Saturated Six-Point Tripods and the Subdivision–PC Gap
Record
Date: 2026-08-14
Status: exact six-point subdivision correction and a narrowed worldsheet gap. The direct rule which sends every barycentric simplex to the Pochhammer/Cousin summand of its maximal face is not a chain map on all of (C_*^{\rm simp}(\operatorname{sd}K_6)) as presently defined in entry 38. Barycentric edges may skip face dimensions, whereas the established Cousin differential uses only codimension-one face inclusions.
The entry-21 tripod contains exactly six such jumps. Each has a unique rational, (D_6)-equivariant saturated replacement: the half-sum of the two vertex–edge–facet flags in the relevant square physical facet. The repaired tripods have the exact required boundary and use only codimension-one tangential/Cousin incidences. Their physical residue-freeness is not yet proved. Consequently the proposed complete eight-point PC polarity primitive remains conditional on a six-point occurrence/Gysin residue theorem, even though the coefficient-level (G/R/K) decomposition has no additional unmarked remainder.
Why the naive top-stratum rule fails
Let
[ \sigma=[b(F_0),\ldots,b(F_k)] ]
be an oriented simplex of the barycentric subdivision, with
[ F_0<\cdots<F_k. ]
Its relative interior lies in (F_k^\circ). In the simplicial boundary
[ \partial\sigma
\sum_{j=0}^{k}(-1)^j [b(F_0),\ldots,\widehat{b(F_j)},\ldots,b(F_k)], ]
all terms with (j<k) retain top stratum (F_k) and are tangential. The term (j=k) has top stratum (F_{k-1}) and must be the Cousin contribution.
This is compatible with entry 38 when
[ \operatorname{codim}{F_k}F{k-1}=1. ]
It is not compatible for a nonsaturated flag. Entry 38 defines (d_{\rm Cousin}) as the signed sum over codimension-one face inclusions; it contains no direct map which jumps two or more strata. Assigning the last simplicial face directly to the deeper stratum would silently add a higher exit-path specialization not present in the established PC complex.
The six-point tripod makes this issue concrete. Its last segment is
[ [T_i,b(F_i)], ]
where (T_i) is a vertex and (F_i) is a square facet. The face-dimension jump is two. Therefore the direct maximal-stratum rule does not yet map the entry-21 tripod to entry 38’s PC complex.
The saturated tripod replacement
Fix a corner (T<F) of a square facet. There are exactly two square edges
[ T<E_0<F, \qquad T<E_1<F. ]
Define
[ \boxed{ \lambda_{T,F}
\frac12 \sum_{a=0}^{1} \left( [b(T),b(E_a)] + [b(E_a),b(F)] \right). } ]
Each summand is a sequence of cover relations. Its boundary is
[ b(F)-b(T), ]
so the same is true of (\lambda_{T,F}).
The coefficient is forced. The reflection fixing the corner/facet pair exchanges (E_0) and (E_1). If their path weights are (a,b), equivariance gives (a=b), while the endpoint boundary gives (a+b=1). Hence
[ \boxed{a=b=\frac12.} ]
There is no integral equivariant choice on the downstairs tripod. The doubled chain is integral, and the rational half-sum is canonical over the characteristic-zero nonresonant PC coefficient field.
The replacement is explicitly homotopic to the original barycentric jump. If
[ \tau_a=[b(T),b(E_a),b(F)], \qquad j=[b(T),b(F)], ]
then
[ \partial\tau_a
\bigl([b(T),b(E_a)]+[b(E_a),b(F)]\bigr)-j. ]
Consequently
[ \boxed{ \lambda_{T,F}-j
\partial\left(\frac12(\tau_0+\tau_1)\right). } ]
The difference of the two integral saturated routes is likewise the boundary of
[ \tau_0-\tau_1. ]
Thus the half-sum selects a symmetric representative of an already explicit homotopy class. A strict zero statement after a tubular/current realization would not be invariant under this representative freedom; the natural derived target is a specified null-homotopy.
Replace the last segment of every entry-21 leg by (\lambda_{T_i,F_i}). Denote the resulting leg by (\widetilde\gamma_i^\epsilon), and set
[ \widetilde\eta_6^\epsilon
\sum_i c_i\widetilde\gamma_i^\epsilon, \qquad \sum_i c_i=0. ]
Then
[ \boxed{ \partial\widetilde\eta_6^\epsilon
\sum_i c_i b(F_i)
q_{6,+}-q_{6,-}. } ]
One-step rotation exchanges the two parity centers. The saturated half-sum is equivariant under all twelve elements of (D_6), including the reflections which force the averaging.
PC typing on saturated edges
For a cover relation (F_0\prec F_1), orient the barycentric edge from (b(F_0)) to (b(F_1)). Its boundary is
[ \partial[b(F_0),b(F_1)]
b(F_1)-b(F_0). ]
The first term is tangential in (F_1^\circ). The second is the unique codimension-one Cousin face. If (e) is the new normal divisor, entry 38 supplies
[ \partial_{\mathscr L}\ell_e=(q_e-1)p_e, \qquad h_e=\frac{\ell_e}{q_e-1}. ]
Together with
[ \operatorname{or}(N_{F_0}) \simeq \operatorname{or}(N_e)\wedge\operatorname{or}(N_{F_1}), ]
this fixes the Cousin sign and lower Pochhammer term. Reordering two normal steps gives the ordinary Koszul sign. Thus entry 38’s local normal-crossing construction applies termwise to every edge of the saturated tripod without a cellular collapse or a higher-stratum jump.
