2005, pp. 91–92; Siderits 2017, pp. 289–292). In this essay, I will defend a different interpretation, according to which an identification with an enduring self is, in fact, a requirement for engaging in the ethics and life plan of Buddhism.3 In this stance, those who adhere to ethics are conventional moral agents, and agency, in turn, requires a sense of personhood.

My observations concerning the implications of the principle of no-self for ethics and agency will rely on the AKBh, and a large portion of my discussion will refer to the ninth chapter of the work, the Refutation of the Doctrine of Self (Ātmavādapratiṣedha, henceforth ĀVP). The ĀVP is arranged as a set of debates between Vasubandhu and his philosophical opponents. Specifically, in the final part of the chapter, Vasubandhu considers questions related to agency and addresses various objections raised by a non-Buddhist opponent, who defends the thesis that an enduring self exists.4

Vasubandhu’s Strategy for Dealing With Questions of Agency

Vasubandhu’s theory of agency in the ĀVP rests on the Buddhist model of the five aggregates (skandha). These are the different functions to which a person can be reduced. In his explanation, Vasubandhu argues that a person is not an independent and enduring entity (dravya) existing above and beyond the psycho-physical complex, as his philosophical opponents tend to believe. Rather, a person is a provisional designation (prajñapti) that refers to the conglomeration of mental and physical constituents itself, or, in other words, to the collection of the five aggregates. The aggregates, in turn, encompass factors (dharma) of a certain type. According to the Buddhist views expressed in the AKBh, beings, as well as the inanimate world, are structured by dharmas, which are basic physical and mental elements and are said to be the building blocks of things in their ultimate mode of existence.5

These, in brief, are the functions of each of the aggregates: the aggregate of materiality (rūpa-skandha) includes all factors of the material world, including factors that establish the physical bodies of living beings. The aggregate of feeling (vedanā-skandha) includes the affective tone of every experience a human being has, which is either unpleasant, pleasant, or neutral. The aggregate of cognition (saṃjñā-skandha) encompasses all events in which a person cognizes an object through its distinct property. The aggregate of conditional factors (saṃskāra-skandha) includes all mental factors that are not part of the other four aggregates, that is, all elements of human psychology except for feelings, cognitions, and consciousness (the fifth aggregate), for example, personal dispositions or psychological patterns, such as compassion, patience, anger, or jealousy. Finally, the aggregate of consciousness (vijñāna-skandha) is the collection of the various sense impressions that beings obtain through the contact between the sense organs and perceived objects.6 Since the five aggregates are, in fact, collections of momentary factors that stand in a relation of causality to each other in such a way that the factors of a given moment ensue from the factors of the previous moment, Vasubandhu often refers to them collectively as the “stream of aggregates” or “series of aggregates” (skandha-saṃtāna).

According to Vasubandhu’s theory, then, persons have a nominal existence, but they do not exist as a separate ontological entity—the kind of entity that is ordinarily thought to be the “self” or the “I.” Vasubandhu’s argument in favor of the no-self stance is simple, if controversial7: if an ultimate entity of an independent and separate self-existed, he argues, one would be able to know it either through direct perception or through inference. This is how people acquire knowledge about everything else that exists.8 However, neither of the two means of knowledge provides proof of the existence of a permanent self which is independent and separate from the five aggregates. Therefore, Vasubandhu concludes, a self beyond the five aggregates does not exist.9

This selfless analysis of the person defies the prevalent notion of an enduring moral agent. Consequently, it raises, and has to address, a number of philosophical difficulties. Specifically, Vasubandhu’s reductionist theory is required to explain who or what a moral agent is, given that the agent is not a persisting self, and to account for moral and normative conventions without asserting an enduring self. Indeed, the non-Buddhist interlocutor raises a series of questions and objections with regard to the prospect of accounting for moral agency and ordinary agential conventions through the conceptual framework that Vasubandhu uses: the schema of the five aggregates and the idea of “provisional designations” alone.

Considering the nature of the philosophical exchange, I take Vasubandhu’s project in the ĀVP as an attempt to fully translate ordinary agential conventions into impersonal language; in other words, to portray the various aspects of agency, as maintained by the realist, under the terms of the five aggregates without asserting a permanent self.10 Nevertheless, my account also shows that at two particular moments, Vasubandhu seems to have difficulties in translating ordinary agential conventions into impersonal language. These are the passages dealing with the sense of autonomous individuality and with self-interested concern for the future. Furthermore, in one particular place, Vasubandhu openly admits that his understanding of agency is different from that of the realist and that the realist’s notion of agency cannot be fully maintained in light of the impermanent and interdependent nature of reality. This is the passage in which Vasubandhu and his opponent are divided in their opinions about the agent’s self-capacity to perform actions.

The main reason that Vasubandhu does not succeed in translating the full range of ordinary agential conventions is that his project involves an intrinsic tension that arises from the two goals he aims to accomplish: first, he seeks to preserve, to the last one, the different conventions that he has his opponent present to him throughout the debate. At the same time, he strives to corroborate the Buddhist view of no-self, having both a philosophical and a soteriological motivation in mind. The tension between the two goals erupts in light of conventions that turn out to involve, inherently and by definition, a conception of an enduring self. In such cases, the consistency of the two sides of Vasubandhu’s project must be demonstrated, as the opponent in the ĀVP indeed requires him to do.

How does Vasubandhu deal with the problems of agency under the no-self premise? A close reading of his treatment of agency in the ĀVP reveals that it is characterized by two particular ways of approaching the issue. First, he chooses to explore agency from a descriptive perspective only. In accordance with this approach, Vasubandhu confines the discussion to the details of how different aspects of agency can be depicted using the terms of the five aggregates. What he is trying to avoid, it seems, is deriving normative conclusions from his description. Such normative conclusions could have been, for example, prescriptions for how we ought to act, whether we ought to modify our moral practices, or in what ways we ought to reformulate what matters to us in agency.

Vasubandhu may have had various reasons for this choice. It can be proposed that he assumes that the new understanding of what a person is will inevitably be followed by a normative shift, without the need to state the conclusions explicitly. Alternatively, it may be argued that Vasubandhu does not believe that absolute normative principles can be derived from the metaphysical view (an indication of this will appear later in the discussion). It may also be suggested that he was still unaware of the potential relation between the descriptive and the normative—between facts and values, the “is” and the “ought.” I suggest the contrary: that in the ĀVP, Vasubandhu was trying to avoid this issue entirely, and that he does so because had he derived normative conclusions from his descriptive account, it would have revealed his inability to preserve the entire range of ordinary normative conventions, for the simple reason that he would have had to modify ordinary normative values and principles in one way or another—something which, I believe, he was attempting to avoid in this case.

A second way in which Vasubandhu chooses to approach the problem of selfless agency is by examining it from a third-person perspective. Accordingly, agency is treated primarily through the eyes of an observer who is external to the agent himself. The other approach—that of the first-person perspective of agency, i.e., agency through the eyes of the agent himself—does not occupy a central place in the discussion. In this case also, different explanations can be proposed for Vasubandhu’s choice to examine agency from a third-person perspective. However, I believe that the reason that Vasubandhu refrains from the first-person perspective is that, contrary to the third-person perspective, it reveals the essential differences between agency under the realist view and agency under Vasubandhu’s view, which cannot capture the entire range of ordinary agential conventions. Our ordinary first-person perspective involves a sense of identity, a sense of being an enduring self or subject, and this element of the first-person perspective threatens to sabotage Vasubandhu’s project.

