AFTD Auto-Formalizing Theoretical Domains

linear_flattening_potency_gt_of_rank_le

theoremverified

If a linear flattening L on (ℂⁿ)⊗³ satisfies rank L(x ⊗ y ⊗ z) ≤ ρ for all x, y, z and rank L(T) > r * ρ for some tensor T, then L has potency strictly greater than r.

Statement

theorem linear_flattening_potency_gt_of_rank_le {n u v : ℕ}
    (L : (Fin n → Fin n → Fin n → ℂ) →ₗ[ℂ] Matrix (Fin u) (Fin v) ℂ) (r ρ : ℕ)
    (hρ : ∀ x y z : Fin n → ℂ, (L (tensor3_pure x y z)).rank ≤ ρ)
    (T : Fin n → Fin n → Fin n → ℂ)
    (hT : r * ρ < (L T).rank) :
    linear_flattening_potency_gt L r
Source
step towards linear_flattening_testable_order_gap_four
Verified
23 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 2
  • linear_flattening_potency_gtA linear flattening L : ℂⁿ ⊗ ℂⁿ ⊗ ℂⁿ → Mat_{u×v}(ℂ) has potency greater than r: some tensor T has rank L(T) > r · ρ(L), where ρ(L) =…
  • tensor3_pureThe pure tensor x ⊗ y ⊗ z in ℂⁿ ⊗ ℂⁿ ⊗ ℂⁿ, with entries x_i y_j z_k.

The proof, explained

To show that the linear flattening $L$ has potency strictly greater than $r$, it suffices to exhibit a tensor $T$ such that $r \cdot \operatorname{rank} L(x \otimes y \otimes z) < \operatorname{rank} L(T)$ for all $x, y, z$. The given tensor $T$ serves directly as the witness: multiplying the uniform rank bound $\operatorname{rank} L(x \otimes y \otimes z) \le \rho$ by $r$ gives $r \cdot \operatorname{rank} L(x \otimes y \otimes z) \le r \cdot \rho$, which is strictly less than $\operatorname{rank} L(T)$ by hypothesis.

  1. $T$ witnesses that $L$ has potency strictly greater than $r$.

    By definition, $\operatorname{linear\_flattening\_potency\_gt}\ L\ r$ is an existential statement asserting that there is a tensor $T$ such that $r \cdot \operatorname{rank} L(x \otimes y \otimes z) < \operatorname{rank} L(T)$ for all $x, y, z$. Providing the given tensor $T$ as the witness, for any vectors $x, y, z$, we multiply the uniform bound $\operatorname{rank} L(x \otimes y \otimes z) \le \rho$ (from $h\rho$) on the left by $r$ via `Nat.mul_le_mul_left` to get $r \cdot \operatorname{rank} L(x \otimes y \otimes z) \le r \cdot \rho$. Combining this with the hypothesis $hT$ ($r \cdot \rho < \operatorname{rank} L(T)$) using `lt_of_le_of_lt` yields $r \cdot \operatorname{rank} L(x \otimes y \otimes z) < \operatorname{rank} L(T)$.

    line 8

    uses lt_of_le_of_lt, Nat.mul_le_mul_left

      ⟨T, fun x y z => lt_of_le_of_lt (Nat.mul_le_mul_left r (hρ x y z)) hT⟩
lt_of_le_of_lt
If $a \le b$ and $b < c$, then $a < c$ (transitivity of $\le$ and $<$).
Nat.mul_le_mul_left
Multiplication on the left by a natural number preserves order: if $n \le m$, then $k \cdot n \le k \cdot m$.

Written by a model (gemini-3.8-flash) from the Lean source alone (no goal states were available). Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.

Lean source view module on GitHub

/-- A linear flattening L achieves potency strictly greater than r if there is a tensor T with rank L(T) > r * ρ for a uniform upper bound ρ on the rank of L on all pure tensors. -/theorem linear_flattening_potency_gt_of_rank_le {n u v : ℕ}    (L : (Fin n → Fin n → Fin n → ℂ) →ₗ[ℂ] Matrix (Fin u) (Fin v) ℂ) (r ρ : ℕ)    (hρ : ∀ x y z : Fin n → ℂ, (L (tensor3_pure x y z)).rank ≤ ρ)    (T : Fin n → Fin n → Fin n → ℂ)    (hT : r * ρ < (L T).rank) :    linear_flattening_potency_gt L r :=  ⟨T, fun x y z => lt_of_le_of_lt (Nat.mul_le_mul_left r (hρ x y z)) hT⟩
See it in the graphDiscuss this resultChallenge it