# Split off from EPR paradox revisited

1. Feb 9, 2006

### ttn

Apparently I didn't make clear what bothers me. I'm happy to set aside the silliness of there being two Bobs, one in each branch that occurs after Bob makes a measurement. And I concede that your idea of choosing, at random, by the Born rule, one of those branches for Bob's consciousness-token to follow, allows for a consistent picture from Bob's POV.

You suggest that if I'm unhappy with the mindless hulks, I should just add additional consciousness tokens during each branching. That's also not really what I was getting at, but wouldn't this become a serious flaw for this model then? Suppose Bob makes a measurement on some state such that there is 90% probability for outcome A and 10% for outcome B. Well, if I add the extra consciousness token, now both branches will with certainty contain such a token. In other words, it is now 100% certain that both outcomes are equally real, equally experienced by Bob. How are you going to then explain the meaning of those original 90/10% probabilities? Your way of giving such probabilities meaning in the original theory (there is a 90% chance that Bob's consciousness moves from an "i'm about to perform this experiment" state to a "i just performed it and got outcome A" state) is ruined by making all the subsequent branchings equally blessed by a token. Or do I misunderstand something?

Anyway, what was really bothering me about this was how fast and loose it plays with the idea of what "really happens". Basically, this notion is cashed out in this theory in terms of the consciousness tokens following one or the other of the branches. So, going back to the original one-token model, what "really happens for Bob" (i.e., what Bob actually experiences to be real) is (say) outcome A. And one can make a similar comment about what "really happened" for Alice. And then you want to tell this story about Alice and Bob getting together later to compare notes, and you want to claim that they agree just the right fraction of the time, so voila, we have a local explanation of the correlations. My point was just to flesh out with more clarity what exactly this involves. Usually MWI people talk as if it "merely" involves rejection of the idea that the outcomes on each side are definite. But look at the consequences of that: at the coffee shop, the "real outcomes" of Alice and Bob are *never compared*. And you can't get around having some notion of "real outcomes" becuase it's just an empirical fact that if a person makes one of these measurements, he finds himself afterwards believing that it had a definite outcome (a belief which is false according to MWI, but so what -- we can just use that belief to redefine "real"). So that's my point. It's not just that Alice is in the uncomfortable situation of being still in a superposition until she gets to the coffee shop, at which time all the superpositions are resolved and everything is OK. No. Rather, the "real" Alice goes to the coffee shop, but she never ends up having the intended conversation with the "real" Bob. They never meet (because, at least sometimes, their cs-tokens are in different branches). So the intended comparison of outcomes from the two sides never in fact happens.

I don't think this is a new point to you, and I certainly don't expect you to change your mind about MWI because of it. I just wanted to clarify for others how extreme a price you pay for going down this route.

Yes, it does tend to come back to that solipsistic attitude, doesn't it? As long as all we're ever *really* talking about is my beliefs (how things "appear" to me, etc.) then the theory "works". But then unfortunately it no longer makes any sense to describe the theory as local. If everything I take to be an external world is really a (deluded) belief in my head, then it makes no sense to claim that causal interactions in the external world are always sub-luminal.

Not for me. But that's subjective. But the point we should in principle be able to agree about is this: claims about "locality" are literally meaningless if what the theory turns out to be fundamentally *about* is mere appearances/beliefs.

Trust me, I know! :tongue2:

2. Feb 9, 2006

### vanesch

Staff Emeritus
There are two Bobs now, but the original one is in the 90% branching, and the new one in the 10% branch. As if I put you in a copying machine. Your copy will *think* it is you, but *you* will still be *you*. You just went into the copying machine (as if you went in your car, or in your bed). Imagine I equip your car with a copying sensor. Next time you go into your car, you're (without knowing) scanned, and a copy of you is fabricated.
That copy will of course start out with all of your memories, and hence *think* it was leading your life up to then (it might just have a deja-vu experience when it comes out of my lab, and said to itself: huh, I could have sworn that I already stepped into my car - huh, I can be distracted these days!). But in fact it was just fabricated. You will still continue your life as usual.
Somebody meeting your copy will, during the conversation, be convinced it is you (and so will the copy). But you, of course, will not be aware of any of this. You were really driving your car. Nothing special happened to you.

Consider the "you" in the car as the one following the Born rule, and the copy as all the others. Both live of course in the same body now, but that body is in different states. And there is now not this differentiating aspect of sitting in the car or not.

3. Feb 9, 2006

### vanesch

Staff Emeritus
I split this off from the other thread in order not to hijack it.

