1. The Introduction of Persons into Buddhist Philosophy

The no-self view is a fundamental precept of the Buddha; accordingly, all schools and philosophers within the Buddhist tradition defend the no-self doctrine. But these various schools and philosophers disagree about the Buddha's stance toward persons. There is no doubt that the Buddha did talk about persons. The Bhārahārasūtra (the sūtra on the bearer of the burden) is one of the scriptures most frequently alluded to in this connection. In translation, the original Sanskrit reconstruction reads as follows:

The sūtra has been interpreted as a variation on the topic of the four noble truths (Eltschinger Reference Eltschinger2014). The definitions provided for the burden, the taking up of the burden and the laying down of the burden coincide with those of the truths of suffering, of the origin of suffering, and of the cessation of suffering. The originality of the sūtra lies in introducing the notion of a burden-bearer (bhārahāra) defined as a ‘person’ (pudgala). In the text, the term ‘burden’ is interpreted as desire or craving that is the root cause of morally wrong action, which in turn is the cause of rebirth and suffering according to the Buddhist doctrine of karma. The person is introduced in the original sūtra as the bearer of moral desert.

The controversy over persons raged among Indian Buddhists for more than a millennium. The question is whether the ‘person’ in the sūtra is to be interpreted as ultimately real, as Pudgalavādins (Personalists) did, or, as only conventionally real, as the mainstream Abhidharma Buddhists did. This Buddhist distinction between two kinds of truth or reality is sometimes motivated by the ‘two truths’ doctrine in the Samādhirāja-sūtra. However, recent studies suggest that the two truths distinction is an innovation by the Ābhidharmikas that came into prominence as a heuristic device useful for later interpreters to reconcile apparent inconsistencies in the Buddha's teachings (Karunadasa Reference Karunadasa1996: 25–26 and 25n139; The Cowherds 2011; 5). In the Abhidharmakośa-bhāṣya, Vasubandhu characterizes the distinction as ‘That of which one does not have a cognition when it has been broken is real in a concealing way (samvṛti-sat); an example is a pot. And that of which one does not have a cognition when other [elemental qualities (dharma)] have been excluded from it by the mind is also conventionally real. That which is otherwise is ultimately real (paramārtha-sat)’ (Abhidharmakośa-bhāṣya 6.4, translated by Ganeri [Reference Ganeri2012: 67]).

According to the Abhidharma doctrine, the only things that are ultimately real are the indivisible, momentary physical and mental dharmas (best understood as tropes; see Goodman Reference Goodman2004). Everything else that can be decomposed into parts, physically or conceptually, is only conventionally real. Siderits explains the Ābhidharmika position thus: ‘A statement is said to be ultimately true iff it corresponds to how things are independently of the concepts one happens to employ. Such a statement can neither assert nor presuppose the existence of any composite entity. . . . Many entities in people's folk ontology are not ultimately real: chariots, forests, trees, pots, and so on. Such entities are said to be conventionally real, mere conceptual fictions’ (Reference Siderits and LoLordo2019: 314).

Different schools in the Indian Buddhist tradition have different strategies for making sense of the Buddha's thought and talk of persons. The Pudgalavādins are alone in holding that persons are ultimately or substantially real. The dispute between the Pudgalavādins and their opponents is not just an attempt to lay claim to ‘what the Buddha taught’, but it is also motivated by philosophical considerations about consistency with the no-self doctrine and other important Buddhist theses, for example impermanence, dependent origination, and the karma theory. Both Abhidharma and Madhyamaka traditions hold that persons exist only conventionally. There are differences, though. According to the Madhyamakas, the pragmatic usefulness of person-talk is important as a catechetic device for teaching the doctrine to the uninitiated. The Buddha's talk about persons and other continuing entities in the sūtras is aimed at ordinary people who, due to ignorance are beset with the false view of a self. As Vincent Eltschinger puts it, ‘these preliminary and merely provisional teachings are meant to offer a transition between the adhesion to worldly beliefs and the intuition of universal emptiness’ (Reference Eltschinger2014: 470). The Ābhidharmikas, on the other hand, did not think of ‘person’ as a useful pragmatic device or skillful means but emphasized that it is nothing but a conventional designation, a mere name for a group of psychophysical aggregates, ultimately the collections of dharmas. As I discuss below in section 3, according to Vasubandhu, persons are beset with the same problems as selves, and therefore they are not useful at all.

