Collins, Parfit, and the problem of personal identity

·期刊原文
Collins, Parfit, and the problem of personal identity in two philosophical traditions
Matthew Kapstein
A review of:
Selfless Persons.
By Steven Collins
Cambridge: Cambridge University press, 1982. Pp. IX + 323. U.S. $44.50
Reasons and Persons.
By Derek Parfit.
Oxford: Clarendon Press, 1984. Pp. XV + 543. U.S. $29.95
Philosophy East and West
Vol.39 (1989.01 )
pp.289-298
Copyright by University of Hawaii Press

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

Matthew Kapstein is in the Department of Philosophy, Brown University.

Author's Note: I am grateful to Professor James Van Cleve and Professor Michael Zimmerman of the Department of Philosophy, Brown University, for their criticism of an earlier draft of this article

p. 289

The problem of personal identity through time has been much debated in modern Angle-American metaphysics, while, coincidentally, Buddhist Studies has struggled with an analogous problem in its confrontation with the canonical doctrine of anattaa ("no-self," Sanskrit anaatman) and its later formulations in Buddhist commentarial literature. Because the two fields are effectively isolated from each other by the prevailing state of what C. Gudmunsen has aptly termed "East-West philosophical apartheid," [1] they have derived very little indeed from their separate contributions to the study of persons. It is, therefore, refreshing to find two significant new books treating persons and personal identity from the perspectives of philosophy and Buddhist Studies, respectively, that defy convention and intellectual taboo each by nodding in the other's direction. Derek Parfit's Reasons and Persons and Steven Collins' Selfless Persons are concerned not only with personal identity, to be sure, but because it is a central theme in both books and also their point of intersection, it will be the focus of this essay. Reasons and Persons represents the radical elaboration of a theory of persons derived ultimately from Locke, which holds the facts of personal identity to be wholly explicable in terms of continuity and connectedness, while Selfless Persons is a remarkably thorough analysis and interpretation of the themes relating to the concept of persons in Theravaada Buddhist discourse. In what follows, I will first consider them separately, and then offer some reflections on the significance of Parfit's work for Buddhist Studies and on that of Collins for contemporary philosophy.

Reasons and Persons
As his title suggests, Parfit has two major, ultimately intertwined, themes, namely, our reasons for adopting given actions, and our nature as persons, that is, as those who act. The close relation between these themes is stressed in a phrase Parfit (p.336) draws from Rawls, namely, that "the correct regulative principle for anything depends upon the nature of that thing." [2] This is to say in the present context that, if my nature is very different from what I had otherwise thought it to be, I may have to revise my notions of just what I have reason to do. Parfit argues with sparkling philosophical virtuosity and wit that we are indeed in

p. 290

a profound state of error with respect to our nature, so that we must rethink our notions of reasoned action.

Various theories attempt to indicate "what we have most reason to do." Some of these are concerned above all with moral reasons, while others focus upon which courses of possible action are rational. All moral theories agree in holding that we must act morally, whatever else they may say of which particular actions are to be adjudged moral; and all theories of rationality concur that we are to act rationally, however they go about determiningjust what it is rational for us to do. A major target of Parfit's assault, for instance, is the theory of rationality he terms the Self-interest Theory (S). S "gives to each person this aim: the outcomes that would be best for himself and that would make his life go, for him, as well as possible" (p. 3). There are a great many objections to the unqualified assumption of this theory, and the first two parts of Reasons and Persons -- "Self-Defeating Theories" (pp. 3-114) and "Rationality and Time" (pp. 117-195) -- argue in essence that the qualifications which must be introduced in response to these objections amount to abandoning S.

A self-defeating theory is one that fails in its own terms; that is, the results of successfully applying it are that its aims are not well achieved. S is not in this sense directly self-defeating, but it may be indirectly self-defeating, and so must modify its claims. It is indirectly self-defeating, for instance, when it directs us to be not self-denying (if I am in principle self-denying, how can that be best for me?) in those cases when it is in fact worse for us to be so (should I deny myself the pleasure of driving when I'm very drunk?). Here it is indirectly self-defeating because, without modifying S's claims in any way, its injunction that we are to make our lives go as well as possible precludes self-denial; and this has the consequence that S's aims will be very poorly achieved, or perhaps not achieved at all, in cases where self-denial is in fact a condition for attaining the best possible result. But the theory is not directly self-defeating because S can be modified by the claim that to be sometimes self-denying is one of the conditions for my life's going as well as possible. That is, the successful application of S does not guarantee its own failure.

