Response to Mark Siderits Review

·期刊原文
Response to Mark Siderits' Review
By Paul Williams
Philosophy East and West
Vol. 50, No. 3 (July 2000)
pp. 424-453
Copyright 2000 by University of Hawaii Press
Hawaii, USA

--------------------------------------------------------------------------------

p. 424 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

I am enormously grateful to Mark Siderits both for his kind comments on my Altruism and Reality and for the eminently courteous and intelligent way in which he has developed his disagreements with the two chapters of my book in which I apply the tools of contemporary analytical philosophy to assessing some of Śāntideva's arguments concerning issues related to personal identity, rebirth, and altruism. In writing

p. 425 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

those chapters -- each an independent paper -- I had intended to be provocative, for Śāntideva's arguments and their Buddhist context involve philosophical issues that have been vital down through the ages and are very much at the center of contemporary discussions in metaphysics, the philosophy of mind, and ethics. Our Buddhist thinkers put forward a position and arguments for that position. Those arguments beg to be analyzed, assessed, and developed. They may not work. If we are interested in truth (that sadly unfashionable concept), there is no avoiding wrestling with the arguments. In doing this we transcend the history of philosophy -- the understanding of those thinkers who provide our material -- and we do philosophy itself. We aspire to do philosophy just in the sense that Parfit or Strawson do philosophy. This wrestling with the arguments is, if you like, "analytic meditation." At least, it is what I take analytic meditation to be, and it certainly is an essential ingredient in what analytic meditation was and is in, for example, the dGe lugs tradition of Tibetan Buddhism.

On Bodhicaryāvatāra 8: 97-98
My Argument
Where Mark Siderits and I disagree I have yet to be convinced that he is right. I don't suppose he thought I would be (which is not to say I might not become convinced in the future -- I keep changing my mind). Let me begin with his discussion of my second chapter, the reprint of my paper "On Altruism and Rebirth: Philosophical Comments on Bodhicaryāvatāra 8:97-8." As I recall, the relevant philosophical argument is this: Śāntideva wants to argue (1) that in all pertinent respects (ontologically and morally) the relationship between me now and contemporary others is exactly the same as the relationship between me now and "my" next rebirth. That is, "my" next rebirth is in all pertinent respects other to me now, and this will remain the case so long as I am in the present incarnation. (2) I do, as a matter of fact, have altruistic care and concern for "my" future rebirth. (3) Therefore, in order to be rationally and morally consistent I should also have altruistic care and concern for contemporary others. Morally, the care and concern I have for one other (my future rebirth -- the argument does not require reference here to more than one future rebirth) must be universalized to all others, including contemporary others. Śāntideva's argument is clever, and in practice may well be convincing for a Buddhist. If so, who could doubt the value of the results? But I have responded with the suggestion that in order to restore rational and moral consistency it is open to Śāntideva’s opponent (or Śāntideva himself in his meditation, qua opponent) to reply that he or she had not realized that the relationship between me now and "my" future rebirth was one of difference, otherness, and to resolve to cease caring about his or her future rebirth(s) along with not caring about contemporary others.

The Reborn Being as Conventionally a Different Person from the One Who Died
It should be clear from all this that Śāntideva's argument requires that the reborn being be different in all relevant respects from the being who died. The reborn being

p. 426 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

must certainly be in all relevant respects as different as contemporary others are different, otherwise Śāntideva’s argument would lose all plausibility. And Śāntideva does indeed hold, quite unequivocally, in Bodhicaryāvatāra 8:98, that the being who is reborn is different from the being who died. rGyal tshab rje and the dGe lugs tradition have always been very clear that this means that the reborn being is a different person (pudgala/gang zag) from the one who died, and a great deal of my paper in fact treats different commentarial ways of understanding Śāntideva’s verses. It seems to me that rGyal tshab rje is right in his understanding of Śāntideva here. Śāntideva must mean that the reborn being is a different person from the one who died; otherwise Śāntideva could not establish his desired conclusion of altruism. And I also think that Śāntideva and rGyal tshab rje are right to hold this. I can make no sense of the reborn being (supposing there is such a thing) actually being the same person as the one who died.

For our purposes here I am prepared to take "being a person" in a very wide sense, converging with "subjectivity" when applied to sentient beings. In other words, I am prepared, for the purposes of our discussion, to give personhood to animals and even cockroaches (or blobs, providing they are not mindless blobs), since any hypothetical sense in which the reborn being might be the same person as the one who died in the context of Buddhism must include that possibility. Now, I am not sure exactly what factors go into creating a person, but among the factors must be some sort of psychological coherence and continuity, plus a place ascribed as a person in the social life-world (whichever social life-world is relevant) -- and, crucially, some sort of physical coherence and continuity as well. While there might be some debate over borderline cases, I take it that this embraces a minimum for ascribing person hood (note: it is not to be taken as a definition). But all of these, it seems to me, break down in attempting to ascribe same-personhood over lifetimes.

Buddhists frequently explain that the relationship between the reborn being and the one who died is actually the same relationship as that between, say, a three-year-old and the fifty-year-old he subsequently becomes (see, e.g., Milindapañho 40). [1] Now, this relationship could be taken as referring simply to one of causal continuity, in which case it might be true that the fifty-year-old is on a causal continuum with the three-year-old. But this is unhelpful. That the reborn being is on a causal continuum with the one who died can be accepted without any implication that the reborn being is in any sense the same person as the deceased. To explain Y as causally contiguous (i.e., on a causal continuum) with X is a particular way of explaining their relationship. It explains how Y came about. But it is not an explanation in terms of identity or difference. Thus it is perfectly consistent to say that Y is causally contiguous with X, but that Y is different from X. In fact, we do it all the time. My son is causally contiguous with me, but he is very definitely different from me. As a matter of general principle the common Buddhist claim that if two things are causally contiguous they are neither the same nor different is true only if "different" is taken as meaning "different in the way that things which are not causally contiguous are different."

But this would not enable us to deny (if evidence suggested otherwise) that in

p. 427 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

our normal, less-restricted sense of "different" the reborn being is different from the deceased. And to claim that the relationship between the reborn being and the one who died is the same as the relationship between a three-year-old and the fifty-year-old he subsequently becomes in a way that is more significant than simply claiming causal continuity -- that is, there is no significant ontological difference between the two cases -- is simply false. There is a significant difference between the two cases, a difference that is crucial to what it is to be a person. Between the being who died and the being who is reborn there is a crucial break in physical continuity and coherence. There is also incidentally a break in their roles in the life world (i.e., apart from perhaps certain cases in, e.g., Tibetan monasteries, we do not treat the being who died and the reborn being as if they are the same person), and it seems clear to me that there is even a break in significant mental continuity and coherence between, say, an elderly dying lawyer and an embryo, not just because elderly lawyers and embryos "think" in radically different ways -- different ways that seem to me, in cases where elderly lawyers within a fairly short period of time become embryos, to suggest different persons (or, as some would have it, potential persons with respect to embryos) -- but also because it seems pretty likely to me that the embryo could not even remember being an elderly lawyer.

Without wanting to follow Locke in directly equating continued memory with personal identity, it nevertheless seems to me that where there is absolutely no memory of me saying (to use Geach's [1969] example) "The die is cast" and crossing the Rubicon -- nor could there even possibly be such a memory of me (this person, Williams, son of Rob, wife of Shad, employee of Bristol University) doing this -- I could scarcely claim to be the same person as Julius Caesar. [2] Everything that would make for personhood in the case of me and personhood in the case of Caesar is different. In other words, the description of what it means to be a person would fail if we claimed that the person who dies and the person who is reborn are the same person. "Being a person" no longer makes sense, no longer has any meaning. In case this still seems unclear, consider my (very likely) rebirth as a cockroach in South America (which, it seems to me, is not significantly different here from being an embryo). It makes no sense whatsoever -- it is to strip words of all meaning-to say that a cockroach in South America (or an embryo in Kathmandu) is the same person as Williams, a fully grown human who was in England. And we could scarcely maintain that the question of whether or not they are the same person is simply a matter of language and convention. There can be no sense of meaningful spatiotemporal, physical, or even mental continuity that will allow us to say we have the same person in a case like this. If Williams is reborn as a cockroach, then any basis for speaking of the same person fails. It seems to me that Williams will have perished, irrespective of whether we can make any sense of causal and/or karmic continuity. Note this. It seems to me that according to the Buddhist model, I, the person who is Williams, perish notwithstanding karmic continuity. I (i.e., this person -- this has nothing to do with the ātman) might be happy to perish, but that is how it is. We ourselves and all our hopes and fears are an almost infinitely insignificant moment in infinite time.

p. 428 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

If I were Śāntideva's opponent, I could indeed ask: what is it to me that there is some other person who is on a causal continuum with me? If I am already a kind and generous person, it will make a difference to me that someone else will succeed me who bears to me now a relationship such that I can have an effect on their well-being. But as an argument for why I should become a kind and generous person, Śāntideva's argument fails. That Buddhists in practice do think of future rebirths as in some relevant sense them, reborn, is a confusion. Perhaps it involves insufficient imagination. One tends to imagine oneself in a future rebirth as a human being -- in fact, the person one is now, maybe in a different fleshly body. I cannot even imagine it being me in any sense as a cockroach, or becoming a cockroach by passing, instant by instant, through whatever process a cockroach goes through when it emerges from the last moments of a dying Williams. How am I supposed to understand it being the same person? [3]