Two Notions of Moral Agents

The notion of the moral agent that Vasubandhu’s opponent expects him to explain in the ĀVP has three essential characteristics. First, the agent (kartṛ) is the one who performs the action (yaḥ karoti). Second, the agent is the owner (svāmin) of the action. And third, the agent has a certain self-capacity (svātantrya11) to perform actions, unaided by and independent of other factors.12 The first section in Vasubandhu’s examination of agency concerns the status of the agent of memory. In reply to his non-Buddhist opponent, who inquires who the agent of memory is, if there is no enduring self, Vasubandhu explains that the attribution of a memory to a particular person named Caitra involves two cognitive steps. First, one labels a stream of aggregates with the name “Caitra.” Then, at the moment in which a remembering thought arises from a perceiving thought in this stream of aggregates, one says that Caitra remembers.13 Vasubandhu immediately extends this account to other instances of momentary agential events. He argues that this cognitive process of attribution takes place when we state that a certain person apprehends an object and engages in similar cognitive activities.14 When certain recognition, apprehension, and so on arises in one of the five aggregates due to previous causes and conditions, it is said that the person designated on the basis of the five aggregates recognizes, apprehends, and so forth. What all the above cases have in common, it seems, is that they are temporally confined; that is, they take place during a single moment.

Vasubandhu, however, also refers to the model of the five aggregates when explaining the occurrences of continuous actions, such as walking, which take place over time. Here, he adds an epistemological layer to his theory by providing an explanation of how precisely it happens that we label the aggregates with a certain name despite the fact that they change from moment to moment. The reason for elaborating on this matter at this point seems to be that another factor has now been added, namely, the factor of time and continuity. In this section of the debate, Vasubandhu’s opponent claims that if persons are not selves, they cannot walk. Vasubandhu responds by claiming again that a personal name—in this case Devadatta—refers only to the stream of aggregates. Persons are “uninterrupted streams of momentary causally conditioned factors (saṃskāra).”15 Thus, similarly to momentary agential conventions, a continuous action is also attributed to a person on the basis of its taking place in that person’s stream of aggregates; it is only that in the case of a continuous action, the attribution is justified by the stream of aggregates being causally conditioned and uninterrupted, thus creating the appearance of a single entity. According to Vasubandhu, common people see one solid entity within this stream, namely, a sentient being, which moves from one place to another. In truth, however, Devadatta’s “walking” is simply the fact of the arising of his stream of aggregates in different places.16 The momentary and continuous conventions of agency are thus explained on the basis of the five aggregates of the person that is identified as the agent. The agent is the collection of five aggregates, designated by a certain name, in which a certain event or continuous action occurs.

Alongside this epistemological account of the nature of agency, Vasubandhu introduces an alternative view of the agent (kartṛ). In several places, he argues that agency—in accordance with his opponent’s definitions—is to be attributed to the primary cause (hetu, kāraṇa) of the agential act or event in question. This alternative notion of agency, he proposes, satisfies the first two characteristics of agency mentioned above, namely, the ownership of an action and the performing of an action.17

The problem of agency as ownership is discussed when Vasubandhu’s opponent raises the objection that if there is no permanent self, then the identity of the owner of memories is unclear. In order to argue that the cause of remembering can also be referred to as the owner of the memory, Vasubandhu questions his opponent on the latter’s definition of the meaning of ownership. In this part of the debate, which resembles a Socratic dialectical scrutiny, Vasubandhu draws from his opponent an example of ownership according to which a person is the owner of a memory in the same sense that a person, Caitra, is the owner of a cow. The essential expression of Caitra’s ownership of the cow is explained by the opponent as the position in which Caitra is found (adhīna) to employ or use the cow (tasyā viniyogaḥ) according to his own wishes. Vasubandhu then follows this example of ownership and argues that in each and every case, the cause of an action satisfies the opponent’s definition of ownership. The cause is that which exercises control over the action. Therefore, Vasubandhu claims, the cause by itself is sufficient as an owner and there is no need to assert a distinct self on top of it to fulfill this role. In the words of Vasubandhu,

If so, it is the cause, “the master,” which obtains, and it is the effect which is “the servant.” In this manner, the cause governs its effect and by that very effect, the cause is its owner.18

To conclude his argument, Vasubandhu returns to the model of the five aggregates and reminds his opponent that Caitra and the cow are in fact only two streams of aggregates, and that Caitra’s ownership of the cow boils down to the five aggregates of Caitra being the cause of transformations in the five aggregates of the cow. Vasubandhu then applies this account to all other actions attributed to agents, beginning with the momentary events of recognition, apprehension, and the like, through continuous acts such as walking, and ending with actions that carry positive or negative karmic potential. In the same way, later in the chapter, the owner of the “sense of individuality” (ahaṃkāra) is also explained by Vasubandhu to be its cause.19

Jonardon Ganeri (2007) is right, in my opinion, in saying that “Vasubandhu’s way of dealing with the objection [against reducing facts about ownership to facts about causal connection] is less than convincing” (p. 175). I would add that this is not only because this account leaves untreated various issues concerning the notion of ownership, but also because it seems to misrepresent the notion of the agent-owner itself as Vasubandhu’s realist opponent conceives of it. It seems that what the opponent has in mind is a continuous owner, who owns his memory not at a particular point in time, but over many moments in time, even after this memory has ceased to be actively present. Moreover, this owner seems to be one who is able to possess several memories, actions, and so forth, not a momentary owner who exercises control over one particular event and then perishes away.20 Vasubandhu’s suggestion that the owner is a certain cause in the collection of aggregates that comprises the person seems, on the face of it, a sophistic move, which does not provide a genuine solution to his opponent’s notion of ownership, but rather redefines the original notion. However, I believe that this problem may be solved if one recalls the interplay between the conventional and ultimate notions of agency. Since any event that occurs within the person’s stream of aggregates is also attributed to the person himself (the five aggregates designated by a particular name), it can be inferred that the conventional person whose five aggregates contain the owner-cause of a certain memory is also the owner of that memory. This solution is supported to a certain degree by Vasubandhu’s final account of the nature of Caitra’s ownership of the cow. The nature of the relation between Caitra and his memory, however, is not explicitly spelled out in the same way.

Vasubandhu reformulates the same metaphysical notion of agency when he treats its second characteristic, namely, the performing of actions. Thus, as part of his account of memory, Vasubandhu claims that “the one who performs it [the act of remembering] (yas tāṃ karoti) has been explained: the cause of remembering (smṛtihetu) is a special mind moment (cittaviśeṣa).”21 Similarly, when his opponent asks him to explain who the doer of actions or creator of karmic potential (karmaṇāṃ kartā) is, Vasubandhu ascribes the agency to the cause of the action in the stream of aggregates. In Vasubandhu’s words, “that which is the chief cause (pradhāna-kāraṇa) of it [i.e., of the action], is said to be its agent (kartṛ).”22 Finally, in this section of the debate, Vasubandhu also rejects the third characteristic of agency that his opponent puts forth, i.e., an independent power (svātantrya) to act, claiming that there is nothing in the process of performing an action that is independent of other factors. In addition, he argues that a self that is independent of other factors does not participate in causality, and therefore logically cannot cause anything, or, in other words, cannot perform any action. Vasubandhu, therefore, concludes that a single entity, a producer dependent on itself, cannot be ascertained.23

Even though the cause as an agent has no independent power to produce the effect, one passage in the discussion about primary causes (kāraṇa-hetu) in the second chapter of the AKBh suggests that in moral contexts the cause serves to distinguish between the agent and other individuals. In this passage, Vasubandhu claims that primary causes can be divided into two categories: (1) the chief (pradhāna) cause, which leads to the result, and (2) causes that are potentially capable of posing a hindrance to the arising of the result, yet which in practice do not.24 The second category includes all factors (dharma) outside of the chief cause, which are involved by way of not interfering with the arising of the result. At this point, an objection is raised as to the reason why when a murder is committed, not all sentient beings, like the murderer himself, are morally responsible for that murder. In reply, Vasubandhu explains that all the factors are acknowledged as efficient causes because they do not constitute an obstacle. It is not the case that they are all agents (kāraka).25 In other words, the chief cause, which is plainly the cause that yields the effect, defines who the agent is. All other factors, which are causes involved in the production of the effect by not obstructing it, are not the agent and therefore do not carry responsibility for that action.

