SORITES ISSN 1135-1349
Issue #15 -- December 2004. Pp. 87-93
Johnston on Fission
Copyright © by SORITES and Brian Garrett
Johnston on Fission
Brian Garrett

  1. In this discussion paper, I want to evaluate some arguments of Mark Johnston's which appear in his articles `Fission and the Facts' (FF) and `Reasons and Reductionism' (RR). My primary concern will be with his description of fission cases, and his assessment of the implications of such cases for value theory.
  2. Johnston presents the case of fission as a paradox. Following Parfit's (1971) discussion,Foot note 1_1 he begins by describing Sydney Shoemaker's example in which Brown's brain is transplanted into Robinson's debrained body.Foot note 1_2 Call the resulting person `Brownson'. The dominant response is that Brown is the same person as Brownson. Brownson is a unique psychological continuer of Brown, and the psychological continuity has its normal cause (the continued existence of the brain).
  3. Consider a different example. Suppose that Brown is one of the (few) actual people whose brain hemispheres are equipollent (i.e., they support the very same mental functions), and work in tandem. Suppose that Brown undergoes a hemispherectomy. His left hemisphere, say, is removed and destroyed, whilst his right hemisphere continues to function as normal. We would have no hesitation in saying that Brown survived the operation and now has only one hemisphere. The resulting person is psychologically just like Brown, and his psychology is supported by the very same physical organ which (in part) supported Brown's earlier psychology.
  4. Given the above, if Brown's left hemisphere had been transplanted into Robinson's debrained body (and Brown's old body and right hemisphere destroyed), the resulting person would have been Brown. But now -- this is the fission case -- imagine that both of Brown's hemispheres are simultaneously transplanted into two debrained bodies. We now have two excellent, and equally good, candidates for being Brown. So it seems that we must conclude that Brown survives as two people. This conclusion is bizarre; yet each step in the reasoning seemed plausible -- so we have a paradox.
  5. As Johnston notes, the paradox arises conditionally upon the following five plausible assumptions:
    1. Persons are not spooky entities; persons are thoroughly material.
    2. The identity through time of persons is an intrinsic matter.
    3. A person never survives spatially separated from himself in the fashion of a universal.
    4. No two persons can be in exactly the same place at the same time.
    5. At no time is a person constituted by two independently functioning human bodies. (FF: 379)
  6. Responses to the paradox typically involve rejecting one or more of these assumptions. Thus Swinburne rejects (A) -- thereby allowing that Brown might survive as just one of the two off-shoots.Foot note 1_3 Parfit, and many others, have denied (B): the relation of personal identity is the relation of non-branching physical and/or psychological continuity.Foot note 1_4 Hence, acknowledging that Brown survives in a non-branching situation (such as the single hemisphere transplant case) does not force us to say that he survives in a branching situation (such as fission). On a view of persons as universals, the consequence that Brown has two instances should not strike us as absurd -- hence we should reject (C). On David Lewis' view of fission, two people occupy Brown's pre-division body, and so share a common temporal part.Foot note 1_5 They subsequently become spatially distinct. On the four-dimensional view of persons, it ought to be no more remarkable for two people to share a common temporal part than to share a common spatial part. So we should reject (D). Finally, one might think it acceptable to suppose the products of fission to constitute one whole (big) person, and so reject (E).
  7. I will not here be concerned with responses that involve giving up (A), (C), (D), or (E). Johnston has an alternative response, which I shall presently consider, but I want first to look at his reasons against the response which recommends jettisoning (B).
  8. One point Johnston makes in passing, which might be thought to lend some support to (B), is this. He notes that, in the actual world, it suffices to determine who is who over time if we trace only lines of bodily continuity. «We do not look elsewhere and elsewhen.» (FF: 381) Only intrinsic factors are relevant to determining the identity over time of an actual person. As an attempt to support (B), this point is irrelevant. Just because, in the actual world, we can always conclusively determine whether or not Jones robbed the bank by tracing the world-line of Jones' body, it does not follow that personal identity can never be extrinsically grounded, in unusual contrary-to-fact circumstances. To think otherwise would be to let a mere contingency (the fact that fission of persons doesn't occur) determine a metaphysical necessity (the supposed necessity of intrinsic grounding). It would also be to let an epistemological claim settle a metaphysical thesis.
  9. Johnston then suggests that if we embraced Extreme Haecceitism about persons (EH), we would have a reason to reject (B). EH is the view that «... facts of personal identity can float free of any other layer of facts, whether they be facts about the total world process, facts about the persistence of brains, bodies or minds, or facts about processes of intermediate extent.» (FF: 391).

