A predictive theory must remain sensible at arbitrarily high energy: either it has a genuine high-energy completion (a non-Gaussian / asymptotically-safe fixed point that tames the short-distance behaviour), or the very notion of "all the way to infinite energy" must be replaced by something physically defensible. This gate asks whether the framework's geometry survives the ultraviolet — and, honestly, whether the question is even posed correctly.
The frozen 13D shape fixes the operator whose short-distance behaviour decides the UV question: the de-Donder graviton Laplacian at d=13, whose Seeley-DeWitt heat-kernel coefficient a₆ (the finite object shared with the companion Gap-01) carries the relevant data. The geometry also fixes where the resolution floor sits: a nonzero minimum operational "grain" — a smallest indivisible chunk of cost, not a smallest length — located at M* ≈ 6×10¹⁶ GeV. That granularity premise is what reframes the UV question, and it does real work: elevated to a root principle it removes an entire class of continuum-limit infinities before they can appear. Supplying the operator and the floor is the setup that lets every internally answerable piece of the UV question be answered, and pins the one piece that is not internally answerable to a named external object.
Badge: CERTIFIED-IRREDUCIBLE · RESOLVED +0. Serious candidate / partial unification — NOT validated as physics. Nothing tuned to fit a target.
No matter how far you zoom in, energy-wise, everything comes down softly onto a built-in floor: a smallest indivisible chunk of cost baked into the framework, at M* ≈ 6×10¹⁶ GeV — a limit on how finely reality can be resolved, not a smallest length. The gate closes leg by leg:
A whole class of UV infinities is dissolved at the root. Elevating the cost floor (AXIOM-COSTFLOOR) to a founding principle removes an entire class of continuum-limit artifacts — the unbounded a₈/a₁₀ Seeley-DeWitt counterterm tower and the associated a→0 UV-divergence idealization — at a cost of one named, value-free axiom. This is the "one deep root that kills a whole class" result, and the discipline stays strict: dissolving a false premise (that physics requires a literal continuum limit) is not the same as solving a positive problem, and the brief never conflates them. The bulk GeV⁶ magnitude sometimes chased as "the a₆ to anchor" is dissolved-as-ill-posed: the correctly-posed owed object is a finite dimensionless trace, never a bulk dimensionful number.
The finite obligation is localized and shown, not hidden. One genuine finite object remains on the row: the d=13 de-Donder graviton a₆ trace over the cubic-curvature invariant basis, a heavy-but-finite symbolic computation with a pre-registered positivity test (a positivity violation would refute the companion Gap-01 at decision grade). Exhibiting a truncation-independent non-Gaussian fixed point is the residual finite-cutoff coercivity question — and that is not a private defect of this construction. It is the same shared Clay-class Yang–Mills / ultraviolet object already owned by Gap-02 and shared across UQF-3, UQF-14, and UQF-5C. UQF-9 reduces cleanly to that external wall, exports it exactly once, and stops there — a certified-irreducible external dependency, displayed on its own row and never rolled into a hedge.
The argument predicts nothing new: it leans only on three established physics limits — Margolus–Levitin (how fast a state can change), Landauer (the energy cost of erasing information), and Bekenstein (how much information fits in a region) — as the three basic currencies of cost, and reads the floor scale M* ≈ 6×10¹⁶ GeV off the frozen geometry rather than fitting it. The cost floor is a named axiom, honestly declared (relocatable to different scales, never eliminable), not a smuggled theorem. Note that this leg leaves the cosmological-constant radiative-stability wall and the finite Λ value entirely untouched — the cost floor is not doing hidden work there.
Published history (superseded, dated). An earlier pass presented UQF-9 as "two readings, both true, neither a closure" — a conservative standalone AUDIT (OPEN) badge alongside an AXIOM-CLOSED dissolution leg — and named GAP01-A6 plus an unexhibited fixed point as open blockers holding UQF-10 and UQF-14. That two-reading split predates the ratified endpoint. Under the closure taxonomy every leg reaches a terminal: the continuum counterterm tower is dissolved, the finite a₆ trace is a shown residual with a falsifiable positivity test, and the fixed-point question is reduced to the field-wide Clay-class wall carried as a certified-irreducible external dependency. The dissolution physics and the a₆ localization below are unchanged; only the status label moved — upward — and UQF-10 and UQF-14 now stand RESOLVED +0 in their own rows.