Two Conventions That Do Not Withstand Impersonal Articulation

Earlier, I discussed the strategy that Vasubandhu adopts in accounting for the notion of the moral agent under the no-self premise, which he uses to explain various agential conventions. Some of these have already been discussed: momentary agential conventions, such as remembering and perceiving, and continuous conventions, such as walking. I also mentioned at the beginning that Vasubandhu appears to be interested in keeping the debate focused on descriptive account of agency and on the third-person perspective, and I suggested that the reason for this is that the two complementary approaches—developing a normative discussion on agential conventions and adopting a first-person perspective—would expose the inner tension at work in Vasubandhu’s project. However, there are two passages which are exceptional in that they do mention the aspect of a first-person perspective. For this reason, they reveal further layers of Vasubandhu’s notion of agency under a no-self premise and the implications that selfless agency has for ethics.

The first passage that constitutes an exception to Vasubandhu’s overall strategy (and hence, I suggest, threatens the coherence of his project) concerns the basis for distinguishing between the agent and other individuals. The opponent requests that Vasubandhu clarify the following problematic: if the “sense of individuality” (ahaṃkāra) arises when one conceives one’s aggregates—one’s body, for instance—why is it that this conception does not arise with regard to the bodies of other people? In this question, the opponent seems to point to the assumption that there must be some qualitative difference between oneself and others, and to suggest that this difference can be accounted for only by independent selves, which Vasubandhu rejects. In reply, Vasubandhu explains that there is a unique relation between the “sense of individuality” and one’s own five aggregates, which does not obtain for the aggregates of others.26 Vasubandhu does not provide more details about the nature of this relation, except that it is a relation of cause and effect. He offers, however, a certain etiology by saying that it is the result of a mental habit which has no point of beginning.27 This beginningless habit of conceiving individuality involves ignorance: Vasubandhu explains that the cause of “the sense of individuality” is “a mind moment accompanied by ignorance, whose object is its own stream [of aggregates] and which is pervaded by a former ‘sense of individuality.’”28 In other words, the cause of the sense of individuality, according to Vasubandhu, is a mental event which conceives the stream of aggregates in which it occurs and is clouded by a misunderstanding of the true nature of the person. The cause of the sense of individuality itself displays that same sense of individuality, which indicates that such a mental state is caused by a previous mental state of the same kind, and so on with no identifiable point of beginning.

Two things should be noted in this last account of individuality. First, that Vasubandhu adheres to the descriptive level. His response does not include a further normative step of justifying or rejecting this convention, and he does not theorize normative implications that may stem from his account. This is in stark contrast to the approach of other Buddhist thinkers, such as Śāntideva and Buddhaghosa, who purposefully utilized metaphysics to modify ordinary normative conventions in order to better fit with the Buddhist worldview.29 A second point to be noted is that ignorance with regard to the true nature of the self plays a central role in Vasubandhu’s account. Here again, Vasubandhu is compelled to describe agency using an idea—the deluded “sense of individuality”—that he had been trying all along to avoid and eradicate from our epistemology. In doing this, the passage reveals the inescapable tension in Vasubandhu’s project: if one rejects the idea of a conventional permanent self, then not all ordinary conventions can be accounted for. In this case, Vasubandhu was required to add a foreign concept, which seems prima facie to contradict his soteriological view. Otherwise, he could not explain how a particular agent could distinguish himself from other individuals.

What can be seen is that an inconsistency arises when the first-person perspective comes into the picture. Vasubandhu could use his strategy if the question were about distinguishing between one agent and another from a third-person perspective, such as in the case of distinguishing between Caitra and the cow as two different streams of aggregates and individuals. This was accomplished by applying only the concepts of aggregates and provisional designations. However, here the issue in question is how to distinguish between myself and another agent, not merely between two different agents, and in order to solve this issue, it is required that an ordinary first-person perspective, along with ignorance regarding the true nature of the self, be called into action. Here, Vasubandhu seems to face a dilemma: he either has to admit of being unable to fully account for ordinary agency without a notion of a permanent self, or he has to dismiss an essential aspect of agency which is also central to his own soteriology. In this case, Vasubandhu takes the first route and maintains the convention of distinguishing between moral agents (or between moral agents and moral subjects) from a first-person perspective. However, in order to do so, he is required to resort to the idea of a “sense of individuality,” which is pervaded by ignorance. One crucial point to observe in this exceptional section of the dialog is that maintaining a subjective sense of agency requires the acceptance of a certain unified self.

In the second of the two passages, Vasubandhu’s opponent raises the problem of self-interested concern for the future. This problem, which is also known in Western philosophy (see, for example, Schechtman 1996, pp. 52–53), concerns the ordinary intuition that our anticipation for the future rests on the assumption that we possess an enduring self, which unifies our existence at different times. Without such a persisting identity, it is difficult to explain why we anticipate our future experience in a different way than, say, the experiences of our best friend and what reasons we have to act in order to shape our own future. Vasubandhu’s opponent’s question, then, is also about the end or reason (artha) for undertaking actions for one’s own future welfare in the absence of a self. In the ĀVP, Vasubandhu replies that the purpose for which self-interested actions are taken can be described as “so that I shall be happy and not suffer.”30 It is important to note that this exchange, as it is phrased here, is open to two levels of interpretation, one descriptive, one normative. According to the descriptive interpretation, the opponent asks for an account of the process that takes place when people undertake actions for the future, given that there is no permanent self. However, according to the normative interpretation, the opponent requires much more than that: he requires that Vasubandhu provide the reasons and motivations for undertaking those actions. In other words, what the opponent asks is why we ought to take actions, not why people do so in practice.

It should be remembered that in the context of this debate, the question of self-interested concern for the future must have had a broader import than establishing reasons for worldly egoistic concern; the spiritual path, both Buddhist and non-Buddhist, is founded on the principle that one can achieve spiritual liberation as a result of continuous practice and on the idea that seeking self-liberation from saṃsāra is a legitimate motivation for undertaking the spiritual path, as Vasubandhu himself indicates at the outset of the chapter.31 This double descriptive-normative meaning is maintained in the Tibetan translation.32 What follows after this part of the dialog, however, can only be interpreted as a descriptive account:

Question: What is it that is called I?

Vasubandhu: That which is the object of this sense of individuality (ahaṃkāra).

Question: What is the object of this sense of individuality?

Vasubandhu: The object is the aggregates.

Question: How is it known [to be the aggregates]?