Let us note also another problem. If the reborn being is the same person as the one who died, then it would follow that throughout the infinite series of rebirths there is only one person. This is because if it is the same person in lifetime Y as it was in lifetime X, and the same person in lifetime Z as it was in lifetime Y, then it should follow that it is the same person in lifetime Z as it was in lifetime X. And so on, throughout all rebirths. But it makes absolutely no sense at all to speak of the same person throughout all rebirths (even if we do not understand "same" in the sense of "identical"), for we would again face, in a rather acute form, the stripping of all meaning from the concept "person." Moreover, in the Buddhist context it is difficult to see how Siderits' interpretation will avoid becoming on the conventional level a form of pudgalavāda. There must at least conventionally be a person, a pudgala, in addition to the aggregates, the skandhas, since clearly if, between the reborn being and the deceased, there is the same person there are nevertheless not the same aggregates (even the continuum of the physical aggregates is in a crucial sense different, for example). Thus, the person must exist at least conventionally over and above the aggregates. And it is the same person throughout all lives. [4]

I see no reason to think that Śāntideva and rGyal tshab rje would disagree with what I have stated above. And Siderits also accepts that this is Śāntideva's position, but "[t]o say that it is one person who dies and another person who is reborn is incompatible with the orthodox Buddhist view of rebirth." Thus, Siderits wants to claim, Śāntideva's position (and that of rGyal tshab rje and the dGe lugs tradition) is atypical. Inasmuch as my paper was solely concerned with the assessment of Śāntideva's argument, perhaps I should be happy with this (although Siderits, in the light of Buddhist orthodoxy, subsequently takes back the concession he had granted). Unfortunately, however, I am not convinced that Siderits' view of Buddhist orthodoxy is correct. Indeed, I am not sure what the "orthodox" Buddhist view of rebirth is. I assume Siderits is thinking of the formula in Milindapañho 40, where it is said that the reborn being is neither the same as nor different from the one who died. Siderits thus takes it that the orthodox view is that, as it would be expressed in Abhidharma analysis, the formula "neither the same nor different" can be split

p. 429 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

according to the two truths. So the reborn being is conventionally the same person as the one who died, but ultimately, in terms of the continuum of dharmas, there is no being who dies or is reborn. There is just a causally conditioned flow. Thus, Siderits tells us that "[t]he Buddhist tradition had generally taken personal identity to extend over rebirth."

One can see what Siderits means. In Buddhist tradition, to speak of the same person in the light of no ātman is just to employ a word, or a concept, that has no fixed, unchanging referent. We are familiar with cases in the Jātakas, for example, where the Buddha might say "I was at that time the deer Such-and-such." But nevertheless, if this is the orthodox Buddhist position, that the reborn being is the same person as the being who died, then I have suggested above and in my book that this is incoherent. It would make no sense for the Buddha to say he is the same person as the deer Such-and-such. Deer Such-and-such was born from deer parents. The Buddha was, as Buddhists have always said, a human being. It is part of what we mean by the ordinary everyday usage of "persons" that they are born, and part of what we mean by "being a human person" that one is born (one way or another) from human parents. That person who was the Buddha was thus born from human parents, not deer parents. What must be meant in the Jātaka stories, even conventionally, is that the Buddha is on a direct causal continuum with the deer Such-and-such. It seems to me that, in spite of sayings like those in the Jātakas, if the Buddhist position is going to have even a remote chance of plausibility it is going to have to maintain that the reborn being is not the same as the one who died even conventionally.

Note, however, that I do not want to commit myself at this point to maintaining (with Śāntideva and rGyal tshab rje) that the common Buddhist position is that the reborn being is different from the one who died. I myself do think that the reborn being (if there is such a thing as rebirth, in some sense) must be a different person from the one who died, and I think that it is difficult to see on the Buddhist premises how there could be causal connections of the right sort to maintain even a rudimentary personal identity between Williams and either an embryo or a cockroach. But I do not want to maintain that this would follow as the Buddhist position simply from denying that the reborn being is the same as the one who died, even conventionally. For the Buddhist position also denies -- for me even conventionally -- that the reborn being is different from the one who died. Actually the reborn being is on a causal continuum with the one who died even conventionally. And this is how the Buddhist avoids annihilationism and eternalism.

Am I right in applying to the Buddhist position the formula "neither the same nor different" even conventionally? I think so, if we want even the rudiments of conceptual precision. If conventionally the reborn being were the same person as the one who died then we have seen, by an argument I employed above, that the Buddhist would be committed to maintaining that conventionally there is just one person (per continuum) throughout the infinite series of rebirths. Thus, conventionally at least, there would be some sort of ātman, or perhaps some sort of pudgalavāda. Note

p. 430 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

that there would have to be just one person per continuum throughout the series of rebirths. It would not be enough to argue that there is one continuum but not one person. If the relationship conventionally between the one who dies and the one who is reborn is that of one causal continuum, then I accept that. As I understand it, what Siderits wants to argue, however, is that the reborn being is conventionally the same person as the one who died, not conventionally a future spatiotemporal stage of the same continuum (or perhaps a very similar person, though what "very similar" could mean in the case of Williams and the cockroach escapes me). Examples used to illustrate the relationship between the reborn being and the one who died seem to indicate precisely that we are dealing even conventionally with a continuum where it is held to be impossible to use either the locution "same" or "different." Thus Buddhadatta, in his Abhidhammāvatāra (603-611), states what seems to me to be the "orthodox" Buddhist position, if there is one, when he argues that rather than complete identity or complete difference what we have in the case of rebirth is a continuum, and he illustrates this with the relationship of milk to curds. [5]

This same example is often used in such a context -- it is also found in Milindapañho 40-41, for example. Buddhadatta explains that if there were complete identity between milk and curds then the latter would not be produced from the former. But if they were completely different, then one could not establish the legal relationship of ownership between the two. These are homely points, made using a homely example. It would seem clear that what it tells us is that where one thing is a transformation of another, or evolves out of the other by some sort of continuum, normal everyday practical responses in terms of complete identity or complete difference are inappropriate. The point here clearly cannot be one of (what would be for Buddhadatta) ultimate truth -- that the continuum of dharmas that are constructed into milk and curds do not allow us to speak of identity or difference. That may well be true, but it would be to explain the obscure by the obscure. What Buddhadatta wants to say is that in everyday life (i.e., on the conventional level) where there is a continuum of transformation we do not speak of complete sameness or complete difference. In everyday life, curds are neither the same nor different from milk. Rather, they are on a causal continuum with milk. Would Siderits want to say that conventionally curds are the same as milk? Or that conventionally, to use another common example, the sprout is the same as the seed? [6] Why, therefore, should one say that in the case of rebirth -- the very topic that these examples are intended to illustrate -- being neither identical nor different applies only ultimately and not conventionally?

In the Sarvāstivāda Vijñānakāya there is a discussion with the Pudgalavādin in which the "orthodox" Sarvāstivādin asks his opponent to agree that it makes no sense for someone to maintain that the reborn being is the same person as the one who died, since the Buddha quite clearly taught a range of different ("and they are established as different") destinies -- hells, animal realms, and so on. This, again, is a matter of the conventional level. In other words, one cannot say that we have the same person in a hell realm as is (after dying in hell) the reborn being in, say, the animal realm. These are different, and the context of the discussion makes it clear

p. 431 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

that the problem with saying that we have the same person in each case is that descriptions of what is to count as being a person in each of these realms are also completely different (see Watanabe 1983, pp. 177 if.).

Conventionally curds are not the same as milk, and conventionally the reborn being is not the same as the one who died. The reborn being is for Buddhist "orthodoxy" not different from the deceased either, although Śāntideva, rGyal tshab rje, et al. say he or she is. I have argued that Śāntideva is right to say this, for being on a causal continuum does not rule out difference, and we have here a case of patent difference. Perhaps Śāntideva is telling us how it really is for us. But in his practical context of reasoning into compassion Śāntideva appears to have adopted a strategy that he hoped would actually work. His fellow Buddhists -- and he himself -- do concern themselves with "their own" future rebirths. Śāntideva was entitled to hope that on this basis he could persuade them to concern themselves with contemporary others. Given his premises, I think Śāntideva's position is philosophically coherent as far as it goes, and his moral hope and perhaps even his moral strategy appropriate. In the last analysis, however, I think he simply fails to establish his conclusion.

