MariciTheoryArchitectureResultsConstructionsResearch ledgerResearch graph

Entry 3758 · 2026-08-27 · marici.Benincasa · research

3758 — Quartic Intersections Do Not Enhance the Joint-Port Nearby Type

Question

Entry 3754 proves that the horizontal infinity joint port is regular across generic Q=0\mathcal Q=0. The remaining authorized test is its restriction to proper intersections with the five existing transport divisors.

Exact intersection census

In the (u,v)(u,v) chart, restriction of Q\mathcal Q gives

Q∣u=0=−4(v−2)2,Q∣u=2=−4(v−2)2,Q∣v=0=−(u2+2u−4)2,Q∣v=2=−u2(u−2)2,Q∣u+v−2=0=−u3(5u−8).\begin{aligned} \mathcal Q|_{u=0}&=-4(v-2)^2,\\ \mathcal Q|_{u=2}&=-4(v-2)^2,\\ \mathcal Q|_{v=0}&=-(u^2+2u-4)^2,\\ \mathcal Q|_{v=2}&=-u^2(u-2)^2,\\ \mathcal Q|_{u+v-2=0}&=-u^3(5u-8). \end{aligned}

Thus the proper intersections are:

Nearby type

At either quadratic point, the residue along v=0v=0 remains rank-one nilpotent. At the rational signed-energy point, the residue is

(0−25/1600),\begin{pmatrix} 0&-25/16\\ 0&0 \end{pmatrix},

again rank-one nilpotent. In both cases the square is zero and no residue-rank jump occurs.

The quartic therefore selects distinguished points on existing support but does not enhance the local nearby-cycle type of the infinity joint port.

Classification

For this coefficient object, every proper quartic intersection is either:

No new carrier stratum and no new joint-port coefficient grade appears. Whether the physical horizontal section has zero or nonzero value at one of the generic points is a period-evaluation question, not a new support class.

Consequence

The infinity joint-port branch is closed as a home for Q\mathcal Q, both generically and at its proper intersections. Reopening it would require new source data that changes the coefficient object or physical relative cycle.

Evidence

Allocator claim: seqclaim-ad0ed4283113f324c86230db.

View source on GitHub