>>he usual way of looking at quantum mechanics is to associate a wavefunction
>>with a quantum entity. The wavefunction evolves according to Schroedinger's
>>equation when no one is looking. But as soon as someone tries to measure a
>>property of that entity, the wavefunction collapses (jumps) to another
>>wavefunction, an eigenfunction of the equation.

>That is not quite correct in EEQT. The jump happens not "as soon as
>someone tries to measure", but after osme time (time of jump is a random
>variable), and as the result of interaction between quantum object and
>a monitoring device.

Well, to misquote Galileo, "and yet it jumps". EEQT, whatever it
accomplishes, doesn't get rid of that. If one has a wavefunction extending
the length of the universe, then from the universe's point of view it
doesn't make a whole lot of diffence whether it is a mind or an apparatus
that provokes a universe-wide collapse, nor does it matter much whether it
is slow or fast.

>>Both parts of this picture are endlessly bothersome. 

>Why?

See above. Plus all the usual reasons. The wavefunction is supposed to be
able to completely describe reality, but it doesn't because there are no
events with the continuous (Schroedinger) dynamics. So something is needed
to collapse the wavefunction (slow or fast) and this something is just a
piece of the classical world that QM was meant to explain. So the
wavefunction is a very unclassical entity in what is otherwise a desperate
(understandably so) attempt to hang on to classical imagery.

>> The continuous
>>evolution of the wavefunction when unobserved describes a world without
>>events, i.e. not the one we live in. But the break in continuity that takes
>>place during observation, and that is triggered by observation, is an even
>>greater embarrassment. 

>Observation is just another "coupling". New coupling implies new
>dynamics.

>> The wavefunction can have arbitrarily great spacial
>>and temporal extension, and jumps instantaneously at observation time.

>Why does it bother you?

Because of my mammalian nervous (perceptual) system! :) 

>>That there is instantaneous correlation (but not influence) at a distance is
>>reality, and must be incorporated in any interpretation of QM. 

>It is incorporated in a mathematical model.

Yes, but we are still left with the task of finding a context which
minimizes weirdness. I can't believe you are as blase about this stuff as
you appear to be. You must be a practicing physicist or something like 
that. :)

>> But observer
>>created reality is not fact. It results from an error in my opinion. 

>In this model observer plays no role. But a monitoring device does. 

Let's say I concede that your approach accomplishes that. The weirdness
remains that an apparatus constructed by an earth ego can, slowly by our
standards but instantaneoulsy by universe standards) collapse a universe wide
wavefunction.

[CUT]

>>The wavefunction is indeed necessary, and the calculation of probabilities
>>of outcomes would be impossible without it. What isn't necessary however is
>>attributing the wavefunction to quantum entities. 

>Not everybody does it. We do not attribute coordinates to space time
>points, yet coordinates are useful in calculations. We attribute
>"positions". We attribute state vectors to quantum entities, and 
>why is it bothering you?

Ok, we attribute positions, and one knows how to use coordinates describing
those positions to get distances and separations. With the wavefunction
we know how to calculate probabilities. Complete the analogy: coordinates to
position allowing calculation of distance, wavefunction to ???? allowing us
to calculate probabilities. And beyond that, position is but one attribute
of a classical particle; the wavefunction is (supposed to be) a complete
description of a quantum entity. So in that sense the wavefunction IS (or is
supposed to be) the quantum entity, not just one property of it. 

>> Well, you ask, if not to
>>quantum entities, then what? To the experiment! To the experiment as a
>>whole, not to be divided up into setup and execution (registration).

>No experiment is also an experiment. And experimenter can freely decide
>whether or when to swich on interaction. Therefore wave function is
>attributed to the quantum object, but its dynamics depends on the actual
>situation, on actual experimental setup that can change in time.

>Of course the term "actual" becomes rather suspicious in a relativistic
>situation, but that is another subject.

I don't fully understand you here. But the wavefunction itself, never mind
the dynamics, is determined ENTIRELY by the experimental setup.

>>The wavefunction applies to the experiment, the whole experiment and
>>nothing but the experiment. 

>Not really. It describes the situation between experiments as well.

THERE IS NO BETWEEN EXPERIMENTS. If two quantum experiments are combined in
a quantum way, there is then ONLY ONE experiment. See below. That's one of
my main points.

>> It allows us to calculate the outcome
>>probabilities. Therefore there is NO COLLAPSE. 

>That depends on the model. How can you model any experiment
>without collapse? What if your experiment gives you a series of data:
>a click today, a click to morrow, and another click next week.
>How will you model this series of clicks without collapse? What
>is your definition of a "click?" What is your definition of experiment?

I have a rather narrow definition of a quantum experiment (what I like to
call a weakly replicable experiment {WRE}). And that is simply that it have
a set of possible outcomes and that the statistics of the outcome set depend
on experimental parameters in a HS vector way rather than a set theoretic
way. That's a pure WRE. So in order for the experiment you describe to be a
WRE, one has to consider the experiment as one (or more exactly as an
instance of one experiment - which would be repeated to get statistics).
Therefore a series of clicks (and intervals) is one outcome. The statistics
of these outcomes can theoretically be calculated using the wavefunction
that describes the experiment.

Of course, there are imperfectly executed WREs, variants, and therefore
"mixtures". I.e. what might be intended as a series of instances of a pure
WRE might turn out to series classically selected from a set of variant
WREs, and those outcomes are described by the apparatus (density matrix or
whatever) used to describe mixtures.

So, no collapse. The wavefunction isn't doing anything. It's simply (a
description of) a property of the experiment. There are other problems, I
admit. You say "model", I say "picture". No picture, true. Unless you accept
the wavefunction as a model with the problems I complained about before. But
I think that the price paid for avoiding collapse via this approach is less
than paid by some others.

-Dave Davis    dave@panix.com     www.panix.com/~dave/quantum