Who are you?

There are two things that people debate with regards to continuation of personhood. One is whether edge cases to our intuitions of what ‘me’ refers to are really me. For instance if a simulation of me is run on a computer, is it me? If it is definitely conscious? What if the fleshy bloody one is still alive? What if I’m copied atom for atom?

The other question is whether there is some kind of thread that holds together me at one point and some particular next me. This needn’t be an actual entity, but just there being a correct answer to the question of who the current you becomes. The opposite is a bullet that Eliezer Yudkowsky does not bite:

…to reject the idea of the personal future – … that there’s any meaningful sense in which I can anticipate being myself in five seconds, rather than Britney Spears. In five seconds there will be an Eliezer Yudkowsky, and there will be a Britney Spears, but it is meaningless to speak of the currentEliezer “continuing on” as Eliezer+5 rather than Britney+5; these are simply three different people we are talking about.

The two questions are closely related. If there’s such a thread, the first question is just about where it goes. If there’s not, the first question is often thought meaningless.

I see no reason to suppose there is such a thread. Which lump of flesh is you is a matter of definition choice as open as that of which lumps of material you want to call the same mountain. But this doesn’t mean we should give up labeling mountains at all. Let me explain.

Why would one think there is a thread holding us together? Here are the reasons I can think of:

1. It feels like there is.

2. We remember it always happened that way in the past. There was a me who wondered if I might just as well experience being Britney next, then later there was a me looking back thinking ‘nope, still Katja’ or some such thing.

3. We expect the me looking back is singular even if you were copied. You wouldn’t feel like two people suddenly. So you would feel like one or the other.

4. Consciousness seems like a dimensionless thing, so it’s hard to imagine it branching, as if it could be closer or further from another consciousness. As far as our intuitions go, even if two consciousnesses are identical they might be in a way infinitely distant. What happens at that moment between there being one and there being two? Do they half overlap somehow?

1 is explained quite well by 2. 2 and 3 should be expected whether there is any answer to which future person is you or not. All the future yous look back and remember uncertainty, and currently see only themselves. After many such experiences, they all learn to expect to be only one person later on. 4 isn’t too hard to think of plausible answers to; for instance, perhaps one moment there is one consciousness and the next there are two very similar.

Eliezer goes on to describes some more counterintuitive aspects:

…I strive for altruism, but I’m not sure I can believe that subjective selfishness – caring about your own future experiences – is an incoherent utility function; that we are forced to be Buddhists who dare not cheat a neighbor, not because we are kind, but because we anticipate experiencing their consequences just as much as we anticipate experiencing our own. I don’t think that, if I were really selfish, I could jump off a cliff knowing smugly that a different person would experience the consequence of hitting the ground.

These things are all explained by the fact that your genes continue with your physical body, and they design your notions of selfishness (Eliezer disagrees that this settles the question). If humans had always swapped their genes every day somehow, we would care about our one day selves and treat the physical creature that continued as another person.

If we disregard the idea of a thread, must every instantaneous person just as well be considered a separate, or equally good continuation, of you? It might be tempting to think of yourself randomly becoming Britney the next moment, but when in Britney only having her memories, so feeling as if nothing has changed. This relies on there being a you distinct from your physical self, which has another thread, but a wildly flailing one. So dismiss this thread too, and you have just lots of separate momentary people.

Imagine I have a book. One day I discover the pages aren’t held together by metaphysical sticky tape. They have an order, but page 10 could just as well precede page 11 in any book. Sure, page 11 in most books connects to page 10 via the story making more sense, but sense is a continuous and subjective variable. Pages from this book are also physically closer to each other than to what I would like to think of as other books, because they are bound together. If I tore them apart though, I’d like to think that there was still a true page 11 for my page 10. Shouldn’t there be some higher determinant of which pages are truly the same book? Lets say I accept there is not. Then must I say that all writing is part of my book? That may sound appealingly deep, but labeling according to ordinary physical boundaries is actually pretty useful.

The same goes for yourself. That one person will remember being you and act pretty similar and the rest won’t distinguishes them interestingly enough to be worth a label. Why must it distinguish some metaphysically distinct unity? With other concepts, which clusters of characteristics we choose to designate an entity or kind is a matter of choice. Why would there be a single true way to choose a cluster of things for you to identify with any more than there is a true way to decide which pages are part of the same story?

