Home·The theorem·The proof in one reading·Architecture·Units·Manuscript·Glossary·FAQ·Work items·Data·Method·Provenance·Changes
← V.1 Decomposition of the limit into phases (notes)Part V·Unit outlineV.3 Reconstruction for every phase and the reconstructed theory (notes) →

V.2. Osterwalder–Schrader reconstruction of a Wightman theory — detailed notes

Contents

1. What this unit proves 2

2. What it uses 12

3. Strategy 14

4. Steps 15

4.1 Tools 15

4.2 The strict-time continuation 16

4.3 Boundary values and spectral support 17

4.4 Poincaré covariance 18

4.5 Continuation to the full tube 20

4.6 Locality 21

4.7 Real-time vectors, Gram limits, nulls 23

4.8 Assembly (VIII.1) and the field realization 24

5. Where the difficulty lies 25

6. Notation 26

7. What remains unexplained 27

The full account of unit V.2: every proposition with its proof sketch, checked against the Lean in two rounds and edited. The condensed outline and the map of the Lean are on the unit page.

Unit page · All units · Overview

Part: V, phases and reconstruction.

Lean:

Size: 6,456 lines in 705 declarations.

part lines
proofs collected by reconstructionArcs 4,279
Fields 438
Compact 218
Distribution 189
Spark 120
Hilbert 85
Boundary 66
Overlay and Vac (ultraviolet material) 558

Three bundles in Integrated.lean are also ultraviolet material placed here by the dependency cut; they are listed at the end of §1.

1. What this unit proves

Conventions.

Unless stated otherwise, arities \(n, m\) range over \(\mathbb{N}\), including \(0\), and label tuples range over \(\mathcal{F}^n\). Supports of functions are closed supports.

Tables and the OS space.

\(\displaystyle M_{d + 1} (s) \partial_{d, b} = \sum_a s_{ab} \hspace{0.17em} \partial_{d, a} M_d (s) .\)

The induced star satisfies \(J \text{pin}_d v = \text{pin}_d (M_d (r_0) C_d v)\), where \(r_0\) is time reflection.

Continuations and boundary families.

The thirty reconstruction statements are conditional results for arbitrary tables, OS data and continuations. VIII.1 applies them to the mixed limit table and to a selected phase (§4.8).

Proposition V.2.1 (the reconstruction bundle). The statements A16 and C01–C29 (V.2.2–V.2.8) hold simultaneously. Each retains its own hypotheses. Status: theorem. Lean: Integrated.reconstructionArcs, Bid.Arcs.

Proposition V.2.2 (strict-time continuation; C01–C03). For the tables \(S, T\), the OS chamber \(V\) with Hilbert space \(\mathcal{H}\), a differential plan \(P\) and a spread \(A\), assume the slab hypotheses (H):

Then:

(a) (C01) There are linear maps \(B_p : \mathcal{F} \to \mathcal{B} (\mathcal{H})\), \(p \geqslant 0\), such that:

(b) (C02) Let \(B\) be such a family. Put

\(\displaystyle G_p (w, \vec{a}) = e^{- (\text{Im} \hspace{0.17em} w - 2 p) H} \hspace{0.17em} e^{i \hspace{0.17em} \text{Re} \hspace{0.17em} w \hspace{0.17em} H} \hspace{0.17em} U (\vec{a}),\)
\(\displaystyle M_p (z ; \vec{x}) (v) = \left\langle \Omega, \hspace{0.17em} B_p (v_1) \hspace{0.17em} G_p (z_2 - z_1, \vec{x}_2 - \vec{x}_1) \hspace{0.17em} B_p (v_2) \cdots G_p (z_n - z_{n - 1}, \vec{x}_n - \vec{x}_{n - 1}) \hspace{0.17em} B_p (v_n) \Omega \right\rangle .\)

For \(n = 0\) set \(M_p = \| \Omega \|^2\). Call \(p\) admissible for \(z\) if \(0 < p < 1\) and \(\text{Im} (z_j - z_i) > 4 p\) for \(i < j\). Then:

(c) (C03, from (b) alone) There is a continuation \(F\) with \(F_n (z ; \vec{x}) = M_p (z ; \vec{x})\) for \(z \in \Sigma_n\) and every admissible \(p\), satisfying:

Status: theorem. The derivative and spatial regularity hypotheses remain part of the stated input package. Lean:

Proposition V.2.3 (Cauchy–Riemann identities and tempered boundary values; C04). Let \(F\) satisfy (SC) and (PG).

(a) (C04) Three properties hold:

\(\displaystyle \tfrac{d}{dt} \psi_v (y + ta ; f) |_{t = s} = \psi_v \left( y + sa ; - i \hspace{0.17em} \partial_{\underline{a}} f \right), \qquad \partial_{\underline{a}} = \sum_k a_k \hspace{0.17em} \partial / \partial x^0_k .\)

(b) If \(F\) satisfies (I), (HC), (CR) and (PG), there is a boundary family \(W\) with (BV): the integrands defining \(\psi_v (y ; f)\) are integrable for \(y \in E_n\), and \(\psi_v (y ; f) \to W_n (v) (f)\) as \(y \to 0\) within \(E_n\), for all \(n\), \(v\) and \(f\). In particular, the radial boundary limit (BV\(_{\text{ray}}\)) holds: \(\psi_v (\varepsilon u ; f) \to W_n (v) (f)\) as \(\varepsilon \downarrow 0\) for each \(u \in E_n\).

Status: theorem. Part (b) lives in Shift/Breath and Shift/Osc, which are filed under VI.1. It is used through Show.Rhythm.opens (VIII.1). The boundary family is an output of the reconstruction. Lean:

Proposition V.2.4 (Laplace identity and spectral support; C05, C06). Let \(F\) satisfy (SC), (PG), (CR) and (HC), and have boundary family \(W\) (BV).

