argument-selection.md

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

3111 bytes, as of the version from 2026-11-18 03:54 (current). Recorded here, not run.

# T744 argument selection

The one free parameter is `k`, type R, the actual coefficient in
R_k(x,y)=x^3+y^3-x-y+kxy, with domain k>=0. Negative k adds no new value,
since R_{-k}(x,y)=-R_k(-x,-y). The positive roots and the threshold below
are exact algebraic arguments; decimals do not define any input.

The complete selection is sqrt(n), 0<=n<=36, together with
kappa=16/(3sqrt(3)), ordered by increasing k (kappa follows 3).
Square radicands reduce to integer identities. `n` is an enumeration rule,
not a stored independent parameter. There are 38 identities and no Cartesian
product or continuous mesh.

This parameter has a mixed selection with two disjoint, exhaustive groups:

* Intrinsic: `0`, `sqrt(8)`, `3`, `16/(3*sqrt(3))`.
  The generic cubic degenerates at 0 and 3. Lemmas 1–4 in “The third family”
  distinguish the larger-root transition at sqrt(8), discriminant change at 3,
  and the final smaller-root contact at kappa. Including these alongside
  the low regime checks both Jensen roots rather than extrapolating a theorem.
* Arithmetic structure: `sqrt(n)` for 1<=n<=36, n not in {8,9}, with square
  cases represented by integers. The source Introduction explicitly singles
  out k^2 integral, with integral model
  E_n: Y^2=X^3+(n-24)X^2-16(n-9)X.
  Its discriminant is 4096 n (n-9)^2(n+16), checked by exact expansion.
  The family therefore includes integral and nonsquare quadratic coefficients;
  it is not justified merely by the spelling of radicals or by special fame.

The noncanonical height bound n<=36 is a finite cost choice, retaining a complete
small-height prefix, integer k through six, nonsquare coefficients, both singular
fibers, the low two-root regime and the theorem's threshold. There are no sampled interior points in the short intervals sqrt(8)<k<3
and 3<k<kappa; their distinguished endpoints are retained. No continuous
coverage of those intervals is claimed. No individual application
or L-value theorem at every input is claimed. The mathematical domain remains
all k>=0; `complete: no` explains the finite extent. Higher heights extend this
same table rather than form a new quantity. No exploratory group is hidden
inside the intrinsic group. The arithmetic group is supported by the integral
elliptic model, not a general prescription to sample radicals for other functions.

Checked source: Bertin–Zudilin, On the Mahler measure of hyperelliptic families,
DOI 10.1007/s40316-016-0068-4, Introduction, Theorem 2, “The third family”,
Lemmas 1–4 and Propositions 3–4 (journal pp. 200–201, 205–209); matching
arXiv:1601.07583 Introduction and third-family section. Source bytes are retained
in this directory. The independent critic must inspect these claims and enumerate
every normalized identity in its own argument-review sidecar; this builder has
not issued that verdict. The generator, enumeration.json, Programs and completeness
note all use the selection above.

Skill manifest: ../20261118T033405Z-build-skills/manifest.json;
matching local/live bundle SHA-256:
d6e85d49873fac5bc748d068b0288cf242b0f9aebf3a9a0fff70295019227420.