Quandle twisted Alexander invariants (Q2159509)

In this paper, the authors define a version of the twisted Alexander polynomial using representations of quandles, rather than groups. In the ordinary version of twisted Alexander polynomials, the polynomials are defined by using a group representation of the knot group, [\textit{X. S. Lin}, Acta Math. Sin., Engl. Ser. 17, No. 3, 361--380 (2001; Zbl 0986.57003); \textit{M. Wada}, Topology 33, No. 2, 241--256 (1994; Zbl 0822.57006)]. The group-twisted Alexander polynomials may be calculated by using a twisted version of the Alexander matrix, and are an invariant of the knot group and the chosen representation. In order to define a twisted version of the Alexander polynomial using quandles, the authors introduce equivalence relations on matrices, which will seem familiar to readers accustomed to the usual Alexander polynomial calculations from presentations of the first homology module of the covering space corresponding to the commutator subgroup. The twisted matrix itself was previously defined by the authors in [Fundam. Math. 259, No. 1, 1--31 (2022; Zbl 1497.57010)], and is recalled in this paper, along with a type of \textit{derivative} which parallels the Fox free calculus. The authors show that the resulting polynomial which can be calculated in this manner is an invariant of the knot quandle and the chosen representation (Theorem 5.2 in this paper). They use this result to show that a certain pair of knots known as \(11n38\) and \(11n102\) may be distinguished in a very straightforward way by using the quandle twisted polynomials. It would be very interesting to see more work on this topic. In particular, the natural question to ask is how much stronger the quandle twisted Alexander invariants are than group-twisted Alexander polynomials. Since quandles form an almost-complete invariant for knots, and since the knot group is determined by the quandle (in particular, the knot group is the smallest group containing the knot quandle as a sub-quandle with conjugation as the quandle operation), the quandle-twisted Alexander invariants must be at least as strong as the group-twisted invariants. However, it is not clear how much stronger they are.