Home·The theorem·The proof in one reading·Architecture·Units·Manuscript·Glossary·FAQ·Work items·Data·Method·Provenance·Changes

The work items

Contents

What they are 1

Why they are a natural skeleton 2

Families 2

Family descriptions 3

Family D: infinite-volume extraction (Part IV) 3

Family B: phase decomposition and Osterwalder–Schrader positivity (Part V) 3

Family C: Osterwalder–Schrader to Wightman reconstruction (Part V) 3

Family A: stress tensor, Ward identity and rotations (Part VI) 4

Family E: energy and pressure estimates for a renormalization-group step (Part III) 4

Family F: stationary point and Hessian of the renormalization-group step (Part III) 4

Family G: label polynomials (Part III) 5

Family H: fusion of forced activities (Part III) 5

Family I: marker supports and widths (Part III) 5

All work items 5

Family A (Part VI): stress tensor, Ward identity, rotations and boosts 6

Family B (Part V): phase decomposition, positivity in each phase, cutoff estimates 7

Family C (Part V): Osterwalder–Schrader to Wightman reconstruction 8

Family D (Part IV): extraction of the infinite-volume limit 9

Family E (Part III): energy and pressure estimates for one renormalization-group step 9

Family F (Part III): stationary point, Hessian, positivity margin 11

Family G (Part III): label polynomials 12

Family H (Part III): fusion of forced activities 12

Family I (Part III): marker supports and widths 13

What they are

The Lean development is organized around about one hundred stated obligations. Each is a proposition YM.Bid.Xnn in the directory YM/Bid, stated for all gauge data, label systems and tables, and each is proved by a theorem answer : YM.Bid.Xnn in its own file YM/Work/Xnn.lean (sometimes with a folder YM/Work/Xnn/ of auxiliary modules). The code calls these obligations proof requests (YM.Bid); this page calls them work items. There are 105 of them, in nine families A to I, with 28,654 lines in total (the file together with its auxiliary modules). Three further entries in YM/Work/ (MY1/, MY3ActiveGeometry, MY3SmoothSeedSectors) are not work items.

The proofs are written as constraints, with docstrings of the form "no such-and-such is assumed", so the docstrings are often negative. The descriptions below are read from the statement of the proposition together with the docstring. The statements are assembled in the files of YM/Bid into named structures (for example Bid.station, Bid.Looms, Bid.Arcs) which take the propositions as hypotheses and produce the objects used by the main theorem.

Why they are a natural skeleton

Each work item states one mathematical fact about the limit theory, in terms of objects (tables, phases, rotation actions, Hilbert spaces) that are defined elsewhere, and has a self-contained proof. Parts IV to VI of the proof (infinite volume; phases and reconstruction; symmetry) are covered almost entirely by the families D, B, C and A, so the work items give a ready division of an exposition of those parts into statements of moderate size, with the dependencies between them stated explicitly. The families E to I belong to Part III, the renormalization group; they have the same form and are listed here for completeness, but they cover only a part of that construction, namely the estimates for one renormalization-group step.

The families build on each other in the following order within the limit theory: D (the limit table, with derivatives) comes first; then B (phases, positivity in each phase); then C (reconstruction, for each phase); and A (Ward identity, rotations, Lorentz group), where the boost items A16 and A17 enter C09 and the rotation layer A10–A15 feeds B06 through the rotation and orientation data.

Readings of the statements are marked certain, probable or tentative in the sense of the glossary: certain where the statement is short and self-explanatory, probable where the statement and docstring were read but not the full proof, tentative where the vocabulary (the code's nicknames) is not resolved. Lean names in code spans are explained in the glossary where they occur there.

Families

family part items number lines content
A VI A01–A17 17 5,022 stress tensor, Ward identity, rotations and boosts
B V B01–B18 18 3,603 phase decomposition, positivity in each phase, cutoff estimates
C V C01–C29 29 6,193 Osterwalder–Schrader to Wightman reconstruction
D IV D01–D06 6 1,035 extraction of the infinite-volume limit
E III E01–E09 9 3,370 energy and pressure estimates for one renormalization-group step
F III F01–F09 9 5,259 stationary point, Hessian, positivity margin
G III G01–G09 9 2,865 label polynomials
H III H01–H04 4 847 fusion of forced activities
I III I01–I04 4 460 marker supports and widths
105 28,654

Part VI is Symmetry: stress tensor and Ward identities (family A); Part V is Phases and reconstruction (families B and C); Part IV is Infinite volume (family D); Part III is Ultraviolet construction (renormalization group) (families E to I).