2. Abhidharma Philosophers on the Status of Persons

In the final chapter of the Abhidharmakośa-bhāṣya, Vasubandhu, offers a defense of the Abhidharma doctrine of no-self and no-persons. Vasubandhu's argument against the Pudgalavādins who deny the existence of the self but accept the reality of a so-called person (pudgala) is based on the causal efficacy principle: everything that is real or substantial (dravya) is causally efficient, having specifiable cause-and-effect relations with other entities (see, for example, Gold Reference Gold2015). Everything else is a conceptual construct, a mere convention (prajñapti), and thus should be rejected as unreal. Vasubandhu raises a dilemma for the Personalists: Does such a person exist ‘substantially’ (dravyataḥ) or ‘nominally’ (prajñaptitaḥ)? If the first, then the person would have to be different from the five aggregates after all, for it would have its own distinct nature. The Pudgalavādins deny this, because for them the nature of the person depends on the aggregates. If the latter, then the Pudgalavādin must agree with Vasubandhu: the person has mere nominal existence (Abhidharmakośa-bhāṣya 461, 14–18. Unless otherwise noted, all citations to this work are to Pruden's four-volume translation [1988–1990]). The Pudgalavādin response is to deny the proposed dichotomous division of reality. For them, the person exists substantially although it depends on the aggregates. Vasubandhu then asks the opponent to clarify the sense of ‘depending on’. If, by this, the Pudgalavādin means ‘dependent on the aggregates as a conceptual object’, then, argues Vasubandhu, they are conceding that persons are mere conceptual constructions. It is like the taste, touch, and smell of milk generating the conceptual construction ‘milk’. If by ‘depending on’ they mean that the person is causally dependent upon the aggregates, then, Vasubandhu argues, they are saying that the person is caused by the aggregates. But if the aggregates are the causes of the person, then the person, too, must change as the aggregates change. The Personalists do not want to admit this because they think that a person persists across the changing aggregates. So, Vasubandhu restates his complaint as the demand for the cause of persons. The Personalists have no satisfactory answer. This argument may be reconstructed thus:

• P1: The person is presupposed to be a persisting being that exists ‘depending on’ the psycho-physical aggregates (Pudgalavādin premise)

• P2: Everything that is real or substantial (dravya) is causally efficacious, having specifiable cause-and-effect relations with other entities (Abhidharmakośa-bhāṣya 461, 19–20; Gold Reference Gold2015).

• P3: The person (as presupposed by the Pudgalavādin) is causally inefficacious since it does not have specifiable causes and effects (Abhidharmakośa-bhāṣya, 461–472, 14).

• C: Thus, there is no person (as presupposed by the Pudgalavādin).

Vasubandhu's Pudgalavādin opponents object to P3 by claiming that there are many scriptural texts that imply the existence of persons. More importantly, they argue that we need to postulate persons to causally explain moral responsibility, rebirth, and so forth: If there is no person, how can we make sense of the talk of ‘bearer of the burden’ (Abhidharmakośa-bhāṣya 468, 1–9)? The questions continue. If there is no person, how can we make sense of claims that assert transmigration (Abhidharmakośa-bhāṣya 471, 24–472, 3)? Again, what are we to make of Buddha's claims that he remembered being this or that person in a previous existence (Abhidharmakośa-bhāṣya 472, 3–7)? It is important to keep in mind that each of these points—responsibility, rebirth, memory—are exactly things that the self is posited to explain in Indian philosophy. The issue between the Personalists and Vasubandhu is whether there exists a permanent thing that can be the agent of action and bearer of responsibility. The Personalist is using the notion of person to salvage the notions of agency and responsibility that are central to the Buddhist doctrine of karma, which do not seem to make sense in the absence of selves. Vasubandhu intends to show that there is no need to posit persisting persons and that we can make sense of agency and responsibility in terms of future aggregates suitably causally related to some previous aggregates. He argues that none of the statements of the canonical Buddhist texts (about responsibility, rebirth, memory, soteriology, salvation and the nature of the liberated saint) is to be taken as referring to, or implying that there are, persons. Each of these can and should be interpreted so that they refer only to psychophysical aggregates that in turn are ultimately composed of indivisible dharmas. In the discussion with Vasubandhu, the hypothetical Pudgalavādin opponent, quotes the sūtra of the burden-bearer and asks if the aggregates were the person, how we make sense of the Buddha's talk of the burden bearer:

Vasubandhu's strategy is to translate the talk of burdens and burden-bearers into a linear, causal series: it is because the former collection of psychophysical aggregates (the burden is the cause) torment the next ones (the burden-bearers are their effects) that they are called burden-bearers. The rebuttal of the Pudgalavādin's objection is to say that the person (the burden-bearer) is not distinct from the psychophysical aggregates (the burden) just as craving (the taking up of the burden) is not distinct from the psychological aggregates. After offering similar explanations for karma, moral responsibility, rebirth, and transmigration, Vasubandhu concludes that none of these phenomena offers any reason to postulate persons. The point of these explanations is not just that there is a better alternative explanation of phenomena, like moral responsibility, in terms of linear causal series. Rather, the point is that this explanation shows there is no need to postulate a person to explain these phenomena as long as we can explain these in terms of a linear causal series of evanescent aggregates. A set of aggregates at a later time will suffer depending on the nature and the quality of action of the set of aggregates that exist now to which they are causally connected.

3: Conventional Reality of Persons

What does Vasubandhu mean by saying that persons are only conventionally real? In the text Vasubandhu illustrates this by a series of examples: ‘The designation pudgala occurs by taking elements as an object: this is to recognize that pudgala is a word designating the skandhas, the same as the designation “milk” occurs taking as its object the constituents of milk—colour etc’ (Abhidharmakośa-bhāṣya 461, 21–463, 9).

