Home·The theorem·The proof in one reading·Architecture·Units·Manuscript·Glossary·FAQ·Work items·Data·Method·Provenance·Changes
← III.11 Interacting rank, Gaussian comparison and the stress tensorPart III·Detailed notesIII.13 Alternative construction: one Wilson family for all labels →

III.12. Bounds uniform in the volume and convergence as the spacing tends to zero

Contents

1. What this unit proves 1

2. What it uses 3

3. Strategy 4

4. Steps 5

5. Where the difficulty lies 6

6. Notation 6

7. What remains unexplained 6

Code dependencies 7

Lean source 7

Additional source locations 8

Key statements 13

Part III — Ultraviolet construction (renormalization group).

Combines bounds uniform in volume with fixed-volume convergence as the lattice spacing tends to zero.

Status: edited. Condensed from the unit's detailed notes, which were checked against the linked Lean in two rounds and edited; the local limitations still apply. Status meanings.

Overview · All units · Detailed notes · ← Previous: III.11 · Next: III.13 →

Results · Inputs · Strategy · Proof steps · Notation · What remains unexplained · Code dependencies · Lean source · Key statements

Source inventory and conventions

Part: III, Ultraviolet construction (renormalization group). Lean: unit III.12, YM/FullVolume/. Size: 18,973 lines in the source allocation.

1. What this unit proves

Proposition III.12.1 (kernel bounds uniform in the volume). Let \(R\) be a run for a simple gauge group \(G\), with Wilson family \(C\). Fix \(n\), labels \(v_i \in \text{Lab}_{d_i}\), a finite list \(\lambda\) of pairs (slot, axis), and \(r > 0\). There are \(B \geqslant 0, N_0\) such that, for every \(N \geqslant N_0\), every \(p\) and every integer configuration \(x\) with \(a_N \text{dist} (x_i, x_j) \geqslant r\) for \(i \ne j\),

\(\displaystyle \left| \Delta_{\lambda} \mathbb{E}_{N, p} \prod_{i = 1}^n O_{N, p} (v_i ; x_i) \right| \leqslant B.\)

Each difference acts in its indicated slot; distance is periodic \(\ell^{\infty}\) distance. Constants may depend on \(R, n, (v_i), \lambda, r\), but not \(p\). Status: theorem for every run. Lean: genuine.

Proposition III.12.2 (derivative closure). For a run with simple \(G\), every sequence \(w_{N, p} \in \text{Lab}_d\) bounded by some \(K \geqslant 0\) eventually in the mesh, and every finite axis list \(l\), the trial \(\Delta_l O (w)\) is funded. Status: theorem. Lean: Run.financed.

Proposition III.12.3 (closure lemma). Fix \(d\). Assume a consistent Wilson family \(C\) and:

  1. the centred budget;

  2. the scale-selection data;

  3. sound sections \(V_e\) and linear automorphisms \(W_e\) of abstract labels for every degree \(e\), satisfying the section, symmetry and exact local identities described below, with Portfolio;

  4. a presentation \(\pi\) and the full section-comparison data, including quantitative comparison with \(V_e\), identification of \(\ker \pi\) with the dressed equation ideal, and \(\sup_{N, p} \| \pi_{N, p} \| < \infty\) in each degree;

  5. degree-\(d\) detector rows with the placement and bounded-lift conditions and short-distance comparison of VitalRank.

If \(F\) is funded, every \(F_{N, p} (o)\) is continuous in the gauge field, \(l\) is an axis list, and \(x\) is a degree-\(d\) coin with \(O (x) - \Delta_l F\) owed at cost \(u_{N, p} \to 0\) in the mesh, then \(\Delta_l F\) is funded. Status: conditional theorem; a run's ties and short-distance estimates supply these hypotheses. Lean: improved.

Proposition III.12.4 (uniform rank detection). Under hypotheses 1–5 of III.12.3, fix \(l\). For all large \(k\) and all sufficiently small \(b > 0\), there is \(Z \geqslant 0\) such that, eventually in the mesh,

\(\displaystyle \| \pi_{N, p} x\| \leqslant Z \|\mathcal{P}_{k, b, l ; N, p} (\pi_{N, p} x)\| \quad (x \in \text{Lab}_d) .\)

The detector takes correlations with fixed partners and successively averages the first slot in the directions of \(l\) over lengths between \(b\) and \(2 b\); repeated axes give repeated averages. Status: conditional theorem. Lean: trained.

Proposition III.12.5 (negligible equation labels). Assume consistency of \(C\) and hypotheses 2–4 of III.12.3. Every degree-\(d\) coin \(x\) with \(\pi_{N, p} x_{N, p} = 0\) for all \((N, p)\) has \(O (x)\) owed at cost \(a_N^{1 / 32}\). Status: theorem. Lean: banked.

Proposition III.12.6 (coupling power-series bounds). Let \(R\) be a run with a Campaign, Pillars and bounds Sky. Let \(\bar{\beta}\) be the target level-0 coupling and \(w > 0\) the fan width. Fix \(p\), an \(n\)-tuple of labels \(V_0\) and \(r > 0\). There exist \(b, m \in \mathbb{N}\), \(m \geqslant 1\), \(B \geqslant 0\) and \(c \in (0, 1]\) such that the following holds for every \(m \leqslant j < N\) and \(a \in [- 1, 1]\).