"Rationality and Time" focuses primarily on the peculiar analogy between other persons and myself at other times as the beneficiaries or victims of contemplated actions. Parfit endeavors to show that S involves the conjunction of a bias towards oneself with temporal neutrality -- that is, it avoids a bias towards oneself now -- and that this conjunction involves an arbitrary asymmetry: if temporally neutral, then why not neutral with respect to persons, and if biased with respect to persons, then why not biased with respect to times? One possible defense for such an asymmetry would require making a strong metaphysical assumption, namely, that there is something ultimate about the synchronic differences between persons that is without temporal analogy. That is, a person p1 at time t1 is identical to some person p2, at t2, but is not identical to any other person at time t1. It is the third and longest part of Reasons and Persons,

p.291

"Personal Identity" (pp. 199-347), which seeks to demonstrate that such an assumption is false.

Parfit's contention is that theories of personal identity can be shown to be of two basic types. His logical strategy is then to generate a rich array of puzzles that cannot be satisfactorily resolved by one of these theories, wherefore the other is proven by disjunctive syllogism. The two rival theories are the Reductionist View, which holds that

A person's existence just consists in the existence of a brain and body, and the occurrence of a series of interrelated physical and mental events. (P. 211)

and the Non-Reductionist View, which rejects the reductionist claims that "the fact of a person's identity over time just consists in the holding of certain more particular facts" and that "[t]hese facts can be described in an impersonal way." The most familiar sort of Non-Reductionist View is represented by the Cartesian doctrine of spiritual substance; but there are other versions as well, for example, the Further Fact View, which maintains that "though we are not separately existing entities, personal identity is a further fact, which does not just consist in physical and/or psychological continuity" (p. 210). Parfit holds that we have good reasons for rejecting the Non-Reductionist View, in whatever form, as false.

It will not be possible to review here the many ingenious puzzles Parfit advances in support of his arguments. One of the pleasures of reading Reasons and Persons derives from its abundance of unusual, often empirically impossible, but always entertaining, problem cases, which provide honing-stones for our philosophical intuitions. An example that will win the hearts of the Star Trek generation asks that you consider in what sense, if in any, you are the same person after a teletransporter beams you to Mars. Your terrestrial brain and body are destroyed and an exact replica appears a few minutes later on the Red Planet. The replica has your memories, intentions, habits, and tastes. He will carry out your projects, care for your family, and so on. In short, he is continuous with you in all relevant respects. But have you survived your teletransportation? Parfit argues that "this kind of continuity is just as good as ordinary continuity," and that, on certain criteria of personal identity at least, "my Replica ... would be me" (p. 209).

The Non-Reductionist will object that this account does not do justice to certain key facts about persons, which are only adequately explained by some form of Non-Reductionist View. [3] Among the facts alleged by some philosophers to count against Reductionism are the "strict and philosophical" identity of persons, which is contrasted with a "loose and popular" notion of identity that applies to, for example, your old VW before and after the new transmission was installed; the "unity of consciousness," which is held to demonstrate that there is but one subject of my many perceptions, thoughts, feelings, and so on; and the phenomenon of first person reference, which is sometimes said to require that the

p.292

pronoun "I" have a unique referent, the self, Parfit clearly holds that psychological and/or physical continuity and connectedness are adequate to the facts of personal identity without any notion of strict identity being posited, though those who do not share his intuitions about his puzzles will no doubt take exception to this. [4] His answers to the last two and other related objections (pp. 219-226) are too cursory, I think, to satisfy those who have rigorously argued that the phenomena in question entail some form of Non-Reductionism: Parfit admits that the problem raised by the unity of consciousness might give him trouble, but then tables it for future consideration; and he follows Lichtenberg and Williams in arguing that while "[p]ersons must be mentioned in describing the content of countless thoughts, desires, and other experiences ... such descriptions do not claim that these experiences are had by persons." Though Parfit argues that is possible to describe impersonally our thoughts and experiences, he fails to elaborate a full and systematic account of the proposed redescription. Future debate on this point will have to focus on the question of whether, for example, the sentence "there is depression going on here" really succeeds in conveying the content of "I am depressed." If not, then just what is missing?

