Thought Experiments, Ontology, And Concept-dependent Truthmakers

THOUGHT EXPERIMENTS, ONTOLOGY, AND CONCEPT-DEPENDENT TRUTHMAKERSI 1. What thought-experiments do Thought experiments are usually employed by philosophers as a tool in conceptual analysis. We pose ourselves questions such as "Would it be the same F if pT or "Would it count as knowledge if q," where p and q state some bizarre circumstances that are unlikely actually to occur and may even be beyond current technical possibility.^ The answers we are inclined to give to such questions are held to throw light on the nature of our concepts of, in these cases, identity and knowledge. But the facts about our concepts that are unearthed in this way £ire implicitly assumed to be deep, not superficial, facts. They are not meant to be facts contingent upon our current linguistic usage, psychology, or social structure, where these could easily be otherwise. If they were just facts of this superficial kind, it would hardly be worth the effort of uncovering them, for they would bind no-one who preferred a different convention or practice. The conceptual truths revealed are meant to be unavoidable, in some sense, and not merely conventional: there is something Platonic or Kantian in the background. The argument of Sections 2-8 of this essay is that, in the case of the thought experiments used to throw light on our concepts of person and personal identity, the results do not seem to be deep or hard to revise, and that this is so largely because of the ontological assumptions shared by more or less all participants in the debate.^ I shall be arguing that it is primarily these ontological assumptions, rather than the insight into our concepts that the thought experiments are supposed to bring, that determine the answers to the questions about persons and their identity. In the final two sections I shall make some cautious qualifications to this conclusion.

"Thought: Experiments, Ontology, and Concept-dependent Truthmakers" by Howard Robinson, The Monisl, vol. 87, no. 4, pp. 537-553. Copyright © 2004, THE MONIST, Peru, Illinois 61354.

538

HOWARD ROBINSON

2. The "Accepted Ontology" The ontological assumptions in question constitute what I shall call the "Accepted Ontology." This ontology is as follows: (1) There are human bodies. (2) There are mental states "associated with" or "belonging to" these bodies. (It does not matter for present purposes whether these mental states are themselves thought of physicalistically or in a property-dualist fashion.) (3) The mental states of a given person are related with each other over time roughly in the way that Parfit characterizes as continuity and connectedness, (which, following others I shall call R-relatedness) and, if they were not so related there would, at the very least, not be a normal person there. (4) There is no further entity—such as an immaterial substance—responsible for the identity at or through time of the person. My claim is that, given that these ontological decisions are already made, further revelations about what our concepts determine can only be trivial. In order to bring this out, I shall distinguish between two kinds of truthmakers for factual claims. 3. Crucially concept-dependent truthmakers The distinction that I wish to draw is between those truthmakers that are crucially world-dependent and those that are crucially conceptdependent. If I claim that there are two billiard balls on the table, concepts are, of course, involved in this claim. But, in ordinary circumstances, the nature of the world entirely determines, for all practical circumstances, how those concepts apply to the world. The same is true under normal circumstances if I say that there are two people in this room or that I am the same person as I was yesterday. But the situation is different under the following circumstances: (i) it is contentious whether a concept applies in certain dif^cult cases:

CONCEPT-DEPENDENT TRUTHMAKERS

539

(ii) this contentiousness is not thought to be a reflection of any substantive uncertainty about what is out there in the world, except in so far as such uncertainty reflects doubt about the proper use of the concept: (iii) even if there is a fact to be discovered about whether the concept actually applies in these cases, there is no reason why there should not be or have been a counterpart concept which delivers the same result in uncontentious cases, but that delivers the opposite decision in the contentious ones.* When these three conditions are met, the tnithmaker is crucially concept-dependent. When a fact is concept-dependent in this way, it is merely a form of conceptual conservatism to insist on the factual status of claims using the present concept, rather than allowing equal propriety to a "fact" articulated by using one of the counterpart concepts. The connection between the Accepted Ontology and crucially concept-dependent truthmakers is shown in condition (ii) above. A tnithmaker is crucially concept-dependent if the concepts employed in articulating that truthmaker do not figure in the articulation of the basic ontology, and if the very same states of affairs in the basic ontology are open to different conceptualizations at the non-basic level of the tnithmaker in question. Which of the different conceptualizations one chooses will determine what the tnithmaker is. 4. Thought experiments used in discussing personal identity Let us consider how the thought-experiments used in the discussion of personal identity relate to these two kinds of tnithmaker, in the light of the ontology that most of the major discussants seem to accept. For a normal human being, continuity of mental life and bodily continuity naturally go together. There are real-life cases of breakdown in mental continuity, in, for example, total amnesia and multiple personality, but the thought experiments set the normal bodily and mental signs of personal identity against each other in more systematic ways than in such pathologies. Understanding the role of these thought-experiments is made more complicated, however, by the fact that two issues are at stake in the debate. One issue is whether the psychological criterion or the bodily criterion plays the crucial role in determining personal identity; the other