Siderits' "More Charitable Reading"
Śāntideva requires no more "charitable" reading. I have taken his argument as the one he stated. Siderits' charitable rereading, it seems to me, probably will not help. For Siderits, Śāntideva is drawing attention to the fact that "it is through a convention that there is constituted a person who then exhibits self-interested concern for the future," and that "the future suffering that will occur to me is a socially constructed fact" that can be changed. But while this might have been his intention, Śāntideva's argument as it stands is stronger than that; anyway, I find these claims doubtful. Siderits' grounds for thinking that the claim that future suffering will occur to me is a socially constructed fact appear to be that we have to teach children to anticipate future pleasure and pain. But assuming this to be correct, [7] that we have to teach someone something does not in itself entail that what is taught is "a socially constructed fact." I might teach someone that three plus three equals six, but this is far from claiming that mathematical facts are all "socially constructed." Or I might teach someone that the sun is the center of our solar system and that the planets of this solar system circle around the sun at whatever rates they do, but this, too, does not become a "socially constructed fact" (even if the language in which it is expressed requires social construction, and these facts may also enter into social constructions). Or that William the Conqueror was King of England before Henry VIII. Or that the rocks of the Devonian age are older than those of the Carboniferous (a fact that was true long before there were any societies to know them or such conceptual categories as "Devonian" and "Carboniferous"). Or that my eldest son bears to me whatever genetic relationship he bears to me. [8]

I am not quite sure what it would be for me simply as a socially constructed fact to learn to identify with the past and future states of this causal series ("my having learned ... this causal series"). Why should I do that? Because others have done it

p. 432 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

and I am taught by them? This looks close to question-begging (or circularity). Moreover, I would imagine that there are very solid biological factors (survival) involved in a being exhibiting self-interested concern for its future. To suggest that this biological need (surely it must be a need) was learned simply as a socially constructed fact seems to me to be quite implausible. [9] Was there chronologically or conceptually first a society of those who did not exhibit such a concern so vital for survival? How, chronologically or conceptually, could such social constructions originate? One might have to learn altruism, but I would imagine that there is much less learning involved in acquiring a sense of one's own future welfare. In the book, I discuss in passing (pp. 121-123) what seems to me to be the given identity and distinctness of natural kinds like biological organisms (trees, human beings, etc.) as compared with cultural constructs (artifacts) like tables or trains. To reduce the former to the latter seems very problematic.

I find it difficult to see how the genesis of the person that we are could be explained as simply a convenient socially constructed fact. It is certainly convenient -- indeed, so convenient that I suggest that persons are already given, and were they not there would be no societies at all. The reductionist model tends to presuppose the very person it is trying to reduce, and (it seems to me) invariably fails to explain how to reassemble our common or garden-variety person from its reductive explanation. Thus I find problematic Siderits' assertion that "[t]he convenient designation 'person,' like any other convenient designation, is adopted for its utility, given our interests" (my italics). Whose interests? If we are there, with our own respective interests, then what more do we need? And "[w]hen present psychophysical elements learn to identify with past and future elements in a causal series, this lessens the incidence of what we conventionally call gross imprudence." Why should that be? Since on Siderits'/Buddhist premises present psychophysical elements in a causal series are not identifiable with past and future elements in the series (it is the same series, but not the same elements), why should sheer falsehood lessen the incidence of "gross imprudence" (i.e., our pain, suffering)? [10] It seems to me that persons are given first. The reductionist model comes after, both chronologically and, crucially, conceptually. That is why it constantly presupposes persons.

"[J]ust as the child can learn to identify with past and future elements in the causal series -- to anticipate that future pain, to feel shame at that past action -- so the aspirant to enlightenment can learn to identify with suffering occurring in distinct causal series" (Siderits). Is that right? When the child, supposing that this is what happens, learns that if he or she does not eat now there will be pain tomorrow, is this essentially the same process as a bodhisattva now realizing that she must remove Archibald's pain-in the bodhisattva's case a result of simply learning to see things in an alternative but equally true way? I remain to be convinced. I do not think that all this is nothing more than ultimately arbitrary constructions, a matter of what has been learned through the process of socialization. It seems to me that there is a biological basis to self-concern that is arguably lacking in altruism, but in any case certainly different. Of course, we can learn to overcome self-concern, and we can learn altruism. But this ability to learn otherwise simply does not entail that self-

p. 433 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

concern itself was nothing more than acquiring an ultimately arbitrary social construction. Crucially this is reflected in the essential subjectivity of pain. I think there is a necessary connection between pain and the subject who is in pain. Thus, it simply is not, experientially (qua qualia), the same for me to have a pain as it is for Archibald to have a pain (and I do not, incidentally, think that this necessarily entails a vicious adherence to private objects/privileged access). Of course, I wholeheartedly approve of learning to remove the pain of Archibald, of learning altruism, but I suggest that it cannot be founded on arguments that simply ignore or deny this basic experiential point. Even on Siderits' charitable and intelligent rereading of what Śāntideva actually says, I fail to see how the hoped-for universal altruism of the bodhisattva follows from the argument(s) we have been considering.

On Bodhicaryāvatāra 8:101-103
My Argument
Mention of pains brings us neatly to Siderits' comments on my long fifth chapter (a wholly new chapter written for this book), in which I treat many of these issues at much greater length. There I contend that Śāntideva's argument at Bodhicaryāvatāra 8:101-103 will not work since, in order for an implication from no Self to no selfishness to have any plausibility, no Self would have to entail no conventional self as well as no ultimate, unchanging metaphysical ātman. It seems to me obvious that selfishness is a matter of conventional selves. It involves giving preferential treatment to this person -- the one who is me -- rather than that person, the one who is Other. No selfishness simply does not follow from no Self (no metaphysical ātman) per se. [11] I can quite consistently accept that I have no metaphysical ātman -- say, an independent unchanging Cartesian self -- and be perfectly selfish. I suspect many philosophers and scientists do and are. [12] Thus, if Śāntideva's argument at Bodhicaryāvatāra 8:101-103 is to be plausible, it involves making some sort of ontological distinction between the conventional self and duḥkha such that we can consistently deny the conventional self and hold that it still makes sense to speak of there being duḥkha. Thereby it will also still make sense to exhort the removal of duḥkha, but a duḥkha that must be stripped of any reference to subjects because stripped of any reference to conventional selves. Śāntideva chooses to do this in terms of a model that would be perfectly acceptable to other Buddhists as well, a model familiar from the Abhidharma. There are no selves (even conventionally) because persons are fictions, conventional constructs superimposed upon a spatiotemporal series (of momentary dharmas).

But I argue that this strategy will not work. Take the case of physical pain, the actual sensation felt when treading on a drawing-pin. Such pains form a subclass of mental events coming under duḥkha. Thus, if the bodhisattva is to remove the duḥkha of others, he or she has to remove such physical pains. But I contend that in removing the conventional self in order to argue for altruism, Śāntideva has himself removed the very basis for making sense of pain statements. I argue, using as alternatives an adverbial analysis and an event analysis of pain statements, that

p. 434 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

physical sensations of pain have a necessary dependence on subjects. [13] Śāntideva's argument requires that we can make sense of pains not anchored to subjects, and the concept of pains not anchored to subjects seems to me to be meaningless. Thus I argue that since, granted Śāntideva's argument, the bodhisattva could not remove pains, the bodhisattva could not remove all duḥkha. Thereby Śāntideva has destroyed the bodhisattva path. Inter alia I also offer some critical comments on the reductionist approach to persons, and I suggest against the Buddhist that it is by no means incoherent to think of wholes as not reducible to parts, and to include wholes in one's ontology (perhaps as substances) even though they are neither the same nor different from parts. And I toy with the idea of the person as a whole, but in the end, at the moment, I think I incline toward Strawson's view of the person as a sui-generis irreducible.

Siderits states that he agrees with me that Śāntideva's argument will not work if it involves denying the conventional reality of persons. But my position is that Śāntideva is in a dilemma. His argument will also not work if he does not deny the conventional existence of persons. For Siderits, "Śāntideva is once again reminding us that since persons are ultimately unreal -- that ultimately there is suffering but none who suffer -- the conventional reality of persons can only be grounded con-sequentially, which in turn shows that the utility of the me-construction can be improved upon." But I am not sure what Siderits intends here by drawing this onto-logical distinction between persons, which he says are ultimately unreal, and suffering -- my interest is in physical sensations of pain -- which he can still accept ultimately. Let us put to one side here (it is not particularly relevant) the point that Śāntideva, as a Mādhyamika, would not accept the ultimate existence of pains. Siderits seems simply to repeat Śāntideva's position, a position that I argued lands him in a dilemma. I accept that some sort of ontological distinction like this is one that Śāntideva would want to hold. Śāntideva has to be able to make sense of pains in some sense without subjects. I have argued that the ultimate unreality of persons is irrelevant.

The question is whether pains necessarily imply the subjects of pains -- persons conventionally. I argue that they do. Siderits' defense of Śāntideva involves adhering to the position that there is a level -- surely in spite of what Siderits says it must be the conventional level, since that is the level on which everyday life and the altruistic activity of bodhisattvas takes place -- where we can make sense of pains without subjects. For Siderits, presumably the association of pains with subjects-or at least the particular subjects they have -- is again simply an ultimately arbitrary social construction. He certainly suggests that this, too, "is merely a useful device for realizing the ultimate (impersonal) goal of maximizing overall welfare." Since the association of pains with the particular subjects they have is ultimately a social construct, it can be changed (presumably by the great mental power of a bodhisattva). As far as I can see (although he does not agree), Siderits is committed to the coherence of conceptually (logically) free-floating pains, and this seems to me completely bizarre. To me it flies in the face of all that we know about the conceptual and biological

p. 435 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

framework for understanding pain statements. What must be for Śāntideva a conceptual construct (prajñaptisat), and what Siderits actually argues is a social construction, for me involves a necessary connection. I argue that the explication of pain statements necessarily involves reference to subjects in pain. Subjects are logically necessary for the individuation of pains, and not just subjects as such but also the particular subjects that particular pains have. This is not just a relatively convenient but ultimately arbitrary social construct. Pains are not logically separable from subjects. It seems to me there is no difference in ontological status between pains and subjects. Siderits (and Śāntideva) need to show that it is otherwise.