The term person is just a collective term for a group of aggregates, ultimately collection of physical and mental dharmas, just as milk is a term for a collection of tropes—such as whitishness, liquidness, and potability. For Vasubandhu, only the dharmas or the tropes exist ultimately or substantially. All other things exist in name only. Siderits explains the idea thus: ‘The point about “chariot” and “person” is that they are opaque enumerative expressions: when taken at face value they seem to denote individual entities, only further analysis shows them to be ways of referring to a plurality of entities in a certain arrangement’ (Reference Siderits and LoLordo2019: 313). So, persons, chariots and pots are conceptual fictions. Siderits adds, ‘The idea here is that enumerative term “pot” represents a concept that has proven useful for creatures with certain interests (in this case, for storage) and certain cognitive limitations (such as the inability to track all the many parts). . . . Now conventional truths do typically guide one to successful practice. This would be difficult to explain if the truth-makers for conventionally true statements consisted of nothing but mere conceptual fictions. Thus conventionally real entities are said to supervene on ultimately real entities’ (Reference Siderits and LoLordo2019: 314–15).

The quotation above is Siderits's own gloss on the Abhidharma and, more generally, the Buddhist notion of conventional reality. Conventionally real entities are useful fictions and they supervene on ultimate reality. Siderits is, however, mistaken on both counts. Pots and chariots might seem to be useful fictions because they serve our needs, but I do not think that all conventionally real entities are useful, period. Persons might initially seem to be an attractive addition to the list of conventionally real entities because they give us something, more or less persisting, to harbor practical concerns, for example, responsibility and desert, compensation, social treatment, and emotions. But we need to be more careful here. Ābhidharmikas are concerned to offer a revisionary metaphysics that will force us to rethink our notions of responsibility, desert, and the like rather than just restate them in terms of persons rather than selves. However, note that persons can successfully reinstate the sharp distinctions between me and you, mine and yours, which are likely to produce more suffering than benefit. One of the major consequences of the Buddhist no-self doctrine is to break the barriers between individuals so that these selfish tendencies associated with the I-sense do not get a hold. Persons are not natural kinds (Wiggins Reference Wiggins1976). At most, they are social artifacts. Persons depend in part for their existence on certain kinds of social practice in contrast to some other conventionally real things, like water. Whether someone is a person or not is a matter of legislation, social practice, or convention—or a combination of these. Insofar as persons are conventional realities, they do not locally supervene on ultimately real entities as Siderits would have it.

Abhidharma Buddhists are clear that selves are not ultimately real, but what about the conventional existence of selves? Classical Buddhist philosophers readily admit that selves, too, exist conventionally. Vasubandhu admits that selves are conventionally real. In a discussion often missed in the contemporary literature, in chapter 3 of the Abhidharmakośa-bhāṣya, Vasubandhu opens the discussion of dependent origination (pratītya-samutpāda) by raising an objection from the Hindu opponents who believe in self (ātman):

But Vasubandhu's reply that ātman is a mere name will not satisfy the opponent because the Hindus think of the self (ātman) as a simple, eternal substantially real entity. The Hindu Naiyāykas argue that such a self must be posited in order to causally explain such phenomena as memory, agency, and responsibility. Vasubandhu's argument against a simple eternal self is that such a self is explanatorily redundant; it does not play any role in the causal explanation of memory, agency, and the like (Abhidharmakośa-bhāṣya, 472–78). (I develop Vasubandhu's response elsewhere; for detailed reconstruction and analysis of his argument for the no-self doctrine, see Chadha Reference Chadha2019, Reference Chadha2018, Reference Chadha2017.) Since there is no proof for such a self, the only way we can make sense of the term self is to say that it is a name given to the aggregates. Selves have only a nominal existence. It is in this sense that we can say, on behalf of Vasubandhu, that persons are just like selves. So it seems that for these classical Buddhists selves qua aggregates have as much reality as persons. Surely, these Buddhists would not want to say that selves qua aggregates are useful fictions. The whole of the Buddhist doctrine is built around the denial of self and the extirpation of the sense of self, the root cause of suffering.

Vasubandhu is a revisionist philosopher, not just concerned with describing what we instinctively assume, but also challenging that what we assume. Our ordinary talk and thought misleads us into thinking that persons (and selves) really exist. And, insofar as persons (and selves) are persisting entities we are prone to believing that we continue to exist in the future. This is the basis of the special concern that we have for ourselves. But Vasubandhu shown us the need to be careful here. Persons exist only as conceptual constructions; they are only conventionally, not ultimately, real. There is no future person or self that will continue to exist, and thus our special concern for ourselves is not grounded in anything real. We can, if we like, continue to talk in this way, but must remain conscious of the fact that this is nothing more than a matter of convenience. In what follows, I argue that Buddhists, and philosophers more generally, should be suspicious of persons if they are suspicious of selves. I address the Buddhists first.

The standard Buddhist arguments for no-self derive from the fact of impermanence or momentariness (see, for example, Siderits Reference Siderits and Gallagher2011). But the notion of person seems to be caught up precisely in persistence beyond the momentary. Persons as postulated continue to exist during a lifetime and possibly across lifetimes; they are bearers of karma and so persisting loci of agency and responsibility. Thus, the notion of person seems to share the very features that lead Buddhists to view selves with suspicion. Persons are just as ethically problematic as selves because they can be equally responsible for our sense of being the same thing within and across lifetimes. The central Buddhist teaching is to extirpate the false sense of I and in that endeavor the conventional posit of person cannot be a useful fiction; it will be an obstacle.