    First, EH is a very strong doctrine, with few adherents. Rejectors of (B) are certainly not committed to EH. Second, the Best Candidate theorist (who rejects (B)) has no reason to accept the doctrine. EH about persons is normally understood to be a view about transworld identity: the view that numerical identities or non-identities between persons in different worlds do not supervene on any impersonal identities or non-identities. (So, for example, according to EH, there is a possible world, qualitatively identical to this world -- containing all the same atoms arranged in all the same ways -- yet containing a numerically different set of people.) But the Best Candidate theory is a theory about transtemporal identity. Third, thus understood, there would -- contrary to one of Johnston's claims (FF: 391) -- be no embarrassment if a Best Candidate theorist chose to embrace EH. Perhaps there is a world like this one in all impersonal respects (e.g., in terms of the arrangement of matter), except only that the person occupying my body in that world is not Garrett, but, say, Pamela Anderson. It can still be true that, within any world, the identity of a person over time supervenes upon relevant lines of continuity.

  10. Johnston's central objection to the rejection of (B) as a response to the paradox is that there is no reason why we should reject (B) rather than any of the other assumptions. We could reject any one of the assumptions in the face of fission, so why reject (B)? But. there are reasons why giving up (B) is the most plausible response.
  11. First, on any plausible account of artefact identity we are committed to the extrinsicness of identity. In the Ship of Theseus story, the planks of the original ship a are continuously removed and replaced. The removed planks are used to construct a ship b which is plank for plank identical with the original ship a. At that later time a ship c exists which is spatio-temporally continuous with the original a. We have two later candidates for identity with the original ship. In this case our dominant reaction is to identify a with c: in the case of artefacts, the spatio-temporal continuity criterion outweighs the identity-of-original-parts criterion. But had c not existed (i.e., had the removed planks not been replaced), a would have been identical with the ship composed of a's original parts.Foot note 1_6 (This last situation is not relevantly different from that in which a ship is dismantled, transported across land, and then rebuilt. In such a case, we have no hesitation in saying that the earlier ship is the later ship.) So we are committed to the extrinsicness of artefact identity. If we think that the identity of persons, like that of artefacts, is traced by lines of continuity, we should not find it surprising that the identity of persons can be determined extrinsically.
  12. Second, giving up (B) is not as counterintuitive as giving up any of the other assumptions. Rejecting (A) conflicts with the commitment of many to materialism; and rejecting any of the others involves too great a distortion of our concept of a person. Aside from split-brain cases, it is hard to make much sense of giving up the principle one person: one body.
  13. Third, the consequences of extrinsicness are not bizarre. There is no violation of the necessity of identity.Foot note 1_7 It does follow that properties like being occupied by person B are extrinsic properties. Suppose that A is neither B nor C, but that had C not existed, A would then have been B* (i.e., the person then occupying B's body). In that case, whether a particular body has the property of being occupied by B, rather than by B*, is fixed by extrinsic factors (the existence or non-existence of C). But since the identity-involving property being occupied by B is not a causal property of a body, it should come as no surprise if its possession depends on what happens to objects which exercise no causal influence on it.Foot note 1_8 (The property of widowhood is extrinsic and non-causal, and whether a woman is a widow may depend upon what happens to a person who, at the relevant time, exercises no causal influence on her.)
  14. Fourth, the consequences of extrinsicness are even more palatable if we are impressed by Parfit's Reductionist View of persons. On one way of understanding this view, facts about persons are `conceptual' or verbal.Foot note 1_9 Once we know all the relations of continuity and connectedness holding between A at t1 and B at t2, it is a purely verbal issue whether we choose to call them `the same person', just as it is a purely verbal decision, e.g., whether to call sea-sickness `pain'. Suppose this view is correct. The extrinsicness of personal identity may seem less odd. All it amounts to is the claim that whether or not we call A and B `the same person' depends upon whether there is an equally good or better competitor for identity with A.Foot note 1_10 This linguistic dependency seems less threatening than its ontological cousin.
  15. What of Johnston's positive proposal -- his solution to the paradox?

    Johnston writes: «the fission case shows that we cannot subscribe to the unrestricted versions of all these principles. [(A) -- (E)] There is no consistent way of extending to the fission case just the principles which hold up in everyday cases. So there is no determinate sense to be made of the idea of what we would say about fission .... ...[W]e should regard the fission case as a case of indeterminacy, a case in which there is no fact of the matter about personal identity. ...Personal identity is here an indeterminate matter.» (FF: 393)

