Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

What we see is bogus in MWI

  1. Feb 20, 2007 #1


    User Avatar

    So, in response to certain experimental outcomes which you SEE, you conclude that you can't trust what you see.

    Good logic. :rofl:
  2. jcsd
  3. Feb 20, 2007 #2


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    It's always your same caricatural objection, where you implicitly change "is not all there is" into "cannot be trusted" or "deluded", hence enlarging implicitly the original statement's conclusion, in order to try to show it evidently wrong.

    You change "I see the tip of the iceberg, hence there must be much more below the surface which I don't see" into "I see the tip of the iceberg, hence I must be deluded in my seeing", and from that, you obviously conclude that if you are deluded, then there was no reason to expect there to be an iceberg in the first place. But the two statements are different.
  4. Feb 21, 2007 #3


    User Avatar

    Not so! It's often one of several, different caricatural objections instead!

    The difference is that the two parts of an iceberg are *parts*, and may actually claim therefore a kind of independent existence. Each part is actually *there*. So it's no logical contradiction to be aware of the part above water, and unaware of the part below water.

    But superposed terms in a wave function are not like that, are not each individual "parts" which all have independent existence. If the wave function is all there is to describe the physical states of systems (which is the MWI position) and if the spin state of some spin half particle is

    |psi> = a|+> + b|->

    and you believe that the particle is definitely in the eigenstate |+>, your belief is *wrong*. Not "part truth", but just plain error/delusion. If you think the state is |+> when it is actually |psi>, you are exactly as wrong as you'd be if you thought it was |+> when it is actually |->.

    In any case, to recap, the logical problem I'm pointing out goes as follows: you say we (humans) go into labs and notice certain things, like pointers definitely pointing to certain directions and not others, and infer from this [after getting orthodox QM and then fighting for a while over how to interpret it] that, in fact, those pointers were *not* pointing the way we thought we saw them pointing. At the end of the day, MWI requires we accept that the best way to interpret our having seen certain physical objects in certain positions, is to deny that those objects were in fact in those positions at all.

    (In addition, it requires us to deny that there is even such a thing as physical objects moving and interacting in spacetime. ...which is pretty weird for a theory whose only reason for existing is: (allegedly) to save relativity in the face of the otherwise problematic Bell theorem. Wherever you look, MWI is swallowing itself by the tail, contradicting and undermining its own motivations and foundations.)

    My earlier post may have constituted "drive by shooting", but it was not caricatural.
  5. Feb 21, 2007 #4


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Why would I believe that, any more, than I would believe that all there is to the iceberg, is what is above the water ?

    Because, mind you, if the state of the system is a |+> + b|->, I will NOT see just |+> or |-> ! I will in fact not see anything. The only way I could possibly say that it is in fact in the |+> state, is when I INTERACTED WITH IT, and when the actual state is a |me>|+> + b |mycopy> |->.
    But that is, IMO, analogous to:
    (I above the water) (the tip of the iceberg) + (Joe in his diver suit)(the rest of the iceberg)

    So there's nothing wrong with me saying that I saw |+>. It is an (erroneous) extrapolation, however, to state that the particle IS ONLY |+>. It is |+> for me (and that is correct), while it is |-> for mycopy (and that's correct to, probably, though I have no way to know directly).

    As I said, this is not correct IMO. The particle cannot be in the state (a|+> + b|->) while I saw |+>. I must have entangled with it, otherwise I cannot have measured it.

    Again, no. They ARE pointing in the way we saw them pointing, FOR US, and they are pointing elsewhere for COPYUS. A quantum pointer is a pointer which can point in several directions at the same time (by definition!). This is not a delusion. The extrapolation which brings all the misery is to say that because they point that way for us (which is correct!), that's all there is to it, and they cannot point elsewhere for copyus. Although this is not directly observable (and this is the main incentive to discard this other part of reality), it can have indirect observational consequences in certain cases, such as in EPR-Bell.

    This is IMO exactly analogous as saying that if I am on the surface, that all there is to an iceberg, is its tip, because it is all I see, and denying that Joe in his diversuit can see the rest of it. Now, of course the difference is that the observational consequences of Joe are much more tangible: he'll come at the surface, or I can put on my diver suit etc... while I'll never meet "copyus" directly. But apart from that, my observing just the tip of the iceberg, and my inference from other observations that there must be an invisible part below the surface, are very close to the derivation of the existence of the copyus with the other result.

    Again, not at all. If we *observed* those objects (which, if they are quantum objects, have miriads of positions at the same time) at a certain position, then that means that FOR US, a "classical shadow" of that object WAS in that position state - but that doesn't exclude that for copyus, another classical shadow of that object was indeed elsewhere.

    Spacetime is (as long as we don't do quantum gravity) a background on which quantum objects are defined, but they do not need to have a single position! Only, their interactions with other quantum objects must be such that they can be *decomposed* over spacetime.
    You could just as well require the EM field to have a specific position in spacetime. It doesn't. It is everywhere. However, interactions between the EM field and other systems also occur in a decomposed way over spacetime. This is what relativity is about. Spacetime is the background over which interactions can be decomposed.
  6. Feb 21, 2007 #5


    User Avatar

    A particle can have only one state. If it's |+> it's not |psi> and vice versa. This is equivalent to the assertion that the wave function alone provides a complete description of the physical states. Someone who says "sure, the wave function is |psi> but *really* the particle is definitely spin-up" is someone who rejects the completeness doctrine (and hence rejects MWI).

    "is ... for" is nonsense. We're talking about what the actual physical state is here. This is (by definition) not something that can be "relative" to me/you/mycopy/whoever. If the state is one thing "for you" and another thing "for me" then what we're talking about isn't *the state*.

