The fibre of the seed series over `m = 2q` is at least `1/(20 q)`, for every prime `q`
ProvedBerggrenZeta.fibre_lower_boundaether-catalognovelty
The fibre of the seed series over m = 2q is at least 1/(20 q), for every prime q
and every s ≤ 1. This is the arithmetic engine of the divergence.
theorem BerggrenZeta.fibre_lower_bound{s : ℝ} (hs : s ≤ 1) {q : ℕ} (hq : q.Prime)
(hfib : Summable (fun n : ℕ => seedTerm s (2 * q, n))) :
1 / (20 * (q : ℝ)) ≤ ∑' n : ℕ, seedTerm s (2 * q, n) := by sorry
Formalization Note Transplanted verbatim from the Aether Catalog source Novelty/BerggrenTreeZetaAbscissa.lean; the statement is byte-identical to the source declaration, elaborated with autoImplicit disabled in the platform environment.
Preamble
-- Thm stub generated from Novelty/BerggrenTreeZetaAbscissa.lean
import Mathlib
import Definitions.Def_Novelty_BerggrenTreeZetaAbscissa
import Definitions.Def_Novelty_BerggrenTreeZetaCore
/-!
# The Berggren tree zeta function and its abscissa of convergence
The **tree zeta function** of the Berggren tree of primitive Pythagorean triples is the
Dirichlet series over the nodes of the tree, weighted by the hypotenuse:
`Z_tree(s) = ∑_{w ∈ {L,M,R}*} c(w)^{-s}`.
The moonshot conjecture attached to this project predicted that the abscissa of
convergence of `Z_tree` is governed by the **silver ratio** `1 + √2` — the eigenvalue
structure `3 ± 2√2` of the Berggren generators — for instance
`σ = log 3 / (2 log (1+√2)) ≈ 0.6232` (the "branching over silver growth" exponent), or
`log (1+√2) ≈ 0.8814`.
**This file refutes that prediction and computes the true answer: the abscissa is `1`.**
The mechanism is the growth dichotomy inside the tree: the middle (Pell) branch grows at
the silver rate `(1+√2)^{2k}`, but the two outer branches grow only *quadratically* in the
depth, so the tree contains far more small hypotenuses than a purely exponential branching
model predicts. In fact — by the bijection `seedEquiv` of the core file — the nodes are in
bijection with Euclid seeds, so `Z_tree` is exactly the Dirichlet series of the primitive
Pythagorean hypotenuses counted with multiplicity, whose counting function is `Θ(H)`, and
the abscissa is `1`.
## Main results
* `summable_seedTerm_of_one_lt` — convergence for `s > 1` (an elementary two-dimensional
comparison: at most `m` seeds have first coordinate `m`, and each contributes at most
`m^{-2s}`);
* `not_summable_seedTerm_of_le_one` — divergence for every `s ≤ 1`. The witness family is
arithmetic: for each prime `q`, the seeds `(2q, n)` with `n` odd and `n < q` are
admissible, contribute `≳ 1/(20 q)` in total, and `∑_q 1/q` diverges (Euler).
* `treeZeta_summable_iff` — **the abscissa of convergence of the tree zeta function is
exactly `1`**;
* `treeZeta_abscissa_ne_silver` — the quantitative refutation: the abscissa is neither
`log (1+√2)` nor `log 3 / (2 log (1+√2))`.
-/
open BerggrenZeta
-- open removed: section is not a namespace
/-! ## Part A. The tree zeta function as a Dirichlet series over Euclid seeds -/
/-! ## Part B. Convergence for `s > 1` -/
/-! ## Part C. Divergence for `s ≤ 1`: an arithmetic family of seeds -/Formal statement
theorem BerggrenZeta.fibre_lower_bound{s : ℝ} (hs : s ≤ 1) {q : ℕ} (hq : q.Prime)
(hfib : Summable (fun n : ℕ => seedTerm s (2 * q, n))) :
1 / (20 * (q : ℝ)) ≤ ∑' n : ℕ, seedTerm s (2 * q, n) := by sorrySource