It is also useful to consider this issue in the context of Buddhist soteriology, which aims at escape from existential suffering. The core Buddhist idea is that we suffer because we mistake there to be a persisting being that is the referent of I and is the sole locus of meaning and value. The Buddhist insight is that there is no such persisting thing: everything that exists is transitory. If there is no persisting being, there is no locus of value and meaning, irrespective of whether we think it is a simple self (that exists independently of the aggregates) or a persisting, though composite, person (that exists dependently of the aggregates). This revisionary Abhidharma metaphysics also grounds an argument for compassion. Ultimately, we are collections of physical and mental dharmas causally related to future and past dharmas within a series and with other series of dharmas. This way of looking at the world undermines the basis me and mine and others that are distinct from me. Thus, there is no basis for discriminating between my suffering and that of others. The aim of the Buddhist teaching is to reduce all suffering, impersonally and selflessly. The self is to be rejected as it is likely to distort this vision, create artificial distinctions between me and others, and increase suffering. It seems that the person will do the same and should therefore be rejected for the same reasons: persons are fictions to be sure, but for the Buddhist they are dangerous fictions.

I have argued that in critical respects, selves and persons are at par for Buddhists. Both selves and persons are conceptual constructs and so ultimately not real. In addition, I have argued that, at least for Abhidharma Buddhists, selves and persons seem to be equally problematic from a soteriological perspective: they are a result of flawed ways of thinking about the world and these ways of thinking create suffering. So, I think, contra Siderits, eliminativism about persons and selves both is the right attitude for the Buddhists. For the Buddhist, persons, like selves, should be regarded as dangerous fictions. If selves are to be extirpated from our metaphysics, so should persons be.

4. Persons in Contemporary Philosophy

If the Buddhist, as argued above, ought to be eliminativist about persons, what about the rest of us? Since most of us do not have the soteriological and metaphysical commitments of the Buddhist, we continue to be enthusiastic about persons. Like it or not, I argue, the source of this enthusiasm is our implicit commitment to descriptive metaphysics. We assume that the cost of wholesale revision of our person-related practices, concerns, and attitudes is too high.

Parfit believes that eliminativism about persons is defensible. Eliminativism results from the combination of reductionism about persons with what he calls the ‘extreme claim’, which says that one's moral and prudential concerns cannot be grounded without a deep separate fact about personal identity. Parfit thinks that another view is also defensible. This view results from the combination of reductionism with the ‘moderate claim’, which says that our moral and prudential concerns may well be grounded in what does matter in personal identity. This, according to Parfit, is psychological connectedness and/or continuity (Reference Parfit1984: 311). Adopting reductionism combined with the moderate claim, however, requires a major reconfiguration of how we think about the immorality of great imprudence, desert, and commitments. Parfit's ‘identity doesn't matter’ view, I argue below, puts us on the path to eliminativism. More recently, David Shoemaker has developed what he calls the ‘identity really really doesn't matter’ view (Reference Shoemaker, Blatti and Snowdon2016: 325). Adopting Shoemaker's view requires us to endorse a wide-ranging pluralism about what matters in personal identity and theories of the relation between personal identity and our person-related concerns and attitudes. Shoemaker's view takes us closer—but not quite all the way—to endorsing eliminativism about persons. Why is that? Because of their continued allegiance to descriptive metaphysics.

To show why, I first return to Parfit. In the concluding chapter of Reasons and Persons, Parfit summarizes his position thus: ‘On this Reductionist View, persons do exist. But they exist only in the way nations exist. Persons, are not, as we mistakenly believe, fundamental. This view is in this sense more impersonal’ (Reference Parfit1984: 445). Parfit's critique is aimed at Strawson's conception that persons are basic particulars: ‘a type of entity such that both predicates ascribing states of consciousness and predicates ascribing corporeal characteristics . . . are equally applicable to an individual entity of that type’ (Reference Strawson1959: 101–2). Parfit's critique of the entity-hood of persons, what he calls ‘non-reductionism’, is well known. Strawson's concept of person is that of an entity thought of as embodied subject of experience. This descriptive concept is much thinner than the normative Kantian concept of persons, according to which person refers to entities possessing moral capacities and moral status (because of having those capacities). Parfit's no-entity view strips persons of their fundamental status not only in descriptive metaphysics but also in domains of ethics and morality. Parfit says that while this does not totally undermine our prudential and ethical concerns, they need to be reconfigured in light of the reductionist view (to which I return later). What does Parfit mean by saying that ‘persons do exist’? He writes, ‘We do not deny that people exist. And we agree that we are not series of events—that we are not thoughts and actions, but thinkers and agents. But this is true only because we describe our lives by ascribing thoughts and actions to people. As I have argued, we could give a complete description of our lives that was impersonal: that did not claim that persons exist’ (Parfit Reference Parfit1984: 341, my emphasis).

The talk of persons, according to Parfit, is a matter of convention. Just as Vasubandhu says. Persons are nothing more than an artifact because of the way we think and talk. There are no real persons. It seems to me that Parfit comes very close to endorsing eliminativism about persons. We can continue to talk talking about persons in our ordinary everyday language, but given that we need to change drastically our views about the nature of persons and personal identity over time, we should note that persons are only nominal existences. As Vasubandhu would put it, person is only a name given to a collection of aggregates.

