- The first key idea (see this, this and this) was that one can interpret octonions O as Minkowski space M8 (see this) by using the number theoretic inner product defined by the real part Re(o1o2) of the octonion product. Later I gave up this assumption and considered complexified octonions, which do not form a number field, but finally found that the original option is the only sensible option.
- The second key idea was that if either the tangent spaces T or normal spaces N of Y4 ⊂ M8 are quaternionic and therefore associative, and also contain a commutative subspace C, they can be parameterized by points of CP2 and mapped to H=M4× CP2. This would be the first half or M8-H duality.
- How to map the M4 ⊂ M8 projection to M4× CP2? This question did not have an obvious answer. The simplest map is direct identification whereas the inversion with respect to the cm or tip of causal diamond cd⊂ M4⊂ H is strongly suggested by Uncertainty Principle and the interpretation of M8 coordinates as components of 8-momentum (see this. Note that one can considerably generalize the simplest view by replacing the fixed commutative subspace of quaternion space M4 with an integrable distribution of them in M8.
- I considered first the T option in which T was assumed to be associative. The cold shower was that there might be very few integrable distributions of associative tangent spaces (see this and this). As a matter of fact, M4 and E4 were the only examples of associative 4-surfaces that I knew of. On the other hand, any distribution of quaternionic normal spaces is integrable and defines an associative surface Y4. This led to a too hasty conclusion that only the N option might work.
- If M8 is not complexified, the surfaces Y4 in M8 are necessarily Euclidean with respect to the number theoretic metric (see this). This is in sharp conflict with the original intuitive idea that Y4 has a number theoretic Minkowski signature. Is it really the normal space N , which must have a Minkowskian signature? Is also T possible.
- The minimal option in which M8-H duality determines only the 3-D holographic data as 3-surfaces Y3⊂ M8 mapped by M8-H duality to H. The images of Y3 could define holographic data consistent with the holography = holomorphy (H-H) vision (see this, this, this, this and this). Both M8 and H sides of the duality would be necessary.
There are also interpretational problems.
- The proposed physical interpretation for the 4-surface Y4⊂ M8 was as the analog of momentum space for a particle identified as a 3-D surface. In this interpretation the Y4 would be an analog of time evolution with time replaced with energy. A more concrete interpretation of the 3-D holographic data would be as a dispersion relation and Y4 could also represent off-mass shell states. Momentum space description indeed relies on dispersion relations and space-time description to the solutions of classical field equations.
- For N option Y4 must be Euclidean in the number theoretic metric. Therefore the momenta defined in terms of the tangent space metric are space-like. What does this mean physically?
Could the problem be solved if the momentum assignable to a given point of Y4 is identified as a point of its quaternionic normal space as proposed in (see this).
Or should one accept both T and N options and interpret the Euclidean Y4 as as a counterpart of a virtual particle with space-like momenta and of CP2 type extremals at the level of H. At vertices belonging to Y41∩ Y42 me, quaternionic N(Y41) would contain points of quaternionic T(Y42) so that the earlier proposal would not be completely wrong.
- A further criticism against the M8-H duality is that its explicit realization has been missing. For N option, the distributions of the quaternionic normal spaces N are always integrable but their explicit identification has been the problem. For T option even the existence of integrable distributions of T has remained open.
Consider now how the view to be described could solve the listed problems.
- There are two options, which could be called T and N: either the local tangent space T or normal space N is quaternionic. Which one is correct or are both correct?
Any integrable distribution of quaternionic normal spaces is allowed whereas for tangent spaces this is not the case. This does not mean that non-trivial solutions would not exist. Perhaps the rejection of the T option was too hasty.
Furthermore, for X4⊂ H both Minkowskian and Euclidean signatures of the induced metric are possible: could T and N option be their M8 counterparts?
- Holography= holomorphy vision (H-H) allows an explicit construction of the space-time surfaces X4⊂ H. For Y4⊂ M8 the situation has been different. The very nature of duality concept suggests that the explicit construction must be possible also at the level of H.
- Is M8-H duality between 4-D surfaces Y4⊂ M8 and space-time surfaces X4⊂ H or only between the 3-D holographic data Y3⊂ H and X3⊂ M8-H?
