The Equals Sign as a Rotation

This is the proven floor of the whole project. Where the Millennium essays carry conditional steps and open conjectures, this one does not. It is a short theorem in a Hilbert space, and it is true outright. Everything bolder in the volume is an application of it; this is where to start if you want to know what is actually nailed down. The full preprint is on Zenodo; judge it there.

The paper opens with one question: what does the equals sign do? The answer it gives is that for a particular and common kind of equality — between two quantities that share a fixed budget — the equals sign is secretly a rotation.

Start with the definition. Two quantities are complementary when they partition a conserved resource: gaining one costs the other, with nothing left over. Call them V and D, normalized so each lives in [0,1]. The claim is that complementarity alone forces

V² + D² = 1.

The proof, in plain steps

"Partition a conserved resource" has an exact meaning. It says there are two orthogonal subspaces H_V and H_D of a Hilbert space H, with H = H_V ⊕ H_D. Orthogonality is the whole content of the trading relation: a vector living in H_V has zero component in H_D, and vice versa — that is what "nothing left over" means.

Let P_V and P_D be the projections onto those subspaces, and let Ω be the unit vector that is the state of the system. Then by the Pythagorean theorem in Hilbert space — equivalently, Parseval's identity for a two-part decomposition —

‖Ω‖² = ‖P_VΩ‖² + ‖P_DΩ‖².

The cross-term ⟨P_VΩ, P_DΩ⟩ vanishes precisely because the two projections are orthogonal. Set V = ‖P_VΩ‖ and D = ‖P_DΩ‖, and you have V² + D² = ‖Ω‖² = 1. That is the entire proof.

Why a circle and not a line

The result that matters is not just the identity but its shape. Because the constraint is the ℓ² norm, the solutions trace a circle, and you can write V = cos θ, D = sin θ. The angle θ ∈ [0, π/2] is the single degree of freedom — one number names the whole state. The equals sign, here, is the rotation that moves θ.

Compare the naive guess, V + D = 1. That is the ℓ¹ constraint, and it draws a straight line segment, not a circle. The paper is sharp about when each holds: the ℓ¹ line is what you get when interference survives — when the cross-term ⟨P_VΩ, P_DΩ⟩ is nonzero and must be carried. Strict complementarity means that cross-term is exactly zero, and the moment it is, the norm is ℓ² and the geometry is a circle.

Why this one is the foundation

Every other essay on this site leans on this identity, so it is worth being clear about its status:

• It is unconditional. No pair correlation conjecture, no continuum limit, no separator assumption. Two orthogonal subspaces and Parseval — that is all it needs.
• It is general, and that cuts both ways. The theorem holds for any two complementary quantities. The work — and the risk — in the other papers is the claim that a particular V and D (zeta zeros, Wilson loops, tissue states) genuinely form orthogonal sectors. This essay proves the consequence; each application has to earn the premise.
• It tells you exactly where to push. If you doubt one of the bolder results, the question is never "is V²+D²=1 true" — it is "are these two things really orthogonal." That is the honest seam, and this proof is what isolates it.

None of that is a dismissal of the reach. A clean, general theorem that names the one degree of freedom in any complementary pair — and makes plain that the geometry is a circle the moment interference dies — is the right floor to build a project on. It is the part that is simply true.

Related: the identity near zeta zeros · the identity in gauge theory