Monday, April 22, 2013

A Somewhat Impractical Plan for Immortality

... and arguably evil, too, though let's set that aside.

My plan requires the truth of a psychological theory of personal identity, a "vehicle externalist" account of memory, and some radical social changes. But it requires no magic or computer technology, and arguably we could actually implement it.

Psychological theory of personal identity. Most philosophers think that personal identity over time is grounded by something psychological . Twenty-year-old you and forty-year-old you are (or will be) the same person because of some psychological linkage over time -- maybe continuity of memory, maybe some other sort of counterfactual-supporting causal connectedness between psychological states over time. Maybe traits, values, plans, and projects come into the picture, too. In practice, people don't have the right kind of psychological connectedness without having biological bodily continuity. But that, perhaps, is merely a contingent fact about us.

Vehicle externalism about memory. What is memory? If a madman thinks he is Napoleon remembering Waterloo, he does not remember Waterloo, even if by chance he happens upon exactly the same memory images as Napoleon himself had later in life. Memory requires, it seems, the right kind of causal connectedness to the original event. But need the relevant causal connectedness be entirely interior to the skull? Vehicle externalists about memory say no, there is nothing sacred about the brain-environment boundary. External objects can hold, or partly hold, our memories, if they are hooked up to us with the right kind of reliable causal chains. Consider Andy Clark's and David Chalmers's delightful short paper on Otto, whose ever-available notebook serves as part of his mind; or consider a science-fiction case in which part one's memory is temporarily transferred onto a computer chip and then later recovered.

Implementation. Could one's temporary memory reservoir be another person? I don't see why not, on a vehicle externalist account. And could the memories -- and the values and projects and whatever else is essential to personal identity -- then be transferred into another human body, for example, over the course of a decade or two into the body of a newborn baby as she grows up? I don't see why not, if we accept at least somewhat liberal versions of all the premises so far, and if we assume the most excellent possible shaping of that child.

By formatting a new child with your memories, your personality, your values, your projects, your loves, your hopes and regrets, you could thus continue into a new body. Presumably, you could continue this process indefinitely, taking a new body every fifty years or so.

As I said, a madman's dream of being Napoleon is no continuation of Napoleon. But the situation would be very different from that. There would be no madness. The memories would have well-preserved causal traces back to the original events; the crucial functional role of memory, to save those traces, would be perserved; everything would be held steadily in place by the person or people implementing this plan on your behalf, as a stable network of correctly functioning cognitive processes. And the result would be not just something on paper or in a memory chip but a consciously experienced memory image, felt by its owner to be a real authentic memory of the original event.

This could, it seems, be done with existing technology, using clever mnemonic and psychological techniques. One would need mnemonists who knew everything possible about you, who observed the same events and shared your same memories, and who were exceptionally skilled at preserving this information and transferring it to the child. The question then would be whether it would be true that the child, when she grew up, would really be you, with your authentic memories, instead of a mad Napoleon. And the answer to that question depends when whether certain theories of personal identity and memory are true. If the right theories are indeed true, then immortality -- or rather, longevity potentially in perpetuity -- would be possible for sufficiently wealthy and powerful people now, if they only chose to implement it.

I have written a story about this: The Mnemonists.


P.D. Magnus said...

This only works if there is no important distinction between declarative and episodic memory; roughly, the distinction between remembering facts about my fourth birthday and remembering having been there. Liberal vehicle externalism really only seems plausible for declarative memory.

Eric Schwitzgebel said...

Why not episodic, too? The mnemonist saw and remembers his visit with his uncle six years ago. She conveys this memory to the new young boy effectively enough that he experiences the memory as his own, including with visual imagery as accurate as our ordinary autobiographical visual imagery is (which is to say, not too accurate).

Neal Tognazzini said...

What a fun reductio! Of exactly what, I'm not sure, but it definitely smells like a reductio. The one big thing that seems to be missing from the story, to me, is the ability of an Oligarch to anticipate having the experiences of the new young boy. If it weren't for the reflexivity of identity, I'd almost want to say that in your story, the new boy is identical to the old Oligarch, but not the other way around. That is, the details of your story make it plausible for the young boy to say, "I am he", but implausible for the dying Oligarch to say "He is I."

P.D. Magnus said...

If that counts as transmission of episodic memory, then we should similarly admit that the madman might remember being defeated at Waterloo. All that is required is that it causally trace back to Napoleon (which certainly it does, since he's not some creative genius who has merely invented a character who is coincidentally named "Napoleon") and that the imagery is rather reliable (which it could be, even if for some particular madman it is not). It is unclear whether this is a reductio of your proposal or an exciting consequence.