    I don't think it changes anything for you to be entangled. It's the same point to say that the state is

    a|+>|me> + b|->|mycopy>

    and you believe the particle to be definitely spin up. It isn't.

    "Classical shadow" is just a made up term to cover what you really mean: that your belief doesn't correspond to the actual facts (it corresponds only to the "shadow", which isn't actually real, so, in other words, your belief corresponds to nothing... i.e., isn't true).

    This really confuses me. Are you saying that the wave function (which is all there is, according to MWI) is a field on spacetime? You know that's not true. It's a field on some giant-dimensional configuration space (-time). And if the wave function is all your theory posits, there is literally nothing in space-time (and so I can only assume you don't bother positing any empty spacetime, either). Maybe you should take a look at the fantastic paper here:


    MWI is like "bare GRW" in the relevant respect (except it's even worse, since it's lacking in the non-linear terms which make bare GRW at least a step toward a solution to the measurement problem).

    As to the EM field not existing in spacetime, I don't know what you're talking about. Sure, the field as a whole is in no one particular place (duh), but still such a field is a field on 3-space such that one can consider the field values at each point a "local beable" (in the sense of Bell). What pray tell are the local beables in MWI? That is, what/where is the matter -- the stuff we see moving and interacting in 3-space?
  7. Feb 21, 2007 #6


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Again, I think you do not realise the difference between the state:

    a |+> + b |->

    on one hand,


    a |u> |+> + b |v> |->

    on the other.

    It would be an error to claim that in the last case, the state of the particle is

    a |+> + b |->

    without any reference to u or v.

    And indeed, in quantum theory, a subsystem doesn't necessarily have a state, only the overall system has. If you insist on giving a "state" to a subsystem, you have to specify also the relative states of the other subsystems in order for this to make sense, and that is exactly what we do. One of those subsystem states is "me", and in relation to me (which is part of u, say), the state of the particle is |+>, while in relation to v, the state of the particle is |->, and the particle, BY ITSELF, has no overall state. It is certainly NOT a |+> + b |->.

    Well, apart from you not liking that, I don't see any fundamental argument why each individual subsystem should necessarily need to have one and unique state. It is a basic axiom of quantum theory that it doesn't, and that only the state of the overall system has a state. This sounds maybe strange and maybe against our intuition, but relativity of simultaneity is also one of those strange notions, which is against our intuition, but therefore not logically false.

    In the same way, there is no logical requirement (but only an intuitive one) for each subsystem to have its own independent ontological state independent of the state of other subsystems. In classical physics, this is the case, and in quantum physics, this isn't.
    So, if you INSIST on giving a state to a particle in quantum physics, you have to describe it relatively to the states of other subsystems.

    Exactly: it isn't "spin up", it is only "spin up" to me. That is, relative to me, if I insist on giving it a unique state, I have to say that it is "spin up", knowing that relative to mycopy, it is "spin down". Now, there is a consistency, in that if there is a third subsystem, it will be in such a state, relative to me, that it is in consistency with the particle being in "spin up" to me.
    And given that in most cases, I will not hear anything anymore from "mycopy", I can, for EASINESS, now assume that the whole particle state is "spin up", because all I will ever observe, all other states that will ever be relative to me of other subsystems, will be consistent with the "spin up" state of my particle. So this is why I have the impression that ALL THERE IS, now, is this particle in a spin up state, because all the other "sub states" of sub systems that will be relative to me, will be consistent with this particle being in the spin up state.

    Well, my *belief* does correspond to the actual facts, namely that the particle has SEVERAL "classical shadows", and that I see one of it. But my observation corresponds to only one. In what way is a "shadow" not real ? Again, let us not forget what is our discussion:

    I claim that we only see PART of reality, but that part is there, all right, only there "is more than meets the eye", while you claim that what we see is simply delusion, and hence not thrustworthy. You have to make that claim in order for you to be able to make the argument that MWI is bogus, namely that it is some theoretical construct which takes as basic tenet that observations are bogus - while of course it has been set up to explain observations (as are all scientific theories).

    There's a big difference between our claims of course, because claiming that "what is observed is not all there is" is not an argument that would allow your reasoning to hold. Indeed, it is not unthinkable, that from the PARTIAL INFORMATION we get by our observations, we can nevertheless DEDUCE the existence of a more involved reality than what is ONLY observed - indeed, that is my point. There wouldn't be any logical inconsistency, any more than there would be a logical inconsistency in the deduction, from our sole observation of the tip of an iceberg, and some other observations (like Archimedes' law), that we can DEDUCE that there must be a much bigger part of the iceberg which we don't see. Extrapolating from the observed to the (more involved) real is not, by itself, logically inconsistent.

    However, in order for your argument to hold, you need to require that observations are BOGUS, are *erroneous*, and are not just "a part of reality" in MWI.

    Well, I claim that this is not the case, and that, what we observe are parts of reality, which are good enough for us to allow to derive the "invisible" part, after some thinking. Now, we can give different names to these parts, but "shadows" are quite convinient, no ? With one difference: all the shadows make up the total state description of the object, which is not the case of an optical shadow.

    It is better to switch to the Heisenberg picture for that, and then the field operators are indeed a field on spacetime, but an operator-valued field.
    You can then "go back" to the Schroedinger representation, and this will bring you extra structure to the wave function's evolution. In other words, spacetime has nothing to do with the *state* of a system, but with its *evolution*. Indeed, spacetime being a "bag of events", it is a structure of *interaction* and not of "the state" of nature.