Take any family \(F (s)\), \(s \in [- 1, 1]\), of admissible campaign members with common block sequence, sections and labels, whose level-0 inverse coupling at \((N, p)\) is \(\bar{\beta} + was\). For every \(r\)-separated configuration \(o\) in \(\Lambda_{j + 1, p}\) embedded into \(\Lambda_{N, p}\),

\(\displaystyle \mathbb{E}^{F (cs / 2)}_{N, p} \prod_i O (V_{0, i} ; o_i) = \sum_{i \geqslant 0} P_i s^i, \qquad \sum_i |P_i | \leqslant Ba_{j + 1}^{- b} \quad (s \in [- 1, 1]) .\)

Expectation and representatives belong to the attached auxiliary transfer family; its correlations agree with the successful trajectory at positive cutoffs. Constants depend only on campaign data, \(p, V_0, r\). Status: conditional theorem; VIII.1 supplies the campaign records. Lean: Face.financed.

Proposition III.12.7 (fixed-volume Cauchy property). Fix simple \(G\), one RG construction with stationary-path bounds, free amplification, observable-transfer bounds and common reference sections, and assume its uniform perturbation estimates. Fix \(p\). Let \(C_n\) come from histories at cutoffs \(f (n) \to \infty\). Assume:

For every finite label list and every configuration \(x\) with \(x_i - x_j \not\in L_p \mathbb{Z}^4\) for \(i \ne j\), lying exactly on some common lattice \(x = a_r z\), \(z\) integer, the cutoff-\(f (n)\) correlations at coordinatewise ceiling addresses of \(x\) are Cauchy. Status: conditional theorem. A run with campaign, pillars, fixed-level convergence and uniform perturbation estimates gives \(C_n = C\), \(f (n) = n\). Lean: Rising.cinema.

Proposition III.12.8 (positional ultraviolet limit). Let \(C\) be a Wilson family and \(e\) a strictly increasing mesh sequence. Assume the kernel-bound conclusion of III.12.1 and that ceiling-address correlations are Cauchy along \(e\) whenever \(x_{i \mu} - x_{j \mu} \not\in L_p \mathbb{Z}\) for every \(i \ne j\) and axis \(\mu\). For every \(p, n\) and label list there is a function \(f\) on \((\mathbb{R}^4)^n\), smooth on \(\text{NoHit}_p = \{x : x_i - x_j \not\in L_p \mathbb{Z}^4 (i \ne j)\}\), with these properties:

Status: conditional theorem; the run hypotheses following III.12.7 give the full sequence \(e (s) = s\). Lean: positioned; full-sequence instance Campaign.realUV.

2. What it uses

The section hypotheses require the prescribed polynomial realization and finite regularity bounds; dressings commute with signed lattice symmetries and conjugation. Dressed equation labels have zero Wilson expectation against every bounded measurable spectator depending only on links beyond a fixed lattice distance, for every nonzero inverse coupling and sufficiently large torus; spectators need not be gauge invariant.

Full section-comparison data include terminal-coordinate automorphisms \(A_{d ; N, p}\), auxiliary columns and sub-power \(E_d (N)\). Eventually, uniformly in \(p\), \(A_{d ; N, p}\) is the symmetry-averaged column map and its inverse increases coefficient size by at most \(2\). For each column and its signed or conjugated versions, the centred root of the auxiliary-minus-chosen section evaluation is at most \(C_d (m, k) E_d (N) a_N^{1 / 8}\) for every admissible \(k < N\) and \(m \leqslant N\); scale selection chooses \(k\) for each large \(m\). The kernel is the image of the Maxwell equation ideal under dressing, threshold rescaling, terminal map and \(A_{d ; N, p}\). Projection symmetries and finite-cutoff equation and difference comparisons are also required.

Detector rows use test configurations with distinct projections on every axis in a box of side \(a_{\text{width} (k)}\), and lifted partner labels bounded uniformly in \(N, p\) and finite indices for fixed \(k\). Their short-distance cumulant comparisons approach the prescribed free Maxwell tests.

3. Strategy

Normalized differences introduce \(a_N^{- 1}\), beyond the centred budget for bounded labels. Local matching replaces each differenced observable by a primary with sub-power inverse terminal coordinates and vanishing root error. Averaged partner correlations bound its projected physical label through an injective free Maxwell detector; equation labels contribute a vanishing root. This funds differences, and chessboard estimates bound their separated products.

At fixed volume, compare full correlations with scalar-normalized correlations at an intermediate level. Fixed-level convergence makes the latter Cauchy, while an exponentially small tail in the budget radius tends to zero. Positional bounds extend convergence from common-lattice configurations; Taylor estimates identify derivatives. For coupling dependence, polynomial coefficient bounds and smaller parameter intervals turn level differences into summable coefficient errors.