I’ve had various arguments about this recently, however I remain puzzled about what others’ views are. I’m not sure that anyone disagrees about the physical facts, and I don’t think most of the people who disagree are dualists. However many people insist that if a certain thing happens, such as their brain is replaced by a computer, they cease to exist, and believe others should agree that this is the true point of no longer existing, not an arbitrary definition choice. This all seems inconsistent. Can someone explain to me?

Added: it’s interesting that the same problem isn’t brought up in spatial dimensions – the feeling of your hand isn’t taken to be connected to the feeling of the rest of you through anything more complicated than nerves carrying info. This doesn’t make it just as well anyone else’s arm. If you had a robotic arm, whether you called it part of you or not seems a simple definitional matter.

18 responses to “Who are you?

  1. Neurologists describe conscious states due to brain injuries that it’s scarcely possible to imagine – most famously in Oliver Sacks’ books such as “The Man Who Mistook His Wife for a Hat.” I think those case studies are a better source of insights than philosophical thought experiments. In particular, it becomes clearer that consciousness is not a unitary, dimensionless thing. And if intuition gets that wrong, I’m not sure I trust it about other hypothetical edge cases.

  2. Very good. I got into a long argument with a friend about this one a little while ago with a friend who in fairness was somewhat drunk… anyway he seemed to completely retract all the views he’d been vehemently advocating the previous night once confronted with a thought experiment the next day.

    My view is that we do meaningfully identify with our future selves – sure because evolution programmed us to be that way, but that doesn’t invalidate it as an approach; meaning is as the existentialists have it self-constructed, so we may as well construct it in accordance with such deeply held intuition. Likewise I hold morality is still important even though its “merely a byproduct of evolution”, and so on.

    I’m a functionalist, so physical continuity is not what matters, a sufficiently advanced computer simulation of me is still me (although pragmatically it may prove very difficult to physically replicate my neural architecture faithfully enough.)

    I think that if you hold the first two views it is incoherent to only identify with a single “thread”, and so there may potentially be multiple future copies of you that are all you. More precisely if three copies are made of me at age 11, then each “is the same person” as me at age 10. However they cease to be the same person as each other as their experiences start to diverge, so personal identity is non-transitive… 12 year old me #1 is 10 year old me, and 10 year old me is 12 year old me #2, but this does not imply 12 year old me #1 is 12 year old me #2.

    So in a universe with person copying technology, identity legitimately branches, and we should identify with the whole tree of all selves proceeding forward from our current consciousness.

    Of course this leads to bizarrely counter-intuitive results. If I am to be copied in 10 seconds time, and then one of the clones is to be tortured to death and the other is to live in bliss for ever, should I be terrified or ecstatic? Well, both. Our intuition isn’t built with copying in mind, and there’s no way around that as far as I can see.

    I can’t explain others’ views for you, sadly. I think most people just abide entirely by their intuition and unconsciously or even somewhat consciously have blatantly contradictory ideas. The philosophically inclined remainder more or less pick arbitrarily just how much intuition failure they can deal with and make that decision axiomatic, as I have done.

    In light of this post I’ll blog my thought experiment argument to my friend about identity branching.

  3. I certainly don’t believe in threads, but I still go around thinking of *trees*. It’s the nonexistence even of trees – of any sense in which I can be said to anticipate becoming even some *group* of Eliezers that excludes Britney Spears – that is a bullet I still have trouble biting.

    • Sorry, I misplaced a paragraph – I removed it so your position is more clear.

      I still don’t understand why anticipation should consist of anything other than the belief there will be future people who remember being you.

    • It seems like we can reduce this to an artifact of reinforcement learning, just as you reduce ‘free will’ to an artifact of avoiding a recursive loop in calculating the consequences of decisions.

      I recommend just biting the bullet and converting anticipation to ‘preference plus cognitive illusion.’

  4. It all seems pretty manageable if we stay within the class of creatures who remember being me as much as I remember being Robin of decades ago, and whose personality hasn’t changed much more than that change. The smaller or more distorted their memories, or the larger and more varied their personality changes, the harder it is to decide how much I care about them, or expect to be them.

    • Sure, but still a practical matter of choosing a definition you like.

      • Pablo Stafforini

        It cannot be a mere definitional matter if you believe, as Robin and others seem to believe, that important things turn on whether you and the future individual who shares your memories and personality are or aren’t the same person. For these people, the decision to sign up for cryonics, say, will partly depend on whether cryopreservation is identity-preserving; and whether it is cannot be settled by simply choosing a convenient definition.

