The strong topological Rokhlin property is a condition on groups that comes from ergodic theory and dynamics. Roughly speaking, a group has this property if its actions on compact spaces can be approximated in a particularly well-controlled way by actions that factor through finite structures. The paper develops a new, purely combinatorial way to describe this property using finite symbolic data, specifically by characterizing it through what the authors call "globally realizable tuples." This reformulation makes the property easier to work with because it replaces an abstract dynamical condition with a concrete finite checking procedure.
Using this new characterization, the authors prove an important structural result: if a finite-index subgroup of a finitely generated group has the strong topological Rokhlin property, then the whole group does too. In other words, the property "lifts" from a well-behaved piece of a group to the entire group, as long as the piece is not too far from the whole (in the sense of having finite index). This kind of lifting result matters because it lets mathematicians transfer knowledge about simpler or better-understood subgroups to more complex overgroups.
The paper also investigates how complicated the collection of all groups with this property is, using tools from descriptive set theory, which is a branch of mathematics that classifies sets by how difficult they are to describe or define. The authors place this collection of groups within a precise complexity hierarchy, showing it sits in a class called "Pi-0-4" and is at least as hard as a class called "Sigma-0-2." They also find a structural barrier: if the collection turns out not to be describable at a simpler level of the hierarchy, then there must exist a group with the property that cannot be described by finitely many generators and relations, which would be a significant and surprising discovery.