To return to the way we need to reconfigure our claims about morality and rationality in the light of Parfit's reductionism. Adopting this view about persons, Parfit suggests, will require a major reconfiguration of how we think about the immorality of great imprudence, desert, and commitments. Since psychological connectedness and continuity are a matter of degree, the strength of these connections will determine the degree to which one should hold people responsible for acting imprudently, for their past crimes and the promises they make for the future. In a sense this undercuts Siderits's reasons for why Buddhists should not eliminate persons. Siderits writes,

Siderits appeals to the practical considerations to advance the cause of Buddhist reductionism about persons. Moral responsibility no doubt is an important practical concern, but Parfit has shown us that that any notion of responsibility that presupposes a false notion of personhood is ungrounded. The degree of responsibility should be proportional to the degree of psychological connectedness. If the connections between the criminal now and individual at the time of the crime are weaker, he deserves less punishment. So, too, with our commitments and promises, and our self-concern about our own future and finally the fear of death. The more weakly connected our future time slices are to our current time slice, the more reason there is to reduce our overwhelming obsession with the well-being and longevity of our future time slice. We have no reason to be especially concerned about our own futures, but we have reason not to do to our future time slices what would be wrong to do to other people. Buddhists and Parfit would agree that one important consequence of the no deep further fact view is that it succeeds in breaking barriers between people. Is this a change for the better?

So far so good. Parfit helps to show that “identity doesn't matter” for some practical considerations, such as moral responsibility. What matters instead is psychological continuity and connectedness. Shoemaker argues that psychological continuity is not what matters when it comes to other practical concerns, for example, prudential concern and compensation. He illustrates the point with the following example:

Similarly, consider Alistair, who is suffering from Parkinson's disease and who has made an Advance Care Directive refusing life-sustaining treatment, including antibiotics. Alistair's Parkinson's disease has deteriorated significantly over the years and he is now suffering from end-stage dementia. When Alistair contracts a life-threatening community-transmitted virus, his health care team, in accordance with his wish of refusal of life-sustaining treatment, does not call an ambulance or administer antibiotics. His prudential concern in signing the directive is clearly for his future time slice, but now there is no one who is psychologically continuous with him. Biological continuity alone seems to be what matters here; psychological continuity is irrelevant.

Shoemaker offers another example: ‘Consider someone who, due to some traumatic brain injury, undergoes radical psychological discontinuity. She will still be treated as the owner of the pre-transformation-person's car and other property, and she will also be treated as the spouse of the pre-transformation-person's spouse, the daughter of her parent, and so forth’ (Reference Shoemaker, Blatti and Snowdon2016: 316). It is possible, however, as Hannah Tierney et al. (Reference Tierney, Howard, Kumar, Kvaran, Nichols and Sytsma2014) note, that both social treatment and compensation can also track psychological continuity. They offer a few thought experiments to make the point. Consider this variant on their example: suppose we were able to transfer an elderly adult's entire psychological profile from her body to a new body, and the old body, which is in any case deteriorating because of aging, is destroyed. Her children would presumably treat this ‘new’ individual as their own mother despite the lack of biological continuity. And, if an individual had caused some psychic trauma to the elderly adult before the transfer, but rectified the trauma after the transfer occurred, the elderly adult would still surely be compensated, again, despite the lack of biological continuity. In this way, Tierney et al. conclude that ‘it looks as though these identity-related practical concerns fail to track a singular, monistic criterion of identity, but rather follow two distinct criteria in different contexts’ (Reference Tierney, Howard, Kumar, Kvaran, Nichols and Sytsma2014: 200).

Indeed, if we consider the variety of practical concerns—anticipation, first-person recognition and concern, third-person recognition and concern, general social treatment, emotional patterns (such as pride and shame), compensation, and responsibility—we realize that our practical concerns do not consist of a monolithic set. There are different types of practical concerns, and while some are clearly grounded on psychological relations, some are actually grounded on others, including animalistic and humanistic relations (Shoemaker Reference Shoemaker, Blatti and Snowdon2016: 304). What this suggests, then, is that, identity matters not only for our practical concerns but also the relations that Parfit thought mattered are irrelevant to some of our practical concerns (Shoemaker Reference Shoemaker, Blatti and Snowdon2016: 325).

The upshot of this discussion for our purposes is that Strawsonian persons considered as one two-sided thing not as two conjoined one-sided things—mind and body—are no longer available as one single thing to ground all of our person-related concerns. Parfitian reasons lead us to question the identity of Strawsonian persons, Shoemakerian reasons lead us to question the supposed unity of Parfitian persons. We need persons to ground our person-related concerns and practices, but what we find are disunified psychological and physical states playing the person role. What is important is that different clusters of mental and/or physical states play the person role in response to the various practical concerns. These different clusters need not be unified to form a whole, a person, at a time, nor over time.