Siderits' talk about the goal of maximizing overall welfare appears to beg the question as an argument for altruism. If we first of all accept that "our thinking of ourselves as persons each living a separate life -- is merely a useful device for realizing the ultimate (impersonal) goal of maximizing overall welfare," then no doubt we will indeed become altruistic. But why should we think this? Even if reductionism is true, this requires a coherent link between the reductionist model of persons and impersonal altruism, and this link seems to me to be lacking. Is it logically contradictory and/or psychologically impossible to be a reductionist with regard to persons and yet still be selfish? [14] This would only follow if the reductionist model of persons entailed that we could not make what Buddhists call the "conventional" distinctions between persons. If we wish to be consistent we could not logically and/or psychologically (and morally?) privilege ourselves. But I do not see why this follows from the reductionist position. Thus, why not, in fact, privilege myself? Siderits' concern with impersonal overall suffering will only follow if one is already altruistic. Śāntideva's argument, even in Siderits' version, unfortunately will not work as a reason for becoming so.

The Sneeze, and the "Impersonal Formulation"
Which brings us to Siderits' sneeze. [15] I love the example, but I am not convinced! There are just too many gaps in the explanation. I am sure Siderits would want to say that this is only a plausible sketch for a theory, rather than a fully worked-out model. Actually, in the book (p. 133) I used the example of my wife's feeling happy because I feel good. It is my actual feeling of well-being that causes her actual feeling of happiness. Thus, this is a case where a mental event in what we normally call the psychological continuum of one person causes a mental event in what we normally call the psychological continuum of another. How is it (without begging the question) that "my wife's" mental event is not integrated into "my" psychological continuum, given (as is commonly held to be the Buddhist view) a merely causal account of how we construct persons out of psychophysical continua? Siderits' example of the sneeze is perhaps easier for him to handle, since he can appeal to physical proximity or otherwise of nose, ear, and hand. Mental states do not in themselves have physical locations, and any attempt to anchor mental states in relation to bodies will have to do so without appealing specifically to my body and my wife's body. [16] But even with Siderits' sneeze l am not sure that the startlement,

p. 436 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

and the relief -- both mental events (let us say) -- can themselves be related to the nose and the ear qua moving out of proximity to one another. They are, of course, causally related, but that causal relationship is not in itself affected by spatial relocation. At least I do very much want to say, as Siderits surmises I would, that this sneeze model is "arrived at in total reliance on our pre-theoretic intuitions about persons and personal identity, so that it does after all smuggle in the concept of a person." It appears to be a reduction from the everyday givenness of persons and to rely very much on bodily cohesion in supplying a foundation for personal identity in everyday life (cf. our discussion of rebirth above).

There are also some other problems that spring to mind with Siderits' discussion of the sneeze.

1. Siderits' explanation would seem to rely on pinning personal identity to contingent events such as moving out of proximity to body parts. Supposing two people remained for some reason throughout their lives in exactly the same spatial location to each other. We could take two Siamese twins, although one could think of other possibilities. These two Siamese twins are joined at the waist. They have just one body between them below the waist, but different chests, arms, heads, brains, main nervous systems, and so on. They also have rather different interests and personalities. Archibald likes kite flying, Freda likes absailing. It is thus quite clear that Archibald and Freda are different persons. Now, the nose-state of Siamese twin Archibald causes a sneeze in Archibald, and this causes startlement in Siamese twin Freda, who was asleep at the time. Archibald feels relief. It is by no means clear to me, without begging the question, that including the feeling of startlement in "Archibald's" psychophysical series would "diminish the degree of overall causal connectedness," at least on the grounds given by Siderits, while "including the feeling of relief in this set [that of Freda], on the other hand, would increase the degree of overall causal connectedness of the set." Why, given the Siderits version of the reductionist model, are Archibald and Freda constructed into two persons rather than one?

2. Siderits wants to tag the inclusion of the feeling of startlement into a different personal series from that of the feeling of relief on the basis of what happens subsequently (bodily sets move out of proximity, etc.). But this seems paradoxical. The feeling of startlement in "Lawrence's" series occurs at t3, and the feeling of relief in "Gertrude's" series occurs at t3, and yet we can only integrate the feeling of startlement into the person Lawrence, and the feeling of relief into the person Gertrude at tn, where tn is later than t3, on the basis of contingent events that may or may not happen subsequently. What is the precise relevant connection between events that happen subsequently and the integration of feelings into the respective personal series? At the time that the events occur, they have no connection at all with Lawrence and Gertrude, and they may never have if events turn out differently. If the end of the world occurs at t3, the startlement and the relief will never have been Lawrence's startlement and Gertrude's relief. This may indeed be Siderits' and the Buddhist position, but it seems counterintuitive.

3. What do we mean by "increasing (or diminishing) the degree of overall causal

p. 437 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

connectedness of the set"? What is "greater overall causal connectedness"? Is it perhaps simply more causal connections taking place? It is difficult to see a priori that including the feeling of startlement in "Lawrence's" series rather than "Gertrude's" series necessarily leads to more causal connections taking place. I am not even sure what the criterion for the identity of a causal connection is, and therefore how to go about counting causal connections. Is it, on the other hand, that including the feeling of startlement in "Lawrence's" series rather than "Gertrude's" series leads to more coherent causal connectedness? But what do we mean by "coherent causal connectedness," without begging the question by relating coherence to persons and the lived world we all occupy? And according to Siderits "[t]he goal of the construction is to include as many elements as possible consistent with the maximizing of causal connectedness." But the goal for whom? And what is to count as the maximum of causal connectedness? I'm not even sure I understand what the expression means. The maximum of coherence expressed in causal terms?

If understand him correctly, Siderits' impersonal formulation for adopting one particular construction rather than another is as follows (with comments).

(a) The point of the construction of persons is to maximize overall utility. But according to Siderits' own model, the construction of persons does not maximize overall utility. That is why with Śāntideva we need to adopt the impersonal model and practices of the bodhisattva. Thus, the construction of persons fails in its point, and, indeed, it is going in quite the wrong direction. But why, then, did it ever take place? What is its foundation -- or is it just brute fact? How could things be such that a construction of persons plausibly might maximize overall utility. How can this be accounted for without presupposing persons already? And from where (in fact, and in the Buddhist tradition) do we get the suggestion that the point of the construction of persons is the maximizing of overall utility? [17]

(b) "[S]ystems with the capacities of self-scrutiny, self-control, and self-revision turn out to be quite effective at local maximizing." Why should this be the case? Utility is maximized by systems with the capacities of self-scrutiny, and so forth. [18] Presumably this is not just pure chance, in spite of Siderits' use of the expression "turn out" (as Siderits says later, "This fiction is useful" -- that is, it just so turns out that the fiction of persons has its value). If we ask what in the nature of things makes this to be the case, the temptation is to say that it is because persons are actually prior to construction, and the reduction employed by the reductionist presupposes the conceptual primacy of persons. And what is meant by "local" maximizing? I take it (perhaps wrongly) that this refers to spatiotemporal location. But is it possible to explain spatiotemporal location without presupposing unities provided by physical and psychological sets such as bodies or (dare I say it) persons?

(c) Systems with the capacities for self-scrutiny and so forth can only be constructed out of sets exhibiting maximal causal connectedness. Thus, a person is constructed "as a series of psychophysical elements with maximal causal connectedness." As I stated above, I am not sure what is meant by "maximal causal connectedness," and I am also not at all clear what is the connection between

p. 438 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

maximal causal connectedness and the capacity for self-scrutiny and so forth. Why should it be that only systems exhibiting maximal causal connectedness show the capacity for self-scrutiny and so forth?

Fission, Moral Responsibility, and Justice
Siderits is happy in the reductionist model to embrace fission, and I am not sure that if the reductionist were correct I would object psychologically to fission taking place in me, assuming that first-person psychological attitudes and so forth survived. [19] But fission would have some strange repercussions. Consider the following thought experiment, granted the reductionist model of personal identity (I hope no one else has thought of it first).

What we want is a case where a crime is committed that could only have been done by one person. [20] Perhaps a king is paranoid about the possibility of assassination, and lives in one room with many strategies and devices to ensure that he can only ever be accompanied by one person in the room at a time (various drawbridges, barriers that will only allow one person to pass, and so on). Soon after a visit by Archibald, the king is found to be dead in a way that shows that it must have been murder. Only one person could have murdered the king, and it must have been Archibald. Archibald escapes capture, and ten years later, for reasons important to him but of no interest to us, Archibald undergoes fission. There are now two people claiming to be Archibald, each with Archibald's memories and having a causal connection "of the right sort" with the last significant moment of the single Archibald. For my example we can take it that each of the two continuants with Archibald, taken separately, would normally be considered to be Archibald in whatever way we consider human continuants to be the same person. Let us now call the earlier, single Archibald "Archibald1," and the two later continuants with this Archibald "Archibald2" and "Archibald3." Now, one day Archibald2 and Archibald3 meet to talk about old times, and while they are talking the law finally catches up with them and they are arrested and tried for the murder of the king. The case is overwhelming that Archibald1 did the crime, and both Archibald2 and Archibald3 are clearly shown to be Archibald, continuous in the right way with Archibald1. Truth drugs were applied, both Archibald2 and Archibald3 were found to be lying when they denied committing the crime. Truth drugs showed that when under hypnosis they said they did commit the crime they were speaking the truth. It is an open and shut case. But....

