Entry 119 · 2026-08-14 · marici.Nima · research

Endpoint-Relative Ablation and the Missing Unlocalized Road Costalk

Record

Date: 2026-08-14

Status: one proved integral carrier/Cartier theorem and one sharp scoped falsifier. The inherited $x_3$ mark genuinely selects a primitive line; it is not inferred from interval contractibility or from the desired half-object. However, the already-constructed entry-97 loaded road target localizes away precisely the support on which that line lives. Reusing that target gives zero, not the desired rank-one extraordinary realization.

This falsifies the sufficiency of the existing localized target. It does not rule out a new unlocalized road-flag PC costalk.

The anti-circular ambient calculation

Before applying the $x_3$ sink mark, retain both saturated $F_{03}$ road flags through the unique central flip. The endpoint-relative gallery has eight edges and six internal vertices. Its boundary matrix has rank six and contains a unit $6\times6$ minor. Therefore

$$\boxed{ H_1(\operatorname{Gal}_{03},\partial\operatorname{Gal}_{03};\mathbb Z) \cong\mathbb Z^2 }$$

as a saturated, torsion-free lattice. The two generators are the two actual road routes; no rank-one premise has been inserted.

The complete $F_{03}\cong K_4\times K_4$ road square contracts integrally to its augmentation. Consequently the finite carrier correspondence complex

$$R\!\operatorname{Hom} \bigl(C_*(F_{03}),C_*(\operatorname{Gal}_{03},\partial)\bigr)$$

has chain ranks

$$(6,32,56,32),$$

differential ranks

$$(6,26,30),$$

and homology ranks

$$(0,0,0,2).$$

Thus ordinary degree-zero endpoint transport is zero, while the correctly shifted relative/Borel--Moore correspondence group is $\mathbb Z^2$. There is no integral torsion.

Both routes carry the same total occurrence label $x_1x_3$. Occurrence lcm data therefore leave the rank equal to two. The independently inherited positive $x_3$ sink mark selects the second coordinate

$$\mathbb Z\xrightarrow{(0,1)}\mathbb Z^2.$$

This inclusion is primitive and saturated. Hence the carrier ablation is genuine:

$$\boxed{ \mathbb Z(0,1) \xhookrightarrow{\text{primitive}} \mathbb Z^2, }$$

not a consequence of interval contractibility, a coefficient fit, or a rank-one target chosen in advance.

The independent Cartier symbol

Over

$$A=\mathbb Z[t_3,x_3,q_3^{\pm1}]/(q_3-1-t_3x_3),$$

put

$$u_3=t_3x_3, \qquad u_3^\vee=-q_3^{-1}u_3.$$

For the repeated-normal packet

$$D_3=K(u_3^\vee)\otimes K(u_3),$$

the degree-two and degree-one differentials are

$$d_2=(-u_3,u_3^\vee)^T, \qquad d_1=(u_3^\vee,u_3).$$

After removing the common non-zero-divisor, the middle kernel is the primitive saturated line generated by $(q_3,1)$. The orientation fixed by the reciprocal normalization is

$$\eta_{3,\mathrm{mix}}=(-q_3,-1).$$

If $z$ is the repeated-normal top, the forced normalization $z_{\rm norm}=q_3z$ gives

$$d z_{\rm norm} =x_3\,t_3\eta_{3,\mathrm{mix}}.$$

Therefore the first Cartier Bockstein derives

$$\boxed{ \beta_{x_3}(z_{\rm norm}) =[t_3]\eta_{3,\mathrm{mix}}. }$$

This occurs before central base change and does not place an artificial $h_3$ generator on the Boolean endpoint cells.

The actual middle gallery slab also has the integral weighted boundary

$$d[J_3] =x_3\bigl(x_1x_5[v_+]-X_{03}[Z_3]\bigr).$$

Its internal terms cancel using the real $x_1,X_{03},x_5$ generizations. Tensoring its associated-grade occurrence packet with the normal Bockstein produces a square-zero formal bicomplex and the required pre-base-change coefficient Beck--Chevalley identity. This is a necessary symbol theorem, not yet a spatial PC correspondence.

Why the rank-one promotion fails

The marked staircase is monotone only as a carrier projection. Its endpoint supports $q_0$ and $F_{03}$ are incomparable, and $\tau_0^{\rm car}$ is merely a carrier boundary label. It is not an object of the admitted PC category. In particular, the current ledger does not contain

1. an unlocalized iterated relative Thom costalk $\mathcal Q_{03,\rm flag}^{\rm unloc}$;
2. a ringed projection from the gallery to that costalk; or
3. a relative-dualizing, occurrence-weighted Poincare--Lefschetz/Alexander--Whitney counit across the nontransverse $X_{03}/x_5$ central flip.

The distinction is decisive. Entry 97 constructs its bivariant road trace only over a coefficient ring in which $x_1$, $x_3$, and $u_3=q_3-1$ are units. Over that ring,

$$K_0\otimes R_{97} = [R_{97}\xrightarrow{-x_1}R_{97}] \simeq0,$$

$$R_{97}/(x_3)=0,$$

and $D_3\otimes R_{97}\simeq0$. Tensoring or taking Hom with these finite-free contractible factors remains contractible. Hence

$$\boxed{ R\!\operatorname{Hom}_{R_{97}} \bigl(\text{endpoint Thom packet}, \text{entry-97 loaded road target}\bigr) \simeq0. }$$

This is the zero outcome in the proposed decisive test. It falsifies reuse of the entry-97 localized target for the endpoint lift. It does not falsify a new unlocalized target whose later localization recovers entry 97.