Revising a pre-philosophical Strawsonian concept of person to be more like the Parfitian unified bundle of psychological states looks unstable, since several of our practical concerns turn out to put pressure on the unity of the bundle. No revision of the ‘person’, then, can capture all of our practical concerns. And as Nichols and Shoemaker (Reference Nichols and ShoemakerMS) caution, ‘don't forget the biggest practical concern of them all: fear of death (or the hoped-for anticipation of survival of death)’. They write, ‘The practical concerns that often justify preservation for other concepts have much more difficulty doing so in the case of the self. And recall as well that there could be a significant practical advantage to elimination of the “self,” namely, the therapeutic effect articulated by Buddhism’. Though Nichols and Shoemaker (Reference Nichols and ShoemakerMS) are concerned with the self and want to draw a totally different conclusion from this, I believe that they would not have trouble agreeing that self can be replaced with person in this quotation. The conclusion they draw is a provocative suggestion: what they call a radical pluralism, according to which, we can choose to be pluralists rather than preservationists or eliminativists about the self depending on the context. I do not agree with that suggestion, but rather than argue against it here, I press on with the argument for the claim that not only Buddhists but all of us have good reason to settle for eliminativism about selves and persons.

Other practical concerns, as noted above, depend on what Shoemaker calls ‘animalistic and humanistic relations’ (Reference Shoemaker, Blatti and Snowdon2016: 304). Psychological continuity is (at least sometimes) not what is at issue when we are thinking of compensation or social treatment; a different criterion, for example bodily continuity or biological continuity, may work in these cases. Following Shoemaker, I add to this list: third-person reidentification and its associated sentiments (for example, why is my happiness at seeing my parents after a long time appropriate?); first-person reidentification (for example, why is it appropriate that when I look at certain photos on my mother's coffee table I feel nostalgic?) (Shoemaker, Reference Shoemaker2007: 318). What explains these feelings?

Some might suggest, as do Tierney et al., that we do not need to give up on personal identity. Rather ‘empirical evidence and philosophical thought experiments indicate that judgments about personal identity are regimented by two different criteria, one in terms of psychological traits and one that largely conforms to biological criteria’ (Tierney et al. Reference Tierney, Howard, Kumar, Kvaran, Nichols and Sytsma2014: 198). We may go further with Shoemaker and settle for a wide-ranging pluralism in the face of the disunity of our practical concerns. But Shoemaker stops short of giving up on persons and person-related concerns in the hope that there is a theory (or theories!) of the relation between personal identity and our person-related practices and concerns. He says, ‘several concessions may be required, including admission, perhaps, of (a) the irrelevance of certain powerful and popular criteria of personal identity for (at least some of) our practices and concerns, (b) the ultimate disunity of these practices and concerns (such that multiple types of theories of the relation between them and the metaphysics may be called for), and/or (c) the possibility of different types of rational grounding—justification and rendering-possible—where justification may actually be off the table altogether for some practices and concerns’ (Reference Shoemaker2007: 354).

Pluralism about criteria of personal identity, theories of the relation between personal identity, and person-related practices might seem to be revisionist, but they really are not. This is because they do not question our person-related practices, concerns, and attitudes. These are considered sacrosanct; they are not up for revision. Why not? Strawson would balk at such a question. He writes, ‘I shall reply, first, that such a question could seem real only to one who had utterly failed to grasp the purport of . . . the fact of our natural human commitment to ordinary inter-personal attitudes. This commitment is part of the general framework of human life, not something that can come up for review’ (Reference Strawson1974: 14, my emphasis). Strawson is not only claiming that it is hard for us to give up our interpersonal attitudes and concerns, but he also thinks to give up these would be to give up on our humanity. Parfit, in contrast, is not content to be a descriptive philosopher. In the preface to Reasons and Persons he writes, ‘Descriptive philosophy gives us reasons for what we instinctively assume, and explains and justifies the unchanging central core of our beliefs about ourselves, and the world we inhabit. I have great respect for descriptive philosophy. But by temperament, I am a revisionist. . . . I try to challenge what we assume. Philosophers should not only interpret our beliefs; when they are false, they should change them’ (Reference Parfit1984: x).

Parfit has revisionist intentions, but it is this respect for descriptive philosophy that stops him in his tracks, just shy of endorsing eliminativism about persons. His argument for endorsing reductionism combined with the moderate claim gains traction from saving ordinary practices and attitudes like awarding credit or blame to others or feeling resentful towards others who have wronged you in the past. Contemporary philosophers like Shoemaker and Tierney et al. settle for pluralism about criteria of personal identity and pluralism about theories of the relation between personal identity and person-related concerns and practices. I think this is also what hinders contemporary philosophers interested in Buddhism, especially Siderits and Ganeri, from endorsing eliminativism about persons All this effort is in the service of hoping to ground our person-related concerns, practices, and attitudes. But why do we regard these as sacrosanct? Why do we not want to consider revising or even jettisoning some of our ordinary person-related concerns and practices and interpersonal attitudes? We have come so far as to believe that they cannot simply be grounded in persons or what matters in personal identity. What if they cannot be explained or justified because they are ungrounded, period? A thoroughgoing revisionist should consider revisiting some of our person-related concerns and practices and attitudes. That is exactly what the Abhidharma Buddhist urges us to do. Eliminating selves and persons might also result in eliminating some of our practices and revising others. I show this in the next Section.