1. We now have two persons guilty of committing a crime that only one could have committed. It is unjust to punish two people for a crime that cannot have been committed by two people. One must be innocent.

2. It is also unjust to punish either Archibald2 or Archibald3 individually for the crime, since if two people cannot have committed the crime, each can claim correctly that it was the other one who did the deed and consequently not himself.

3. It is unjust to let a criminal get away with murder for a reason that is extrinsic to the actual issue of the crime and personal responsibility (i.e., that there happen to be two persons continuous in the right way with Archibald1). We can take it that

p. 439 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

both Archibald2 and Archibald3 each bear the right relationship to Archibald1 to be held in their own right morally responsible and guilty. Therefore it is unjust to punish no one, for it would be unjust not to punish the guilty.

Thus, whatever happens, something unjust will take place. But supposing overnight, unable to bear the strain, Archibald2 commits suicide. We now have just one person, Archibald3, who is continuous in the right way with Archibald1. We therefore punish Archibald3 for the crime of murder. This would seem unproblematic, and yet surely Archibald3 cannot now become morally and legally responsible for a crime simply because of something Archibald2 has done (i.e., gone out of existence).

What does all this show? Siderits claims that in the reductionist model "personal identity is all in the numbers: my now being the same person as someone at some distinct time just consists in there being enough connections (of a certain specified sort) between me and that person" (his italics). Well, one thing my thought experiment shows is that if personal identity is all in the numbers, where fission takes place it may prove very difficult to preserve our normal intuitions concerning the link between personal identity and moral and legal responsibility. This may be of concern, and encourage us to avoid the reductionist model unless forced by very strong arguments to the contrary, at least where that model is loose enough to allow for fission. Such moral problems could also prove to be of some concern to Buddhists. [21]

On Pains, and Functional Analysis
Siderits states that I find "incoherent the Madhyamaka claim that nothing bears its own intrinsic nature (which [I find] equivalent to metaphysical nihilism ...)." This relates to a couple of points in which I try inter alia to see why it was that followers of "mainstream" (e.g., Vaibhāṣika) Abhidharma would understand Madhyamaka -- in spite of its protestations otherwise -- to amount to nihilism. For nothing at all to have a svabhāva, in the sense in which this term is used in Vaibhāṣika Abhidharma (where it does not mean intrinsic nature, if "intrinsic" is taken to mean inherent or uncaused), would indeed amount to metaphysical nihilism. It would mean that literally everything is a fictional construct, and it is contradictory to hold that everything is a construct (for there would be nothing left for things to be constructed out of). That is why mainstream followers of Abhidharma insisted that Madhyamaka was nihilism (and, I would argue, that is why Yogācāra, too, came up with a fundamental substratum that was not -- in the Madhyamaka sense-a construct [i.e., prajñaptisat]).

On these particular principles I think I, too, agree. And what I see as "sheer sophistry" is the suggestion that what is dependently originated (1) thereby has no svabhāva in the sense in which this term is used in, for example, Vaibhāṣika Abhidharma, and (2) thereby bears something less than what we all know and love as normal existence. A pain is dependently originated, yet it fully exists in the only way a pain can exist. To quote myself (p. 245): "A pain appears as a pain. It hurts. The hurting is the pain. There is nothing more to the appearance of a pain than the pain itself." There is no such thing as an illusory pain as such. [22] And the bit about propositional attitudes arises out of an attempt I make to try and understand what is meant by the common dGe lugs pa claim that when we say in Madhyamaka that

p. 440 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

something is an illusion, what we mean is that it is like an illusion in that to us unenlightened folk it appears one way (as inherently existent) and exists another (as dependently originated). I argue that it makes no sense to say that a pain appears one way and exists another. Its appearance is its hurting, and that is the way it exists. The sensation of pain itself simply does not involve (even if it entailed, which I do not think it does) a propositional attitude, perhaps a belief like "pain inherently exists, i.e., exists independently of causes and conditions." Therefore, pain does not become experienced one way and exist another. Thus, pain is not (in this technical sense) illusory. Pain is thus "found under analysis" (and to say that it is not found under analysis just because it results from causes and conditions is "sheer sophistry").

This commits me to taking pains as different things from propositional attitudes. Actually this seems to me to be fairly uncontroversial. But I do not think it commits me to the "so-called myth of the given" as such, or at least in any vicious sense. Does Siderits himself think that part of what it is to be a sensation, or at least the sensation called "pain," is a propositional attitude? [23] And is that propositional attitude a belief that, for example, pains exist independently of causes and conditions? And how, then, would Buddhas experience sensations without that attitude (whether or not they experience pain)? All this seems to me far more controversial!

Siderits takes me as holding that pains are to be given a functional analysis. This is a teensy-weensy bit unfair to me (although perhaps it is my fault for not expressing myself unambiguously). I make it quite clear that I hold to pains as qualia. There is an irreducible phenomenological "what-it-feels-like" about pain. Like a true Mādhyamika, my use of adverbial and event analyses is as alternatives (some sort of combination may be possible, for all I know) that are intended to show one point only for the benefit of my opponent and others: that we cannot make sense of pains (as qualia) without necessary reference to subjects in pain. Pains cannot even conceptually, logically, be free-floating. Likewise in my quote from Robert Melzack. I require (at this stage and for these purposes) no other commitment.

Siderits refers to my discussion of a robot (in my note 84). I argue that I at least would be happy to take a robot as feeling pain, even if it were not made of flesh and blood like us. Pain in the case of a robot would still necessarily entail a subject, made rather differently from the way we are made (with apologies to my robot readers). Discussion of robots in this context often does go with a functionalist analysis, for one explains the robot's sensations in terms of the appropriate inputs and outputs. From a functionalist point of view, the particular hardware (flesh and blood etc. versus metal and computer chips) in which the appropriate stimuli and responses are realized is irrelevant to issues of whether we can say that the robot is in pain or not. Yet I do not refer to the example of a robot in order to support either an overt or a covert functionalism. I introduce the robot in the context of arguing that the relationship I wish to support between pains and subjects is a conceptual one, and not simply a contingent matter of pains being realized in bodies broadly like ours.

Thus, I do not appeal to the human physiology of pain in support of my position that pain requires subjects. That is all I am committed to with my robot. That I would

p. 441 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

be very wary of adopting a functionalist analysis here should be evident from the fact that in my use of the robot example I also granted for my purposes that the robot had consciousness: "supposing nevertheless we granted that the robot had consciousness (whatever that might be, at least as much consciousness as I might grant to a cat)." Just because I refer to a robot, and say about it some of the things that functionalists say about robots, I hope that does not make me a functionalist.

I believe that neither an adverbial analysis nor an event analysis nor my reference to robots nor even my quote from Melzack commits me to functionalism. Melzack, as far as I can see, is simply characterizing pain with reference to the body. He points out that we characteristically speak of pain as "throbbing, burning, or sharp," which seem like straightforward descriptions of sensations to me, indicating the "bodily" foundation of even their phenomenological description, and not committed per se to a functionalist analysis of pain statements. Siderits has taken it that Melzack at least would want to adopt a functionalist model, and his quote from part of my reference to Melzack looks as if this is true. But from what Melzack says, I simply do not know if he is a functionalist. As for myself, inasmuch as it is taken as eliminativist with regard to pains, I would not want to adopt a functionalist analysis. [24] But there are some versions of (soft) functionalism that, as I understand it, would not be eliminativist but could rather be taken as explicating, for example, pains (as physicists are not necessarily eliminativist in their analyses of, e.g., boulders), and quite compatible with qualia. As regards such a version of functionalism I remain agnostic. And if Siderits is right that (certain sorts?) of functionalism would entail pain as a conceptual construction, then that would be one major reason for my rejecting those forms of functionalism. Pain as a conceptual construction I take as axiomatically wrong. And I do not believe that Siderits' (very clever) reading of my position and analysis here forces me into that.

Reductionism, Realism, and the Imputing Mind
Finally, a comment on Mark Siderits' clarification of the reductionist position with regard to mountains and their status as conceptual constructs. There is a sense in which this issue, as an issue of the relationship of putative wholes to parts (mereology), is central to a great deal of the Buddhist ontology. If it could be shown that it makes sense to include in one's ontology wholes (as such, or particular examples) as well as parts then the Buddhist ontology would have problems. This is far too large a topic to deal with here. It should be clear that I am a philosophical Realist. I hold inter alia to the existence quite independently of minds of many composite entities, and I certainly hold that it does not follow because something is a composite entity (i.e., an entity with parts) that its existence as a thing is (1) dependent on mind(s), or (2) dependent on social construction or convention. [25] I hold that what it is to be a composite entity depends upon what sort of thing it is, and I am definitely sympathetic to some sort of substance doctrine, where substances are generally mind-independent. The Madhyamaka (Prāsaṅgika) position is that all things, no matter what, are dependent on minds. It seems clear to me that this commits the Mādhyamika to the claim that before there were minds there were no things at all. Siderits'

p. 442 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

reference to parts having mind-independent existence, and his reference to explaining composites through the behavior of their parts is not compatible with (at least Prāsaṅgika) Madhyamaka. Among some further problems with this radical Madhyamaka position would be the following.