Family descriptions

Families are listed in the order in which they build on each other in the limit theory (D, B, C, A), followed by the renormalization-group families E to I.

Family D: infinite-volume extraction (Part IV)

D01–D06, 6 items, 1,035 lines. These show that the ordered limit (first the cutoff, then the volume, along a chosen subsequence) is a limit of the lattice tables in a sense strong enough to inherit smoothness. D01–D03 give the coefficient budget, control of the discrepancy between separated tuples and smeared test functions, and convergence of one-slot forward differences to derivatives of the limit kernel. D04 assembles them into Org.Arrives: the limit table is a spanning limit, so label derivatives pass to the limit. D05–D06 give the one-cap derivative bound and the compatibility of the limit differential system with reflection and signed symmetries; these are used by Bid.healthy to show that the limit table satisfies the Euclidean hierarchy and the orientation law. The family is small: the uniform-in-volume estimates (compactness, Arzelà–Ascoli) are elsewhere.

Family B: phase decomposition and Osterwalder–Schrader positivity (Part V)

B01–B18, 18 items, 3,603 lines. This family builds the Osterwalder–Schrader Hilbert space structure of the mixed limit theory and splits it into clustering phases. B01–B03 give the time-slab word operators (Spread), their norm laws, and the Theme (permutation invariance and linearity through the field quotient). B04–B05 pass from lattice grids to continuum points and prove reflection invariance of the phases; B06 upgrades this to invariance under all of O(4). B07 produces the phase decomposition itself: the Gelfand characters of the commutative algebra of vacuum-compressed operators (Garland), giving S^mix = ∫ S^ξ. B08–B14 then show that every phase inherits reflection positivity, derivatives and stress laws, the one-cap smoothness bound, separability and the operator bound. B15–B18 supply estimates on cutoff families (cutoff trains with bounded moments, polynomial limits, convergence of character moments, corrected detectors) used to move cluster operators between the cutoff and the limit; they sit at the boundary of Parts IV and V. All are assembled by Bid.Looms into Phase.Loom, hence into Materializes (via Phase.Loom.cleanse); B15–B18 first pass through Bid.episode and Bid.fees. Internal chain: B04 → B05; B06 → B07, B09; B15–B18 are used jointly.

Family C: Osterwalder–Schrader to Wightman reconstruction (Part V)

C01–C29, 29 items, 6,193 lines. This family proves analytic continuation of the Schwinger functions to Wightman distributions with Poincaré covariance and locality, for any table with the Osterwalder–Schrader data. C01–C03 construct the Osterwalder–Schrader vectors and the strict-time continuation Waves; C04–C06 give its Cauchy–Riemann identities, the absolute-frequency identity and the support in the cone; C07–C09 transfer the Ward boost identities to the real Poincaré action on the boundary values (C08: the Poincaré push on joint Schwartz test functions; C09 combines A16 and C08); C10–C16 construct the open-tube emission (cone thickening, damping, Aurora, matching with Waves, Lorentz covariance); C17–C21 glue the two tubes at adjacent spacelike pairs by imaginary boosts (the step of Bargmann–Hall–Wightman type), which establishes the input for local commutativity; C22–C24 pass from the Euclidean Gram forms to real-time operators; C25–C29 give the tensor density, nulls, locality (C27), strong continuity and word insertion. The statements are collected by the structure Bid.Arcs (A16 and C01–C29), which is proved outright as Integrated.reconstructionArcs. Internal dependencies: C09 on C08 and A16; C14 on C10; C26 on C25; C28 on C08.

