Erdős #1049: the Bézout–Plücker normal form for endpoint jets
DefinitionErdosProblems_Erdos1049_BezoutPluckerJetsbezout-pluckererdos-1049formalization
A Bézout coordinate change turns endpoint jet determinant constraints into a Plücker normal form, using the shared arithmetic of selector rows rather than counting independent ambient coordinates. The submitted module contains the source declarations bezoutPluckerEquiv, bezoutPluckerEquiv_apply, bezoutPluckerEquiv_anchor, exists_scalar_eq_of_det_eq_zero, anchor_det_zero_forces_all_det_zero, among others. Source topic: Erdős #1049: the Bézout–Plücker normal form for endpoint jets.
Definition code
import Lean.Elab.Tactic.Omega
import Mathlib.Algebra.BigOperators.Group.Finset.Basic
import Mathlib.Algebra.Module.BigOperators
import Mathlib.Algebra.Module.Prod
import Mathlib.Data.Fintype.Pi
import Mathlib.Data.ZMod.Basic
import Mathlib.RingTheory.Coprime.Basic
import Mathlib.Tactic.Choose
import Mathlib.Tactic.Push
import Mathlib.Tactic.Ring
/-!
# Erdős #1049: the Bézout–Plücker normal form for endpoint jets
`ZudilinConeArithmetic.fourJetSignature_card` counts the *whole ambient*
endpoint-jet target, `(3^R)^2 · (2^S)^2`, and its binary-selector pigeonhole
therefore needs rank `4R + 2S`. That count is wasteful whenever the rows
share arithmetic: the four jets are two coordinates modulo `N = 2^S·3^R` by
CRT, and one determinant-one shear makes one of those coordinates *free*
while the other consists exactly of exterior minors.
## The normal form
Fix an anchor `(a,b)` that is unimodular modulo `N`, i.e. `u·a + v·b = 1` for
some `u, v`. The shear
`T = ![![u, v], ![-b, a]]`, `det T = u·a + v·b = 1`
sends the anchor to `(1,0)` and a general row `(A,B)` to `(u·A + v·B, a·B - b·A)`.
Its second coordinate **is** the exterior minor against the anchor. Hence the
subgroup generated by the rows is `R × ⟨minors⟩`, so
`|H_N| = N²/d` with `d = gcd(N, minors)`, and the cokernel is cyclic of
order `d` — every `2`-adic anchor-minor depth saves one selector and every
`3`-adic depth saves two, with no logarithmic estimate.
At maximal collapse (`d = N`, every anchor minor vanishing) the target has
only `N` elements instead of `N²`, so the ambient rank threshold `4R + 2S`
drops to `2R + S`. `binary_row_collision_of_anchor_det_zero` is that halving,
proved from `card R < 2 ^ card ι` rather than from `card (R × R)`.
## Claim ceiling
**Erdős #1049 remains open.** This is the linear-algebraic half only. It
supplies a smaller jet target; it does not supply the analytic remainder
non-vanishing, and `ZudilinConeArithmetic` continues to record that the
collision theorem yields a kernel element, not a contradiction.
-/
namespace ErdosProblems.Erdos1049.BezoutPluckerJets
section BezoutPlucker
variable {R : Type*} [CommRing R]
/-- The determinant-one Bézout shear based at the unimodular anchor `(a,b)`.
Its second coordinate is the exterior minor `a·y - b·x` against the anchor. -/
def bezoutPluckerEquiv (a b u v : R) (h : u * a + v * b = 1) : R × R ≃+ R × R where
toFun z := (u * z.1 + v * z.2, a * z.2 - b * z.1)
invFun z := (a * z.1 - v * z.2, b * z.1 + u * z.2)
left_inv := by
rintro ⟨x, y⟩
apply Prod.ext <;> dsimp
· calc a * (u * x + v * y) - v * (a * y - b * x) = (u * a + v * b) * x := by ring
_ = x := by rw [h, one_mul]
· calc b * (u * x + v * y) + u * (a * y - b * x) = (u * a + v * b) * y := by ring
_ = y := by rw [h, one_mul]
right_inv := by
rintro ⟨x, y⟩
apply Prod.ext <;> dsimp
· calc u * (a * x - v * y) + v * (b * x + u * y) = (u * a + v * b) * x := by ring
_ = x := by rw [h, one_mul]
· calc a * (b * x + u * y) - b * (a * x - v * y) = (u * a + v * b) * y := by ring
_ = y := by rw [h, one_mul]
map_add' := by
rintro ⟨x₁, y₁⟩ ⟨x₂, y₂⟩
apply Prod.ext <;> dsimp <;> ring
end BezoutPlucker
section BinaryPigeonhole
end BinaryPigeonhole
section MaximalCollapse
variable {R ι : Type*} [CommRing R] [Fintype R] [Fintype ι]
end MaximalCollapse
/-! ## The `ZMod N` instance -/
end ErdosProblems.Erdos1049.BezoutPluckerJets
Source
Pinned Lean source: https://github.com/wcook04/plectis-erdos-lean/blob/c93c2e4dd86a2e317e0cb650ea244fee1afd59c2/ErdosProblems/Erdos1049/BezoutPluckerJets.lean#L55-L327
Paper exposition and prior-art bibliography: https://github.com/wcook04/plectis-erdos/blob/551bae6dc6e732cf85172d66323c8d2bc77ba962/paper/1049/erdos-1049-rational-base-lambert.tex#L1252-L1257
Correspondence: Erdős #1049 formal source; Zudilin is credited for the relevant rational-base Lambert-series method, without a novelty or all-rational-base claim.