The Present Holds · Phronesis

A reading, offered as one. Under the Hilbert–Pólya program the Riemann hypothesis stops being a statement about where certain zeros sit and becomes a statement about whether a state survives its own time evolution intact: a zero off the critical line is an energy that is not real, an evolution that is not unitary, and a present that frays as it goes. Everything in the chain is standard mathematics except its first link, which is a conjecture nobody has proved — and this note is careful about which is which, because the reading is worth having only if it is not mistaken for a result.

The coordinates

Write a nontrivial zero as ρ = ½ + ε + it. The two coordinates are not symmetric in what they do. t is the height up the strip; ε is the departure from the critical line. v, the small library this comes from, names them outright: t is time and ε is space — and once they are named that way, the critical line ε = 0 is the pure-time axis.

A zero on the line is all now and no elsewhere.

That is a naming, not a theorem. It earns its keep only if something downstream cares about the distinction, and something does.

The conjecture the reading rests on

Hilbert and Pólya are supposed to have asked, independently and without publishing it, whether the numbers t — the heights of the zeros — are the eigenvalues of some self-adjoint operator H. If they are, the Riemann hypothesis follows immediately, because a self-adjoint operator has a real spectrum, and a real spectrum puts every zero on the line.

No such operator is known. This is a programme, not a theorem, and it has been open for a century. The nearest thing to a candidate is the Berry–Keating suggestion that the classical Hamiltonian H = xp sits behind it, which reproduces the right counting function but has never been made into a rigorous quantum operator with the right spectrum. Everything below is conditional on the programme being right, and that conditional is the whole reason this is a note and not a paper.

What follows, if it is

Grant the operator. Then the rest is textbook, and it is worth walking because the conclusion is not obvious from the premise.

A self-adjoint H has real eigenvalues.
A self-adjoint H generates time evolution by U(t) = e^−iHt, and that U is unitary.
Unitary evolution preserves the norm. Nothing in the state grows, nothing decays; the whole of it is carried forward.

Now let one eigenvalue go complex — equivalently, let one zero sit off the line, ε ≠ 0. Then H is not self-adjoint, U is no longer unitary, and e^−iHt acquires a real exponent. Some modes grow without bound and others decay to nothing. The state does not persist; it frays, and what is left after a while is not what set out.

So, on this reading, the Riemann hypothesis is the condition that there is a stable present to be in — a now that carries forward whole rather than tearing into parts that blow up and parts that vanish. v compresses it to four words: real ↔ ground state.

What is proved, and what is read

Stated flatly, so nobody has to guess:

Proved. Self-adjoint implies real spectrum implies unitary evolution implies conserved norm. Standard functional analysis, not in question.
Proved. If the Hilbert–Pólya operator exists, RH follows.
Open. Whether such an operator exists. A century of work, no construction.
A reading. That “unitary evolution” is well described by the words a present that holds. Nothing forces that description. It is chosen because it is faithful to what the mathematics does, and it could be declined without touching a single result.

The honest summary is: if the programme is right, RH is equivalent to a stability condition on time evolution — and calling that stability “the present” is an interpretation laid on top of that equivalence, not a further theorem.

Why it is worth writing anyway

Because it changes what the question feels like, and that occasionally matters. “Do all the zeros lie on a line” is a question about the location of things. “Does this system evolve without tearing” is a question about stability, and stability questions come with a different toolkit and different intuitions about what a counterexample would even look like. A hypothetical zero off the line stops being a dot in the wrong place and becomes a mode that runs away.

It also lands on the same structure as everything else here. The critical line works as a reference precisely because it is not a stored value — it is the fixed set of an involution, ρ ↦ 1 − ρ̄, and so it cannot drift, be recomputed, or be quietly kept current. That is the one property a frozen reference must have, and the zeros are the only case in that note’s table with no recorded failure. Not a coincidence worth mysticism — a consequence of being a fixed point rather than a variable.

What would make this more than a reading

One thing: a construction. An explicit self-adjoint operator whose spectrum is the zero heights, or a proof that none can exist. Everything above turns on that single open link, and no amount of reframing supplies it.

Short of that, the honest test is narrower and still worth running: does the stability framing ever produce a statement that the location framing does not? If it only redescribes, it is a good sentence and nothing more. If it suggests a constraint — something about the growth of modes that has a counterpart in the zeros — then it has earned more than the word “reading”. It has not yet, here.