Universally and existentially definable subsets of global fields (Q1720115)

This paper belongs to the area studying definability and decidability in Number Theory. One can arguably call the solution to Hilbert's Tenth Problem by Matiyasevich, building on results of Davis, Putnam and J. Robinson the event that brought the area into being. The solution left open the question whether there was an algorithm for solving diophantine equations over \(\mathbb Q\), and whether there was an existential definition of \(\mathbb Z\) over \(\mathbb Q\) that would imply absence of such an algorithm. Some conjectures by B. Mazur imply that a definition does not exist. At the same time a definition of \(\mathbb Z\) over \(\mathbb Q\) using both existential and universal quantifiers due to J. Robinson has been known for a long time. More recently, Robinson's definition was simplified in various ways. Poonen produced a first-order definition of algebraic integers over number fields using only two universal quantifiers in the definition. Koenigsmann produced a universal (i.e. using only universal quantifiers) definition of \(\mathbb Z\) over \(\mathbb Q\). He also constructed a definition of \(\mathbb Z\) over \(\mathbb Q\) using only one universal and several existential quantifiers. J. Park generalized the universal definition to number fields, that is she constructed a universal definition of the ring of algebraic integers of a number field \(K\) over \(K\) for all number fields \(K\). The question posed by Hilbert and related problems of definability make sense over many other rings, such as a global function fields. We know that Hilbert's Tenth problem to be undecidable over all global function fields. R. Rumely constructed the first version of first-order definitions of polynomial rings and rings of \(S\)-integers over global function field. A. Shlapentokh produced an analog of Koenigsmann's one universal quantifier definition for global function fields. The paper under review constructs the analog of the universal definitions of Koenigsmann and Park over these function fields. It also contains other interesting definability results known over number fields. The authors give a Diophantine (existential) definition of non-squares and non-norms. Many tools used by Park and Koenigsmann to construct their definitions transfer to the global function field setting, but not all. This leads to several technical difficulties successfully handled by the authors. For a student in the area reading side by side the Park paper and the paper under review would make for a very nice introduction to many technical aspects of definability.