Home·The theorem·The proof in one reading·Architecture·Units·Manuscript·Glossary·FAQ·Work items·Data·Method·Provenance·Changes
Part I·Detailed notesII.1 Field labels and their lattice representatives →

I.1. Lattice, symmetries, gauge groups and the Wilson measure

Contents

1. What this unit proves 1

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 6

Key statements 15

Part I — Setting.

Sets up the periodic lattices, gauge groups and Wilson probability measure used throughout the construction.

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 · Next: II.1 →

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

Source inventory and conventions

Part: I, Setting. Lean: YM/Lattice (1,215 lines), YM/Gauge (1,131), YM/Space (430), YM/Corona (405), YM/Mink (363), YM/Wilson (157). Size: 3,701 lines in 595 line-counted declarations in the source allocation.

1. What this unit proves

Throughout, \(G \subset U (N)\) is a closed subgroup, \(N \geqslant 1\), \(\ell \geqslant 1\) is integral, and \(\Lambda_{\ell} = (\mathbb{Z}/ \ell \mathbb{Z})^4\). Configurations assign \(U (x, i) \in G\) to links \(x \to x + e_i\); transport multiplies these variables, using inverses on backward steps.

Proposition I.1.1 (torus geometry). Every site has a unique lift \(r_{\ell} (x) \in (- \ell / 2, \ell / 2]^4 \cap \mathbb{Z}^4\), and \(| \Lambda_{\ell} | = \ell^4\). The metric \(|x - y|_{\ell} = \max_i |r_{\ell} (x - y)_i |\) is translation and signed-permutation invariant; \(| [y]_{\ell} |_{\ell} \leqslant |y|_{\infty}\) for every \(y \in \mathbb{Z}^4\). Status: theorem. Lean: lift_eq_iff, separation_triangle.

Proposition I.1.2 (transport and symmetries). On \(\Lambda_{\ell}\) or \(\mathbb{Z}^4\), let \(A = (\rho, b)\), \(Ax = \rho x + b\), with \(\rho = (\pi, \epsilon) \in B_4\) a signed permutation, \((\rho x)_{\pi (i)} = \epsilon_i x_i\), \(b \in \Lambda\). Pullback assigns image-link transport and satisfies \(1^{\ast} = \text{id}\), \(B^{\ast} A^{\ast} = (A \circ B)^{\ast}\) and \((A^{\ast} U) (x ; w) = U (Ax ; \rho w)\). For every \(g : \Lambda \to G\), \(U^g (x, i) = g (x) U (x, i) g (x + e_i)^{- 1}\) gives

\(\displaystyle U^g (x ; w) = g (x) U (x ; w) g (x_{\text{end}})^{- 1} .\)

Transport depends only on links with endpoints visited by \(w\). For \(p = (x ; i < j)\) put \(U_p = U (x ; e_i, e_j, - e_i, - e_j)\). Affine maps injectively send plaquettes to image squares \(A_{\sharp} p\), bijectively on the torus; every conjugation- and inversion-invariant \(\phi : G \to M\) satisfies \(\phi ((A^{\ast} U)_p) = \phi (U_{A_{\sharp} p})\). Gauge invariance means \(F (U^g) = F (U)\) for all \(g\); locality on \(X\) means dependence only on links with both endpoints in \(X\). Periodic extension \(E_b U (y, i) = U ([y]_{\ell} + b, i)\) respects transport, gauge transformations and locality, and \(E_b ((\rho, a)^{\ast} U) = (\rho, 0)^{\ast} E_{\rho b + a} U\). Status: theorem. Lean: travel_pull, travel_gauge, Face.send_price, expandAt_pull.

Proposition I.1.3 (spacings and blocks). For \(m, p \in \mathbb{N}\), odd integers \(L_m \geqslant 3\) and strictly increasing positive integers \(P_p\) divisible by \(4\), set \(D_0 = 1\), \(D_{m + 1} = D_m L_m\), \(a_m = D_m^{- 1}\), \(\ell (m, p) = P_p D_m\), \(s_m = \log (1 / a_m)\). Then \(a_m \downarrow 0\) strictly, \(s_m \to \infty\), and \(a_m \ell (m, p) = P_p\); placements \(a_m r_{\ell (m, p)}\) lie in \((- P_p / 2, P_p / 2]^4\) and are preserved by subdivisions. If \(\log L_m / s_m \to 0\) for \(m \geqslant 1\), some \(c > 0\) satisfies \(a_{m + 1} \geqslant ca_m^2\) for every \(m\).

For every integer \(L \geqslant 1\), \(\ell' = \ell L\), the equivariant injective homomorphism \(\iota x = [Lr_{\ell} (x)]_{\ell'}\) satisfies \(r_{\ell'} \iota = Lr_{\ell}\). The bijection \((x, y) \mapsto \iota x + [r_L (y)]_{\ell'}\) defines parent and residue maps. For odd \(L\), parent is \(B_4\)-equivariant with cubic fibres of radius \((L - 1) / 2\) centred at \(\iota x\). For odd \(L\), \(\text{par} (z + e_i) = \text{par} (z) + e_i\) if \(r_L (\text{res} \hspace{0.17em} z)_i = (L - 1) / 2\), and equals \(\text{par} (z)\) otherwise. For all fine sites \(z, z'\),

\(\displaystyle |z - \iota \text{par} \hspace{0.17em} z|_{\ell'} \leqslant L / 2, \qquad L | \text{par} \hspace{0.17em} z - \text{par} \hspace{0.17em} z' |_{\ell} \leqslant |z - z' |_{\ell'} + L.\)

Status: schedule definition, chosen in III.2; consequences theorems. Lean: Schedule, mesh_inequality, Ratio.assemble_bijective, Corona.foreshorten.

Proposition I.1.4 (cover numbers). Let \(\text{halo}_M (D) = \{y : \exists x \in D, |x - y|_{\ell} \leqslant M\}\), and let \(\text{cov}_M (X)\) minimize \(|D|\) subject to \(X \subset \text{halo}_M (D)\) and halo connectivity by steps of sup-distance at most \(1\); empty sets are allowed. For all sets and \(M \in \mathbb{N}\), cover numbers are monotone and subadditive for intersecting sets,

\(\displaystyle |X| \leqslant (2 M + 1)^4 \text{cov}_M (X), \qquad |x - y|_{\ell} \leqslant (4 M + 2) \text{cov}_M (X) + 2 M \quad (x, y \in X) .\)

A halo is connected exactly when its centres are chained by steps at most \(2 M + 1\), and \(| \text{halo}_M (D) | \leqslant (2 M + 1)^4 |D|\). For \(b, k, m \in \mathbb{N}\), \(b \geqslant 1\), \(\text{cov}_b (\text{halo}_{kb} \{x\}) \leqslant (2 k + 1)^4\); a scale-\(mb\) cover \(D\) gives \(\text{cov}_b (X) \leqslant (2 m + 3)^4 |D|\). For subdivision ratio \(L \geqslant 1\), \(X \subset \Lambda_{\ell L}\), \(M, B, s \in \mathbb{N}\), \(B \geqslant 1\), \(M \leqslant 2 B\),

\(\displaystyle L \text{cov}_B (\text{halo}_{sB} (\text{par} \hspace{0.17em} X)) \leqslant (L + 10 \text{cov}_M (X)) (2 s + 13)^4 .\)

Status: theorem. Lean: coverNumber, Corona.collapse.

Proposition I.1.5 (cost and Haar measure). \(G\) is compact and second countable. Put \(\sigma (u) = 1 - N^{- 1} \text{Re} \hspace{0.17em} \text{tr} \hspace{0.17em} u\), \(\chi (u) = \|u - 1\|_F / \sqrt{N}\). Both are continuous, conjugation- and inversion-invariant, vanish at \(1\), and satisfy \(\chi^2 = 2 \sigma\), \(0 \leqslant \sigma, \chi \leqslant 2\), \(\chi (uv) \leqslant \chi (u) + \chi (v)\). Haar probability \(\mu_G\) is invariant under inversion and both translations. Status: theorem. Lean: chord_price, probHaar_rightInv.

Proposition I.1.6 (Lie algebra and Casimir). \(\mathfrak{g}= \{X \in M_N (\mathbb{C}) : e^{tX} \in G \forall t \in \mathbb{R}\}\) is a closed real Lie subalgebra of skew-Hermitian matrices, invariant under conjugation. Every curve \(f : \mathbb{R} \to G\) differentiable at \(0\) with \(f (0) = 1\) has \(f' (0) \in \mathfrak{g}\). The invariant inner product \(- N^{- 1} \text{tr} (XY)\) identifies \(\mathfrak{g}\) with colours \(\mathbb{R}^d\), \(d = \dim \mathfrak{g}\), \(c \mapsto X_c\), with \(\|X_c \|_F = \sqrt{N} |c|\). Projection satisfies \(X_{\Pi A} = P_{\mathfrak{g}} A\), \(| \Pi A| \leqslant \|A\|_F / \sqrt{N}\); conjugation induces orthogonal \(\text{Ad}\), and \(\text{ad} (x) y\) represents \([X_x, X_y]\).

If \(\mathfrak{g}\) is nonabelian with no nonzero proper ideals, every symmetric real bilinear form satisfying \(B (\text{ad} (x) y, z) = - B (y, \text{ad} (x) z)\) for all \(x, y, z\) is a scalar multiple of the inner product. Writing \(C_{\mathfrak{g}} = - d^{- 1} \sum_i \text{tr} (\text{ad} (b_i)^2)\) in the orthonormal colour basis, for all colours \(x, y\),

\(\displaystyle - \text{tr} (\text{ad} (x) \text{ad} (y)) = C_{\mathfrak{g}} \langle x, y \rangle, \qquad C_{\mathfrak{g}} > 0, \qquad \kappa = 11 C_{\mathfrak{g}} / (24 \pi^2) > 0.\)

Status: theorems; constants definitions. Lean: lie, invariant_scalar, trace_adjoint, casimir_pos.

Proposition I.1.7 (small balls). For every \(0 < r \leqslant 1\), \(\mu_G \{\chi \leqslant r\} \geqslant (r / 4)^{2 N^2}\). Status: theorem. Lean: shell_ge.

Proposition I.1.8 (local group geometry). There exist open neighbourhoods \(V \ni 1\), \(W \ni 0\) and \(\log_G : M_N (\mathbb{C}) \to \mathbb{R}^d\), real-analytic on \(V\), with \(\log_G (1) = 0\), \(\log_G (e^{X_c}) = c\) for \(c \in W\) and \(e^{X_{\log_G u}} = u\) for \(u \in G \cap V\). Every matrix \(X\) has a Frobenius-nearest point \(u \in G\), satisfying \(u^{\ast} X - 1 \in \mathfrak{g}^{\perp}\). Nearest points respect left and right translations by \(G\) and are unique in some tube about \(G\); near \(1\) this retraction is analytic with strict derivative \(P_{\mathfrak{g}}\). For every \(\varepsilon > 0\), some \(\delta > 0\) gives

\(\displaystyle \|u - v - P_{\mathfrak{g}} (X - Y)\|_F \leqslant \varepsilon \|X - Y\|_F\)

whenever \(\|X - 1\|_F, \|Y - 1\|_F < \delta\) and \(u, v\) are nearest to \(X, Y\). Curvature \(F (u) = \Pi ((u - u^{- 1}) / 2)\) is continuous, odd under inversion, satisfies \(F (vuv^{- 1}) = \text{Ad} (v) F (u)\), \(F (1) = 0\), \(|F (u) | \leqslant \chi (u)\), and has derivative \(c\) along \(e^{tX_c}\) at \(0\). Status: theorem. Lean: plog_right, nearest_approximate, leading_curvature.

Proposition I.1.9 (Wilson law). For product Haar \(\mu_{\ell}\) and plaquette holonomies \(U_p\), set

\(\displaystyle E_{\ell} = \sum_{x, i < j} \sigma (U_{(x ; i < j)}), \qquad Z_{\ell} (\beta) = \int e^{- \beta E_{\ell}} d \mu_{\ell}, \qquad d\mathbb{P}_{\ell, \beta} = Z_{\ell} (\beta)^{- 1} e^{- \beta E_{\ell}} d \mu_{\ell} .\)

For every real \(\beta\), \(E_{\ell}\) is continuous, \(0 \leqslant E_{\ell} \leqslant 12 \ell^4\), \(Z_{\ell} (\beta) > 0\), and \(\mathbb{P}_{\ell, \beta}\) is a probability measure integrating continuous or bounded measurable complex functions. Status: definition with theorems. Lean: law, partition_pos.

Proposition I.1.10 (invariance). For every real \(\beta\), affine lattice map \(A\) and \(g : \Lambda_{\ell} \to G\), pullback and gauge transformation preserve \(\mu_{\ell}\), \(E_{\ell}\) and \(\mathbb{P}_{\ell, \beta}\). Status: theorem. Lean: law_pull, law_gauge.

Proposition I.1.11 (continuum symmetries). \(B_4 \to O (4)\) respects coordinate actions; skew-matrix exponentials lie in \(O (4)\). With \(x = (\vec{x}, x_t)\), set \(\textsf{m} = x_t^2 - | \vec{x} |^2\), \(\bar{V}_+ = \{| \vec{x} | \leqslant x_t \}\), \(V_+ = \{| \vec{x} | < x_t \}\). The topological group \(\mathcal{L}= \{g \in GL (4, \mathbb{R}) : \det g = 1, \textsf{m} \circ g = \textsf{m}, g \bar{V}_+ = \bar{V}_+ \}\) preserves Lebesgue measure and \(V_+\) and contains time-fixing elements of \(SO (4)\). For \(R (\vec{x}, x_t) = (- \vec{x}, x_t)\), \(g^{\vee} = RgR = (g^{- 1})^T\) satisfies \(\langle gx, g^{\vee} p \rangle = \langle x, p \rangle\). For every unit spatial vector \(e\), the boost \(B_e (r)\) is a continuous one-parameter subgroup of \(\mathcal{L}\), with generator \((\vec{x}, x_t) \mapsto (x_t e, \vec{x} \cdot e)\). The closed cone is self-dual and equals \(\{x : (g^{\vee} x)_t \geqslant 0 \forall g \in \mathcal{L}\}\). For \(x \in V_+\) and \(p \in \bar{V}_+ \setminus \{0\}\), \(\langle x, p \rangle > 0\). The Poincaré group \(\mathcal{L} \ltimes \mathbb{R}^4\), acting by \(x \mapsto \Lambda x + b\), preserves Lebesgue measure diagonally on every \((\mathbb{R}^4)^n\). Distinct-point configurations form an open set; pairwise distinct coordinates on every axis, or positive pairwise sup-separation, imply distinct points. Status: definitions and theorems. Lean: Mink.Proper, Lorentz.from_half, Poincare.

Proposition I.1.12 (cover boxes and routes). For every \(r \in \mathbb{N}\), \(B_r = \{- r, \ldots, r\}^4\) is \(B_4\)-invariant with \((2 r + 1)^4\) points. For cover sites \(x, y\), axis routes from \(x\) to \(y\) end at \(y\), give the reverse route from \(y\) to \(x\) when axis order is reversed, and transform covariantly under affine maps. For \(a \in \mathbb{Z}^4\) and an axis set \(J\), slabs \(\{y \in B_r : y_i = a_i (i \not\in J)\}\) decompose into pinned layers when an axis \(i \not\in J\) is freed; their sums decompose accordingly for every function into a commutative additive monoid. Pinned layers \(z, w \in \{- r, \ldots, r\}\) correspond by translation \((w - z) e_i\). For every cover configuration and integer \(d\), \(U (x ; \text{straight}_j (d + 1)) = U (x ; \text{straight}_j (d)) U (x + de_j, j)\), where \(\text{straight}_j (d)\) has displacement \(de_j\). Status: theorem. Lean: route_end, slab_sum.

2. What it uses

3. Strategy

The spacing schedule separates physical periods from mesh sizes; centred residues define the metric. Connected covers turn set size into path length, allowing coarsening by sparse sampling. Image-link transport reduces Wilson invariance to Haar invariance and conjugation/inversion invariance of the cost.

The exponential definition of the Lie algebra supplies tangent directions. Normal complements and the inverse function theorem give local coordinates and nearest points; simplicity forces invariant symmetric forms to be scalar.

4. Steps

  1. Centred division gives the metric and subdivision; odd residue windows give parent equivariance. Ratio.parent_signed.

  2. Traverse a fine cover by a path, sample parents every \(LB\) steps, then tile the resulting coarse halo; this proves the coarsening bound. Corona.collapse.

  3. Walk induction gives transport covariance; gauge factors telescope. Reversal and change of starting corner give plaquette inversion and conjugation. travel_gauge, Face.send_price.

  4. For the curve of I.1.6, approximate \(e^X\) by \(f (1 / n)^n\) to prove the tangent criterion. Isolation of normal exponentials gives the local logarithm. tangent_generates, isolates.

  5. Minimize distance and use normality with local coordinates to identify nearest points. The symmetric operator associated with an invariant form has ideal eigenspaces; simplicity gives the Casimir identity. nearest_is_focus, invariant_scalar.

  6. Pack ambient matrix balls for I.1.7; normalize the positive Wilson density and use link permutations, inversions and two-sided translations for invariance. shell_ge, law_preserved.

  7. Boosts and Cauchy–Schwarz establish the cone characterizations. cone_test, Lorentz.from_half.

5. Where the difficulty lies

Local group geometry must handle disconnected groups. Coarsening is uniform in torus size. Real simplicity suffices without irreducibility of the defining representation. Reflections shift link base points; even-side lift ties prevent unrestricted placement covariance.

6. Notation

\([y]_{\ell}\): reduction; \(r_{\ell}\): centred lift; \(\text{par}\): block parent; \(\| \cdot \|_F\): Frobenius norm; \(P_{\mathfrak{g}}\): real Frobenius orthogonal projection; \(\text{ad}\): Lie bracket in colour coordinates. Time is coordinate \(3\).

7. What remains unexplained

Helpers outside the six principal folders remain uncovered. The condition \(M \leqslant 2 B\) for the consecutive cover scales in III.5 and III.10 remains to be traced. Exact bibliographic theorem locations remain unchecked.

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: none above the threshold.

Referenced by: III.10, III.11, III.4, VII.2, III.7, III.13, III.12, III.6, III.2, II.1, III.9, III.14, VI.2, III.5, VII.4, III.1, VII.1, VIII.1, IV.1, VI.1, III.3, III.15, V.2, VII.3, VII.6, III.8, V.1, V.3, III.16, II.2, VII.5.

Lean source

Source allocation: 3,701 attributed lines; 595 line-counted declarations; 1,136 assigned declaration names. Of the assigned names, 949 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

Part I·Detailed notesII.1 Field labels and their lattice representatives →