real_cubic_im_le_of_discriminant_bound
For a real monic cubic X^3-a X^2+b X-c, if a^2-3b is at least -3 and its discriminant plus 4*(a^2-3b+4)^2 is nonnegative, every complex root has imaginary part of absolute value at most one.
Statement
theorem real_cubic_im_le_of_discriminant_bound
(a b c : ℝ) (z : ℂ)
(hz : z ^ 3 - (a : ℂ) * z ^ 2 + (b : ℂ) * z - (c : ℂ) = 0)
(hD : -3 ≤ a ^ 2 - 3 * b)
(hH : 0 ≤ a ^ 2 * b ^ 2 - 4 * b ^ 3 - 4 * a ^ 3 * c - 27 * c ^ 2 +
18 * a * b * c + 4 * (a ^ 2 - 3 * b + 4) ^ 2) :
|z.im| ≤ 1- Source
- Supporting cubic algebra for standardized_laplacian_order_four_im_le (OP-465; arXiv:2610.14100 Questions 1 and 3). The two-chord family is from Section 5; experiment 4367 records its explicit factorization.
- Verified
- 24 Oct 2026
- Axioms
Classical.choiceQuot.soundpropext- Used by 1
- two_adjacent_chord_cubic_im_leEvery complex root of the nonzero cubic factor of the unscaled directed four-cycle with adjacent chord weights t,u has imaginary part of…
The proof, explained
The theorem is proved by contradiction: assume the imaginary part y = Im(z) satisfies |y| > 1, so y ≠ 0 and y^2 > 1. Writing the root as z = x + iy allows the cubic equation's real and imaginary parts to express the coefficients b and c in terms of a, x, and y. Substituting these into the discriminant condition factors it into 4*(1 - y^2)*B ≥ 0, where B is a sum of squares; since 1 - y^2 < 0, this forces B = 0 and in particular (a - 3*x)^2 = 0. Finally, substituting b into the bound a^2 - 3*b ≥ -3 simplifies it to -3*y^2 ≥ -3, i.e., y^2 ≤ 1, contradicting y^2 > 1.
The real part of the cubic equation evaluates to x^3 - 3*x*y^2 - a*(x^2 - y^2) + b*x - c = 0, where x = Re(z) and y = Im(z).
Setting x = Re(z) and y = Im(z), we take the real part of both sides of z^3 - a*z^2 + b*z - c = 0 using congrArg Complex.re. Expanding powers and complex operations using basic arithmetic lemmas for ℂ and applying non-linear arithmetic yields the equation for the real part.
uses
congrArg,Complex.re,Complex.add_re,Complex.sub_re,Complex.mul_re,Complex.mul_im,Complex.ofReal_re,Complex.ofReal_im,pow_succ,pow_zero,Complex.one_re,Complex.one_im,Complex.zero_relet x := z.re let y := z.im have hre : x ^ 3 - 3 * x * y ^ 2 - a * (x ^ 2 - y ^ 2) + b * x - c = 0 := by have h := congrArg Complex.re hz simp only [Complex.add_re, Complex.sub_re, Complex.mul_re, Complex.mul_im, Complex.ofReal_re, Complex.ofReal_im, pow_succ, pow_zero, Complex.one_re, Complex.one_im, Complex.zero_re] at h dsimp [x, y] nlinarith [h]
The imaginary part of the cubic equation evaluates to y*(3*x^2 - y^2 - 2*a*x + b) = 0.
Similarly, taking the imaginary part of both sides of z^3 - a*z^2 + b*z - c = 0 using congrArg Complex.im and expanding with complex arithmetic identities, non-linear arithmetic yields the factored imaginary equation y*(3*x^2 - y^2 - 2*a*x + b) = 0.
uses
congrArg,Complex.im,Complex.add_re,Complex.add_im,Complex.sub_re,Complex.sub_im,Complex.mul_re,Complex.mul_im,Complex.ofReal_re,Complex.ofReal_im,pow_succ,pow_zero,Complex.one_re,Complex.one_im,Complex.zero_imhave him : y * (3 * x ^ 2 - y ^ 2 - 2 * a * x + b) = 0 := by have h := congrArg Complex.im hz simp only [Complex.add_re, Complex.add_im, Complex.sub_re, Complex.sub_im, Complex.mul_re, Complex.mul_im, Complex.ofReal_re, Complex.ofReal_im, pow_succ, pow_zero, Complex.one_re, Complex.one_im, Complex.zero_im] at h dsimp [x, y] nlinarith [h]
Assuming |y| > 1 for contradiction implies y^2 > 1 and y ≠ 0.
We rephrase the goal |z.im| ≤ 1 as |y| ≤ 1 and assume for contradiction that |y| > 1. Squaring gives y^2 = |y|^2 > 1, which in particular forces y ≠ 0 since 0^2 is not greater than 1.
uses
lt_of_not_ge,sq_abschange |y| ≤ 1 by_contra hbad have hy2 : 1 < y ^ 2 := by have ha : 1 < |y| := lt_of_not_ge hbad nlinarith [sq_abs y] have hy : y ≠ 0 := by intro h; rw [h] at hy2; norm_num at hy2
The coefficients b and c satisfy b = y^2 - 3*x^2 + 2*a*x and c = x^3 - 3*x*y^2 - a*(x^2 - y^2) + b*x.