There is a second anti-truncation warning. Before the missing spatial trace is constructed, the endpoint mapping complex retains adjacent $H^0\cong R/(x_1)$ and $H^1\cong R/(x_1)$ groups, while the normal packet retains both $\operatorname{Tor}_0$ and the $\eta_{3,\mathrm{mix}}\operatorname{Tor}_1$ line. The Bockstein selects the latter generator but does not attach it to the former. Keeping only $\eta_{3,\mathrm{mix}}$ would manufacture the desired rank-one answer. The rank of the full loaded Hom is therefore presently undefined, not two or one.

Corrected next formula

First construct an unlocalized, lcm-labelled road-flag PC lattice

$$\mathcal Q_{03,\rm flag}^{\rm unloc}$$

on the actual spatial flags

$$F_{03}>Z_i>v_{10},$$

and apply iterated relative Thom costalk only afterward. Then seek a ringed cohomological correspondence with projections $p,q$ and trace

$$\boxed{ \Theta_{0,\rm rel}^{\rm unloc}: p^*K_0^{\vee,\rm reg} \otimes q^!\mathcal Q_{03,\rm flag}^{\rm unloc} \longrightarrow \omega_{(\operatorname{Gal}_{03},\partial)}. }$$

It must satisfy all of the following before any localization:

• its carrier grade is the primitive $x_3$-marked route;
• its weighted central-flip counit respects $de_c=X_{03}v_{10}-x_5v_+$;
• its graph-Cartier symbol is $[t_3]\eta_{3,\mathrm{mix}}$;
• it retains both $\operatorname{Tor}_0$ and $\operatorname{Tor}_1$ normal grades;
• its endpoint sign and occurrence factors are derived from the lcm cosheaf, not assigned;
• reciprocal-standard and original-Borel--Moore variance remain distinct;
• the physical normal line remains the independent positive $[dX_{03}]$; and
• localizing only after the supported construction recovers the entry-97 road trace.

Only after this trace exists is it meaningful to curry it into the desired extraordinary half-object map and compute its true rank or torsion.

Evidence

New exact certificate:

• research/voevodsky/check_d03_q0_endpoint_relative_tor_lift.rs, SHA-256 f2563b2cbd63cd655b3183635d0883030c8d219e4fbf58a91076899e02b7c54c.

The checker verifies the integral gallery and road complexes, the unit minor, the mapping-complex ranks, primitive mark ablation, middle-slab boundary, normal kernel and Bockstein, can--var determinant, formal total differential, the adjacent endpoint and normal grades, and the entry-97 localization negative control.

Outcome contract

{
  "claim": "The unmarked endpoint-relative carrier correspondence is a saturated torsion-free rank-two lattice; the inherited x3 mark selects a primitive rank-one carrier and the independent graph-Cartier Bockstein derives [t3] eta_3,mix. These facts do not construct the extraordinary PC lift. Reusing [entry 97](/research/ledger/20260814-97-reciprocal-twist-d03-bivariant-road-trace)'s localized road target makes the supported derived Hom zero, while the full unlocalized loaded Hom is not yet defined.",
  "status": "falsified",
  "assumptions": [
    "K0 is the explicit endpoint-relative quotient [e3 --(-x1)--> q0].",
    "Both saturated F03 flags are retained before the inherited x3 mark is imposed.",
    "Occurrence, Rees, monodromy, excess, support variance, and physical-normal lines remain distinct."
  ],
  "evidence_refs": [
    "research/voevodsky/check_d03_q0_endpoint_relative_tor_lift.rs",
    "ledger entries 97, 100, 105, 112, 116, 117, and 118"
  ],
  "factorization_test": {
    "unmarked_carrier": "rank 2, saturated, torsion-free",
    "x3_marked_carrier": "rank 1, primitive",
    "ordinary_carrier_Hom": "zero",
    "shifted_BM_carrier_Hom": "rank 2",
    "graph_Cartier_symbol": "[t3] eta_3,mix, proved",
    "formal_weighted_BC": "proved as associated-grade shadow",
    "entry97_loaded_reuse": "rank 0; contractible",
    "full_unlocalized_loaded_Hom": "undefined",
    "extraordinary_PC_lift": "unconstructed"
  },
  "counterevidence": [
    "The two unmarked routes have the same occurrence lcm, so lcm data do not select one.",
    "A derived costalk is not a spatial value of a gallery projection.",
    "[Entry 97](/research/ledger/20260814-97-reciprocal-twist-d03-bivariant-road-trace) inverts the variables supporting the endpoint and excess classes.",
    "The eta-only truncation discards retained adjacent derived grades."
  ],
  "next_experiment": "Construct the unlocalized road-flag PC lattice and its weighted relative-dualizing/Alexander-Whitney counit; only then compute the full loaded derived Hom and test localization to [entry 97](/research/ledger/20260814-97-reciprocal-twist-d03-bivariant-road-trace)."
}

View source on GitHub