1. What this unit proves 2
2. What it uses 4
3. Strategy 4
4. Steps 4
5. Where the difficulty lies 5
6. Notation 5
7. What remains unexplained 5
Code dependencies 5
Lean source 5
Additional source locations 6
Key statements 9
Part VII — Identification as Yang–Mills.
Relates the running coupling to its leading logarithmic asymptotic under the construction hypotheses.
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: VII.4 · Next: VII.6 →
Results · Inputs · Strategy · Proof steps · Notation · What remains unexplained · Code dependencies · Lean source · Key statements
Reading scope. The coupling result is in YM/Pert/Asymptotics; the allocation also includes auxiliary algebra and limit lemmas from other folders, as explained below.
Part: VII, identification as Yang–Mills.
Lean: YM/Pert/Asymptotics,
YM/Extra/WEAK_SCALING_ACTLAW, YM/TraceExtension/Foundation, YM/TraceExtension/Soot.
Size: 232 lines, 43 declarations.
Source inventory and conventions
For VII.5.1–3, let \(G \subset U (q)\) be compact, \(q \geqslant 1\). Write \(2 b_0 = \kappa_G = 11 C_{\mathfrak{g}} / (24 \pi^2)\), \(a_m = (\prod_{k < m} b_k)^{- 1}\), \(s_m = \log (1 / a_m)\), with odd \(b_k \geqslant 3\). The trajectory has inverse couplings \(\beta_{N, p, m}\), bare endpoint \(\beta_{N, p} = \beta_{N, p, N}\), reference value \(\beta_{N, p, 0} = \bar{\beta} > 0\), and \(g_{N, p, m} = \beta_{N, p, m}^{- 1 / 2}\). Simple means real simple and nonabelian. The running clause says that for every \(\varepsilon > 0\), eventually in \(m\),
uniformly for \(N \geqslant m\) and \(p \in \mathbb{N}\).
Proposition VII.5.1. If the family satisfies the construction conditions (II.2), then for every \(\varepsilon > 0\) there is \(m_0 \geqslant 1\) such that
Status: theorem; only the running clause is used, without simplicity. Lean: YM.Pert.Asymptotics.zeroUp_bound.
Corollary VII.5.2. Under the construction conditions and simplicity of \(\mathfrak{g}\), the Wilson family satisfies (Reg1): for every \(\varepsilon > 0\) there is \(N_0 \geqslant 1\) such that \(| \beta_{N, p} / s_N - b_{\ast} | < \varepsilon\) for all \(N \geqslant N_0\) and \(p \in \mathbb{N}\), where \(b_{\ast} = 2 b_0 > 0\). Status: theorem in VIII.1. Lean: YangMills.regular.
Proposition VII.5.3. Under the same hypotheses, for any lattice-level and volume sequences \((N_i), (k_i)\) with \(i \leqslant N_i\) eventually,
Status: theorem; the second limit follows by squaring. Lean: YM.Pert.Asymptotics.comparison.
For VII.5.4–5, \(\mathcal{F}_d, \mathcal{F}_d^{\circ}\) are finite-dimensional complex label spaces of degree at most \(d\) in Euclidean hierarchies on \(S\) and its free Maxwell comparison table, on the same reference labels. Derivative plans have inclusions \(\iota_{d \to e}\) and derivatives \(\partial_{d, a}\). Exact compatibility means preserving inclusions, signed permutations \(\Gamma_4\), conjugation and units.
Proposition VII.5.4. Fix the free Maxwell data of VI.2, including the finite separating tests and boundary, polynomial and internal-rotation conditions. Its plan \(W\) satisfies the inclusion, derivative and orientation axioms; \(J^{\circ}\) satisfies Ward identities with translation part \(\kappa \partial^W\) and rotation part \(\kappa\) times the infinitesimal rotation action, where \(\kappa \ne 0\) and \(\partial^W_4 \ne 0\). Let \(V\) satisfy the inclusion and derivative axioms. Suppose \(J \in \mathcal{F}_4^{4 \times 4}\) is symmetric, real, \(\Gamma_4\)-covariant and satisfies \(\sum_a \partial^V_{4, a} J_{ab} = 0\) for every \(b\), and \((J, p, Y)\) satisfies the local Ward identities. If some \(c \in \mathbb{R}\) satisfies \(p_{4, a} = c \partial^V_{4, a}\) for all \(a = 1, \ldots, 4\), there exist \(D, N, M \in \mathbb{N}\), \(R_0 \geqslant 1\) such that \(c / \xi_k \to \kappa\) for every tracking scaling sequence at every level \((n, d, j, R) \geqslant (N, D, M, R_0)\) componentwise (VI.1). Also, for any hierarchy and local Ward identities, \(p \ne 0\) implies \(J \ne 0\). A scaling sequence has complex-linear isomorphisms \(U_k (d) : \mathcal{F}_d^{\circ} \to \mathcal{F}_d\), positive strictly decreasing \(r_k \to 0\) and \(\xi_k > 0\). Tracking \((V, W, J)\) at \((n, d, j, R)\) requires exact compatibility through \(\max (d, 4)\), convergence of transported correlations at \(r_k x\) for each fixed list of at most \(n\) fields of degrees at most \(d\), to the free correlations, with all Fréchet derivatives through order \(j\), uniformly on \(\{x : \|x\| \leqslant R, \|x_a - x_b \|_{\infty} \geqslant R^{- 1} (a \ne b)\}\), and \(\xi_k^{- 1} r_k^4 U_k (4)^{- 1} J_{ab} \to J^{\circ}_{ab}\).
Status: theorems. Lean: YM.Extra.WeakScaling.tracks_ratio, flux_nontrivial.
Proposition VII.5.5. Let \(V, W\) satisfy the inclusion and orientation axioms. For every \(d\) there is a complex-linear, \(\Gamma_4\)-equivariant, conjugation-preserving retraction \(P_d\) of \(\iota^W_d := \iota^W_{d \to d + 1}\). For any complex-linear isomorphisms \(A_d : \mathcal{F}_d^{\circ} \to \mathcal{F}_d\), define
Fix \(D \in \mathbb{N}\). If \(A_d\) preserve symmetries and units for \(d \leqslant D\), then \(X_d\) are exactly compatible through degree \(D\). For sequences of such isomorphisms \(A_d^{(k)}\) satisfying \((A_{d + 1}^{(k)})^{- 1} \iota^V_d A_d^{(k)} v \to \iota^W_d v\) for every \(d < D, v\), one has \((A_d^{(k)})^{- 1} X_d^{(k)} \to 1\) in operator norm for every \(d \leqslant D\). These \(X_d^{(k)}\) are eventually bijective; replacing nonbijective terms by \(A_d^{(k)}\) gives isomorphisms \(\tilde{X}_d^{(k)}\) with \((A_d^{(k)})^{- 1} \tilde{X}_d^{(k)} v \to v\). If the original frames preserve symmetries and units through \(D\), the replacements are eventually exactly compatible simultaneously through \(D\). Status: theorem. Lean: retraction, adjusted_exact, perturb_limit, relative.
Proposition VII.5.6. Along any filter: (a) for an injective linear map \(f : E \to X\) between finite-dimensional real normed spaces, \(f (a_y) \to f (e)\) implies \(a_y \to e\); (b) continuous real-linear maps \(f_y, g : (\mathbb{R}^4)^n \to \mathbb{C}\) satisfy \(f_y \to g\) in operator norm if \(f_y (e_{i, a}) \to g (e_{i, a})\) at all \(4 n\) coordinate vectors. Status: theorem. Lean: norm_fit, norm_fit_deriv.
Definition VII.5.7. In gauge-invariant radial-jet polynomials, with curvature weight 2 and derivatives adding 1, for \(t \in \mathbb{R}^4\) set
Here \(F, \partial F\) are linear curls of degree-0 and degree-1 radial Taylor coefficients. These are polynomial identities before quotienting by null fields. Put \(T^{\text{can}} = T^{(0, 1, - 1 / 4, 1)}\), \(\mu^1_b = \sum_i \langle F_{bi}, \partial_b F_{bi} \rangle\), \(\mu^2_b = \sum_{i, j} \langle F_{ij}, \partial_b F_{ij} \rangle\), \(E_b = \sum_{a, i} \langle F_{bi}, \partial_a F_{ai} \rangle\), and, for \(w \in \mathbb{R}^3\),
Both \(\mu^1, \mu^2\) belong to the real space \(\mathcal{R}_1\) of real, \(\Gamma_4\)-covariant vectors of degree at most 5. Every degree-0 radial jet is realised by a constant-curvature free Maxwell field with affine potential. Status: definitions and two lemmas. Lean: raw, basisV, bareVector, shell_affine.
Definition VII.5.8. Fix a family, nondecreasing \(K : \mathbb{N} \to \mathbb{R}\) with \(K \geqslant 1\), degree \(d\), stages \(\ell_d \leqslant k \leqslant N\), volume \(p\), and \(u \in \mathbb{R}\). The sharp read-out bound asserts \(u \|c_k \|_1 \leqslant 4 (1 + n_k) w\) for every \(w \geqslant 0\) and every chain of insertion presentations \(A_i\) at the origin, with coefficient vectors \(c_i\), remainder costs \(q_i\) and densities \(\mathcal{D}_i\), satisfying the admissibility conditions of III.12 (regular and forced remainders, with square-root forced allowance), \(c_{\ell_d} = 0\), and, for every \(\ell_d \leqslant i < k\),
For each \(\ell_d \leqslant i < k\), costs \(q_i, q_{i + 1}\) are nonnegative and \(B_i\) is a complex-linear automorphism commuting with \(P\), with \(\|B_i^{\pm 1} c\|_1 \leqslant K_i \|c\|_1\) for every \(c\). For \(\ell_d \leqslant i \leqslant k\), the admissible remainder cost is at most \(q_i\), and \(q_k \leqslant w + n_k (w +\|c_k \|_1)\). Status: definition. Lean: YM.MergeProof.Accent.SharpRead.
I.1: Casimir positivity and divergent logarithmic scales (Schedule.divergent); II.2, III.1–2: construction conditions, selected histories and shooting (YM.Family.Makes).
III.7, III.9: one-loop shift and coupling reset.
IV.1, VI.1–2: hierarchies, derivative plans, Ward identities, A04–A05 and free Maxwell data (MaxwellData); III.12: insertion estimates.
Standard results: Cesàro summation, continuity of square root and reciprocal, Maschke averaging, finite-dimensional continuity and openness of invertibility.
The inverse coupling follows a linear reference trajectory in logarithmic scale. Uniform one-step errors sum to \(o (s_m)\) around \(\bar{\beta} + 2 b_0 s_m\). Division by \(s_m \to \infty\) removes the reference coupling. The auxiliary arguments compare two limits proportional to the same nonzero derivative, which determines the stress normalisation. Averaged retractions split each inclusion; correcting frames recursively on the included subspace makes compatibility exact, while the relative correction tends to the identity.
Given \(\beta_{N, p, 0} = \bar{\beta}\) and transfer errors bounded by \(\delta_{m + 1}\) with \(\delta_{m + 1} / \log b_m \leqslant J (m + 1)^{- 2} \to 0\), summation gives the uniform running clause. Split off a finite initial segment and bound the remaining error by a small multiple of \(s_m\) (III.2). trajectory.
Divide by \(s_m\); specialise to \(m = N\) for (Reg1), and use continuity of inverse square root for VII.5.3; positivity uses simplicity. zeroUp_bound, comparison.
A04–A05 give derivative and Ward limits; a linear functional nonzero on \(\partial^W_4\) extracts \(c / \xi_k \to \kappa\). tracks_ratio.
Average retractions over \(\Gamma_4\) and conjugation; the recursive correction preserves inclusions and tends to the original frames in relative operator norm, ensuring eventual invertibility. retraction, perturb_limit.
At the polynomial level, sign cancellation gives covariance of \(\mu^1, \mu^2\); constant curvature realises every degree-0 jet on a Maxwell solution. momentum1_mem, shell_affine.
The analytic work lies upstream in the one-loop estimate, coupling reset and shooting. This unit establishes the leading lattice logarithm; the main theorem retains only a positive slope through (Reg1).
\(C_{\mathfrak{g}}\) is the Casimir for \(\langle X, Y \rangle = - q^{- 1} \text{Re} \hspace{0.17em} \text{tr} (XY)\). For an orthonormal basis \((e_i)\), \(C_{\mathfrak{g}} = - (\dim \mathfrak{g})^{- 1} \sum_i \text{tr}_{\mathfrak{g}} ((\text{ad} \hspace{0.17em} e_i)^2)\), with value 0 if \(\mathfrak{g}= 0\). In VII.5.8, \(\ell_d\) is the family’s reference stage, \(P\) removes the unit coordinate, \(\|c\|_1 = \sum_{\alpha} |c_{\alpha} |\), \((S_i c)_{\alpha} = b_i^{d - d_{\alpha}} c_{\alpha}\), \(K_i = K (J (i + 1))\); \(n_k\) is the column number. Here \(d_{\alpha}\) is label weight and \(J\) the scale-stage index. In VII.5.4, \(\xi_k > 0\) is stress normalisation.
The one-loop integral belongs to III.7. A declaration establishing cutoff removal, infinite-volume limits and comparison with a coupling from continuum correlations has not been identified (tentative). Tracking-sequence existence belongs to VI.1 and VII.1; the read-out bound requires upstream one-step hypotheses.
Detailed notes: the full account of this unit, with proof sketches for every step, is on the detailed notes page.
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: IV.1, I.1, III.16, VI.1, II.1.
Referenced by: VII.2.
Source allocation: 232 attributed lines; 43 line-counted declarations; 43 assigned declaration names. Of the assigned names, 43 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.
work package YM/Extra/WEAK_SCALING_ACTLAW/ — 162 lines.
work package YM/TraceExtension/Foundation/ — 37 lines.
work package YM/Pert/Asymptotics/ — 18 lines.
proof of comparison — 14 lines.
work package YM/TraceExtension/Soot/ — 1 lines.
Further pinned locations for the definitions and comparisons discussed in this unit.
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.
K-VII.5.1 comparison (key result) — The running regulator coefficient times the square root of log inverse lattice spacing tends to 1/sqrt(2 b0), giving the leading-log asymptotics of the running coupling.