Topology of singular fibers of differentiable maps (Q1890496)

The book is a systematic summary of the author's research done in the past decade, partly with co-authors. The goal is to investigate differentiable maps \(f:M\to N\) through their singularities \(f:(M,f^{-1}(y))\to (N,y)\). The set \(f^{-1}(y)\) is generically discrete (often finite) in the \(n\leq p\) case, hence the \(n\leq p\) case nicely fits into the study of multi-singularities. The author, however, concentrates on the ``\(-1\)-codimensional'' case, i.e. \(p=n-1\). In Part I of the book, Saeki deals with \(n=2,3,4\) and carries out the first standard step of singularity theory: he classifies the generic smooth map germs up to diffeomorphisms of the source and of the target. As a result, the target manifold of a generic map \(f:M\to N\) is stratified according to these right-left equivalence classes of \(f\)-fibers. A careful study of these strata and their hierarchy (i.e. incidences) results in statements comparing numerical invariants of some of these strata. A prototype result is the following: for a stable smooth map of a closed 4-manifold into a 3-manifold the number of singular fibers containing both a cusp and a fold point is always even. In Part II of the book, the abstract algebraic framework of the coexistence results of Part I is developed. A Vassiliev type universal complex is defined and studied, that is, a cochain complex whose generators are the right-left equivalence classes of map germs, and whose differentials are determined by the hierarchy of the singularities. In low dimensional cases the author computes the cohomology of the universal complex---with respect to a certain equivalence relation. Then a ``Thom polynomial'' type map is defined (and partially computed): given a generic map \(f:N\to P\), a cohomology class of the universal complex is mapped into the cohomology of the target manifold. In the last three chapters of Part II the author studies cobordisms of smooth maps with a given set of singularities, and shows how a cohomology class of the universal complex gives rise to cobordism invariants. The book is undoubtedly a milestone in global singularity theory of smooth maps of negative codimension: a very promising research topic. This clearly written and nicely illustrated volume is suitable for anybody in the field from graduate students to researchers.