540

HOWARD ROBINSON

is whether identity is what matters for "survival." These issues tend to be treated as overlapping because the psychological criterion tends to push one towards accepting that fission or reduplication preserve what matters to the human subject, whilst being inconsistent with identity proper. Opponents of the psychological criterion, therefore, tend also to be defenders of the importance of identity. There are, therefore, two main initial kinds of imaginative case. First, there are those that seem to challenge the necessity of identity for "personal survival." These are mainly cases of the splitting of mental life, for fission of this sort notoriously leads to contradictions with the logic of identity. Second, there are cases that purport to show the superiority of one of the criteria. Those that are intended to support the psychological criterion illustrate various ways in which R-relatedness can survive body swapping. Whole-brain transplants are an instance of this. The brain flssion with transplants cases support both the priority of the psychological criterion and the lack of importance of identity. Those that are intended to support the bodily criterion are meant to illustrate how mentality can be replicated without preserving the subject of experience. Williams's cases (1956-57; 1973) are meant to be instances of this. There is a third kind of case, however, that brings together, from the perspective of the psychological criterion, the supposed results of the other two. If one accepts the priority of the psychological criterion and the non-essentiality of identity, then straightforward reduplication or "xeroxing" of persons will preserve "what matters." For some (e.g., Johnston: 1997) this is a reductio of the psychological approach, for others, such as Parfit, it is a triumphant liberation.5 5. The thought experiments and the Accepted Ontology It seems to me that, given the Accepted Ontology, there is no relevant dispute about the facts in any of these cases, and so the truthmakers in disputes about personal identity must be crucially concept-dependent. In the case of fission, all agree about what happens to the body, and what the R-relations are after the flssion. With replication, it is agreed that there is a different body with apparent R-relations to the original. With Williams's cases, it is agreed that the identity of the body remains unchanged, but certain mental states are altered. What, then, in terms of the Accepted Ontology, is in dispute in any of these cases? On the ontology in (1) to (4),

CONCEPT-DEPENDENT TRUTHMAKERS

541

if A and B are psychologically connected, and A and C are not, but A and C are somatically connected and ^4 and B are not, those are all the facts. The further question "but which of 5 or C is the same person as A?" has no further factual content; at most, it reflects the concept we happen to possess.* 6. The thought experiments and crucial concept-dependence To see whether this is the correct diagnosis, let us look at the imaginary cases in the light of the three criteria for being crucially concept-dependent. Criterion (i)—that the application of the concept is, in some cases, contentious—plainly applies because it is contentious how the concept of 'person' applies in the thought-experiment cases; (ii)—that this contentiousness does not arise from any disputes about "the facts"— is what was argued in Section 5. Only by appealing to intuitions about the content of the concept of a person cein the issue be further decided. It seems, then, that (iii)—whatever our concept says, there would be an equally workable one that delivered different verdicts in the contentious cases—is the only possible area for dispute. It must be claimed that it is not merely accidental or arbitrary that we have the precise concept that we do. Prima facie, it is difficult to see why this should be so. It is agreed on all sides that the bodily criterion and the psychological criterion give the same results for all normal cases: this is why science-fictional thought experiments are required to bring out the difference. Suppose we somehow discovered or decided that our linguistic-conceptual practices, at a deep level, really confirmed the bodily criterion. I can see no reason in theory or practice why those of another community could not similarly confirm the psychological one. As they both work for all the actual cases that have so far occurred, neither could be dysfunctional. For similar reasons, if teletransportation, brain-transplanting, etc., became actual practices, a Parfitian might recommend that we change our linguistic practices and hence our concepts so as to accommodate them. If the ontology stated in our assumptions is correct, then the protest "but they would still not really be the same person," can mean no more than "that is not exactly how we have so far used the concept 'person'." If these considerations are correct, the dispute can be about nothing more than how our present concept happens to legislate the matter. If it were to be about something more than such a crucially concept-dependent decision, this could only be because there is what Parfit terms a "further