        Of course, if you really thought it was a definitional matter (as I think it is), then that would be a reason to drop the view that personal identity has rational or moral significance.

  5. mitchell porter

    I “think there is a thread holding us together” – that is, that there is something which is me, persisting in time; and that even if there comes a time when there are multiple subjective duplicates of myself, there will be a definite fact as to whether any of them is the physical continuation of who I am now, and that at most one of them will be such a continuation (if we neglect the possibility of time travel).

    I see this contemporary enthusiasm for denying the existence of a self, the continuity through time of the self, etc, etc as merely a symptom of incomplete knowledge. The world looks atomistic; maybe one also believes in many quantum worlds or the unreality of time; there’s a sorites problem if I am supposed to be a pile of atoms – removed one by one, when does the pile stop being me? So instead we have the blithe suggestion that “Which lump of flesh is you is a matter of definition choice.”

    I would propose instead that the path ahead involves having a LOT more respect for subjective appearances. Don’t just say “the world looks like this, or I can imagine this thought-experiment, and it’s inconsistent with how things feel; well, so much the worse for these feelings, I will deny them or dismiss them as illusion”. That is all just a triumph of intellectualization over reality.

    I will concede this: it may be *extremely* difficult to get this right. Asking someone who’s currently getting it wrong to start getting it right may be asking them to make a shift in perspective as big, and as difficult to invent, as anything in human history. I think reaching the truth will require both very very subtle shifts of perspective *and* radically new “technical” concepts. I suspect we already have the kernel of it in the idea that consciousness is “in the brain”. But what’s wrong is to then sacrifice every subjective intuition that conflicts with current physical *theory* and current neuroscientific *theory*. Appearances also have their own reality, deserve more respect than this, and in fact will be our best guide towards the next fundamental reconception of the world – *if* we respect them enough.

  6. I highly recommend reading “Phantoms of the Brain” by V.S. Ramachandran, in particular the chapter “Do Martins See Red” which is about qualia and the sense of self. A quote:

    “The embodied self: My Self is anchored within a single body. If I close my eyes, I have a vivid sense of different body parts occupying space […] And yet the body image, as we have seen, is extremely malleable, despite all its appearance of stability. With a few seconds of the right type of sensory stimulation, you can make your nose three feet long or project your hand onto a table (Chapter 3)! […] Partial damage to these structures can cause gross distortions in body image; the patient may say that her left arm belongs to her mother or […] claim that the left half of her body is still sitting in the chair when she gets up and walks! If these examples don’t convince you that your “ownership” of your body is an illusion, nothing will.

  7. Really interesting post, lots of thinking to do with it, I digg it, sounds pretty related to my last entry where I try to say something coherent about consciousness itself like:I thought that consciousness as well as space and time could be thought as constructed frameworks, artificially generated contexts, and also as boundaries were existing entities can be identified. Actually one common definition of consciousness is a certain context
    you can check the rest:


  8. Nice article. This problem has been solved, among others, by Daniel Kolak, who has written a book called “I Am You” and a shorter paper called “Room to a View”.

    The solution is that we’re everyone. It was called “open individualism”. It may seem intuitively absurd, but it’s logically true.

  9. Conscious Observer

    “Can someone explain to me?”


  10. Pingback: Population ethics and personal identity | Meteuphoric

  11. Alexander Kruel

    A few days ago I wrote down some quick thoughts on this topic over at Less Wrong..

  12. Pingback: On the Anthropic Trilemma | Meteuphoric

  13. The main question this raises for me is: OK, but what things *aren’t* like that? I’m not saying there are none, but I suggest you think about what sorts there might be. I’d be interested to hear your thoughts too.

    Does this no-thread idea apply to all things that take up space and persist through time a bit? And does it apply to other things? You might think a similar no-thread thing applies to instantaneous spatiotemporal things, but just modally rather than modally-and-temporally – i.e. that with the instanteneous spatiotemporal things, there is no one true answer about which counterfactual scenarios would have involved those things and which not.

    And does it ever not apply to spatiotemporal things? And does it ever apply to abstract (non-spatiotemporal) things? You might think it sometimes applies to, say, games considered as abstract entities.

    (I’ve looked at your blog before but came here this time from listening to some interesting podcast episodes with you and Robin Hanson, since he just announced some new ones at OB. I listened to Signalling and School.)

Leave a Reply to mitchell porter Cancel reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.