Since the imaginary equation is y*(3*x^2 - y^2 - 2*a*x + b) = 0 and y ≠ 0, the second factor must vanish, giving b = y^2 - 3*x^2 + 2*a*x. Rearranging the real part equation then isolates c.
uses
mul_eq_zero.mphave hb : b = y ^ 2 - 3 * x ^ 2 + 2 * a * x := by have h := (mul_eq_zero.mp him).resolve_left hy linarith have hc : c = x ^ 3 - 3 * x * y ^ 2 - a * (x ^ 2 - y ^ 2) + b * x := by linarith [hre]
The discriminant expression factors as 4*(1 - y^2)*[((a - 3*x)^2 + y^2 - 4)^2 + 16*(a - 3*x)^2] ≥ 0.
Substituting the formulas for b and c into the modified discriminant expression, polynomial ring normalization proves that it factors identically as 4*(1 - y^2)*B, where B = ((a - 3*x)^2 + y^2 - 4)^2 + 16*(a - 3*x)^2. Rewriting hypothesis hH with this identity gives 0 ≤ 4*(1 - y^2)*B.
have hfactor : a ^ 2 * b ^ 2 - 4 * b ^ 3 - 4 * a ^ 3 * c - 27 * c ^ 2 + 18 * a * b * c + 4 * (a ^ 2 - 3 * b + 4) ^ 2 = 4 * (1 - y ^ 2) * (((a - 3 * x) ^ 2 + y ^ 2 - 4) ^ 2 + 16 * (a - 3 * x) ^ 2) := by rw [hc, hb] ring rw [hfactor] at hH
The factor B = ((a - 3*x)^2 + y^2 - 4)^2 + 16*(a - 3*x)^2 is zero.
As a sum of two squares of real numbers, B is nonnegative. If B were nonzero, it would be strictly positive; but since y^2 > 1, the factor 4*(1 - y^2) is strictly negative. The product of a negative and a positive number would be strictly negative, contradicting 0 ≤ 4*(1 - y^2)*B. Hence B = 0.
uses
lt_of_le_of_ne,Ne.symm,mul_neg_of_neg_of_poslet B := ((a - 3 * x) ^ 2 + y ^ 2 - 4) ^ 2 + 16 * (a - 3 * x) ^ 2 have hB : 0 ≤ B := by dsimp [B]; positivity have hBzero : B = 0 := by by_contra hn have hp : 0 < B := by exact lt_of_le_of_ne hB (Ne.symm hn) have hneg : 4 * (1 - y ^ 2) < 0 := by nlinarith have hprod := mul_neg_of_neg_of_pos hneg hp change 0 ≤ 4 * (1 - y ^ 2) * B at hH linarith
(a - 3*x)^2 = 0.
Since B = ((a - 3*x)^2 + y^2 - 4)^2 + 16*(a - 3*x)^2 = 0 and squares of real numbers are nonnegative, the term 16*(a - 3*x)^2 must vanish, so (a - 3*x)^2 = 0.
uses
sq_nonneghave hsq : (a - 3 * x) ^ 2 = 0 := by dsimp [B] at hBzero nlinarith [sq_nonneg ((a - 3 * x) ^ 2 + y ^ 2 - 4)]
A contradiction is reached with the hypothesis -3 ≤ a^2 - 3*b.
Substituting b = y^2 - 3*x^2 + 2*a*x into a^2 - 3*b rewrites it as (a - 3*x)^2 - 3*y^2. Since (a - 3*x)^2 = 0, this expression is -3*y^2. The hypothesis -3 ≤ a^2 - 3*b then asserts -3 ≤ -3*y^2, which implies y^2 ≤ 1, contradicting y^2 > 1.
rw [hb] at hD nlinarith [hsq]
congrArg- If a₁ = a₂, then f(a₁) = f(a₂) for any function f.
Complex.re- The projection from ℂ to ℝ giving the real part of a complex number.
Complex.im- The projection from ℂ to ℝ giving the imaginary part of a complex number.
Complex.add_re- The real part of a sum of complex numbers is the sum of their real parts: Re(z + w) = Re(z) + Re(w).
Complex.sub_re- The real part of a difference of complex numbers is the difference of their real parts: Re(z - w) = Re(z) - Re(w).
Complex.mul_re- The real part of a product of complex numbers: Re(z * w) = Re(z) * Re(w) - Im(z) * Im(w).
Complex.mul_im- The imaginary part of a product of complex numbers: Im(z * w) = Re(z) * Im(w) + Im(z) * Re(w).
Complex.ofReal_re- The real part of a real number cast to ℂ is the real number itself.
Complex.ofReal_im- The imaginary part of a real number cast to ℂ is zero.
pow_succ- In a monoid, a^(n + 1) = a^n * a.
pow_zero- In a monoid, a^0 = 1.