542

HOWARD ROBINSON

fact" in the world, such as would be constituted by a Cartesian ego; that is, only if one rejected (4) in the Accepted Ontology. 7. Smuggling in "furtherfacts" It is worth labouring this latter point because it is so natural to take the further fact for granted, without noticing it. It is, I believe, for example, quite natural to think in the following way of the difference between fission and reduplication: In the case of fission, complete continuity of consciousness through the process of division would seem to be possible; experiences could, in James's phrase "flow into" one another across the traumatic moment, as they do in normal experience. With reduplication, however, the overlapping of contents from moment to moment that seems to characterize normal experience could not cross the divide between the last experience had by the subject before duplication, and the first experience of the duplicate. Even if there was no temporal gap between the former and the latter, and even if the contents entirely matched, there could be no actual overlapping of content.'' It is natural to take this fact as evidence that it is or might be really you that survives division, but it is not in duplication. The possibility of real overlap of content in the one case and its impossibility in the other, is taken as a sign of the possibility of real survival in the one case and its impossibility in the other. But, on the Accepted Ontology, what could such "real survival" in the fission case consist in? There are only different facts about how the R-relations operate, and about relations of bodies. Given that, ex hypothesi, it is not evidence for the existence of something further, how could the fact that one situation could, in principle, sustain unbroken continuity and the other could not, be, of and in itself, a matter of any particular significance? It can only be a fact about kinds of continuity, and nothing else. Nevertheless, one thinks that it indicates something further. In fact, slippage of this kind goes back to the language Locke used in the Essay Concerning Human Understanding (Locke 1690/1894/1959) when first formulating the memory criterion of personal identity. He adapts it from his account of the identity through time of organisms. That being then one plant which has such an organization of parts in one coherent body, partaking of one common life, it continues to be the same plant as long as it partakes of the same life, though that life be communicat-

CONCEPT-DEPENDENT TRUTHMAKERS

543

ed to new particles of matter vitally united to the living plant . . . {Essay 11.27.5) One might wonder what this 'common life' is supposed to be, especially in the light of his rejection of Aristotelian 'substantial forms'. The important nineteenth century editor of the Essay, Alexander Campbell Fraser, adds the simultaneously unhelpful and revealing note: It is only in a loose sense in which the 'organisation' which is visible, can be identified with the 'life' which is invisible. (Locke 1959: 443) The comment is unhelpful because it is unclear what the 'loose sense' is. It is revealing because it draws attention to the slippage in Locke's thinking. Locke's language reifies 'the same consciousness' in much the same way as it does 'the same life'. the same consciousness uniting those distant actions into the same person . . . {Essay \\.21.\0) For it is by the consciousness it has of its present thoughts and actions, that it is self to itself now, and so will be the same self, as far as the same consciousness can extend to actions past or to come... . {Essay 11.27.12.) Locke's language, on both these issues, is sometimes reductionist, but sometimes suggests that some further, real entity supervenes on, or is immanent in, the continuities that underlie it. No doubt this slide happens unconsciously in Locke, and a similar shift is liable to occur in us when we think we are thinking in terms of a reformed and reductive ontology. The reductionism obliges one to think of experiential unity at a given time as a construct from the relations between mental contents, but one can hardly avoid the thought that that very same unity—a unity of apperception—is passed on as the contents change, at least for as long as there is unbroken consciousness: the unity itself becomes reified. 8. Consequences for the psychological criterion/bodily criterion debate How do the conclusions so far affect the debate between the psychological and the bodily criteria? They might seem to be neutral because they only show that nothing much is gained by trying to unearth the identity conditions implicit in our current concept of a person and so favours the claim of neither to be the key to this non-issue. If one.