Family A: stress tensor, Ward identity and rotations (Part VI)

A01–A17, 17 items, 5,022 lines. This family turns the Euclidean limit table into one carrying a conserved stress tensor and an action of O(4), and then of the Lorentz group, on the labels. It has four layers. (i) A01–A03 and A05–A06 take the Ward premises (Climb.Pre: the shell maps p and spin rows Y that appear when the angular weight of a cutoff is differentiated) and prove their algebra: compatibility with caps, the unit, mirrors, shell commutation and transversality; A04–A06 also identify their free-Maxwell scaling limits. (ii) A07–A09 show that the limit rows detect the internal Ward kernel, that the differentiated Ward identity gives the so(4) brackets, and a pure Lie-algebra lemma. (iii) A10–A15 integrate the brackets to a group action (Port, with O(4) generated by axis reflections, A13), prove the rotation laws, preserve the Casimir polynomial (A11, A12), and average the stress tensor to a conserved tensorial J' (A15). (iv) A16–A17 do the same for boosts. The statements are assembled by Bid.station, which builds the Org.Twist.Commission (A01–A08, A11, A13–A17 and A09, A10), and by Bid.masts, which builds MaxwellData (A09, A10, A12). Dependencies: A06 needs A04; A11 and A15 need A10; A14 needs A13; A17 needs A16.

Family E: energy and pressure estimates for a renormalization-group step (Part III)

E01–E09, 9 items, 3,370 lines. These give one step of the renormalization group an energy bound per seed. E01 packs macro charges; E02 compares the fine link with the tree predictor; E03–E04 bound the row, column and lane sums of the block kernel; E05 gives coarse-row lower bounds in tubular charts; E06–E08 bound the F/I rows, the departures from affected cones and the chord size on unaffected cones; E09 assembles them, with the coarse-row lower bound Belly as a remaining input (Bid.EBare), into Pressure.Teller.Grants. Order: E03 → E04 → E06; E01, E03, E04 → E07; E01, E02 → E08; E09 last. The assembling statement is Bid.issues, proved as Integrated.pressure_grants; EBare is the one input not supplied by the family.

Family F: stationary point and Hessian of the renormalization-group step (Part III)

F01–F09, 9 items, 5,259 lines. F01–F05 (with E02) describe the saddle point of the block-spin integral: the tree-slice sensitivity (F01), the analytic uncut main term (F02), the canonical Hessian on the tree slice (F03), the critical branch and its quadratic value (F04), and independence of the choice of gauge-fixing tree (F05). F06–F09 construct the positive quadratic-form data of the stack of towers: repair of the positivity margin on the three-form row (F06), square roots of positive spectral symbols by Laurent series (F07), the stack at uniform widths (F08), and the period-by-period algebra (F09). Assembling statements: Bid.cospons (E02, F01–F04) gives Scene, used by Integrated.stationary_scene; Bid.one_for_all (F05) transfers station matching between trees; Bid.pinning (F06–F09) pins the horizon data.

Family G: label polynomials (Part III)

G01–G09, 9 items, 2,865 lines. These show that the renormalized lattice observables (labels) are given by invariant polynomials in the local curvature slots, with controlled cost. G01–G02 give gauge and signed-permutation covariance and analytic recovery of link holonomy from radial connection data (Peg); G03–G04 give the linearized plaquette sampler and its equivariant splitting; G05 builds the real invariant slot polynomials keeping the lower degrees; G06 averages them to an equivariant section; G07 bounds the polynomial cost. Bid.teaching assembles G01–G07 into Bid.Learns (Integrated.source_sections). G08–G09 are different: they propagate cost bounds along renormalization-group histories between matched trajectories, and feed the ultraviolet extension rather than Bid.

Family H: fusion of forced activities (Part III)

H01–H04, 4 items, 847 lines. Combinatorial identities for the activity expansion at the start of a renormalization-group step: the gas identity for fused activities (H01), the budget of the merger (H02), the forced-family and move receipts (H03), and continuity in the parameter at a fixed period (H04). They are consumed together by Bid.Oust.begins.

Family I: marker supports and widths (Part III)

I01–I04, 4 items, 460 lines. Bookkeeping for the markers of the cluster expansion: support price by covering number (I01), resale of budgets to price tagged weights (I02, using I01), the geometry of the centred link tiling (I03) and stop-width bookkeeping (I04). Bid.installs assembles I01–I04 into Input.Metered.

All work items

Lines are the file together with its auxiliary modules. "Builds on" lists the other work items used in the proof or named in the statement. The last column is the reading of the description.

