Home·The theorem·The proof in one reading·Architecture·Units·Manuscript·Glossary·FAQ·Work items·Data·Method·Provenance·Changes
← III.7 Matching the Wilson energy to fourth orderPart III·Detailed notesIII.9 The marginal flow: the running coupling and its reset →

III.8. Connected activities and backward dependence cones

Contents

1. What this unit proves 2

Part A 2

Part B 3

2. What it uses 4

3. Strategy 4

4. Steps 5

5. Where the difficulty lies 5

6. Notation 5

7. What remains unexplained 5

Code dependencies 6

Lean source 6

Additional source locations 7

Key statements 12

Part III — Ultraviolet construction (renormalization group).

Explains connected activity expansions and the regions of earlier fields on which an observable depends.

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.7 · Next: III.9 →

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: YM/RG08/Steam/, YM/RG08/Forms/, YM/RG08/Skiff/, YM/RG08/Jetty/, YM/Push/GRID_APPROXIMATION/. Size: 4,147 lines in the source allocation.

1. What this unit proves

Part A

Let \(I\) be any set, \(d\) a pseudometric, \(\sigma \geqslant 0\). Rank-\(n\) tensors are real functions on \(I^n\) with ordered slots; \(\text{diam}_d (x)\) is the largest pairwise distance, zero for an empty tuple. Use

\(\displaystyle \|F\|_{\sigma} = \max_i \sup_z \sum_{x : x_i = z} |F (x) | e^{\sigma \text{diam}_d (x)}, \qquad [\Gamma]_{\sigma} = \max \left( \sup_u \sum_v | \Gamma (u, v) |e^{\sigma d (u, v)}, \sup_v \sum_u | \Gamma (u, v) |e^{\sigma d (u, v)} \right) .\)

Empty maxima and suprema are zero, including rank-zero tensor norms. Propagators need not be symmetric. A pair is \((B, \delta)\)-controlled if both norms are at most \(B\) and its difference norm is at most \(B \delta\). A polynomial majorant satisfies \(0 \leqslant P (x) \leqslant Cx^q\) for \(x \geqslant 1\), for some \(C \geqslant 0\), \(q \in \mathbb{N}\).

A formal potential \(V = (V_{e, n})\) is eligible if \(V_{0, n} = 0\) and nonzero coefficients satisfy \(n \leqslant e + 2\), \(e + n\) even. Let \(\Delta_{\Gamma}\) and \(\star_{\Gamma}\) contract and join slots, averaging over output permutations:

\(\displaystyle \Delta_{\Gamma} H (x) = \text{Sym} \sum_{u, v} \Gamma (u, v) H (u, v, x), \qquad (H \star_{\Gamma} G) (x, y) = \text{Sym} \sum_{u, v} H (u, x) \Gamma (u, v) G (v, y) .\)

Define \(V^{(0)} = V\) and

\(\displaystyle V^{(j + 1)}_{e, n} = \frac{1}{j + 1} \left[ \binom{n + 2}{2} \Delta_{\Gamma} V^{(j)}_{e, n + 2} + \sum_{c = 0}^n \sum_{i = 0}^j \sum_{a = 0}^e \frac{(c + 1) (n - c + 1)}{2} V^{(i)}_{a, c + 1} \star_{\Gamma} V^{(j - i)}_{e - a, n - c + 1} \right] .\)

Proposition III.8.1. Fix \(j, e, n \in \mathbb{N}\). Put

\(\displaystyle T_j (e, n) = \{(a, r) : 1 \leqslant a < e, 1 \leqslant r \leqslant a + 2\} \cup \{(e, r) : 1 \leqslant e, n < r \leqslant e + 2\},\)

adding \((e, n)\) if \(j = 0\), \(e \geqslant 1\), \(1 \leqslant n \leqslant e + 2\). There is a polynomial majorant \(P\), depending only on \((j, e, n)\), such that for all \(W \geqslant 1\), \(\delta \geqslant 0\), eligible \(V, V'\) and propagators \(\Gamma, \Gamma'\), control \((W, \delta)\) of the propagators and coefficient pairs in \(T_j (e, n)\) implies control \((P (W), \delta)\) of the output pair \(V^{(j)}_{e, n}\). The majorant is uniform in \(I, d, \sigma, \delta\). At \(n = 0\) this is vacuous. Status: theorem. Lean: navigation.

