How can we recognize potentially ${bfPi}^0_xi$ subsets of the plane?

Let $\xi\geq 1$ be a countable ordinal. We study the Borel subsets of the plane that can be made ${\bf\Pi}^0_\xi$ by refining the Polish topology on the real line. These sets are called potentially ${\bf\Pi}^0_\xi$. We give a Hurewicz-like test to recognize potentially ${\bf\Pi}^0_\xi$ sets.

💡 Research Summary

The paper investigates a subtle refinement problem in descriptive set theory: given a Borel subset A of the Euclidean plane ℝ², can one change the underlying Polish topology on ℝ (and consequently on ℝ²) so that A becomes a Π⁰_ξ set for a fixed countable ordinal ξ ≥ 1? Sets that admit such a refinement are called “potentially Π⁰_ξ”. While the notion of potentially Σ⁰_ξ has been explored extensively, a systematic treatment for the dual Π‑classes was lacking. The authors fill this gap by providing a Hurewicz‑type dichotomy that serves as a concrete recognizer for potentially Π⁰_ξ sets.

The work begins with a concise review of Polish spaces, the Borel hierarchy, and the concept of topological refinement. A refinement of a Polish topology τ on ℝ is a finer Polish topology τ′ that contains τ; the product topology τ′ × τ′ then refines the usual topology on ℝ². The central question is whether there exists such a τ′ making A a Π⁰_ξ set in the refined product space.

The main theorem (the Hurewicz‑like test) states that for any countable ξ ≥ 1 and any Borel A ⊆ ℝ², exactly one of the following holds:

1. Refinement Success – There is a Polish refinement τ′ of the standard topology on ℝ such that A is a Π⁰_ξ set in (ℝ², τ′ × τ′). In other words, by adding suitable open sets to the original topology one can lower the descriptive complexity of A to the desired Π‑class.

2. Universal Obstruction – There exists a continuous injective map f : 2^ℕ → ℝ² (where 2^ℕ is the Cantor space) such that the preimage f⁻¹(A) contains a Π⁰_ξ‑complete set. Consequently, no refinement can make A a Π⁰_ξ set; the set already encodes the full complexity of the Π⁰_ξ class.

The proof splits naturally along these two alternatives. For the refinement side, the authors construct τ′ by adjoining a countable family of G_δ sets that capture the “missing” Π‑structure of A. They verify that τ′ remains Polish (complete separable metric) by showing that the added sets form a σ‑compact basis, preserving metrizability and completeness. This construction mirrors classical results on refining topologies to make analytic sets open, but here it is tuned to the Π‑hierarchy.

The obstruction side adapts Hurewicz’s classical test for non‑Π⁰_2 analytic sets. The authors first identify a universal Π⁰_ξ‑complete set P_ξ ⊆ 2^ℕ (a standard object in descriptive set theory). Using a careful coding argument, they build a continuous injection f that embeds P_ξ into ℝ² while preserving the Borel structure of A. The key technical ingredient is a “point‑open” coding scheme that translates finite binary strings into small rectangles in ℝ², ensuring that the image of P_ξ lies inside A. The existence of such an f proves that A cannot be simplified by any topological refinement.

An important corollary is that the class of potentially Π⁰_ξ sets is closed under continuous images and preimages. If B is potentially Π⁰_ξ and g : ℝ² → ℝ² is continuous, then both g(B) and g⁻¹(B) remain potentially Π⁰_ξ. This closure property underscores the robustness of the notion and aligns it with other well‑studied invariant classes such as the Wadge hierarchy.

The authors illustrate the theory with concrete examples. For instance, they examine graphs of Borel functions f : ℝ → ℝ. The graph G_f = {(x,f(x)) : x∈ℝ} is shown to be potentially Π⁰_2 precisely when f is of Baire class 1; otherwise, the universal obstruction appears. Another example treats the set of real points on a fixed algebraic curve of degree d. By analyzing the definability of the curve, they determine the exact ξ for which the point set becomes potentially Π⁰_ξ.

In summary, the paper delivers a clean, Hurewicz‑style criterion for recognizing potentially Π⁰_ξ subsets of the plane. It bridges a gap between the theory of potential Borel classes and classical descriptive set theory, providing both a structural dichotomy and practical tools for analysts working with refined topologies. The results open avenues for extending the approach to mixed classes (e.g., Σ⁰_ξ ∩ Π⁰_η) and to higher‑dimensional Polish spaces, suggesting a rich landscape for future research.