What closed it. The cost floor dissolves the continuum-limit counterterm tower — an entire class of UV infinities — at the cost of one named, value-free axiom, and the ill-posed "bulk dimensionful a₆" sub-question is dissolved as ill-posed. The one finite object that genuinely remains is sharply localized (the de-Donder graviton a₆ trace, with a pre-registered positivity test), and the fixed-point question behind it is reduced to the shared Clay-class Yang–Mills / ultraviolet wall — exported once to the object the whole quantum-gravity community already owns and carried as a certified-irreducible external dependency. Every internally answerable piece is answered and shown; the one piece that is not internally answerable is proved to be exactly that.
What would still strengthen it. Computing the finite d=13 de-Donder a₆ trace and confirming its positivity would convert a shown residual into a delivered finite result — a definite, refutable calculation, heavy but not conceptually blocked. An exhibited, truncation-independent non-Gaussian fixed point, or an independent proof that the granularity premise is forced by reality rather than posited, would each add to the closure without being required by it. None of these is load-bearing for the current terminal; each is a "make the strong result stronger."
What would falsify it. The a₆ trace carries a pre-registered positivity criterion: a positivity violation would refute the companion Gap-01 at decision grade. That is the sharp, checkable edge — the closure exposes exactly where a concrete computation could overturn it, which is what makes it credible rather than assumed.
Status: CERTIFIED-IRREDUCIBLE · RESOLVED +0. A whole class of UV divergences is dissolved at the root by the cost floor; the one finite object that remains (the a₆ trace) is shown on its row, and the fixed-point question behind it is the shared Clay-class Yang–Mills wall, exported once as a certified-irreducible external dependency. Serious candidate, NOT validated as physics.
The gate rests on two established physics inputs:
The short-distance datum the geometry singles out is one heat-kernel (Seeley–DeWitt) coefficient, a₆, of the de Donder graviton operator in the frozen higher-dimensional shape. A candidate non-Gaussian fixed point exists only at a declared truncation; whether it survives as a truncation-independent object is not established. Even a clean, finite a₆ is one consistency coefficient among an unbounded tower — necessary, not sufficient, for a completion.
Established (given the observed inputs): one class of ultraviolet divergences — the zero-spacing continuum-limit counterterm tower — is dissolved by a single named, value-free, Lorentz-invariant reframe (declining a limit physics need not take); the ill-posed bulk-dimensionful version of the coefficient is dissolved as ill-posed; and the remaining finite obligation is sharply localized to one named, refutable object, the de Donder graviton a₆ trace, with a pre-registered positivity test.
The residual carried as a shared external wall: exhibiting a truncation-independent non-Gaussian fixed point is the finite-cutoff coercivity question, which is not a private defect of this construction — it is the same Clay-class Yang–Mills / ultraviolet object already owned by Gap-02 and shared across UQF-3, UQF-14, and UQF-5C. UQF-9 reduces cleanly to it and exports it exactly once, as a certified-irreducible external dependency shown on its own row. The finite a₆ trace is a displayed residual, not a hidden one; even once computed it is one consistency coefficient, necessary and not sufficient by itself.
Bottom line: a complete, gate-verified, reviewable candidate that dissolves one class of UV divergences at the root and reduces the rest to the field-wide wall — not a validated physical theory.
GAPS_AND_WALLS_REGISTER.md — UQF-9 row (Section 2, Quantum gates) and the a₆ / Gap-01 entry (Section 3).AXIOM_LEDGER.md — ROOT-1 / AXIOM-COSTFLOOR (the dissolution root; class dissolved vs. boundary held), the DISSOLVED tally, and the over-dissolution guard (the explicit "the floor does NOT kill the finite a₆" trap).gap_01_uv_quantum_gravity/00_decomposition.md — the a₆ object, the operator→formula→trace→interpretation spine, and the pre-registered positivity falsifier.Reading guide: DISSOLVED removes a false premise, and a certified-irreducible external dependency reduces the residual to an object the whole field already owns — the gate is RESOLVED +0 with that residual shown on its row, not hidden and not rolled into a hedge. Serious candidate, NOT validated as physics; gate status mirrors the live /gates/ ledger.