    Yes, but so is the state in classical particle dynamics: it is a point in a giant phase space. However, the FLOW in that phase space is such, that one can find back the illusion of some spacetime, although the actual state is an element in a huge-dimensional phase space. Now, the difference between classical physics and quantum physics, is that one can distill more easily individual substates for subsystems (take slices in the phase space which correspond to subsystems), while the quantum state cannot be sliced up in a juxtaposition of individual states.
    But again, it is the specific structure of the evolution of that state which gives you the illusion of spacetime "as a place to be", while it is just a structure of interactions ; just as in classical phase space.

    Let's not forget that spacetime is the SET OF EVENTS. Not the "set of places to be". Events are interactions.

    Uh. I don't see the EM field as "consisting of local beables" ! Imagine you write the EM field in a Fourier decomposition: where are your local beables now ? The state of the EM field is a point in a very big phase space! However, its *interactions* are such that it gets a spacetime structure, and that's all that is needed for us to think that that we see a "space of 3 dimensions" around us, because what we see is a result of interactions.
    Now, indeed, one can, in classical field theory, if one wishes, slice up the phase space of the EM field into "subspaces" which correspond to "pieces of EM field" - these subspaces correspond to your local beables, I suppose.
    As I said, this cannot be done in quantum theory, because "the subsystem state" doesn't exist there.

    Indeed, in quantum theory, there are no local beables. What is local (that is, what gives a structure of spacetime) are interactions between subsystems.
    Last edited: Feb 22, 2007
  8. Feb 22, 2007 #7


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I'm reading this paper now. I'll continue, but I already have a serious problem with the following:

    That is what I call "sneaking out". We now divide "ontology" in two parts, namely "primitive ontology" and "algorithmic ontology". All the nasty stuff is put into "algorithmic ontology", which doesn't have to respect certain rules, while what we need, is just the "primitive ontology", which can then respect the rules.

    But if such a thing is allowed, then I don't see why we bother with any interpretational stuff at all: call "observations" the PO, and quantum theory the "algorithmic ontology" and we have the standard "shut up and calculate" interpretation (which is different from Copenhagen, btw). The whole point in providing a rigorous ontology for the formalism of quantum theory is to be able to give a precise ontological picture in which NO "algorithmic" part is foreseen, but in which all elements of the ontology are treated on the same level.

    If I'm allowed to play that game too, then I can say that the PO of MWI are the different states of my awareness (different body states which correspond to different awarenesses), and the whole wave function stuff is just an algorithm which generates probability distributions over my different states of awareness. It produces perfectly what I am aware of then.

    In a certain way, I'm not surprised that GRW and Bohmian mechanics turn out to be related. They have similar advantages and weaknesses: the cleavage in the "ontology of the world" ; the relativistic invariance of the results, but not of the internal workings (which are done away with by putting the non-relativistic part in the "algorithmic" section, so that we shouldn't care about it) etc...

    (reading on...)

    After reading further, I think that the difficulty is this:

    I think it is this basic assumption which is giving troubles with quantum theory. After all, matter doesn't need to exist in some kind of 3-dim space, as long as OBSERVATIONS are consistent with such an IMPRESSION. This is, again, like in classical physics, if we take it as ontologically correct that the world is 6N-dimensional, but that the interactions are such, that it looks like a 3-dim space with SEVERAL points in it. In that case, the impression of a 3-dim space is just a consequence of a special requirement on the Hamiltonian flow. There is moreover, no need for "extra" ontology in this hamiltonian vision on things, because our very measurement procedures would also be defined as a function of this requirement on the hamiltonian flow. The only thing we would need to specify (again), is to what points in phase space correspond what subjective impressions (in other words: what configuration of "brain-particles" - which is a part of the phase space corresponds to me having the visual impression of "a green frog at 2m from me, slightly to the right", for instance).

    Now, in classical physics, it doesn't give any formal PROBLEM to reformulate mechanics, instead of on a 6N dim space (which we could consider as ontologically real), as a "mechanics of several points" on a 3-dim space. So for people having such desire, they can do so. But in quantum mechanics, it DOES give some formal problem, and insisting, such as the authors of the paper, of wanting to have an ontology which consists PURELY of local beables in space, will necessarily clash at one point or another with the formalism of quantum theory. In fact, they even have to split their ontology into two parts, a PO one which can be "local-beablisised", and an "algorithmic ontology" which doesn't.

    So I think that the difficulty comes from a requirement which is not necessary, namely, the requirement that "ontology" (or at least, the speakable part of ontology) must live, as local beables, on spacetime. I think the correct way of seeing things from a quantum formalism viewpoint, is not to put any requirement on what kind of mathematical object "ontology" must be, and to see spacetime as a source of symmetries of the ontology (the ENTIRE ontology).

    EDIT 2:

    I finally reached the conclusions of the paper. I re-iterate my comments above: I think it is a "cheap trick" to separate the "ontology of the world" into two different ontologies. Or, the ontology of the world is entirely, always and with no exception, represented by local beables on spacetime, or this isn't the case. Now, the first point in the conclusion seems to be that this ought to be the case. Fine. But next, they talk about the wavefunction in hilbert space. I object to that. In as much as this wavefunction has some ontological character, according to the first point, it must live on spacetime (which it doesn't). I think that from that point on, one can stop. There's no point in requiring PART of the ontology to live on spacetime, and changing names into "primitive ontology" and "algorithmic ontology" is a verbal game. If some part of the ontology doesn't need to live on spacetime, then there's no point in requiring that ANYTHING lives on spacetime. Spacetime can then become a requirement of symmetry in that other place, where the ontology lives (say, Hilbert space) ; and from that symmetry can follow our *impression* of spacetime.