Vasubandhu: Because of the attachment (sneha) to them [i.e., to the five aggregates and not to a self].33

It seems, then, that whereas the opponent’s initial question is best understood as requiring a normative explanation, Vasubandhu’s reply must be taken as providing a descriptive account. According to this account, self-interested actions are performed by agents because they are attached to a self, which is in fact the stream of aggregates, and the aim of their actions is for this non-existing self to be happy and not suffer. But in taking this route and only discussing the problem of self-interested concern for the future from a descriptive perspective, Vasubandhu does not provide a satisfactory solution to the normative question of reasons and motivations to act, since he only describes the current state of affairs, which is characterized by ignorance. He does not describe the way in which a person who has become disillusioned with the belief in a self comes to act.
On the other hand, if Vasubandhu’s reply is to be taken as an attempt to come up with a reason or justification for performing actions driven by self-interested concern for the future, then his solution involves an apparent paradox. This paradox results from the clash between the requirement to let go of the notion of a permanent self as a condition for liberation and the necessity to assume such a self as a motivation for acting in the interest of one’s future and also from the clash between the requirement to eradicate attachment to the self as a condition for liberation and the necessity to have attachment to the self and its happiness as a motivation for acting in the interest of one’s future. As has already been noted, this paradox does not only concern ordinary egoistic actions. Rather, it is present in the injunctions of the spiritual path itself, as advocated by Vasubandhu. The paradox goes as follows: In order to attain liberation from suffering, the Buddhist path instructs the practitioner to let go of the sense of individuality and of the attachment to the happiness of one’s self. At the same time, the realization of the Buddhist goal, nirvāṇa, presupposes that the practitioner is attached to his self and to his future happiness, whose highest embodiment is spiritual liberation, and this requires one to accept a sense of enduring individuality. However, as is well-known, maintaining attachment and a sense of individuality leads to suffering and to the perpetuation of rebirth in saṃsāra, which then contradicts the former injunction. In short, the paradox can be framed in two propositions in the following way:
  1. (1)

    In order to attain liberation, one needs to relinquish the sense of enduring individuality and let go of attachment.

  2. (2)

    In order to follow the injunction expressed in proposition (1), one needs to maintain a sense of individuality and be attached to one’s future happiness.


As mentioned, Vasubandhu does not develop his treatment of this problem further, and thus leaves his opponent (and the reader) either without a normative answer or with a normative paradox. However, in another section in the AKBh, Vasubandhu touches once again on the paradox of self-interested concern for the future, as it may be called, and shows both the ethical significance of the subjective belief in an enduring self and the conditions under which this assumption can be accepted. He also shows what normative conclusions can be drawn from the negation of an ultimate enduring self for the way in which one ought to care for one’s own future happiness.

The Paradox of Self-Interested Concern for the Future

What I call “the paradox of self-interested concern for the future” consists of two elements. The first element is related to the simultaneous maintaining and letting go of the view of an enduring self, while the second element is related to the simultaneous maintaining and letting go of the attachment involved in wishing for future happiness. A closer look at the second element reveals that this is in fact already a well-known and oft-discussed problem in Buddhism, the so-called “paradox of desire.” This problem, already acknowledged by traditional Buddhist thinkers, has been the topic of several essays, although none of them discuss it with particular reference to Vasubandhu’s thought. This paradox arises from the apparent contradiction in the principle that prescribes to “desire to end all desire.” As Herman 1979 nicely summarized the problem, “[t]he paradox of desire points to the practical contradiction or frustration involved in the desire to stop all desiring and states simply that those who desire to stop all desiring will never be successful” (p. 91). From a different perspective, this paradox finds expression as the problem that arises from the idea of desireless action.34 It is this aspect of the paradox of desire that makes it so relevant, in my opinion, to the paradox of self-interested concern for the future. For, as already mentioned, one of the two questions at the heart of the paradox of self-concern for the future is how can one act with an attachment to the happiness of one’s future self while aspiring to eliminate all attachment?

Modern thinkers have raised a number of solutions to the paradox of desire.35 Here, however, I want to examine how Vasubandhu treats this problem—a problem which he himself links to the other aspect of the issue in question, namely, to the contradiction between the requirement to simultaneously maintain and eradicate the sense of an enduring self. The context of the discussion is the classification of latent tendencies (anuśaya), mental afflictions in their dormant state, or in other words, dispositions in a state of potency (as opposed to their manifest state, when the dispositions are expressed). The text goes on to explain that unwholesome latent tendencies, such as attachment, hostility, and conceit, which have “impure” (sāsrava)36 cognitive objects (ālambana), tend to “stick” and grow (anuśerate), whereas latent tendencies, which have “pure” (anāsrava) cognitive objects, behave differently; they do not stick.37

What does it mean for a latent disposition to be directed at a pure cognitive object, or alternatively at an impure cognitive object? To put it plainly, pure cognitive objects are objects that are related to the cessation of suffering and to the path leading to this cessation (to ethics and a correct understanding of reality, to name but two elements of the path). They are said to be harmless (anapakāra), peaceful (śānta), pure (śuddhi), and excellent (agra). Impure cognitive objects, on the other hand, are connected to suffering and its causes; they harm, they are not peaceful, they are impure, and they hold as excellent that which is low.38 The reason that latent tendencies which concern the pure do not stick and grow is that their objects cannot be considered as I or “mine.”39 In other words, with regard to these objects, one does not develop an attitude of possessiveness either through the afflicted view of an enduring self or through attachment. At the same time, pure cognitive objects oppose the mental afflictions, insofar as they do not sustain them.40 Impure cognitive objects do the opposite: they cause the mental afflictions to remain and to grow.

In accordance with the abovementioned distinction between latent tendencies with pure cognitive objects and latent tendencies with impure cognitive objects, Vasubandhu identifies two types of desire, which are qualitatively different:

But that which is standing here, [the latent tendency] which asks for that [high] level [of the path, i.e., asks for the pure]—this is a wish for wholesome factors [kuśaladharmacchanda; not the latent tendency of attachment].41

Because of this difference, unwholesome attachment, whose cognitive object is impure, should be abandoned, but the wish for wholesome factors, whose cognitive object is pure, should not be rejected.42 Attachments of the first kind, it can be deduced, result in a denser presence of attachment and suffering in the mind. They constitute a hindrance to the path and so distance one from liberation. However, the wish for wholesome factors, having a pure cognitive object, opposes the mental afflictions and is conducive to the path and to liberation.

Vasubandhu’s method for solving the paradox of desire is, then, to distinguish between those attachments which are unwholesome, and hence lead to suffering, and their wholesome counterparts, wishes that concern the cessation of suffering and the path that leads to this cessation. Vasubandhu’s solution to the paradox of desire is, hence, that the wish for liberation from saṃsāra is not the kind of attachment that must be eradicated as part of the spiritual path. If I understand Vasubandhu correctly, his view about the convention of self-interested concern for the future is that this ordinary convention need not be rejected. The injunction to eliminate attachment is not inconsistent with being concerned for future happiness. What this injunction does entail is that one should modify the kind of happiness one searches for and the way one works to achieve this happiness.

According to the ordinary worldview, one pursues happiness by following attachment, which grasps at the self and at impermanent objects with a possessive mind. Searching for happiness in such a way, according to Vasubandhu’s analysis, perpetuates the presence of mental afflictions in the stream of aggregates and reaffirms the belief in a self (in fact, according to Buddhist premises, this approach should not be at all regarded as pursuing happiness, but rather as pursuing suffering). The selfless way of being concerned about one’s future happiness, on the other hand, aspires to attain liberation from saṃsāra and to follow the spiritual path that leads there. It means acting on another type of motivation, “the wish for wholesome factors,” which opposes and erodes the mental afflictions and the belief in an enduring self rather than maintaining them.43 Thus, one can be attached to one’s future happiness, wish not to suffer, and wish to ultimately attain liberation; it is only the kind of happiness to which one ought to be attached and the way happiness ought to be sought that need to change.

