A supplement to the monodromy theorem (Q923783)

Let \(f: ({\mathbb{C}}^{n+1},0)\to ({\mathbb{C}},0)\) be a holomorphic function germ with an isolated singularity at 0. Take \(\epsilon\),\(\eta\in {\mathbb{R}}\) with \(0<\eta \ll \epsilon \ll 1\) and put \[ X=\{z\in {\mathbb{C}}^{n+1}/\quad | z| <\epsilon \text{ and } | f(z)| <\eta \},\quad X_ t=X\cap f^{-1}(t)\text{ and } X^*=X\setminus X_ 0. \] Then \(f| X^*: X^*\to \{t\in {\mathbb{C}}| 0<| t| <\eta \}\) is a \(C^{\infty}\) fiber bundle. Let \(X_ f\) denote a typical fiber of this. Then \(X_ f\) has the homotopy type of a wedge of \(\mu\) n- spheres, where \(\mu\) is the Milnor number of f. Let h: \(X_ f\to X_ f\) denote the characteristic homeomorphism of this Milnor fibration. The monodromy transformation T of f is defined as \(T=h^*{}^{-1}: \tilde H^ n(X_ f,{\mathbb{C}})\to \tilde H^ n(X_ f,{\mathbb{C}}).\) The following theorem is well-known. Monodromy Theorem: The eigenvalues of T are roots of unity. The Jordan blocks of T are of size at most \(n+1\). Jordan blocks for eigenvalue 1 of T are of size at most n. In this paper the following supplement to this theorem is given. Theorem 1: If T has a Jordan block of size \(n+1\) (necessarily for an eigenvalue \(\neq 1)\), then T also has a Jordan block of size n for the eigenvalue 1. Two proofs of this supplement are given. The first works only for nondegenerate functions with respect to their Newton diagram, but has independent interest, because it expresses the Hodge-Steenbrink numbers \(h^{n,n}_{\neq 1}\) and \(h_ 1^{n,n}\) in terms of the Newton diagram. The second proof works in the general case and Steenbrink's theory of the mixed Hodge structure on the cohomology group of \(X_ f\) is applied in it. In the latter half of this paper a related conjecture is given and a proof of the conjecture in some restricted cases is given. The reviewer thinks that the second proof of Theorem 1 does not work if the set V of \((n+1)\)-fold points of \(D\cup \tilde X_ 0\) is not sufficiently large. (We follow the notation in the paper.) Moreover, in the proof of the conjecture the inequality \(c,d>1\) on the line 13 from below on page 232 does not always hold. For example for the plane curve singularity \(x^ 3+xy^ 3=0,\) we have \(e_ 1=9\), \(e_ 2=5\), \(e_ 3=3\) and thus \(c=\delta_ 2=\gcd (e_ 2,e_ 1)=1\ngtr 1\) and \(d=\delta_ 3=\gcd (e_ 3,e_ 1)=3\). However, with some minor supplements, we can easily make all the proofs complete.