Screening as Syntactic Limitation: Nine Puzzles of Quantum Mechanics from a Single Hilbert-Space Constraint
Fabien Dechery
Abstract
We demonstrate that nine foundational puzzles of quantum mechanics — superposition, the Born rule, state reduction, decoherence, entropy, entanglement, the arrow of time, the black-hole information paradox, and the Unruh effect — are not independent problems requiring separate solutions but nine facets of a single structural limitation: the inability of the linear Hilbert-space habitat $\mathcal{H}$ to express the nonlinear, non-separable dynamics of the generative substrate $\mathcal{G}$. The argument rests on the axiomatic closure of the descriptive interface $\mathcal{D}$ established in the companion articles [20, 22]: the informational axioms that define $\mathcal{D}_{\mathrm{QM}}$ as a coherent syntax force its linearity as a theorem, and this linearity constitutively screens all nonlinear content arriving through the coarse-graining map $\Pi: \mathcal{G} \to \mathcal{D}$. We formalize four structural properties of $\Pi$ (non-injective, non-surjective, non-linear, non-factorizing) and show that each property generates a specific subset of the nine phenomena. Crucially, the screening is not a binary on/off filter: quantitative results from the Emergence Programme [25, 26] reveal an *anatomy of screening* resolved by isospin channel, Dirac bilinear structure of the insertion operator, and proper-time depth $\tau$, with deficit magnitudes ranging from $< 2\%$ (mass, Class A) to $100\%$ (isoscalar magnetic quark form factor, Class D). The Gisin-Polchinski no-signaling constraint, which appears to forbid any ontological nonlinearity, is dissolved as a category error: it prohibits nonlinearity at $\mathcal{D}$, not at $\mathcal{G}$. The spectral invariants of $\mathcal{G}$ (including the mass gap) propagate faithfully through $\Pi$ as global properties of the coarse-graining equivalence classes.