Jonathan
Thanks for the good review. I did not break out fractal geometry in this, but it lurks in the wings. I indicate first what underlies this structure and then indicate how this leads to fractal geometry. I will get to loop variables and related systems at the end.
The underlying construction here involves the Leech lattice. The 27 dimensional Jordan exceptional algebra describes the 26 dimensional bosonic string when a light cone condition is imposed on the system. This involves three copies of the octonions related to each other by triality, which are 24 dimensional, plus two independent scalar terms. Modulo the triality condition this is a gauge-like theory in 8 plus 2 = 10 dimensions, which is embedded in 11 dimensions according to the light cone condition, or the infinite momentum condition as Susskind calls it. The system is then in greater generality a 24-dimensional Leech lattice system with two time dimensions. That might sound a bit odd, but I then break the symmetry of the system. One of the octonions is broken into H_4xH_4+permutations of 8 according to the Weyl group. H_4 is the four dimensional analogue of the dodecahedron or its dual icosahedron. This is the 120-cell, called the hyperdodecahedron or dodecachoron, which is a polychora with 120 octahedron boundaries, 720 pentagons, 1200 edgelinks and 600 vertices, and Schlafli index {5; 3; 3}. The dual is the 600 cell, with 120 vertices which define a group under quaternionic multiplication. This group is sometimes called the binary icosahedral group, which is a double covering of the icoshahedral group. The symmetry group of the 600-cell is the Weyl group H_4 ~ {3; 3; 5}, a group of order 120_2 = 14400.
The octahedron will tessellate a hyperbolic three dimensional space, see attachment. I worked out how the H_4 or dodecachoron will tessellate a four dimensional hyperbolic space. Now, these hyperbolic spaces might sound intimidating, yet the two dimensional version is the Poincare disk, which I attach as well in an Escher print. I need to write up this proof and get it published --- I did this last August, but other concerns have kept me away from doing that write up. This hyperbolic space then exists in the anti-deSitter spacetime AdS_5, with group structure SO(4,2) ~ SU(2,2). This space has a duality with conformal fields on the boundary of the AdS, which is the celebrated AdS/CFT duality of Maldacena. I am currenlty working out how the Jordan matrix derives the Born-Infeld action, which is the cornerstone of M-theory and AdS/CFT duality. The conversion of the closed string into an open string on the M_2 brane I work in my paper is a simply stringy aspect of M-theory.
Fractal geometry enters into the picture (finally) with the branching which occurs in the tessellation. This is easily seen with the image of the Poincare disk. There are arcs which leave the boundary, reach into the interior and then return to the boundary. In the AdS spacetime these are particles which leave at an energy close to infinity, and approach the boundary again in the limit of zero energy. This describes a renormalization group (RG) flow. RG flows have fractal structure to them with Hausdorff dimensions, such as if you work the Polchinski-Wilson renormalization group. The tessellations describe this according to the branching of the tessellation. At each vertex there are other flows which intersect a particular flow. The renormalization group flow then has a self-similar structure to it, which quantum mechanically is related to the so called cauliflower fractal.
The RG flow stops at a finite energy where particles assume masses comparable to their energies. This stopping is related to how it is there is a quantum critical point in the determinant of the cosmological constant. It is also related to something called the zitterbewegung, or the confined vibrational motion of particles, where the electron is the standard case.
Where Causal Dynamical Triangulation (CDT) and loop quantum gravity is of interest is when one takes this structure and Wick rotates it to a de Sitter spacetime. We obviously do not live in a universe with two time directions. So the t - -> -it on one of the time dimensions maps this structure into the de Sitter spacetime we actually live in. Well, as a caveat I should say we live in a universe which is asymptotically approaching a de Sitter vacuum. There are a lot messy stuff going on with local clumps of matter and the breakdown of RG flow going on. Now loop variables and related systems come into this because what I have been working up has twistor construction in it. Yeah, let's not forget them! They are Penrose's original combinatorial system of doing quantum gravity. So assume we have in a low energy setting the AdS spacetime that maps into a dS spacetime under the Wick rotation. Yet the endpoints of RG flows perturb this spacetime into another configuration with local bumps, waves, black holes and so forth. Now black holes are most important, for black holes in the universe have connections to a BTZ black hole in the AdS. I sort of ignored that bit here, but that is an important aspect of this system. There are I think connects with these other combinatorial aspects of quantum gravity on the dS spacetime after the Wick rotation from AdS to dS is performed.
I have an extended paper in this blog dated sept 22, where I start to lay out how I think LQG might fit into this picture. I became blocked in this effort and felt I needed to post this promised extension. I think somehow the constraints of LQG, and maybe as well the CDT structures, fit into this picture somehow. I think these are really versions of twistor theory, or some aspect of it, in some way. If you read my extended paper you will see how I work on the AdS-black hole, and I started to work up a LQG (strictly spacetime) version of this.
Cheers Lawerence B. CrowellAttachment #1: 2_Hyperbolic_orthogonal_dodecahedral_honeycomb.JPGAttachment #2: Escher_Circle_Limit_III.jpg