CRITICAL STUDY

Thinking Clearly About Death, by Jay F. Rosenberg. Englewood Cliffs: Prentice-Hall, 1983. xv + 235 pp. H.E. BABER and JOHN DONNELLY

Some sixteen years ago, Robert Olson in his article "Death" in

The Encyclopedia of Philosophy wrote (rather accurately) that "We are likely to discover more about the topic [of death] in the writings of men of letters than in those of technical philosophers." Since then philosophical writing has taken a decided thanatological turn, and one would be hard pressed to find a more outstanding contribution to this field than Rosenberg's study. Rosenberg characterizes his book, rightly, as a work of "hard-core analytic philosophy." He is, however, at pains to make it accessible to thoughtful readers without any previous philosophical training and in this project, we believe, he has been largely successful. His prose is clear, compact and hvely; he eschews logical symbolism and scrupulously avoids the use of unnecessary technical terms. In addition, Rosenberg prefaces his discussion of the metaphysics and ethics of death with an account of the analytic method in philosophy which illuminates the material that follows and stands on its own as an excellent mini-essay in meta-philosophy. I

Rosenberg's discussion of the metaphysics of death is especially interesting insofar as it incorporates a great deal of material that has appeared piecemeal in journal articles over the past twenty years as well as some original material, into a unified theory of the self incorporating both a theory of mind and an account of personal identity. The theory of mind is a modified "person theory" reminiscent of Strawson. On Rosenberg's account, persons are subjects of both "banal properties," b-properties, which they share with inanimate objects, and of certain "'special properties," s-properties, which have to do with life, motility, sentience and rationality. The same entity 79

CRITICAL STUDIES is the subject of both b- and s-properties. Indeed, Rosenberg argues, there could not be any entity that had only s-properties: s-properties are "powers" and nothing can have such dispositional properties without having some non-dispositional properties as well; thus there can be no Cartesian soul. As for his account of personal identity, Rosenberg defends a version of the spatio-temporal continuity criterion which seems to have been inspired by Wiggins. To trace a person "through time" we trace that part of the organism which is causally responsible for the continuity of those functions which consitute his "persona" cognition, perception, desire, memory, volition and the like. "We cannot pry the notion of a person loose from the notion of a living organism," writes Rosenberg. "The significant corollary of this result is that we cannot make coherent sense of the supposed possibility that a person's history might continue beyond that person's death. ...[W] e cannot even describe such a supposed happening without misusing language, mistaking linguistic appearances for linguistic realities." ( 9 6 - 9 7 ) Some readers will no doubt object to this "hard-core analytic" approach to the mind/body problem and the problem of personal identity; we however do not wish to take issue with Rosenberg's sectarian stance and method. Nevertheless we find some of his substantive claims unacceptable and, therefore, continue to look for the resurrection of the dead and the life of the world to come. II First, we suggest that Rosenberg relies too heavily on what might be called the "paraphrase test" for sorting out genuine things from merely "nominal" ones. Rosenberg argues at length that a variety of puzzles in philosophy of mind are the product of linguistic confusion. In particular, he suggests, we are inclined to assume wrongly that nouns invariably refer to things, that verbs always pick out events, and that the only function of adjectives is to assign conditions or attributes to things, which assumption leads us to posit bogus things, pseudoevents, and attributes that are, at best, queer. Among these merely "nominal" things, events, and attributes, Rosenberg suggests are the soul or mind, the supposed momentary event of dying as distinct from the process of dying, and the condition of being dead. Naturally, some things, events, and conditions are not merely 80

CRITICAL STUDIES

nominal. How then are we to distinguish them from merely nominal things, events, and conditions? Rosenberg suggests that merely nominal things, events and conditions can be "paraphrased away" while the genuine articles cannot. That is to say, we test putative entities in the following way: if a sentence which seems to commit us to the existence of a certain object admits of an adequate paraphrase which carries no such commitment then we shall conclude that the putative object is merely nominal. If no such paraphrase is available, then we shall consider the object in question genuine. For example, (1)

Aunt Ethelis dead.

can be paraphrased as (1')

Aunt Ethelhas died.