4. Steps

  1. Match differences. For a run with simple \(G\) and an eventually bounded degree-\(d\) label sequence \(w\), normalized differences preserve regularity with sub-power bounds after increasing degree. Local matching gives a degree-\((d + |l|)\) coin \(x\) with \(O (x) - \Delta_l O (w)\) owed at \(a_N^{1 / 16}\). prepared_rows.

  2. Remove equation labels. Under III.12.5's hypotheses, auxiliary equation sections have zero root: their exact distant-spectator identity annihilates the reflected product. Section comparison leaves cost \(a_N^{1 / 32}\). banked.

  3. Detect projected labels. Under III.12.3's hypotheses, averaging differences telescopes and bounds tests by \((2 / b)^{|l|}\) times undifferenced correlations. Short-distance comparison approaches the injective free Maxwell matrix; finite-dimensional stability and the rank bound give III.12.4. trained.

  4. Close the budget. Under III.12.3's hypotheses, set \(X = \pi x\). Apply negligible-equation control to \((x - X) / (1 +\|X\|)\), obtaining \(\|X\| \leqslant Zc + \tfrac{1}{2} (1 +\|X\|)\). Thus \(X\) is bounded; the centred budget funds \(O (X)\), and vanishing errors fund the desired difference. cap, improved.

  5. Bound kernels. Once all fixed-label differences are funded, place separated insertions in distinct cells with interior margins; offset comparison and the chessboard inequality bound products. Slot differences act on their insertions, giving III.12.1. funding.

  6. Control coupling series. Under III.12.6's campaign hypotheses, composed transfer bounds give coefficient sums polynomial in \(a_{j + 1}^{- 1}\). With fixed-level convergence and perturbation inputs, interval reduction converts real \(O (a_j)\) level differences to summable \(C \sqrt{a_j}\) coefficient bounds. For fixed \(p\), labels, compact \(K \subset \text{NoHit}_p\) and \(q \in \mathbb{N}\), this gives a positive parameter width independent of cutoff and a neighbourhood of \(K\) with limits \(C^q\) in position and analytic in the complex parameter; kernels with spatial stencils of length at most \(q\), their interpolations and admissible smeared sums have analytic continuations whose parameter derivatives of every fixed order converging uniformly on compact subdiscs and compact spatial subsets. Fan.standing.

  7. Take the mesh limit. Under III.12.7's hypotheses, fixed-level effective insertions and normalized densities are uniformly Cauchy; Haar integration and positive partition-function bounds give Cauchy correlations. For large fixed \(j\), the full correlation differs by at most \(2 F_0 e^{- B_j / 64}\), uniformly in sufficiently large cutoffs, with \(F_0\) independent of \(j\) and cutoff and \(B_j \to \infty\). Sever.echoes, Rising.arrives.

  8. Recover spatial derivatives. Under III.12.8's hypotheses, equicontinuity extends dense-set convergence. Taylor comparison with differences over a fixed physical interval identifies successive derivative limits; stencil averaging, multilinear interpolation and dominated convergence give the remaining conclusions. positioned.

5. Where the difficulty lies

Matching controls inverse terminal coordinates; rank detection must bound physical labels while section comparison controls the projection kernel. Bounds are volume-uniform, but convergence fixes \(p\) first. Every derivative order is fixed before choosing the cutoff threshold; analytic widths may depend on volume, compact spatial set and derivative order. The additive action scalar need not converge. Collision extension belongs to VII.4.

6. Notation

7. What remains unexplained

The local matching estimate, construction of auxiliary equation sections and relation to lattice Schwinger–Dyson identities remain imported or unverified here. One-step transfer capacity, intermediate weighted analytic estimates, spatial derivative identification for analytic limits, and multiscale remainder and omission bounds need their source analyses.

Detailed notes: the full account of this unit, with proof sketches for every step, is on the detailed notes page.

Code dependencies

Code dependencies (from the compiled Lean). These lists follow references between declarations allocated to units; they can include shared definitions, record fields and alternative constructions. They are not a mathematical reading order. Only unit pairs with at least 25 references are shown; see the counting rules.

References to: III.1, III.10, I.1, VIII.1, II.1, III.2, III.16, III.11, III.9, II.2, III.5, IV.1, III.15, III.6, III.7, V.2, III.4.

Referenced by: VIII.1, III.2, III.14.

Lean source

Source allocation: 18,973 attributed lines; 3,138 line-counted declarations; 4,449 assigned declaration names. Of the assigned names, 4,315 are in the compiled closure of the seventeen headline results. These quantities have different scopes.

The folders and result chains below are source inventories. Their line counts cover each whole group, including any share allocated to another unit; percentages use attributed lines plus one per assigned name. They need not sum to the unit total.

Additional source locations

Further pinned locations for the definitions and comparisons discussed in this unit.

Key statements

In the selected statement order. K- identifiers name this generated list; proposition and lemma numbers in the outline are separate. Each sentence is an automatically drafted summary of a Lean statement; follow its link for the exact hypotheses and conclusion. The outline status does not certify these summaries. Main result and key result describe importance, not correctness. Full names are in the statement table.

All units · Overview of the proof · Repository

← III.11 Interacting rank, Gaussian comparison and the stress tensorPart III·Detailed notesIII.13 Alternative construction: one Wilson family for all labels →