(a) (C05; Laplace identity (L)) For \(y \in E_n\) and compactly supported \(f \in \mathcal{S} (\mathbb{R}^{4 n})\),

\(\displaystyle W_n (v) (\hat{f}) = \psi_v (y ; \widehat{e^{\langle \underline{y}, \cdot \rangle} f}) .\)

(b) (C06, from (SC), (PG), (I) and (a)) \(\mathcal{F}W_n (v)\) is supported in

\(\displaystyle \mathcal{P}_n = \left\{ p : \left. \sum_k y_k p^0_k \geqslant 0 \text{ for every nondecreasing } y \right\} = \left\{ p : \sum_k p^0_k = 0, \sum_{k \geqslant j} p^0_k \geqslant 0 \forall j \right\} . \right.\)

(c) (Lorentz covariance) If \(W\) also satisfies (Cov) (V.2.5), then \(\mathcal{F}W_n (v)\) is supported in

\(\displaystyle \mathcal{V}_n = \bigcap_{\Lambda \in \mathcal{L}^{\uparrow}_+} \tilde{\Lambda}^{- 1} \mathcal{P}_n,\)

where \(\tilde{\Lambda} = \textsf{r} \Lambda \textsf{r}\) is the dual action (\(\textsf{r}\) reverses the spatial components). Every \(p \in \mathcal{V}_n\) has \(\sum_k p_k = 0\) and \(\sum_{k \geqslant j} p_k \in \bar{V}_+\) for all \(j\). This is the cone-support hypothesis of C14–C16 and C19.

Status: theorem. (SC) is not used in C05 or C06, and (I) is not used in C06. Lean:

Proposition V.2.5 (Poincaré covariance; A16, C07, C08, C09, C28).

Data.

(a) (A16) \(\mathcal{L}^{\uparrow}_+\) is generated, as an abstract group, by the spatial rotations and the coordinate boosts.

(b) (C08, no hypotheses) For all \(n\) and \(f \in \mathcal{S} (\mathbb{R}^{4 n})\):

(c) (C07) Assume \(S\) and \(T\) satisfy the hierarchy, the inclusions of \(P\) are compatible with \(S\), \(T\) respects \(S\), and Euclidean covariance, the rotation laws of VI.1 and spin compatibility hold. Then every \(F\) with (SC), (PG), (I), (CR) and (HC) satisfies the boost identity (B\(_y\)): for all \(y \in E_n\), tags \(v\), spatial axes \(a\) and \(f \in \mathcal{S} (\mathbb{R}^{4 n})\),

\(\displaystyle \sum_{k = 1}^n \psi_{v^{(k)}} (y ; f) = \psi_v (y ; D_{a, y} f), \qquad D_{a, y} = \sum_k \left( x^a_k \hspace{0.17em} \partial_{x^0_k} + (x^0_k + iy_k) \hspace{0.17em} \partial_{x^a_k} \right) .\)

Here \(v^{(k)}\) is \(v\) with \(v_k\) replaced by \(- i \rho (e_0 \wedge e_a) v_k\). Among the table, plan and spin inputs, the boost argument uses the Euclidean hierarchy and rotation data.

(d) (C09) Assume A16, C08 and the same table, plan, rotation and spin hypotheses as in (c). If \(F\) satisfies (SC), (PG) and (B\(_y\)) and has boundary family \(W\) (BV), then \(W\) satisfies (Cov):

(e) (C28) Assume C08, \(S\) symmetric, the inclusions of \(P\) compatible with \(S\), spin compatibility, and \(J\) the induced star. Let \(g = (a, \Lambda)\) act on words letterwise by \((u, f) \mapsto (K (\Lambda) u, g_{\ast} f)\). Then (Cov) implies:

Status: theorem. Lean:

Proposition V.2.6 (Fourier–Laplace continuation in the forward tube; C10–C16).

Definitions.

\(\displaystyle \mathcal{W}_n (v) (z) = (2 \pi)^{- 4 n} \hspace{0.17em} (\mathcal{F}W_n (v)) (e^{i \langle \cdot, z \rangle} \chi_n), \qquad z \in \mathcal{T}_n .\)

Statements.

(a) (C10, cone duality) For \(p \in \mathcal{V}_n\) and \(\eta \in \bar{V}^+_n\), \(\langle p, \eta \rangle \geqslant 0\). For \(\eta \in V^+_n\) there is \(c > 0\) with \(\langle p, u \rangle \geqslant c \lvert p \rvert\) for all \(p \in \mathcal{V}_n\) and all \(u\) near \(\eta\).

(b) (C11) Cutoffs exist.

(c) (C12) Given (a), every cutoff satisfies the damping properties (D):

(d) (C13) Given (a), \(\mathcal{W}_n (v)\) is holomorphic on \(\mathcal{T}_n\).

(e) (C14, the integral representation) Given (c), for \(\eta \in V^+_n\) the identity \(W_n (v) [\eta] (f) = \int \mathcal{W}_n (v) (x + i \eta) f (x) \hspace{0.17em} dx\) holds, with integrable integrand.

(f) (C15) Suppose:

(g) (C16) Given \(\text{supp} \mathcal{F}W \subseteq \mathcal{V}_n\) and (Cov), \(\mathcal{W}_n (v) (z) =\mathcal{W}_n (K (\Lambda) v) (\Lambda z)\) for \(z \in \mathcal{T}_n\) and \(\Lambda \in \mathcal{L}^{\uparrow}_+\), with \(\Lambda\) acting on real and imaginary parts.

Status: theorem. Several listed hypotheses are not used (§5). Lean:

Proposition V.2.7 (locality; C17–C21, C27).

Setting.

Statements.

(a) (C17) There are spectator profiles \(y_m \to 0\) with the following property. For every \(m\) and every \(x \in \Gamma_Q\) there are a unit direction \(e\), \(t > 0\), and neighbourhoods \(U_{\text{o}}, U_{\text{t}}\) of \(x + i \underline{y_m}\) such that, for every \(b \in [0, t]\):

(b) (C18) Under spin compatibility, \(t \mapsto K_d (\Lambda_e (t))\) extends to an entire function of \(t \in \mathbb{C}\), for each \(d\) and \(e\).

