Quantum Interactions on Null Surfaces

What is the meaning of physical locality in Minkowski spacetime, given 
that for any two events A and B there exist other points C such that 
the spacetime intervals CA and CB are both zero.  This leads to the 
question of whether physical effects can operate "in both directions" 
along a null interval ("is locality transitive?").  The space-time 
separation between the transmission and absorption of a photon is 
zero, so does a photon already 'know' how it will be absorbed when 
it is emitted?  This is closely related to the Wheeler-Feynman
"absorber theory" of advanced and retarded electromagnetic waves,
although the Minkowski metric has similar implications for massive 
particles as well, noting that Schrodinger's wave equations, like 
Maxwell's equations, work equally well forwards and backwards in 
time.  Of course, on a macroscopic level we seem to only observe 
outward "retarded" waves, not inward "advanced" waves.

Several people have suggested interpretations along these lines, but
it seems to me that the actual content of this kind of interpretation
hasn't yet been fully and clearly articulated.  The existing 
descriptions remind me somewhat of the old Lorentz theory of the 
electron around the turn of the century, in which they had the length
and time contractions, etc., but hadn't quite figured out that they 
were dealing with a fundamental aspect of space and time that could 
be explained (and even deduced) very simply from a sound set of
fundamental principles.

Surprising as it seems, nearly a century after special relativity 
was first put forward, physicists have yet to fully grasp the 
physical significance of the Minkowski structure of spacetime,
particularly the singularities in the pseudo-metric that represent
null surfaces.  For any two spacetime points A,B the intersection of 
the corresponding nullcones is a quadratic surface: a hyperboloid if 
the interval AB is spacelike, an elipsoid if AB is timelike, and a 
paraboloid if AB is lightlike.  Notice that of these three surfaces 
only the elipsoid (corresponding to the intersection of time-like 
separated nullcones) is finite.

In conventional terms, suppose an electron is emitted from System 
X at point A and absorbed by System Y at point B, with no intermediate
interactions.  The interval AB is necessarily timelike, and the 
intersection between >A < and >B < is a closed elipsoidal surface
in spacetime.  Relative to the frame of the interval AB this surface 
is simply a sphere of radius r = cT/2 (where T is the time interval 
between A and B) at the instant half-way between A and B.

The interaction between Systems X and Y is symmetrical in time, 
and can be considered to originate on the surface of intersection 
between >A < and >B <.  This surface consists of precisely all 
the points in spacetime that are null-separated from both A and B.  
Two equal and opposite electron waves "emanate" from this surface.  
The positive wave converges along the nullcone >B < to point B, 
and the negative wave converges along the nullcone >A < to point 
A. The net effect is to deduct an electron from System X at point A 
and add an electron to System Y at point B.

This basic model can be used to represent all physical interactions, 
and gives results entirely consistent with observation.  However, 
this view has the distinct advantage that, because interactions occur 
along null absolute separations, all the familiar "quantum paradoxes" 
of locality vanish.  To illustrate, consider the traditional EPR 
experiment in which two initially coupled particles are discharged 
from emitter A in opposite directions and are absorbed by spin-
detecting sensors at locations B and C.  If w denotes the difference 
between the spin-orientation test angles at sensors B and C, we 
expect the probabilities of the four possible outcomes to be as 
shown below
                            spin at C
                       up              down
  spin    up      (1+cos(w))/2      (1-cos(w))/2
  at B   down     (1-cos(w))/2      (1+cos(w))/2

The time-symmetric interpretation has no difficulty describing how 
such a correlation can be realized, because the discharge emanates 
from the surfaces of intersection >A<|>B< and >A<|>C<, both 
of which are on  the nullcone of A.  The coupled discharge at A takes 
place only if/when two suitably correlated negative electron waves 
reach it simultaneously.  These electron waves are correlated with 
the absorptions of the corresponding positive electron waves at 
sensors B and C because the two emanating surfaces are null-
separated from points B and C.