Proposition III.8.2. For every eligible \(V\) and arbitrary \(\Gamma\), all \(V^{(j)}\) are eligible, nonzero coefficients satisfy \(n + 2 j \leqslant 3 e\), and positive-\(j\) outputs are symmetric. For every real \(t\),

\(\displaystyle \mathcal{E}_t (V)_{e, n} = \sum_{j = 0}^{3 e} t^j V^{(j)}_{e, n} = V_{e, n} +\mathcal{R}_t (V)_{e, n} .\)

Here \(\mathcal{R}_t\) sums \(j \geqslant 1\), and \(\mathcal{E}_t (V)\) is eligible. The series vanishes for \(n > 3 e\); otherwise its degree is at most \(\lfloor (3 e - n) / 2 \rfloor\). Status: theorem. Lean: preserves, live, coal.

Proposition III.8.3. Let \(\mathcal{S}\) consist of smooth real functions on \(\mathbb{R}^4\) with polynomially growing derivatives, with

\(\displaystyle \|f\|_{k, \sigma} = \max_{| \alpha | \leqslant k} \sup_y | \partial^{\alpha} f (y) | e^{- \sigma |y|} .\)

For marked tensors valued in real-linear functionals on \(\mathcal{S}\), use \(\| \lambda \|^{\ast}_{k, \sigma} = \sup_{\|f\|_{k, \sigma} \leqslant 1} | \lambda (f) |\) and \(\|M\|_h = \sum_x \|M (x)\|^{\ast}_{k, \sigma} e^{\sigma h (x)}\), where \(h : I \to [0, \infty)\) is 1-Lipschitz and \(h (x) = \max (0, \max_i h (x_i))\). Set \(M^{(0)} = F\); the recursion replaces the contraction input by \(M\) and the join by \(M^{(i)} \star_{\Gamma} V^{(j - i)}\), doubling its coefficient.

For each \(j, e, n \in \mathbb{N}\) there is a polynomial majorant \(P\), depending only on these indices, such that for all \(k \in \mathbb{N}\), \(\sigma > 0\), \(W \geqslant 1\), \(U, \delta \geqslant 0\): if \(V, V'\) are eligible, propagators and unmarked inputs \(1 \leqslant a \leqslant e\), \(1 \leqslant r \leqslant a + 2\) are \((W, \delta)\)-controlled, and marked inputs with \(a < e\), or \(a = e, r > n\), or \(j = 0, (a, r) = (e, n)\) are \((U, \delta)\)-controlled, the marked output pair is \((P (W) U, \delta)\)-controlled and real-linear.

For every \(l \in \mathbb{N}\), eligible \(V\) and \(F\) supported on \(r \leqslant l + a\) give nonzero outputs only if \(n + 2 j \leqslant l + 3 e\) and \(n \leqslant l + e\). For fixed \(l, e, n\), there is a polynomial majorant \(P\) depending only on these indices such that the paired sums \(\sum_{j = 1}^{l + 3 e} t^j M^{(j)}_{e, n}\) are \((P (W) U, \delta)\)-controlled for \(|t| \leqslant W\), under the preceding bounds omitting the marked input \((e, n)\). They comprise the whole positive-time part if both marked inputs satisfy that support condition. Status: theorem. Lean: journey, registry, full.

Part B

Use I.1's gauge group and Haar probability measures. A block step has odd \(L \geqslant 3\), \(J \geqslant 4\), patch radius \(\rho = (J + 3) L\), embedding \(\iota\), and \(\mathcal{B}F (U) = \int k_{\beta} (U, V) F (V) \hspace{0.17em} dV\). Locality on \(X\) means dependence only on links with both endpoints in \(X\); \(\mathcal{I} (X)\) adds gauge invariance, measurability and boundedness.

