It is epistemic relativism.

Question 1 and 3 are explicitly about values, so I don't think they do amount to epistemic relativism.

There seems to be a genuine question about what happens and which rules govern it, and you are trying to sidestep it by saying "whatever happens - happens".

I can imagine a universe with such rules that teleportation kills a person and a universe in which it doesn't. I'd like to know how does our universe work.

There seems to be a genuine question here, but it is not at all clear that there actually is one. It is pretty hard to characterize what this question amounts to, i.e. what the difference would be between two worlds where the question has different answers. I take OP to be espousing the view that the question isn't meaningful for this reason (though I do think they could have laid this out more clearly).

Oh, it's much worse. It is epistemic relativism. You are saying that there is no one true answer to the question and we are free to trust whatever intuitions we have. And you do not provide any particular reason for this state of affairs.

Nice challenge! There's no "epistemic relativism" here, even if I see where you're coming from.

First recall the broader altruism analogy: Would you say it's epistemic relativisim if I tell you, you can simply look inside yourself and see freely, how much you care, how closely connected you feel about people in a faraway country? You sure wouldn't reproach that to me; you sure agree it's your own 'decision' (or intrinsic inclination or so) that decides how much weight or care you personally put on these persons.

Now, remember the core elements I posit. "You" are (i) your mind of right here and now, including (ii) it's tendency for deeply felt care & connection to the 'natural' successors of yours, and that's about what there is to be said about you (+ there's memory). From this everything follows. It is evolution that has shaped us to shortcut the standard physical 'continuation' of you in coming periods, as a 'unique entity' in our mind, and has made you typically care sort of '100%' about your first few sec worth of forthcoming successors of yours [in analogy: Just as nature has shaped you to (usually) care tremendously also for your direct children or siblings]. Now there are (hypothetically) cases, where things are so warped and that are so unusual evolutionarily, that you have no clear tastes: that clone or this clone, if you are/are not destroyed in the process/while asleep or not/blabla - all the puzzles we can come up with. For all these cases, you have no clear taste as to which of the 'successors' of yours you care much and which you don't. In our inner mind's sloppy speak: we don't know "who we'll be". Equally importantly, you may see it one way, and your best friends may see it very differently. And what I'm explaining is that, given the axiom of "you" being you only right here and now, there simply IS no objective truth to be found about who is you later or not, and so there is no objective answer as to whom of those many clones in all different situations you ought to care how much about: it really does only boil down to how much you care about these. As, on a most fundamental level, "you" are only your mind right now.

And if you find you're still wondering about how much to care about which potential clone in which circumstances, it's not the fault of the theory that it does not answer it to you. You're asking to the outside a question that can only be answered inside you. The same way that, again, I cannot tell you how much you feel (or should feel) for third person x.

I for sure can tell you you ought to behaviorally care more from a moral perspective, and there I might use a specific rule that attributes each conscious clone an equal weight or so, and in that domain you could complain if I don't give you a clear answer. But that's exactly not what the discussion here is about. 

I can imagine a universe with such rules that teleportation kills a person and a universe in which it doesn't. I'd like to know how does our universe work.

I propose a specific "self" is a specific mind at a given moment. The usual-speak "killing" X and the relevant harm associated with it means to prevent X's natural successors, about whom X cares so deeply, from coming into existence. If X cares about his physical-direct-body successors only, disintegrating and teleporting him means we destroy all he cared for, we prevented all he wanted to happen from happening, we have so-to-say killed him, as we prevented his successors from coming to live. If he looked forward to a nice trip to Mars where he is to be teleported to, there's no reason to think we 'killed' anyone in any meaningful sense, as "he"'s a happy space traveller finding 'himself' (well, his successors..) doing just the stuff he anticipated for them to be doing. There's nothing more objective to be said about our universe 'functioning' this or that way. As any self is only ephemeral, and a person is a succession of instantaneous selves linked to one another with memory and with forward-looking preferences, it really is these own preferences that matter for the decision, no outside 'fact' about the universe.

All agreeable. Note, this is perfectly compatible with the relativity theory I propose, i.e. with the 'should' being entirely up to your intuition only. And, actually, the relativity theory, I'd argue, is the only way to settle debates you invoke, or, say, to give you peace of mind when facing these risky uploading situations.