Given the availability of this paraphrase, Rosenberg concludes that in asserting the former sentence we are not ascribing a current condition to Aunt Ethel, but rather alluding to an event which occurred to her in the past. Similarly to say (2)

She has a generous soul.

is to say (2')

She is generous.

Thus, he suggests, souls, at least insofar as they are mentioned insentences like (2), are merely nominal objects. On the other hand, some entities cannot be paraphrased away in this manner. We cannot, for example, paraphrase (3)

He is overly belivered.

(4)

She has a red Pontiac.

or

Hence, Rosenberg suggests, we are forced to conclude that livers and Pontiacs are not merely nominal. The trouble is that for all his examples, Rosenberg does not provide the reader with any account of what constitutes an acceptable paraphrase of a given sentence. Why, for example, should we not paraphrase (3) and (4) as (3')

He is overly believered.

and (4')

She Pontiacdrives redly. 81

CRITICAL STUDIES respectively? Rosenberg rules out any sentences which, like (3) and (4), purport to paraphrase away livers and automobiles on the grounds that "an enlarged liver and a red Pontiac are things in their own right, with histories o f their own, which, for better or for worse, a person can own or possess." (39) This, however, will not do. The paraphrase test is supposed to enable us to determine precisely whether putative entities are "things in their own right," that is, genuine rather than merely nominal objects. To rule out a proposed paraphrase on the grounds that it does away with objects that are not merely nominal is to beg just the question that the paraphrase test was supposed to settle. Of course the obvious reason for allowing (1 ') and (2') while rejecting (3') and (4') is that the former are good ordinary English while the latter are not; so we might say that, in general, a sentence is an acceptable paraphrase o f an English sentence only if it is itself a good ordinary English sentence. Even if it is not always clear whether a sentence is good ordinary English, it is clear that some sentences, like (1') and (2'), are, and that others, like (3') and (4'), are not. Given this account of paraphrase, however, the paraphrase test yields counterintuitive results in a variety of cases. Consider the following sentences and their paraphrases: (5) (5') (6) (6')

He has blonde hair. He is blonde. She has large breasts. She is well-endowed.

(5') and (6') are perfectly good, ordinary English and adequately convey the meanings of (5) and (6) respectively. Thus it would seem that both hair and breasts fail the paraphrase test and so must be regarded as merely nominal objects. We, however, do not think that they are. Rosenberg's case against dualism does not rest on any crude or naive use of the paraphrase test to eliminate the soul and other objectionable entities. Nevertheless, ff Rosenberg wants to use the paraphrase test as a procedure for sorting out genuine things, events, and conditions from merely nominal ones, as he does, it would seem incumbent upon him to give us some account of what constitutes an adequate paraphrase.

82

CRITICAL STUDIES

III Rosenberg's arguments against dualism, even if successful, do not of course settle the question of whether persons can survive death. The mind/body problem and the problem of personal identity (to cite Perry's maxim), though related, should not be conflated. Indeed, several contemporary philosophers have suggested that some mentalistic accounts of personal identity which allow for survival are compatible with materialistic theories of mind. Rosenberg, however, rejects mentalistic criteria for personal identity, concluding that "we cannot make coherent sense of the supposed possibility that a person's history might continue beyond that person's death." (96) His arguments against the mentalistic theories which embody memory criteria for personal identity are, however, less than persuasive. As is customary in discussions of the memory criterion, Rosenberg distinguishes between genuine or actual remembering "which connotes success or correctness - and (merely) seeming to remember ("S-remembering") - w h i c h comments on the character of one's experience without implying correctness or SU c c e ss.

