standardized_laplacian_order_four_sector_bounds
Every complex eigenvalue of an order-four standardized Laplacian satisfies both the sector inequality at zero and its reflection at one.
Statement
theorem standardized_laplacian_order_four_sector_bounds (L : Matrix (Fin 4) (Fin 4) ℝ)
(hL : is_standardized_laplacian 4 L) (z : ℂ)
(hz : z ∈ spectrum ℂ (L.map (algebraMap ℝ ℂ))) :
|z.im| ≤ z.re ∧ z.re + |z.im| ≤ 1- Source
- Sector-bound formalization for IN-166; arXiv:math/0508176, Theorem 5.
- Verified
- 27 Oct 2026
- Axioms
Classical.choiceQuot.soundpropext- Built on 7
- matrix_order_four_centered_shiftAdd a scalar multiple of the identity minus the averaging matrix to an order-four real matrix.
- matrix_order_four_centered_shift_row_sumA centered scalar shift preserves zero row sums.
- matrix_order_four_centered_shift_mulVecThe action of a centered scalar shift on a complex vector is its original image plus the scalar multiple of the vector minus its average.
- matrix_row_sum_zero_shifted_eigenvectorFor real order-four matrices with zero row sums, an eigenvector of one with nonzero eigenvalue can be adjusted by a constant to give an…
- matrix_order_four_mem_spectrum_of_eigenpairA nonzero complex eigenvector witnesses membership in the spectrum of an order-four real matrix.
- standardized_laplacian_order_four_im_le_reFor any standardized Laplacian matrix L of order 4, the imaginary part and real part of every complex eigenvalue z of L satisfy |Im(z)| ≤…
- is_standardized_laplacianA standardized Laplacian matrix of order n is a real n x n matrix with zero row sums and off-diagonal entries in [-1/n, 0].
- Used by 1
- standardized_laplacian_order_four_two_fractional_sectorFor an order-4 two-fractional standardized Laplacian, every complex eigenvalue satisfies the sector bound |Im z| ≤ Re z, and dually Re z +…
The proof, explained
The theorem proves that every eigenvalue z of an order-four standardized Laplacian L satisfies both |Im(z)| ≤ Re(z) and Re(z) + |Im(z)| ≤ 1. The first inequality is the known sector bound at zero. For the second, after treating z = 0 and z = 1 trivially, the proof constructs the transformed matrix K = -L + I - (1/4) J, which is shown to be another order-four standardized Laplacian. An eigenvector of L with eigenvalue z yields an eigenvector of K with eigenvalue 1 - z, and applying the sector bound at zero to K produces |Im(1 - z)| ≤ Re(1 - z), which rearranges directly to Re(z) + |Im(z)| ≤ 1.
|z.im| ≤ z.re, reducing the goal to proving z.re + |z.im| ≤ 1
The sector bound at zero (|z.im| ≤ z.re) is supplied directly by the helper theorem standardized_laplacian_order_four_im_le_re. We then split the conjunction to leave only the upper sector bound z.re + |z.im| ≤ 1 to be proved.
uses
standardized_laplacian_order_four_im_le_rehave hleft := standardized_laplacian_order_four_im_le_re L hL z hz refine ⟨hleft, ?_⟩
z.re + |z.im| ≤ 1 holds when z = 0
If z = 0, both the real and imaginary parts vanish, and 0 + 0 ≤ 1 holds trivially by simplification.
by_cases hz0 : z = 0 · simp [hz0]
z.re + |z.im| ≤ 1 holds when z = 1
If z = 1, Re(1) = 1 and Im(1) = 0, so 1 + 0 ≤ 1 holds by simplification.
by_cases hz1 : z = 1 · simp [hz1]
The matrix K, defined as the centered shift by 1 of -L, has row sums equal to zero
We define K_ij = (-L)_ij + δ_ij - 1/4. Since each row of L sums to 0, negating each entry shows -L has zero row sums as well. Applying matrix_order_four_centered_shift_row_sum shows that the centered shift operation preserves this zero row-sum property.
