AFTD Auto-Formalizing Theoretical Domains

three_hadamard_clean_return_phases

theoremverified

For unit complex phases u and v, vanishing of the unwanted numerator 1-u+v+u*v of an alternating three-Hadamard chain forces u=v and v=i or v=-i.

Statement

theorem three_hadamard_clean_return_phases (u v : ℂ)
    (hu : Complex.normSq u = 1) (hv : Complex.normSq v = 1)
    (hreturn : 1 - u + v + u * v = 0) :
    u = v ∧ (v = Complex.I ∨ v = -Complex.I)
Source
Supporting algebra for OP-3, arXiv:2610.01024, Section E.1; developed in deep session IN-184. These helpers do not establish a lower bound for arbitrary circuit gate lists.
Verified
25 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Used by 1

The proof, explained

The equation 1 - u + v + u * v = 0 rearranges to 1 + v = u * (1 - v). Taking the squared complex norm of both sides and using |u|² = 1 shows |1 + v|² = |1 - v|², which expands to reveal Re(v) = 0. Since |v|² = 1, this forces Im(v) = ±1, meaning v = Complex.I or v = -Complex.I. Finally, using v² = -1 and 1 - v ≠ 0, cancelling 1 - v from u * (1 - v) = 1 + v = v * (1 - v) yields u = v.

  1. |1 + v|² = |1 - v|²

    The given relation 1 - u + v + u * v = 0 rearranges algebraically to 1 + v = u * (1 - v). Applying the squared complex norm Complex.normSq to both sides via congrArg, and simplifying the right-hand side using multiplicativity (map_mul), the unit modulus assumption |u|² = 1 (hu), and identity multiplication (one_mul), yields |1 + v|² = |1 - v|².

    lines 5–7

    uses congrArg, map_mul, one_mul

      have he : 1 + v = u * (1 - v) := by linear_combination hreturn
      have hn := congrArg Complex.normSq he
      rw [map_mul, hu, one_mul] at hn
  2. Re(v) = 0

    Expanding |1 + v|² and |1 - v|² with Complex.normSq_add and Complex.normSq_sub, and substituting |v|² = 1 (hv), |1|² = 1 (map_one), and Re(v̄) = Re(v) (Complex.conj_re), simplifies |1 + v|² = |1 - v|² to 2 + 2 Re(v) = 2 - 2 Re(v). Linear arithmetic then forces Re(v) = 0.

    lines 8–11

    uses Complex.normSq_add, Complex.normSq_sub, map_one, Complex.conj_re, one_mul

      have hr : v.re = 0 := by
        simp only [Complex.normSq_add, Complex.normSq_sub, hv, map_one,
          Complex.conj_re, one_mul] at hn
        linarith
  3. Im(v) = 1 ∨ Im(v) = -1

    Rewriting the hypothesis |v|² = 1 in terms of real and imaginary parts using Complex.normSq_apply and substituting Re(v) = 0 yields Im(v)² = 1, which factors as (Im(v) - 1) * (Im(v) + 1) = 0. By mul_eq_zero.mp, at least one factor must be zero, so Im(v) = 1 or Im(v) = -1.

    lines 12–17

    uses Complex.normSq_apply, mul_eq_zero.mp

      have hi : v.im = 1 ∨ v.im = -1 := by
        rw [Complex.normSq_apply, hr] at hv
        have hmul : (v.im - 1) * (v.im + 1) = 0 := by nlinarith [hv]
        rcases mul_eq_zero.mp hmul with h | h
        · left; linarith
        · right; linarith
  4. v = Complex.I ∨ v = -Complex.I

    In either case for Im(v), the real part is 0 and the imaginary part is 1 or -1, which match the real and imaginary parts of Complex.I and -Complex.I respectively. By Complex.ext, comparing components establishes that v = Complex.I or v = -Complex.I.

    lines 18–21

    uses Complex.ext

      have hvv : v = Complex.I ∨ v = -Complex.I := by
        rcases hi with hi | hi
        · left; apply Complex.ext <;> simp [hr, hi]
        · right; apply Complex.ext <;> simp [hr, hi]
  5. u = v

