When Bernard Williams discusses whether immortality is desirable, he considers three possible scenarios. In the first, you don’t change, and you eventually become intolerably bored. In the second, you change enough to acquire new interests and avoid boredom, but at the price of eventually changing so much that you evolve into a new person, and thereby cease to exist. In the third scenario, you change just as much as you do in the second, but you don’t cease to exist (not because you change in a different way, but for other reasons). Williams thinks immortality is undesirable in the first scenario, impossible in the second, and undesirable in the third.
In chapter 2, I discussed only the first two scenarios. I did this because the objection I consider successful against the second scenario also scores against the third and because the third scenario is quite involved and blows us somewhat off course. I will discuss it here.
As I said, in the third scenario you change enough to avoid permanent and intolerable boredom, and yet you don’t cease to exist. How is it that you change so much but still exist? The reason is not that there’s something different about you. The reason is that the right theory of personal identity is different than most philosophers think it is. Most philosophers who have thought about personal identity over time have subscribed to some version of the theory that you still exist only if your later self has enough continuity and connectedness with your earlier self. To put it simply, you must have enough psychological features in common with your earlier self. Williams seems to think that this theory is correct, but he also considers the possibility that he and others are wrong and that some other theory of personal identity is correct.
On other theories, you can continue to exist even while changing dramatically over time. This might happen if, for example, you have an immortal soul that is not the same as your personality and temperament, and your survival is simply a matter of that soul’s survival, or if (according to another theory) you still exist as long as your brain does, regardless of your psychological properties over time. If one of those theories is true, then you can be immortal and avoid permanent and intolerable boredom. Williams wants to show that immortality is not desirable even if one of these theories is correct, so he asks whether that kind of existence is desirable. This is not the same question as whether you continue to exist; it’s a question about whether you would want to continue existing in that way: “That is, in my view, a different question from the question of whether it will be him” even after centuries of change.1
Williams thinks you have no reason to desire a future in which you continue to exist forever but evade intolerable boredom only by changing your character, personality, and values dramatically over time. He doesn’t claim it would be an unpleasant life, just not one that it makes any sense to desire. Why not? Williams’s answer to this question starts with his answer to another question: When, in general, do we have reason to continue living? It’s not enough that we desire simply to keep on living—“that desire itself . . . has to be sustained or filled out by some desire for something else, even if it is only, at the margin, the desire that future desires of mine will be born and satisfied.”2 In other words, it makes no sense to say that you simply desire to be alive. You desire to continue living because you need to remain alive in order to satisfy some other desires you have or will have.3 We must desire that our continued life have features that make it worth living, and we must desire those features.
According to Williams, desiring those features involves a special kind of desire. He distinguishes between desires that are conditioned on your being alive and desires that are not conditioned on your being alive, which he calls “categorical desires.” A desire for painkillers would be a conditional desire: getting such painkillers is not a reason to desire to continue living, but if you’re going to be here anyway, you want to be free of pain. Presumably a categorical desire would be something like writing a book you’ve always wanted to write—it gives you a reason to want to go on at least long enough to see it in print. According to Williams, the more categorical desires you have (and perhaps the stronger they are), the more reason you have to keep on living.4
A categorical desire is a reason for me to keep on living. A desire can’t play that role unless the object of the desire—whatever fulfills the desire, in other words—requires that I be there when it happens. That can happen either because (a) my participation will help bring about the object of that desire (I want to mentor my niece so that she graduates from high school on time), or because (b) my desire for that object includes the desire that I be there to see it happen (I want to watch her graduate), or both.
Williams doesn’t mention (a) and (b), but I think they represent what he has in mind, for they are crucial to why I have no reason to desire Williams’s third scenario of what immortality would be like—that is, no reason to desire to remain alive forever while endlessly evolving my psyche so much that I seem like a different person, even though (unlike the second scenario) I still exist. There are two conditions for me to rationally desire that I continue existing. First, whoever is living that future life must clearly be me. That makes sense; if the survivor isn’t me, then I didn’t get the future I desired. On the theories of personal identity assumed under the third scenario, this condition is met.
