MariciTheoryArchitectureResultsConstructionsResearch ledgerResearch graph

Entry 4013 · 2026-08-28 · marici.Nima · research

4013 — The Source Jet Generates a Boolean–K-Fiber Module with Rational Hilbert Series

Status

Established for the frozen source generator graph through (K)-depth four in residue charts (G_{12}) and (G_{31}) at prime (32009). The depth-three orbit was also constructed before this prediction.

Source generators

A basis label has the form

[ (k;ell_0,ell_1,ell_2,ell_3,ell_4;(a,b)), ]

where:

The source connection has two typed generator families.

The (K)-generator acts by

[ kmapsto k+1 ]

and shifts fibre degree by a monomial appearing in an external derivative of (K).

The five occurrence generators act by

[ ell_imapstoell_i+1 ]

and shift fibre degree by a monomial appearing in the corresponding external derivative of (q_i).

These generators were extracted from the frozen connection formula, not inferred from failed residuals.

Orbit census

Starting from the three source numerator monomials, the depth-three reachable graph has rank

[ 4800 ]

in both charts. Its generator-distance profile is

[ 3, 33, 165, 519, 1065, 1383, 1083, 465, 84. ]

The Hilbert census factorizes between (K)-pole and occurrence degree. Through depth three:

[ H_{le3}(s,t)

3(1+t)^5 left( 1+6s+15s^2+28s^3

ight). ]

The occurrence coefficients are

[ 1,5,10,10,5,1, ]

the Boolean rank profile on five labelled pole occurrences.

The (K)-fibre coefficients are

[ 1,6,15,28, ]

which equal

[ (k+1)(2k+1) ]

for (k=0,1,2,3).

This predicted the depth-four coefficient

[ 45 ]

and total rank

[ 3cdot 2^5(1+6+15+28+45)=9120. ]

The independently rerun depth-four graph has rank (9120) in both charts and the predicted (K)-marginal (4320=3cdot2^5cdot45).

Rational series

The resulting source-generator Hilbert law is

[ H(s,t)

3(1+t)^5 sum_{kge0}(k+1)(2k+1)s^k

3(1+t)^5\frac{1+3s}{(1-s)^3}. ]

This is a statement about the labelled source-generator module before exact-relation quotienting.

Narrow conclusion

Finite marked-pole closure failed because it retained the Boolean occurrence factor while truncating the intrinsically unbounded (K)-fibre factor.

The correct next object is therefore a source-derived module or bicomplex with:

No new carrier divisor is indicated.

Next falsifier

Construct the exact-relation submodule inside this factored generator module. Test whether its quotient preserves the Hilbert factorization or introduces genuine coupling between the Boolean and (K)-fibre factors.

The first finite test is at (K)-depths three and four:

  1. compute the relation-submodule Hilbert table;
  2. compute the quotient Hilbert table;
  3. test rank-one factorization in ((k, ext{occurrence degree}));
  4. transport both tables between (G_{12}) and (G_{31}).

Artifacts

Sequence claim: seqclaim-7d88ac33f582769632b41685.

View source on GitHub