Non-faithful representations of surface groups into \(\mathrm{SL}(2,\mathbb C)\) which kill no simple closed curve (Q2516425)

The paper under review proves the existence of representations of the fundamental group of a closed, oriented surface \(S\) of genus at least 4 into \(\mathrm{SL}(2,\mathbb C)\) as in the title: the kernel of the representation is nontrivial but it does not contain any simple closed curve. The existence or not of those representations was asked by \textit{Y. N. Minsky} [RIMS Kokyuroku 1153, 1--19 (2000; Zbl 0968.57500)], as the non existence would have implied the simple loop conjecture for hyperbolic three manifolds, which is still open. The proof uses character varieties. The authors show that there are uncountably many such representations by constructing an affine subvariety \(Z\) of the character variety of \(S\), so that the required properties are satisfied for every character in \(Z\) away from a countable union of proper subvarieties (hence of positive codimension). Fix a simple closed curve \(C\) in the surface \(S\) that bounds a punctured torus, then \(Z\) is defined to be the set of characters of \(S\) such that their evaluation at \(C\) is 2. Since a representation of a group generated by two elements is reducible precisely when the trace of their commutator is 2, representations in \(Z\) are noninjective (they are noninjective already when restricted to the punctured torus). Then for each simple closed curve remove the set of characters of representations that are trivial at this curve, and for each element of the fundamental group, remove all characters that are real but not \(2\) at this element. This amounts to removing countably many proper algebraic subsets, so the result is an open subset of \(Z\) consisting of characters of representations that are not injective, nontrivial on simple closed curves, and so that all finite order elements in their image are trivial. It is worth noticing that irreducibility of \(Z\) is already a non-elementary result interesting by itself. The authors also give a new proof of a theorem of \textit{W. M. Goldman} [Adv. Math. 54, 200--225 (1984; Zbl 0574.32032)]: an irreducible representation of a surface group in \(\mathrm{SL}(2,\mathbb C)\) is a smooth point of the character variety. As explained in a footnote in the introduction, other approaches to Minsky's question have appeared after submission of this paper.