Anonymous said...

"consider a science-fiction case in which part one's memory is temporarily transferred onto a computer chip and then later recovered."

Eric, I'd be interested to hear your reaction to this:

Actually, another person as the intermediary makes more sense to me than using a machine. We are already hard-wired to absorb influences from other human beings. However, if I may use the story below for examples, there is the question of the intermediate status. We have Oligarch I. His memories are transferred into Mnemonist I. After the death of the Oligarch, the memories are tranferred to a young boy who becomes Oligarch II. Mnemonist II starts her training.

This is in many ways analogous to the Tibetan Tulku system of succession-by-reincarnation, by the way. Tibetan history shows that it is an incredibly inefficient system: less than half the Dalai Lamas ever manage to wrest power from the hands of their regents. But we are not interested in the PolSci aspects right now. So, here are my problems:

1. After the death of Oligach I, in what way is the Mnemonist NOT the Oligarch? She has the memories. Law and custom, I suppose, but philosophically, I don't see how we can deny her that position.

2. In the same vein, immediately before the death of Oligarch I, how many Oligarchs are there?

3. Suppose the Mnemonist broke law and custom and secretly starts training a new Oligarch while the old one is still alive. NOW how many do we have? Could she train the Oligach's twin brother to usurp his place?

4. I don't see any requirement for Mnemonist I to commit suicide once Mnemonist II graduates and starts working. There are going to be an awful lot of copies of Oligarchical memories floating around. I'm all for triple redundancy, but if memory determines identity, then we now have a surfeit of oligarchs!

- Michel Clasquin-Johnson, posting as Anon because the Google ID link is down.

Pete Mandik said...

This technique plus a dash of cloning is a major plot point in SPOILER ALERTS...the live action Aeon Flux film as well as various episodes of the Venture Brothers (the titular brothers being highly death prone).

Eric Schwitzgebel said...

@ Neal: I'm inclined to agree that it's naturally read as a reductio, but as with my work on group consciousness, I think we should give non-trivial credence to the truth of the conclusion. I'm inclined to think the Oligarch can in some sense anticipate his experiences as a future boy: "Again, I will relive the awakening of sexuality; again, I will hear my mnemonist remind me of this treasured memory." Maybe that's not the right *kind* of anticipation, but if not, why not?

I hope the value of it as a thought experiment is that it forces us to clarify our theories of memory and personal identity to exclude Oligarchs, if we want to. Most current accounts are too vague to definitively exclude them, unless they require bodily continuity or deny the possible externalization of memory.

Eric Schwitzgebel said...

PD: I think in the true mad Napoleon case, the transmission is too haphazard and depends upon a failure of proper functioning rather than a success of proper functioning. The question, in my mind, is whether the same can be said of the Oligarch.

Eric Schwitzgebel said...

Michel: Terrific questions!

On 1: The mnemonist doesn't think of herself as Oligarch and has different life projects, e.g., to train the boy rather than to (say) renew acquaintance with his friend. The Oligarch's memories don't play the same functional role for her as they do for the Oligarch.

On 2: Still one Oligarch, but sleeping or on hold, as it were.

On 3 and 4: Here we have the standard objection to psychological continuant theories of personal identity, e.g., in double-transporter scenarios. (I assume you know these cases.) This is not a problem for this theory in particular. I have arranged the plot so there is a unique best continuant. I had toyed around with changing that, but I thought that was too much for one story and would muddy the message. My own inclination on these questions is broadly Parfitian.

Eric Schwitzgebel said...

Oh, and thanks for the link, Michel! That's a nice list of concerns, but I don't see that any of them are clearly fatal. 1, 4, and 5 and minority positions. 2 and 3 I'm inclined to agree with, actually, but are by no means universally accepted. I'm setting them aside for this thought experiment, but one possible reaction to the thought experiment is to think it supports something in a skeptical direction. The remaining concerns are ethical and practical, rather than in principle.

Eric Schwitzgebel said...

Pete: Thanks for the tip! What I need to do is download your entire science-fiction library directly into my brain. Are we going to do that conference sometime?

D said...

Interesting idea. It reminds me a lot of Cyteen by C.J. Cherryh. (spoilers) A powerful older woman has herself cloned and wills her clone all her material possesions, and does everything in her power, including prerecorded messages on the computer, to ensure the clone grows up to be just like her. In the story memories can be implanted by means of "tapes."