Say, you can overnight destructively upload, with 100% reliability your digital clone will be in a nicely replicated digital world for 80 years (let's for simplicity assume for now the uploadee can be expected to be a consciousness comparable to us at all), while 'you' might otherwise overnight be killed with x% probability. Think of x as a concrete number. Say we have a 50% chance. For that X, will you want to upload.

I'm pretty certain you have no (i) clear-cut answer as to threshold x% from which on you'd prefer upload (although some might have a value of roughly 100%). And, clearly (ii) that threshold x% would vary a lot across persons.

Who can say? Only my relativity theory: There is no objective answer, from your self-regarding perspective. 

Just like it's your intrinsic taste who determines how much or whether at all you care a lot about the faraway poor, or the not so faraway not so poor or for anyone really: it's a matter of your taste and nothing else. You're right now imagining going from you to inside the machine, and feel like that's simply you being you there w/o much dread and no worries, and looking forward to that being - or sequence of beings - 'living' essentially with nearly certainty another 80 years, then yes, you're right, go for it if the physical killing probability x% is more than a few %. After all, there will be that being in the machine, and for all intents and purposes you might call it 'you' in sloppy speak. You dread the future sequence of your physical you being destroyed and to 'only' be replaced by what feels like 'obviously a non-equivalent future entity that merely has copied traits, even if it behaves just as if it was the future you', then you're right to refuse the upload for any x% not close enough to 0%. It really is, relative. You only are the current you, including weights-of-care for different potential future successors on which there's no outside authority to tell you which ones are right or wrong.

Thanks! In particular also for your more-kind-than-warranted hint at your original w/o accusing me of theft!! Especially as I now realize (or maybe realize again) your sleep-clone-swap example, which indeed I love as an perfectly concise illustration, had also come along with at least an "I guess"-caveated "it is subjective", i.e. which some sense is really already included a core part of the conclusion/claim here.

I should have also picked up your 'stream-of-consciousness continuity' vs. 'substrate/matter continuity' terminology. Finally, the Ship of Theseus question thus, with "Temporal parts" vs. "Continued identity" would also be good links, although I guess I'd be spontaneously inclined to dismiss part of the discussion of these as questions 'merely of definition' - just until we get to the question of the mind/consciousness, where it seems to me indeed to become more fundamentally relevant (although, maybe, and ironically, after relegating the idea of a magical persistent self, maybe one could say, also here in the end it becomes slightly closer to that 'merely question of definition/preference' domain).

Btw, I'll now take your own link-to-comment and add it to my post - thanks if you can let me know where I can create such links; I remembered looking and not finding it ANYWHERE even on your own LW profile page.

Yep.

And the clue is, the exceptional one refusing, saying "this won't be me, I dread the future me* being killed and replaced by that one", is not objectively wrong. It might quickly become highly impractical for 'him'** not to follow the trend, but if his 'self'-empathy is focused only on his own direct physical successors, it is in some sense actually killing him if we put him in the machine. We kill him, and we create a person that's not him in the relevant sense, as he's currently not accepting the successor; if his empathic weight is 100% on his own direct physical successor and not the clone, we roughly 100% kill him in the relevant sense of taking away the one future life he cares about.

*'me being destroyed' here in sloppy speak; it's the successor he considers his natural successor which he cares about.

**'him' and his natural successors as he sees it.

Good decisions need to be based on correct beliefs as well as values.

Yes, but here the right belief is the realization that what connects you to what we traditionally called your future "self", is nothing supernatural i.e. no super-material unified continuous self of extra value: we don't have any hint at such stuff; too well we can explain your feeling about such things as fancy brain instincts akin to seeing the objects in the 24FPS movie as 'moving' (not to say 'alive'); and too well we know we could theoretically make you feel you've experienced your past as a continuous self while you were just nano-assembled a mirco-second ago with exactly the right memory inducing this beliefs/'feeling'. So due to the absence of this extra "self": "You" are simply this instant's mind we currently observe from you. Now, crucially, this mind has, obviously, a certain regard, hopes, plans, for, in essence, what happens with your natural successor. In the natural world, it turns out to be perfectly predictable from the outside, who this natural successor is: your own body. 