It turns out that a modification of the original form of the M8-H duality, formulated in terms of a real octonion analytic functions f(o):O→ O, leads to a possible solution of these problems.
- All the conditions f(o)=0, f(o)=1 and Imf)(o)=0, and Ref)(o)=0 are invariant under local G2 acting as as a dynamical spectrum generating symmetry group since fº g2= g2º f holds true. The task reduces to that of finding the 4-surfaces with a constant quaternionic T or N.
- In particular, M4⊂ M8 has been hitherto the only known Y4 of type T and the action of local G2 generates a huge number of Y4 of type T. Both T and N option are possible after all! G2 symmetry applies also to the N option for which E4 is the simplest representative!
- The roots of Im(f)(o)=0 resp. Re(f)(o)=0 are unions ∪o0S6(o0) of 6-spheres, where o0 is octonionic real coordinate o0. The 3-D union Y4= ∪o0S6(o0) ∩ M4=S2(o0)⊂ M4 has quaternionic tangent space T=M4. The interpretation as holographic data and the M8 counterpart of a partonic orbit is suggestive.
- The Euclidean 3-surface Y3= S6(o0) ∩ E4(o0)=S3(o0) could serve as a holographic data for Y4 with quaternionic normal spaces and with an Euclidean number theoretic signature of the metric. Obviously, the option Y4=∪o0 Y3(o0) fails to satisfy this condition. The interpretation would be as the M8 counterpart of CP2 type extremal with a Euclidean signature of the induced metric. The identification as the M8 counterpart of a virtual particle with space-like momentum is suggestive.
If T resp. N contains a commutative hyper-complex subspace, it corresponds to a point of CP2. Hence Y4 can be mapped to X4⊂ H=M4× CP2 as M8-H duality requires.
- What could be the counterpart of H-H vision in M8? One can choose the function f(o) to be an analytic function of a hypercomplex coordinate u or v of M4 and 3 complex coordinates of M8. The natural conjecture is that the image X4 of Y4 has the same property and satisfies H-H.
- The proposed physical interpretation for the 4-surface Y4⊂ M8 was as the analog of momentum space for a particle identified as a 3-D surface. The interpretation the Y4 as the analog of time evolution with time replaced with energy looks range. A more concrete interpretation of the 3-D holographic would be as a dispersion relation emerges and Y4 could also represent off-mass shell states. Momentum space description indeed relies on dispersion relations and space-time description to the solutions of classical field equations.
Number theoretic discretization as a selection of points as elements of the extensions of rationals defining the coefficient field for f(o) and the replacement of fermions to the "active" points of discretization would realize many fermion states at the level of H. Galois confinement (see this and this) stating that the total momenta are rational numbers would provide a universal mechanism for the formation of bound states.
- For N option Y4 must be Euclidean in the number theoretic metric. Therefore the momenta defined in terms of the tangent space metric are space-like. What does this mean physically?
Could the problem be solved if the momentum assignable to a given point of Y4 is identified as a point of its quaternionic normal space as proposed in (see this).
Or should one accept both T and N options and interpret the Euclidean Y4 as as a counterpart of a virtual particle with space-like momenta and of CP2 type extremals at the level of H.
At 2-D vertices belonging to the intersection Y41∩ Y42, quaternionic N(Y41) would contain points of quaternionic T(Y42) so that the first proposal would not be completely wrong.
- Could the TGD analogs of Feynman diagrams be built by gluing together T and N type surfaces Y4 along 3-surfaces Y3 defining analogs of vertices. In the role of consciousness theorist, I have called them "very special moments in the life of self" (see this) at which the non-determinism of the classical field equations in H-H vision is localized. At these 3-surfaces the smoothness of Y4 fails and could give a connection to the notion of exotic smooth manifold (see this, this, and this), conjectured to make possible particle vertices and fermion pair creation in TGD despite the fact that fermions in H are free (see this, this and this).
For a summary of earlier postings see Latest progress in TGD.
For the lists of articles (most of them published in journals founded by Huping Hu) and books about TGD see this.