4. Feb 9, 2006

### vanesch

Staff Emeritus
Well, it's one step less than true solipsism (but it goes in that direction, I agree). We still take it that there *is* an objective world (which is denied by true solipsism). Only, what is observed subjectively is put in a complex relation with the real world. That's not exactly the same as saying that what is subjectively observed is NOT originating in any kind of objective world.

As such, the constructed objective world can still satisfy locality properties or not in its dynamics. And the constructed objective world does not contain any ambiguities: it is the wavefunction.

But honestly, it really goes beyond the utility of MWI to discuss these things, because they are for ever unobservable. The difference between a "mindless hulk", a "copy of a person" or "the person itself" is unobservable, and we usually only _assume_ consciousness in other people by social convention. We can hence do that too with the "mindless hulks" or the "copies". They will not behave differently.

5. Feb 9, 2006

### ttn

Yes, I think we agree about the status when things are put carefully. Like I said, my point wasn't really to convince you of anything you didn't already know, but just to make clearer (for people who maybe haven't thought about this as much) the extent of what we have to give up to "save locality" ala MWI.

Yes, I know.

There was one lingering point that I didn't agree with you about, though. You suggested before that it was perfectly permissable to add extra consciousness-tokens so that, say, when Alice makes a spin measurement and splits the universal wf in two, *both* branches are "inhabited" by an Alice's-consciousness-token. I'm not so sure this is really possible. Isn't the whole point of the consciousness tokens in the first place basically to give some meaning to the whole probabilistic structure of QM -- that, when the state prior to a measurement is such-and-such, the probabilities for various outcomes are thus-and-so? With a single consciousness token, you can make sense of this in terms of relative frequencies of results in the past. Suppose Alice has a bunch of particles all in the state

|psi> = \sqrt(90%) |up> + \sqrt(10%) |down>

and starts measuring their spins one at a time. Well, after measuring, say, the first hundred, her "token" will have followed some path down the branching structure such that she believes she saw a sequence like: up, up, down, up, up, up, up, up, up, down, up, up, up, up, up, up, up, up, up... or whatever. Something where about 90% of the outcomes were "up" and about 10% "down". So the probabilities of which QM speaks are cashed out, in this theory, in terms of the relative frequencies with which various outcomes are *experienced* (in the past). Are we on the same page about all that?

I don't see how you can do anything like this if there are two consciousness tokens, one in each split branch for each split. Considering the same sequence of experiments as above, we have, with certainty, that there is a branch in which some Alice experienced "down, down, down, down, down, down, and so on..." some other branch in which some Alice experienced "up, down, down, down, down....." and so forth for *all* of the 2^N branches. So then what meaning can you give to a statement like "the probability to find "up" for a spin measurement on a particle in the above state |psi> is 90%"? You *can't* any longer cash this out in terms of the relative frequencies of experienced outcomes -- now *all* possible sequences of outcomes are equally experienced (by one or the other of the Alices). Indeed, the theory is no longer stochastic in any sense whatever (that I can see). I suppose you could try to give meaning to the probabilities by counting the number of worlds in which a given sort of sequence appears, but this suffers from a standard litany of problems, not the least of which is that it clearly gives the wrong answer for the example considered above.

So... how exactly is one supposed to add extra consciousnesses to MWI? It seems to me that the version you standardly advocate (a single consciousness token which follows the branching according to Born's rule) is the only really viable version of this. (And I cringe to even admit that much! :tongue2: )

6. Feb 9, 2006

### Sherlock

The discussion of copies wrt the simultaneously existing, alternative realities of MWI has me a bit confused.

In my readings on this, I gathered that the different worlds are not regarded as copies of some basic reality, but are, rather, independent and simultaneously existing realities --- and that there are an inifinite number of realities, and an infinite number of realities are being generated each instant.

Nevertheless, accepting the above notion of objective reality, the laws of physics are presumably the same in each branch. So, we may behave as if there is only one branch --- and we're back to the usual notion of definite measurement outcomes, and a probability interpretation of quantum theory.

7. Feb 9, 2006

### vanesch

Staff Emeritus
Yes. But...