In situations like those imagined with cloning thought experiments instead, it suddenly is less obvious from the outside, whom you'll consider your most dearly cared for 'natural' (or now less obviously 'natural') successor. But as the only thing that in reality connects you with what we traditionally would have called "your future self", is your own particular preferences/hopes/cares to that elected future mind, there is no objective rule to tell you from outside, which one you have to consider the relevant future mind. The relevant is the one you find relevant. This is very analogous for, say, when you're in love, the one 'relevant' person in a room for you to save first in a fire (if you're egoistic about you and your loved one) is the one you (your brain instinct, your hormones, or whatever) picked; you don't have to ask anyone outside about whom that should be.

so if there is some fact of the matter that you don't survive destructive teleportation, you shouldn't go for it, irrespective of your values

The traditional notion of "survival" as in invoking a continuous integrated "self" over and above the succession of individual ephemeral minds with forward-looking preferences, must be put into perspective just as that of that long-term "self" itself indeed.

There's a theory that personal identity is only ever instantaneous...an "observer. moment"... such that as an objective fact, you have no successors. I don't know whether you believe it. If it's true , you epistemically-should believe it, but you don't seem to believe in epistemic norms.

There's another, locally popular , theory that the continuity of personal identity is only about what you care about. (It either just is that, or it needs to be simplified to that...it's not clear which). But it's still irrational to care about things that aren't real...you shouldn't care about collecting unicorns...so if there is some fact of the matter that you don't survive destructive teleportation, you shouldn't go for it, irrespective of your values.

Thanks. I'd be keen to read more on this if you have links. I've wondered to which degree the relegation of the "self" I'm proposing (or that may have been proposed in a similar way in Rob Bensinger's post and maybe before) is related to what we always hear about 'no self' from the more meditative crowd, though I'm not sure there's a link there at all. But I'd be keen to read of people who have proposed theoretical things in a similar direction.

There's a theory that personal identity is only ever instantaneous...an "observer. moment"... such that as an objective fact, you have no successors. I don't know whether you believe it.

On the one hand, 'No [third-party] objective successor' makes sense. On the other hand: I'm still so strongly programmed to absolutely want to preserve my 'natural' [unobjective but engrained in my brain..] successors, that the lack of 'outside-objective' successor doesn't impact me much.^[1]

1. ^{^}

   I think a simple analogy here, for which we can remain with the traditional view of self, is: Objectively, there's no reason I should care about myself so much, or about my closed ones; my basic moral theory would ask me to be a bit less kind to myself and kinder to others, but given my wiring I just don't manage to behave so perfectly.

[edit: pinned to profile]

"Hard" problem

That seems to rely on answering the "hard problem of consciousness" (or as I prefer, "problem of first-person something-rather-than-nothing") with an answer like, "the integrated awareness is what gets instantiated by metaphysics".

That seems weird as heck to me. It makes more sense for first-person-something-rather-than-nothing question to be answered by "the individual perspectives of causal nodes (interacting particles' wavefunctions, or whatever else interacts in spatially local ways) in the universe's equations are what gets Instantiated™ As Real® by metaphysics".

(by metaphysics here I just mean ~that-which-can-exist, or ~the-root-node-of-all-possibility; eg this is the thing solomonoff induction tries to model by assuming the root-node-of-all-possibility contains only halting programs, or tegmark 4 tries to model as some mumble mumble blurrier version of solomonoff or something (I don't quite grok tegmark 4); I mean the root node of the entire multiverse of all things which existed at "the beginning", the most origin-y origin. the thing where, when we're surprised there's something rather than nothing, we're surprised that this thing isn't just an empty set.)

If we assume my belief about how to resolve this philosophical confusion is correct, then we cannot construct a description of a hypothetical universe that could have been among those truly instantiated as physically real in the multiverse, and yet also have this property where the hard-problem "first-person-something-rather-than-nothing" can disappear over some timesteps but not others. Instead, everything humans appear to have preferences about relating to death becomes about the so called easy problem, the question of why the many first-person-something-rather-than-nothings of the particles of our brain are able to sustain an integrated awareness. Perhaps that integrated awareness comes and goes, eg with sleep! It seems to me to be what all interesting research on consciousness is about. But I think that either, a new first-person-something-rather-than-nothing-sense-of-Consciousness is allocated to all the particles of the whole universe in every infinitesimal time slice that the universe in question's true laws permit; or, that first-person-something-rather-than-nothing is conserved over time. So I don't worry too much about losing the hard-problem consciousness, as I generally believe it's just "being made of physical stuff which Actually Exists in a privileged sense".

