X72-03: Relational Geometry and the Coercivity Gap

Continuing from Round 02's three critical-carrier barriers, this round no longer relies on a single critical amplitude, and instead keeps the full relational geometry — the strain tensor S, vorticity ω, eigenvalues, and alignment angle. Proves that 'amplitude-only observation fails': S_grow = diag(−2a,a,a) and S_decay = diag(−a,−a,2a) have the same amplitude but opposite-sign determinants, so a single scalar amplitude cannot preserve the sign generated by the nonlinearity — a single-observable rejection theorem X_{Γ_amp} within a restricted context. Establishes a conditional closure criterion from the middle eigenvalue λ2 and the aligned strain σ, but proves that a constant geometric dissipation factor does not change the superlinear closure class. Finally stops at STOP-C06 (the relational-geometry/coercivity gap): the geometry has recovered the sign and a conditional criterion, but no theorem yet forces the Navier–Stokes dynamics itself into the safe region, handing off to the next round's direct study of the geometric evolution dynamics.