I think there are 2 ways out of this. The first one is the one I suggested: although all these branches are "inhabited" by conscious Alices (many of them who don't observe the Born rule) the ORIGINAL Alice did follow the Born rule. I tried to suggest that with the car example: it is not because I'm making a conscious copy of you when you step into your car, that you experience suddenly NOT being in your car. A NEW COPY is created somewhere else (who will have a rather strange experience, because it will think it was you, stepping in your car, and then suddenly found itself elsewhere, in my Dr. Frankenstein lab). You could even postulate that from the moment it gets created, it also follows the Born rule, and it are ITS copies that get the non-born terms. As such, it will only experience one "non-Born" event in its remembered life: namely when it got created (split off from its original).
So if you remember a weird event in your life, maybe you are simply a copy! :tongue2: And the original Travis got convinced in being an MWIer somewhere else by one of my previous posts (which should have had a high probability) :rofl:

Not from the outside, indeed. But for the original Alice, it is. Her path is still the same, following the token going through the successive splits, as if there were only one. I mean, in nothing, you are affected, on your own subjective path, by the fact (or not - that's why I don't care) that there might be copies of you, which are created at a certain moment, no ? Call it the "principle of distinguishability of consciousness from the inside" :-)

Yes, that's where I think "standard" MWI-ers do themselves a lot of unnecessary troubles. If you want to go this road, I don't see why one should have equal probabilities for each branch, in other words, why one should do "branch counting". It seems natural, but there's nothing wrong with assigning OTHER a priori probabilities, and the obvious choice goes to the Born rule of course.
I mean, in the scheme of things, where you are "one of several", there is no absolute need for the probability for you to be in a state A to be given by a uniform distribution over all states. It could be another one, like the Born rule. It depends on how "you are thrown into the game". Maybe states with a big Hilbert norm have higher "cross section" to capture you (and your mates if you want to) .
Another way out, where you can do the "counting" stuff, is this (it gets weirder and weirder ) Imagine the Great Spaghetti Monster throwing out little souls which will lead a Patrick's life for a few milliseconds, say he has 10^2000 of them. He's not obliged to distribute them *equally* over the *number* of branches. There's no problem with having, say, multiple "souls" to one single branch: a body state could be a host to more than one (identical) subjective experience, in the same way as a body state could not be a host to a subjective experience (mindless hulk). States with large hilbert norm are then supposed to have more souls attached to them than states with smaller norm, so now when you do the counting, you arrive at the right probabilities. In as much as in a discrete setting, this sounds a bit strange (which is nothing else but a pictoreske way to illustrate non-uniform a priori probabilities!), in a continuous setting, this starts to make more sense.

I advocate it because it is the only thing that *matters* for an observer is HIS/HER little path through the successive states. And all the rest (such as written above) are funny speculations which don't add much to the story (except for the fun of it :-)

You're right, though, that one has to go through a whole stuff of weirdness to save locality!

8. Feb 10, 2006

### ttn

I don't see this as helping. It only resolves the problem by retreating, in essence, to the original (one consciousness token) version. I mean, you still need some kind of distinction between the one "special" branch that is "really" Alice or whatever, and all those other branches. Without some notion of specialness which only one of the branches possesses, there is literally nothing for the probabilities to be probabilities *of*. If everything happens and all those options are equally real, equally really happening, there is literally no meaning to a claim like "the probability that such and such happens is thus and so." So, yes, you can add extra consciousness tokens as much as you want -- but only if the ones you add are somehow inferior -- somehow not the "real" Alice.

Would you agree with that?

But if you're going to go to this "from the inside perspective", then I fail to understand what role probability can possibly play in the theory -- what meaning something like the Born rule can have. What are those probabilities supposed to be probabilities *of*? With a single consciousness token, there is a clear (if bizarre) meaning to the Born rule, and the theory can actually (if bizarrely) account for why I have the Born-rule-respecting experiences that I have always had. With exponentially splitting consciousness tokens, it's true that the consciousness in each branch experiences a "normal" set of experiences, at least in so far as they think their experiments have had definite outcomes and whatnot; but their reality is *likely* to be quite *abnormal* in the sense that their experiments' outcomes don't tend to obey the Born rule! So then one is left asking: how come *my* experiences conform to the Born rule if, according to this theory, it's overwhelmingly likely that they *shouldn't*? That's the kind of thing that scientists normally use as a good reason to reject the theory in question (namely, it predicts the wrong things).

Sure, and like you said the awarenesses that result when a token hits a branch need be only momentary. Such a momentary awareness would, after all, presumably come along with a physical brain state such that the person believed he had lived and experienced a perfectly continuous, long, normal life up to that point, and would have no reason to expect that (a) that wasn't true and (b) it was about to end anyway. Then again, this is the kind of thinking that can get you locked away in the loony bin, and for good reason. To whatever extent the MWI path leads here, I say this just makes clear we've taken the wrong path. In other words, if this isn't a reductio ad absurdum, I don't know what is.

