selfAdjoint said:
For example in our conversation about the RQM interpretation of an Aspect type experiment, on the Beyond the Standard Model... forum, you said that the near observer sees a superposition and the far observer doesn't, and these states have to have enough reality for the observers to compare notes when they get back together.
Uh, that's what (I understood) RQM says, not me.
In MWI, this is understood by the fact that *by definition* in MWI an "observer" being a STATE of an observer body, it always sees only one state of itself, but it is entangled with other states: whether these other states are now "observer states" or superpositions thereof makes one say that "the other observer is still in a superposition with respect to this one".
In all this, of course, it depends on what one calls "an observer state", which is, I guess, the "preferred basis problem" you refer to. One way to say what "an observer state" is, is one with a definite entry in a notebook !
As you cannot compare notes of notebooks which are not in "a definite state of notebook", you automatically, when talking about "comparing notes" ASSUME that these states are somehow special.
Now, as I said somewhere else, in how much, these states have some naturalness to them in that they are robust against environmental interaction, and in how much we have to postulate them, is not yet entirely clear for me: if I believe people like Zeh, it seems that environmental decoherence gives them a natural status. But EVEN if I have to postulate them, I do not do anything worse than what's done in RQM, when it is ASSUMED from the start that there are "states of definite knowledge" (namely the vectors of yes/no answers) - which correspond to "definite entry in a notebook" states I referred to above. So, if you don't mind RQM to introduce the *intuitive* concept of "states of knowledge" with a series of definite answers y/n, then you shouldn't object in MWI introducing these states by postulate too. However, if moreover they have some naturalness to them (as indicated by environmental decoherence), then at least we have, in MWI, a *natural* explanation, for an irreducibly intuitive concept in RQM.
So RQM doesn't address any more the "preferred basis" problem than MWI does - but at least MWI has a possible mechanism to it to give some natural explanation for the preferred basis (of definite notebook states).
I say that the state vector is not observable and the near observer does not observe that superposition. She infers it on the basis of
1) The experimental setup, which is designed to produce a superposition.
2) Her later comparisons of what she DID observe with what the far observer observed.
Yes, no problem with that. As we infer the existence of the Earth by looking around us, by seeing pictures of it from satellites etc...
But the basis problem says that there is no definite set of numbers describing the near superposition. Rather there is an infinite set of equivalence classes of sets of numbers anyone of which provides a complete description from which the probablility densities of the whole experiment, near and far, can be computed, but no one of which is to be preferred over any other.
Well, again, I refer to what I said above. Or we have to postulate such a basis (as does implicitly RQM, by giving a special status to these vectors of yes/no answers) where we say that observer states are those with "definite memory states" - or they have some naturalness to them, in that they emerge as the robust states throughout interaction with environment.
Near Alice did not see them and cannot remember them. The only thing that has to stay real to do the experiment is the hardware and the observers' memories of what they actually saw. And I think RQM can do that.
Yes, but by intuitively introducing "observer memory states" (namely vectors of yes/no answers). Let me introduce them as preferred states in MWI too ! We DEFINE observer states as those with definite memory content. But what's (maybe) nice, is that these states ALSO appear as the robust states that "survive" in environmental interaction. So that's maybe the *reason* why we take them as "memory states". If they weren't robust, they wouldn't *keep* their memory, and it would be useless to consider them as "observer memory states", because the very next moment, they'd be gone. So if any "observer experience" which has to stretch over an (even very short) amount of time requires a "robust memory state", then it would be *natural* to have "observer experience states" which relate to states which are robust over (an even very short) amount of time. If we believe the decoherence people, this is exactly what happens: some states (which ressemble classical states) are "robust" against environmental interaction, can hence serve as "memory states" and can hence serve as states associated with observers.
And coming back to your paragraph above, I think when you make the innocent proposal to investigate the measurement process you are in fact imputing stable reality to something that does not enjoy mathematical uniqueness.
No, I hope not. I hope that we can assign "observer status" to certain states by a natural phenomenon (namely stability over time against environmental interaction) ; these states are then "memory states". And we can then investigate how they arise in a specific process.
As I said, I don't know how much "needs to be put in by hand" and how much is "generated naturally", but at least there is some hope to generate it naturally, while in RQM, it is, from the start, put in by hand. So you can't possibly object to that!
Einstein faced an analogous problem in GR when he conceived the hole argument. But he had the advantage of tensors whose equations remain stable when you vary the coordinate system. Unfortunately, spinors don't behave that way.
I don't know if they don't behave that way, when environmental interactions are taken into account. If we believe the decoherence guys, things surely seem to turn out that way.
