Octahedral developing of knot complement. I: Pseudo-hyperbolic structure (Q1622908)

For any link in $S^3$ and any diagram $D$ thereof without a component consisting entirely of over- or under-crossings, D. P. Thurston has given a partially-ideal decomposition of $S^3 \setminus L$ into octahedra canonically associated to $D$, with one octahedron per crossing, in [\textit{D. P. Thurston}, ``Hyperbolic volume and the Jones polynomial'', Unpublished manuscript, \url{http://pages.iu.edu/~dpthurst/speaking/Grenoble.pdf}]. Decomposing the octahedra further into tetrahedra, one obtains a partially-ideal triangulation of $S^3\setminus L$. W. P. Thurston's famous gluing equations are still well-defined for partially-ideal triangulations, and solutions to the equations parametrize pseudo-hyperbolic structures. This system of equations canonically associated to a link diagram $D$ has $4\, c(D)$ or $5\, c(D)$ variables according as one triangulates the octahedra with four or five tetrahedra. \par The key results of the paper under review are two reformulations of these particular instances of Thurston's gluing equations. One system has $2 \, c(D)$ variables and equations, corresponding to segments of the diagram $D$. The other system has only $2 + c(D)$ variables and equations, corresponding to complementary regions of $D$. In both cases there are extra nondegeneracy conditions one may impose, one such condition for each crossing. \par Using these reformulations, the authors construct, for every pseudo-hyperbolic structure $d$ on $S^3\setminus L$ (given as a solution to one of the above reformulations), an explicit representation $\rho: \pi_1(S^3\setminus L) \to SL_2\mathbb{C}$ yielding $d$, in terms of the Wirtinger presentation of $\pi_1(S^3\setminus L)$ coming from $D$. They thereby derive a formula in terms of the segment-variables for the cusp shapes of boundary-parabolic such structures, using the segment- and crossing-labels introduced in [\textit{A. Tsvietkova}, Hyperbolic structures from link diagrams. PhD thesis, Univ. of Tennessee (2012)]. \par Motivated by R. Kashaev's well-known volume conjecture, the paper concludes with various computations: of all boundary-parabolic structures on all $T(2,N)$ and $J(M,N)$ knots (among other examples), after appealing to a result in [\textit{J. Cho}, Proc. Am. Math. Soc. 144, No. 4, 1803--1814 (2016; Zbl 1341.57014)]; of their cusp shapes as above; and of their volumes and Chern-Simons invariants, using a formula from [\textit{J. Cho} et al., J. Knot Theory Ramifications 23, No. 9, Article ID 1450049, 32 p. (2014; Zbl 1320.57015)]. This formula for the Chern-Simons invariant is, modulo $\pi^2$, of the form $-Re(V_0(\mathbf{z}))$ for a function $V_0$, evaluated at gluing equation solutions $\mathbf{z}$. These computations strongly suggest that the real parts of $V_0$ evaluated at boundary-parabolic solutions, omitting the reduction modulo $\pi^2$, fit into some uniform pattern, as opposed to the Chern-Simons invariant itself. \par The reader well may ask what happens to the above framework if, instead of Thurston's gluing equations, one uses the more recent Ptolemy equations defined in [\textit{S. Garoufalidis} et al., Duke Math. J. 164, No. 11, 2099--2160 (2015; Zbl 1335.57034)]. The authors have studied this in their preprint [\textit{H. Kim}, \textit{S. Kim} and \textit{S. Yoon}, ``Octahedral developing of knot complement. II: Ptolemy coordinates and applications'', Preprint, \url{arXiv:1904.06622}].