    But once we're there, there's no point in fiddling with the standard quantum formalism, which does exactly that.

    If we're allowed to algorithmic "deux ex machina", then quantum theory is good enough an algorithm to calculate outcomes in the lab, which we can then take as "primitive ontology".

    In other words, all this is a lot of work for no formal results. The aim of interpretative work is to help understand the formalism, not to fiddle with it. We can deduce no new fundamental principle from all this, which might guide us in, say, having an idea about how to deal with the borderline between quantum theory and GR.
    Last edited: Feb 22, 2007
  9. Feb 22, 2007 #8


    User Avatar

    I understand all this just fine. You're still just missing the point: if the wave function is all there is, and if the wave function is something like

    a|+>|I believe +> + b|->|I believe ->

    and I *think* the particle has state |+> and I *think* I believe it has state +, i.e., I *think* the combined state is

    |+>|I believe +>

    then I am just wrong. I am wrong about the particle, wrong about myself, and wrong about the combination.

    MWI forces me to accept that I am almost always wrong about everything in this way. Included in the "facts" I am wrong about, are all of those "facts" which made me believe in the quantum algorithm in the first place. So MWI asks me to accept that the best way to interpret the physical meaning of that algorithm, is to accept that I was deluded into accepting it in the first place... and yet somehow still I should still accept it.

    Paraphrasing Monty Python, you may call this logic, but I call it crap.

    It has nothing to do with the separability of states. It has to do with whether there is anything actually posited by the theory which corresponds to my beliefs -- i.e., whether the theory leaves any room for beliefs (including beliefs about how certain experiments came out in the 1920's) to be true.

    You introduced the term "classical shadow" so you should define it, not me. I took it as clear, though, that you weren't meaning to posit these "classical shadows" as things that really exist *independent* of the wave function. They are just a certain "narrow subjective perspective" on what is already in the wave function, right? But then you haven't actually addressed my point. If all that really exists is the wf, then what really exists doesn't correspond to my beliefs about what exists. Or if you are positing some new beables here (the "classical shadows") you better flesh it out a bit and give some dynamical laws for how those things behave and interact.

    You give up the whole game when you say "what is observed is not all there is". The fact is, "what is observed" directly contradicts the posited nature of "all there is." That's the point I keep raising above. What is observed is a world with entities with certain definite properties. Yet according to MWI the real world is nothing like that. Unless, of course, you think that the plus signs in superpositions mean "and/or" or some such. But only naive people who beleive in the "ignorance interpretation of superposition" accept that -- not people who accept the completeness doctrine.

    Your whole case here hangs on interpreting one term in a superposition as a "part" of the whole. Right? The problem is, that just isn't what the word "part" means (which is why your iceberg analogies don't work). It might work, granted, if MWI actually meant "many worlds". Then the one term would describe a certain physical reality that is realized in one universe, with other terms also describing realities in other universes. Then it would be true that the one term is a part, and hence that *belief* that the state of things is as described by that one term is a "partial truth" (like your iceberg analogy) rather than just false.

    But I think you know the problems with interpreting terms in a superposition as living in different worlds, literally. So you're stuck.

    You want the beables of the theory to be mathematical operators? I don't even know what that means.

    So... do you accept that there is matter (e.g., tables, chairs, planets) in 3-space?

    That's probably what this comes down to. I say: if you don't accept that, you are just crazy and not a serious (or, at least, psychologically healthy) physicist. And if you do accept that, you have a problem because such stuff is nowhere to be found in MWI.

    I'm sorry, but this is just silly. Yes, it's possible to formulate classical particle dynamics in phase space. But each particle actually exists in 3-space, yes? So there is no problem here like the one faced by MWI.

    Imagine you take the fourier transform of the mass density of a tree. Oh my god! The tree is gone! There are just an infinite number of abstract waves each spreading across the entire universe!

    But we're not talking about that phase space. The EM field lives in 3-space. As do trees. Period. If you don't accept that, we can't possibly have a serious discussion.

    If what you mean by "quantum theory" is orthodox or Copenhagen, you're wrong. There are local beables: classical objects (including especially apparatuses). Those are just *posited* by the theory as the local beables. Bohr then said that's *all* that really exists physically ("there is no quantum world...").

    This theory of course suffers from the measurement problem, essentially because it never says what is "classical" and what isn't, i.e., it's "unprofessionally vague and ambiguous." But at least it has the virtue of accepting that there are real physical things moving around in 3-space!
  10. Feb 22, 2007 #9


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Indeed ! That's why I don't believe that!
    It is not because I SEE a car over there, that I believe that there is car over there ! I only believe that there must be a term

    |I see a car over there> |car is over there>

    in a superposition.

    Absolutely. If you keep your beliefs the way you state them, then you are indeed wrong. But that is your problem, not mine.

    Again, only if you believe that what you see, is all there is. If you believe that what you see, only corresponds to a term amongst others (which you won't see), then there's no problem, right ?

    Well, often with initially wrong beliefs, one arrived at the right answers ! It is not because many astronomical observations were made based upon totally erroneous beliefs about astrology, that all that work (including Kepler's) was wrong.

    Again, that depends on your beliefs. If you believe that the entire reality is what you observe, then indeed, there's a problem.

    No, it are those parts of the term of the wavefunction, corresponding to other subsystems, which is entangled with that part of the body wavefunction from which emerges *my* subjective experience (and not the potential subjective experience of my copy, for instance).