The thing is, this answer to the hard problem of consciousness has kind of weird results relating to eating food. Because it means eating food is a form of uploading! you transfer your chemical processes to a new chunk of matter, and a previous chunk of matter is aggregated as waste product. That waste product was previously part of you, and if every particle has a discrete first-person-something-rather-than-nothing which is conserved, then when you eat food you are "waking up" previously sleeping matter, and the waste matter goes to sleep, forgetting near everything about you into thermal noise!

"Easy" problem

So there's still an interesting problem to resolve - and in fact what I've said resolves almost nothing; it only answers camp #2 [LW · GW], providing what I hope is an argument for why they should become primarily interested in camp #1 [LW · GW]. In camp #1 [LW · GW] terms, we can discuss information theory or causal properties about whether the information or causal chain that makes up the things those first-person-perspective-units ie particles are information theoretically "aware of" or "know" things about their environment; we can ask causal questions - eg, "is my red your red?" can instead be "assume my red is your red if there is no experiment which can distinguish them, so can we find such an experiment?" - in which case, I don't worry about losing even the camp #1 form of selfhood-consciousness from sleep, because my brain is overwhelmingly unchanged from sleep and stopping activations and whole-brain synchronization of state doesn't mean it can't be restarted. 

It's still possible that every point in spacetime has a separate first-person-something-rather-than-nothing-"consciousness"/"existence", in which case maybe actually even causally identical shapes of particles/physical stuff in my brain which are my neurons representing "a perception of red in the center of my visual field in the past 100ms" are a different qualia than the same ones at a different infinitesimal timestep, or are different qualia than if the exact same shape of particles occurred in your brain. But it seems even less possible to get traction on that metaphysical question than on the question of the origin of first-person-something-rather-than-nothing, and since I don't know of there being any great answers to something-rather-than-nothing, I figure we probably won't ever be able to know. (Also, our neurons for red are, in fact, slightly different. I expect the practical difference is small.)

But that either doesn't resolve or at least partially backs OP's point, about timesteps/timeslices already potentially being different selves in some strong sense, due to ~lack of causal access across time, or so. Since the thing I'm proposing also says non-interacting particles in equilibrium have an inactive-yet-still-real first-person-something-rather-than-nothing, then even rocks or whatever you're on top of right now or your keyboard keys carry the bare-fact-of-existence, and so my preference for not dying can't be about the particles making me up continuing to exist - they cannot be destroyed, thanks to conservation laws of the universe, only rearranged - and my preference is instead about the integrated awareness of all of these particles, where they are shaped and moving in patterns which are working together in a synchronized, evolution-refined, self-regenerating dance we call "being alive". And so it's perfectly true that the matter that makes me up can implement any preference about what successor shapes are valid.

Unwantable preferences?

On the other hand, to disagree with OP a bit, I think there's more objective truth to the matter about what humans prefer than that. Evolution should create very robust preferences for some kinds of thing, such as having some sort of successor state which is still able to maintain autopoesis. I think it's actually so highly evolutionarily unfit to not want that that it's almost unwantable for an evolved being to not want there to be some informationally related autopoietic patterns continuing in the future.

Eg, consider suicide - even suicidal people would be horrified by the idea that all humans would die if they died, and I suspect that suicide is an (incredibly high cost, please avoid it if at all possible!) adaptation that has been preserved because there are very rare cases where it can be increase the inclusive fitness of a group the organism arose from (but I generally believe it almost never is a best strategy, so if anyone reads this who is thinking about it, please be aware I think it's a terribly high cost way to solve whatever problem makes it come to mind, and there are almost certainly tractable better options - poke me if a nerd like me can ever give useful input); but I bring it up because it means, while you can maybe consider the rest of humanity or life on earth to be not sufficiently "you" in an information theory sense that dying suddenly becomes fine, it seems to me to be at least one important reason that suicide is ever acceptable to anyone at all; if they knew they were the last organism I feel like even a maximally suicidal person would want to stick it out for as long as possible, because if all other life forms are dead they'd want to preserve the last gasp of the legacy of life? idk.

But yeah, the only constraints on what you want are what physics permits matter to encode and what you already want. You probably can't just decide to want any old thing, because you already want something different than that. Other than that objection, I think I basically agree with OP.

