The Yamagishi Time-Operator and Clock-Subsystem Theory: Pauli’s theorem needs the Weyl relation, and the internal clock only moves the obstruction. Third edition.
Against the non-existence of a time operator in quantum mechanics (Pauli’s theorem) and the disappearance of time in quantum gravity (the Wheeler-DeWitt equation HΨ=0), this paper organises established content — Pauli’s theorem, POVMs, and the Page-Wootters mechanism — in the vocabulary of computer architecture. It claims no new physics whatever. What the third edition adds is an audit of what the second edition’s proof and closing paragraph actually rested on.
First, the proof of Theorem 2.1 uses the Weyl relation, not the commutator. The second edition imposed the canonical commutation relation [H,T]=iħ and derived U†HU = H+ε from it, but that line does not follow. Formally the Baker-Campbell-Hausdorff expansion terminates because [A,B] is a c-number; but that termination requires [T,H] to be a multiple of the identity on the whole space, and the commutator holding on a dense invariant domain does not give that. What licenses moving an unbounded operator out through an exponential is precisely the Weyl relation. And the canonical commutation relation on a dense domain does not imply the Weyl relation (Nelson’s example; Reed and Simon, vol. I, Section VIII.5). The gap is not empty: Galapon (2002, Proc. R. Soc. Lond. A 458, 451-472; arXiv:quant-ph/9908033), making Pauli’s implicit assumptions explicit, constructs bounded self-adjoint time operators canonically conjugate to Hamiltonians with semibounded or discrete spectrum. So saying that no self-adjoint time operator exists says too much; what is right is that none satisfying the Weyl relation exists. The third edition states Theorem 2.1 in that form.
Second, [H,T]=iħ has no finite-dimensional solution: tr[H,T]=0 for any matrices, while tr(iħI)=iħN is nonzero (verified numerically: over 200 random Hermitian 6×6 pairs the trace of the commutator vanished throughout). Hence H and T are necessarily unbounded and no operator identity can be written without specifying domains. The first point is not a formality but a remark about where this theorem lives.
Third, the Page-Wootters mechanism does not remove Pauli’s obstruction. For the constraint H = H_C tensor I + I tensor H_S with H|Psi⟩=0 to have a non-trivial solution, the spectrum of H_C must meet minus the spectrum of H_S; if the latter lies in [E0, infinity) and is unbounded above, then H_C is not bounded below. Indeed the textbook ideal clock takes the spectrum of H_C to be all of the reals. That is exactly dropping the boundedness-below that Theorem 2.1 requires. The ideal clock does not dissolve Pauli’s obstruction; it works where the obstruction does not reach. And a physical clock is bounded below — for such a clock the pointer is not a self-adjoint operator but a covariant POVM. Pauli’s obstruction has not vanished. It has moved into the clock.
Fourth, this is the same shape as the criticism this paper made of itself. The second edition rejected the first edition’s external rendering loop on the ground that it raises at once the question of what time the loop itself obeys, so that it is a relocation and not a solution. That judgement is correct. But the internal clock the second edition adopted is, with respect to Pauli’s obstruction, also a relocation. Only the destination differs — the first edition moved it outside the system, the second into the clock. The former is clearly worse, generating an infinite regress; but the latter being written as though it were a solution must be corrected. Apply your own criticism to your own next move.
Fifth, the two difficulties have distinct roots. The second edition’s closing paragraph said both arise from trying to treat time as a canonical variable internal to the system, and that Page-Wootters dissolves both at once. That unification is corrected. Pauli’s obstruction comes from the spectrum of H being bounded below — a spectral fact about a single system. The Wheeler-DeWitt difficulty comes from H not being a generator of time evolution at all but of gauge transformations, that is from its being a constraint — a fact about the structure of the theory. The former is a constraint imposed by stability, the latter one imposed by general covariance; the same word, time, simply occurs twice. To write that one mechanism solves both, one must first show that both have the same root. This paper does not show it.
Sixth, the name Read-Only attribute is withdrawn. What POVM-ness means is that the observable is unsharp — that it has no sharp, repeatable, projection-valued representation — which is standard vocabulary in quantum measurement theory (Busch, Grabowski and Lahti) and a different concept from an asymmetry between reading and writing. Read-Only misses in three places: position and momentum are also described by POVMs in real measurements, so POVM-ness does not single time out; no operation corresponding to writing to time exists anywhere in the formalism, so the prohibition has no content; and read-only memory implies that reading leaves the object unchanged, whereas quantum measurement generally disturbs the system. What remains is the fact itself — time observables exist as unsharp observables — an internal fact of quantum mechanics that needs no metaphorical reinforcement.
It is also made explicit that the conditional state obeys the Schrödinger equation exactly only in the absence of an interaction term coupling clock and system. Smith and Ahmadi (Quantum 3, 160 (2019); arXiv:1712.00081) show that with an interaction term the conditional state obeys a time-nonlocal Schrödinger equation, the system Hamiltonian being replaced by a self-adjoint integral operator. The H used here has no such term, so the statement stands, but exactly is a conditional word. The second edition’s corrections are all retained: the withdrawal of the first edition’s external loop, the deletion of the reference to Paper 29’s Read-Only attribute (that concept being withdrawn in the second edition of Paper 29), and the weakening of the claim that the correspondence table guarantees the picture is not a mere metaphor (a correspondence is not explanatory power).
IMPORTANT (scope): a structural interpretation and self-audit of established standard facts (Layer 2). Everything used is established (Pauli 1933 and its later refinements, time observables as POVMs, the Page-Wootters mechanism of 1983 with the experimental illustration of Moreva et al. 2014 and the refinement of Giovannetti, Lloyd and Maccone 2015). NO NEW THEOREM AND NO PREDICTION. What is not claimed: any new physical law (there is none) / that Pauli’s theorem follows from the commutator alone (the Weyl relation is needed) / that Page-Wootters removes Pauli’s obstruction (it works outside the hypothesis, and for a physical clock the obstruction reappears as a covariant POVM) / that the two difficulties share a root (not shown) / that the Read-Only metaphor is accurate (the correct name is the unsharp observable) / that the analogy with computer architecture is a physical necessity (Section 7, quarantined as Layer 3). What remains: apply your own criticism to your own next move.
On the making of this work: The ideas and content of this work stem from the author’s own considerations. Assistance from an AI (a large language model) was used for structuring, English translation, and checking the algebra. Any remaining errors or misinterpretations are solely the author’s. Feedback and corrections are sincerely appreciated.
Reference
Yamagishi, Y. (2026). The Yamagishi Time-Operator and Clock-Subsystem Theory [Paper 54]. Zenodo.
DOI: 10.5281/zenodo.21603325
ORCID: 0009-0006-8587-3450