In discussing the varieties of attachment, Vasubandhu addresses one aspect of the paradox of self-interested concern for the future. However, the conundrum of maintaining and letting go of the sense of individuality (ahaṃkāra) is yet to be resolved. To tackle this issue, Vasubandhu discusses the moral status of another latent disposition, the afflicted view of an enduring self (satkāyadṛṣṭi), where he adopts a different strategy to the one he employed in addressing the paradox of desire. It should be noted here first that although in the AKBh Vasubandhu uses various expressions to discuss the belief in an enduring and unitary self, his discussion of the notion indicates that they all denote the same idea. Thus, the concept of the “afflicted view of an ‘I’” (ātmadṛṣṭi), which appears in Vasubandhu’s reductionist argument, the concept of the “afflicted view of an enduring self” (satkāyadṛṣṭi), which Vasubandhu uses in the discussion that follows below, and the concept of the “mistaken view of a self” (ātmaviparyāsa), which appears in other places in the AKBh all refer to the same idea of grasping at an enduring and unitary self (with satkāyadṛṣṭi having a somewhat wider extension, which also includes the view that the self is the owner of the aggregates).44 Moreover, in his commentary to the AKBh, Yaśomitra equates these concepts with the concept of a “sense of individuality” (ahaṃkāra) found in the discussion of the agential conventions of self-interested concern for the future and agential autonomy, which appeared earlier.45

Vasubandhu argues that unlike certain other latent dispositions, which are inherently unwholesome, the afflicted view of an enduring self is morally neutral (avyākṛta), that is, this latent disposition in itself is neither wholesome nor unwholesome. Likewise, the ignorance that accompanies the view of an enduring self is also morally neutral.46 What this means is that this view and the accompanying ignorance do not necessarily lead to the accumulation (upacaya) of negative (akuśala, aśubha) karman. One of the reasons provided by Vasubandhu for this neutral moral qualification is that maintaining the afflicted view of an enduring self is not contradictory to acting morally—practicing generosity, for example. Ignorance with regard to the true nature of the person can be wholesome in that it motivates us to pursue our future happiness and to act morally in order to achieve it.47 Here, Vasubandhu’s explanation echoes the problem from which this discussion stemmed:

[Question:] What is the reason [for the afflicted views of an enduring self and of holding to extremes, and the ignorance which is connected with them, being morally neutral]? [Vasubandhu:] It is because they are not incompatible with giving and so on. [With the thought] “I shall be happy in the next life,” one gives a donation, one observes moral conduct.48

In other words, one may believe that one exists as an enduring self, and consequently live egoistically and immorally, creating negative karman; but one may also think in this way and thereby be motivated to act morally, in such a way that brings real happiness in the form of liberation.

The question may be raised of why Vasubandhu did not develop these last points in the dialog with his philosophical opponent. As I have argued before, I suggest that the reason he did not do this is that it would have undermined his overall project in the ĀVP by revealing that the Buddhist view of no-self does modify ordinary agential conventions in a certain way, contrary to his attempt to show how all conventions can be retained under the no-self premise.

Conclusion: Vasubandhu’s Notions of Moral Agents and Their Normative Significance

The final section of the discussion is instructive with respect to the way in which Vasubandhu understands the concept of moral agency. It is also revealing with respect to the relations between the ontological status of the moral agent and normative considerations. First, Vasubandhu takes the view that persons are enduring selves to be morally insignificant, in the sense that definite normative values and principles cannot be conclusively derived from it. He particularly emphasizes the undetermined karmic value of this view. Due to this indeterminacy, it is implied, the view can justify different moral theories: it may be associated with a lifestyle of negativities (guided by egoistic self-interest or by false beliefs)49 or it may encourage an ethical view which defeats the clinging to an enduring self. Because the afflicted view of an enduring self is open to a range of different, even contradictory, normative interpretations, this renders it ethically and normatively futile, as ultimately no definitive moral principles or reasons to act can be derived from it.

However, moral principles and reasons to act can be extracted from the ultimate notion of the moral agent. A normative moral theory which assumes ultimate selfless and momentary moral agents would prescribe principles that embody this assumption, such as moral sentiments and motivations to act that do not assert an enduring self. Another normative consequence of Vasubandhu’s ultimate notion of the agent is the redefinition of happiness and the way in which one ought to be concerned about one’s future happiness. Happiness is redefined in a way that does not involve the view of an enduring self (even if this forms a motivational factor). From a meta-ethical point of view, Vasubandhu’s concept of the ultimate agent redefines the good as that which does not reinforce the belief in, and attachment to, an enduring self.

Finally, while clinging to a sense of individuality can lead to a variety of moral theories and views, the discussion shows that holding to an enduring self is a necessary condition for two essential conventions: for a subjective distinction between the moral agent and recipients of action, and as a motivation to pursue the spiritual path and to observe ethics. The way in which this issue is treated by Vasubandhu sheds a different light on the prevailing presupposition that identification with a self is always a hindrance to being moral whereas the realization of no-self is a requirement for perfect morality. What I have shown here is that, at least in the AKBh, a certain reconstruction of identity is a requisite for engaging in actions for the future in general, and in ethical behavior in particular. Whether Buddhist ethics is seen as a consequentialist moral theory (as Goodman (2009a) advocates) or as a form of eudaimonian ethics (as maintained, for example, by Keown (1992)), the agent needs to have an outlook for the future. Under both theories, one undertakes actions because one cares for one’s future—because one wants to achieve liberation, for example—and this care is embedded in a sense of an enduring self.


  1. 1.

    References to the AKBh in Sanskrit are to the critical edition by Pradhan (1975). References to the AKBh in Tibetan (henceforth AKBhT) are to Zhu-chen Tshul-khrims-rin-chen (1985). In translating from Sanskrit, I was aided by Yasunori Ejima’s notes in Lee (2005) and by Lambert Schmithausen (personal communication). References to Yaśomitra’s Sphuṭārthavyākhyā on the AKBh (henceforth AKVy) are to Shastri (1970-1973). All translations are mine unless otherwise mentioned. Verses (karikā) from the Abhidharmakośa appear in the translations and transliterations from Sanskrit in bold letters.

  2. 2.

    AKBh IX, p. 461; AKBhT Khu 82a1–2.

  3. 3.

    As has also been recently proposed by Meyers (2014), who suggests that in the Pāli suttas and the Abhidharma traditions, certain elements that are involved in practicing the path—goal-oriented actions, effort, and initiative—require that one regard oneself as an autonomous agent through self-grasping, and McGarrity (2015), who argues that the works of Mādhyamika philosophers, such as Nāgārjuna, Āryadeva, and Candrakīrti, advocate the reconstruction of the person as an agent and with it a sense of personhood, which serve as the basis for a teleological orientation towards future goals.

  4. 4.

    The exact philosophical affiliation of Vasubandhu’s opponent, who in the ĀVP is simply called a tīrthika (a non-Buddhist thinker), is not fully clear. Duerlinger (2003, pp. 117–118, n. 60) suggests that the opponent in the later part of the chapter—the part I will consider below—is a Nyāya-Vaiśeṣika philosopher. La Vallée Poussin (La Vallée Poussin and Sangpo 2012, pp. 2632–2633, n. 166), on the other hand, holds that certain passages of the same part of the debate are directed towards a Buddhist Pudgalavādin philosopher. Charles Goodman (2009b, pp. 297–299) comments that some of the passages concern the ideas of the Nyāya-Vaiśeṣika school, while others seem to be directed towards a proponent of the Sāṃkhya school; as a whole, they concern only non-Buddhist schools. On the topic of agency in Nyāya thought, see Dasti (2014a). On the notion of agency in Sāṃkhya philosophy, see Bryant (2014).

  5. 5.

    As with other Buddhist concepts, the nature of the dharmas was a contested issue among the different Buddhist schools. The AKBh presents the accounts of the Sarvāstivāda and the Sautrāntika schools. On the historical and theoretical development of the dharma theory, see Gethin (2004); Ronkin (2005), pp. 34–85; and Williams (1981).