(c) (C19) Assume (TC), cone support V.2.4(c), spin compatibility, \(S\) symmetric, the inclusions of \(P\) compatible with \(S\), (b), and the real covariance V.2.6(g). Let \(v\) be labels, \(Q\) a pair, and \(y\) a spectator profile admitting the paths of (a). Then every \(x \in \Gamma_Q\) carries two germs:

(d) (C20) Assume \(S\) and \(T\) symmetric, the inclusions of \(P\) compatible with \(S\), \(T\) smooth, \(T\) respects \(S\), (SC) and (TC). For any such germs and any \(x \in \Gamma_Q\) with all \(x^0_k = 0\), \(g^{\text{o}}_x = g^{\text{t}}_x\) near \(x + i \underline{y}\).

(e) (C21) Assume (D) and (TC). Each \(x \in \Gamma_Q\) has a radius \(r > 0\) such that, for compactly supported \(f\) with support in \(B (x, r)\):

\(\displaystyle W_n (v) [\underline{y}] (f) = \int g^{\text{o}}_x (\xi + i \underline{y}) f (\xi) \hspace{0.17em} d \xi, \qquad W_n (v \circ s) [\sigma \underline{y}] (f \circ \sigma) = \int g^{\text{t}}_x (\xi + i \underline{y}) f (\xi) \hspace{0.17em} d \xi .\)

(f) (Boundary exchange) Suppose:

(g) (C27) Joint locality implies word locality: \(W (a \hspace{0.17em} (u, f) (u', g) \hspace{0.17em} c) = W (a \hspace{0.17em} (u', g) (u, f) \hspace{0.17em} c)\) for all words \(a\), \(c\), labels \(u\), \(u'\), and tests \(f\), \(g\) such that \(\text{supp} f - \text{supp} g\) consists of spacelike vectors.

Status: theorem. Lean:

Proposition V.2.8 (real-time vectors, Gram limits, nulls; C22–C26, C29).

Heated vectors. For \(n \geqslant 1\), take \(v \in \mathcal{F}^n\), heights \(y \in E_n\) with positive entries, \(p \in (0, 1)\) satisfying \(4 p < y_k\) and \(4 p < y_j - y_i\) for \(i < j\), and real points \(r \in \mathbb{R}^{4 n}\). Set \(z_k = r^0_k + iy_k\). The heated vector is

\(\displaystyle \Psi_p (v ; y ; r) = e^{- (y_1 - p) H} e^{ir^0_1 H} U (\vec{r}_1) \hspace{0.17em} B_p (v_1) \hspace{0.17em} G_p (z_2 - z_1, \vec{r}_2 - \vec{r}_1) \hspace{0.17em} B_p (v_2) \cdots G_p (z_n - z_{n - 1}, \vec{r}_n - \vec{r}_{n - 1}) \hspace{0.17em} B_p (v_n) \Omega .\)

For \(n = 0\) set \(\Psi_p = \Omega\). For a word \(b = ((u_1, f_1), \ldots, (u_n, f_n))\) and \(t > 0\), the heat-cup vector is

\(\displaystyle \Phi_t (b) = \int f_1 (x_1) \cdots f_n (x_n) \hspace{0.17em} \Psi_{p (t)} (u ; (t, 2 t, \ldots, nt) ; x) \hspace{0.17em} dx, \qquad p (t) = \min (t, 1) / 8.\)

(a) (C22, the pointwise Gram identity) Assume (H), \(J\) the induced star, the outputs of V.2.2(a)–(c), the orientation of \(P\), and \(T\) respects \(S\). Take \(x \in E_n\) and \(y \in E_m\) with positive entries, and widths \(t\), \(u\) admissible with all heights exceeding \(4 t\) (resp. \(4 u\)). Then

\(\displaystyle \langle \Psi_t (v ; x ; r), \Psi_u (w ; y ; q) \rangle = F^h_{n + m} (\varrho) \hspace{0.17em} (Jv_n, \ldots, Jv_1, w_1, \ldots, w_m),\)

with \(h = (- x_n, \ldots, - x_1, y_1, \ldots, y_m)\) and \(\varrho = (r_n, \ldots, r_1, q_1, \ldots, q_m)\).

(b) (C23, integrability and translation covariance of heat cups) Assume (H), the slab-insertion properties V.2.2(a), and spin compatibility. For every word \(b\) and \(t > 0\), the integrand of \(\Phi_t (b)\) is Bochner integrable, and \(\Phi_t (g \cdot b) = U (a) \Phi_t (b)\) for translations \(g = (a, 1)\), where \(U (a)\) is the unitary translation of \(\mathcal{H}\).

(c) (C24, the heat-cup Gram limit) Assume (H), (SC), (BV), integrability and translation covariance of heat cups and the pointwise Gram identity. Then \(\langle \Phi_s (b), \Phi_{s'} (c) \rangle \to \langle b, c \rangle_W\) as \((s, s') \to (0^+, 0^+)\) independently.

(d) (C25) For every \(n \geqslant 0\), tensor products span a dense subspace of \(\mathcal{S} (\mathbb{R}^{4 n})\).

(e) (C26, label faithfulness) Assume C25, \(N_d (T) = N_d (S)\) for all \(d\), the hierarchies for \(S, T\), compatibility of the inclusions of \(P\) with both tables, (SC) and (TC). A label \(u \in \mathcal{F}\) with \(W (a \hspace{0.17em} (u, f) \hspace{0.17em} c) = 0\) for all words \(a\), \(c\) and all \(f\) is zero. This concerns single labels, not words.

(f) (C29, no hypotheses) For every boundary family \(W\), every antilinear map \(J\), all words \(a, c\) and all labels \(u, u'\), the following hold:

Status: theorem. Lean:

Proposition V.2.9 (the target: a Wightman structure realizing \(W\)).

A Wightman structure consists of:

It realizes \(W\) if:

The field axioms without vacuum uniqueness (VIII.1) are the same list without the last clause. Status: definition (a specification). It is produced in VIII.1 by the construction of V.2.10. Consequence: Fields.general and Fields.sees show that a realization has Gram form \(\langle \varphi (b) \Omega_W, \varphi (c) \Omega_W \rangle = \langle b, c \rangle_W\), so \(W\) is positive on words. Lean: Wightman, Wightman.Realizes, Fields, Fields.Total, Symmetries, Action, FieldCore.