5. Revisionism about Practices

Abhidharma Buddhists are not content to do descriptive metaphysics in the spirit of Strawson and other contemporary philosophers. Their aim is to propose a better structure, in terms of what it would take to reduce all suffering, impersonally and selflessly. Therefore, Buddhists, I have argued, must reject not only selves but also persons. And, if the consequence of this rejection is that we have to radically revise, or even discard, some of our ordinary everyday person-related practices concerns and attitudes, so be it. Abhidharma Buddhists do not have much respect for descriptive metaphysics and so they are open to reviewing our person-related practices, concerns and interpersonal attitudes in accordance with the normative goal of reducing suffering.

I begin with self-concern and the special concern that we have for those we love dearly, then turn to other interpersonal concerns like responsibility, compensation, and social treatment. The Buddhists do recognize that it is built-in precondition of our form of life that we have self-concern and special concern for our loved ones. That is why the Buddhists do not recommend giving up on this self-concern or special concern for our loved ones. Rather they recommend extending similar concern to others. And they do not think that such an extension comes easy to us given our human nature: it has to be inculcated by extensive meditation practices. That said, this is not to be thought of as the giving up of one's humanity but rather enlarging it. As Parfit thinks, this discovery is liberating and consoling (Reference Parfit1984: 281). The Buddhists agree.

To suggest how these meditation techniques are supposed to work and how we should consider revising our interpersonal attitudes, I focus on Buddhaghosa (fifth century CE), an important philosopher in the Pāli Abhidharma (Abhidhamma) tradition. In what follows, I depend on Maria Heim (Reference Heim and Ganeri2017) for explaining Abhidhamma meditation practices and techniques. Buddhaghosa recommends various analytical techniques for the practice of loving kindness concentration. To begin with, Buddhaghosa says, one should consider the practical advantages of becoming filled with loving kindness, most important of which is that one becomes more tolerant of one's own and others’ shortcomings and gets rid of self-loathing. It is important for beginners to the practice of loving kindness meditation to choose the first object on which to concentrate carefully. This is tricky, and advice from one's teacher is helpful. Buddhaghosa recommends choosing a concrete real particular person, not someone dead. Furthermore, he warns against choosing someone one intensely dislikes or likes too much, a hostile person, or a person to whom one feels a sexual attraction, as the feelings one has toward such persons can be distracting, making it harder to begin. Since the content of loving kindness meditation is ‘may this being be happy’, Buddhaghosa recommends starting with oneself as perhaps the best strategy to allow a smooth entry into the practice. The meditator can then progress in the practice by moving beyond the self to direct loving kindness toward a dearly loved one, then to a neutral person, and finally to an enemy. It is at this last point that the meditator is going to be faced with serious obstructions in the practice. Contemplating one's enemy, the meditator is likely to be overcome by anger and resentment. Buddhaghosa recognizes that these feelings are hard to dislodge, but he does offer practical techniques for overcoming such feelings. One of the recommended techniques is what Heim calls ‘resolution into elements’ (2017: 179). The technique recommends that one breaks down an enemy to a bundle of psychophysical aggregates or, further, into the material components of the body (head hair, body hairs, nails, etc.). Now ask yourself, which part is resented? There is no sensible answer to this question. If a person is seen as nothing more than a heap of constantly changing material and mental dharmas, then the anger and resentment cannot get a foothold.

Other meditations like compassion and sympathetic joy and finally equanimity meditations are to be thought on the same model and are to be undertaken after completing the loving kindness meditation and in the order in which they are listed. Equanimity is different in that it retracts from the happiness and pleasure taken in the forms of love developed in the first three. Pleasure is considered a ‘danger’ because of its association with desires and proximity to aversion and attraction. Equanimity, on the other hand, promotes impartiality that brings with it peace, although it is not suggested that the meditator develop this with apathy or indifference. Heim does mention though that with equanimity ‘we may have reached the edges of what we can mean by “love”. The breaking down of barriers it effects is coming to see all beings—dear and despised alike—as neutrals’ (Heim Reference Heim and Ganeri2017: 182). Equanimity is the ideal to be reached, and I am not suggesting that all human beings are to go that far; rather making progress on the Buddhist path through successive meditations is the way to go.

The point is that though attitudes of resentment and blame are difficult to dislodge because they are deeply ingrained, it is not the case that they cannot come up for review. Strawson would say reviewing attitudes of resentment and blame which are at the core of our practices of treating each other as participants would be to give up our humanity. Perhaps we could tinker with some reactive attitudes around the margins. The Buddhists disagree. They suggest a wholesale, but principled, revision of our practices to jettison all the negative reactive attitudes including those at the core. The principle guiding this revision is the normative goal of reducing suffering. The idea is to break down barriers among persons, to identify with others and think of others as me and mine. After all, it is causal relations that undergird the special concern I have for my future self. But I also have the same kind of causal relations with my loved ones and other contemporary beings. Why not extend the special concern I have for my future self to others? There is no suggestion this is going to be easy, but the idea is to slowly expand the circle of me and mine to include all others. This does not require giving up on all the normal sentiments that we find ourselves with, such as anger, love, resentment, and joy, but only the negative reactive attitudes like resentment, hatred, and anger. The idea is to develop the moral sentiments like empathy, sympathy, and compassion, all of which can be aided by identifying with other persons.

