Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 4.9 — Properties of the staircase constructor

Open
TarskiBootstrap.Dimensional.Supplement.lemma_4_9

by mikedeng1 · Nov 15, 2026 · Mathlib 0df444a (Lean v4.33.1)

adversary-methodp2o-batch-stoc27sp2o-gran-per-chapterp2o-plan-paperp2o-v1quantum-query-complexitytarski-fixed-points

Let F\mathcal FF, AAA be as in Lemma 4.3 and m≥2m\ge2m≥2. An element (s,f)(s,f)(s,f) of the composite domain D\mathcal DD consists of an ordered search instance s=(s1,…,sm)s=(s_1,\dots,s_m)s=(s1​,…,sm​) with hidden position i0i_0i0​ (sr=1s_r=1sr​=1 for r<i0r<i_0r<i0​, si0=0s_{i_0}=0si0​​=0, sr=−1s_r=-1sr​=−1 for r>i0r>i_0r>i0​) and f=(f1,…,fm)∈Fmf=(f_1,\dots,f_m)\in\mathcal F^mf=(f1​,…,fm​)∈Fm. For the staircase constructor E^\hat{\mathcal E}E^:

  1. for every (s,f)∈D(s,f)\in\mathcal D(s,f)∈D, E^(s,f):S→S\hat{\mathcal E}(s,f):S\to SE^(s,f):S→S is monotone;
  2. for every (s,f)∈D(s,f)\in\mathcal D(s,f)∈D, the unique fixed point of E^(s,f)\hat{\mathcal E}(s,f)E^(s,f) is ((i0,fi0∗),(i0,fi0∗))((i_0,f_{i_0}^*),(i_0,f_{i_0}^*))((i0​,fi0​∗​),(i0​,fi0​∗​)), where si0=0s_{i_0}=0si0​​=0;
  3. for any (s(1),f(1)),(s(2),f(2))∈D(s^{(1)},f^{(1)}),(s^{(2)},f^{(2)})\in\mathcal D(s(1),f(1)),(s(2),f(2))∈D and any u=((i,x),(j,y))∈Su=((i,x),(j,y))\in Su=((i,x),(j,y))∈S, with q1(u)=(i,x)q_1(u)=(i,x)q1​(u)=(i,x), q2(u)=(j,y)q_2(u)=(j,y)q2​(u)=(j,y) and h(a)(r,z)=Hsr(a)(fr(a))(z)h^{(a)}(r,z)=H_{s^{(a)}_r}(f^{(a)}_r)(z)h(a)(r,z)=Hsr(a)​​(fr(a)​)(z),
E^(s(1),f(1))(u)≠E^(s(2),f(2))(u) ⟹ ∃ℓ∈{1,2}, h(1)(qℓ(u))≠h(2)(qℓ(u)).\hat{\mathcal E}(s^{(1)},f^{(1)})(u)\ne\hat{\mathcal E}(s^{(2)},f^{(2)})(u)\ \Longrightarrow\ \exists\ell\in\{1,2\},\ h^{(1)}(q_\ell(u))\ne h^{(2)}(q_\ell(u)).E^(s(1),f(1))(u)=E^(s(2),f(2))(u) ⟹ ∃ℓ∈{1,2}, h(1)(qℓ​(u))=h(2)(qℓ​(u)).

These properties make the embedded family a valid Tarski family with unique fixed points and give the locality constant c=2c=2c=2 for the adversary reduction lemma.

Differences from the source:

