The Shafarevich conjecture predicts that the universal cover of a smooth complex projective variety is holomorphically convex — a deep statement relating topology (the fundamental group) to complex analysis. Aguilar and Garay refine this conjecture for varieties defined over the reals, and the refinement reveals a structural bifurcation.
When the variety has real points — a nonempty real locus X(R) — the universal cover should be real holomorphically convex. When it has no real points, the cover should be dianalytically holomorphically convex, a fundamentally different notion. The arithmetic condition of having real points forces two different analytic geometries on the same topological object.
This is surprising because the Shafarevich conjecture in its complex form treats all varieties uniformly: the topology of the fundamental group determines the analytic geometry of the cover, full stop. The real refinement breaks this uniformity. Whether or not real solutions exist — a purely arithmetic/topological condition determined before any analysis begins — selects between two qualitatively different convexity structures.
Aguilar and Garay prove the refined conjecture in two cases: when X is a curve, and when the fundamental group is nilpotent. The proof introduces the notions of real holomorphic convexity and transverse holomorphic convexity, formalizing the geometric differences dictated by the real locus. The transverse notion captures how the real structure interacts with the complex holomorphic functions.
The purely complex theory cannot see this distinction. Complex holomorphic convexity is a single concept; the bifurcation into real and dianalytic variants only appears when you ask whether the variety has real points. The answer to a yes-or-no arithmetic question reshapes the analytic geometry of the universal cover.
The presence or absence of real points bifurcates the analytic geometry of the universal cover, forcing a structural distinction between two types of holomorphic convexity that the complex theory alone cannot detect.