What effect is the Buddhist practice likely to have when we think other person-related concerns like responsibility, compensation, and social treatment? For moral responsibility, the Abhidharma Buddhists would agree in principle with Parfit that psychological continuity is what matters, though the details of their account of moral responsibility are different. Vasubandhu is clear that it is intention or volition associated with the physical act that determines the moral quality of the act. Actions, good and bad, perfume other mental dharmas that belong to what he calls ‘the subtle mind’, which is another evanescent series of mental dharmas. The subtle mind is sort of holding the supposed karmic residues (or seeds) of the action in a dormant state until they ripen, that is, are ready for the results of the action to take effect. The idea is that present mental states determine the character and quality of future mental states depending on the moral quality of the action. The Buddhist account of agency and moral responsibility involves a strategy of stepwise reduction (Repetti: Reference Repetti and Davis2017). First, the discourse of enduring agents is reductively analyzed in terms of impersonal psychophysical aggregates. On this analysis, there are no enduring selves or persons and no centralized locus of agency. Second, the psychophysical system itself is analyzed into momentary mental and physical events, the dharmas. At this level of fundamental ontology agency is explained in terms of the causal connections between the dharmas and between the dharmas and the external environment. Responsibility for the actions performed by aggregates of ancestor dharmas is borne by the aggregates of descendant dharmas in virtue of the causal chains that bind them.

Perhaps we can then think of responsibility in terms of evanescent dharmas, but we cannot think of compensation on this model. Someone might object: if we give up all notions of persons, how do we go about our daily lives of compensating one individual for the work he has done? If there is no concept of person, this practice starts to seem hard to follow. Imagine that we actually eliminated the concept of a persisting person, such that we did not even have a notion that we could work with, it seems that we couldn't have a market economy. Why would I produce a good and sell it to you if I did not think that you would regard me as the same individual when it comes to paying me for the good?Footnote ¹ Before I turn to think about Abhidharma accounts of compensation and social treatment, I return to the thought embedded in Buddhaghosa's elaboration of the meditation techniques. These Buddhists urge us to think about interpersonal relations as extensions of intrapersonal relations. Parfit, too, wants us to think of our future selves as not that different from other people. Parfit concludes the section of ‘Immorality of Imprudence’ by saying that “We ought not to do to our future selves what it would be wrong to do to other people” (Reference Parfit1984: 320). Parfit wants us to think of intrapersonal relations as extensions of interpersonal relations. This is the other way around to what I take Buddhaghosa to be suggesting. Both Buddhaghosa and Parfit are thinking of our ordinary everyday way of thinking of relations between persons very differently from the way we ordinarily do. We tend to think or our relations with others on the model of transactions. Since the 1950s, the popularity of transactional analysis promotes thinking of all communication exchanges between people, even our intimate relations, as transactions (Berne Reference Berne1958). Many of our problems, these psychologists and psychotherapists claim, come from transactions that are unsuccessful. I propose, on behalf of the Abhidharma Buddhists, that we reject this paradigm and that we should not think of relations between persons as transactions. Think of our more intimate relations, with one's children, for example; these are not transactional. Parents do not expect something in return. Should your children compensate you with filial piety for all you have done and are doing for them? That certainly does not seem to be the right way to think about this. When the Buddhists talk about breaking barriers between people, the idea is to think of other people—loved ones, friends, strangers, even enemies—as one thinks of one's own future self. The Buddhist practice of dāna, best thought of as ‘giving away’ (rather than ‘gift-giving’, which is standard in the Buddhist literature) as an ethical and soteriological virtue, is both a duty of the laity as well as an exalted ‘perfection’ of the Bodhisattva. This practice forms the basis of lay renunciation for relinquishing attachment to material possessions and supporting the community and as a moral ideal. The ideal for the Buddhist is that we live in a community (saṅgha) with no concept of private property, relinquishing all material possessions. Think of ‘giving away’ (dāna) whatever goods and services one can offer to the community rather than ‘transacting’ or ‘compensating’ as a starting point for our ordinary everyday dealings with others. It is perhaps better to think of expanding our circle slowly. Perhaps from giving to one's own children to giving to one's students and so on. The university pays academics for teaching, but I do not think academics think that the students owe them compensation for giving them goods; it is rather a gift of knowledge that is shared with students. Markets and other financial institutions allow ordinary people like us to function in an interdependent society, but the Buddhist ideal is to take this fact of interdependence seriously so that it would lead—ultimately, in an ideal Buddhist society—to a dissolution of markets and other financial institutions. A similar conclusion is endorsed by Graham Priest (Reference Priest2018).

Again, when it comes to social treatment, the Abhidharma Buddhist urges us to think of all others as one would think of one's own future self. So setting aside money for a future self is no different from sharing it with our contemporaries. These consequences sound startling and so they should. Does Buddhism succeed in offering a better structure in the way that it promises? That is not a question I can answer here. My task was to present a thoroughly revisionary metaphysics and its entailments for our person-related practices, concerns and attitudes. I leave the reader to ponder whether the proposed alternative is indeed a better structure.