Callan S. said...

I'm not sure why this counts as living on?

What if we had a sealed house, and we birth a child into it (using an artificial womb). The child can't leave the house, ever, and experiences the precisely decided set up of the house (no other people are in it, lets say, except precisely decided recordings). Let's say it has machines to nurture it so it grows, etc.

Eventually the child dies (wow, now I feel evil...).

So we reset the house back to the very set up we had previously. And we birth a clone of the prior child into the house.

Does that qualify as living on? It's going to be the exact same life experiences, isn't it?

Or will the child, in small ways, act differently - potentially marked differences (like a missplaced pebble can become avalanche)?

And perhaps somewhat like some other peoples examples here, what happens when actually were running two houses at once, each exact duplicates (as much as our technology allows). If that's living on, is that also living in paralel?

Unknown said...

I think Parfit's approach to personal identity might be useful here. I'm sympathetic with your use of cognitive extension, and the idea that memory might be broadly constitutive of personal identity, but when it comes down to it my suspicion is that our understanding of selfhood is just too ambiguous. In one sense this would allow for immortality, and in another it wouldn't, and it's basically just indeterminate what we mean by personal identity.

As a means of testing our intuitions, and perhaps even demonstrating their incoherency, it's a great thought experiment!

(I actually considered something similar in my undergrad dissertation, but I'm no longer particularly convinced by the arguments:

clasqm said...

"The mnemonist doesn't think of herself as Oligarch and has different life projects, e.g., to train the boy rather than to (say) renew acquaintance with his friend. The Oligarch's memories don't play the same functional role for her as they do for the Oligarch."

OK, my other questions really pivot on this one. Let's play around with this a little longer. Note to university IT snoops: this is not goofing off at work, this is a serious academic conversation!

You are the famous philosopher, so famous that society has decided to perpetuate you by providing you with a Mnemonist: me. (I think we will agree that no conceivable society will consist of 50% citizens and 50% Mnemonists - this is always going to be an elitist thing). We are the same gender and there is a slight physical resemblance between us.

I shadow you for years, absorbing your memories. Then I realize something: you are really not that much smarter than I am. I learn, though your memories, all the little mental and rhetorical tricks of the trade. I see that had I been born into slightly different circumstances, that Nobel, that Pulitzer, could have been mine! My hatred grows and I hatch a terrible plan.

One day, you and I go off into the mountains on one of your accustomed searches for solitude. We get caught in a rockslide, engineered by me. You die, and I am horribly disfigured, but when they dig me out, I gasp "Is my Mnemonist okay?"

I assume the position of Eric Schwitzgebel. After extensive plastic surgery, even your family accepts me. I undertake extensive auto-hypnotic exercises to suppress my former memories. In time, these exercises become as automatic as breathing. One might say I have induced a psychotic breakdown in myself, a Multiple Personality Disorder with the one personality completely squashed. By the time of my death (for reasons I have long "forgotten" I have always declined the services of another Mnemonist and developed a phobia about fingerprints and DNA tests), I firmly believe myself to be Eric Schwitzgebel. The only memories I have are yours.

Who am I?

My life project here, while close to yours, was not identical. Yours was to do philosophy. Mine was to steal the glory of having done the philosophy. Or would the urge to do philosophy take over as the Schwitzgebel personality asserted itself?

Eric Schwitzgebel said...

Thanks for the awesome continuing comments, folks!

@ D: Thanks for the tip. I'll check it out!

@ Callan: It might have something to do with counterfactual dependence and causation. If Kid A *had* chosen to do X, then Y would have followed for Kid A, but not for Kid B. Hanging it on counterfactuals in this way might seem weird and tenuous, but that's my inclination at this point. (I also found myself forced into counterfactuals and causation several years ago contemplating Greg Egan's "Dust Hypothesis" a few years ago on this blog.)

Eric Schwitzgebel said...

@ Joseph: My own views are very close to Parfit's. In that spirit, you are welcome to read this story and post as a reductio ad absurdum of too rigorous an adherence to the flawed folk-psychological assumption of sharp facts about personal identity.

Eric Schwitzgebel said...

@ Michel: Wow, I love your follow-up story! See my response to Joseph immediately above this response on the metaphysical question. I don't think there needs to be a sharp fact of the matter.

Howie Berman said...