In the final analysis, Parfit's version of Reductionism argues for a revision of our moral claims, a revision that includes the decisive rejection of the SelfInterest Theory:

It becomes more plausible, when thinking morally, to focus less upon the person, the subject of experiences, and instead to focus more upon the experiences themselves. It becomes more plausible to claim that, just as we are right to ignore whether people come from the same or different nations, we are right to ignore whether experiences come from the same or different lives. (P. 341)

The last section of Reasons and Persons is entitled "Future Generations" (pp. 351-454). Its intricate web of utilitarian arguments concerning our moral responsibilities to future persons proceeds from the conclusions of the preceding sections, namely, the abandonment of self-interest and the relativization of personal identity. Not only is it bad for me to pollute even if it will affect no one for years to come, but in a crucial sense it is just as bad as my polluting the water I intend to bathe in shortly. Mere temporal and personal distinctions do not ground moral evaluations: Suppose that I shoot an arrow into a distant wood, where it wounds some person. If I should have known that there might be someone in this wood, I am guilty of gross negligence. Because this person is far away, I cannot identify the person whom I harm. But this is no excuse. Nor is it any excuse that this person is far away. We should make the same claims about effects on people who are temprally remote. (P. 357)

Parfit, in sum, has argued that the asymmetry of the Self-Interest Theory must give way to both synchronic and diachronic neutrality with respect to persons.

p.293

Readers familiar with Buddhist moral theory will perhaps greet this with some sense of recognition.

Selfless Persons
Steven Collins sets out to tackle two "classic problems" in this profound and subtle study. First, there is the anattaa doctrine itself, concerning the examination of which he says:

I shall wish to elucidate how it appears in the texts, what it asserts, what it denies, and what it fails to assert or deny; and perhaps most importantly, I shall wish to study what role or roles it plays in the varieties of Buddhist thought and practice, what function or functions it might have for those who profess allegiance to it and whose religious activity is patterned on it. (P. 5)

Second, in addition to this elucidation of anattaa as revealed in the normative teaching of refined Theravaada, he considers "the problem of the relation between the content of Buddhist doctrine as it is found in the scriptural tradition of the pall Canon and the other kinds of religious thought and practice found in what we call 'Buddhist societies'." Though it is the first which primarily concerns us here, the second is not without interest in this context. For Collins shows that although the "no-self" discourse of Theravaada Buddhism is of concern almost exclusively to the learned Buddhist specialist, it is systematically related to the world of the ordinary lay Buddhist through the patterns of imagery which inform Buddhist discourse at all levels.

Collins' work is impressive partly owing to his exceptional command of therelevant primary literature in Pali: his conclusions are supported by impeccable textual authority. But beyond this, Collins evidences the methodological sophistication required to present his conclusions to a contemporary learned readership that may (and I hope, in this case, will) include some non-Buddhologists. The disciplines in which he has chosen to ground his exposition are classical Indology, social anthropology (both in general and with particular reference to specialized work on India and Buddhist South Asia), and philosophy. Dumont and Silburn, Durkheim and Geertz, and Bergson and William James all contribute to the conceptual background of Selfless Persons, with the welcome result that this exploration of the continuities of a Buddhist tradition is quite appropriately expressive of the continuities of the tradition in which its author was trained, and to which he addresses himself.

Throughout Selfless Persons Collins wisely refrains from becoming entangled in any of the questions surrounding the hypothetical reconstruction of "pre-canonical" or "original" Buddhism. The pali Canon and its commentaries aretaken, as they are by the Theravaada tradition itself, as an organic whole. This is not to say, however, that the corpus in question can be considered in pure isolation from its own cultural and historical background. Indeed, Collins' point of departure is "The cultural and social setting of Buddhist thought" (pp. 29-84),

p.294

a setting which includes much of the cosmological landscape of late Vedic Indianthought. The themes rightly scrutinized in detail are the pervasive trio of sa.msaara, karman, and mok.sa (that is, nirvaa.na). It was within a universe delimited by these concepts that the classical Hindu notions of the self (aatman) and their Buddhist rejection took shape:

Just as socially the Buddhist tradition has provided an alternative to the Brahmanical religion of the sacrifice, with its supposed cosmic significance, so, too,psychologically Buddhism has refused to recognise the microcosmic correlate of the sacrifice in Brahmanical thought, the 'self' or 'person' within. The absolute indescribability of nirvaa.na along with its classification as anattaa,'not-self', has helped to keep the separation intact, precisely because of the impossibility of mutual discourse.