544

HOWARD ROBINSON

therefore, abandons the search for identity in a pure sense and instead sees the dispute as one between two rival criteria for what matters in the continuity of persons ("personal identity" loosely so-called), then it might seem that our argument has nothing to tell you about which to prefer, and so to be neutral. The contest between the two criteria tends not, however, to be represented in this way. Parfit claims that it is R-relatedness that matters, not identity, but his opponents do not make the parallel claim that it is bodily continuity, not identity, that matters. Rather, the kind of bodily continuity they are interested in is strict one-to-one continuity, with no branching or fusion, so they can assimilate it to strict identity for the body. They then further assimilate the identity of the body to that of the person, perhaps with the proviso that the body must be living. These moves are often implicit, and options need distinguishing. A supporter of the bodily criterion can be thought of as saying any of three things. First, and most modest, accepting the neutrality postulated above, would be the claim that, though he accepts that my arguments show that nothing is at dispute of a deeply factual nature, he prefers the bodily criterion. This position seems unmotivated. To prefer being the same body when one does not see this as the key to being the same person, but simply being the same body, per se, seems to lack any rationale. The second is that bodily continuity without (bodily) branching constitutes bodily identity, and that that identity counts, per se as personal identity. This seems hardly better motivated than the first. Within the constraints of the Accepted Ontology, the suggestion that bodily identity alone—^that is, in the absence of any R-relatedness—should count as personal identity amounts to the claim that causal dependence on the same body (or, if one is reductionist, being a property of the same body) constitutes being experienced by the same subject. But the idea of being experienced by the same subject carries with it something further than the claim that all the relevant mental states causally belonged to the same body—even if the more that it carries along with it is thought to supervene on this fact. Williams wants to convince you that, when your body is tortured after all your mental characteristics have been obliterated, it will really be you that is in agony, where this is not another way of saying that it will be the body that once housed your psychology that feels the pain. If it were just the latter, it would remain an open question whether you, as prudent, should care about the prospect. The third option is that sameness

CONCEPT-DEPENDENT TRUTHMAKERS

545

of body constitutes personal identity because there necessarily supervenes upon it—irrespective of our concepts—sameness of subject. This is why it is a real fact that it will hurt you when your psychologically renovated body it tortured. It is this last position which, I think defenders of the bodily criterion tacitly, if unconsciously, assume. But, as with the case of continuous overlapping consciousness considered above, this is actually a hidden version of a "further fact" theory, and, as such, not compatible with the accepted ontology.* This gives us good reason for thinking that, within the constraints of the Accepted Ontology, Parfit must be right. First, there is the realization that, given the ontology, truthmakers for identity-claims will be crucially concept-dependent and so trivial. But there is also a further consideration that strongly favours him, which has emerged in the paragraphs immediately above, namely that the bodily criterion has no account of subjecthood. With the Accepted Ontology there are only two forms of connexion between mental contents that are possibly relevant to subjecthood. One is their causal dependence on the body, the other is their relations one to another that are summed up as R-relatedness. R-relatedness is generally thought of as being a diachronic relation, but synchronic co-consciousness—sometimes equated with the unity of apperception— must also be something of the same sort: that is, it must be a construction from a basic co-consciousness relation. The animalist believes that the same subject can survive total obliteration of R-relatedness and, presumably, by parity of argument, atomization of current consciousness. This leaves him with no alternative but to claim that causal dependence on the same body constitutes sameness of subject. Or, at least, it leaves him with no alternative, if he thinks that sameness of person entails sameness of subject. He might argue that the same body is equivalent to same person, but allow that a person could be—or house—a series of different subjects. This manoevre has the danger that it seems to transfer at least some of the problems we re interested in to einother point. I am interested in myself as a subject of experience, and if the person that I am gets detached from this, then perhaps it is not myself qua person that most concerns me. Assuming that the animalist wants to tie sameness of person to sameness of subject, and that he cannot, therefore, explain sameness of subject in terms of R-relatedness, he has no other resources with which to explain it except as belonging to the same body. But this, without some

