Derived Modular-Envelope Lift and the Physical Descent Obstruction

Record

Date: 2026-08-13

Status: the resolved tree counit now has a canonical cyclic, all-topology, strictly Cut-monoidal lift on the universal surface-presentation complex. The construction is the derived modular envelope of the tree operation, tensored with the closed-circuit augmentation of entry 46.

This separates two questions which entry 46 had left entangled:

  1. extending the tree counit to resolved surfaces;
  2. descending that extension to a particular already-summed physical surface-function model in the variables (X_C).

The first question is solved by the universal construction below. The second is an exact kernel-inclusion problem and remains open beyond the existing tree, one-loop closed-circuit, and all-loop leading-singularity checks.

Subsequent resolution: entry 48 proves this kernel inclusion for canonical surface functions using the Cut Equation and the cubic-scalar ultraviolet boundary condition. The present entry records the intermediate factorization of the problem and the universal construction on which that proof depends.

Reproducible convention audit:

research/nima/check_modular_envelope_counit.rs

Tree datum to be completed

Let

[ \mathcal E^{\rm YM}_{\rm tree} ]

denote the cyclic resolved tree coefficient object generated by the pairwise trace sectors (U_{ef}), their deletion-simplex homotopies, and the physical amplitude quotient used in entries 42–45. Let

[ \mathbf 1_{\phi,\rm tree} ]

be the corresponding scalar coefficient line. The established tree counit is the augmentation

[ u_0: \mathcal E^{\rm YM}{\rm tree} \longrightarrow \mathbf 1{\phi,\rm tree}. ]

Its relevant properties are:

  • it is a chain map on the deletion resolution;
  • its degree-zero class is independent of the retained pair;
  • cyclic relabelling preserves the class, with the integral representative torsion of entry 44 retained as homotopy data;
  • for every tree channel, [ \Delta_e u_0=u_0\boxtimes u_0 ] in the tensor product of lower physical amplitude quotients.

The state object used in sewing is the resolved Brauer category

[ \operatorname{Br}^{\rm res}_D. ]

Entry 46 constructed its cyclic monoidal augmentation

[ \epsilon_{\rm Br}: \operatorname{Br}^{\rm res}D \otimes{\mathbb Z[D],,D\mapsto1}\mathbb Z \longrightarrow \operatorname{Br}_1. ]

These are exactly the genus-zero and state-trace data required by modular completion.

The derived modular-envelope construction

Let (\mathbb L\operatorname{Mod}) denote the derived modular envelope. It replaces trees by stable graphs while retaining the full cellular resolution of changes of graph, cyclic order, and contraction history. Define

[ \mathfrak S^{\rm YM}_{\rm univ}

\mathbb L\operatorname{Mod} \left( \mathcal E^{\rm YM}_{\rm tree} \otimes \operatorname{Br}^{\rm res}D \right) \otimes{D\mapsto1}\mathbb Z ]

and

[ \mathfrak S^{\phi}_{\rm univ}

\mathbb L\operatorname{Mod} \left( \mathbf 1_{\phi,\rm tree} \otimes \operatorname{Br}_1 \right). ]

Functoriality gives the promised surface operation:

[ \boxed{ u^{\rm univ}

\mathbb L\operatorname{Mod} \left(u_0\otimes\epsilon_{\rm Br}\right): \mathfrak S^{\rm YM}{\rm univ} \longrightarrow \mathfrak S^{\phi}{\rm univ}. } ]

The word derived is essential. A free span of circles with arbitrarily drawn contraction chords is not automatically a modular operad: one of the mixed composition/contraction axioms can fail. The derived modular envelope contains the higher cells imposing precisely these coherences.

Explicit stable-graph formula

Let (\Gamma) be a connected stable ribbon graph, let (a_v) be the resolved tree coefficient at vertex (v), and let (\kappa) be a resolved polarization-contraction cover. Then

[ \boxed{ u^{\rm univ} \left[ \Gamma;\bigotimes_{v\in V(\Gamma)}a_v;\kappa \right]

\epsilon_{\rm Br}(\kappa) \left[ \Gamma;\bigotimes_{v\in V(\Gamma)}u_0(a_v) \right]. } ]

The graph itself is unchanged. Open tree data are mapped vertex by vertex, the orientation line supplies the usual Koszul sign, and each resolved closed polarization circuit is sent from (D) to the scalar loop value one.

This formula applies to every stable graph and therefore to arbitrary genus, number of boundary components, and number of marked boundary intervals. It is not an induction in loop order.

Strict Cut monoidality

Cut an internal edge (e) of (\Gamma) and expose its two half-edges as new flags. There are two cases.