    They are "pieces of wavefunction", selected by two things:
    1) the separation of "my body" and "the rest of the universe"
    (which gives us a more or less unique decomposition of the overall wavefunction in different terms, which correspond to different "worlds" according to me)
    2) my subjective experience emerges from ONE of these terms (randomly selected by the Born rule) - eventual other subjective experiences emerge from copies of myself, but I don't experience that.

    This completely selects the unique term in the wavefunction, which consists of two factors: a body state, and a "rest of the universe" state. What I observe is of course directly, my body state, and indirectly, the "rest of the universe" state which is compatible with this body state.

    Now, because of the nature of the interactions between "my body" and "the rest of the universe", by decoherence, each different term contains state vectors of subsystems which look very much like classical systems. I observe ONE such classically-looking term, which gives me the impression that things are localised in space, and have more or less well-defined momentum.

    Well, these properties seem to be well defined, but in MWI, that's explained by the decoherence (that is, the quick split into different entangled terms) of states which are NOT so, in many cases (but not in all! That's how we found out about quantum theory in the first place!). This comes about because of the specific structure of INTERACTIONS, which needs to respect this famous spacetime structure. In other words, our IMPRESSION of living in a spacetime, only comes about, because interactions between different subsystems have to be organized in such a way that they are "labeled by spacetime" and can interact only "locally".

    Indeed, but it also explains why we get the IMPRESSION that things are so.

    Yes, and in what way is this not so ? By "part", I mean: "mathematical part". There's a quite definite way in which the overall wavefunction is sliced up into terms (namely, the "my body/rest of the universe" decomposition). If we now accept that my subjective experience emerges from one of those "my body" components, then the component corresponding to "rest of the universe" in that term is well-defined. I call that "my world", but there's no reason to assume that the other terms in the overall wavefunction are qualitatively different, so I can call them "their worlds".

    No, I don't know those problems... each of these terms (mathematical parts) of the wavefunction can live as a wavefunction all by itself (because of the linearity of the time evolution operator).

    In quantum theory ? No. Of course not. However, I can accept that I make observations which give me the impression that there are chairs, planets and so on in 3-space. And I also understand that I can set up a classical model in which there ARE tables and chairs in 3-space, which can, most of the time, also predict my impressions correctly. In that classical theory (in its Newtonian version), of course, I accept that there are tables and chairs and planets in 3-space. But not in quantum theory.

    Remember that what is to be considered "real" is depending on the theory in which one works. With each physical theory goes a different ontology (= hypothesis about what's real).

    Half. In MWI, there are statevectors (the components of that part of the wavefunction which is supposed to describe the chair and so on which are in "my" term of the wavefunction) which come very close to a mathematical model which corresponds, what will concern my impressions, to a classical description of the chair and so on. So, given that there is a mathematical piece of the ontology which I could bring into correspondence with "chair in 3-dim space", in as far as that piece is well-defined, I could give some form of "existence" to that "chair in 3-dim space". A bit like you can give some form of existence to a character in a film, or even better, to icons on your computer screen.
    And in as far as I can make that approximation, I can then just as well work with the classical model, in which I can say that there ARE chairs and planets and so on in 3-space, because there, there is no problem in considering them "genuinly real".

    In the Newtonian view on classical mechanics, yes. In the Hamiltonian view, no. I have no problem with considering that the world is actually 6N dimensional (is a single point in a 6N dimensional space - hey, maybe there are even OTHER points, corresponding to other worlds, in that same phase space), if it is organized in such a way that I get the impression that there are different particles in a single 3-dim space. I even consider that a good exercise to get a feeling for quantum theory.

    Exactly. However, the way these waves interact with my "body waves" is such that I get the subjective impression of seeing a tree. I have no difficulty in principle with such an approach. However, it would not be a representation of reality in which I could easily deduce what brain states corresponds to my "subjective impressions". The "states corresponding to my subjective impressions" would be rather defined in a complicated way.

    I do not require this as an absolute prerequisite, indeed. At a certain point in time, it was not a bad idea (because inspired by our intuition). But I think we are now far beyond that. The only thing we still need, is that somehow we get a subjective IMPRESSION of 3-space. And I think that this comes about because of the specific structure of interactions, which needs to be a representation of spacetime.

    I think that that is the reason why it has so many "paradoxes".
    You cannot keep the superposition principle, and require things to be localised in space. However, you CAN keep the superposition principle, and require interactions to be "indexed on spacetime".
  11. Feb 22, 2007 #10


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I moved our discussion to a new thread, because it started to hijack the "OK Corral..." thread.
  12. Feb 23, 2007 #11
    I was wondering if I could interupt this discussion with a question about MWI?

    According to MWI everything is quantum and has as associated state space and that "measurements" are just entangling unitary operations between systems (one of which will be the system of interest and the other will be an observer).

    I can see how this works in the case of measuring an observable of a system but how does it work when measuring a POVM?
  13. Feb 23, 2007 #12


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I'm only acquainted in a rudimentary way with POVMs. I'm not sure that they are physically relevant in this discussion, in that in MWI, we're supposed to use the pure state of the entire universe. As such, the only measurements which are physically executable should be of the projective kind, no ?
    But, as I said, I probably don't know enough of this to give any meaningful answer...
  14. Feb 23, 2007 #13


    User Avatar

    I don't have time to get into this at great length (and it's probably pointless anyway since we've had this same discussion a thousand times), but a few brief comments:

    But neither of those factors actually represents the state of the relevant degrees of freedom. That's the whole problem. The factor representing what you call "a body state" in the one term (evidently the one you subjectively experience) refers to the same physical degrees of freedom (the same particles or whatever) as some factors in other terms. *That* is why it is invalid to say your subjective experience actually corresponds to the real state of those (or any) degrees of freedom, in MWI. Those particles (or whatever) simply are not in the state you experience them to be in.