Very interesting question to me coming from the perspective I outline in the post - sorry a bit lengthy answer again:

According to the basic take from the post, we're actually +- in your universe, except that the self is even more ephemeral than you posit. And as I argue, it's relative, i.e. up to you, which future self you end up caring about in any nontrivial experiment.

Trying to re-frame your experiment from that background as best as I can, I imagine a person having an inclination to think of 'herself' (in sloppy speak; more precisely: she cares about..) as (i) her now, plus (ii) her natural successors, as which she, however, qualifies only those that carry the immediate succession of her currently active thoughts before she falls asleep. Maybe some weird genetic or cultural tweak or drug in her brain has made her - or maybe all of us in that universe - like that. So:

Is expecting to die as soon as you sleep a rational belief in such a universe?

I'd not call it 'belief' but simply a preference, and a basic preference is not rational or irrational. She may simply not care about the future succession of selves coming out at the other end of her sleep, and that 'not caring' is not objectively faulty. It's a matter of taste, of her own preferences. Of course, we may have good reasons to speculate that it's evolutionarily more adaptive to have different preferences - and that's why we do usually have them indeed - but we're wrong to call her misguided; evolution is no authority. From a utilitarian perspective we might even try to tweak her behavior, in order for her to become a convenient caretaker for her natural next-day successors, as from our perspective they're simply usual, valuable beings. But it's still not that we'd be more objectively right than her when she says she has no particular attachment for the future beings inhabiting what momentarily is 'her' body.

When your body sleeps, you die never to experience anything ever again.

This is an infohazard that can fuck up your sleep cycle. :D

Pretty much everyone in the society remembers "continuing on" many thousands of times.

Many people probably also remember that their previous selves remembered something they no longer remember, or felt enthusiastic about something they can't find motivation to do. So there is some evidence in both directions.

The upload +- by definition inherits your secret plan and will thus do your jumps.

Not in this case.

a) If you anticipated continuity of experience into upload and were right, then you experience being an upload and remember being you and and you believe that your prediction is borne out.

b) If you were wrong and the upload is conscious but isn't you, then you're dead and nothing is borne out to you. The upload experiences being an upload and remembers being you and believes that your prediction is borne out.

c) If you were wrong and the upload is not conscious, then you're dead and nothing is borne out to you. Nothing is borne out to the upload either, since it was never able to experience anything being borne out or not. The upload unconsciously mimics everything you would have done if your prediction had been borne out.

Everyone else sees you continuing as you would have done if your prediction had been borne out.

So in all cases, everyone able to experience anything notices that your prediction is borne out.

The same is true if you had predicted (b).

The only case where there is a difference is if you predicted (c). If (a) or (b) was true then someone experiences you being wrong, whether or not that person is you is impossible to determine. If you're right then the upload still behaves as if you were wrong. Everyone else's experience is consistent with your prediction being borne out. Or not borne out, since they predict the same things from everyone else's point of view.

The point is, "you" are exactly the following and nothing else: You're (i) your mind right now, (ii) including its memory, and (iii) its forward-looking care, hopes, dreams for, in particular, its 'natural' successor. Now, in usual situations, the 'natural successor' is obvious, and you cannot even think of anything else: it's the future minds that inhabit your body, your brain, that's why you tend to call the whole series a unified 'you' in common speak.

Now, with cloning, if you absolutely care for a particular clone, then, for every purpose, you can extend that common speak to the cloning situation, if you want, and say my 'anticipation will be borne out'/'I'll experience...'. But, crucially, note, that you do NOT NEED to; in fact, it's sloppy speak. As in fact, these are separate future units, just tied in various (more or less 'natural') ways to you, which offers vagueness, and choice. Evolution leaves you dumbfounded about it, as there is no strictly speaking 'natural' successor anymore. Natural has become vague. It'll depend on how you see it.

Crucially, there will be two persons, say after a 'usual' after cloning, and you may 'see yourself' - in sloppy speak - in either of these two. But it's just a matter of perspective. Strictly speaking, again, you're you right now, and your anticipation of one or future person.

It's a bit like evolution makes you confused about how much you care for strangers. Do you go to a philosophers to ask how much you want to give to the faraway poor? No! You have your inner degree of compassion to them, it may change at any time, and it's not wrong or right.*^[1]

1. ^{^}

   Of course, on another level, from a utilitarian perspective, I'd love you to love these faraway beings more, and its not okay that we screw up the world because of not caring, but that's a separate point.