Proposition V.2.10 (field realization of \(W\)). Fix rotation data \(R\) and a real structure on \(S\). Let \(J\) be the induced involutive star and assume spin compatibility. Let \(W\) satisfy word bilinearity, continuity in the two tests, word locality, label faithfulness, continuity of word matrix elements and invariance of the word Gram form. Suppose vectors \(\Phi_s (b) \in \mathcal{H}\), \(s > 0\), satisfy the Gram limit V.2.8(c), \(\Phi_s (\varnothing) \to \Omega\) as \(s \downarrow 0\), and, for every word \(b\) and translation \(a\), \(\Phi_s (a \cdot b) = U (a) \Phi_s (b)\) for all sufficiently small \(s > 0\). Assume \(T\) satisfies the normalized, unital, permutation-symmetric Euclidean laws. Then there is a Wightman structure satisfying every clause of V.2.9 except possibly vacuum uniqueness. If the common fixed vectors of all \(e^{- tH}\), \(t > 0\), are multiples of \(\Omega\), it realizes \(W\). Status: theorem. The route used by the main theorem is Next.BOUNDARY.establishing followed by Compatible.vacuum (VIII.1). Boundary.harvest states the same in one lemma and is used only by an alternative route. The ingredients are in §4.8.

Not reconstruction. Integrated.lean also proves three ultraviolet bundles, and the folders Overlay and Vac are ultraviolet material; they belong to Part III (III.6, III.10–III.13):

2. What it uses

Inputs from other units.

  1. The Euclidean hierarchy (IV.1; for each phase, V.1): Table.Hierarchy. It consists of permutation symmetry, normalization, the unit-insertion law, compatibility with signed symmetries and conjugation, translation homogeneity, reflection positivity on words and on smeared joint tests, smoothness off coincidences, and capped polynomial bounds near collisions. Its uses include:

  2. The OS space and the slab operators (V.1, the B-items; files YM/OS, YM/Heat).

  3. Rotation, spin and star data (VI.1).

  4. The label structures of I.1 and II.1: the gauge group, the register, the quotient labels, \(\text{pin}\), and the common level of finitely many labels.

Standard mathematics.

Imported constants and assumptions.

3. Strategy

The slab bound controls a field insertion \(B_p (v)\) between two heat factors. The relation \(B_{p + u} = e^{- uH} B_p e^{- uH}\) makes products of these insertions independent of the admissible width. Their vacuum matrix elements give a continuation to strictly ordered imaginary times, with polynomial growth as the time differences shrink. Cauchy–Riemann identities and Taylor's formula along \((0, 1, \ldots, n - 1)\) then give tempered boundary values for arbitrary approaches through ordered heights. A Laplace identity yields total energy zero and nonnegative suffix energies.

Euclidean covariance gives Poincaré covariance of the boundary family: translations and spatial rotations pass through a one-variable identity theorem; infinitesimal rotations in a time–space plane give the boost identity, which is integrated along the real boost groups. Lorentz covariance improves the spectral support to \(\mathcal{V}_n\). Fourier–Laplace transformation extends the continuation to the forward tube. Locality follows from two holomorphic germs near each spacelike adjacent pair: imaginary boosts construct the germs, Euclidean permutation symmetry identifies their jets on the Wick slice, and continuation along a rescaling of real times propagates their agreement. Partial boundary values then give exchange of adjacent spacelike fields.

Finally, inner products of heated OS vectors equal the continuation at a glued, reflected configuration. Their limit gives positivity of the word kernel and an isometric embedding of its Kolmogorov space into the OS Hilbert space. Left multiplication defines the fields on the common word domain. The embedding intertwines translations, transferring positive energy; Lorentz covariance gives the forward-cone spectrum condition. A one-dimensional OS heat-fixed space gives vacuum uniqueness. These are the operator and local-germ forms of Osterwalder–Schrader reconstruction used here.

4. Steps

4.1 Tools

Lemma V.2.a (Kolmogorov). Let \(K\) be a positive semidefinite kernel on a set \(X\). There are a Hilbert space and vectors \((\psi_x)_{x \in X}\) with \(\langle \psi_x, \psi_y \rangle = K (x, y)\) and dense span. Proof sketch. Complete the finitely supported functions modulo the null space of \(\sum \bar{f} (x) g (y) K (x, y)\). Lean: Kernel.host, Kernel.Gram.

Lemma V.2.b (projection-valued measure). Let \(X\) be a compact Hausdorff space and \(A : C (X) \to \mathcal{B} (\mathcal{H})\) a unital complex \(\ast\)-algebra homomorphism. There is a strongly countably additive projection-valued measure on the Borel sets of \(X\) representing \(A\). Proof sketch.

  1. For each \(\psi\), the positive functional \(f \mapsto \langle \psi, A (f) \psi \rangle\) gives a Riesz–Markov measure \(\mu_{\psi}\).

  2. \(f \mapsto A (f) \psi\) extends to an isometry \(L^2 (\mu_{\psi}) \to \mathcal{H}\).

  3. \(E (s) \psi\) is the image of \(1_s\); for disjoint Borel sets, the corresponding indicator sums converge in \(L^2 (\mu_{\psi})\), giving strong countable additivity.

It is used (V.1, SNAG/Escape) on the one-point compactification of \(\mathbb{R}^4\) to build the spectral measure of the translations. Lean: Spectral.Rep.ships (proof read for additivity; Compact/Borel statement only).

Lemma V.2.c (Schwartz calculus; tensor density, C25).

  1. Let \(f_{\lambda}, g \in \mathcal{S} (\mathbb{R}^d)\). If every jet of \(f_{\lambda}\) converges locally uniformly to the corresponding jet of \(g\), and every Schwartz seminorm of \(f_{\lambda}\) is eventually bounded uniformly in \(\lambda\), then \(f_{\lambda} \to g\) in \(\mathcal{S}\).

