Symplectic symmetries of 4-manifolds (Q882480)

The authors study symplectic finite group actions on a symplectic 4-manifold. The key tool afforded by the symplectic hypothesis is the use of a \(G\)-equivariant version of Seiberg-Witten-Taubes theory [\textit{C. H. Taubes}, Math. Res. Lett. 1, 809--822 (1994; Zbl 0853.57019) and J. Am. Math. Soc. 9, 845--918 (1996; Zbl 0867.53025)] in studying the fixed point set of the symplectic action. The key point is that the equivariant version of Taubes' results can be used to show that the canonical class of the symplectic 4-manifold is represented by a set of 2-dimensional symplectic subvarieties invariant under the group action. This is most effectively applied in the case of a minimal symplectic 4-manifold with \(c_1^2 = 0\): in this case, the set \(C = \cup_i C_i\) of the 2-dimensional subvarieties representing the canonical class is relatively simple. The main result of the paper is a complete description of the structure of the fixed-point set for a symplectic cyclic action of prime order on a minimal symplectic 4-manifold with \(c_1^2 = 0, b_2^+ \geq 2,\) where the induced action on \(H_2(M;\mathbb Q)\) is trivial. In this case the connected components of the set \(C\) of subvarieties are described and the possible representation data for fixed points within \(C\) is described, with fixed points outside \(C\) having local representation contained in \(SL_2(\mathbb C).\) These techniques are combined with application of the \(G\)-signature theorem (which gives its own restrictions even in the topological locally linear case and had previously been used effectively for pseudofree, homologically trivial actions by \textit{A. L. Edmonds} [Trans. Am. Math. Soc. 299, 155--170 (1987; Zbl 0623.57027)] and \textit{A. L. Edmonds} and \textit{J. H. Ewing} [Am. J. Math. 114, 1103--1126 (1992; Zbl 0766.57020)]) to give the results here. As a special case (with a simpler independent proof given), it is shown that for a symplectic 4-manifold \(M\) with a trivial canonical class (over \(\mathbb Q)\) and nonzero signature, if \(b_2^+ \geq 2,\) then any homologically trivial (over \(\mathbb Q)\) symplectic action of a finite group \(G\) on \(M\) must be trivial. As a corollary, this implies that a homologically trivial symplectic symmetry of a \(K3\) surface is trivial. This result should be contrasted with an open question by Edmonds (Problem 4.124 from the Kirby Problem List) as to whether any homologically trivial, smooth finite group action on a \(K3\) surface must be trivial.