The patch hypothesis requires unique minimising frames and best predictors whenever their branches are used, and nonempty fallback tags for every cover configuration.

Proposition III.8.4. Assume the patch hypothesis, \(\beta \in \mathbb{R}\). For bounded measurable complex \(F\) local on \(X\), \(\mathcal{B}F\) is local on any \(Y \supseteq \bigcup_{x \in X} \Lambda_{10 \rho} (\iota x)\). If also \(k_{\beta} (U^g, V^{g \circ \iota}) = k_{\beta} (U, V)\) for every fine gauge transformation \(g\), then \(F \in \mathcal{I} (\Lambda_R (o))\) implies \(\mathcal{B}F \in \mathcal{I} (\Lambda_{LR + 10 \rho} (\iota o))\), with the same sup bound. Status: theorem. Lean: back_local, back_cylinder.

Histories satisfy the patch hypothesis, gauge covariance and covariance under translations and signed axis permutations at every step. For each \(N, p\), \(\mathcal{B}_m\) uses coupling \(\beta_{N, p, m + 1}\); levels grow finer, \(a_{m + 1} = a_m / L_m\), \(J_{m + 1} \leqslant J_m + 1\), \(\rho_m = (J_{m + 1} + 3) L_m\). Define

\(\displaystyle \mathcal{B}_{k \to m} =\mathcal{B}_{m - 1} \circ \cdots \circ \mathcal{B}_k, \quad r_{k \to k} (r) = r, \quad r_{k \to m + 1} (r) = L_m r_{k \to m} (r) + 10 \rho_m,\)

with \(L_m\) odd, \(L_m \geqslant 3\), \(J_m \geqslant 4\), identity at \(m = k\), \(\tau_m = 30 a_m (J_m + 5)\), and \(\text{room} (h, k, r)\) meaning \(2 (a_k r + \tau_k) < a_h\).

Proposition III.8.5. Under the patch hypothesis, for every \(\beta\), bounded measurable \(F, G\) local on disjoint coarse sets satisfy \(\mathcal{B} (FG) =\mathcal{B}F \hspace{0.17em} \mathcal{B}G\); likewise finite products of \(F_i \in \mathcal{I} (C_i)\) on distinct tiling cells. For every history, \(N, p\), \(h \leqslant k \leqslant m\), finite level-\(h\) set \(l\), and \(F_i \in \mathcal{I} (\Lambda_r (\iota_{h \to k} i))\), \(\text{room} (h, k, r)\) implies

\(\displaystyle \mathcal{B}_{k \to m} \prod_{i \in l} F_i = \prod_{i \in l} \mathcal{B}_{k \to m} F_i .\)

This also holds for tile copies of one \(F \in \mathcal{I} (\Lambda_r (0))\), with level-\(m\) copies on the right. Status: theorem. Lean: back_factors, backTo_copies.

Proposition III.8.6. For every history, \(N, p\), \(k \leqslant m\), pull-back preserves bounded measurability, is linear, fixes constants, commutes with conjugation and decreases the sup norm. For bounded measurable complex \(F\) at level \(k\) and integrable complex \(g\) at level \(m\), the composite push-forward from level \(m\) to \(k\) is integrable and

\(\displaystyle \int (\mathcal{B}^{\ast}_{k \to m} g) F = \int g\mathcal{B}_{k \to m} F, \qquad \int \mathcal{B}^{\ast}_{k \to m} g = \int g.\)

For \(h \leqslant k\) and every level-\(h\) site \(i\), pull-back sends \(\mathcal{I} (\Lambda_r (\iota_{h \to k} i))\) into \(\mathcal{I} (\Lambda_{r_{k \to m} (r)} (\iota_{h \to m} i))\). On bounded measurable functions it commutes with signed axis permutations followed by translation by \(\iota_{h \to \cdot} i\), and centred level-\(h\) copy maps. Status: theorem. Lean: pushTo_dual, backTo_local, backTo_copy.