Complex.one_re- The real part of 1 ∈ ℂ is 1.
Complex.one_im- The imaginary part of 1 ∈ ℂ is 0.
Complex.zero_re- The real part of 0 ∈ ℂ is 0.
Complex.zero_im- The imaginary part of 0 ∈ ℂ is 0.
Complex.add_im- The imaginary part of a sum of complex numbers is the sum of their imaginary parts: Im(z + w) = Im(z) + Im(w).
Complex.sub_im- The imaginary part of a difference of complex numbers is the difference of their imaginary parts: Im(z - w) = Im(z) - Im(w).
lt_of_not_ge- In a linear order, if b is not at most a, then a < b.
sq_abs- In an ordered ring, |a|^2 = a^2.
mul_eq_zero.mp- If a product is zero, at least one of the factors is zero: a * b = 0 implies a = 0 or b = 0.
lt_of_le_of_ne- In a partial order, if a ≤ b and a ≠ b, then a < b.
Ne.symm- Inequality is symmetric: if a ≠ b, then b ≠ a.
mul_neg_of_neg_of_pos- The product of a strictly negative number and a strictly positive number is strictly negative: a < 0 and 0 < b imply a * b < 0.
sq_nonneg- The square of any element in an ordered semiring is nonnegative: 0 ≤ a^2.
- The algebraic identity in lines 37–42, which factors the discriminant-like expression into 4(1 - y^2)B, is verified by polynomial normalization (the ring tactic).
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
Choose a tactic — a marked line — to see the goals Lean had before and after it, as in Lean’s infoview. Step with ← → or j k.
set_option maxHeartbeats 1500000theorem real_cubic_im_le_of_discriminant_bound (a b c : ℝ) (z : ℂ) (hz : z ^ 3 - (a : ℂ) * z ^ 2 + (b : ℂ) * z - (c : ℂ) = 0) (hD : -3 ≤ a ^ 2 - 3 * b) (hH : 0 ≤ a ^ 2 * b ^ 2 - 4 * b ^ 3 - 4 * a ^ 3 * c - 27 * c ^ 2 + 18 * a * b * c + 4 * (a ^ 2 - 3 * b + 4) ^ 2) : |z.im| ≤ 1 := by let x := z.re let y := z.im have hre : x ^ 3 - 3 * x * y ^ 2 - a * (x ^ 2 - y ^ 2) + b * x - c = 0 := by have h := congrArg Complex.re hz simp only [Complex.add_re, Complex.sub_re, Complex.mul_re, Complex.mul_im, Complex.ofReal_re, Complex.ofReal_im, pow_succ, pow_zero, Complex.one_re, Complex.one_im, Complex.zero_re] at h dsimp [x, y] nlinarith [h] have him : y * (3 * x ^ 2 - y ^ 2 - 2 * a * x + b) = 0 := by have h := congrArg Complex.im hz simp only [Complex.add_re, Complex.add_im, Complex.sub_re, Complex.sub_im, Complex.mul_re, Complex.mul_im, Complex.ofReal_re, Complex.ofReal_im, pow_succ, pow_zero, Complex.one_re, Complex.one_im, Complex.zero_im] at h dsimp [x, y] nlinarith [h] change |y| ≤ 1 by_contra hbad have hy2 : 1 < y ^ 2 := by have ha : 1 < |y| := lt_of_not_ge hbad nlinarith [sq_abs y] have hy : y ≠ 0 := by intro h; rw [h] at hy2; norm_num at hy2 have hb : b = y ^ 2 - 3 * x ^ 2 + 2 * a * x := by have h := (mul_eq_zero.mp him).resolve_left hy linarith have hc : c = x ^ 3 - 3 * x * y ^ 2 - a * (x ^ 2 - y ^ 2) + b * x := by linarith [hre] have hfactor : a ^ 2 * b ^ 2 - 4 * b ^ 3 - 4 * a ^ 3 * c - 27 * c ^ 2 + 18 * a * b * c + 4 * (a ^ 2 - 3 * b + 4) ^ 2 = 4 * (1 - y ^ 2) * (((a - 3 * x) ^ 2 + y ^ 2 - 4) ^ 2 + 16 * (a - 3 * x) ^ 2) := by rw [hc, hb] ring rw [hfactor] at hH let B := ((a - 3 * x) ^ 2 + y ^ 2 - 4) ^ 2 + 16 * (a - 3 * x) ^ 2 have hB : 0 ≤ B := by dsimp [B]; positivity have hBzero : B = 0 := by by_contra hn have hp : 0 < B := by exact lt_of_le_of_ne hB (Ne.symm hn) have hneg : 4 * (1 - y ^ 2) < 0 := by nlinarith have hprod := mul_neg_of_neg_of_pos hneg hp change 0 ≤ 4 * (1 - y ^ 2) * B at hH linarith have hsq : (a - 3 * x) ^ 2 = 0 := by dsimp [B] at hBzero nlinarith [sq_nonneg ((a - 3 * x) ^ 2 + y ^ 2 - 4)] rw [hb] at hD nlinarith [hsq]