Re: "spooky action at a distance"

[Joao Leao]

scerir wrote:

Joao Leao:

> The association between non-locality and "retrocausality"
> (for lack of a better word) is anything but simple! In any
> case it has less to do with the flow of time than with its
> negation! [...]

Bell's theorem shows that, given the hidden variable lambda,
the result of the experiment at B is dependent on the angle
of the measurement at A, *or* the the result of experiment at
A is dependent on the angle of the measurement at B, *or* both.
Now, because of symmetry, it must be both. Thus, if there are
"retrocausations" (or "influences", or "weak signals" as Ian
Percival calls them) they are in both directions (and with the
same probabilities).

So yes, it is difficult to show that the flow of time is
involved. Antoine Suarez (and the Geneva Group) speaks of
a-temporal quantum effects.

That is a defensible poin-of-view. The "time symmetric"
approach does not conceive the measurement in these terms
though! It requires the actual symmetrization of the coincidence
measurements, what the Aharonov school calls "pre- and post-selection".
This is a way of symmetrizing the initial with  the final conditions...

Your proposal below is not lunatic in the least bit, though!
It has been mentioned in the literature several many times.
I just don't have the references handy.
(but check http://arXiv.org/abs/quant-ph/9511002  )
It is a tricky point to
reconcile it with the usual description but it can be done.
You have to bear in mind that the correlations can only
be exacted a-posteriori from the coincidence counts.
A single pairwise detection will not provide you with
any retrodictory inference about an actual value
being set before you choose the basis of observation.
(This is an instance of a delayed choice experiment by
the way.)

 
Now let us imagine this set-up.

I suppose it can be useful also within the MWI, at
least as a possible answer to the question "If we live
in all of them can we pick the cheapest one?". So I go
on trying to describe this gedanken experiment (or
perhaps lunacy).

There is the usual SPDC source, two correlated photons,
mirrors m1 and m2, one human observer, polarization detectors
(measuring photon-1) and, very close, 4 boxes to collect photon-2.

Of course the path of photon-1 is shorter that the path of
photon-2, so there is a time-delay, for photon-2 going
into one of those boxes (possible delayed choice here?).

m1
/----------<-----------<--source-->------>- detectors
|
|
|
\------------------>----------------->----- boxes (1,2,3,4)
m2

Now the observer can measure, with his detectors, or the
linear polarization of his photon-1, or the circular
polarization of his photon-1.

Of course the observer, having measured his photon-1,
can predict what is the polarization state of photon-2.
There are 4 possibilities: linear/x, linear/y, circular/+,
circular/-.

Being very short the distance between detectors and
boxes, the observer has time (due to that time-delay)
to move there and pick up the right box (that one with the
right label: linear/x, linear/y, circular/+, circular/-)
and collect, into the right box, the photon-2 which
is arriving.

This is possible because he *knows* what was his *choice*
while measuring, with detectors, the polarization state
(linear *or* circular) of photon-1. And he also *knows*
what was the measurement outcome for photon-1: i.e.
linear/x, or circular/+, or ...

This is also possible because the observer has *time* to
move to the other location and pick up the right box,
to collect photon-2.

But before observer makes his *choice* the photons
(and especially photon-2, which is "late") were
already flying.

So you could ask: what was the polarization state of
photon-2, before the observer made his choice measuring,
with his detectors, the polarization state of photon-1?

The answer seems to be that photon-2 fits equally well
in both categories, that is to say: linear polarization
and circular polarization. Thus neither of these
properties can be ascribed to it as an objective property.

Now you can also ask: what if I cut the path lenght
of photon-2 and I make it equal to the path lenght
of photon-1? It happens that the observer becomes
unable to move from the detectors location to
the boxes location, because there is no time-delay
now. So, in these conmditions, the observer, loses
control of the situation. His information remains
hidden, or useless, ot impossible. But this, imo,
does not mean that photons gain some objectiveness.
Or not?

Not really. I mean, what is at stake is not the objectiveness
of the photons but of the value of their polarization in either
base (or both). The fact that you may have stored a value
that you did not know while you waited to "objectify" it
does not particular help you...

 
Of course you excluded the possibility of (weak or strong)
signals traveling FTL, from detectors or from photon-1
to photon-2. In example making the path lenght of
photon-2 much much longer than the coherence lenght
of the photon(s).

But imagine that your procedure (here above) is not
enough, and actually there is some FTL effect.
The interesting point here is that any FTL effect
from detectors or photon-1 makes actual, objective
the state of photon-2 *before* its measurement.

An FTL effect is definitely a possibility but somewhat
heavyhanded to deal with the "passion-at-a-distance"
business. FTL would indict QM a lot harder than EPR.
Read Svetlichny's paper and the several sequels...

-Joao

-- 

Joao Pedro Leao  :::  [EMAIL PROTECTED]
Harvard-Smithsonian Center for Astrophysics
1815 Massachussetts Av. , Cambridge MA 02140
Work Phone: (617)-496-7990 extension 124
Cell-Phone: (617)-817-1800
----------------------------------------------
"All generalizations are abusive (specially this one!)"
-------------------------------------------------------