(a) If to be is to exist in dependence upon the imputing mind, then whose imputing mind are we talking about? If it is the mind of person X, then all the universe (including ourselves) exists in dependence upon the mind of person X. If I am not X, I do not exist (at least, as an independent center of consciousness). I am just a conceptual imputation by someone else. If I am X, then how is the Madhyamika position to avoid solipsism? And supposing solipsism were avoided, since each thing would exist only relative to each imputer we would each experience different things, and we would thus each live in our own different world with no contact between them. If this reductio is to be avoided, the Madhyamika should be expected to explain, without recourse to something that is not a mental imputation, how we arrive at a common life-world. [26] The only alternative, I suppose, is that to be is to exist in dependence upon a mental imputation, which is that of no mind in particular, or just an abstract mental imputation depending on an abstract mind-in-general. This, too, would entail my own existence simply. as a conceptual imputation, with an additional problem that generally Buddhist thought has been averse to things-in-general, rather than specific, concrete instances.

(b) Actually I (Williams, Siderits, et al.) would have to exist only as a conceptuality in dependence upon the imputing minds of others. I say "of others" because it would be contradictory to claim that I also exist in dependence upon my own imputations. That would mean that I existed (in some sense) prior to the imputations, independently of the imputing mind(s), in order to do the imputing. [27] Thus, I exist only for others and not for myself! If all other sentient beings went out of existence, I would by virtue of that very fact cease to exist myself. And of course the same applies for all others. This would seem to mean that no sentient beings exist.

(c) Moreover, even if we ignore this point, I would have to claim that I am something like the sum total of all visions of myself (including, for the argument, "my own"). I would be a bundle. But then, if someone on the other side of the universe whom I had never met and of whom I had no knowledge-but who had a "vision" of me -- died, I would die (I am the bundle, and the bundle is no longer the same bundle). Even if I said that part of me would die, this seems counterintuitive. How can the continued existence of me as such depend on someone's knowing of me (but not necessarily my knowing of them) on the other side of the universe?

(d) If all things are only conceptualities, then what about (our knowledge of) the conceptualities themselves. Is that only a conceptuality, a construct in dependence upon (my?) imputing mind? In that case, I could not have access to the independent fact that all things are only conceptualities.

And so on, and so on. Even in Siderits' Abhidharma version of reductionism it does not seem to me obvious that the behavior of all composites (things the parts of which we can speak, qua wholes and/or substances) can be explained in terms of the

p. 443 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

lawlike behavior and properties of extremely minute parts. Could this explain all the details of reproduction (a calf from a bull and cow, for example)? Or the behavior of a tree, which responds to the macro-events of the world and its climatic changes, not just in terms of and in response to micro-events? Could it explain love, or altruism itself? I don't know. It certainly seems to me that composites possess properties not possessed by the parts (having the particular shape the composite thing has, or causing our eyes to see red in the case of a tomato, conveying the story of Alice in Wonderland in the case of a book, unable to be broken in the case of a bundle of sticks, and so on). Whether all these can be explained in terms of the behavior of minute irreducible parts not apprehensible by the naked eye, again I do not know. But it does not seem to me to be obviously the case.

I incline once more toward a sortal analysis of wholes, seeing wholes as different sorts of things from parts, and what is to count as a whole in a particular class of cases as requiring further clarification. [28] As I observe in my book, what does follow from seeing wholes as different sorts of things from parts (perhaps as substances) is that there is no immediate ontological absurdity in finding that a whole is neither the same as nor a separate thing on the same level (i.e., another part) from the relevant parts. [29] The reductionist model that would insist that a whole is always only a convenient social construct, reducible in an eliminativist way to the behavior of parts, seems to me to be wrong. Before there were any minds there were not only the atoms (or whatever) that make up mountains. There were mountains. [30] By that, I mean that Mount Fuji was actually there, and there was a true answer to the question "Were there pine trees on Mount Fuji (in year X)?" even if we do not know what that answer was. [31] I do not mean the concept "Mount Fuji," or the concept "pine tree." My question is grammatically well-phrased, and it could not be asking whether the concept "pine tree" plus the concept of plurality applied to it was on the concept "Mount Fuji." Concepts are mind-dependent. It does not follow that their referents necessarily are. I mean the referent Mount Fuji, and so on.

I simply do not hold it at all obvious that before minds were present all the atoms (or whatever the scientific simples currently are, if there are any) of a tree were objectively (mind-independently) there, but not the tree as an objective mind-independent organism. At least there would have to be a very good argument to persuade me that this is a convincing position, abandoning the Realist view that actual mind-independent trees were indeed there to be uprooted by storms many millions of years before the origination of consciousness. [32] I take it that in biological and evolutionary terms there must have been trees, the actual macro-object which corresponds qua trees with the trees we all see and sit under now, long before minds appeared on the scene. To argue that, in order for something that has parts nevertheless to be one thing, the principle of individuation and identity can only be explained through the intervention of the conceptualizing activity of consciousness seems far too weak an argument for such radical repercussions as Constructivism. But Siderits knows as well as I do that these are intensely debated questions in contemporary (and not so contemporary) metaphysics. All we can do here is state our different positions on the battle lines -- and, crucially, indicate that these are difficult

p. 444 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

questions, for which there is at the moment no clear-cut and unproblematic answer accepted by all. [33]

Let me finish at this point. [34] I have found reading and replying to Mark Siderits' review stimulating, challenging, and great fun. Scholarship should be fun, and I am grateful to Siderits for the friendly way in which he has responded to my book. I am not convinced yet by his arguments, as he is not (I imagine) by mine. But we do both agree that these are serious and perennial philosophical issues that require reasoned argument and analysis. I recently heard of a protest at someone pointing out problems in one of Nāgārjuna's arguments: "But Nāgārjuna was a realized Master!" To which we can only reply that even realized Masters (what a ghastly use of the English language!) appear to be capable of some very bad arguments. Arguments, grounds, coherence, and (I would argue) truth stand or fall independent of one's being a realized Master. I have taken it that Śāntideva is putting forward an argument for a conclusion. If he is not, he should have expressed his purposes more clearly and not obscured them with the form of an argument. And my concern, then, becomes anyone who might put forward this argument. Whether the conclusion follows or not is independent of its being put forward by Śāntideva. It could have been uttered by my robot. We owe it to these Masters (Realized or not), and to our own intelligence, to take their arguments seriously. We may do it badly, but there is no avoiding doing serious philosophy. There it should be obvious that Mark Siderits and I are in complete agreement.

Notes
1. Let us be quite clear about the following: (1) As Plato points out in his Theaetetus, there is simply no such thing as something that changes all the time in every respect. Change must be relative to some stability. Some thing changes, and qua that thing it remains stable. If ex hypothesi something changed all the time and in every respect, that thing could not exist. Period. (2) How much something can change before it ceases to be that thing depends on what sort of thing it is; that is, what are the criteria for identity (remaining the same) applied to that sort of thing. There are some sorts of things that by definition undergo a very considerable degree of constant or near-constant change and still remain the same. Take the case of a river. If it did not change constantly ("new waters forever flowing"), it would not be a river. It would be a stagnant pond. Thus, for something to be the same river as we saw upstream or last week cannot by definition require that it be unchanging in terms of the stuff that composes it. One could argue -- I think I would -- that persons are also in this respect like rivers. (3) Thus, before we conclude that change over time entails that something is no longer the same, we need to know what sort of thing it is. Pace Hume (and Buddhists), it is simply false that any thing that changes over time de facto ceases to be the same thing. And once we have recourse to sortals, this should not be thought paradoxical. (4) Again, it is simply false to say that because the sort of thing something is allows it to change over time and still re-

p. 445 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

main the same thing, then the principle of identity possessed by that thing which makes this possible must be nothing more than an ultimately arbitrary social or linguistic convention (i.e., its existence over time as that thing becomes nothing more than "conventional truth"). There is indeed a genuine issue of what provides the principle of identity in the case of different sorts of things. Arguably the principle of identity with regard to (many, if not all) artifacts is a matter of social convention. But, equally arguably, the principle of identity underlying natural kinds (such as trees and human beings) is, for example, some form of holistic homeostatic biological regulation. But all this requires further research once the previous principles have been recognized.

2. People sometimes seem to think that purported cases of "memories" (hypnotically induced or otherwise) of past lives might prove rebirth. To call these "memories" is, of course, to beg the question. They could be any one of a number of alternatives (fantasies, access to a "racial subconscious," demonic possession, subconscious cultural projections, etc., etc.). But even if we could show that these are genuine memories of past lives, they would not prove rebirth. The very most they would prove, even on the most Lockean models for personal identity, is that these particular cases show cases of rebirth. To think that one can draw an induction over the enormous numbers of sentient beings throughout history on the basis of so few cases is astonishing. And to think, as some might suggest, that actually embryos and little babies must really be remembering their past life but, for biological reasons, they are unable to tell us and simply forget all about their previous life when they grow older and are subject to the overwhelming influences of the new life strikes me as an act of faith that it is difficult to see could be proved or refuted. But it doesn't seem very plausible (in my view).