  6. 6.

    Vasubandhu’s explanation of the five aggregates echoes earlier depictions of this taxonomy, such as its depiction in the Pāli Canon and in earlier Sarvāstivāda accounts. On the five aggregates in the Pāli Nikāya, see Hamilton (1996). On the five aggregates in the Sarvāstivāda tradition, see Dhammajoti (2007), pp. 30–32, 242–272.

  7. 7.

    For critical examinations of this argument, see Duerlinger (2003), pp. 123–130; Gold (2015), pp. 99–104; and Kellner and Taber (2014), pp. 719–727.

  8. 8.

    Although Vasubandhu generally accepts three means of knowledge (direct perception, valid inference, and scriptural authority; see, e.g., AKBh II:46b, p. 76; AKBhT Ku 81a7), in his argument, he acknowledges only direct perception and valid inference as reliable sources for proving or disproving that a self exists. La Vallée Poussin (La Vallée Poussin and Sangpo 2012, p. 2590, n. 11) suggests, following Yaśomitra, that in this argument, proof by scriptural authority is not mentioned because it is included within the means of valid inference. Duerlinger (2003, p. 128) mentions another explanation, provided originally by the Chinese commentator P’u-kuang, according to which scriptural authority is not mentioned because the argument is directed towards non-Buddhist thinkers. Vasubandhu and his non-Buddhist opponents cannot come to an agreement on scriptures, since they follow different corpuses, but they can argue on the basis of perception or inference, which are two principles on which they do agree.

  9. 9.

    AKBh IX, p. 461; AKBhT Khu 82a2–5.

  10. 10.

    Cf. Chadha 2017, who argues that Vasubandhu rejects the notions of person and agent and that a reconstruction of selfhood and a sense of agency is not in line with the spirit of Buddhism, and Duerlinger (2003), p. 240, who holds that Vasubandhu had no pretensions to translate, without loss of meaning or information, sentences about persons to sentences about the five aggregates.

  11. 11.

    According to Matthew Dasti (2014b, p. 3), svātantrya in Indian thought is the best equivalent Sanskrit term to the notion of free will in Western thought. He explains that this concept suggests the capacity for self-determined action. If this is true, then Vasubandhu’s rejection of the idea of svātantrya has implications for how we ought to understand his conception of free will. However, I will not develop this point further here. For further discussion surrounding the topic of free will and determinism in Vasubandhu’s thought, see Gold (2015), pp. 180–188, and Meyers (2010), Sects. 3 and 5.

  12. 12.

    It is interesting to note that this notion of a moral agent as someone who performs actions and owns them may sound too thin in other philosophical contexts. It does not encompass certain characterizations, such as the capacity for deliberation, having reasons and motivations to act, being accountable for actions, or possessing the freedom to act, which, according to certain thinkers, are essential for agency.

  13. 13.

    AKBh IX, p. 473; AKBhT Khu 91a5.

  14. 14.

    AKBh IX, p. 473; AKBhT Khu 91b2–3.

  15. 15.

    AKBh IX, p. 473: kṣaṇikā… saṃskārā abhinnasaṃtānā. AKBhT Khu 91b5: ‘du byed skad cig pa rgyun tha mi dad pa rnams.

  16. 16.

    AKBh IX, p. 473; AKBhT Khu 91b5–6.

  17. 17.

    While Vasubandhu’s main interlocutor seems to be a proponent of the Nyāya-Vaiśeṣika schools or the Sāṃkhya school, the context of the debate also arises from the notion of agency expounded by the Grammarians, where the agent (kartṛ) is said to be endowed with svatantra; the term svatantra is understood, according to George Cardona (2014, pp. 86–87), as “one who has himself as the principle person,” although the meaning is not clarified further. The relevance of the Grammarians’ understanding of agency to this particular debate of Vasubandhu is, moreover, lucidly demonstrated by Salvini (2008), pp. 101–121, esp. 115–118. I thank Mattia Salvini for sharing his dissertation with me.

  18. 18.

    AKBh IX, p. 473; AKBhT Khu 91a7–91b1.

  19. 19.

    AKBh IX, p. 476; AKBhT Khu 93b5.

  20. 20.

    From a Western philosophical perspective, notions of agency, which resemble the ultimate notion of agents as proposed by Vasubandhu, have received a great deal of criticism through what is known as the “disappearing agent” objection. The essence of this objection is that by leaving out the agent (as a persisting entity), theories such as Vasubandhu’s fail to capture agency, since all instances of agency turn into events that causally happen to us rather than actions that are performed by someone. On the problem of the disappearing agent, see Lowe (2008), pp. 159–161; Mele (2003), pp. 215–220; and Steward (2013).

  21. 21.

    AKBh IX, pp. 472–473; AKBhT Khu 91a5.

  22. 22.

    AKBh IX, p. 476; AKBhT Khu 94a3.

  23. 23.

    AKBh IX, p. 476; AKBhT Khu 94a3.

  24. 24.

    AKBh II:50a, p. 82; AKBhT Ku 86a5.c AKBh II:50a; AKBhT Ku 86b2.

  25. 25.

    AKBh II:50a, p. 83; AKBhT Ku 86b3–4.

  26. 26.

    AKBh IX, p. 476; AKBhT Khu 93b4.

  27. 27.

    AKBh IX, p. 476; AKBhT Khu 93b4–5.

  28. 28.

    AKBh IX, p. 476: pūrvāhaṃkāraparibhāvitaṃ svasantativiṣayaṃ sāvadyaṃ [Schmithausen emends to sāvidyaṃ] cittam. AKBhT Khu 93b5–6: sngon bdag tu ‘dzin pas yongs su bsgos pa rang gi rgyud kyi yul can ma rig pa dang bcas pa’i sems yin no.

  29. 29.

    On deriving normative values from the principle of no-self in Śāntideva, see Williams (1998), Sects. 2 and 5, and following discussions by Clayton (2001), Garfield (2012), Garfield et al. (2016), Harris (2011), Siderits (2000, 2016), and Wetlesen (2002). For this topic in Buddhaghosa, see Goodman (2002) and Jenkins (2016).

  30. 30.

    AKBh IX, p. 476: ātmany asati kim arthaḥ karmārambhaḥ| ahaṃ sukhī syām ahaṃ duḥkhī karmārambhaḥ na syām ity evam arthaḥ. AKBhT Khu 93b1: bdag med na ci’i phyir las rtsom | bdag bde bar gyur cig | bdag sdug bsngal bar ma gyur cig ces bya ba de’i phyir ro.

  31. 31.

    James Duerlinger (2003, p. 279) hypothesizes, in addition, that the problem of self-concern for the future has a bearing on altruistic actions. He argues that the context in which Vasubandhu’s argument is presented suggests that according to the Nyāya-Vaiśeṣika opponent, as well as according to Vasubandhu, performing actions for the sake of others requires that we first have reasons to act in our own interest.

  32. 32.

    It is also maintained, to a certain degree, in the translations of La Vallée Poussin and Sangpo (p. 2571) and Duerlinger (p. 104), although my impression is that both lean towards a normative understanding of the question. Matthew Kapstein’s translation (Kapstein 2001, p. 372) follows the normative understanding wholeheartedly and unambiguously, whereas Charles Goodman (2009b, 303) seems to understand the entire exchange according to the descriptive sense. Yaśomitra’s AKVy, unfortunately, does not shed light on this point.

  33. 33.

