The paper develops new mathematical tools for studying how geometric shapes can develop singular points, meaning places where smoothness breaks down in specific ways. These singularities appear naturally when you map one space into another, and a key question in geometry is: how often does a particular type of singularity appear, and what determines that count? A classical approach uses objects called Thom polynomials, which encode these counts in terms of characteristic classes, a standard bookkeeping device in topology. The challenge addressed here is that existing theory works well when the target space is large relative to the source, but breaks down or becomes poorly understood in the opposite regime, which the authors call negative relative codimension.
To handle the unstable regime, the authors introduce the concept of a "Thom series" for a class of singularities called contact function singularities. The idea is to organize an infinite family of related singularities, obtained by a systematic procedure called quadratic stabilization, into a single generating object. This Thom series encodes all the Thom polynomials in the family at once, expressed in terms of combinatorial objects called Schur polynomials. The authors prove two important structural results: the index sets labeling which Schur polynomials appear follow a clean stabilization pattern as the family grows, and the coefficients of those polynomials vary in a controlled, polynomial way. This gives a unified and computable picture across all cases, stable and unstable alike.
On the computational side, the authors carry out extensive explicit calculations. They compute stable Thom polynomials for a large collection of singularity types, and also compute what they call Legendre Thom polynomials, which relate to a somewhat different geometric setup involving contact geometry. They introduce a new class called multi-binary singularities and work out their Thom polynomials as well. The paper is partly preparatory: the authors note that a companion paper will apply these results to concrete problems in enumerative geometry, the branch of mathematics concerned with counting geometric objects satisfying given conditions.