  2. Let \(K\) be compact metric, \(\mu\) a finite Borel measure, and \(u \mapsto f_u\) continuous from \(K\) into \(\mathcal{S} (\mathbb{R}^d)\). If \(g \in \mathcal{S}\) satisfies \(g (x) = \int_K f_u (x) \hspace{0.17em} d \mu (u)\), then the sums \(\sum_j \mu (E_j) f_{u_j}\) over finite measurable partitions, with \(u_j \in E_j\) and \(\max_j \text{diam} (E_j) \to 0\), converge to \(g\) in \(\mathcal{S}\).

  3. The linear span of tensor products is dense in \(\mathcal{S} (\mathbb{R}^{4 n})\).

Proof sketch. For 3:

4.2 The strict-time continuation

Lemma V.2.d (slab insertions; C01). As in V.2.2(a). Proof sketch.

Lean: C01.single, C01.inner_raw, C01.at_width, C01.answer.

Lemma V.2.e (Euclidean values and width independence; C02). As in V.2.2(b). Proof sketch.

Only gaps \(> 2 p\) are needed; the factor \(4\) in "admissible" is slack. Lean: C02.train_pure, C02.pad_bar, C02.matrix_pad.

Lemma V.2.f (holomorphy and polynomial growth; C03). As in V.2.2(c). Proof sketch.

Lean: C03.defined_exists, C03.defined_region, C03.defined_power.

4.3 Boundary values and spectral support

Lemma V.2.g (C04). As in V.2.3(a). Proof sketch.

Lean: C04.cr_height, C04.deriv_bound_near, C04.lines.

Lemma V.2.h (boundary values). As in V.2.3(b). For every nondecreasing \(y\), including \(y = 0\), the boundary value is given by an explicit formula. Let \(\varrho = (0, 1, \ldots, n - 1)\) and \(D = - i \partial_{\underline{\varrho}}\), and let \(m\) be the exponent of (PG). Then

\(\displaystyle W^y_n (v) (f) = \sum_{k = 0}^m \frac{(- 1)^k}{k!} \hspace{0.17em} \psi_v (y + \varrho ; D^k f) - \int_0^1 \frac{(- s)^m}{m!} \hspace{0.17em} \psi_v (y + s \varrho ; D^{m + 1} f) \hspace{0.17em} ds.\)

Proof sketch.

Lean: Waves.desired, Waves.meant, Waves.maintains, Waves.cost, Waves.seeks.

Lemma V.2.i (Laplace identity; C05). As in V.2.4(a). Proof sketch. Put \(A (s) = \psi_v (sy ; \widehat{e^{s \langle \underline{y}, \cdot \rangle} f})\) for \(s > 0\).

  1. \(A\) is constant. By (CR) along \(y\), and since \(\frac{d}{ds} \widehat{e^{s \langle \underline{y}, \cdot \rangle} f} = i \hspace{0.17em} \partial_{\underline{y}} \widehat{e^{s \langle \underline{y}, \cdot \rangle} f}\) in \(\mathcal{S}\) (here compact support is needed), \(A' (s) = 0\). The product rule is justified by the uniform bound \(\lvert \psi_{y'} (g) \rvert \leqslant \text{toll} (g)\) of V.2.h.

  2. The limit. As \(s \to 0^+\), \(A (s) \to W_n (v) (\hat{f})\), by the radial limit and the same bound.

Only the radial limit is used. Lean: C05.hasDerivAt_zero_psi, C05.tendsto_psi.

Lemma V.2.j (support in \(\mathcal{P}_n\); C06). As in V.2.4(b). Proof sketch. Let \(y\) be nondecreasing and \(f\) compactly supported in \(\{p : \langle \underline{y}, p \rangle < - 1\}\).

  1. Apply V.2.i at the ordered heights \(Y (t) = ty + \varrho\), whose gaps are all \(\geqslant 1\).

  2. (PG) with \(k = 1\) gives \(\lvert \psi_{Y (t)} (\hat{g}) \rvert \leqslant C \| \hat{g} \|_{L^1}\).

  3. Absorb \(e^{\langle \underline{\varrho}, \cdot \rangle}\) into the compact test. Every seminorm of the remaining factor \(e^{t \langle \underline{y}, \cdot \rangle}\) times that test is bounded by a polynomial in \(t\) times \(e^{- t}\), and tends to zero.

  4. Hence \(W (\hat{f}) = 0\). Rescaling \(y\) removes the \(- 1\).

Constant profiles of both signs force \(\sum_k p^0_k = 0\). Lean: C06.patch, C06.lift_falls.

4.4 Poincaré covariance

Lemma V.2.k (Poincaré action on \(\mathcal{S}\); C08). As in V.2.5(b). Proof sketch.

Lean: C08.size_affine_le, C08.continuous_gpush, C08.probe_slope_of_pointwise.

Lemma V.2.l (boost identity at strict heights; C07). As in V.2.5(c). Proof sketch.

  1. Euclidean input. Differentiating the \(O (4)\) covariance of \(T\) along \(\exp (s \hspace{0.17em} e_0 \wedge e_a)\) at a non-coincident configuration gives the rotation Ward identity: the derivative of \(T (x ; v)\) along the rotation field equals \(- \sum_k T (x ; \ldots, \rho (e_0 \wedge e_a) v_k, \ldots)\).

  2. The analytic function \(E (z)\). Fix \(f\) with compact support and introduce a complex parameter \(z\) and the times \(z \hspace{0.17em} x^0_k + iy_k\). Let \(E (z)\) be the integral of

\(\displaystyle F \cdot \left( \sum_k [x^a_k \partial_{x^0_k} + z (zx^0_k + iy_k) \partial_{x^a_k}] f \right) - z \sum_k F (v^{(k)}) f.\)

The times stay strictly ordered on \(\text{supp} f\) for \(z\) in a convex open set containing \(\mathbb{R}\), and \(E\) is holomorphic there. 3. \(E\) vanishes on a segment of the imaginary axis. At \(z = iw\), \(w > 0\) small, the configuration is Euclidean, with times \(y_k + wx^0_k\). There the vector field is \(w\) times the pull-back of the \((0, a)\)-rotation field, and \(F\) is the Schwinger function. By step 1 the integrand is the divergence of the compactly supported vector field \(fT\) times the divergence-free rotation field, so \(E (iw) = 0\). 4. Conclusion. The identity theorem gives \(E (1) = 0\), which is (B\(_y\)) for compactly supported \(f\). Density (cutoffs) and the continuity of \(f \mapsto \psi (y ; f)\) (V.2.h at fixed \(y\)) give all \(f\).