Family A (Part VI): stress tensor, Ward identity, rotations and boosts

id part what it proves lines builds on reading
A01 VI Ward premises (Pre) on the limit table imply the two raw shell maps are compatible with the label structure: the translation/colour map p commutes with inclusions of caps, kills the unit label and is covariant under signed axis permutations (mirrors); the spin rows Y are real, inclusion-compatible, kill the unit and are mirror-covariant. 675 – certain
A02 VI Moving the centre of the angular weight inside a fixed cutoff gives commutation relations between the derivative on labels and the two shell maps: Y(d+1)(v)(deriv u) = deriv(Y v u) + sum_c v_c p(u), and deriv(p u) = p(deriv u) (ShellCom). 309 – certain
A03 VI All positional jets of the transverse-component vector Ward tests: given the colour law, the shell commutation and the line-sum identity, both p and the label derivative are transverse (annihilate the longitudinal direction). 204 – probable
A04 VI For the stress tensor J and free-Maxwell scaling systems X, the Ward test written with the derivative of the label (ConLap: conservation versus Laplacian of the rescaled stress read-out) holds along every scaling system; the limit identity is the one for the free Maxwell field. 184 – probable
A05 VI Using finitely many free-slot tests, along every scaling system the rescaled p converges to kappa times the Maxwell label derivative and the rescaled spin rows converge to kappa times the Maxwell Lie action on spin; both shell maps are recovered from the free-Maxwell comparison. 617 – probable
A06 VI Transport of the finite collinear Ward test matrix: along scaling systems the aligned rows converge at fixed free coordinates, and in the limit the transverse leaves (derivative and p) vanish on the test aisle. 258 A04 probable
A07 VI Internal Ward kernel detection: given the action law tying the stress to the Maxwell data (ActLaw), the limit rows separate points of the internal finite-dimensional space (erases), so a spin-row defect that vanishes on all separating rows is zero. 291 – probable
A08 VI Lie brackets after Ward propagation: the differentiated Ward identity (a directional derivative equals minus a spin substitution sum), together with erases, forces the spin rows Y to satisfy the so(4) bracket relations (Brackets Y). 249 – certain
A09 VI Algebra lemma: any bracket-satisfying family of spin rows obeys the relations rel and the pairwise commutation of the associated wave polynomials (Lie/polynomial lemma used for Maxwell data). 91 – probable
A10 VI For a rotation action R on the table integrating the spin rows (Port), the Lie map satisfies the brackets and is equivariant: R.act A (lie g v) = lie (A g A^-1) (act A v). 117 A09 certain
A11 VI (Assuming A10.) Reality and signed subgroup: transporting the Maxwell wave polynomial by a real change of basis (signed permutation) keeps it in the preserved heat-rank space. 207 A10 probable
A12 VI Limiting Casimir selection polynomial: at every degree cap there is a real polynomial p in the Lie generators with Works p (selects the limiting Casimir), for a Port rotation action. 338 – probable
A13 VI The orthogonal group O(4) is generated by the coordinate-axis reflections and rotations (AxisGens = top): every hyperplane reflection is conjugate to a coordinate mirror by rotations. 202 – certain
A14 VI (Assuming A13.) Continuation from Port: if the configuration-space Ward identity holds at all separated points, then the integrated rotation action U satisfies the rotation laws for the derivative plan and is oriented, i.e. the table is O(4)-covariant at label level. 422 A13 probable
A15 VI (Assuming A10.) Averaging the stress over the rotation group (given the Crystal premise lie = c^-1 Y) yields a stress tensor J' that is conserved, O(4)-tensorial and satisfies the stress laws. 653 A10 certain
A16 VI The proper orthochronous Lorentz group is generated by spatial rotations and coordinate boosts (beamGens = top); boosts suffice to bring a future unit timelike vector to rest. 103 – certain
A17 VI (Assuming A16.) Boost covariance of the vacuum: for the Ward data Tab and the transported actions at each cap, which are representations, the vacuum label space is invariant under boosts (wand), by ODE uniqueness along boosts. 102 A16 probable

Family B (Part V): phase decomposition, positivity in each phase, cutoff estimates