This proves the boundary identity for the saturated PC image, up to the same filtered chain-homotopy strength as the facewise normal-cone construction:
[ \boxed{ d_{\rm PC}, \mathbb P^{\rm sat}_{\alpha’} (\widetilde\eta_6^\epsilon)
\mathbb P^{\rm sat}{\alpha’} (q{6,+}-q_{6,-}). } ]
It does not define a direct map on every nonsaturated simplex of (C_*^{\rm simp}(\operatorname{sd}K_6)). A full such map requires a flag/exit-path PC resolution and a filtered quasi-isomorphism from that resolution to entry 38’s codimension-one PC complex.
The residue test remains open
Every repaired leg has four nonzero saturated edge terms supported inside its physical square facet. Therefore residue-freeness does not follow from the simplicial boundary or from support. One must compute the occurrence-decorated Gysin image, including the term where the leg first enters the physical facet and the lower normal-Koszul terms.
Entry 85 sharpens the required statement. Literal chain-level vanishing depends on the chosen flag/collar representative. The invariant requirement is
[ \boxed{ \operatorname{Res}^{\rm PC}{D} \mathbb P^{\rm sat}{\alpha’} (\widetilde\eta_6^\epsilon) =d_{\rm PC}s_D^\epsilon, \qquad \text{equivalently }[\operatorname{Res}^{\rm PC}{D} \mathbb P^{\rm sat}{\alpha’}(\widetilde\eta_6^\epsilon)]=0. } ]
Entries 32 and 37 prove strict physical coaction and transverse scalar base change. They do not directly prove this formula: the first tripod segment entering (F_D) changes the physical core, and is not an independent transverse scalar-refinement edge. The finite audit therefore does not encode the residue as zero.
Consequence for the eight-point decomposition
Entry 23 proves the exhaustive pole-grade decomposition
[ q_Q^\epsilon
G_Q+ \sum_{D\in Q}R_{Q,D}^\epsilon +K_Q^\epsilon. ]
It follows that there is no fourth, unmarked coefficient sector in the eight-point polarity difference:
- (G_Q) is polarity independent;
- the (R)-difference is supported entirely on the eight physical factorization triangles;
- the (K)-difference is the marked contact boundary closed by entries 24 and 83.
Formally the desired primitive is
[ H_8^{\rm PC}
\sum_D G_D^{\rm PC} (\widetilde\eta_6^{\rm PC}) + H_{\rm ct}^{\rm PC}. ]
Its coefficient-level boundary is exactly
[ \sum_Q(q_Q^+-q_Q^-). ]
The marked second term is already a residue-free PC chain. The first term becomes a proved PC factorization primitive only after the displayed six-point residue class is shown to vanish. Accordingly:
The proposed unmarked octagonal remainder is absent from the exhaustive polarity comparison, but the complete PC polarity homotopy is still conditional on the six-point residue map.
This does not settle the stronger problem of gluing the primitive local half-lines (g_Q), nor does it construct a global Jordan-valued square or octagonal higher-coherence cell. The bare Möbius (H_1) may still govern that distinct atlas problem.
Reproducible certificate
Run:
rustfmt --check research/nima/check_six_point_subdivision_pc.rs
rustc --edition=2021 -D warnings -O research/nima/check_six_point_subdivision_pc.rs -o "$env:TEMP\\marici-six-subdivision-pc.exe"
& "$env:TEMP\\marici-six-subdivision-pc.exe"
The checker verifies:
- all fourteen hexagon triangulations;
- both parity centers and all three square physical facets;
- all six original codimension-two tripod jumps;
- the two saturated flags at every corner/facet pair;
- uniqueness of the rational weights (1/2,1/2);
- the exact tripod boundary for a general sum-zero coefficient vector;
- covariance under all twelve (D_6) elements;
- nonzero physical-facet support of every saturated tail.
Certificate SHA-256:
46021191c34034bf4cd64f5f80e6fe9f0fb39316f86b263fdfeaae9785a310d4
Decision
Reject:
Every barycentric simplex maps directly to entry 38’s PC complex merely by assigning it to its maximal face.
Promote:
The six-point scalar tripods possess a unique rational, dihedrally equivariant refinement into codimension-one face flags. Entry 38’s local Pochhammer/Cousin construction therefore gives them a correctly typed saturated PC boundary.
Retain as the immediate bounded frontier:
Compute the occurrence-decorated physical residue of one saturated tripod leg, including the entry-face Cousin term and its normal Koszul contraction. Rotate the result through the three channels and both sheets. Entry 85 reduces its class to one scalar (\lambda_D^\epsilon[g_4\boxtimes g_4]); its vanishing, or an explicit null-homotopy, closes the full eight-point PC polarity primitive.
Epistemic-graph event:
ev-000000000022-fee2a075-f030-4f1a-875a-5dd2f44bfded
Internal dependencies
- Entry 21: barycentric scalar tripods.
- Entries 23–24: exhaustive (G/R/K) decomposition and marked contact primitive.
- Entries 32 and 37: strict physical coaction and transverse base change.
- Entry 38: facewise PC/Cousin complex and normal Koszul contractions.
- Entries 82–83: local target-first dependent coherence and marked loaded octagon.
research/nima/check_six_point_subdivision_pc.rs.