If (e) is a bridge, then

[ \Gamma\setminus e=\Gamma_L\sqcup\Gamma_R ]

and the stable-graph formula gives term by term

[ \boxed{ \Delta_e u^{\rm univ}

\left( u^{\rm univ}{\Gamma_L} \boxtimes u^{\rm univ}{\Gamma_R} \right) \Delta_e. } ]

If (e) is not a bridge, cutting it lowers the first Betti number by one and keeps the graph connected. The two exposed flags retain the same local tree counits, while any state trace opened by the Cut is evaluated by the same Brauer cap. Hence

[ \boxed{ \Delta_e u^{\rm univ}

u^{\rm univ}_{\Gamma\setminus e}\Delta_e. } ]

For several edges, the result is independent of Cut order. This follows from the modular associativity axioms and from the circuit cocycle proved in entry 46. Thus separating and nonseparating Cuts are parts of one graph identity, not two unrelated amplitude checks.

Canonical one-corolla formula

The construction can also be written without choosing a pants decomposition. For a connected oriented surface with genus (g) and (b) boundary components, set

[ G=2g+b-1. ]

Choose a genus-zero cyclic word (a_\Sigma) with (2G) auxiliary flags and a canonical family of pair contractions. The surface operation is

[ \boxed{ u^{\rm univ}_\Sigma

\epsilon_{\rm Br} ,\xi_{e_1e_2}\xi_{e_3e_4}\cdots \xi_{e_{2G-1}e_{2G}} ,u_0(a_\Sigma). } ]

The modular-envelope theorem says this is independent of:

  • the names of the auxiliary flags;
  • the order of the contractions;
  • cyclic basepoints on boundary words;
  • the placement and order of boundary and handle blocks;
  • the chosen stable-graph or pants presentation.

For the chain-level surface complex, the higher ribbon-graph cells implement these equalities as homotopies rather than erasing their provenance.

The one-holed-torus (3S) cell

The vertices of the pants complex for a one-holed torus are unoriented primitive slopes. Take oriented primitive vectors (a,b\in\mathbb Z^2) with

[ \det(a,b)=1 ]

and put

[ c=a+b. ]

Then (a,b,c) are the vertices of a Farey triangle, hence a Hatcher (3S) cell. Choose the three oriented cut-chart frames

[ F_a=(a,b), \qquad F_b=(b,-a), \qquad F_c=(c,-a). ]

Each has determinant one. With

[ T_{j\leftarrow i}=F_j^{-1}F_i, ]

the three-step chart holonomy telescopes exactly:

[ \boxed{ T_{a\leftarrow c} T_{c\leftarrow b} T_{b\leftarrow a} =1. } ]

There is therefore no topological, orientation-line, or resolved-state curvature around (3S). The Rust audit checks this for 308 oriented Farey triangles. This bounded number audits signs and conventions; the displayed matrix identity is the all-slope proof.

The same conclusion is intrinsic to the derived modular envelope: (3S) is a 2-cell, so its boundary is sent to a boundary. Likewise, the mixed (6AS) hexagon is enforced by the mixed modular composition/contraction axioms. Thus neither is an additional axiom for (u^{\rm univ}).

What remains for physical surface functions

Let

[ q_{\rm YM}: \mathfrak S^{\rm YM}{\rm univ} \twoheadrightarrow \mathfrak S^{\rm YM}{X} ]

be the comparison which evaluates a resolved presentation as the chosen YM surface function, and let

[ q_\phi: \mathfrak S^{\phi}{\rm univ} \twoheadrightarrow \mathfrak S^{\phi}{X} ]

be the scalar comparison. Assuming the first map is surjective, a physical surface-function counit (u_X) making

[ q_\phi u^{\rm univ}=u_Xq_{\rm YM} ]

exists if and only if

[ \boxed{ \left.q_\phi u^{\rm univ}\right|{\ker q{\rm YM}}=0. } ]

Equivalently,

[ u^{\rm univ}(\ker q_{\rm YM}) \subseteq \ker q_\phi. ]

This is the exact remaining obstruction. It is narrower and better typed than an unspecified (3S) holonomy: any failure must be caused by a relation introduced by physical surface evaluation, not by surface topology, cyclic order, Cut order, or polarization-state sewing.