It is partially, at least, in the light of such considerations that Collins develops his view "that the denial of self in fact represents a linguistic taboo" (p. 12). Uncharacteristically, Collins waits until halfway through the book to explain to us just what he means by "linguistic taboo" (p. 149, endnote I, which appears on pp. 285-286). There, following F. Steiner, he makes it clear that "[t]aboo is concerned ... with all the social mechanisms of obedience which have ritual significance." There can be little doubt that Collins' analysis of the anattaa doctrine in terms of taboo has much to be said in its favor: the Buddhist virtuoso, after all, does set himself apart from the lay world by affiliating himself with the sa.mgha, from sa.msaara by directing himself to the attainment of nirvaa.na, and from mundane gods by seeking his refuge in the Buddha; and the doctrinal correlate of this particular orientation is anattaa. But it must be recalled that this doctrine derives its power in large measure from the assertion that that which it denies, the Brahmanical soul-theory in its various instantiations, represents a false ontology.

I think that Collins would agree with this, that is, that his concept of a "linguistic taboo" is by no means intended to deprecate Buddhist doctrinal discourse by denying it philosophical content; rather it underscores the fact that such discourse, in addition to that content, has a precise role in Buddhist social and soteriological contexts. That the anattaa doctrine is meant to be taken seriously from a philosophical perspective is clear both because the texts Collins discusses present arguments intended to undermine the soul-theories (pp. 95-103), and because they attempt to elaborate alternative explanations with which to replace those theories (pp. 103-110). An especially valuable contribution which Collins makes to our understanding of such Buddhist alternatives is his analysis of the theory of "bhava^nga-mind" (pp. 238-261), a theory which has many affinities with the "ground-" or "store-consciousness" (aalayavij~naana) developed in Mahaayaana idealism. The bhava^nga here is that which "provides the opportunity for, and links together, a series of mental processes, in one connected and conscious existence" (p. 239). It is evident that the formulators of

p.295

the Theravaada canon were concerned to ground our sense of continuity, while rejecting a substantial self as its basis.

The picture that emerges has the semblance, at least, of Parfit's scheme: a NonReductionist account of persons is denied and the apparent facts of personal identity are redescribed in terms of continuity and connection. What of the implications for Theravaada ethics? Collins explains:

The world of sa.msaara represents, as it were, a four-dimensional throng of 'individualities': some of these happen to be connected with 'him' in a linear temporal series, and so represent past and future 'selves'; some are not thus connected, and so remain for ever 'others'. The crucial point is this: unless he is an omniscient Buddha, or a monk who has acquired the memory of former lives . . . any given individual cannot know which of these are which. [Italics original.] Accordingly, the rationale for action which acceptance of Buddhism furnishes provides neither for simple self-interest nor for self-denying altruism. The attitude to all 'individualities' ... is the same-loving-kindness, compassion, sympathetic joy, and equanimity. (P. 193)

A CRACK IN THE WALL
In the remarks introducing this essay I noted that Parfit and Collins have each to some extent rejected the prevailing segregation of philosophies according to cultural and geographic origins. What I wish to discuss here is just how they have done this and what philosophy and Buddhist Studies might have to gain from further interaction along these lines.

Parfit's nod to Buddhism is as follows:

... I have not considered views held in different ages, or civilizations. This fact suggests a disturbing possibility. I believe that my claims apply to all people, at all times. It would be disturbing to discover that they are merely part of one line of thought, in the culture of Modern Europe and America.
Fortunately, this is not true. I claim that, when we ask what persons are, and how they continue to exist, the fundamental question is a choice between two views. One one view, we are separately existing entities, distinct from our brain and bodies and our experiences, and entities whose existence must be all-or- othing. The other view is the reductionist view. And I claim that, of these, the second view is true.... Buddha would have agreed. [P. 273, italics original. An appendix (pp. 502-503) quotes several short Buddhist texts drawn from Collins and Stcherbatsky in support of this position.]