Isn't that how traditional cultures perpetuate their identity?
Let's suppose I do this while I'm still alive; would I then like Pythagoras was purported to be once in a while, in two places at the same time?

Scott Bakker said...

How much does this have to do with our need to simplify (and to keep our simplications simple)? Instead of asking whether the second Oligarch is in fact the first due to the machinations of the Mnemonist, why not look at all three - Oligarch, Mnemonist, Oligarch - as a single, diachronic system that becomes progressively more complicated over time? This way the granularity of the picture you're invoking is always indexed to the substantive facts of the systems involved. Saying Oligarch II 'is' Oligarch I becomes a simplifying shorthand for the far larger system that includes the Mnemonist.

clasqm said...

Scott Bakker: I think you've put your finger on exactly how the people in Eric's hypothetical society would see it. For them, it would be axiomatic that there is a line of oligarch-ness stretching back into the past and potentially, at least, stretching into the future. Just, as I noted above, as Tibetan Buddhists accept it as axiomatic that there is a line of tulkus. The individual currently occupying the place in that line is the least important aspect of it.

Whether we can stretch our western individualist worldviews (I think it was Alan Watts who came up with the term "skin-encapsulated ego") to really encompass such views is the question. Which is why I enjoy Eric's thought experiments so much.

Acid test: How many of us would take that hemlock? Personally, I'd need some more time to think it over. Like another fifty years, perhaps ...

P S Eric, sorry to have killed you in my previous post. But you'll be happy to know that you live on in me, won't you?

Eric Schwitzgebel said...

@ Howie: Perhaps!

@ Scott: Yes!

@ Michel: Man, I don't know about that hemlock. I claim the skeptic's privilege of complete befuddlement! No waiting 50 years, though, the Oligarchs need their replacement now and there's a backup candidate waiting in the wings.

Callan S. said...

Scott: So what supports the far larger systems continued repetition?

Is it passing the buck to further and further systems to presume an identity? Kind of like passing the buck to the strings that run from the arms, then passing the buck to the wooden cradle the strings are tied to above, to keep up the pretence the little wooden boy puppet is alive?

Callan S. said...

Eric - I'm not sure if I dig the dust hypothesis (pun unintended). Say the Mona Lisa also had a thinking brain built into it - but someone comes along and replaces it with a counterfit. Well, would you say it's still the Mona Lisa?

I'll grant I've envisioned conciousness riding atop the material base before. But really it's not actually seperate. Or atleast I don't think it's seperate and thus can somehow be subject to the various permutations of the dust hypothesis. The dust hypothesis seems to say 'see these mechanical interactions, well, they create this other thing called conciousness - now we have this other thing, lets look at that seperately (slice by slice)...etc etc'. But...there is no other thing involved. There can be no other Mona Lisa (only paralel development).

Eric Schwitzgebel said...

Callan: No other thing required on Dust -- at least there's not intended to be. Why think so?

Duplication/counterfeit is of course a huge issue in personal identity, e.g., with the help of transporter thought experiments. My own feeling is that if the duplicate/counterfeit is the same in every relevant qualitative respect and has the right sort of causal/counterfactual dependency, it shouldn't matter. I wouldn't even mind two distinct continuers, although that problematizes the strict sense of "identity".

Callan S. said...


Well the dust hypothesis blog entry mentions material AND concious thing.

Surely it's requiring that second thing, the conciousness thing, for it's theory?

In the blue planet RPG (spoilers) there are aliens that capture a woman and for some reason duplicate a copy of her at the atomic level, then for unknown reasons cast the unconcious woman into the ocean depths to be eaten by creatures. The copy escapes, but doesn't actually know what happened. Is this continuing, by your measure?

What if you found out such a thing had happened to you - would you pretend the copy event didn't occur? How can it be a coninuance when that copy event is glaringly something that isn't part of the originals life. Say the original lives - well, you've certainly ceased to copy their life at all now, theirs goes one way, yours, from the copy event, goes another.

Eric Schwitzgebel said...

It's weird and I don't feel at all confident about it, but setting aside my general reservations about materialism and functionalism, I'm inclined to think that the copy is me, and the drowning being was also me -- I split onto two paths.

Linda Nagata has some interesting plot explorations along these lines in Bohr Maker.

Callan S. said...

Okay - that seems rather altruistic!

It doesn't seem right to me though. Somehow I feel the balls back in my court on supporting a claim, so I have to go figure what the claim is first...

WinstonSmith714 said...

OMG!!! i've come across a bunch of geniuses!!! :O)