Yup. I'm sure we don't agree about exactly how to quantify such things, but my opinion is that what we end up having to accept to save locality, is far far worse/weirder/crazier/stupider than the nonlocality we were trying to avoid. Locality is mere peanuts (some apparent implication of a relatively recent theory in one small branch of science) compared to what we have to give up to save it (which basically includes every item of common-sensical and scientific knowledge accumulated over millenea).

9. Feb 10, 2006

### vanesch

Staff Emeritus
It's what makes you so special :!!) :rofl:

Apart from the loony bin , it is yet another (totally hypothetical and speculative) possible view: consider that you have actually only one consciousness, but that it "multitasks" over your different possible bodystates, with an "intensity" (a kind of time sharing) which is given by the Born rule.
I once read a paper on the arxiv where it was argued that this kind of argument makes you in fact immortal, and was spiced up with a 'quantum suicide' party, where everybody plays the Russian roulette driven by a quantum experiment (the Schroedinger cat, version Scream or so) :uhh: Another version of it was a kind of "improbability drive": if you want to live an improbable event, make sure you die if it doesn't happen

But all these specific "schemes" speculate way beyond the simple statement, which is:
1) objective nature is described by the wavefunction
2) I'm subjectively aware of ONE of my bodystates, picked out by the Born rule.

It's all you need. All the rest (what happens to others, your copies, etc...) is of no real importance, and is utter speculation - which can, of course, be very funny :tongue2:

I don't think it is *so* weird as being totally absurd (only partly absurd ). It cuts in with specific philosophical ponderings, but which we usually could put aside, as having nothing to do with science, and here, they roar up their ugly head. But as such, these ponderings are not totally absurd.

No, for a simple reason. I already said several times I only adhere to MWI for the intuition and explanatory power it has concerning the formalism of quantum theory, and not as some kind of new age mysticism. As such, I prefer to limit myself to the rules 1. and 2. enunciated above.
If I would led go locality, for sure I could also explain some EPR effects, but I would be puzzled by two things, for which I would not have the slightest intuition: 1) how come we cannot build a FTL phone ; and 2) why do all equations need to be written in a Lorentz invariant way. And I'd get totally lost in spacetime, as its geometrical meaning would fall apart. As such, _as an interpretational scheme that HELPS me devellop an intuition for the formalism of current physics_, I would find every scheme that requires non-locality a counter-productive view.

Why do I keep locality ? It helps my intuition about relativity. Why do I keep parallel worlds ? It helps my intuition about the superposition principle. Why do I NOT adhere to world counting (as do most MWI-ers) ? Because that doesn't help me in understanding the right probabilities of observation (which are given by the Born rule).
To me, an ontological picture is simply a tool to devellop an intuition for a formalism.

10. Feb 10, 2006

### ttn

I laughed out loud when I got to the end of that paragraph -- and remembered that it began with "apart from the loony bin". :rofl: :yuck:

Yes, yes, I agree "it's all you need." My point was just that the rest isn't completely up for grabs. It's possible to supplement the "minimal" theory in such a way that you end up with a theory that no longer works in the way you were supposed to require. But it's you, not me, who has to really worry and care about exactly how to make this loony bin stuff work!

Fair enough, and I do understand your sentiments here; I just don't agree with them. And i think that's probably because we put different levels of faith in the quantum *formalism*. For me, the formalism (including such things as the requirement that laws be lorentz invariant) is just some fancy math to help us organize what we learn, over the cumulative centuries, from perceptual data (including perceptual data of such things as where pointers on apparati point). The formalism itself is completely negotiable, and if some theory can explain all the same observed facts with a different formalism, we should consider that and compare the alternatives based on their physics (not formal) merits.