The term \(iy_k \partial_{x^a_k}\) is the imaginary part of the complex time multiplying the spatial derivative. Lean: C07.oriented_differential, C07.integral_orbit, C07.error_imag, C07.error_one, C07.compact_boost, C07.answer.

Lemma V.2.m (real Poincaré covariance; C09, A16). As in V.2.5(a), (d). Proof sketch.

Lean: C09.time_compare, C09.restricted_translation, C09.restricted_spatial, C09.boost_corner, C09.boost_invariant, A16.answer.

Lemma V.2.n (word covariance and cone support; C28). As in V.2.5(e) and V.2.4(c). Proof sketch.

Lean: C28.answer, tents.

4.5 Continuation to the full tube

Lemma V.2.o (cone geometry, cutoff, damping; C10–C12). As in V.2.6(a)–(c). Proof sketch.

Lean: C10.telescope, C10.cone_bound, C11.answer, C12.cone_estimate, C12.attenuate_estimate.

Lemma V.2.p (holomorphy in the tube; C13). As in V.2.6(d). Proof sketch.

  1. Exponential reserve. For \(z \in \mathcal{T}_n\), on \(\text{supp} \chi_n\) one has \(\text{Re} \hspace{0.17em} (i \langle p, z \rangle) = - \langle p, \text{Im} \hspace{0.17em} z \rangle \leqslant A - c \lvert p \rvert\), by C10 and distance \(\leqslant 3\). So \(e^{i \langle \cdot, z \rangle} \chi_n\) has finite seminorms weighted by \(e^{c \lvert p \rvert / 2}\).

  2. Differentiability. Write \(L_z (p) = e^{i \langle p, z \rangle} \chi_n (p)\). For each Schwartz seminorm \(q\), exponential decay gives \(q (L_{z + h} - L_z - i \langle \cdot, h \rangle L_z) \leqslant C_q |h|^2\) for all sufficiently small \(h\). Thus \(z \mapsto L_z\) is complex Fréchet differentiable into \(\mathcal{S}\).

  3. Analyticity. Compose with the continuous functional \(\mathcal{F}W\). Complex differentiability on an open set gives analyticity.

The argument applies to every Schwartz test with the prescribed exponential form; C13 asserts the existence of such tests and the resulting holomorphy. Lean: C13.lens_reserve, C13.analytic_interior; ExpReserve.of_cut, hasFDerivAt_shift.

Lemma V.2.q (partial boundary values as integrals; agreement with \(F\); C14, C15). As in V.2.6(e), (f). Proof sketch.

Lean: C14.modulation_eq, C14.emit_integral, C15.blanketed_eq, C15.partialD_psi.

Lemma V.2.r (Lorentz covariance in the tube; C16). As in V.2.6(g). Proof sketch.

Lean: C16.lens_support, C16.answer.

4.6 Locality

Lemma V.2.s (imaginary boosts at a spectator profile; C17). As in V.2.7(a). Proof sketch.

  1. Spectator profiles. The profiles \(y_m = d / (m + 1)\) are spectator profiles. Monotone heights give \(\underline{y_m} \in \bar{V}^+_n\), and \(\sigma\) fixes \(\underline{y_m}\). The closure statements follow by approaching through strictly ordered heights.

  2. Choice of direction. Let \(\xi = x_{i + 1} - x_i\) be spacelike, so \(\vec{\xi} \ne 0\), and take \(e = \vec{\xi} / \lvert \vec{\xi} \rvert\). Then

\(\displaystyle \text{Im} (\Lambda_e (ib) w)_k = \text{Im} \hspace{0.17em} w_k + \sin b \cdot c_e (\text{Re} \hspace{0.17em} w_k) + (\cos b - 1) \hspace{0.17em} s_e (\text{Im} \hspace{0.17em} w_k),\)

with \(c_e (\vec{X}, t) = (te, e \cdot \vec{X})\) and \(s_e (\vec{X}, t) = ((e \cdot \vec{X}) e, t)\). 3. The pair difference. Its imaginary part is \(0\) at \(b = 0\), and \(c_e (\xi) = (\xi^0 e, \lvert \vec{\xi} \rvert) \in V_+\) exactly because \(\lvert \xi^0 \rvert < \lvert \vec{\xi} \rvert\) (the Jost condition). So the difference enters \(V_+\) for small \(b > 0\). 4. The other consecutive differences. They are already in \(V_+\) and stay there by continuity, uniformly near the base point. 5. The swapped tube. After the swap the pair difference is \(- \xi\), and the opposite rapidity \(- ib\) compensates.

For \(n = 2\) the profile is identically \(0\), so the germs sit at real points. Lean: C17.height_path, C17.fleck_of_spectator, C17.choose_direction, C17.enters, C17.local_paths.

Lemma V.2.t (complexified labels; the two germs; C18, C19). As in V.2.7(b), (c). Proof sketch.

Lean: C18.eq_exponential, C18.entire, C19.paneAlong.

Lemma V.2.u (agreement at Euclidean points; C20). As in V.2.7(d). Proof sketch.

  1. The Wick slice. Let \(M (\vec{x}, \tau) = (\vec{x}, i \tau)\). Its complex span is \(\mathbb{C}^{4 n}\). If \(x\) has zero times, then \(M (\vec{x}, y) = x + i \underline{y}\).

  2. The original germ. On the open set of strictly ordered \(\tau\), \(M\) maps into \(\mathcal{T}_n\), and \(g^{\text{o}} \circ M =\mathcal{W} \circ M = F\) at imaginary times. That equals the Schwinger function \(T ((\vec{x}_k, \tau_k) ; v)\) (V.2.6(f), (SC)).

  3. The tilted germ. On the swapped-ordered set, \(g^{\text{t}} \circ M\) equals the same Schwinger function, by permutation symmetry of \(T\).