  16. So, for Johnston, if I divide into Lefty and Righty, it is indeterminate whether I'm Lefty and whether I'm Righty. That is, it is neither true nor false that I'm Lefty, and neither true nor false that I'm Righty. Is this a satisfactory response to the paradox?
  17. First, and contrary to what he thinks, Johnston's response does involve jettisoning (B). Suppose I divide into Lefty and Righty. Then, for Johnston, it is indeterminate whether I'm Lefty. But, Johnston agrees, had Righty not existed, I would have been Lefty (and this identity is determinate). This means that whether it is determinately true that I'm Lefty depends upon whether or not Righty exists. But this is just to give up (B). So Johnston cannot consistently criticise the extrinsicness solution and endorse the indeterminacy response.
  18. Second, there are cases of lop-sided fission to which Johnston's response is not applicable. One example is the Ship of Theseus. Another, involving persons, would be the following. Suppose that my hemispheres are not equipollent. My right hemisphere sustains all (or most) of my distinctive psychology. My left hemisphere sustains only basic psychological functioning. Suppose that my hemispheres are divided and transplanted. Righty has my distinctive memories, beliefs and character. Lefty is, in contrast, a very impaired person, with the mental age of a 4 year old. The most plausible description of this case is that I'm Righty and not Lefty. Righty is, by far, the best candidate for being me. Nonetheless, had Righty not existed, I would then have been Lefty -- Lefty's candidature for identity with me is good enough in that circumstance. As in the symmetrical fission case, we have a commitment to the extrinsicness of identity. Since there is a best candidate in cases of lop-sided fission, the rationale for Johnston's line is missing (there is no indeterminacy in such cases). This is an objection to Johnston, I take it, because any account of fission should cover cases of lop-sided fission.
  19. Third, as well as unwittingly violating (B), Johnston's solution also violates an analogue of (D). If I divide into Lefty and Righty, then it is indeterminate whether I'm Lefty and whether I'm Righty. Suppose that I exist at t1, and that Lefty and Righty exist at a later time, t2. How many people exist at t1? At least one -- me. But if it is indeterminate whether I'm Lefty, then it is indeterminate whether Lefty exists at t1. The same is true of Righty. So it is indeterminate how many persons exist at t1. But the thought underlying (D) is that one and only one person occupies the pre-fission body. Johnston's description violates this principle. In doing so, the implausibility of his description is revealed.
  20. There is a fourth reason for dissatisfaction with Johnston's response. Cases of indeterminacy in identity over time are typically cases in which something is missing. For example, in Star Trek Teletransportation, it is plausible to hold that it's indeterminate whether Kirk 1 is Kirk 2 because one half of the two strands that make for personal identity in the normal case is missing (viz., normal physical continuity). But in the case of fission, nothing is missing; on the contrary, everything is present, twice over. This underscores the oddness of Johnston's preferred description of fission.
  21. Parfit has argued that if we change our metaphysics of persons (e.g., reject Cartesianism and embrace Reductionism), we should change our value theory (in particular, we should come to see that personal identity is not `what matters').Foot note 1_11 One of Parfit's main arguments for the thesis that identity is not what matters relies upon the case of fission. I want here to focus on Johnston's criticism of this argument.Foot note 1_12
  22. In describing the case of fission in RR, Johnston rehearses points made in FF. But he also makes some new ones. He writes: «... relative to our practice as it stands the fission case (i) violates the ordinary presupposition of essential unity, (ii) is as a result an indeterminate case, and (iii) also violates a presupposition of our future-directed self-concern by providing more than one future person to continue an earlier person's mental and bodily life.» (RR: 603)Foot note 1_13
  23. In this quote, (i) is proposed as a reason for (ii). But what are we to make of (i)? Fission demonstrates that persons lack `essential unity': «entertaining fission involves entertaining the idea that each person has two (or more) subparts such that the survival of either one of these subparts in the right environment can secure the survival of the person. Part of the importance of the fission case is that by imaginatively violating essential unity it illustrates how we might not be mental or physical substances.» (RR: 602)
  24. It is the last step which is controversial. Why should fission be thought to undermine the `essential unity' or substance-status of persons? Consider an obvious analogy. The human body has two kidneys, which work together in tandem. If one kidney is removed, the other takes over, continuing its normal functioning. No one thinks that these facts about our kidneys show that human bodies are not substances; and this would remain true even if most organs in the human body worked in pairs. So why think that analogous facts about the human brain show that persons are not substances? Failing an answer to this question, we should leave (i) to one side -- it can provide no support for (ii).
  25. Parfit's Argument from Fission runs as follows. Suppose that I am about to divide. The prospect of division is not as bad as that of ordinary death; so my relation to my off-shoots contains what matters. But I am not identical to either off-shoot. Hence, personal identity cannot be what matters.
  26. Johnston concedes Parfit's central intuition about fission: «... I would not make a significant sacrifice to have someone intervene in my upcoming fission to ensure that only the transplanting of my left hemisphere proved viable. And this is my reaction even though I believe that only then would I determinately survive the procedure.» (RR: 610) Despite this, Johnston does not think it follows that personal identity never matters. He writes: «It is one thing to conclude that in the fission case (neurally-based) R [the relation of psychological continuity and/or connectedness] and not identity is the relation in terms of which one should extend one's special concern. But it is quite another to conclude that quite generally it is (neurally-based) R that matters.» (RR: 610)
  27. The reason for this is that in fission cases two presuppositions of our special concern are violated: (1) determinate identity; and (2) having only one future person continue one's mental life (this was (iii) above). «When such presuppositions are violated, future-directed concern neither determinately applies nor determinately fails to apply. It is reasonable to try to find a natural extension of such concern for such cases.» (RR: 610) Hence, «... were we ever to face fission, it would be reasonable to care about our fission products as we would care about a future self. But this is not because identity is never what matters. Instead, this is because caring in this way represents a reasonable extension of self-concern in a bizarre case.» (RR: 611)
  28. The lynch-pin of this argument is the premise that fission violates two presuppositions of our special concern ((1) and (2) above). There are two ways of arguing against this premise: show that (1) and (2) are not presuppositions of our special concern or show that fission cases do not violate them. I have argued that fission does not violate (1). What of (2)? By definition, fission cases violate (2). Is (2) a presupposition of our special concern? What does this mean? Does it mean that, normally, when I have `special' concern for a future person there is only one person for whom I have that concern? But then why suppose that having a unique continuer is a genuine presupposition (i.e., necessary condition) of special concern, rather than just a contingent accompanier of special concern in the actual world? We need an argument if we are to accept the former view -- and that Johnston fails to provide. Parfit's Argument from Fission has not been confounded.
  29. There is an additional worry about the claim that (2) is a presupposition of special concern. If `special concern' means `self-interested concern', then the claim is trivially true: by definition, I can only have self-interested concern for myself. If `special concern' means `strong concern I can have for myself and others', then the claim is false, but for reasons that have nothing to do with fission: I have strong concern for many people who are not R-related to me. The fact that Johnston's criticism is open to this dilemma reinforces my claim that Parfit's argument has not been properly addressed.
  30. Johnston has advanced the following three claims:

    1. (1) Rejecting (B) is an arbitrary response to the paradox of fission;
    2. (2) Fission cases involve indeterminate identity;
    3. (3) Contra Parfit, fission cases have no implications for value theory in the actual world.

I have argued that (1) and (2) are false, and that (3), if true, is not true for any reason that Johnston gives.


Brian Garrett
Australian National University

[Foot Note 1_1]

D. Parfit `Personal Identity'.

[Foot Note 1_2]

S Shoemaker Self-Knowledge and Self-Identity (Cornell University Press, 1963) pp, 23 ff.

[Foot Note 1_3]

See Swinburne's contribution in S. Shoemaker and R. Swinburne Personal Identity.

[Foot Note 1_4]

See Reasons and Persons, 266-72.

[Foot Note 1_5]

See his `Survival and Identity'.

[Foot Note 1_6]

This ship, though qualitatively identical with b, is not numerically identical with b -- on pain of violating the necessity and transitivity of identity. (Note that even if we judged that a is b and not c, we would still concede that, had b not existed, a would then have been the continuously repaired ship. Either way, we are committed to extrinsicness.)

[Foot Note 1_7]

Why not? If A divides into B and C (where all designators are rigid), isn't the extrinsicness theorist committed to the claim that, e.g., although A is not C, had B not existed, A would have been C? No -- the extrinsicness theorist accepts the necessity of identity and holds that A is not B and that A is not C; so he cannot accept that counterfactual. But he is committed to the following counterfactual: had B not existed, C wouldn't have existed. Some find the truth of this counterfactual hard to accept.

We can make a claim similar to the first counterfactual if we use non-rigid terms to refer to B and C. Suppose that `Lefty' and `Righty' are non-rigid terms referring to B and C, respectively. Then we can truly say: had Lefty not existed, A would have been Righty. (The terms `Lefty' and Righty' should be so understood in what follows.)

[Foot Note 1_8]

Lefty and Righty are in different rooms of the hospital, and exercise no causal influence on each other.

[Foot Note 1_9]

See D. Parfit `Who do you think you are?'

[Foot Note 1_10]

This version of the Reductionist View will also have implications for value theory: how can the fact of a person's identity over time be significant if personal identity is a verbal matter?

[Foot Note 1_11]

See Reasons and Persons, esp. Chs. 13-15.

[Foot Note 1_12]

My own criticism of it can be found in `Persons and Values', Philosophical Quarterly, July 1992, pp. 337-344, and in Ch. 6 of my Personal Identity and Self-Consciousness.

[Foot Note 1_13]

As it stands, (iii) is not correctly formulated: cases of lop-sided fission do not violate any `presupposition of future-directed self-concern' (since identity is preserved), yet in such cases more than one future person continues an earlier person's mental life.

maintained by:
Lorenzo Peña
Editor of SORITES