Descent on the maximal-Cut quotient

The kernel condition is already satisfied after restricting to leading singularities. Let

[ q^{\rm LS}{\rm YM} \quad\text{and}\quad q^{\rm LS}{\phi} ]

evaluate a resolved stable trivalent graph by on-shell gluing its three-point vertices, respectively in YM and (\operatorname{Tr}\phi^3). On every vertex, the tree counit converts the YM building block to the scalar one. On every internal edge, (\epsilon_{\rm Br}) converts the resolved polarization state sewing to scalar sewing. Therefore, graph by graph,

[ \boxed{ q^{\rm LS}_{\phi},u^{\rm univ}

u^{\rm LS},q^{\rm LS}_{\rm YM}. } ]

Carrôlo and Figueiredo independently identify the on-shell-gluing result for an arbitrary graph with the maximal residue of the gluon surface integral. Before cancellations, every polarization-contraction pattern corresponds with the same sign to a non-overlapping curve cover of the fatgraph; the full sum then agrees as well. At higher topology, maximal residue automatically selects the relevant mapping-class orbit, and every contributing closed curve has the resolved value

[ \nu_\gamma-\Delta_\gamma=D ]

used by entry 46.

Consequently

[ \boxed{ u^{\rm univ} \left(\ker q^{\rm LS}{\rm YM}\right) \subseteq \ker q^{\rm LS}{\phi} } ]

at arbitrary graph topology. The universal counit therefore descends to the physical maximal-Cut/leading-singularity quotient. Any obstruction to the full (X_C)-surface-function descent must have vanishing maximal residues. It is confined to contact completion, total derivatives, hereditary scaleless terms, or other non-maximal-Cut data.

Hatcher’s (3S) and (6AS) cells still give the first local comparison tests when (q_{\rm YM}) is described in cut charts. But they now test whether the published (X_C)-coefficient system realizes the universal modular envelope; they do not test whether the universal lift exists.

Additional generators of (\ker q_{\rm YM}) may come from:

  • total derivatives in a curve-integral presentation;
  • topology-local contact functions invisible to all physical Cuts;
  • hereditary scaleless terms;
  • specializations identifying distinct surface variables by ordinary momentum homology.

Hatcher simple connectivity does not remove these analytic or kinematic relations.

Correction to the previous frontier

Entry 46 correctly identified the resolved Brauer augmentation, but its final wording made (3S) and (6AS) sound like obstructions to the existence of any surface lift. The modular-envelope construction sharpens this:

[ \boxed{ \text{universal resolved surface lift: constructed}, \qquad \text{physical }X_C\text{ descent: open at this stage}. } ]

Also, the published one-loop punctured-disk calculation of entry 46 is a closed-circuit normalization test. It is not a one-holed-torus (3S) test: the planar one-loop surface is an annulus/punctured disk, whereas (3S) is supported on a genuine genus-one subsurface.

Executable evidence

The Rust certificate verifies:

  • 18 canonical boundary/handle presentations through genus two and three boundary components;
  • 156 cyclic presentation squares;
  • 4,793 contraction histories;
  • 308 oriented Farey (3S) triangles;
  • 51,105 Cut/counit squares in connected multigraphs through four vertices;
  • 1,012 separating and 2,637 nonseparating one-edge Cuts, including the exact first-Betti-number alternatives.

These checks are not the proof of the derived modular-envelope theorem. They are executable audits of the combinatorial conventions used in the explicit formula.

Next physical falsification test

Construct the comparison (q_{\rm YM}) on the smallest genus-one surface chart carrying one (3S) cell. If (r_{3S}) denotes the resolved boundary of that cell, compute

[ \mathfrak o_{3S}

q_\phi u^{\rm univ}(r_{3S}). ]

The possible verdicts are now exact:

  1. (\mathfrak o_{3S}=0) strictly in surface kinematics;
  2. it is a total derivative;
  3. it lies in the hereditary scaleless ideal;
  4. it is a nonzero admitted obstruction to physical descent.

If the first three cases hold, repeat the comparison on the mixed (6AS) cell. A successful comparison then leaves only non-topological generators of the physical evaluation kernel.

Primary sources

  • Getzler and Kapranov, Modular operads, for the graph-versus-tree completion: https://arxiv.org/abs/dg-ga/9408003.
  • Costello, The A-infinity operad and the moduli space of curves, for the derived modular envelope and ribbon-graph surface complex: https://arxiv.org/abs/math/0402015.
  • Doubek, The modular envelope of the cyclic operad Ass, for the direct surface theorem, canonical contraction formula, and the warning that naive chord diagrams do not by themselves satisfy all modular axioms: https://arxiv.org/abs/1312.5501.
  • Hatcher, Pants Decompositions of Surfaces, for the (3S), (6AS), and simple-connectivity relations: https://arxiv.org/abs/math/9906084.
  • Backus and Figueiredo, Surface Gauge Invariance, Soft Limits and the Transmutation of Gluons into Scalars, for the tree and one-loop physical transmutation evidence: https://arxiv.org/abs/2505.17179.
  • Carrôlo and Figueiredo, How gluon leading singularities discover curves on surfaces, for graph-by-graph equality of on-shell gluing and surface maximal residues at arbitrary topology: https://arxiv.org/abs/2512.17019.