9 Someone actually remembers something just in case he or she s-remembers it and is right." (78) Now the memory theorist must hold either that actual remembering is criteriat for personal identity or that S-remembering is, and Rosenberg argues that neither version of the memory criterion is acceptable. Rosenberg rejects actual memory as a personal identity criterion on the grounds that it is unsatisfactory epistemically. Using the actual memory criterion we can never know for certain that A is the same person as B unless we establish that A actually remembers (or can actually remember) an action or exoerience of B's. But we cannot know for certain that A actually remembers an action or experience of B's unless we can determine that A is the same person as B: hence the memory criterion is sometimes said to be "circular." Thus, actual remembering cannot be an independent criterion for personal identity since, as Butler put it, it "presupposes" personal identity. Rosenberg argues that S-memory cannot be the criterion for personal identity since the relation of identity is necessarily one-toone while the relation which obtains between S-rememberers and the subjects of S-remembered experiences and actions is not. Consequently, if S-memory were criterial for personal identity we should 83

CRITICAL STUDIES be required to read some puzzle cases - those commonly known as "fission" or "branching" cases - as situations in which two different persons are each identical to the same person. 9 "there can be many lunatics, all o f whom seem to remember having lived through the events o f Napoleon's life. If Smemory-connectedness were sufficient to make any o f these poor wretches identical to Napoleon then, it would be sufficient to make all o f them identical to Napoleon - and thus to make each of them identical to (that is, one and the same person as) all o f the others, a conclusion which is simply incoherent. . . . No relation which multiplies entities can preserve identity through time." (80) Thus, Rosenberg argues, neither version o f the memory criterion is acceptable: the actual-memory version does not provide us with a conclusive test for the correctness o f oersonal identity claims and the S-memory criterion yields logically incoherent results. We suggest, however, that the mentalists may grasp either horn of this supposed dilemma since, we shall argue, both versions of the memory criterion are acceptable. It is true that, for practical pruposes, actual memory is not a useful identity test. People can be mistaken about the genuineness o f their own S-memories as well as those o f others. Once we establish that an S-memory is genuine, we have indeed established personal identity, but establishing that an S-memory is genuine in the first place, without 'presupposing" personal identity, presents difficulties. This is not however to say that it is, as Rosenberg suggests, impossible. Perry, for example, in "Personal Identity, Memory and the Problem o f Circularity" argues that given the proper account o f remembering it is possible in principle - though not in practice, given the primitiveness o f current neurophysiological techniques to determine when genuine remembering occurs without making any illegitimate assumptions about the identity o f the rememberer. But even if Perry is incorrect and the memory criterion is in principle epistemically unsatisfactory, it does not follow that we must reject it in favor o f some bodily criterion which precludes survival. It may be that personal identity must be taken as primitive and unanalysable, either in terms o f m e m o r y o r in terms of bodily continuity. Butler seems to have been suggesting this when he rejected Locke's memory criterion as "circular" - not that we reject the 84

CRITICAL STUDIES memory criterion in favor of some bodily criterion for personal identity. Furthermore it is not clear that the fission objection rules out Smemory or any other mentalistic criterion for personal identity. This objection to mentalistic accounts of personal identity has been discussed extensively in recent years and it is unfortunate that Rosenberg does not mention the variety of responses available to the mentalist. As a preliminary, we must point out that the spectre of fission should worry Rosenberg no less than the mentalist. Rosenberg suggests that personal identity "tracks" with sameness of brain "because there is a logical or conceptual connection between the concept of a person and the concept o f an entity having a persona, and because we have discovered, not in point of logic but as a matter of fact, that the brain is the organ of those functions which constitute a person's persona." (94) As a matter o f fact information is reduplicated in the brain and we can easily imagine dividing a brain and transplanting the resulting portions, each o f which carried the putative memories, capacities, desires, beliefs and character traits which constitute the original subject's "persona" into two different bodies to yield, it would seem, two different persons. If, as Rosenberg suggests, personal identity tracks with whatever physical structure is causally responsible for an individual's persona, then such versions of the fission case should give him pause. Mentalists, however, do not regard the logical difficulties posed by such apparently "branching" cases of identity as insurmountable. There are at least three strategies for coping with such cases. The first way is simply to deny that what matters for survival is identity: what matters is "psychological connectedness" between "person-stages," temporal slices of the "histories" or "careers" of persons, which may be a one-many relation. On this account, the relation by which we trace persons through time is not, strictly speaking, an identity relation at all. Where fission occurs, one object does not literally become two. Rather the path of the relation o f psychological connectedness branches so that one person-stage may be psychologically connected to two person-stages which are not psychologically connected to one another. Such a response has been developed by Derek Parfit. The second way is to deny that the relation by which we count persons at any given time is, strictly speaking, the relation of 85