3. I seem to recall a story where Milarepa is said to have projected his consciousness into a sheep, and found the experience of being a sheep a bit depressing. But it still appears to have been Milarepa in the body of a sheep. Of course, this could only make sense with a non-bodily mental criterion for personal identity.

4. Incidentally, in dGe lugs thought the person (gang zag) is the "mere-I," which is a conceptual imputation imputed in dependence upon the aggregates. I discuss this in chapter 5 of my book. This gang zag is thus indeed a conventional entity that exists over and above the aggregates, and is not reducible to them. So the dGe lugs view of the (of course, conventional) person is non-Reductionist, and indeed Tsong kha pa gives arguments against a reductionist model of persons. This topic is treated in a philosophically interesting way by Geshe Thupten Jinpa in his recent and as yet unpublished Cambridge University doctoral thesis.

5. Concerning the Abhidhammāvatāra reference, I am grateful to the summary of this text by Lance Cousins; it will appear in the ninth volume of the Encyclo-

p. 446 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

pedia of Indian Philosophies, ed. Karl H. Potter et al. Lance Cousins has also produced an unpublished translation of (most?) of this work.

6. In other words, when Siderits says, in his endnote 3 to his review of my book, that "Nāgasena would claim that adult and infant cannot be said to be either the same person or different persons because ultimately persons do not exist, that is, because no statement concerning persons can be ultimately true," I am not sure he is right. It seems to me that Nāgasena would claim that the infant and the adult could not be said to be the same or different because we cannot say that sort of thing about a causal continuum. Nevertheless I have to admit that we do in everyday life talk about the infant and the adult as the same person. Whether that can be justified on Buddhist grounds is another issue. But if Siderits substituted "Nāgasena would claim that reborn and deceased cannot be said to be either the same person or different persons because ultimately persons do not exist, that is, because no statement concerning persons can be ultimately true," then I would definitely disagree. Here it seems clear to me that Nāgasena would claim that the deceased and the reborn could not be said to be the same or different because we cannot say that sort of thing about stages in a causal continuum.

7. Is it, though, always? Do we always have to teach a child to anticipate that something done now may lead to pain in the next moment, or some time in the not-too-distant future? Interestingly enough, in the Indian context of belief in rebirth I recall reading (from Bhart.rhari?) of the claim that a baby cries for food because it knows from previous lives the consequences of not eating (see below). It is in this sense "innate." So, perhaps Śāntideva himself would reject some aspects of Siderits' defense.

8. Yes, I am familiar with the arguments of, for example, Kuhn and Feyerabend, which in their extreme form would entail -- or are taken to entail-that there is no objective (conceptual-framework-independent) truth in scientific "discoveries" (or no greater truth than in myths, etc.). I find their case in this respect completely unconvincing. For a critical discussion of this and other issues that is very much being discussed at the time of this writing see Sokal and Bricmont 1998, especially pp. 67-79.

9. Take the case of a spider. When I put my finger in front of it, it runs away. We say that it is frightened. It is certainly responding in a way that in general conduces to self-preservation. We can say that at some level it "knows" that my finger could be a danger to its continued survival. Perhaps more precisely a spider is biologically encoded to avoid danger. And this biological encoding only makes sense if the spider is an enduring entity with a self-identity at least sufficient for self-preservation over time to make biological sense. But this biological encoding is not learned, it is not a product of spider "social constructions," and it is not an ultimately arbitrary structure that spiders superimpose on the world as given to them (cf. what is referred to in Tibetan

p. 447 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

discussions of Madhyamaka as the "innate" [lhan skyes] sense of self). It is part of what it is to be a spider. I suggest they always had it. We call it "instinct," and a biological instinct is surely the exact opposite of a social construction. Although humans have much greater self-awareness, I find it difficult to think that what is an instinct in spiders is an ultimately arbitrary social construction in humans. Of course, in the Buddhist context the spider's "instinct" would be explained with reference to previous lives. In other words, it is open to a Buddhist to claim that what we call an "instinct" is actually the product of social constructions learned previously (would the Buddhist want to say, then, that there is no such thing as "instinct"?). But I'm not sure 1 can make much sense of this. At what point in the past did spiders make this social construction? Or is this instinct for self-preservation learned in a previous incarnation as a member of, for example, a human society where this sort of social construction is possible, and then automatically superimposed on the experiences of spiders? Why should a spider do that (especially if ultimately this is all ontologically arbitrary)? It is, in this rebirth a spider, not a human. It doesn't seem very likely. Anyway, such a response would require an injection of faith into our philosophical and scientific discussion that many Buddhists -- at least modern Buddhists -- might be unhappy with. Perhaps Siderits himself would balk at trying to explain the Buddhist position here by appealing to the Buddhist understanding of the influences of previous lives, since it looks question-begging and merely puts the problem back one stage. And any appeal, maybe, to infinite previous lives and any claim that it is therefore not necessary to explain either conceptually or chronologically how the arbitrary social construction that operates in terms of self-preservation over time came about would seem to me to be unhelpful.

10. And note that for Siderits the very fact that we can do this is simply the result of "an accident of location"! In other words, that I can eat now in order to prevent my pain tomorrow is simply due to the accident that my current psychophysical makeup will have spatiotemporal contiguity (location) with tomorrow's psychophysical makeup. This seems both highly implausible and also is not very Buddhist, since the psychophysical continuum that "1 am" for a Buddhist is not an accident of location but held together very firmly by causal links of the karmic sort. Incidentally, the fact that the Buddhist model relies heavily on karmic causation belies the suggestion that the Buddhist model is close, if not identical, to that of Parfit. I seem to recall that roughly the same point is made in a most interesting and philosophically aware University of Bristol doctoral thesis (unpublished) on Parfit and Buddhism by Nigel Tetley.

11. Note this, because a lot of people seem to think it must do. They seem to think that, in order to be consistent, the Buddhist doctrine of no Self must entail no selfishness.

12. A common Madhyamaka strategy would be to argue that these philosophers and scientists may have got rid of philosophically acquired or learned senses of

p. 448 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

Self but they still have the innate (lhan skyes) sense of Self, and that is why they are still selfish. I am not totally sure what an innate sense of Self is supposed to be. It certainly requires philosophical (and psychological?) unpacking. Whether we all have it (assuming we are unenlightened) or not would seem to need some sort of investigation. Anyway, unless the innate sense of Self is equivalent to the conventional sense of self, that is, to lack an innate sense of Self amounts to lacking even a conventional sense of self, then it does not seem to follow that no Self either as an innate or a learned Self entails altruism. So long as I have a sense of "me," a sense of subjectivity, I can consistently accept no Self in either sense and still privilege myself.

13. In this particular context -- but only in this particular context -- I would also not be averse to a functionalist analysis, inasmuch as functionalism, too, glosses pain statements with reference to the behavior of subjects. My position does not require a choice between these alternative plausible ways of reading pain statements. All my position requires is that pains without subjects are incoherent. But actually and in general I do not hold to a functionalist position with reference to mental events. I shall return to this point later.

14. I am tempted to add "morally impossible," too. But can one morally be selfish? On reflection, I suppose one can morally privilege this person, oneself, which is the point at issue here. For example, if I am attacked I have the right to defend myself. In so doing I privilege my own well-being over that of my attacker. But presumably we would say that in a case like this I am morally right (or at least not morally wrong) to exert my right to defend myself.

15. Siderits seems to find it strange that I should be discussing the reductionist account of persons given that I hold that Śāntideva must here deny the conventional as well as the ultimate existence of persons. But at this point in his argument, for Śāntideva himself the reductionist account would seem to entail that we cannot make sense of the conventional existence of persons either. Śāntideva may be wrong, but that is a different point. In addition Śāntideva does make use of the model that would see the person as a fiction, a mental superimposition upon a psychophysical continuum. That common Buddhist view, whether it involves denying the person only ultimately or conventionally as well, was one that I also wished to examine.

16. And while I would probably be happy with a derivative location for at least some mental states, like pains, with reference to the bodily locations with which they are associated, it is not clear to me that Siderits would welcome this strategy (at least with reference to pains). I doubt he could do it without begging the question.

17. I might add that I really do not think I can make much sense of the idea of maximizing overall utility in abstract. For example, we talk of maximizing happiness for X, Y, or Z where these variables refer to subjects capable of experiencing happiness. Thus I am not sure I can make much sense of the

p. 449 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

conceptual priority of maximizing happiness, which, in general, it just turns out, is facilitated by the construction of persons. (One might even argue, incidentally, that this tendency to speak of the abstract and general as conceptually prior to the particular seems rather un-Buddhist.)

18. Does the very notion of self-scrutiny etc. beg the question by presupposing the concept of subjectivity (via reflexivity) and thus, in the sense in which I am using the term when applied to psychophysical organisms, persons? If the argument is that systems that have a sense of subjectivity are in some sense connected with the generation of the sense of person, then there is a plausible case for question-begging here, too.