  4. Smoothness at the base point. The point \((\vec{x}, y)\) is non-coincident: the pair has equal times but \(\vec{x}_i \ne \vec{x}_{i + 1}\) by spacelikeness. So \(T\) is smooth there.

  5. Jets. Both open sets accumulate at \((\vec{x}, y)\). Hence all real jets of \(g^{\text{o}} \circ M\) and \(g^{\text{t}} \circ M\) at that point equal those of \(T\).

  6. Conclusion. Real jets on a slice that spans over \(\mathbb{C}\) determine all complex derivatives, so \(g^{\text{o}} = g^{\text{t}}\) near \(x + i \underline{y}\).

No Schwarz reflection and no edge-of-the-wedge theorem are used. Lean: C20.jets_on_open_side, C20.complex_germ, C20.emit_table.

Lemma V.2.v (local realization and exchange; C21). As in V.2.7(e), (f). Proof sketch.

Lean: C21.realize, Panes.reaches, Panes.forgets, NearPair.actor; Pane.propagate, Blanket.flatD.

Lemma V.2.w (locality for words; C27). As in V.2.7(g). Proof sketch. If \(\text{supp} f - \text{supp} g\) is spacelike, the tensor test of the word \(a \hspace{0.17em} (u, f) (u', g) \hspace{0.17em} c\) is supported in \(\Gamma_Q\), with \(Q\) the slots of \((u, f)\) and \((u', g)\). Composing with \(\sigma\) swaps the two factors. Joint locality applies to all Schwartz tests supported in \(\Gamma_Q\), so no density argument is needed. The pair may be at any position in a word of any length \(\geqslant 2\). Lean: C27.coordinate_tsupport, C27.exchange.

4.7 Real-time vectors, Gram limits, nulls

Lemma V.2.x (Gram form of heated vectors; C22). As in V.2.8(a). Proof sketch.

  1. Imaginary times. At imaginary times each heated vector is the GNS vector of a positive-time Euclidean word.

  2. Continuation in real times. Introduce one complex parameter \(c\) and the times \(iy_k + c \hspace{0.17em} r^0_k\). Both sides are holomorphic on a strip \(\lvert \text{Im} \hspace{0.17em} c \rvert < \delta\); the left side is antiholomorphic in the first slot, handled by conjugation.

  3. Conclusion. The two sides agree at \(c = i \tau\) for small \(\tau\), hence at \(c = 1\).

The width formula of C03 is not used. Lean: C22.euclideanRef, C22.onImag, C22.pointJoined.

Lemma V.2.y (heat cups and their Gram limit; C23, C24). As in V.2.8(b), (c). Proof sketch.

Lean: C23.shot_control, C23.heatCup_shift, C24.toward, C24.integrated.

Lemma V.2.z (nulls; word insertion; C26, C29). As in V.2.8(e), (f). Proof sketch.

Lean: C26.marker_on_joint, C26.strict_zero, C26.read_zero, C29.answer.

4.8 Assembly (VIII.1) and the field realization

The reconstruction is assembled in Next/BOUNDARY/Stage.lean, in this order.

  1. rhythm chains C01 → C02 → C03 → C04.

  2. illuminated applies C07, C09 (with A16 and C08), C05 and C06. spectra uses Lorentz improvement of spectral support to get cone support.

  3. blankets (C10, C12), visible (C13) and neon_any (C13–C15) give (TC). emit_eq shows that the continuation on \(\mathcal{T}_n\) does not depend on the cutoff.

  4. windows combines C17, C19 (with C18 and C16), C20 and C21. Show.uses then gets joint locality from the boundary exchange lemma and Lorentz improvement of spectral support.

  5. revealing applies C22, C26 with C25, and C24 with C23.

  6. establishing applies C28, C29 and C27, then builds the field system (step 7).

  7. The field system is built (V.2.10) in the following order.

    1. The heat cups converge: the two-variable Gram limit makes \(s \mapsto \Phi_s (b)\) Cauchy (norms_kernel, Cups.out). The limit vectors \(\Phi_0 (b)\) satisfy \(\langle \Phi_0 (b), \Phi_0 (c) \rangle = \langle b, c \rangle_W\) and \(\Phi_0 (\varnothing) = \Omega\).

    2. Hence \(\langle \cdot, \cdot \rangle_W\) is positive (Outfield.safe), and V.2.a gives its Kolmogorov space.

    3. Fields act by left multiplication of words. The identity \(\langle (u, f) b, c \rangle_W = \langle b, (Ju, \bar{f}) c \rangle_W\) shows that a null linear combination of words remains null after left multiplication, so the fields are well defined (Soil.starter).

    4. The same identity gives the formal adjoint. The other properties of \(W\) give the remaining clauses of V.2.9: temperedness (continuity in the two tests); local commutativity (word locality); field faithfulness (label faithfulness).

    5. A unitary representation \(U\) of \(\mathcal{P}\) is defined by \(U (g) [b] = [g \cdot b]\), using invariance of the word Gram form and continuity of word matrix elements (Soil.mobilize).

    6. \([b] \mapsto \Phi_0 (b)\) is an isometry into \(\mathcal{H}\) that intertwines translations and sends the vacuum to \(\Omega\) (universal_embed, Outfield.daylight).

    7. Spectrum condition (Action.spectrum). The isometric intertwiner identifies each scalar spectral measure with that of its image in \(\mathcal{H}\), by equality of characteristic functions. These measures are supported in \(\{p^0 \geqslant 0\}\) by positivity of \(H\). Lorentz covariance of \(U\) upgrades this to support in \(\bar{V}_+\).

    8. Vacuum uniqueness (Action.only). Under one-dimensionality of the heat-fixed space, an invariant vector is mapped to the ground space of \(\mathcal{H}\), which is \(\mathbb{C} \Omega\).

  8. Compatible.vacuum adds uniqueness for the selected phase. after_materializes runs the chain for the mixed table (\(S = T\), without vacuum uniqueness) and for the selected phase.