CRITICAL STUDIES personal identity. This sort of response has been developed by John Perry, David Lewis and others. On this account, at any given time that we take the census, what we count are person-stages that occur at that time. Normally, at any given time there will be the same number of person-stages as there are persons at that time. In certain bizarre circumstances however this will not be the case. Where fission occurs the histories of different persons partially overlap so that, at some times two or more different persons share one personstage. At such times we shall count them as one. We may describe such a situation as a case in which one person has become two but this is not strictly correct: in cases o f fission there are really two or more persons before branching as well as afterwards. Finally, the most radical way with the fission objection is to modify the traditional account of identity in order to accommodate cases in which one object becomes two. Prior, for example, has suggested that in light o f such cases we might be forced to modify or abandon Leibniz's Law. It is disputed, of course, whether any of these responses to the problem of branching is acceptable. Nevertheless, in light of the extensive attention such strategies have received in the literature, it will not do simply to cite the problem of branching as a conclusive reason to reject any account of personal identity which allows for the possibility of survival, as Rosenberg does, particularly since branching is a threat to accounts of personal identity which preclude survival also. IV Rosenberg claims that a typical human adult has no positive, active rights (i.e. a right to have certain things done for him when exercised by the bearer, a right to obtain a service or benefit)vis-avis another specific individual. That is, unlike the rights of children, no one is morally obligated to feed, shelter or provide medical care for an adult person. I do, according to Rosenberg, have certain negative, passive rights not to have certain things done to me, such as the respect accorded me by others not torturing, enslaving, or killing me, etc. Granted we have correlative moral obligations to respect other persons' negative, passive rights. Indeed Rosenberg maintains the stronger claim that we have "a responsibility to see to i t . . . that no one else" (130) violates such negative, passive rights. Inasmuch as Rosenberg admits such a stringent duty of enforcement here, we are puzzled why he will not also acknowledge a moral obligation of per86

CRITICAL STUDIES formance and not mere forbearance to respect a person's positive, active rights, say, to various forms of medical care, to die with dignity, etc. What is the point of respect inherent in the concept of passive rights if not the enhancing of the right-holder's freedom of action, which Rosenberg sometimes terms an active right? We find the distinction drawn by Rosenberg between active and passive rights to be quite nebulous. Rosenberg's phenomenological analysis of our "implicit moral practice" reveals our adherence to two normatively primitive intrinsic values, namely, the exercise of moral autonomy and freedom from suffering. Rosenberg's implicit ethic reveals that life itself (i.e. "the self-sustaining syntropic powers of an entity to preserve and increase its structural organization," (127)) is not intrinsically worthwhile per se, but only a causally necessary condition of rationality and sentience and what is intrinsically valuable. Unlike non-human animals (animal liberationists will find much to fault in the book), human beings are potential sources of intrinsic worth and as a result have negative, passive rights to protection from any imposed constraints on such autonomy. This provides the theoretical support for our prima-facie, absolute right to life, which remains defeasible in light of considerations of intrinsic disvalue (i.e. suffering). Since for Rosenberg, no case of refraining could logically violate a negative, passive right, refusals to provide dialysis treatment or insulin for otherwise well-functioning persons are forbidden only on the grounds that such behavior violates their exercise of moral autonomy. The specific moral agents who are to countermand any possible refusals is not made clear by Rosenberg, and would seem to present the same sort of difficulty of assigning moral agent responsibility that led him to reject positive (passive or active) rights. That is, a curious consequence of Rosenberg's "implicit moral theory" is that no one has to acknowledge any person's positive, passive absolute rights. No moral agent can be picked out as the right-ower, and the appeal to "society in general" gambit will not suffice. Yet Rosenberg apparently finds no great difficulty in selecting rightowers for protecting and enforcing our negative passive rights. Any "implicit moral theory" that does not acknowledge any positive, passive absolute rights seems suspect. Suppose (1) Smith refuses Jones his dialysis treatment or some similar, non-scarce fife-supporting technology. We normally believe that (2) society is morally obligated to supply everyone, where possible, with the minimum essential for life. As a result, we would normally conclude that (3) Smith has violated Jones' alleged 87