546

HOWARD ROBINSON

further or emergent feature, cannot do justice to the fact that subjecthood is not a matter of logical ownership alone—as a ball possesses its roundness —but has an essential experiential aspect. Once the Cartesian ego is abandoned, conscious subjecthood has to be understood in terms of some basic relation between the only consciousness-possessing elements left, namely mental states. Negating these, leaves one with notfiing.^ In the next two sections I shall discuss two recent accounts that, in different ways, claim that our ordinary conceptions of personal identity are untouched by reductionist accounts. From these will emerge a cautious reservation about the argument so far. 9. Johnston's defence of "further facts " Quite counter to the main argument above, Mark Johnston has argued in various places, but especially his (1992) and (1997), that the existence of a "further fact" about identity—and other things—is consistent with a reductionist ontology. According to Johnston, ordinary commonsense facts are, in general, autonomous and not dependent on one's preferred theoretical ontology. It is proper to hold that normal people are free agents whether or not one believes in "uncaused volitions," and whether or not one is a determinist. One believes that tomorrow is yet to come, whether or not one accepts "a moving NOW" or a block universe. Similarly, issues about the identity of human beings can be fixed in the normal ways whether or not one accepts Cartesian egos or just bodies with associated mental events. Johnston calls belief in the resilience of ordinary conceptual practice in the face of metaphysical differences, "Minimalism." Johnston's position is complex and, perhaps, sometimes overgeneralized. One might be tempted to think that some of our conceptual practices are dependent on a certain metaphysical background, and others are not. Johnston, via his doctrine of "minimalism," talks as if none of them are. Johnston, for example, seems to believe that adopting a "redundancy" theory of truth, and rejecting Parfit's claims that his theory of personal identity has ethical consequences, are both instances of "minimalism," and, therefore, parts of the same general strategy. It is difficult to believe that the kinds of considerations that enter into these questions have anything interesting in common. But, in the present context, I am only concemed with whether adopting the Accepted Ontology gives a

CONCEPT-DEPENDENT TRUTHMAKERS

547

rationale to the psychological criterion, or whether, as Johnston claims, our common-sense views on identity are untouched. In fact, the point I am making in this section is not essentially ad hominem against Johnston. It concems a quite general thesis which is widely and naturally held. It is that we have many vital conceptual practices that do not concern themselves with basic ontology and which are not affected by different views on basic ontology, and talk about persons is one such vital conceptual practice. It can hardly be denied that what is said in this general thesis is often true: not all our conceptual practices are dependent on our views about basic ontology—indeed, perhaps few of them are. But the emergence of novel cases—like fission or teletransportation—obliges us to consult our ontology to see how to decide what we should do in those cases. Why this consultation of our ontology should be required is clear from what has been said above. The issue is whether the conceptual practices that are not directly dependent on our basic ontology are not thereby rendered crucially concept-dependent, and so open to revision. I have already argued that in the case of the concept of a person, it clearly is conceptdependent. Perhaps my basic point could be re-expressed as follows. In so far as a conceptual practice is not determined or required by one's basic ontology, the reason for adopting it cannot be that it is required if one is to be true to the facts. But the disputants on the issue of personal identity do, I think, believe that they are arguing about the truth. Nevertheless, Johnston raises an important issue. The third clause in the definition of crucially concept-dependent truthmakers required that there be different concepts that delivered the same verdict in non-controversial cases and different verdicts in the contentious cases. There is an implication here that the adequacy of a concept is entirely a matter of its descriptive adequacy: if two concepts deliver the same verdict in the uncontroversial cases, then they are equally acceptable candidates. But there might be non-descriptive reasons for choosing one of the candidates. Locke said that the concept of a person was a forensic concept. In other words, the conditions for the unity of a person are determined by the requirements for holding someone responsible. Rovane (1998), rather similarly (though there are also serious differences) thinks that a person is that to which agency can be ascribed. Parfit believes that the flexibility that accompanies the psychological criterion enables one to avoid the morally pernicious priority of prudence. On the other side, Wiggins (2001:193-244)