id part what it proves lines builds on reading
B01 V For the unselected (mixed) theory and a time axis, there is a time-slab operator family (a Spread of word operators on the OS Hilbert space) with the all-copy property, from a width supplied by the chessboard bench. 72 – probable
B02 V Operator-norm laws of that Spread: derivatives in the positions are controlled uniformly by bounded operator derivatives (A.Laws P). 619 – probable
B03 V On separated words, the compressed operators M are permutation invariant, respect unit labels and are linear in each slot, i.e. depend on words only through the all-spectator field quotient (Theme). 417 – probable
B04 V Grid passage: for generic grids, the lattice expectation of a point word along the ordered limit converges to the continuum point limit (Grid.Expects). 187 – certain
B05 V (Assuming B04.) The mixed theory is reflection invariant in the phase sense (Phase.reflects): the time-slab operators commute with time reflection. 116 B04 probable
B06 V (With phase reflection.) Individual invariance of the compressed operators M under every orthogonal transformation, in particular all reflections and time-space rotations. 244 – probable
B07 V (Assuming B06.) Existence of the phase decomposition: a family of characters (Garland), i.e. points of the Gelfand spectrum of the commutative algebra of compressed vacuum operators, indexing phase tables p(i). 125 B06 certain
B08 V Every phase table p(i) is reflection positive, including smeared reflection positivity and for all axes. 210 – certain
B09 V Every phase table inherits the label derivatives, inclusions and the stress-tensor laws of the mixed table. 516 – certain
B10 V Fixed-arity, fixed-label bounds on all derivatives (Capped) follow from the one-cap statement (OneCap) for a table respecting the derivative plan. 107 – probable
B11 V Each phase table p(i) satisfies the one-cap derivative bound. 91 – probable
B12 V The compression algebra of the Spread is norm separable (a countable dense sequence in the algebra). 93 – certain
B13 V Operator bound in a character chamber: for each phase chamber there is an all-copy Spread, because a train estimate on the unselected vacuum space controls the character chamber. 106 – probable
B14 V There is one countable list of future words whose span is dense in every phase chamber (Cluster.horizon), so countably many tests select the ground state. 149 – probable
B15 V Translated cutoff trains obey ‖E.fromCut‖ <= (g+eps)^(2n) at late cutoffs, under a centred episode budget g (Sold). 131 – probable
B16 V For fixed train and degree, the finite polynomial expectation along cutoffs converges to the limit expression wanted (finite polynomial limits). 195 – probable
B17 V Character moment limit: the limiting train expectation equals <Omega, ((cluster)^* cluster)^n Omega> in the OS Hilbert space (moment convergence to the vacuum moments of the cluster operator). 91 – probable
B18 V Episodes for corrected detectors: polynomial detectors encoded as episodes send the cluster-end quantity to the cutoff-end quantity (Sends), for both the one-step and two-step differences. 134 – probable