CRITICAL STUDIES positive, passive absolute right to the essentials of life. Rosenberg does not concur. He wants to claim that there are no collective, societal obligations here involved, because there is no moral agent who is individually obligated. Moreover, Rosenberg finds these putative positive, passive rights "useless" to morality. However, apart from the fact that our implicit morality, pace Rosenberg, does acknowledge the obligation to respect such rights, if Smith is an executive at Medicare, then (3) does follow. Rosenberg does allow that matters are different with so-called conditional positive, passive rights, where the correlative obligation is not a matter of simple natural kind membership, but instead involves a special relationship to other specifiable individuals. Unfortunately Rosenberg is fuzzy on the exact demarcation between negative, passive rights and positive, passive rights, seeming to list on page 151 the right to dialysis treatment under both categories. At times Rosenberg provides the needed corrective against the conventional wisdom of rights-advocacy (that often is oblivious to the claims of duty, the role of the virtues, and various goal-based considerations in moral experience) but all too often, like Bentham who thought that rights involved "nonsense on stitls," Rosenberg imperils many fights which our implicit morality acknowledges. Granted it is often difficult to specify an individual right-ower, but it is not impossible. The deontic buck stops somewhere. Rosenberg's implicit moral theory provides only rights against not rights to. He makes no allowances for any positive rights that serve as claims on society or the state for specific assistance or benefit (e.g. some of the welfare rights listed in the 1948 Universal Declaration of Human Rights). Rosenberg may have a point in insisting that all rights stem ultimately from our implicit (or explicit) moral theory. (Rosenberg's theory does allow for some non-derivative obligations that do the job normally covered by positive rights, but this only raises the quagmire of epistemic priority.) However, in his phenomenology of moral practice, Rosenberg has drawn only a skeleton-depiction of a moral theorythat would be needed to justify a theory of rights. Rosenberg's account of moral experience at best recognizes positive rights as ideals and not actual entitlements. To be sure some rights operate like, that, but not all do, such as the right to a minimum standard of food and shelter. It may well be that not all alleged rights entail obligation to provide a benefit, but some do (of course, I can have a right to do x, but it hardly follows that x is always the right thing to do). Rights as ideals do not yield a COl-re88

CRITICAL STUDIES

sponding duty of perfect obligation, but rather a correlative duty of imperfect obligation (e.g. the duties of beneficence). However,pace Rosenberg, we are at times non-conditionally obligated not just to do no harm, but instead to positively prevent or remove any harm. If so, then not all positive rights would require supererogatory conduct. V Like so many philosophers (and non-philosophers) writing about thanatology, Rosenberg conflates the logic of suicide and euthanasia. His paradigm of Darcy serves as a case in point. Darcy is in her sixties, suffering from acute and progressive bone cancer. The chances of radiation and chemotherapy putting the cancer into remission are slim, although such treatments will in all likelihood postpone her inevitable (and otherwise temporally imminent)death by several years. Darcy refuses the prescribed oncological treatment program except for the use of analgesics and pain controlling narcotics. And when the latter prove basically ineffectual, given her diminished capacities, she plans to commit suicide. Rosenberg is a highly skilled analytical philosopher, yet he subscribes herein to a definition of suicide as voluntary self-annihilation that is woefully nebulous and open-textured. Contra Rosenberg, suicide has to do with a person's willing that her life be preternaturally extinguished, and excludes from its purview cases of heroic selfsacrifice where a person can preserve her life only by performing a morally dishonorable act, as well as cases of euthanasia involving a medically diagnosed terminal illness. To invent plausible, rational scenarios involving heroic self-sacrifice or euthanasia, and then use these as counter-examples against the traditional moral censure of suicide is an egregious non-sequitur. In euthanasia, it is a patient's condition which is hopeless; in suicide, it is the suicidist who is (temporally) hopeless and not her condition. Rosenberg does have some interesting things to say about paternalistic intervention, but again such remarks don't touch the fundamental issue of suicide per se. In his case of the unfortunate Elbert who "freely" wills himself into perpetual servitude - only to (we assume) later regret it Rosenberg argues we have good moral reason to prevent this situation from happening (if not on slippery-slope considerations alone) because of Elbert's desperate state and the strong likelihood of eventual remorse over his "decision." Elbert is obviously not deliberating in Butler's cool hour. 89

