Operational Success Stories

Why precedent is worth consulting

Chapter 42 itemised the reasons to keep looking for an operator whose spectrum is the Riemann zeros. Chapter 43 priced the discovery in commitments we would have to release. Both chapters were written in the conditional. This one is not. Four times in the modern record, a spectrum was known or strongly suspected before anyone could write down the operator behind it, and each time the operator eventually arrived. The record does not tell us whether the Riemann operator exists. It tells us what tends to happen when such an object is found.

Read the cases as case law rather than as encouragement. Precedent constrains expectation; it does not supply evidence about the case at hand.

One: the function-field Riemann Hypothesis

Over finite fields, the analogue of the Riemann Hypothesis is a theorem. Weil proved it for curves; Deligne proved the general case. The proof works by showing that the zeros are literally eigenvalues of the Frobenius endomorphism acting on étale cohomology. A spectral interpretation that had been conjectural became fact.

What the neighbourhood gave up: the clean separation between arithmetic and geometry. Number theory absorbed a large part of algebraic geometry's machinery, and the traffic ran both directions. New cabinets had to be built for objects that belonged to neither discipline as previously drawn.

This is the one place where the Hilbert-Pólya intuition is fully vindicated. It is vindicated in a different arithmetic universe, and no transfer to the integers is known. That caveat is the whole reason Chapter 42 listed this entry as convergence rather than as progress.

Two: Selberg's trace formula and the Laplacian

The Selberg zeta function of a hyperbolic surface has zeros and poles corresponding exactly to eigenvalues of the Laplace-Beltrami operator on that surface. A formal resemblance to the explicit formula for the Riemann zeros became a precise dictionary: primes correspond to closed geodesics, zeros to eigenvalues.

The unusual feature of this case is that the operator was already known. The Laplacian is the least exotic object in differential geometry. The surprise was that so ordinary an operator generates a spectrum with an arithmetic-looking distribution. Spectral geometry acquired an arithmetic flavour, and the trace-formula style of argument spread well beyond its origin.

What the neighbourhood gave up: the assumption that arithmetic-looking regularity requires an arithmetic source. Sometimes the geometry alone produces it.

Three: quantum mechanics itself

Empirical atomic spectra existed for decades before anyone wrote down the operators. Balmer fitted his series in 1885; Lyman, Paschen and others extended the catalogue. Schrödinger and Heisenberg supplied the self-adjoint Hamiltonians whose eigenvalues those spectral lines are. This is the archetype of the situation: precise numbers first, mechanism later.

What the neighbourhood gave up: a great deal. The planetary-orbit picture — the Bohr proxy Chapter 41 retired — had to go. Determinism at the fundamental level was released. Probability amplitudes, non-commuting observables, and the measurement problem entered the permanent furniture of physics and have not left. Several thoughts held inviolate did not survive the arrival of the operator that explained the numbers.

It is worth being exact about the chronology, because the moral depends on it. The operators did not merely reproduce the spectra; they reproduced them from a formalism nobody would have accepted on aesthetic grounds beforehand.

Four: Wigner and the nuclear spectra

Heavy nuclei showed statistical level-repulsion patterns long before there was a theory of them. Wigner recognised that the statistics matched the eigenvalues of large random Hermitian matrices — not one operator, but an entire ensemble. Random-matrix theory followed, and with it the concept of universality.

What the neighbourhood gave up: the expectation that explanation means naming the specific operator. Sometimes the detailed Hamiltonian is inaccessible and less important than its symmetry class. A whole class of operators became a respectable answer.

This case cuts both ways for the Riemann problem. It legitimises the Montgomery-Dyson observation as real physics-style knowledge. It also demonstrates precisely the ceiling Chapter 42 insisted on: symmetry class fixes statistics and never identifies the member.

The pattern

In every genuine success the spectrum was known or strongly suspected first, and the operator was constructed later. Each time, a cherished separation had to be relaxed or redrawn: arithmetic from geometry, classical from quantum, specific from universal. And each time the world did not end. It acquired new permanent furniture and lost a few comfortable illusions.

The pattern is suggestive and it is not an argument. Four cases where the operator was eventually found is a sample selected by the fact that the operator was eventually found. The cases where a suspected spectral structure never resolved into an operator do not get written up, which means this chapter's evidence base is survivorship-shaped by construction.

Still on the lam

For the Riemann zeros over the integers, none of these successes has been repeated. Candidates appear regularly. None has met the four discharge conditions set out in Chapter 42: self-adjointness on a specified Hilbert space, the correct spectrum with correct multiplicities, the correct counting asymptotics, and a trace formula that returns the explicit formula.

The operator remains honest in every fingerprint and unpaid on the ledger, and it still carries the price listed in Chapter 43. The neighbourhood is intact for now. The historical record only says that when these operators finally do arrive, the neighbourhood rarely stays the same.

Which is a reason to read the four cases as warnings rather than as promises. Each success was a good outcome for knowledge and an uncomfortable one for whoever had built a settled position on the ground that shifted.

The 2026 case that does not become a fifth

In July 2026 a group led by Wei, Zhai and Lu, with Nori, Xin and Long among the authors, published in Nature Communications a direct correspondence between the nontrivial zeros of the zeta function and dynamical quantum phase transitions in two engineered quantum many-body systems, one characterised by an average accumulated phase factor and the other by the Loschmidt amplitude. The Riemann Hypothesis is recast there as the occurrence of those transitions at a single temperature, a proof-of-principle run was performed on a quantum processor, and a framework is proposed implementing both systems with polynomial resources.

It is a real result and it is not a fifth success story, and the difference is exactly the one this chapter was written to keep visible. In each of the four cases the spectrum was found in the world first and the operator arrived afterwards to account for it. Here the correspondence runs the other way: the zeta structure is built into the system by design, and the transitions occur because it was put there. The dynamics are then a faithful instrument for reading a function we already had, which is a substantial engineering achievement and not the discovery of an operator that was waiting.

Tested against the four discharge conditions of Chapter 42, the case is therefore silent rather than failed. No self-adjoint operator on a specified Hilbert space is exhibited whose spectrum is the zeros with the correct multiplicities, no counting asymptotics follow from the construction, and no trace formula is returned. What the paper does supply is a new class of physical probe — a way of interrogating the hypothesis numerically with quantum resources — and a genuinely new transition mechanism in nonequilibrium dynamics. Both are worth having. Neither pays the ledger.

The reason to record it here rather than quietly in a footnote is that it is the most likely near-term source of a misreading. A headline saying the Riemann Hypothesis has appeared in a quantum experiment is true as written and will be read as the operator having been found. The distinction between a spectrum discovered and a spectrum engineered is the whole content of this chapter, and this is the first case in the modern record where the two are easy to confuse.