Family C (Part V): Osterwalder–Schrader to Wightman reconstruction

id part what it proves lines builds on reading
C01 V Centred insertions on the common label union give a Small family of vectors in the OS Hilbert space and the Piece property (all-spectator matrix elements). 119 – probable
C02 V (Assuming C01.) The Small family has the slicing property (Slice): one strict-time width can be used for the whole multilinear form. 179 – probable
C03 V Strict-time analytic continuation: there is a family of continuation functions Waves (holomorphic in complex times) directly computed from the OS Gram forms. 260 – certain
C04 V The continuation has Cauchy-Riemann identities at strictly ordered imaginary time heights, with Cauchy bounds on slices and integration by parts in the time line. 224 – probable
C05 V Absolute-frequency identity on compact Schwartz tests: pairing the continuation with the exponentially weighted frequency tests gives the boundary distribution (during). 241 – probable
C06 V Half-space (spectral cone) support: the boundary distribution of the continuation has frequency support in the cone (cone_tests). 159 – probable
C07 V Temporal PDE on joint Schwartz functions: Ward boosts on strict-height imaginary times give the infinitesimal boost identity for the continuation (Boosty). 673 – probable
C08 V The Poincare group acts continuously on joint Schwartz tests (strong topology) and the boost generator is the derivative of the push-forward. 288 – certain
C09 V (Assuming A16, C08.) From Ward boosts on imaginary times to real Poincare motion: the boundary Wightman distributions are translation and Lorentz invariant with the label rotation action (Urs). 443 A16, C08 certain
C10 V The spectral cone is separated by relative gaps (Wide): absolute-coordinate tube dual positivity. 72 – probable
C11 V Existence of a smoothed thickened-cone cutoff (Blanket) with a compactly supported unit-integral bump. 102 – probable
C12 V Damping on the closed shift set: the exponential lens is controlled in every Schwartz seminorm near the cone (Modified). 183 – probable
C13 V Analytic emission in the open tube: there is an emitted function (Aurora) holomorphic on the whole open tube, given the cutoffs. 456 – certain
C14 V Identification of the strictly damped contour as a function (Income): the emission is a genuine function at all real coordinates. 262 C10 probable
C15 V The emission matches the continuation Waves at ordered complex times (FullBody). 133 – probable
C16 V Real Lorentz covariance inside the tube for the emission (wave_turn). 93 – probable
C17 V Small imaginary boosts at an adjacent spacelike pair: heights tending to zero along which a boost path (travelable) connects the two tubes. 286 – probable
C18 V (Assuming Descends.) The boost action on labels extends analytically to complex rapidity (Weighty). 80 – probable
C19 V Germs through each tube from real Lorentz covariance and the boost path: both tube germs (Panes) exist at an adjacent spacelike pair. 228 – probable
C20 V The two germs agree at a Euclidean point (sameGround), by jet agreement on a totally real slice and analyticity. 309 – certain
C21 V Local realization: both partial boundary values are realized as partial distributions on a neighbourhood (Panes.Forms). 88 – probable
C22 V Inner products at separate heated profiles: Gram forms of the OS vectors continue to the Wightman pairing (Hull). 355 – probable
C23 V Fixed-height integrations on the chamber give the real-time operator structure (Roof). 221 – probable
C24 V From the mixed point kernel to Gram convergence of independently heated words (Lamp). 155 – probable
C25 V Tensor density: the span of tensor test functions is dense in joint Schwartz space for each arity. 139 – certain
C26 V Boundary word nulls are Euclidean label nulls: the Wightman functions do not kill more words than the Euclidean theory (Persists). 94 C25 probable
C27 V Local commutativity: if the boundary distribution is antisymmetric-free under exchange on spacelike gaps (Actor) then exchange of causally separated test functions leaves it unchanged (Regional). 94 – probable
C28 V Simultaneous covariance on joint tests gives continuous word motion: Strong and Transforms for the boundary distributions. 145 C08 probable
C29 V Word insertion in joint tempered distributions: the insertion map is a distribution (Symbol) and the kernel is jointly continuous in test functions (Cam). 112 – probable

Family D (Part IV): extraction of the infinite-volume limit

id part what it proves lines builds on reading
D01 IV Top coefficient budget: for the Wilson family, the product-of-columns estimate bounds the pullback of the reference label at a fixed degree, on all periods. 136 – probable
D02 IV From a gain on separated tuples to vanishing smeared discrepancy: if the discrepancy tends to 0 at separated lattice tuples then the smeared discrepancy tends to 0 along the ordered extraction. 85 – probable
D03 IV One-slot forward differences in the ordered extraction: the smeared difference quotient of the lattice correlation converges to minus the derivative of the limit kernel. 250 – probable
D04 IV (Assuming D01-D03.) Transfer from match rents to the all-shot monitor: the limit table is a spanning limit of the cutoff tables (Spans). 242 D01, D02, D03 probable
D05 IV The centred estimate gives a one-cap derivative bound (OneCap) for the limit table. 182 – probable
D06 IV Compatibility of the limit kernels' differential system with the real and signed (reflection) symmetries (orientation of the derivative plan). 140 – probable

Family E (Part III): energy and pressure estimates for one renormalization-group step

id part what it proves lines builds on reading
E01 III Macro charges and packing estimates on a single oriented sector of the RG step: for every delta > 0 there is a bound N on payments so that every call serves charges and macro-store bounds. 339 – tentative
E02 III Fine-link shell comparison: closing a fine link by the two paths of a given tree and lifting to the cover gives a bound Cops (constant C) comparing the predictor with the fine field in the macro gauge. 246 – probable
E03 III Row and column estimates for the physical kernels cut on the cover: bounded row/column sums of the block-spin kernel including transport and torsion multiplicity (Allowance). 258 – probable
E04 III (Assuming E03.) Localized absolute path series: the lane counts, selector and transpose budgets hold separately in each orientation (LaneCounts, Maximum). 308 E03 probable
E05 III Coarse row lower bounds in tubular charts: there is delta > 0 so that the squared chord deficit prices only the normal displacement (CoverStore). 229 – probable
E06 III (Assuming E03, E04.) Lower bounds for the F and I rows in a plot: the fine rows pay for the listed fine plaquettes (Seals). 442 E03, E04 probable
E07 III (Assuming E01, E03, E04.) Bound on all affected M departures (Lapse). 736 E01, E03, E04 tentative
E08 III (Assuming E01, E02.) Chord size on unaffected cones at post-L costs (Stitches). 346 E01, E02 tentative
E09 III (Assuming E01.) Assembly: given Seals, CoverStore, Lapse and the coarse-row bound Belly, the oriented RG-step estimator grants the energy bound per seed (Grants). 466 E01 probable