CRITICAL STUDIES Rosenberg writes: "Our confidence in the practical inevitability o f such a future reversal of outlook is an empirical confidence. It is warranted, that is, by our collective experiences of the actual attitudes o f enslaved persons and others in situations relevantly similar to Elbert's. But such experiences are logically precluded in the case of suicide. We can in principle have no experiences of the thoughts and feelings of persons who have taken their own lives, for such persons have ipso facto ceased to exist as thinking and feeling entities." (186) We suggest, contra Rosenberg, that we are not so logically precluded in the case of suicide. There are numerous case studies of suicide prevention that offer elaborate factual evidence that attempted-butfailed suicides come to regret what they attempted to do. If Darcy were a bona fide but thwarted suicide, she might too. Rosenberg says that not all do. That may be true, but it doesn't detract from the fact that most do so regret their actions. Analogously, not all voluntary enslavement cases (which are rare today) testify to subsequent regret. We are thinking here o f individuals who make unconditional religious vows of poverty, obedience, and celibacy (seemingly surrendering their autonomy). Some come to regret such vocations, but many do not. To cite, as Rosenberg does, cases of terminallyill patients who wish they had terminated their lives before the last agonizing stage of their disease, is again to mistakenly conflate euthanasia with suicide. Rosenberg agrees with Darcy's "rational" decision to commit suicide. But how are we to locate a subject who as a result of the suicide has the first-person psychological state of being "less worse off"? Given a successful suicide, Darcy becomes a cadaver. And if she avers that she would be better off (qua cadaver) killing herself than continuing to live, who is the subject of the state of being better off?. Rosenberg says her choice is between a dreadful future and no future at all. But we wonder how it is in her self-interest to pick no future? VI We conclude with some comments on Rosenberg's epicurean analysis o f death. Rosenberg claims it is not epistemically appropriate to fear or dread one's own death considered as a non-historical event. Goven his materialistic metaphysical outlook, there is no phenomenal content assignable to death in the first person, so that 90

CRITICAL STUDIES death is not an evil (or good) to the entity who has died. Rosenberg does maintain that one can "rationally anticipate, plan, and provide for (it) in prospect" (197). In short, thanatophobia (or thanatophiha) is logically misplaced as a first-person attitude toward my own death. The concept of death serves as a limit, an ordering function to guide our philosophical inquiries. Just as we know from thermodynamics that no body can have a temperature of absolute zero yet employ the concept of zero temperature, so too we can employ the concept o f death while claiming that there are no experienceless selves. Rosenberg may well believe that his epicurean attitude toward death involves no ataraxia, but we beg to differ. Like Epicurus, Rosenberg wants to claim that death is nothing to a reasonable person. But what is the sense o f "nothing"? Nothing may mean that which is insignificant, of no value or importance to a reasonable person, as in the sentence 'nothing you say interests me'. Or nothing may denote a nonexistent state of affairs, as in the locution 'there is nothing in the room'. Rosenberg's claim may be true (given his metaphysical presuppositions) in the latter sense, but it is false in the former usage. Unlike Epicurus who wrote in his Letter to Menoeceus that "there is nothing terrible in life for the man who has truly comprehended that there is nothing terrible in not living," Rosenberg prudently recognizes that life contains its trials and tragedies, so that even if thanatophobia could be eradicated, few if any persons could sustain a life of only catastematic and kinetic pleasures. But like Epicurus, Rosenberg does seem to think that the fear o f that which is not bad for one is groundless. But such a belief is false. For example, theists justifiably fear the wilt of God yet it is not evil, and so too do most people understandably fear an upcoming operation that is designed to cure or alleviate their medical condition. Could it not reasonably be contended that if there are objective truths about evil states o f affairs (and Rosenberg concedes that a person's death may be an objective evil for others), that there are then evil states of affairs simpliciter involving death. Also, if it is rational to fear a premature death, then it would seem to be also rational to fear death per se. A premature death is just as natural as a mature death. Indeed some philosophers (e.g. Harry Silverstein) have even suggested (in brazen ontological fashion) that posthumous events exist timelessly, so that death could still be regarded as an evil in the first-person, as a person could now experience states of affairs 91

