Genus inequalities and four-dimensional surgery (Q1818707)

In a previous paper [Am. J. Math. 119, No. 5, 1119-1137 (1997; Zbl 0886.57027)] the authors showed that for many homology classes one can find a topologically locally flat (TLF) embedded surface whose genus is minimal among all TLF embeddings. The genus of this surface is in general considerably lower than that given all smooth embeddings, and the difference between the minimal genus in the TLF and smooth categories is a quadratic function of the divisibility of the class and so can be arbitrarily large. The authors continue their work on determining the minimal genus \(m(\alpha)\) for TLF embeddings. In this paper, the authors relate the problem of finding the minimal genus for other classes not covered by the previous result to the surgery conjecture of 4-dimensional topology. The earlier results involved a value \(m_0(\alpha),\) which is the least positive number \(g\) satisfying \[ 2g-2 \geq -\chi(X) + \max_{0 \leq j < d} |\sigma(X) - 2j(d-j)(1/d^2)\alpha\cdot \alpha|. \] They showed that for simply connected 4-manifolds, \(m(\alpha) \leq m_1(\alpha)\), and the two are equal for even classes or those of prime power divisibility. Here they generalize their results using group theoretic hypotheses relating to the fundamental group of the complement of the embedding to define another quantity \(m_1(\alpha)\) and use it to connect the question of whether \(m(\alpha) < m_0(\alpha)\) to the failure of topological surgery in dimension 4.