  • encoding: a hard instance family for Tarski(n, k) in which every function has a unique fixed point → F : Set (Grid n k → Grid n k) with every member monotone and having a unique fixed point; reason: Definition 2.1: instances are monotone self-maps of [n]^k; "hard" adds no assumption beyond the stated witness
  • encoding: all possible fixed points lie on an antichain A → A antichain with possibleAnswers F ⊆ A; reason: literal transcription; the set of possible fixed points is the published possibleAnswers F
  • encoding: an integer m ≥ 2 → m : ℕ+ with 2 ≤ m; reason: same values
  • encoding: positions and block indices 1, …, m → Fin m = {0, …, m-1}; reason: order-preserving reindexing
  • encoding: (s, f) ∈ D → labels (i₀, f) with s = osInstance m i₀ and f : Fin m → F; reason: by (2.13) and the labelling of F′, D consists exactly of the oracles of such pairs, and block r has inner answer s_r
  • encoding: Q with the lexicographic order, S ⊆ Q² with the induced product order, Ê(s, f) : S → S → Fin m ×ₗ Grid n k, a subtype of Q × Q, and a map of that subtype; reason: Mathlib’s lexicographic product has exactly the order of (4.8)
  • encoding: ∃ ℓ ∈ {1, 2} → a disjunction of the two cases; reason: two-element existential
Preamble
import Definitions.Def_TarskiBootstrap_Dimensional_Supplement_Staircase
import Definitions.Def_TarskiBootstrap_Dimensional_Supplement_Composite
Formal statement
namespace TarskiBootstrap.Dimensional.Supplement

/-- Lemma 4.9, PDF/printed pp. 21–22. An element `(s, f)` of the composite domain `D`
is labelled by its hidden position `i₀` (so `s = osInstance m i₀`) and `f ∈ F^m`.
(1) `Ê(s, f)` is monotone on the staircase `S`; (2) its unique fixed point is
`((i₀, f_{i₀}*), (i₀, f_{i₀}*))`, where `s_{i₀} = 0`; (3) with `q₁(u) = (i, x)`,
`q₂(u) = (j, y)` for `u = ((i, x), (j, y)) ∈ S`, a difference of the two staircase values
at `u` forces a difference of the composite oracles `h(r, z) = H_{s_r}(f_r)(z)` at
`q₁(u)` or `q₂(u)`. -/
theorem lemma_4_9 {n k : ℕ+} (F : Set (Grid n k → Grid n k)) (A : Set (Grid n k))
    (hF : ∀ f : F, Monotone f.val ∧ ∃! x, f.val x = x)
    (hA : IsAntichain (· ≤ ·) A) (hcontains : possibleAnswers F ⊆ A)
    (m : ℕ+) (hm : 2 ≤ (m : ℕ)) :
    (∀ (i₀ : Fin (m : ℕ)) (f : Fin (m : ℕ) → F),
      Monotone (staircase A (osInstance m i₀) (fun r => (f r).val))) ∧
    (∀ (i₀ : Fin (m : ℕ)) (f : Fin (m : ℕ) → F) (u : Staircase n k m),
      staircase A (osInstance m i₀) (fun r => (f r).val) u = u ↔
        u.val = (toLex (i₀, fixedPoint (f i₀).val), toLex (i₀, fixedPoint (f i₀).val))) ∧
    (∀ (i₁ i₂ : Fin (m : ℕ)) (f₁ f₂ : Fin (m : ℕ) → F) (u : Staircase n k m),
      staircase A (osInstance m i₁) (fun r => (f₁ r).val) u ≠
          staircase A (osInstance m i₂) (fun r => (f₂ r).val) u →
        compositeOracle A (osInstance m i₁) (fun r => (f₁ r).val) (ofLex u.val.1) ≠
            compositeOracle A (osInstance m i₂) (fun r => (f₂ r).val) (ofLex u.val.1) ∨
          compositeOracle A (osInstance m i₁) (fun r => (f₁ r).val) (ofLex u.val.2) ≠
            compositeOracle A (osInstance m i₂) (fun r => (f₂ r).val) (ofLex u.val.2)) := by sorry

end TarskiBootstrap.Dimensional.Supplement
Source
Li, Ma, Yang and Zhao, Bootstrapping Classical and Quantum Query Lower Bounds for Finding Tarski Fixed Points, STOC 2027 submission #6 (2026), frozen preprint, Lemma 4.9, PDF/printed pp. 21–22, SHA-256 320afa8333f95589a388b30ffbfd6d7b7e5d93cdf07fa8630825fbf7d0107d59; https://github.com/sigact-stoc/stoc2027/releases/download/papers-1/6.pdf. Printed page = PDF page. No arXiv identifier/version supplied.

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me