    In my opinion, you essentially concede the whole argument to me when you say:

    That is: really, according to MWI (though not quantum theory, as you suggest, since that *presupposes* the real existence of the classical world) there are not actually any such things as tables and chairs and planets in 3-space. And yet, every possible bit of *empirical evidence* that could in principle be claimed as supporting MWI, is of this form: a table or chair (or pointer) was at some particular place at some time. Thus my criticism of MWI: it asks us to accept that the best way to interpret the fundamental physical meaning of those pointers and chairs having done what we saw them do, is to deny that those things even exist. And that, frankly, is just dumb.

    Yes, sure. That was the point I had in mind in refering you to the paper I cited a few posts back. I actually agree with your criticism that the division between "primary" and "secondary" ontology is a bit dubious. Either something is posited by the theory to really exist physically, or not. There is no half-way house between reality and unreality. So we agree there. But my point in pointing out the paper was to help you see that, if a theory has none of what the authors call "PO" -- i.e., if it doesn't posit anything that can be thought of as familiar *matter* moving in 3-space -- the theory is literally about nothing, or at least nothing physical. It may still of course be about "subjective experience" and may concoct some crazy story about why/how we are *deluded* into believing in matter moving in 3-space, but that isn't physics, it's some kind of bizarre philosophical/psychological craziness.
  15. Feb 23, 2007 #14


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    This is because you have a preconceived notion of what are "degrees of freedom", which are in blunt contradiction with the superposition principle (and the same happens to your notion of 3-dim space). A degree of freedom in quantum theory is something which can take on several values at once, but of which only one can be observed. This is why we have the impression that it has a precise value. If you INSIST on it having only one precise value in order for it to be real, you have denied the very founding principle of quantum theory. But in its very definition, it also includes why we have the IMPRESSION that it can take on precise values!

    A degree of freedom in QM corresponds by definition to a set of orthogonal basis vectors in a (factor) hilbert space, EACH INDIVIDUAL VECTOR corresponding to a POSSIBLE OBSERVATION. Now, saying that the state of a (sub)system will be a vector in this space, means exactly that it will be in a state that corresponds to MANY DIFFERENT POSSIBLE OBSERVATIONS at the same time. So, by its very construction, it is impossible to equate that what is observed, to the state itself. The whole theory is build upon this concept, that the actual state is a mathematical combination of the different possible "states of observation".

    Now, observations are ultimately "subjective experiences". So what the superposition principle essentially says is that, whatever are the "degrees of freedom" on which subjective experiences emerge, they can only emerge on a specific component (in Hilbert space). They cannot emerge on the entire state vector, because that state vector has to be, by postulate, a superposition of "states of subjective experiences". Hence, *by postulate* it is build in from the start that, if quantum theory is going to have a psycho-physical link, it must be "component-wise".

    So, all we call "degrees of freedom", "classical objects" etc... must hence be seen component-wise. A "single quantum degree of freedom" corresponds as such to MANY "classical states" at once (by hypothesis, in the basic postulate of quantum theory). Insisting that the FULL STATE ought to be the observed state is hence killing the initial postulate from the start, which REQUIRED the full state to be a combination of potential alternative and mutually exclusive observable states. From the moment that you could observe "the full state", you made an error in the construction of the hilbert space, which was supposed to contain *combinations of mutually exclusive observational states*.

    Again, in quantum theory is constructed around the principle that mutually exclusive "observational states" are the "real state". So we START by assuming that the "real state" of the universe is unobservable! But its components are, each individually, consistent observational states. This is the very starting point of quantum theory.
    We shouldn't be surprised, then, that what we observe, is one of those components ; moreover, if these components can lead, most of the time, individual lives and have an almost 1-1 correspondence with classical theories (a different classical state for each of the components), then to me that's just fine: it explains why we *observe* essentially a classical world.

    I agree with you with most of what you write, but we clearly disagree upon that last statement. The whole purpose of science is to explain our subjective experiences. If it can do that, based upon a limited number of principles, then that's perfect for me. There is no reason why there should by any 1-1 link between a "reality" and my observations. This could be, but this could just as well not be. Classical physics makes the hypothesis that this is the case, and quantum theory STARTS by assuming that this is NOT so (it is the content of the superposition principle!)

    All, beyond our subjective experiences, is nothing else but hypothesis in any case. We sometimes have strong intuitive convictions that our observations "are real", and in as much as this is a helpful hypothesis, you can make that hypothesis. As I'm trying to point out, in MWI, your observations are "as real as is possible within the framework" (because that very framework takes as basic postulate exactly that "reality" is a combination of mutually exclusive observations).
  16. Feb 23, 2007 #15


    User Avatar

    This will be my last comment on this thread. I really don't have time to continue with it.

    Vanesch, you basically admit that the whole point of MWI is simply to give an account of "subjective experience." Everything else -- the existence of chairs, tables, planets, atoms, matter in 3-space, everything -- can be dispensed with so long as the subjective experiences come out correctly. Indeed, you sloganize that this is the proper and only function of science. Well, perhaps Ptolemy and Ernst Mach would agree with you there, but many scientists wouldn't. Many scientists actually accept (yes, I know, it's shocking) that the external 3-D world full of material entities moving and interacting is *real*, and that there is a *difference* between theories like Ptolemy's and theories like Copernicus's, even though they (arguably) make all the same predictions for "subjective experience".

