The philosophy of mathematics has explored for decades the problem of mathematical efficacy in the description of physical reality, producing answers that range from ontological Platonism to cognitive naturalism. This work argues that these answers share a structural limit, since they frame the issue as a matter of correspondence between two domains and attempt to explain why mathematics describes the real. The question is misguided. What calls for formal explanation is a more specific and more stringent phenomenon: the response of the real to a mathematical inquiry exceeds the structural precision invested in the model. The work introduces a fundamental ontological distinction between M, the structural precision of the mathematical model, and R(M), the precision of the response produced by the real when interrogated through that model. The notation R(M) is not a formal ornament, as it makes explicit that the response of the real depends functionally on the mathematical question while retaining its own ontological status. The central conjecture states that R(M) > M systematically in fundamental physics, with the difference R(M) − M non-decreasing as M increases. A secondary conjecture states that no circular construction yields this property systematically, since the excess constitutes a structural contribution that emerges from the side of the real and carries information absent from the original model. The work develops direct confrontations with the two deepest philosophical positions on the topic. Wittgenstein’s view, which treats mathematical efficacy as a grammatical matter that requires no ontological explanation, is challenged by showing that the directional asymmetry of R(M) arises from the structure of the phenomenon rather than from a choice of use. A grammatical account would allow the excess to move in multiple directions, while the phenomenon advances in a single direction. Gödel’s view, which interprets efficacy as evidence of a direct perception of Platonic structures, is distinguished from the thesis proposed here: R(M) > M can be read as the empirical signature of the Gödelian meta-level, a measurable trace of the fact that something always stands beyond the formal system and expresses itself through the system without requiring a Platonic ontology. The falsifiability of the conjecture is constructed with epistemic care. A single observation with R(M) ≤ M offers no decisive test, since it may reflect the weakness of the model rather than the absence of the structural property. The criterion of falsification concerns the function f such that R(M) = f(M): the conjecture would be refuted if f(M) − M approached zero or became negative as M increased across distinct domains of fundamental physics. The empirical anchoring covers quantum mechanics, quantum electrodynamics and general relativity, showing that the delimitation of the domain (fundamental physics rather than economics or meteorology) is itself a theoretical result. Keywords: philosophy of mathematics; meta-mathematics; analogy; falsifiability; R(M); asymmetric membrane; Wigner; Wittgenstein; Gödel; fundamental physics.