This is both very good and very bad. It is very bad because even if Parfit is correct to assert that the Buddhist view is substantially similar to Reductionism (I shall ask below whether in fact this is so), this in itself lends very little support to the claim that "my claims apply to all people, at all times." The skeptical response is that this would not be the first time individuals in widely differing circumstances had stumbled on similar false views: compare Dignaaga and Berkeley on the thesis that being consists just in being perceived. But Parfit's brief digression into comparative philosophy is very good insofar as it begins to correct the tendency of some contemporary Western philosophy to argue from

p.296

ignorance of the rest of the world to the supposed uniqueness of Western philosophy and even rationality. In the context of personal identity, this tendency appears, for instance, in Am`elie Rorty's introduction to her valuable anthology, The Identities of Persons:

What are the conditions for the identity of the reflective, conscious subject of experience, a subject that is not identical with any set of its experiences, memories or traits, but is that which has all of them, and can choose either to identify with them or to reject them as alien? Of course from a large world-historical perspective [italics added], this version of the problem of personal identity is one that could only appear under very special social and intellectual conditions....
[L]et us sketch the historical conditions that gave rise to the view of the person as the "I" of reflective consciousness, owner and disowner of its experiences, memories, attributes, attitudes. The philosophic conditions: the movement from Descartes's reflective "I" to Locke's substantial center of conscious experience, to Hume's theater of the sequence of impressions and ideas, to Kant's transcen dental unity of apperception and the metaphysical postulate of a simple soul, to Sartre's and Heidegger's analyses of consciousness as the quest for its own definition in the face of its non-Being. The social conditions: the movement from the Reformation to radical individualism. The cultural conditions: Romanticism and the novel of first person sensibility. [5]

Are the historical conditions for the appearance of the view in question really so specific as Rorty maintains them to be? Consider these observations of Collins on the Brahmanical and Buddhist traditions of India:

This discrimination [of Saa^mkhya] refers both to the distinguishing of the different elements within one person, and also to the project of separating, both in theory and in practice, the essential self or person from the composite psycho-physical personality as a whole. (P. 80)

... according to the strictest doctrinal accounts of [Buddhist] psychology, the unenlightened man must by definition have a sense of' I', must find a 'person' within when he introspects: this sense of self, manifested simply as a reaction pattern, is called asmi-maana, 'the conceit "I am"'. (P. 153)

It is worth noting that classical Indian philosophical speculation on the soul, which generated close analogies to all of the major theories of the self elaborated in seventeenth- and eighteenth-century Europe, did so in the absence of any ideological or cultural counterpart to Western individualism. [6] When we in fact (and not merely in pretense) adopt a "large world-historical perspective," we are led to conclude that the problem of personal identity should not be a priori held to be uniquely a modern Western problem, and that any attempt to resolve it solely by diagnosing the historical conditions for its appearance in the West will therefore be incomplete. [7] A genuine universal solution, however, applicable to all human reflection on the self in all societies, at all times, remains still beyond our reach and, I am afraid, beyond that of Parfit.

Collins himself suggests that the real value of his work may be as a corrective to the inbred perspective of the English-speaking philosophical tribe. He is particularly concerned to challenge the usually unspoken assumption that

p.297

... 'the intuitions of the native English thinker' should be the arbiter of philosophical correctness, and that it is the conceptual and linguistic habits of 'common-sense' to which we should look for enlightenment on philosophical issues. The approach I am suggesting, on the contrary, will see these 'intuitions', and the 'common-sense' constructed out of them, as mere problematic data.. .. (P. 3)

As I have indicated in the preceding, the discovery that others have expressed similar "intuitions" may be no less problematic in certain respects than the discovery of crucial dissimilarities. In either case, the refusal to acknowledge that there are philosophically reflective civilizations outside the West should have no place in contemporary scholarship. Selfless Persons brings one such civilization to the philosopher's doorstep: only the misologist will ignore its presence.