    You mentioned the "psycho-physical link". I think that's the key point here. Interestingly, Bell sometimes uses the phrase "psycho-physical parallelism" to mean the kind of view I indicated in the previous paragraph -- that what we experience actually exists, that our "subjective experience" of chairs and tables actually corresponds to really-existing chairs and tables. To me (and to Bell) that is the kind of thing that we learn when we're about 1 year old, and that literally everything else we do subsequently (I mean "do" here in cognitive terms) is based on. For example, note the absurd circularity in the following: I claim to have gone into a laboratory and done some experiments and concluded, based on the outcomes of those experiments, that there are no such things as tables, chairs, or pointers. But that is just absurd. That the positions of certain pointers in the lab were observed to be such-and-such can *never* be taken as evidence that those pointers don't exist. It's a logical contradiction. And what I'm pointing out here is that MWI is doing exactly this -- undercutting itself by denying the reality of the very sorts of things which manifest all the purported evidence for MWI.

    As you know, I grant that in a certain sense, MWI *can* successfully account for our (or really just *my*) subjective experience. It's just that that "experience" is then nothing like what I always thought "experience" was supposed to be. My experience is not just some "inner theater", but an actual *perception*, an actual *recognition* of really-existing external things. This is what all 2 year olds know, but what MWI denies. And it means that MWI is not really a theory about physics in the normal sense (a theory about the hidden/underlying workings of little bits of matter or fields or particles or whatever, attempting to account for the directly observed behavior of things like tables and chairs and pointers). It is rather closer to a theory about brains in vats, except that even that grants too much psycho-physical parallelism to it, because according to MWI there really aren't any brains or vats (at least not in the way we've seen them and conceptualized them). So MWI is more like Leibniz's theory of monads or something -- there are these (or more likely just this one) consciousness(es) which have a certain stream of "subjective experience" fed into them, not by any familiar material things like evil scientists living outside the vat, but by some ineffable "stuff" living in some totally unfamiliar "space". And then the crucial dynamics of the theory (the part where you, vanesch, differ somewhat from the other, stupider, MWI advocates in the world, the part having to do with a "consciousness token" flowing downstream through the branches in accordance with some stochastic Born rule type law) has really nothing whatsoever to do with anything remotely physical. What we have always meant by "physical objects" are literally nowhere to be found in MWI, except inside the mind of the poor soul having all these "subjective experiences." And as I've stressed (and as you've agreed, though only by twisting the words a bit to hide the strangeness) all those experiences are literal delusions, since (eg) when he thinks there's a solid table in front of him which is brown and has 4 legs and exists in 3-space, none of those things are true. There is no such thing with no such properties in no such space.

    As Bell once said in a different context, this (the boundary between systematically delusional subjective experience and some totally unfamiliar pseudo-physical "world") is a very uncomfortable place to try to do physics. Normal, rational, scientific people should recognize this and admit that MWI is simply not on the right track (in the same way and for the same reasons that they recognize that solipsism is not a good scientific way of making sense of some puzzling observations like retrograde motion or dark matter).

    Anyone else who is lurking and agrees with my thinking here is invited to continue the good fight against MWI. :zzz:
  17. Feb 23, 2007 #16


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    That's what I tell myself each time I engage in such kind of discussion too :biggrin:

    Yes. It is the ultimate goal of all forms of intellectual (and even practical) activity, no ? But it takes some philosophy to see this, that's true. After all, why do we conceptualize such a thing as a chair ? It is because we've had repeated combined sensations, both directly and indirectly, which are all rather well explained by making the hypothesis that there is such a thing as a chair. In other words, the only reason to assign some form of ontology to a concept such as a chair, is that it helps us explain and organize our sensations. "Reality" is a hypothesis which helps us organize our sensations.

    Yup. The important thing is: *can* be dispensed with. It is not a philosophical necessity. Now, I agree as much as you do (just to prove my mental sanity which you are now doubting seriously, I'm sure :smile: ) that it would be nice if we could keep it. But, when interrogating myself, I don't see why that should be an absolute necessity.

    Look at the very definition of a scientific theory: AGREEMENT between PREDICTION (of observation!) and ACTUAL OBSERVATION. If you realize that ultimately, an observation is nothing else but a subjective experience (whether you look at the display, or read the computer printout, or ...), then you see that the very definition of science is to find a "hypothesis of reality" (a model) which generates the subjective experiences in agreement with what we actually experience. The only other requirement we put on that model is that it is based upon as few principles as possible (Occam).

    Uh ! Then he has mis-read von Neumann, who was the first to realize that this plays a role in the interpretation of quantum theory, in his very book on "mathematical foundations of quantum mechanics". This is btw why I find von Neumann much deeper than Bohr on the subject.
    It doesn't mean at all what you write: it means that all subjective experiences should have a "physical carrier" of some kind. Maybe I misread that, but to me, it only means that it is the physics, through some or other mechanism, which "generates" the subjective experience. It means that there must be some aspect of physics which "contains" our experiences.

    Indeed, and it needs some work to get rid of it. In the same way as, say, relativity of simultaneity. That's also something you learn when you are 2 years old.

    Yes. However, you forget to add: I did some experiments and concluded, that there are not really such things as tables and chairs (in the way they are usually conceived), but that there is something different, which, when observed, will give me the impression of there being tables and chairs, which will, most of the time, also behave as tables and chairs the way I learned when I was 2 years old.

    You attach too much importance to "exist". "Exist" is a hypothetical concept which only serves to help us organize our sensations.

    Again, reformulated: the OBSERVED POSITIONS of things which look to me like pointers, gave me evidence that all observations cannot correspond to uniquely existing pointers (but to a more complicated concept, which, most of the time gives me the impression that there are pointers pointing somewhere).

