I am anonymous above! Forgot to log in. Here are more thoughts on where this is going, which I have posted elsewhere on FQXI
-----------------------
I am working on a problem involving octonionic black holes, or a way to use the heavenly phere structure of the Jordan exceptional algebra as local systems of 26 dimensional Lorentz spaces. The identification between the three octonions by the triality operation and the light cone structure of the diagonal elements reduces this to 10 and 11 dimensions. The J^3(O) is locally diagonalized by the F_4 group. The F_4 group is the Hurwitz quaternion "1154" representation of the 24-cell. The 24-cell has B_4, D_4 and F_4 representations and the quotient F_4/B_4 determines a short exact sequence between spin(9) and the Moufang plane OP^2. This local Lorentzian system is then given by connection coefficients on this system.
The Jordan algebra of a vector space V, which can be O, is ~ V\oplus R according to the mixed product
(u, α)*(v, β) = (αv + βu, (u,v) + αβ), (u,v) = \langle u, v\rangle
but the inner product gives
(u, α)*(v, β) = (u,v) - αβ,
which is the Lorentz metric. Thus the diagonal entries of the J^3(O) set the three copies of the octonions in a Minkowski geometry. The diagonalization of the Jordan matrix is the defined locally with a "gauged F_4," which defines both the Minkowsi structure and the quantum algebra locally and constructs connection terms between local charts (local inertial frames).
There is for the Cl_8 group 256 dimensions with 9-grading for spins in a cycle of -2, -3/2, -1, -1/2, 0, 1/2, 1. 3/2, 2, which appears as
1 + 8 + 28 + 56 + 70 + 56 + 28 + 8 + 1 = 1 + 8 + 28 + 56 + (35+35) + 56 + 28 + 8 + 1,
which predict eight Rarita-Schwinger fields and a single graviton. E_8 has a 7-grading 8 + 28 + 56 + 64 + 56 + 28 + 8, where the graviton sector and the scalar and pseudo-scalar fields are removed. The extra-octonionic graded structure 1,0,0,0,3+3,0,0,0,1 indicate that the graviton and the scalars (Higgs & dilaton) are derived elsewhere. Yet in this way the Clifford-8 can embed the exceptional E_8. That the graviton and scalars are not independent in E_8 is why there is the extended CL_{16) system, which embeds E_8xE_8, where now the graviton is due to two copies (in string theory interpreted as due to the handedness of the fields on a closed string) .
In particular, in the J^3(O) the two E_8s are related to each other by the "point" corresponding to one octonion, are dual to lines in OP^2, which are themselves OP^1 --- 7 dimensional spaces. The holonomy for the 7-sphere is the G_2 group, which is centralized with respect to F_4 in E_8. The duality here then describes a graded system on O^2 with dual 8_vx8_s representations. This is then one reason why octonionic gravitation involves the F_4 transformation of the octonionic group which diagonalizes the Jordan exceptional algebra. The action of the F_4 and G_2 groups I indicate in the attached jpg file.
BTW, this is one problem with the Lisi program. The graviton was assumed to be framed with the electroweak interactions. Yet the gauge interactions are internal symmetries, while gravitation is an external symmetry. The intertwining of them is through supersymmetric transformations and the graded structure over Lie algebras.
Lawrence B. CrowellAttachment #1: 1_octonionic_curved_space.JPG