5. Where the difficulty lies

The boost argument must control moving labels and tests without imposing a topology on the full label union. All labels are first placed in one finite-dimensional level; Schwartz estimates then justify the derivative and its boundary limit. The locality argument needs equality of complete jets on the Wick slice. Equality of values at zero real times would not determine the two germs. Smoothness of the Euclidean correlations at non-coincident configurations, including equal pair times, supplies those jets.

The Taylor formula in V.2.h gives a bound uniform over all nondecreasing heights. This uniformity permits varying tests in the Laplace identity and independent heating scales in the Gram limit. Positivity of the word kernel then suffices for the field construction; surjectivity of the embedding into the OS space is not required.

Several formal hypotheses are stronger than the corresponding proof needs. The slab-insertion construction uses the slab bound but not the derivative laws of the spread. C05 and C06 use the tested identities and growth bound without Euclidean root values. C14 uses the damping properties without spectral support; C16 uses spectral support and Lorentz covariance without agreement with the strict-time continuation. These distinctions do not remove the hypotheses from the statements above.

The pairing is Euclidean, so momenta transform by \(\tilde{\Lambda} = \textsf{r} \Lambda \textsf{r}\). The tube uses absolute coordinates and imposes no condition on its first imaginary point. For \(n = 1\), it is \(\mathbb{C}^4\) and \(\mathcal{V}_1 = \{0\}\). The polynomial bound is per label tuple, with no uniformity in the arity. The chosen value \(B_0 = 0\) is outside the admissible widths. The Laplace identity is stated on compactly supported tests.

6. Notation

object code name proposed notation
OS Hilbert space of \(T\), time axis OS.Chamber \(\mathcal{H}\), \(\Omega\), \(H\), \(U (\vec{a})\)
slab operators of a spread; slab bound OS.Spread, AllCopy \(A_d (w)\); (SB)
hypothesis package of C01 Bid.Pantry (H)
labels of \(S\); tags; insertion S.Union, S.Tag, pin \(\mathcal{F}\), \((d, v)\), \(\text{pin}_d\)
Euclidean function of \(T\) on labels of \(S\) S.useAt T \(T (x ; v)\)
single insertions between heat factors Small (bar) \(B_p (v)\)
propagator between insertions Small.rip \(G_p (w, \vec{a})\)
strict-time matrix element Small.matrix \(M_p (z ; \vec{x})\)
ordered heights; strict region east, strict \(E_n\), \(\Sigma_n\)
heights placed as times dawn \(\underline{y}\)
continuation; restriction; smearing Waves, restricted, psi \(F_n\), \(F^y_n\), \(\psi_v (y ; f)\)
strict continuation; integrability; height continuity; Cauchy–Riemann Region, Bounds, Along, Lines (SC), (I), (HC), (CR)
polynomial growth in the inverse time separation Breathes (PG)
boundary family (Wightman functions) Surface \(W_n (v)\); \(W (b)\)
boundary-value hypotheses Opens, NonTangentialFree (BV\(_{\text{ray}}\)), (BV)
Laplace identity; half-space support during, cone_tests, half (L); \(\mathcal{P}_n\)
Wightman spectral set coneM, cones \(\mathcal{V}_n\)
imaginary-part cones; tube fields, shores, tube \(V^+_n\), \(\bar{V}^+_n\), \(\mathcal{T}_n\)
cutoff; damping properties; partial boundary value Blanket, Modified, partialD \(\chi_n\); (D); \(W_n (v) [\eta]\)
Fourier–Laplace continuation Aurora.emit, FullBody \(\mathcal{W}_n (v) (z)\); (TC)
rotation structure; spin structure Rotation (\(R\), lie), Spins (rail, shell) \(R\), \(\rho\); \(K (\Lambda)\)
boost identity; Poincaré covariance Boosty; Urs (B\(_y\)); (Cov)
induced star; word star; kernel StarOp, rev, kernel \(J\); \(b^{\ast}\); \(\langle b, c \rangle_W\)
adjacent pair; spacelike set; swap NearPair, gap, switch/name \(Q = (i, i + 1)\), \(\Gamma_Q\), \(\sigma\)
spectator profile; complex boost spectator/Fleck; augment \(y\); \(\Lambda_e (\zeta)\)
germs at a spacelike pair Panes \(g^{\text{o}}_x\), \(g^{\text{t}}_x\)
heated vector; heat cup shot, heatCup \(\Psi_p (v ; y ; r)\); \(\Phi_t (b)\)
reconstructed theory Boundary.Wightman \((\mathcal{H}_W, \mathcal{D}, \Omega_W, \varphi, J, U, \textsf{E})\)
Kolmogorov space of \(\langle \cdot, \cdot \rangle_W\) Soil \(\mathcal{H}_W\)
the bundle Bid.Arcs (A16), (C01)–(C29)

7. What remains unexplained

  1. Upstream slab estimate. The constants in (SB) are independent of the slab width and word length for fixed degree bound \(D\) and representative bound \(R\); they may depend on \(D, R\). VIII.1 obtains the mixed and selected slab laws from the phase-data package. The upstream proof of those laws, beyond their use in reconstruction, is outside this check.

  2. Imported analytic results. The holomorphic heat-semigroup construction and the OS translation spectral measure are used by their stated properties. The construction of the compact-model Borel functional calculus, and the estimates underlying complex differentiation of the Fourier–Laplace tests, have not been checked in full. Their exact uses are identified in V.2.b, V.2.f and V.2.p.

  3. Location of the arguments. The boundary-value construction, the spectral-support improvement and the germ exchange lemmas are filed under VI.1; the field realization and application to phase data are filed under VIII.1. They belong with the reconstruction argument in a written proof.

  4. Vacuum uniqueness. The construction for the mixed table supplies the field axioms without vacuum uniqueness. The selected phase has uniqueness because its OS heat-fixed space is one-dimensional. No uniqueness assertion for the mixed table follows here.

  5. References. The bibliographic pointers in §2 identify the relevant standard methods; their theorem numbering has not been verified.

← V.1 Decomposition of the limit into phases (notes)Part V·Unit outlineV.3 Reconstruction for every phase and the reconstructed theory (notes) →