riemann

The Lean code claims a proof of the Riemann Hypothesis. This site has not checked the claim yet. See what is checked.

ZeroFreeAll from Green, VWin and HypsAlt

drafted

Statement

Assume three inputs. The Lean local name of each input is in parentheses.

  1. (hGh_G) The statement Green. For each prime gg, each label datum LL and each d>0d > 0, the cuspidal characters of GLd(๐”ฝg)\mathrm{GL}_d(\mathbb{F}_g) are in bijection with the monic irreducible polynomials of degree dd over โ„ค/g\mathbb{Z}/g with nonzero constant term, with the degree โˆi=1dโˆ’1(giโˆ’1)\prod_{i=1}^{d-1}(g^i - 1) and an explicit value formula.
  2. (hVh_V) The statement VWin, that is, VWinAt ฯˆ 401 for each primitive Dirichlet character ฯˆ\psi.
  3. (hh) A proof that the type HypsAlt N DefiniteSideCondition is nonempty, where N(๐’ž)N(\mathcal{C}) is Closed ๐’ž fun Fm _ => LinkN Fm. HypsAlt is a record of 46 fields about a class of families, a selection of bundles and many bounds. The second parameter is the condition DefiniteSideCondition. By the definition Closed, Closed ๐’ž P means: there is ฮต1>0\varepsilon_1 > 0 such that for each ฮตโˆˆ(0,ฮต1]\varepsilon \in (0, \varepsilon_1] there are thresholds such that PP holds for each family of ๐’ž\mathcal{C} with parameter ฮต\varepsilon whose constants are above the thresholds. LinkN is a bound on a weighted count of twisted places.

Then ZeroFreeAll holds. By definition, ZeroFreeAll is ZeroFreeOn fun _ => True, and ZeroFreeOn Q says: for each primitive Dirichlet character ฯˆ\psi with Q ฯˆ, ฯˆ\psi satisfies ZeroFreeAbove (1/2). So ZeroFreeAll says the following. Let ฯˆ\psi be a primitive Dirichlet character of modulus Nโ‰ 0N \neq 0 (a value of the structure PChar, which holds NN, the character and a proof that it is primitive). Then for each sโˆˆโ„‚s \in \mathbb{C} with Reโก(s)>12\operatorname{Re}(s) > \tfrac12, the Dirichlet L-function of Mathlib has L(s,ฯˆ)โ‰ 0L(s, \psi) \neq 0.

Lean

theorem UpperTheory.Top.zero_free_all_of_green_v_win (hG : GlobalTheory.GLq.Green) (hV : UpperTheory.Top.VWin) (h : Nonempty (_root_.GlobalTheory.Fc.HypsAlt (fun ๐’ž => Closed ๐’ž fun Fm _ => LinkN Fm) _root_.GlobalTheory.Fc.DefiniteSideCondition)) : Formalism.HC.ZeroFreeAll

The declaration is in Development/AutomorphicArithmetic/ThetaCorrespondence/ThetaPairLifts/ThirdPlaceSeries/TopPlug.lean, line 12. Its page has the full list of what it uses and what uses it, and a link to the source.

Proof

The proof is one term with four parts.

  1. The hypothesis hh is a proof of Nonempty (HypsAlt ...). The proof takes one record out of it with h.elim. This is bookkeeping: HypsAlt is a type in Type 1, and the goal is a proposition, so the elimination is allowed.
  2. The theorem shape_closed_link_n_of_green takes hGh_G and gives TopShapeAlt True True N TopG, with the same NN as in hh. TopShapeAlt V NR N G is a long implication. It says: for each family class ๐’œ\mathcal{A} and about 50 named data and conditions on ๐’œ\mathcal{A} (a selection, bounds, closed bounds of the form Closed, a proof of VV, a proof of NRNR, and more), the proposition GG holds.
  3. The theorem forall_goal_of_shape_of_hyps has the type TopShapeAlt V NR N G โ†’ V โ†’ NR โ†’ HypsAlt N DefiniteSideCondition โ†’ G. The proof applies it with V=TrueV = \mathrm{True} and NR=TrueNR = \mathrm{True}, the result of part 2, trivial twice, and the record from part 1. The theorem passes the fields of the record into the slots of TopShapeAlt. The result is TopG. TopG says: for each ฮป>1\lambda > 1 and each requirement GG (a value of FReq), if HostExW ฮป G.req holds and each family in the class G.cls ฮป satisfies HypV, then the class G.cls ฮป is FamilyClassConnected.
  4. The theorem zeroFreeAll_of_topG takes a number ฮป\lambda with 1<ฮป1 < \lambda, a proof of TopG, and for each primitive character ฯˆ\psi a proof of HostExW ฮป (GChi ฯˆ).req. It gives ZeroFreeAll. The proof uses ฮป=3/2\lambda = 3/2. The fact 1<3/21 < 3/2 is closed by norm_num. For each ฯˆ\psi, the theorem supplyChi turns hV ฯˆ, a proof of VWinAt ฯˆ 401, into HostExW (3/2) (GChi ฯˆ).req.

So the term is zeroFreeAll_of_topG (la := 3/2) _ (forall_goal_of_shape_of_hyps _ _ _ _ (shape_closed_link_n_of_green hG) trivial trivial h) (fun ฯˆ => supplyChi ฯˆ (hV ฯˆ)), inside h.elim. The theta matching chapter describes the proof of zeroFreeAll_of_topG, with the sums GmG_m of the von Mangoldt function and the statement FormB.

Notes

The input hGh_G is used only in part 2. Inside shape_closed_link_n_of_green, it goes into the lemma closed_linkIRR_of_join_count. So a statement on the characters of the finite groups GLd(๐”ฝg)\mathrm{GL}_d(\mathbb{F}_g) enters the proof of a statement on Dirichlet L-functions only through that lemma. This page does not follow that link.

The two slots VV and NRNR of TopShapeAlt are the proposition True here. They are filled with trivial and carry no content.

The input hVh_V is used only through supplyChi, at the fixed value ฮป=3/2\lambda = 3/2.

ZeroFreeAll has no exception at s=1s = 1. For the trivial character of modulus 11, L(s,1)L(s, 1) is the Riemann zeta function. Mathlib gives it a value at the pole s=1s = 1 by a convention, and this value is not zero. The theta matching chapter says more on this point.

The user of this step is EveryCharacter from Green, VWin and HypsAlt, which applies the step imprimitive to the result.

Uses

The proof cites these declarations, but their steps are not written yet. The text above tells what each one gives: Formalism.HC.ZeroFreeAll, Formalism.MainNotions.LinkN, GlobalTheory.Fc.forall_goal_of_shape_of_hyps, UpperTheory.Top.shape_closed_link_n_of_green, UpperTheory.Top.supplyChi, UpperTheory.Top.zeroFreeAll_of_topG.

Used by

Status

Drafted. The text follows the Lean code, but no person has checked it against the code with the rules of spec 05.7 yet. If you check it, tell us.

A model wrote the first draft (writer@claude-opus-5-5). A person checks it before its status goes up.

The source of this page. Markdown. Something is wrong? Report it.