548

HOWARD ROBINSON

believes that the psychological criterion must be rejected on moral grounds, rather than descriptive ones, because it leads to too insubstantial a conception of human nature. Perhaps it could be the case that two alternative conceptions are equally acceptable in terms of simple descriptive ontology, but are not equally acceptable in terms of pragmatic ontology. Both answer with equal accuracy to what is in one's fundamental ontology, but one serves some important purpose that the other fails. Locke believes that taking bodily continuity as the ground of identity would fail to meet requirements of justice: whereas Wiggins thinks that adopting a Lockean or functionalist criterion would leave us too easily victims to the machinations of evil technocrats. Now it is no part of my argument that there could be no reasons for preferring one criterion rather than another, perhaps depending on the kind of case. But the nature of the dispute will have changed from being one about metaphysical fact. Bad theories will no longer be false, but unfortunate, like the enactment of a bad law, which has destructive consequences. The kinds of argument deployed will be different. My claim is only that, in the strict domain of metaphysical truth, the accepted ontology leaves nothing to be disputed. 10. Gendler's argument for the priority of identity over R-relatedness Tamar Gendler (2002) has provided an illuminating defence of the importance of strict identity for the survival of persons. Her argument is that, although R-relatedness may be sufficient for personal survival in some cases, this may be so for reasons that are parasitic on the central "genuine identity" case. The concept she employs is that of "borrowed lustre." An example might be as follows. (It is mine, not Gendler's.) I might believe that human beings have the basic rights they do because of their nature as rational beings. I might also believe that babies and radically subnormal people also have most of the most fundamental of these rights. Someone might argue that my granting rights to babies and the subnormal showed that I do not think that it is rationality that really commands respect, but whatever quality or qualities it is that normal adults, babies, and the subnormal share. This would not be right, however. I acknowledge those rights in the case of non-rational humans because of the relationship that they stand in to the normal human adult. It is because the mature state is what, by their very nature as human beings, they will

CONCEPT-DEPENDENT TRUTHMAKERS

549

be or should be like, that they have those rights. They "borrow lustre" from their essential relationship to the paradigm case. Similarly, according to Gendler, the R-relatedness shown in Parfitian thought experiments has sufficient affinity with standard cases of identity to squeeze in under the same umbrella. It does not follow that identity matters only because it involves R-relatedness. On the contrary, R-relatedness on its own "borrows lustre" from its similarity to normal identity. In Gendler's own words Tbe suggestion is tbis: suppose we had no sense that there could be identity, as opposed to mere R-relatedness—would there be such a thing as prudential concern? I suggest that the answer is 'No'... . Our views about tbe sorts of rational and moral obligations we bave to ourselves and others, considered as beings wbo exist tbrougb time, rest on tbe assumption tbat eacb of us will bave at most one continuer, and tbat tbis continuer will be someone witb wbom we will be identical. (50) There is, of course, as Gendler acknowledges, a difference between the claim that, if fission and fusion were too common, some or all of our rational practices would break down: and the claim that these rest essentially on the idea of identity. There is also a difference between insisting on the importance of one-one continuation of psychological continuity, and of ruling out body swaps. For body swapping is consistent with nonbranching psychological continuity, because an unbranching mind might move from body to body. But the difference between, on the one hand, insisting on identity and, on the other, accepting the ontological sufficiency of stable continuity, can be reconciled by a Humean. Humeanism is an ontological theory—a person is nothing other than a bundle of mental events standing in certain relations. There is no reason, however, why he should not agree that these relations must be such as to sustain the '"I thought'—^that which enables one to treat oneself as the same subject through time, and that this is possible only with a large degree of stability in mental and, preferably, bodily continuity. The necessity of being able to apply the oversimplifying category of identity will not worry a Humean— the necessity of such illusions pervades his whole theory. It is still, nevertheless, an illusion that it is genuine identity that is doing the work. When the adequacy of continuity depends on the ability to treat it as identity, and identity is no more than the reification, under our concept 'person', of an instance of continuity, which is borrowing its lustre from which? (Blackburn [1997] engages helpfully with these issues.)