It remains to say something of the importance of Parfit's work for the study of Buddhism. As Collins relates in his introduction, not just a few modern authorities on Buddhism have found the anattaa doctrine to be impossible to accept at face value. The instances Collins adduces are mostly variations on the "big Self-little self" argument, that is, the textually unfounded view that the anattaa doctrine negates the empirical ego only to affim the transcendental Self Supreme. But there is another tack some have adopted in endeavoring to find some sort of stable soul underlying Buddhist thought, namely, to suggest that without such an entity the Buddhist teachings of karman and rebirth become simply incoherent. Parfit would agree (Reasons and Persons, pp. 227-228). Students of Buddhism ought to take the philosophical arguments involved here very seriously. The remarks which follow provide merely a sketch.

Parfit holds that something like the Cartesian Pure Ego theory might have been true, and that we might have had good evidence for preferring that theory to its rivals. Such evidence would consist in reliable memories of past lives. Now most Buddhists do believe that there have in fact been cases in which an individual truly remembers his past lives. Even if they believe this to have been true of no one else, they believe that the Buddha `Saakyamuni had just such memories as he sat beneath the Bodhi Tree. Should they not then affirm something like the Cartesian Pure Ego theory? The fact that they do not leaves us with just these possibilities: Buddhist belief, without persisting souls, is incoherent; Parfit is wrong to argue that evidence of rebirth would support NonReductionism, for adequate Reductionist explanations can be given; Parfit is wrong to argue that Non-Reductionism and Reductionsim are the two exclusive alternatives, because there is a tertium quid. The first possibility would be a bad break for Buddhism, and the last almost as bad for Parfit. The second would be good for both. [8] To decide which is philosophically to be preferred and to determine just how later Buddhist philosophers resolved these and related difficulties leaves us with two interesting challenges. Both Buddhist Studies and philosophy will profit from the effort to meet them.

p.298

NOTES
1.Chris Gudmunsen, Wittgenstein and Buddhism (London: Macmillan, 1977), p. viii.

2. John Rawls, A Theory of Justice (Cambridge, Massachusetts: Harvard University Press, 1971), p. 29.

3. An excellent introduction to the historical background of the contemporary philosophical debate between Reductionism and Non-Reductionism is provided by John Perry, ed., Personal Identity (Berkeley, California: University of California Press, 1975). For very well argued defenses of the Non-Reductionist position one can do no better than consult Roderick M. Chisholm, Person and Object (La Salle, Illinois: Open Court Publishing Company, 1976), and The First Person (Min-neapolis, Minnesota: University of Minnesota Press, 1981).

4. Cf., in particular, R. M. Chisholm, Person and Object, chap. 3, "Identity through Time."

5. Amelie Oksenberg Porty, ed., The Identities of Persons (Berkeley, California: University of California Press, 1976), p. 11. My misgivings about this one passage should not be construed as reflecting an overall view of the excellent essay in which it occurs.

6. The diversity of classical Indian speculations on the self may be gathered from Troy Wilson Organ, The Self in Indian Philosophy (The Hague: Mouton & Co., 1964); and Michel Hulin, Le Principe de l'Ego dans la Pens`ee Indienne Classique (Paris: Institut de Civilisation Indienne, 1978). A succinct discussion of the significance of the absence of individualism for Sanskrit literature will be found in Daniel H. H. Ingalls, Sanskrit Poetry from Vidyaakara's "Treasury"(Cambridge, Massachusetts: Harvard University Press, 1968), pp. 26-31.

7. It should be noted that Rorty herself never suggests that an historical diagnosis would constitute a solution to the problem of personal identity. My remarks here are addressed to those who, beginning with historical assumptions similar to the ones put forward by Rorty, would argue for a more radically historicist position with respect to the investigation of philosophical problems.

8. Collins' account in his Part III, "Personality and rebirth," suggests that Theravaada theorists would be strongly inclined to support this second alternative. However, the tradition reveals some noteworthy inconsistencies: the description of the being who is about to be reborn as entering the womb in the form of a mythical gandhabba-spirit (pp. 210-213) is clearly Non-Reductionist; and the theory of bhava^nga-mind, referred to above, may well be seen as appealing to a tertium quid.