Can a resonance theory be a renormalizable theory?

In this talk we make an exhaustive analysis of the possible chiral invariant operators that may described the resonance decay S->pi pi. These provide at the same time the only available chiral invariant structures for the loop ultraviolet divergences in this amplitude. Independently of the order in perturbation theory, we find just one single-trace term (four if multi-trace operators are allowed), whose renormalization renders the matrix element finite.