    Since v = ±Complex.I, squaring gives v² = -1. Moreover, 1 - v ≠ 0, since 1 - v = 0 would imply v = 1 and thus Re(v) = 1, contradicting Re(v) = 0. Calculating v * (1 - v) = v - v² = v - (-1) = 1 + v, and recalling 1 + v = u * (1 - v), we obtain u * (1 - v) = v * (1 - v). Cancelling the non-zero factor 1 - v via mul_right_cancel₀ yields u = v.

    lines 22–33

    uses mul_right_cancel₀

      have huv : u = v := by
        have hsq : v * v = -1 := by
          rcases hvv with h | h <;> rw [h] <;> simp
        have hne : 1 - v ≠ 0 := by
          intro h
          have : v = 1 := by linear_combination -h
          rw [this] at hr
          norm_num at hr
        apply mul_right_cancel₀ hne
        calc
          u * (1 - v) = 1 + v := he.symm
          _ = v * (1 - v) := by linear_combination hsq
  6. u = v ∧ (v = Complex.I ∨ v = -Complex.I)

    Pairing the equality u = v with the disjunction v = Complex.I ∨ v = -Complex.I concludes the proof.

    line 34

      exact ⟨huv, hvv⟩
congrArg
Applying a function to both sides of an equality preserves the equality: if a = b then f(a) = f(b).
map_mul
A multiplicative map preserves multiplication: f(x * y) = f(x) * f(y).
one_mul
Multiplying by 1 on the left is the identity: 1 * a = a.
Complex.normSq_add
The squared complex norm of a sum satisfies |z + w|² = |z|² + |w|² + 2 Re(z * w̄).
Complex.normSq_sub
The squared complex norm of a difference satisfies |z - w|² = |z|² + |w|² - 2 Re(z * w̄).
map_one
A unital map sends 1 to 1: f(1) = 1.
Complex.conj_re
The real part of the complex conjugate equals the real part: Re(z̄) = Re(z).
Complex.normSq_apply
The squared norm of a complex number is the sum of the squares of its real and imaginary parts: |z|² = Re(z)² + Im(z)².
mul_eq_zero.mp
If a product is zero, then at least one factor is zero: a * b = 0 implies a = 0 or b = 0.
Complex.ext
Two complex numbers are equal if their real parts and imaginary parts are respectively equal.
mul_right_cancel₀
In a structure with cancellation, if a * c = b * c and c ≠ 0, then a = b.

Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.

Lean source view module on GitHub

theorem three_hadamard_clean_return_phases (u v : ℂ)    (hu : Complex.normSq u = 1) (hv : Complex.normSq v = 1)    (hreturn : 1 - u + v + u * v = 0) :    u = v ∧ (v = Complex.I ∨ v = -Complex.I) := by  have he : 1 + v = u * (1 - v) := by linear_combination hreturn  have hn := congrArg Complex.normSq he  rw [map_mul, hu, one_mul] at hn  have hr : v.re = 0 := by    simp only [Complex.normSq_add, Complex.normSq_sub, hv, map_one,      Complex.conj_re, one_mul] at hn    linarith  have hi : v.im = 1 ∨ v.im = -1 := by    rw [Complex.normSq_apply, hr] at hv    have hmul : (v.im - 1) * (v.im + 1) = 0 := by nlinarith [hv]    rcases mul_eq_zero.mp hmul with h | h    · left; linarith    · right; linarith  have hvv : v = Complex.I ∨ v = -Complex.I := by    rcases hi with hi | hi    · left; apply Complex.ext <;> simp [hr, hi]    · right; apply Complex.ext <;> simp [hr, hi]  have huv : u = v := by    have hsq : v * v = -1 := by      rcases hvv with h | h <;> rw [h] <;> simp    have hne : 1 - v ≠ 0 := by      intro h      have : v = 1 := by linear_combination -h      rw [this] at hr      norm_num at hr    apply mul_right_cancel₀ hne    calc      u * (1 - v) = 1 + v := he.symm      _ = v * (1 - v) := by linear_combination hsq  exact ⟨huv, hvv⟩
See it in the graphDiscuss this resultChallenge it