550

HOWARD ROBINSON

If Hume can accept the pragmatic unavoidability of appeal to identity, perhaps it is less obvious that Parfit can do the same thing. This is an issue to which we shall return in our conclusion. 11. Conclusion: What is at stake in the conflict between the two criteria? At first sight, the point of the debate about personal identity that has been raging over the last forty years seems clear enough. It is about the make-up or constitution of persons. I think that what I have said about the Accepted Ontology shows that, for the most part at least, that issue is akeady begged. What, therefore, was at stake? There are three major areas of conflict in this dispute, two of which have been discussed above and one hinted at, at the end of the previous section. (i). The one to which I have only hinted is Derek Parfit's main motive for interest in the subject. Although not discussed above, it needs noting because it has been one of the major motivations for the whole debate. It is the belief that taking the R-relation, rather than identity, as "what matters in survival" undermines our conviction that ethical rationality must rest on prudence. Parfit seems to think that the logical possibility that people, including myself, might have been R-related to more than one person, or that such a thing might happen in the future (presumably not to me, because it will not happen that soon) is a reason for me now to feel a kind of connection with others, such that I do not distinguish sharply between their interests and my own. It is a moot point how far acceptance of a pragmatic need to employ the concept of identity would undermine this argument. On the one hand, if one cannot avoid thinking in identity terms, perhaps one cannot legitimately avoid accepting its consequences. On the other, recognition that this is a pragmatic need and not one founded in fundamental ontology might leave space for an argument that "really" there is no fundamental difference between oneself and others. There are different versions of a doctrine of pragmatic necessity and this might affect this argument. Of course, none of this affects the basic arguments for rejecting Parfit's ethical claims. His view seems unconvincing (a) because the fact that something is an unsatisfied logical possibility, or that something might happen in the future to someone else, doe's not seem to be a reason why I should revise my conception of rationality in the present. It is also odd (b) because, even if fission were to happen, the fact that R-related-

CONCEPT-DEPENDENT TRUTHMAKERS

551

ness, not identity, is what matters would only give me reason to be concemed about those who are, or could reasonably be expected to be, Rrelated to me. It would give me an extended family, not a general disinterestedness. I need have no more general a sense of civic responsibility than a corrupt politician. (ii) The second concern that lies behind the debate is the pragmatic importance of being able to treat people as "the same" through time, and in what way this pragmatic requirement can be said to have ontological consequences. Does it have the form of a Humean "unavoidable illusion," or (if this is different) a Kantian category, and, if the latter, does adopting this kind of Kantianism have a metaphysical and ontological cost, or can it be reconciled with the Accepted Ontology? (iii) The third concern is the straight ontological one. Can one show that Parfit's theory is defective in ways that force one to abandon the Accepted Ontology in favour of a "further fact" that has direct Cartesian clout? In sum, I think it is helpful to be clear about the role of the Accepted Ontology; about the difference between world-dependent and crucially concept-dependent truthmakers; and about the ethical, pragmatic, and directly ontological issues at stake. Howard Robinson Central European University Budapest

NOTES 1. My thanks to the editors, Tamar Szab6 Gendler and Dean Zimmerman, for helpful comments on a previous version of this essay. 2. There is a different kind of thought-experiment which has the peculiar feature of being essentially counteifactual. In the cases which will be discussed in this essay and which constitute the main body of thought-experiments relating to personal identity, one is asking "what would one say were such-and-such [for example, a brain transplant] to occur?" Although no such thing has happened yet, some day it may. But there are other cases relevant to issues of identity which are irreformably counterfactual. One might ask "Would it have been the same F if pi" where it is essential that p was not in fact the case. An example of such a case is "Would it have been me that was bom if the sperm which fertilized 'my' egg had been different in certain ways from how it actually was?" No future, however wonderful science becomes, could realize such a case, because if someone had been bom from "my" egg and a slightly altered sperm, that would have done nothing

