|
Joined: 06 Oct 2005
Posts: 602
Location: Toon Town
|
The standard hand-waving argument is that local interference fringes
will never be seen at either end no matter how the transmitter is set in
the future either at "0" or "1". One can still see nonlocal fringes in
the hindsight correlations after-the-fact, i.e. no effective practical
"time travel to past" signal nonlocality. The basic idea here is that
the two entangled photons mutually effectively "which-way" measure each
other. But is this really correct in all possible total experimental
arrangements?
One must show this formally and at least quasi-rigorously of course.
The typical argument is to look at say the receiver photon R that can
take paths A or B. If that photon was not entangled the state at the
detector would be
|RA> + U(A-B)|RB>
Where U(A-B) is the unitary operator representing the path difference
for that one photon R.
The FRINGES happen when at position x on the screen for photon R
<x|RA> + <x|U(A-B)|RB>
And the Born probability density fringe terms include
<RA|x><x|U(A-B)|RB> + cc
However, in the case of the entangled photons R and T where T has a
choice of slits C & D landing at x', the usual argument is that the
initial local momentum conservation constrains the pair state to be
correlated into something like
<x|RA><x'|TC> + <x|U(A-B)|RB><x'|U(C-D)|TD>
The argument is then that we must integrate over all future x' to know
what will be seen at a single x in the past.
It is then argued that the dx' integral of
<TC|x'><x'|U(C-D)|TD> vanishes and since it is a coefficient of
<RA|x><x|U(A-B)|RB> there are no local fringes in the past. You can
switch roles and also get no local fringes in the future no matter what
you do.
That is, the orthodox argument is that each photon will quasi-measure
the other photon and the time sequence of the detections will not
matter. That is, the two photons quasi-measure each other. I use
"quasi-measure" here because of Marlan Scully's "quantum eraser" effect.
The basic "reason" for this is that unitary operators U preserve inner
"bra-ket" products, so that if the dx' integral <TC|TD> = 0 initially
they will remain so. However, this argument is not well posed because
the two paths for photon T to land at x' will generally take different
amounts of time.
All we know is that U(t)*U(t) = 1, but we do not know that U(t')*U(t) =
1 when t =/= t'.
On Jun 26, 2006, at 10:45 PM, Jack Sarfatti wrote:
Is anyone able to refute John Cramer's gedankenexperiment for
backwards-through-time reverse causation in which the future creates the
past?
I have not had time enough yet to think hard enough about Cramer's
particular proposal in this attachment from the AAAS USD meeting last week.
<Reverse Causation.pdf>
If Cramer is right here then an important part of Lenny Susskind's (and
recently Stephen Hawking's) theory of information loss down black holes
is shot down because then we can, in principle, see beyond the horizons
to the other worlds of the multiverse. The Cosmic Landscape is then
testable. George Chapline's theory is also shot down if this works
because he rejects closed timelike curve effects in his "dark star".
The signal nonlocality, hitherto thought to violate orthodox quantum
theory, is what explains remote viewing and possibly other paranormal
phenomena.
See Martin Gardner's "Magic and Paraphysics" for the history of this idea. |
|
|
