A Spinning-Top Perspective on the Zeros of the Riemann Zeta Function

If the critical line were merely an average axis that individual windows of zeros happen to share, GUE-type local fluctuations would give each window a slightly different torque, and a weak symmetric torque alone would struggle to keep that axis identical over long ranges. To make this precise, consider a rigid-body rotation model in which the axis evolves from one zero window to the next with Markovian memory, and analyse its stability on the basis of the arithmetic torque. If the functional equation acted as a coercive stabiliser, pinning the axis regardless of the other torques, the stability of the critical line would be imposed rather than explained, and any argument toward the Riemann Hypothesis built on it would dissolve into the artificiality of the symmetry itself. It is more natural to model the functional equation as a soft, non-coercive symmetric torque. Under this assumption, if the axis is the same in every window, a soft torque applied continuously keeps the rotation centred without difficulty. If instead the axis deforms from window to window, each window must absorb the drift left by earlier ones, so a soft torque applied continuously and uniformly always acts on an axis that has already moved, and this lag turns into precession. A restoring torque with a memory of ℓ windows is stable only below K_ℓ = 2 sin(π / (2(2ℓ+1))), a threshold that tends to zero as the memory grows, so for any fixed soft torque long-range Markovian correlation of the axis cannot be sustained. Within this model, the persistence of a single critical line over arbitrarily long ranges therefore favours an axis that is held fixed in every window over one that deforms and is corrected after the fact.