19. Siderits objects to my reductio reference to fission, but in context I am discussing a very particular case where in the reductionist model we could not know in advance whether the next moment of "my" continuum may not be a moment of "someone else's" continuum. I am concerned with offering some critical comments on attempts by Buddhists to explain how and why discrete mental and physical events within continua are united into persons on the basis simply of causal connections. Any causal explanation that is going to avoid begging the question I think is going to have to be much more complex than that usually given in Buddhist sources. I do not think I am treating the philosophical issue of fission (discussed by Parfit et al.) as such. Likewise with Siderits' point in his note 5. I am concerned in particular with the continuum of physical events, and the continuum of mental events, and problems involved in relating the two without begging the question. Incidentally, the impossibility of physical events initiating a mental series (i.e., some sort of crossover of the series) is used in some Buddhist sources (common among Tibetans, I recall from Dharmakīrti) as grounds for rebirth (the mental continuum cannot be created or cease, since this would require a physical intervention that is ex hypothesi impossible). Here, at least, the series are distinct in some sense. The Buddhist position is (I think) committed to a form of Cartesian dualism about mind and body. I do mention that causation outside the "material" series is accepted in (e.g., Vaibhāṣika) Abhidharma. My point is to ask for a much clearer explanation of how this can be, without presupposing persons or even the identity of the streams/series. The problem is one of creation of identities out of discrete phenomena linked only by causal connections.

20. If you do not like my sketch, think up one of your own.

21. One would like to apply the Archibald fission example to questions of karmic responsibility. But I preferred to phrase it in terms of justice and legal responsibility, since the problems seem more clear-cut than with karma, where we are operating with an impersonal "cosmic law," and questions of justice are arguably irrelevant. Still, supposing the application to karmic responsibility (1) where fission occurs to Archibald1 ten years after committing the crime, or (2) where fission is a question of two rebirths connected in the right way with

p. 450 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

Archibald1, who would get the unpleasant karmic results? I suppose each would, although each could still claim that it was unjust. Given that two people could not have committed the crime, prima facie someone is receiving the results of acts not committed.

22. Actually I find all this talk of inherent nature, intrinsic nature, and essence very vague. Clearly, if something is dependently originated I have no intention of arguing that it has inherent existence -- or intrinsic existence if that means the same thing. But it is sheer sophistry to think that anyone is necessarily committed to holding that things do have such inherent existence just because they hold that things exist, that is, have all the existence that anyone would normally want to claim for them (including, for a Vaibhāṣika, exist sasvabhāva). And the claim that "people" are committed to holding that things have more existence than that if they hold that things exist sasvabhāva needs proof. It is this argument, and allied topics, that I treat in my long footnote 86.

23. I am treating here pains. In the book I add, in passing and in parentheses, the observation that I doubt that any sensations involve in themselves propositional attitudes. I still hold this, but to be fair (to me!) the application to sensations in general was nevertheless an afterthought. I do say that this generalization requires further thought.

24. Siderits states that "On the one hand there is the reading of 'subjective state' that makes of pain something that necessarily pertains to a subject whose pain-state/event it is. This reading seems to require that pain be identified functionally." I simply do not see the entailment. I argue that there is a necessary relationship between pain and subjects in the sense in which I am using pain, as an irreducible and non-illusory sensation, qualia, what-it-feels-like, and I do not see that I am committed to a functionalist analysis of pain here.

25. It certainly seems quite implausible to maintain that composite entities were constructed by an act of synthesis out of the perception or apprehension in some way of the parts. We (either individually or as a society in its socially constructing phase) never do see the parts that compose, for example, a lump of gold (how big is a lump of gold, anyway?) and then conceptually construct composite objects. As noncomposite simples these parts are invisible to the naked eye. I suppose we could try some sort of construction out of sense data, although the sense data would have to be complete spatiotemporal minima. That is now a very unpopular epistemological position, and anyway I find the notion of complete spatiotemporal minima problematic in the extreme.

26. Analogous arguments could be used, I think, against other forms of relativism, including moral relativism. But the arguments here are merely a sketch.

27. Could it be argued that all we need is this particular moment of "me," imputing me and others? I doubt (to say the least) that this makes much sense, but anyway it would require this particular moment existing prior to imputation. Or a vicious infinite regress.

p. 451 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

28. This might suggest, therefore, that I would also be inclined to see persons as wholes. I like the idea, but I step back from it in the book.

29. Siderits refers to sorites-type arguments concerning where, precisely, Mount Everest begins. But questions of where something begins are not answered in the abstract. There is no such thing as the answer to the question where something begins as such. It all depends on what sort of thing it is. If it is the sort of thing that has an ill-defined border, for example, then one should not expect a clear-cut answer to the question where does it begin. Cf. "Where does the fog begin?" It simply does not follow that because there is no clearly defined answer to the question where something begins, that thing is merely a pragmatic conceptual construct. The fog does not become a pragmatic conceptual construct because its borders are ill-defined. Try driving through it. Is that (the referent) the result of social construction? And this would be the case even if (perhaps for some legal reason) one had to make an arbitrary decision as to where the fog, or Mount Everest, starts. The fog, or the mountain, still would not become a conceptual construct.

30. Some would argue that my choice of a mountain is unhelpful. Mountains are aggregate kinds, reducible to more basic "forms of being" (in this respect they are like artifact kinds). I am not yet convinced, but possibly this may be correct (although note that it would not follow per se that because an aggregate kind can be reduced to more basic forms of being, examples of the kind occurring are mentally, conceptually, or socially created). Perhaps a better example would be a pine tree. Anyway, against the Mādhyamika my position is established if anything has mind-independent existence, and against the Ābhidharmika my position is established if anything for which we can also speak of parts can be shown to have mind-independent existence. At the moment, incidentally, I am inclined toward accepting a neo-Aristotelian analysis in terms of substance kinds (where substance kinds are kinds under which everyday objects fall, like pine trees and human beings in their hacceity). This is a substance ontology that is related to analysis in terms of sortals, avoids problems both in bundle theories and in theories of bare uncharacterized substrata, and which involves some form of Realism with regard to many everyday objects. For a brief introduction to substance kinds and problems with bundle theories (similar though not always identical to the Buddhist position) and bare sub-strata, see the just-published "state of the art" essay by Michael Loux (1998). See also Hoffman and Rosenkrantz 1997.

31. It is quite good to look at these arguments in terms of natural kinds such as trees. It seems clear to me that trees show a biological unity and identity that is independent of conceptual imputation. What it is to be a pine tree (the thing, the referent) is a biological matter, and there is therefore a mind-independent answer to this question. The precise answer might depend on issues of artificially drawing the boundary for Mount Fuji (it might not), but even then it would not follow that the answer was just a pragmatic conceptual imputation.

p. 452 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

32. Is a pine tree on Mount Fuji that is being uprooted by a storm explicable "in terms of facts about the particular atoms; the mountain [and pine tree and winds themselves playing] ... no autonomous explanatory role" (Siderits)?

33. For the sort of argument I would follow, see Devitt 1997, especially chapter 13.

34. At the end of his review Siderits suggests that Śāntideva's "Abhidharma-like" argument in chapter 8 may be superseded in certain respects by the introduction of Madhyamaka in chapter 9. I am not sure this is right. Śāntideva appears to presuppose a Madhyamaka analysis already in Bodhicaryāvatāra 4:47. At least this is the way it is taken by commentators like Prajñākaramati. Either way, it is irrelevant to my argument. My argument rests on the need for Śāntideva to deny persons conventionally as well as ultimately and on the incoherence of placing persons and pains on different ontological levels. Whether I am right or wrong, this argument can be applied to either a Madhyamaka or an Abhidharma-like analysis. It is Siderits who has the problem of an interpretation of Śāntideva in chapter 8, which would be incompatible with the Madhyamaka of chapter 9. And since I have argued that for Śāntideva persons do not exist, I simply do not accept as coherent Siderits' suggestion that "if the realization of emptiness involves the understanding that existing things can only exist in the manner of conceptual fictions, then it is possible that the bodhisattva's practice of helping others may take on a more personal tone -- may, that is, reflect greater attentiveness to the concrete individuality of the persons involved."

References
Buddhadatta. 1980. Buddhadatta's Manuals, Parts I, II. Edited by A. P. Buddhadatta. London: Pall Text Society. (Unpublished translation by L. Cousins.)

Devitt, Michael. 1997. Realism and Truth. Second edition, with a new Afterword. Princeton: Princeton University Press.

Geach, Peter. 1969. God and the Soul. London and Henley: Routledge and Kegan Paul.

Hoffman, Joshua, and Gary S. Rosenkrantz. 1997. Substance: Its Nature and Existence. London and New York: Routledge.

Loux, Michael J. 1998. "Beyond Substrata and Bundles: A Prolegomenon to a Substance Ontology." In Stephen Laurence and Cynthia Macdonald, eds., Contemporary Readings in the Foundations of Metaphysics, chap. 16. Oxford and Malden (Massachusetts): Blackwell.

Milindapañho. 1986. Edited by V. Trenckner. London: Pali Text Society.

Prajñākaramati. Bodhicaryāvatārapañjikā. Edited by P. L. Vaidya. See under Śāntideva.

Śāntideva. 1960. Bodhicaryāvatāra of Śāntideva with the Commentary Pañjikā of Prajñākaramati. Edited by P. L. Vaidya. Darbhanga: Mithila Institute.

p. 453 Response to Mark Siderits' Review Philosophy East and West, Vol. 50, No. 3 (July 2000)

Sokal, Alan, and Jean Bricmont. 1998. Intellectual Impostures: Postmodern Philosophers' Abuse of Science. London: Profile Books.

Watanabe, Fumimaro. 1983. Philosophy and its Development in the Nikāyas and Abhidhamma. Delhi: Motilal Banarsidass.