The second condition—this is where (a) and (b) come in—is that “the state in which I survive should be one that, to me looking forward, will be adequately related, in the life it presents, to those aims I now have in wanting to survive at all.” He also states the second condition this way: “What is promised must hold out some hopes for” my current categorical desires.5 This takes us back to (a) and (b): I need to be there, in that future, to help bring about the fulfillment of my categorical desires (they wouldn’t be categorical if I didn’t). That will happen only if my future self has the same categorical desires I have. If he lacks my categorical desire for X, then his existence does nothing to improve the odds that X will happen, and I therefore have no reason to desire that he (I) will exist. In other words, for my future state to be “adequately related” to my present categorical desires (Williams’s second condition), I must have those desires in that future. It is therefore not enough that my future self is still me; he must share my categorical desires. However, to avoid intolerable boredom, he must also be so different from me that he has different categorical desires. That means he will not share my current categorical desires, either because he can’t remember the earlier phase of life when he had them or because, even though he remembers having them, he no longer cares.6 Because his existence doesn’t promote my current categorical desires, I have no reason to desire that he exists, even though he is me.7 Therefore, I have no reason to desire the third scenario, even though I would still exist.
As I said, Williams’s argument about the third scenario, like his argument about the second, falls victim to the Tarzan objection. With respect to the second scenario, the Tarzan objection was that I have reason to care about myself in the next 20 years, and 20 years from now, I will have reason to care about my future self 20 years beyond that (imagine Tarzan taking 20 years to swing from one vine to the next), so even if I evolve into a new person and cease to exist, at any given time, I have reason to want to live a while longer. The Tarzan objection works almost the same way for the third scenario.
In the third scenario, the puzzle is not why I might care about my future self when he is not me. Rather, the puzzle is why I might care about my future self even though he doesn’t share my categorical desires. The answer is much the same: I may not have reason to care about my future self 200 years from now, for he won’t care about my desires for that future, but I do have reason to care about my future self 20 years from now, for he will share some (but not all) of my categorical desires to some extent. Stage by stage, like swinging from vine to vine, I have reason to keep going.
In the rest of this appendix, I want to consider some objections to Williams’s arguments about the third scenario that, in my judgment, are not successful. Although I think his conclusion is mistaken (mainly because of the Tarzan objection), I also believe that his arguments are stronger and more subtle than they might appear to be.
The first is an objection that I could have a categorical desire that my future self’s categorical desires be fulfilled. In other words, I don’t share the categorical desires that I’ll have 500 years from now, but nonetheless, I want those desires fulfilled just because they will be mine. Williams concedes that this is possible as a “limiting case.”8
Of course, that limiting case sounds a lot like a desire to keep on living as long as you enjoy being alive, even knowing that you will change radically over time. If that’s possible, then you could have reason to desire a kind of immortality. However, Williams does not pursue this issue and seems not to be worried about it. Why not? Perhaps for this reason: he claims that my only reason for wanting to continue to exist is that my future existence promotes the fulfillment of desires I now have. If that is right, then I can’t desire that I flourish in the future unless that fulfills one of my current categorical desires for something other than my existence. My future categorical desires could be (and eventually will be, if I live long enough) wildly different from anything I desire now. I therefore have no reason to desire their fulfillment based on the content of those desires. What reason do I have, then? The mere fact that they will one day be mine? That makes sense only if I have a reason to care about my future self, regardless of whether he shares my current categorical desires. However, according to Williams, I can’t have reason to care about my future self unless his existence promotes my current categorical desires; therefore, I don’t have a reason to care about my future self regardless of whether he shares my categorical desires. We seem to be caught in a circle, so the objection fails.
The second unsuccessful objection is that I can desire that my distant, radically different future self continue to exist not because I have any desires about his desires but simply because I believe that his existence could be good for him whether he wants it or not. Larry Temkin argues that immortality could be good for you even independently of whether you desire it if living forever is best for you in some objective way that is not tied to your desires.9 For example, life might be good for my future self (who is still me, by the way), objectively and independent of what he wants, if it contains certain things to a substantial degree: moral goodness, rational activity, developing his abilities, having children and being a good parent, knowledge, and awareness of beauty, among other things (along with the absence of various bad things).10 Let’s assume this theory is correct and that my future selves are better off if their lives contain these things.
Certainly you could have a categorical desire that your future self be objectively better off in some way that bears no relation to his categorical desires. But why would you? Because you care about your future self simply by virtue of his being you? Williams is right: that’s not enough. We need categorical desires to make sense of the desire to keep on living. You could have a categorical desire that your future self will be objectively well off, but that’s not a self-interested desire that you continue to exist unless the fact that he is you plays a role in that desire. In William’s account, you care about him because he shares your categorical desires and will therefore help fulfill them. In Temkin’s account, you have no such reason to care about him, so Temkin needs to provide some other reason why you care so much about it. That reason cannot simply be that he is you, for that begs the question. We are looking for a reason why you should care that he still exists—that is, why you should want to live that long. Williams provides one, but Temkin does not.