Hi Michael,
You have raised a number of interesting issues. I will number my responses to them for clarity.
(1) You wrote: "...the assertion of S^7 ONLY precludes the possibility in physics that the two spaces have different origins such that the S^3 is not a physical subspace of S^7."
The "S^7 only" assertion is not strictly necessary for my analysis to go through. However, from the point of view of quantum correlations, separation of S^3 from S^7 in the manner you have suggested seems to be unjustified. In my picture quantum correlations are correlations among events occurring within spacetime---or equivalently among the clicks of a network of detectors. As far as EPR type correlations are concerned these events can be viewed as occurring within S^3. But S^3 is definitely not enough to reproduce quantum correlations beyond those exhibited by the 2-level systems. For example, the correlations exhibited by the GHZ state can only be reproduced as events occurring within a parallelized 7-sphere. To be sure, the clicks we observe appear to us as occurring within R^3. So the "extra" dimensions of S^7 are certainly hidden from us in that sense, but these dimensions are not necessarily compactified as in your work. In fact I tend to view the correlations exhibited by states like GHZ as the *evidence* that the rotation group of the physical space is S^7, not S^3, with the latter being only a special case of S^7. Still, this does not seem to necessitate the "S^7 only" assertion.
(2) Like Tom, I very much like your meta-principle: "make no preference. This means no preferred speed, ie. the speed of light is always the same, no preferred location (homogeneity) and no preferred direction (isotropy) - these also say no boundary to the space." Absolutely marvellous!
(3) But the following separation is potentially in conflict with my analysis: "With space being S^3 and the 'particle space' being S^7..."
For the reasons explained above, in my analysis the separation of S^3 as "physical space" from S^7 as "particle space" is not justified. All measurement events are occurring within S^7, but we only see them as occurring within R^3. This, however, does not seem to be in conflict with your earlier statement that "physical S^7 space at every point x in the locally flat R^3 space appears to give the point at which to start considering comparisons with Joy's work."
(2) Much of what you say in your Part 2 below has to do with the 'flattening' you require to get QT fully consistent and complete. The flattening required for my analysis to go through has to do with "absolute parallelism", as in teleparallel gravity. Since both S^3 and S^7 are simply-connected manifolds, absolute parallelism is equivalent to their curvature tensors vanishing identically, with torsions within them remaining non-zero in general. This is automatically the case if we view S^3 and S^7 as sets of unit quaternions and octonions, respectively. The very algebra of quaternions and octonions then provides means to define orthonormal frames at each point of these manifolds. This however induces torsional twists within them, and it is these twists in the manifolds that are responsible for what we observe as strong quantum correlations. The latter have nothing to do with non-locality or entanglement per se, because the distant events within S^3 and S^7 are now causally linked by distant parallelism in a non-mysterious way. In other words, in my picture the correlations between distant events are no more mysterious than the innocent correlation between Dr. Bertlmann's socks discussed by Bell.
Best,
Joy