    AKBh IX, p. 476: ko ‘sāv ahaṃ nāma yad viṣayo ‘yam ahaṃkāraḥ [Ejima: nāma| yad viṣayo (’)yam ahaṃkāraḥ| < kiṃviṣayo ‘yam ahaṃkāraḥ > |] skandhaviṣayaḥ| kathaṃ jñāyate| teṣu snehāt. AKBhT Khu 93b1–2: yang bdag ces bya ba yang gang zhig yin | bdag tu ‘dzin pa ‘di’i yul gang yin pa’o || bdag tu ‘dzin pa ‘di’i yul gang zhig yin | yul ni phung po yin no || ji ltar shes | de dag la chags pa’i phyir dang. The emphasis and the dialog form are mine. This part can also be understood as consisting of rhetorical questions and answers.

  34. 34.

    On this problem, see Taber (2011).

  35. 35.

    For modern solutions to the paradox, see Alt (1980); Herman (1979), pp. 93–94; and Visvader (1978), p. 463.

  36. 36.

    For the sake of simplicity, I follow La Vallée Poussin and Pruden (1988, 1990) and La Vallée Poussin and Sangpo (2012) in translating the terms anāsrava and sāsrava here as “pure” and “impure,” respectively. A more literal translation would render the first as “without outflows” and the latter as “with outflows.” Outflows (āsrava) are states that taint the mind and prevent the attainment of awakening. Vasubandhu lists three kinds of outflows: the outflow of pleasure (kāmāsrava), the outflow of existence (bhavāsrava), and the outflow of ignorance (avidyāsrava). See AKBh V:35–37, pp. 306–307; AKBhT 246a6–247a3.

  37. 37.

    AKBh V:17–18, p. 289; AKBhT Ku 235a5-235b1.

  38. 38.

    AKBh V:14, p. 288; AKBhT Ku 234b1–2. AKBh V:16, pp. 288–289; AKBhT Ku 235a1–4.

  39. 39.

    AKBh V:18, p. 289; AKBhT Ku 235b1.

  40. 40.

    AKBh V:18, p. 289; AKBhT Ku 235b2–3.

  41. 41.

    AKBh V:18, p. 289: yas tv ihasthas tāṃ bhūmiṃ prārthayate kuśalo ‘sau dharmmacchandaḥ. AKBhT Ku 235b3: gang zhig ‘di na gnas pa las de don du gnyer bar byed pa de ni mi dge ba’i chos la ‘dun pa yin no. The translation follows the Sanskrit.

  42. 42.

    AKBh V:16, p. 289; AKBhT Ku 235a1–2.

  43. 43.

    See also Meyers (2014), p. 44, who proposes a similar idea.

  44. 44.

    AKBh V:16, p. 289; AKBhT Ku 231a3–5. AKBh V:9ab, pp. 283; AKBhT Ku 231a2–3.

  45. 45.

    AKVy V:9ab, p. 778.

  46. 46.

    AKBh V:19 ac, p. 290; AKBhT Ku 236a2. AKBh V:19d-20ab, p. 291; AKBhT Ku 236a6–7.

  47. 47.

    Martin T. Adam (2005) suggests that one ought to distinguish between different types of agents in the Pāli Canon. According to this distinction, the good conduct of ordinary persons (puthujjana) is informed by the delusion of self; however, disciples in higher training (sekha) have penetrated the delusion of self by insight and are drawn to nirvāṇa, but their good conduct is not motivated by the goal of attaining it for themselves. Arhats (including the Buddha) have eradicated delusion completely and so their activity is entirely free from it.

  48. 48.

    AKBh V:19bc, p. 290: kiṃ kāraṇam| dānādibhir aviruddhatvāt| ahaṃ pretya sukhī bhaviṣyāmīti dānaṃ dadāti śīlaṃ rakṣati. AKBhT Ku 236a2–3: ci’i phyir zhe na | sbyin pa la sogs pa dang mi ‘gal ba’i phyir dang | bdag ‘jig rten pha rol du bde bar ‘gyur bar bya’o zhes sbyin par byed || tsul khrims srung bar byed do.

  49. 49.

    For example, the belief in a creator god (AKBh V:7–8, p. 282; AKBhT Ku 230a6–7). Cf. AKBh V:13ab, p. 287; AKBhT 234a2–3.



The article was written with the kind support of a research grant by the Deutscher Akademischer Austauschdienst (DAAD). I would like to thank Charles E. Snyder, Roy Tzohar, David Weinstein, and Michael Zimmermann for their helpful comments on earlier drafts. Thanks go as well to two anonymous reviewers.


Primary Sources

  1. AKBh Pradhan, P. (Ed.), (1975). Abhidharma-kośabhāṣya of Vasubandhu, revised by A. Haldar. Patna: K. P. Jayaswal Research Institute.Google Scholar
  2. AKVy Shastri, Swami Dwarikadas (Ed.), (1970–1973). Abhidharmakośa & Bhāṣya of Acharya Vasubandhu with Sphuṭārthā Commentary of Ācārya Yaśomitra. Varanasi: Bauddha Bharati.Google Scholar
  3. AKBhT Zhu-chen Tshul-khrims-rin-chen (Ed.) (1985). sDe-dge Bstan-’gyur. Vols. 140 & 141. #4090. Delhi: Delhi Karmapae Choedhey, Gyalwae Sungrab Partun Khang. Accessed through and