Proposition III.8.7. For every such history and nonnegative integer levels and radii,

\(\displaystyle \tau_{m + 1} \leqslant \tfrac{2}{3} \tau_m, \qquad \tau_m \leqslant (2 / 3)^m \tau_0 \to 0, \qquad a_m r_{k \to m} (r) + \tau_m \leqslant a_k r + \tau_k \quad (k \leqslant m) .\)

If \(h \leqslant k\) and \(\text{room} (h, k, r)\) holds, every cone box at \(m \geqslant k\) lies in its centred level-\(h\) cell, of physical side \(a_h\). For fixed \(h, r\), room holds for all sufficiently large \(k \geqslant h\). Status: theorem. Lean: scales_budget, room_fits, plenty.

2. What it uses

3. Strategy

For finite \(I\), eligible symmetric \(V\) and symmetric \(\Gamma\), the unmarked series represents \(\log \mathbb{E}_{t \Gamma} e^V\) formally by Wick’s rule. Contractions and joins generate connected coefficients. Diameter and height inequalities make exponential weights multiply; induction and telescoping give polynomial and difference bounds.

Given the fine field, site tags are independent; coarse links are independent conditional on those tags. Unused variables integrate out. Locality iterates through the affine radius recurrence; decay of the physical budget keeps cones within cells.

4. Steps

  1. For finite input norms and positive output rank, \(\| \Delta_{\Gamma} H\|_{\sigma} \leqslant \|H\|_{\sigma} [\Gamma]_{\sigma}\) and \(\|H \star_{\Gamma} G\|_{\sigma} \leqslant \|H\|_{\sigma} [\Gamma]_{\sigma} \|G\|_{\sigma}\). Diameter inequalities absorb internal sums; telescoping gives difference factors \(2, 3\). Held.bridge.

  2. Heat-order induction gives bounds; slot counting gives support. The marked norm also controls scalars. journey.

  3. Frame routes stay within \(9 \rho\); uniqueness makes selected frames local; connectors fit within \(10 \rho\). patch_dep.

  4. Disjoint coarse sets use disjoint tags and bonds, giving factorisation. back_factors.

  5. Physical radius increments are at most \(\tau_m / 3\); combine with budget decay and iterate. scales_next.

  6. For \(h \leqslant k \leqslant N\), bounded measurable \(O\) at level \(N\) and \(F \in \mathcal{I} (\Lambda_w (0))\) at level \(k\) satisfying \(\text{room} (h, k, w)\), expand replicated differences over subsets and apply factorisation and duality; the remaining factors stay inside one push-forward. Joint.restored.

  7. For central-cell observables under IV.1’s chessboard hypotheses, expand differences of products over nonempty subsets and bound each by its chessboard roots. tariff.

5. Where the difficulty lies

Positive heat order uses higher ranks at the same coupling order; the polynomial may depend on the fixed orders. Unmarked scalar bounds are vacuous. Conditional factorisation allows overlapping fine windows without asserting independence under the fine Gibbs law.

6. Notation

\(e, n, j\): coupling order, rank, heat order. \(\Delta_{\Gamma}, \star_{\Gamma}\): lap, bridge. \(\Lambda_r\): periodic \(\ell^{\infty}\) box (window). \(\iota_{h \to m}\): embedded site. \(\mathcal{B}^{\ast}\): iterated push-forward (pushTo). Copies translate, reflect and conjugate on odd tiles.

7. What remains unexplained

III.14 must supply covariance bounds; the relation to polymer activities remains tentative. The marked normalised Gaussian interpretation remains provisional (probable). Non-summable infinite-lattice sums are assigned zero; unmarked scalars require separate convergence arguments; no convergence in \(g\) is asserted. Support packages need separate outlines.

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: I.1, III.1, IV.1, III.16, II.2, II.1, III.6.

Referenced by: III.14, VII.2, III.13, III.4, III.11.

Lean source

Source allocation: 4,147 attributed lines; 871 line-counted declarations; 1,114 assigned declaration names. Of the assigned names, 1,066 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.7 Matching the Wilson energy to fourth orderPart III·Detailed notesIII.9 The marginal flow: the running coupling and its reset →