Well, it's probably too strong to say the formalism is "completely negotiable." Within some given context, it's totally reasonable to extrapolate the formalism, to expect it to hold in roughly the same way in new situations, etc. It's a guide. But it's not written in stone. Perhaps the best example of the importance of this not-written-in-stone attitude toward the formalism is provided by Bohm's theory. Even at the level of non-rel QM, Bohm shows us that the whole fancy formal structure of Hilbert space and operators-as-observables and all that stuff can be completely jettisoned in favor of a very different looking formalism which has (as it turns out) all the same empirical predictions -- and, as it also turns out, allows us to *deduce* the usual formalism (operators as observables, etc.) as logical consequences. I take exactly the same "open" perspective toward things like special relativity (and the subsequent requirement of lorentz invariance). If some very un-relativistic-looking theory is able to account equally well for all the observed facts that motivated relativity in the first place -- and if in addition it solves some other physical riddles -- then we should be seriously open to it, and not reject it just because it's not of the form (i.e., like the formalism) we're used to. For example, a Bohmian theory of Dirac-equation-obeying particles with instantaneous action at a distance and a preferred reference frame (ala a pre-relativistic lorentz ether theory!) is blatantly in conflict with relativity, yet, it turns out, predicts that *observable effects* will have the usual lorentz invariant structure.

Well, I'll stop blathering there. As I was just typing, I had a feeling of deja vu, because it occured to me exactly what you will say in response (because I've heard you say it in response before, presumably in response to more or less the same blathering!). "Why should the wave equation in the Bohmian theory (viz, Dirac's equation) be Lorentz invariant if you're dropping that as a fundamental criterion on laws?" Right? :rofl:

Because quantum equilibrium washes out the FTL effects at the level of direct observations. The randomness in the initial conditions prevents you from harnessing the nonlocal causality for your personal nefarious communicational purposes.

Not all do! But some do, and I have no really good answer as to why they should. So, yes, there are open questions at the boundaries on my side too. To recap: I need, in the next 50 years say, to develop a good explanation for why some of the fundamental laws should be lorentz invariant when others aren't (i.e., when we're dropping lorentz invariance as a fundamental requirement). You need, in the next 50 years, to figure out whether or not the table in front of you is really there or not and whether all your friends are really mindless hulks! :tongue2: Probably better you don' t answer this part...

I think that's really the basic methodological difference between us, which explains why you prefer MWI and I prefer Bohm. To me, a formalism is simply a tool to help me develop an intuition for an ontological picture!

...which reminds me of something I read recently (but I can't remember where... maybe the zajonc/greenstein book "the quantum challenge"??). At any rate, somebody said: The basic difference between Einstein and Bohr was that Einstein thought in pictures and Bohr thought in words. I thought that was a very eloquent statement. Einstein always insisted on a physical picture that made sense physically. Bohr didn't even believe in physical reality, and was only concerned with constructing a coherent verbal strategy for talking about things (and forbidding talk of others).

OK, enough posting for today!

11. Feb 10, 2006

### vanesch

Staff Emeritus
Yes, that's probably the main difference between our different POV. I consider that the formalism comes first, and the vision on it is just a tool to "put it in music", when the very exercise is to make sense of the formalism. And, as I said already several times, when another formalism comes along, I'm takers. The vision, being part of the formalism, goes in the dustbin when the formalism goes in the dustbin. I have no problems in *changing* a vision (and hence, I don't need the vision that will predominate in 50 years).

Yes, but it is far less intuitive to come to that conclusion. That's why I find (crazy as it may sound) the other vision more productive.

Deal! And when you did so in a convincing way, I am takers

I didn't dare to say so before, but THIS is in fact my real motivation for MWI: it gives me a huge explanatory power of the behaviour of people around me :rofl: :rofl:

Indeed, that's the main difference: for me, an ontology is simply a means to get a better intuition for a formalism. The reason I don't take your viewpoint is that I think that hoping to find the _real_ ontology of the world is a hopeless issue.

But it is a dispute that is old as the world! The dispute between Plato and Aristotle already was about that: Plato considered that the true ontology of the world was the world of Ideas (and we simply observed the Sensitive world, a mere shadow of them, thinking that that was "real" - sounds like MWI, no), and Aristotle thought that the only reality resided in the things themselves (which were observed) and that the world of Ideas was a meaningless concept.

12. Feb 11, 2006

### ttn

What about all the aspects of the ontology we've unraveled so far? Like that the earth is round, orbits around the sun, stuff is made of atoms, atoms have electrons in them, living organisms inherit their genes via DNA, there used to exist dinosaurs, fluid flowing past a certain shape produces lift that can be used to build airplanes, blackbody radiation has such and such a spectrum, etc..... You really think all of these things are delusions or that they'll maybe be overthrown tomorrow and that we should just give up on trying to find out what the world is really like? That strikes me as blatantly, embarrassingly unscientific.

Sure, I'm happy to align myself with Aristotle and have you align yourself with Plato. Both were of course great and influential thinkers. But Aristotle is the grandfather of science; Plato inspired religion.

But even Plato would have been scandalized by MWI... :yuck: