Home·The theorem·The proof in one reading·Architecture·Units·Manuscript·Glossary·FAQ·Work items·Data·Method·Provenance·Changes
← III.12 Bounds uniform in the volume and convergence as the spacing tends to zeroPart III·Detailed notesIII.14 The construction run: flat representations and the ultraviolet rate →

III.13. Alternative construction: one Wilson family for all labels

Contents

1. What this unit proves 2

2. What it uses 5

3. Strategy 5

4. Steps 6

5. Where the difficulty lies 6

6. Notation 6

7. What remains unexplained 7

Code dependencies 7

Lean source 7

Additional source locations 9

Key statements 12

Part III — Ultraviolet construction (renormalization group).

Explains an alternative ultraviolet construction that puts all field labels into one Wilson family.

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.12 · Next: III.14 →

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

Reading scope. This is the alternative construction for habitat. The Wilson family used by YangMills.constructed comes from the run in III.14; see VIII.1 for its assembly. The two families are not identified here.

Source inventory and conventions

Part: III, Ultraviolet construction (renormalization group). Lean: YM/SharpGram/Common.lean, YM/SharpGram/Measured.lean, YM/SharpGram/Assembly/, YM/UVExtension/. Size: about 22,100 lines in the unit inventory.

1. What this unit proves

Throughout, \(G\) is a compact matrix group with simple nonabelian real Lie algebra and \(b > 0\) is the Gaussian block-spin penalty. Levels run from the coarsest, \(0\), to the cutoff, \(N\); \(p\) indexes periods. Notation is collected in §6.

Proposition III.13.1 (one ultraviolet and logarithmic family). There exists a Wilson family \(C\), with penalty \(b\), satisfying all the following.

  1. It is calibrated: one full choice of constants and a directed history generate its history, with the per-step estimates of III.9–III.12.

  2. It is consistent: the trajectory is selected, local sections are sound, and, for every degree \(d\), \(d + 3 \leqslant J (\text{ref} (d))\), \(d + 1 \leqslant \text{order}_m\) and \(d \leqslant \text{power}_m\) for \(m \geqslant \text{ref} (d)\). Every degree-\(d\) stencil point satisfies \(20 (|x_i | + 1) \leqslant \text{side} (N, p)\) in each coordinate whenever \(N \geqslant \text{ref} (d)\). Placed basis sections are regular on their footprints, with the prescribed derivative bounds of size \(\text{columns}_N \theta_N^q\) for factor degree \(q\); transfer is calibrated from \(\text{ref} (d)\) onwards.

  3. For every \(d\) and \(R, \varepsilon > 0\), there is \(A \geqslant 0\) such that, for every \(m \geqslant \text{ref} (d)\) and all sufficiently large \(N\), every \(p\) and every \(v \in \mathcal{L}_d\) with \(\|v\| \leqslant R\), the observable at every site is supported in its centred level-\(m\) reflection cell and its chessboard root is at most \(Aa_m^{- d - \varepsilon}\). Here \(N \geqslant \text{ref} (d)\) is required.

  4. It is the centre of the analytic shooting family of III.13.7.

  5. The OPE entrance conditions hold: for every \(d, D, h\) with \(D + h = d + 4\), there exist \(t > 1\) and \(j (i) \to \infty\) with \(\sigma_{j (i)} / \sigma_i \to t\) for which the III.11 shell bounds hold for every choice of labels and extraction order; cover sizes satisfy \(\text{cs}_{i + 1}^2 \leqslant 2 \text{cs}_i^2\) and for every \(\varepsilon > 0\), \((L_i \text{cs}_i^2)^4 (\text{vacuum cost}) \leqslant \varepsilon \sigma_{i + 1}\) eventually; positive radii satisfy \(\log \rho_i / \sigma_i \to - 1\).

  6. There exist coefficient-map bounds \(W\), allowances \(r : \mathbb{N} \to \mathbb{N}\) and transfer data giving the logarithmic OPE package: III.11 transfer/remainder estimates, triangular label steps, auxiliary comparisons, separation bounds, product sup bounds and VII.4 arity bounds. Also \(L_m, \text{columns}_m, \text{columns}_{m + 1} \leqslant \hat{\beta}_m\), \(r \to \infty\), and the support exponent is \(100\).

  7. For these data and every \(d, N, p, m\) with \(\text{ref} (d) \leqslant m < N\), there is a linear isomorphism \(T\) with

    \(\displaystyle \|x_m (v) - T (V_{L_m} x_{m + 1} (v))\|_1 \leqslant \text{barunit} (L_m) \| \text{core} (x_{m + 1} (v))\|_1 \quad (v \in \mathcal{L}_d) .\)

    On a basis vector of engineering grade \(e\) and factor degree \(q\), the only output of grade \(\leqslant e\) is that vector times \((\theta_m / \theta_{m + 1})^q\); no lower factor degree occurs. The map preserves the unit, commutes with core projection and lattice symmetries, and \(\|T\|_1, \|T^{- 1} \|_1 \leqslant K (J (m + 1))\), where \(K\) comes from \(W\).

  8. Coverage holds: some \(g (k) \to \infty\), \(g (k) \leqslant k\), satisfies \(a_{g (k)} \leqslant \hat{\beta}_k a_k\) eventually, with three depth-\(d\) probe panels poised at level \(g (k)\) for every \(d\) and all large \(k\).

  9. Along every extraction order satisfying IV.1's cell hypotheses, grid approximation holds at arbitrarily fine fixed levels for bounded degrees and III.11's allowed reflection pairs. Along each such order, complete weak comparison systems exist for every finite demand and degree horizon, positive radius bound and real-spin structure on the smooth free comparison table.

  10. Every extraction order \(f\) has a refinement \(g\), extraction base tables \(A, (L_k), T\), a bin and derivative plan along \(f \circ g\), and a seed for \((C, f \circ g, T)\) with Maxwell ranks for \(T\).

Status: theorem. Lean: uv_log.

Proposition III.13.2 (coordinated family). There exists one \(C\) of penalty \(b\), consistent, calibrated and floral, with the chessboard bound above, III.13.5's fixed-volume limits, the logarithmic package, calibrated label steps, coverage and complete weak comparison systems along every extraction order satisfying the cell hypotheses. For every extraction order \(F\), there exist a refinement \(g\) and table \(T\) giving the ordered limit along the full mesh sequence and the periods of \(F \circ g\), VII.4's OPE services, and a II.2 construction record for \((C, F \circ g, T)\).

Status: theorem; \(C\) precedes \(F\). Lean: coordination.

Proposition III.13.3 (all phases). For every extraction order \(f\), there exist \(C\), a refinement \(g\), a table \(T\), a construction record \(F\) for \((C, f \circ g, T)\), and a V.3 reconstruction record \(\mathcal{A}\) for every phase of \(F\), satisfying III.13.2 and completed reconstruction: Wightman realization, surjective comparison maps onto the Osterwalder–Schrader Hilbert spaces, unique vacuum, phase-uniform light-cone and heat-kernel bounds, trivial commutant, continued stress tensor, continuous phase dependence of boundary correlations, and barycentre identities. The derivative sector has a prescription \(P\) satisfying VII.4's estimates for every fixed finite cluster tree and separated spectators, with arbitrary positive power loss.

Status: theorem; \(C\) follows \(f\) in the statement, but the proof uses III.13.2's common choice. Lean: common.

Proposition III.13.4 (moving-frame Gram limits). For every extraction order \(f\), there exist \(C, g, T, F, \mathcal{A}, P\) satisfying III.13.3, with coverage at the cutoff of \(F\). For every ultrafilter \(\mathcal{U}\) finer than the filter of \(f \circ g\) (period large, then mesh large) and every depth \(D\):

Status: theorem. Lean: habitat.

Proposition III.13.5 (fixed-volume derivative limits). For \(C\) as in III.13.1 and every \(d, k \in \mathbb{N}\), all sufficiently large \(p\) (depending on \(d, k\)), every \(n\) and labels \(v_i\) of degrees \(\leqslant d\), there exists \(g : (\mathbb{R}^4)^n \to \mathbb{C}\), continuous on \(\Omega_p^n\), such that the following hold along the whole sequence \(N \to \infty\).

Status: theorem. Lean: Poppy.blossoms, Poppy.Margins.

Proposition III.13.6 (separated lattice points). Fix \(p\) and a member \(C^{\mathbf{s}}\) of III.13.7 whose shifts satisfy \(s_{N, p} \to s\), \(|s| < r_{\ast}\). Assume its chessboard bound, slow growth, label-transfer and remainder bounds. For every \(J\), every finite label list \(u_i\) with \(\text{ref} (\deg u_i) \leqslant J\), and pairwise distinct level-\(J\) torus sites \(o_i\), the correlations \(\langle \prod_i O_{u_i}^{N, p} (o_i) \rangle_{N, p}\) converge as \(N \to \infty\). III.13.7 supplies these bounds for every shift assignment; the centre has zero shifts.

Status: theorem. Lean: Hearth.food.

Proposition III.13.7 (analytic shooting family). For \(C\) of III.13.1 there exist a full choice of per-step constants, a shooting structure with an arc, a label-transfer bundle and reference levels such that:

  1. Some \(r_{\ast} > 0\), depending only on those constants, works for every \(N, p\): each real \(|s| < r_{\ast}\) determines exactly one Wilson coupling whose fixed-recursion trajectory stays in the prescribed scalar bands and reaches \(\bar{\beta} + s\) at level \(0\). Assignments \(|s_{N, p} | < r_{\ast}\) define \(C^{\mathbf{s}}\) with the bundle's labels, sections and reference levels, and \(C = C^{\boldsymbol{0}}\).

  2. For \(m \leqslant N\), \(\beta_m (N, p ; s)\) and normalized activity readings extend holomorphically to \(|s| < r_{\ast}\), bounded by \(3 \hat{\beta}_m\) and \(\text{soft} = \text{rim} / 8\), respectively. For fixed \(m, p\), they converge locally uniformly with every shift derivative as \(N \to \infty\).

  3. The centre has the calibration package with constant allowance, priced sup bounds and extraction base data. Every \(C^{\mathbf{s}}\) has label-transfer, remainder and chessboard bounds.

  4. Some common disc \(|z| < a\) lies in the shooting domain and carries holomorphic coefficient and transfer data, including inverse terminal label maps, for all \(d, N, p\) with \(\text{ref} (d) \leqslant N\).

  5. Fix \(d, m\). For all large \(p\), every \(n\) and labels of degree \(\leqslant d\), there exist \(b' > 0\), holomorphic lattice correlations \(W_N (z, x)\) on \(|z| < b'\) and limits \(f (z, x)\). At real \(s\), \(W_N\) equals the correlation whenever \(s_{N, p} = s\) and all reference levels fit. For every complex \(z\), \(f (z, \cdot)\) has continuous iterated coordinate derivatives through order \(m\) on \(\Omega_p^n\); step differences and interpolated differences converge to them with every \(z\) derivative, uniformly on products of separated spatial compacts and disc compacts. Real-shift limits agree with constant-shift fixed-volume limits. For admissible \(t\), \(\int tf (z, \cdot)\) is holomorphic and smeared lattice correlations converge to it locally uniformly with every \(z\) derivative.

Status: theorem, included in III.13.1. Lean: Oasis.Floral.

2. What it uses

3. Strategy

Choose constants in their required order, raising thresholds while preserving earlier estimates. Increasing Taylor orders and powers accommodate each label degree from its reference level onwards. Shooting gives trajectories with prescribed coarse coupling; uniform contraction and decay of dependence on remote ultraviolet levels give full-sequence limits at each fixed level. Uniform holomorphic bounds then permit Vitali's theorem.

Approximate each renormalized observable by a conditional average of a coarse observable with convergent coefficients. Exponential improvement beats subexponential coefficient costs, and chessboard estimates control products. Dense-grid convergence and large-period slicing yield derivative limits. Extraction, OPE and reconstruction results then apply to the same family.

4. Steps

  1. Choose constants. For simple \(G\), combine the per-step estimates, then raise thresholds for shooting, label transfer, logarithmic OPE and weak comparison. Instantiate first stage \(4\), Taylor order \(6 + i\), power \(2 + i\), support exponent \(100\), allowances \(1\) and \(j + 1\), and target rate \(- 7 / 16\); the final wait also supplies coverage. uv_log.

  2. Take fixed-level limits. For the constructed arc, contraction has rate \(1 / 2\) uniformly in cutoff; continuity and kinematic convergence give convergence of each iterate. Dependence on levels above \(m + v\) decays as \(\text{soft}^v\), giving real-shift limits, then Vitali gives every complex-shift derivative. Camp.artistry, Camp.viva (proofs read; contraction and decay estimates used as stated).

  3. Pass to correlations. For convergent real shifts and the bundle bounds, effective densities converge uniformly and partition integrals stay positively bounded below. The RG identity and coarse approximation reduce separated products to convergent effective expectations; chessboard errors tend to zero, proving the Cauchy property. Hearth.food.

  4. Obtain derivatives and analyticity. Large-period slicing bounds lattice differences through one extra order; dense-grid convergence and Arzelà–Ascoli give spatial derivatives and smearing. Holomorphic transfer bounds and Vitali give the analytic correlation limits. Poppy.blossoms, Beach.sparksP (proofs read through stated transfer estimates).

  5. Assemble common outputs. Uniqueness identifies full-mesh smeared limits with extracted limits. Apply OPE services, every-phase reconstruction and Gram comparison; coverage supplies portfolios along every finer ultrafilter. coordination, common, habitat (report proofs read).

5. Where the difficulty lies

Uniform control over arbitrarily many ultraviolet levels and the balance between coarse-approximation gains and coefficient costs are the main analytic difficulties. Derivative limits require large periods; plain smeared limits hold at every fixed period. The coupling/activity disc is uniform in cutoff and period, whereas the correlation disc may depend on labels and period.

6. Notation

\(\mathcal{L}_d\) is the degree-\(d\) label space; \(\text{ref} (d)\) its reference level. \(a_0 = 1\), \(a_{m + 1} = a_m / L_m\), \(\sigma_m = \log (1 / a_m)\); \(\ell_p\) is the torus side. \(\Omega_p^n \subset (\mathbb{R}^4)^n\) consists of configurations distinct modulo \(\ell_p \mathbb{Z}^4\); admissible tests are complex-valued, smooth, compactly supported on distinct configurations, with closed support inside \((- \ell_p / 2, \ell_p / 2)^{4 n}\). \(O_v^{N, p}\) denotes the renormalized observable and brackets Wilson expectation. \(\bar{\beta}\) is the central coarse coupling, \(\hat{\beta}_m\) the reference coupling, \(\theta_m\) the threshold, \(J (m)\) the stage. \(x_m\) are coefficient maps; \(V_L\) multiplies grade \(e\) by \(L^{d - e}\). \(f \circ g\) denotes refinement of an extraction order. A chessboard root is the absolute reflected-product expectation to the reciprocal number of cells.

7. What remains unexplained

Quantitative transfer, remainder, shell and comparison contracts remain those of III.7, III.11 and VII.4. Internal proofs of several constant choices, contraction/remote-level decay, one-step label estimates and auxiliary estimates were used as stated; moving-frame proofs belong to VII.1/VII.3. The penalty's normalization relative to Balaban's parameter remains unverified. The main theorem uses III.14's separate construction: equality of its family with this one, or transfer of III.13.5 to it, is not established.

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.9, III.16, III.7, III.5, III.2, IV.1, III.11, II.2, III.3, III.15, III.8, III.6, V.2, VII.1, VII.4, III.4.

Referenced by: none above the threshold.

Lean source

Source allocation: 22,100 attributed lines; 3,224 line-counted declarations; 4,212 assigned declaration names. Of the assigned names, 4,113 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.12 Bounds uniform in the volume and convergence as the spacing tends to zeroPart III·Detailed notesIII.14 The construction run: flat representations and the ultraviolet rate →