The Fourth Clock: A Flow Parameter Advances Not Inside a Manifold but Through the Space of Manifolds.
Paper 118 showed that the single word “time” covers THREE LOGICAL TYPES: the coordinate dt, the functional of a curve dtau, and the functional of a measure dS. THIS PAPER FINDS THAT THE ROLL-CALL WAS INCOMPLETE – and finds the gap INSIDE THIS SERIES’ OWN PAPERS. The “time” appearing in Ricci flow (dg_ij/dt = -2 R_ij, Paper 103) and in renormalisation-group flow (Paper 110) is NONE OF THE THREE: IT IS A FOURTH TYPE.
FIRST, THE DIMENSION SETTLES IT. Solving Ricci flow on a round n-sphere gives r^2(t) = r_0^2 – 2(n-1)t (solved here; for n=2, r_0=1 the sphere collapses at t* = 1/2). Matching dimensions forces [t] = L^2 – A SQUARED LENGTH, NOT A SECOND. This is the structure of the heat equation du/dt = kappa*Laplacian(u), where [kappa] = L^2/T so [kappa t] = L^2; Ricci flow has kappa = 1, so the parameter itself carries the L^2. Since dt and dtau both carry dimension T, and TWO QUANTITIES OF DIFFERENT DIMENSION CANNOT BE THE SAME QUANTITY (the measuring stick of Paper 116), the identification is closed. The RG parameter log mu is DIMENSIONLESS – a third dimension again.
SECOND, THE DOMAIN DIFFERS. dtau lives on curves INSIDE a manifold: name one worldline and a value follows. The Ricci parameter is not advancing inside the manifold at all; what advances is A PATH INSIDE THE SPACE OF METRICS Met(M). The decisive point: under Ricci flow THE METRIC ITSELF IS CHANGING – THE VERY INSTRUMENT THAT MEASURES LENGTH IS BEING DEFORMED. Since the instrument is moving, this cannot be a length measured by that instrument. dtau advances inside a manifold; the fourth clock advances through the space of manifolds.
THIRD, THE ROOT OF THE ARROW DIFFERS. Ricci flow is parabolic (under the DeTurck trick), and in Fourier space the backward problem gives u_k(t) = u_k(0) exp(-k^2 t): forwards every mode decays, backwards every mode blows up. At |t| = 1 this is e^1 ~ 2.7 for k=1, e^25 ~ 7.2×10^10 for k=5, and e^400 ~ 5.2×10^173 for k=20 (verified numerically here) – the backward problem is ILL-POSED. THE EQUATION ITSELF IS IRREVERSIBLE, with no observer involved. Hamiltonian flow by contrast is REVERSIBLE (Liouville, Paper 77), and the arrow of dS comes from COARSE-GRAINING, a matter of the observer’s resolution (Papers 77, 78). THE SINGLE WORD “IRREVERSIBLE” COVERS TWO DISTINCT ROOTS: an ANALYTIC root (parabolicity) and a RESOLUTION root (coarse-graining). They must not be mixed.
FOURTH, AND THIS IS THE CORE: INSIDE THE FOURTH TYPE, THE (dt, dS) SPLIT RECURS. The parameter t is A RELABELABLE LABEL – rewriting the flow as dg/ds = -2 lambda Ric and setting s = t/lambda traces THE SAME TRAJECTORY in Met(M) on a different scale; what is geometrically invariant is the trajectory, not the ticks along it. WHAT ACTUALLY CARRIES A DIRECTION IS A MONOTONE: Perelman’s W-entropy (non-decreasing along Ricci flow) and Zamolodchikov’s c-function (non-increasing along RG flow in two dimensions). Were these not monotone, the flow would have no “which way”. SO THE FOURTH TYPE IS NOT ONE NEW CLOCK BUT A NEW PAIR (LABEL, MONOTONE), structurally parallel to (dt, dS) but living in a different space. BRAKE: this parallel is not sameness. W is NOT thermodynamic entropy – it is a functional on triples of metric, function and scale, not on probability measures on phase space. The line Paper 88 drew between two entropies is drawn again: W is not S_coarse-grained and W is not S_BH.
FIFTH, A TENSION RECORDED HONESTLY. By Friedan the one-loop beta function of the two-dimensional sigma model is beta^g_ij = alpha’ R_ij, so RG FLOW AND RICCI FLOW ARE THE SAME EQUATION (Paper 110). YET THERE ARE TWO ACCOUNTS OF WHY IT CANNOT BE REVERSED: on the Ricci side PARABOLICITY (an analytic obstruction), on the RG side INTEGRATING OUT MODES (discarded information). MATCHING EQUATIONS DO NOT LICENSE MERGING THE ACCOUNTS – the discipline by which Paper 110 warned that “surgery ~ renormalisation is a rhyme, not a theorem”. Note further that the RG obstruction is “discarded information”, the same word as the root of the dS arrow; but SAMENESS OF WORD IS NOT SAMENESS OF ROOT – dS discards RESOLUTION IN PHASE SPACE, RG discards HIGH-MOMENTUM MODES. Until the addresses are written, they are not joined.
SELF-AUDIT: Paper 103 had already braked, writing that Perelman’s entropy is A CONTROL DEVICE FOR THE FLOW AND NOT THE ARROW OF TIME OF THE UNIVERSE. That judgement is correct – BUT IT DID NOT SAY WHAT TYPE OF QUANTITY A CONTROL DEVICE IS. All this paper adds is that type-name: a monotone along a path in Met(M), whose parameter is a label of dimension L^2. This is not a correction of Paper 118 but A CONTINUATION OF ITS ROLL-CALL: the line 118 drew (separate by type) applied unchanged to the fourth type, and because the instrument was right, the missing cell became visible. Relation to Paper 117: that paper said the series holds no Root-D result and that beyond the wall runs a known passage, “written on the map but not walked”. THAT PASSAGE AND THIS FOURTH TYPE ARE IN THE SAME PLACE – renormalisation-group flow. BUT THIS PAPER HAS NOT WALKED IT EITHER: it only determined the TYPE of the “time” standing in that passage; no fixed point, no critical exponent and no anomalous dimension has been computed.
IMPORTANT (scope): a structural interpretation and roll-call of established standard facts at Layer 2 (Hamilton’s Ricci flow, its explicit sphere solution, backward ill-posedness of parabolic equations, Perelman’s W-entropy, Zamolodchikov’s c-theorem, Friedan’s beta function). NO NEW THEOREM IS PROVED and no prediction is made. W is not identified with thermodynamic entropy (88, 103). The Ricci parameter is NOT called physical time – its dimension is L^2, and no claim is made that the universe evolves by Ricci flow. NOTHING IS CLAIMED ABOUT THE POINCARE OR GEOMETRISATION CONJECTURES, in whole or in part. Parabolicity and mode integration are not fused (110). The fourth type is not made “the real time” – as in Paper 118, electing any one as genuine is itself the error of type. The blank of Paper 117 remains. Layer-3 lenses are quarantined with a non-scientific label.
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 Fourth Clock — A Flow Parameter Advances Not Inside a Manifold but Through the Space of Manifolds: the Dimension $L^2$ Settles It — [Paper 119]. Zenodo.
DOI: 10.5281/zenodo.22036353
ORCID: 0009-0006-8587-3450