552

HOWARD ROBINSON

to show whether the person so produced was the same as would have been produced under the conditions that obtain in our world. Because of this irreformably counterfactual element, such experiments are condemned to remain etemally thought experiments, in a way that speculations about transplanting brains are not. But these thought experiments can still be important for problems in personal identity. For example, Madell (1981), Swinbume (1986), and Robinson (2003) argue that such "would it have been me?" counteifactuals as the one cited above must have a determinate answer, and that this proves the simplicity of the self in a roughly Cartesian manner. Although such counterfactual cases are not our present concern, they raise interesting issues of their own and their importance should not be overlooked. 3. The exceptions are those who defend a broadly Cartesian theory of the self. Examples are Madell (1981), Swinbume (1986), Foster (1991), and Robinson (2003). 4. I say "even if there is a fact to be discovered " because there are different views about whether vagueness in concepts is possible, and, indeed, on whether, if it is not possible, the true contours of the concept can ever be discovered. See Williamson (1994). 5. The case of "teletransportation," as found in the television series "Star Trek," is often cited, for example, by Parfit, in discussions of duplication. There are, I think, two ways of understanding this process, only one of which makes it a straightforward case of duplication. On this interpretation the teletransporter destroys the initial body whilst also extracting from it information as to its structure. This information is transmitted, and, on the basis of it, a duplicate body is constructed. On the other interpretation, the machine transforms the matter of the body into a form of energy that contains all the information about it and "beams down" that energy to the target location, where it is re-materialized. The important difference is whether what is transmitted can be regarded as a form of, and not just data about, the original. If it can, then identity can be preserved. 6. There may be pragmatic reasons for choosing one concept over another, as we shall see in Section 9. 7. I am assuming that experience is not a series of discrete "pulses," but that data or qualia can overlap or "flow into" each other, so that one datum can begin half-way through the existence of another. All views on the nature of experience are controversial, but the one I am assuming seems to be the natural one. If the "pulse" view is correct, with content reproduced moment by moment without genuine overlap of contents, then duplication and continuation cease to be seriously different. Each new moment is rather like waking up in the moming with one's memories intact: there is no real continuity. 8. Thomson (1997) is an animalist who is aware of this problem. But she acknowledges that the problem presents itself as something with which her theory, as presently understood, does not know how to cope. I am suggesting that the animalist is forced to accept Ayer's "no ownership" theory of the self (Ayer: 1963); a theory refuted by Foster (1968), which Ayer felt obliged to abandon.

REFERENCES Ayer, A. J. (1963): "The Concept of a Person" in The Concept ofa Person and Other Essays, London: Macmillan, 82-128. Blackbum, Simon (1997); "Has Kant Refuted Parfit?" in Dancy (1997), 180-201.

CONCEPT-DEPENDENT TRUTHMAKERS

553

Dancy, Jonathan (ed.) (1997): Reading Parfit, Oxford: Blackwell. Foster, John (1968): "Psychophysical Causal Relations," American Philosophical Quarterly, 5, 64-70. (1991): The Immaterial Self, London: Routledge. Gendler, Tamar Szab6 (2(X)2): "Personal Identity and Thought-experiments," Philosophical Quarterly, 52, 34-54. Johnston, Mark (1992): "Reason and Reductionism," Philosophical Review, 101, 589-618. (1997): "Human Concerns Without Superlative Selves," in Dancy (1997), 149-79. Locke, John (1690/1959): An Essay Concerning Human Understanding, ed. A. C. Fraser, New York: Dover Publications. Madell, Geoffrey (1981); The Identity of the Self. Edinburgh: Edinburgh University Press. Parfit, Derek (1984): Reasons and Persons, Oxford: Oxford University Press. Robinson, Howard (2003): "Dualism," in Stich, Stephen and Ted Warfield (eds.) The Blackwell Guide to the Philosophy of Mind, Oxford: Blackwell, 85-101. Rovane, Carol (1998): The Bounds of Agency, Princeton, NJ: Princeton University Press. Swinbume, Richard (1986): The Evolution of the Soul, Oxford: Clarendon Press. Thomson, Judith Jarvis (1997): "People and Their Bodies," in Dancy, 202-29. Wiggins, David (2001): Sameness arid Substance Renewed, Cambridge: Cambridge University Press. Williams, Bemard (1956-57): "Personal Identity and Individuation," Proceedings of the Aristotelian Society, 57, 229-52. Reprinted in Williams (1973): 1-18. (1973); Problems of the Self, Cambridge: Cambridge University Press. Williamson, Timothy (1994): MJ^MCWCM, London; Routledge.