argument-selection.md

back to table · edit · history · where entries came from · files · download

3739 bytes, as of the version from 2026-11-18 17:13 (current). Recorded here, not run.

# Argument selection for the transcendental diagonal-quartic family

Quantity: det(1 - T F_p) on the rank-two characteristic-zero transcendental summand of H^2 of x0^4+a1*x1^4+a2*x2^4+a3*x3^4=0 over Q, transported to smooth reduction. No Tate twist. Parameter order is a (Symbolic), p (Z); a uses signed dotted coefficient identities, displaying the ordered triple. Coefficients are nonzero signed fourth-power-free integers. This represents rational coefficient classes modulo fourth powers, not rational-isomorphism classes of surfaces. Good places are odd primes not dividing a1*a2*a3; T is the polynomial variable.

The source-defined family has no finite coefficient or prime bound. Gvirtz–Skorobogatov v2 Remark 2.10 defines these twists; Section 3.1 and Proposition 3.2 identify the two-dimensional summand; Remark 3.10 gives its product-character twist. Griffon v2 Definition 2.2, Theorem 2.3 and Remark 2.4 supply the untwisted Jacobi normalization, including nonsplit primes. These passages have been inspected in the archived source HTML.

Parameter a has three disjoint roles (mixed selection):

- Intrinsic: 1.1.1, the all-plus Fermat normalization control.
- Exploratory reference sample: 1.1.k for every signed integer 1 <= |k| <= 10 except k=1. Exact identities: 1.1.-1, 1.1.2, 1.1.-2, 1.1.3, 1.1.-3, 1.1.4, 1.1.-4, 1.1.5, 1.1.-5, 1.1.6, 1.1.-6, 1.1.7, 1.1.-7, 1.1.8, 1.1.-8, 1.1.9, 1.1.-9, 1.1.10, 1.1.-10. Purpose: recognition of low-complexity twists beyond the two Fermat sign conventions. The noncanonical absolute-integer bound includes both signs, square and nonsquare classes, odd primes and prime powers. It is independent of decimal spelling, selects one Kummer coefficient slice and avoids a three-coefficient Cartesian box. The sources support the twist construction, not a documented application for every k.
- Documented arithmetic comparisons: 1.2.8, 1.-2.-8, 2.-2.-4, -2.4.8, 1.2.-2, 1.8.-8, exactly the six triples in Gvirtz–Skorobogatov v2 equation (28), in its Brauer-group/Galois-cohomology calculation. We store their transcendental polynomial, not Brauer-group orders.

Parameter p is a discrete_index: a genuine good-reduction place, with every admissible odd prime <=97 for every model. This noncanonical finite prime bound gives both split and inert congruence classes and depth for arithmetic comparisons after covering all model groups. Bad primes are omitted, with no guessed factors. There are 24 odd primes; deleting 12 model-prime pairs dividing k gives 26*24-12=612 entries. The compact degree-two output makes this broader prime range practical. Neither famous-only pruning nor a larger coefficient box improves the stated recognition purpose. Other coefficient models and larger good primes remain uncomputed.

Enumeration order: intrinsic base; reference k by increasing |k|, positive before negative (omitting the repeated base); six documented comparisons in source order; primes increasing within each model. Enumeration, Programs and completeness use this same coverage. The 1.1.-1 model is included here: T729 holds its degree-21 primitive polynomial, a different quantity. There is no published family duplication.

Reference policy (for a fresh critic to assess, not a verdict): purpose as above; bounds signed 1<=|k|<=10 on (1,1,k), k!=1; coverage rationale one simple exact Kummer slice plus intrinsic and documented models; noncanonical_acknowledged=true; notation_independent=true; sources_checked=true. No continuous parameter or probability input is present.

Skill evidence: ../20261118T160132Z-build-skills/manifest.json; matching bundle SHA-256 b5c3c7985b7c9e178b4899d083a9e3bc4284989dff76f05e5b0e52835d47eed2. Independent argument-selection review remains mandatory.