Family F (Part III): stationary point, Hessian, positivity margin

id part what it proves lines builds on reading
F01 III (Assuming E02.) Fixed-L gap in real counting coordinates: sensitivity of the minimizing predictor to the coboundary, giving the Comptroller estimate on a slice. 313 E02 tentative
F02 III Analytic uncut main term: the main term is analytic along analytic matrix inputs, on a full complex ball (Fires). 486 – tentative
F03 III Transport of the canonical Hessian onto the tree slice: the tree-slice quadratic form equals the pulled-back Hessian. 878 – tentative
F04 III Fixed-period critical branch and tree-slice quadratic value: existence of the critical branch and its quadratic value for the transformed action. 408 – tentative
F05 III The critical value and Hessian agree between two gauge-fixing trees up to a chart ratio factor (Critical, PriceGap), so station matching for one tree transfers to any other. 1,053 – probable
F06 III Repair of the all-phase positivity margin: a positive margin on the three-form row can be enlarged by adding the Gram of the outgoing coboundary, with pre-L constants. 804 – probable
F07 III Positive spectral margin implies root rows: a positive symbol in a strip has a Laurent-series square root in a narrower strip (SqActs). 535 – probable
F08 III Stack construction at uniform widths: given donated rows, a stack of towers (Towers) with deep widths exists replacing the old bed. 630 – tentative
F09 III Real period-by-period algebra of the stack: the stack identities and defect factorization give the Roll identity, with no spectral estimates. 152 – tentative

Family G (Part III): label polynomials

id part what it proves lines builds on reading
G01 III Gauge covariance and signed-permutation compatibility of the holonomy of the radial connection: the Taylor data transform covariantly under constant gauge action and axis permutations. 166 – probable
G02 III (Assuming G01.) Real-analytic recovery of link transport in truncated radial connection data (Peg): holonomy is an analytic function of the jet. 280 G01 probable
G03 III Linearized plaquette sampler: finitely many fundamental loops (lifts) recover the degree-d jet of the curvature. 242 – probable
G04 III Splitting of the sampler linearization is equivariant (Equivariant). 141 – probable
G05 III Real invariant slot polynomials with lower degrees retained: the recovery map's d-jet is a polynomial in the label slots (Cumuli). 409 – probable
G06 III Reynolds averaging turns Cumuli into an equivariant polynomial section (SectionData, pop). 608 – probable
G07 III The sections have polynomial cost in the slot (Credit): t^j cost at fixed finite order, via link derivative bounds. 524 – probable
G08 III From the clocks of a RG history to the reference readout: a one-step cone estimate gives a transfer datum TransferData of the section against the holonomy. 260 – tentative
G09 III Propagation of costs for a matched pair of trajectories (Duet): coordinate difference is small relative to the remainder budget along the RG levels, by cones started at the coarsest index. 235 – tentative

Family H (Part III): fusion of forced activities

id part what it proves lines builds on reading
H01 III Fusion gas identity: the fused activities of the forced gathering equal the original gas times the exponential of the sum of the logarithms of the tap. 268 – tentative
H02 III Budget for the forced startup merger: the fused activity obeys the K-bound with price v. 81 – tentative
H03 III Forced-family and Moves receipts for fusion at fixed budget point P: the fused system is a forced family and has the face moves. 336 – tentative
H04 III Fixed-period continuity of the forced gathering along a parameter curve (Curve). 162 – tentative

Family I (Part III): marker supports and widths

id part what it proves lines builds on reading
I01 III Support price for separated markers: B * card F <= 64 * coverNumber for a B^2-separated set of markers. 84 – probable
I02 III (Assuming I01.) Resale from mixed and grain budgets: all tagged marker weights are priced, Radial bound for the marker bands. 182 I01 tentative
I03 III Geometry for the centred link tiling: alignment and pairwise spacing of the markers at stop level to cutoff. 64 – tentative
I04 III Stop-width bookkeeping: positive widths and sublinear column product bound for the schedule of the given Family. 130 – tentative