Secondary Sources

  1. Adam, M. T. (2005). Groundwork for a metaphysic of Buddhist morals: A new analysis of puñña and kusala, in light of sukka. Journal of Buddhist Ethics, 12, 62–85.Google Scholar
  2. Alt, W. (1980). There is no paradox of desire in Buddhism. Philosophy East and West, 30(4), 521–528.CrossRefGoogle Scholar
  3. Bryant, E. F. (2014). Agency in Sāṃkhya and Yoga: The unchangeability of the eternal. In M. R. Dasti & E. F. Bryant (Eds.), Free will, agency and selfhood in Indian philosophy (pp. 16–40). Oxford: Oxford University Press.Google Scholar
  4. Cardona, G. (2014). Pāṇinian grammarians on agency and independence. In M. R. Dasti & E. F. Bryant (Eds.), Free will, agency and selfhood in Indian philosophy (pp. 85–111). Oxford: Oxford University Press.Google Scholar
  5. Chadha, M. (2017). No-self and the phenomenology of agency. Phenomenology and Cognitive Sciences, 16(2), 187–205.CrossRefGoogle Scholar
  6. Clayton, B. (2001). Compassion as a matter of fact: The argument from no-self to selflessness in Śāntideva’s Śikṣāsamuccaya. Contemporary Buddhism, 2(1), 83–97.CrossRefGoogle Scholar
  7. Dasti, M. R. (2014a). Nyāya’s self as agent and knower. In M. R. Dasti & E. F. Bryant (Eds.), Free will, agency and selfhood in Indian philosophy (pp. 112–136). Oxford: Oxford University Press.Google Scholar
  8. Dasti, M. R. (2014b). Introduction. In M. R. Dasti & E. F. Bryant (Eds.), Free will, agency and selfhood in Indian philosophy (pp. 1–15). Oxford: Oxford University Press.Google Scholar
  9. Dhammajoti, Bhikkhu K. L. (2007) Sarvāstivāda Abhidharma, Third Revised Edition. Hong Kong: Centre of Buddhist Studies, The University of Honk Kong.Google Scholar
  10. Duerlinger, J. (2003). Indian Buddhist theories of persons: Vasubandhu’s “refutation of the theory of a self”. London; New York: Routledge Curzon. Google Scholar
  11. Ganeri, J. (2007). The concealed art of the soul: Theories of self and practices of truth in Indian ethics and epistemology. Oxford: Oxford University Press.CrossRefGoogle Scholar
  12. Garfield, J. L. (2012). What is it like to be a bodhisattva? Moral phenomenology in Śāntideva’s Bodhicaryāvatāra. Journal of the International Association of Buddhist Studies, 33(1–2), 333–357.Google Scholar
  13. Garfield, J. L., Jenkins, S., & Priest, G. (2016). The Śāntideva passage: Bodhicāryāvatāra VIII.90–103. In The Cowherds (Ed.), Moonpaths: Ethics and emptiness (pp. 55–76). Oxford: Oxford University Press.Google Scholar
  14. Gethin, R. (2004). He who sees Dhamma sees Dhammas: Dhamma in early Buddhism. Journal of Indian Philosophy, 32(5–6), 513–542.CrossRefGoogle Scholar
  15. Gold, J. (2015). Paving the great way: Vasubandhu’s unifying Buddhist philosophy. New York: Columbia University Press.Google Scholar
  16. Goodman, C. (2002). Resentment and reality: Buddhism on moral responsibility. American Philosophical Quarterly, 39(4), 359–372.Google Scholar
  17. Goodman, C. (2009a). Consequences of compassion. New York: Oxford University Press.CrossRefGoogle Scholar
  18. Goodman, C. (2009b). Vasubandhu’s Abhidharmakośa: The critique of the soul. In W. Edelglass & J. L. Garfield (Eds.), Buddhist philosophy: Essential readings (pp. 297–308). Oxford: Oxford University Press.Google Scholar
  19. Gross, R. M. (1997). Toward a Buddhist environmental ethic. Journal of the American Academy of Religion, 65(2), 333–353.CrossRefGoogle Scholar
  20. Hamilton, S. (1996). Identity and experience: The constitution of the human being according to early Buddhism. London: Luzac Oriental.Google Scholar
  21. Harris, S. (2011). Does Anātman rationally entail altruism? On Bodhicaryāvatāra 8: 101-103. Journal of Buddhist Ethics, 18, 93–123.Google Scholar
  22. Harvey, P. (2000). An introduction to Buddhist ethics: Foundations, values and issues. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
  23. Herman, A. L. (1979). A solution to the paradox of desire in Buddhism. Philosophy East and West, 29(1), 91–94.CrossRefGoogle Scholar
  24. Ives, C. (1992). Zen awakening and society. Honolulu: University of Hawaiʻi Press.CrossRefGoogle Scholar
  25. Jenkins, S. (2016). Walking into compassion: The three Ālambana of Karuṇā. In The Cowherds (Ed.), Moonpaths: Ethics and emptiness (pp. 97–118). Oxford: Oxford University Press.Google Scholar
  26. Kapstein, M. T. (2001). Reason’s traces: Identity and interpretation in Indian and Tibetan Buddhist thought. Studies in Indian and Tibetan Buddhism. Boston: Wisdom Publications.Google Scholar
  27. Kellner, B., & Taber, J. (2014). Studies in Yogācāra-Vijñānavāda idealism I: The interpretation of Vasubandhu’s Viṃśikā. Asiatische Studien/Études Asiatiques, 68(3), 709–756.Google Scholar
  28. Keown, D. (1992). The nature of Buddhist ethics. New York: St. Martin’s Press.CrossRefGoogle Scholar
  29. King, S. B. (2005). Being benevolence: The social ethics of engaged Buddhism. Honolulu: University of Hawaiʻi Press.Google Scholar
  30. La Vallée Poussin, L. de, & Pruden, L. M. (Trans.) (1988–1990). Abhidharmakośabhāṣyam of Vasubandhu. Berkeley: Asian Humanities Press.Google Scholar
  31. La Vallée Poussin, L. de, & Sangpo, Gelong L. (Trans.), (2012). Abhidharmakośa-Bhāṣya of Vasubandhu: The Treasury of the Abhidharma and its (Auto) Commentary. Delhi: Motilal Banarsidass.Google Scholar
  32. Lee, J. C. (Ed.). (2005). Abhidharmakośabhāṣya of Vasubandhu, chapter IX: Ātmavādapratiṣedha. Bibliotheca Indologica Buddhologica 11. Tokyo: Sankibo Press.Google Scholar
  33. Lowe, E. J. (2008). Personal agency: The metaphysics of mind and action. Oxford: Oxford University Press.CrossRefGoogle Scholar
  34. McGarrity, L. (2015). Mādhyamikas on the moral benefits of a self: Buddhists ethics and personhood. Philosophy East and West, 65(4), 1082–1118.CrossRefGoogle Scholar
  35. Mele, A. R. (2003). Motivation and agency. Oxford: Oxford University Press.CrossRefGoogle Scholar
  36. Meyers, K. (2010). Freedom and self-control: Free will in South Asian Buddhism. PhD dissertation: University of Chicago.Google Scholar
  37. Meyers, K. (2014). Free persons, empty selves: Freedom and agency in light of the two truths. In M. R. Dasti & E. F. Bryant (Eds.), Free will, agency and selfhood in Indian philosophy (pp. 41–67). Oxford: Oxford University Press.Google Scholar
  38. Ronkin, N. (2005). Early Buddhist metaphysics: The making of a philosophical tradition. London: Routledge Curzon.CrossRefGoogle Scholar
  39. Salvini, M. (2008). Conventions and agency in the philosophies of the Mahāyāna. PhD dissertation, School of Oriental and African Studies.Google Scholar
  40. Schechtman, M. (1996). The Constitution of Selves. Ithaca, NY: Cornell University Press.Google Scholar
  41. Siderits, M. (2000). The reality of altruism: Reconstructing Śāntideva. Philosophy East and West, 50(3), 412–424.Google Scholar
  42. Siderits, M. (2016). Does ‘Buddhist ethics’ exist? In The Cowherds (Ed.), Moonpaths: Ethics and emptiness (pp. 119–140). Oxford: Oxford University Press.Google Scholar
  43. Siderits, M. (2017). Buddhist Reductionism and the structure of Buddhist ethics. In P. Bilimoria, J. Prabhu, & R. Sharma (Eds.), Indian Ethics: Classical Traditions and Contemporary Challenges, Volume I (pp. 283–296). New York: Routledge.Google Scholar
  44. Steward, H. (2013). Processes, continuants and individuals. Mind, 122(487), 781–812.CrossRefGoogle Scholar
  45. Taber, J. (2011). Did Dharmakīrti think the Buddha had desires? In H. Krasser, H. Lasic, E. Franco, & B. Kellner (Eds.), Religion and logic in Buddhist philosophical analysis: Proceedings of the fourth international Dharmakīrti conference, Vienna, august 23–27, 2005 (pp. 437–448). Vienna: Verlag der Österreichischen Akademie der Wissenschaften.Google Scholar
  46. Visvader, J. (1978). The use of paradox in Uroboric philosophies. Philosophy East and West, 28(4), 455–467.CrossRefGoogle Scholar
  47. Wetlesen, J. (2002). Did Śāntideva destroy the bodhisattva path? Journal of Buddhist Ethics, 9, 34–88.Google Scholar
  48. Williams, P. M. (1981). On the Abhidharma ontology. Journal of Indian Philosophy, 9(3), 227–257.CrossRefGoogle Scholar
  49. Williams, P. (1998). Altruism and reality: Studies in the philosophy of the Bodhicaryāvatāra. Richmond: Curzon Press.Google Scholar

Copyright information

© Springer Science+Business Media B.V., part of Springer Nature 2018

Authors and Affiliations

  1. 1.Group in Buddhist StudiesUniversity of CaliforniaBerkeleyUSA

Personalised recommendations