    Let me tell you why this is not logically impossible.
    Consider (a bit like the cavemen in Plato - which is exactly the kind of situation MWI presents) that all your life you've been attached to some kind of device which keeps your head pointed at a movie screen, and that all you've ever seen and heard were projections of movies on that screen. Clearly, you will have a "picture of reality" which corresponds to an imaginary world - which corresponds in many respects to a kind of world which looks like our classical world (you have seen people, cars, airplanes etc...). Now, one day, they show you a movie in which they film a guy who has been looking at a movie screen for all his life, then they show you, in the movie, that there is a movie projector, etc...
    And suddenly, you realize that this might very well be your situation ! So, from looking at stuff on a screen (which you previously thought, was "reality"), you suddenly came to realize that you're probably just looking at a screen, and that this is not "reality". You now also understand most of what "happened in your "observation of reality". But now you "know" that real reality, is a movie theater, a projector, and a poor guy who's forced to look just at the screen.

    Right ! But that's always a strange experience. Nevertheless, even an old guy like Plato thought of that.

    Yup. As I said, it takes more than a 2-year old to grasp that concept.

    That's exactly it, right. See, it is an old idea. But you have also a kind of problem in classical physics, so in any case, you need some kind of "conversion gate" between your subjective world and the "physical world" (at least, from a dualist perspective) The only difference between classical and quantum physics, is that the conversion gate is slightly less "transparant", and that hence the physical world is less familiar (in subjective terms).

    Yes and no. Indeed, a crucial part is given by what you say, but there is also a part which is "objectively physical". And that is where I disagree with the following:

    Well, there IS a mathematical object, somewhere deep into the overall wavefunction, which DOES correspond to that table. Or better: which corresponds to the relationship between you and the table. That's why I say that what you see is PART of reality (but not "part" in "piece of volume" of course: part in the sense of part of the mathematical structure).

    Sorry, there IS something within the ontology of the world that corresponds to it. That's what I'm repeating now already a few times. In as far as you have a certain liberty in calling what is "real" (after all, it is just an organizing principle of sensations), you could call it real too.

    I agree. It would have been simpler to live in a classical world. It would even have been nicer to live in a Newtonian world.

    The problem with saying a priori what are "good" and what are "bad" principles, is that one might make an expensive mistake by adhering to a principle which is not of this world, no matter how intuitively attractive it may sound. As you see, I don't adhere to MWI for MWI's sake. I adhere to it as a view on quantum theory, because the theory generates good explanations of observations, and respects in the most fundamental way the fundamental postulate of quantum theory (which is the superposition principle). It wouldn't come to my mind to go and fiddle with a FORMALISM (which works well) which is based upon some principles (no matter how crazy at first sight) just to satisfy my intuitive desires of what ontological principles a theory should satisfy. The formalism IS the ontology, for me. So I twist and turn my views on ontology in order to make them fit the mathematical formalism, and not the other way around. If that exercise pushes me in the corners of unintuitiveness, well, then so be it. I am of the opinion that a good formalism has more to teach us about nature than our intuition, so whatever the formalism tells us, must be a better guess at the ontology of the world than whatever we can intuitively dream of.

    To conclude, I can very well accept that some people don't like MWI. As I've pointed out several times, I only adhere to it because I think it is the "most natural" interpretation of the *quantum formalism*. I find it highly unintuitive too. However, it has the advantage that it explains quite naturally (because of its close ties to the formalism of course) certain "paradoxes" which seem to occur in quantum theory, when one tries to fit it upon a "classical world". As such, this stops me from lying awake at night about the EPR paradox or something of the kind. If it gives you more conceptual problems than it solves, then MWI is not for you. But for me, it works fine.

    However, I do not accept your critique of the *logical contradiction* you try to force upon MWI, by saying that "MWI is based upon observations of which it tells you are bogus". I repeat my objection: we make observations. We may think that they correspond to "something real" or not, that's a separate matter. But in the end we need a story which tells us what our observations should look like. If that story tells us correctly what our observations will look like, EVEN IF FOR THAT IT MAKES THE HYPOTHESIS that reality is in fact quite different, then that is a good story. So it is not true that MWI tells us that the observations we made, which led us to set up MWI, are bogus. It only tells us that reality is different, but from that different reality, it can correctly derive the observations.

    This is a bit like "parameter fitting". Imagine that I have a list of pairs of numbers. Now imagine that I come up with a calculational model which can generate those pairs of numbers, but that for that, I have to feed it with parameters which don't look at all like the pairs of numbers. This is nevertheless a good model. It is not because the parameters I "fit" do not look like the pairs of numbers "I observe" that this is a logical contradiction. This is the same in MWI: it is not because the reality that is postulated doesn't look at all like the observations that are performed, that it is a bad model to explain those observations.
    Last edited: Feb 23, 2007
  18. Feb 23, 2007 #17

    there is nothing good about fighting against the universe simply because you have aesthetic problems with it :wink:
  19. Feb 23, 2007 #18

    Doc Al

    User Avatar

    Staff: Mentor

    Perhaps there's a universe where you are the one trying to convince everyone of the silly extravagance of the MWI. :rolleyes:
  20. Feb 23, 2007 #19


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member


    Is this then a collision of worlds, where SetAI and I are copies ? :biggrin:
  21. Feb 23, 2007 #20

    Doc Al

    User Avatar

    Staff: Mentor

    I'll know the original vanesch by the twinkle in your eye! :wink: (But my copy might not! :eek: )
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook