The paper tackles a fundamental question in mathematical logic and geometry: when can we say that a geometric or algebraic concept is truly "intrinsic" to a structure, rather than being an artifact of how we chose to describe it? The classical answer, going back to Felix Klein's Erlangen program, is that a concept belongs to a geometry precisely when it is preserved by the symmetries of that geometry. A recent result by Madarasz made this precise for certain coordinate geometries built over fields, showing that the concepts definable without extra parameters are exactly those invariant under the geometry's affine symmetries. The current paper identifies the core geometric idea behind that proof and generalizes it substantially.
The key new ingredient is a condition the authors call "semilinear faithfulness." Roughly speaking, this condition says that in any ultrapower of the structure (a standard model-theoretic construction that builds a larger, richer version of the original object), every symmetry of the induced geometry can be cleanly decomposed into two simpler pieces: first an affine transformation of the geometry, then a coordinatewise action of a symmetry of the underlying algebraic base. When this decomposition holds, the paper proves a clean correspondence: a relation definable in the ambient language, without parameters, is a genuine concept of the geometry if and only if the affine symmetry group of the geometry preserves it. The authors also prove a complementary "dual" result about collections of concepts.
The payoff is a unified framework that handles several geometries that the earlier Madarasz theorem could not cover directly, including geometries where some scalars are given fixed names and geometries built over fields equipped with an exponential function. The paper closes by noting that the same factorization principle applies even outside geometry, to any one-sorted structure whose symmetry group can be uniformly coded in a suitable way. Overall, the work clarifies why symmetry-based definability results hold and opens the door to applying them in a broader range of mathematical settings.