CRITICAL STUDIES that begin after his death as they akeady exist atemporally. Perhaps less-contentious metaphysically, we might want to hold that a person can partially experience a state o f affairs at some time only if it begins to partially occur before that person's death. The loss of reputation, the breaking of promises, vile rumors and falsehoods could be correctly sensed as developing by the dying person, and their full instantation as evil occur when that person dies. (Tolstoy offers a powerful portrayal o f this scenario in his The Death oflvan Ilych.) That is, such a person has partially experienced posthumous evil states o f affairs as objects of his current psychological attitudes. We would note that it certainly seems reasonable to fear a similar situation in life, say when senility sets in and our inability to offer a counter-explanation to such allegations is apparent. Also, there are medically-documented cases of "brain-dead" women who have given birth to children, so that we arguably have here, contra Rosenberg, a subject to attribute the evil inherent in death to, as cadavers do not give birth. The death of these women robbed them of their motherhood, and could be spoken of as an evil for them. Pace Rosenberg, is it not rational to fear death because of its emotionally-compelling survival value? Because of that fear, we take appropriate precautions, such as monitoring our health, buying insurance policies, etc. To fly in the face of what is often constructive anxiety over death seems to demonstrate a lack o f a robust sense o f reality. As Kierkegaard put it (Journals and Papers, no. 720): "As the captive animal paces around its cage every d a y . . , so I measure the length o f my chain every day by turning to the thought of death - for the sake of movement and in order to endure living." We agree that the issue o f our first-person attitudes toward death often embroils us in an epistemolgoical quagmire. A hard, transcendental posture toward our mortality reveals that it is both rational and irrational to fear it. The fear of death is often quite functional. Fear in itself may not be desirable or even always rational or cognitive - but as a motivational attitude it often safeguards us against harm. We need to be reminded that it is often difficult to epistemically demarcate a fear o f death from a fear of serious but not terminal illness. My fear of death may involve, under Rosenberg's metaphysical presuppositions, art attitude toward the facticity of human mortality and not toward a private, experiential reality. But death remains, in Mary Mothersill's words, the ultimate "deadline for all m y assignments." I may not finish my assignments, but that failure to -

92

CRITICAL

STUDIES

complete will have its effects and consequences, its harms and benefits. Pace Rosenberg, even "thinking clearly about death" does not normally allay our fear of it. To think so is like mistakenly believing that constant dwelling upon my condition of insomnia will relieve my fatigue and sleeplessness. UNIVERSITY OF SAN DIEGO SAN DIEGO, CALIFORNIA 92110 USA

The Philosophicalff_.uarterly Vol. 35 No. 140 July 1985 Edited in the University of St Andrews by an Editorial Board Chairman of the Board of Editors: LESLIE S T E V E N S O N Executive Editor: J O H N C. HALL Reviews Editor: PETER CLARK A. Weir I. McFetridge R. Downie J.J.C. Smart

Against Holism Supervenience, Realism, Necessity Three Accounts of Promising Laws of Nature and Cosmic Coincidences DISCUSSIONS On Vico Berlin on Vico Physicalism and Phenomenal Qualities Names and Possible Objects Hacking on Frege Subscriptions for 1985 Annual Subscription U.K. Overseas Institutions s s Individuals s s

1. Berlin P. Zagorin E. Conee M. Cook M. Dummett USA s s

Canada s s

Orders, remittances and enquiries should be addressed to: Journals Department, Basil Blaekwell Publisher Ltd., 108 Cowley Road, Oxford OX4 1JF, England.

93