uses
Matrix.neg_apply,Finset.sum_neg_distrib,neg_zero,congrArg,Neg.neg,matrix_order_four_centered_shift_row_sumlet K := matrix_order_four_centered_shift (-L) 1 have hrneg : ∀ i, ∑ j, (-L) i j = 0 := by intro i simpa only [Matrix.neg_apply, Finset.sum_neg_distrib, neg_zero] using congrArg Neg.neg (hL.1 i) have hrK : ∀ i, ∑ j, K i j = 0 := matrix_order_four_centered_shift_row_sum (-L) hrneg 1
K is a standardized Laplacian of order four
Its row sums are zero from the previous step. For off-diagonal entries i ≠ j, the diagonal shift is absent, so K_ij = -L_ij - 1/4. Because L is a standardized Laplacian, its off-diagonal entries satisfy -1/4 ≤ L_ij ≤ 0, which implies 0 ≤ -L_ij ≤ 1/4 and therefore -1/4 ≤ -L_ij - 1/4 ≤ 0, verified by linear arithmetic.
uses
matrix_order_four_centered_shift,if_neg,Matrix.neg_applyhave hK : is_standardized_laplacian 4 K := by refine ⟨hrK, ?_⟩ intro i j hij dsimp only [K, matrix_order_four_centered_shift] rw [if_neg hij] simp only [Matrix.neg_apply] have hb := hL.2 i j hij constructor <;> linarith [hb.1, hb.2]
There is a non-zero eigenvector v of L such that L v = z v
We identify the matrix spectrum of L with the operator spectrum of its associated linear map via Matrix.spectrum_toLin'. Because the space is finite-dimensional, z is an eigenvalue, giving a non-zero eigenvector v satisfying L v = z v.
uses
Matrix.spectrum_toLin',Module.End.HasEigenvalue.of_mem_spectrum,exists_hasEigenvector,Module.End.mem_eigenspace_iff.mphave he : z ∈ spectrum ℂ (Matrix.toLin' (L.map (algebraMap ℝ ℂ))) := by rwa [Matrix.spectrum_toLin'] obtain ⟨v, hve, hv0⟩ := (Module.End.HasEigenvalue.of_mem_spectrum he).exists_hasEigenvector have hv : (L.map (algebraMap ℝ ℂ)).mulVec v = z • v := Module.End.mem_eigenspace_iff.mp hve
The matrix K acts on v as K v = (1 - z) v - ((∑ j, v j) / 4) 1
From L v = z v, we have (-L) v = -z v. The formula matrix_order_four_centered_shift_mulVec expresses K v as (-L) v + 1 • v - (1/4)(∑ j, v j) 1, which simplifies by ring operations to (1 - z) v - ((∑ j, v j) / 4) 1.
uses
Matrix.mulVec,dotProduct,matrix_order_four_centered_shift_mulVec,Pi.neg_apply,Pi.smul_apply,smul_eq_mul,Complex.ofReal_onehave hneg : ((-L).map (algebraMap ℝ ℂ)).mulVec v = -(z • v) := by rw [← hv] ext i simp [Matrix.mulVec, dotProduct] have hKv : (K.map (algebraMap ℝ ℂ)).mulVec v = fun i => (1 - z) * v i - (∑ j, v j) / 4 := by rw [matrix_order_four_centered_shift_mulVec, hneg] ext i simp only [Pi.neg_apply, Pi.smul_apply, smul_eq_mul, Complex.ofReal_one] ring
1 - z is an eigenvalue of K with some non-zero eigenvector w
Both L and K have zero row sums, z ≠ 0, and 1 - z ≠ 0 (since z ≠ 1). The affine eigenvector relation K v = (1 - z) v + c 1 with c = -(∑ j, v j)/4 implies, via matrix_row_sum_zero_shifted_eigenvector, that K has an honest non-zero eigenvector w for eigenvalue 1 - z.
uses
matrix_row_sum_zero_shifted_eigenvector,sub_ne_zero.mpr,Ne.symmobtain ⟨w, hw0, hw⟩ := matrix_row_sum_zero_shifted_eigenvector L K hL.1 hrK v hv0 z hz0 hv (1 - z) (-(∑ j, v j) / 4) (sub_ne_zero.mpr (Ne.symm hz1)) (by rw [hKv] ext i ring)
z.re + |z.im| ≤ 1 holds, completing the theorem
Since 1 - z is an eigenvalue of the order-four standardized Laplacian K, the sector bound at zero applied to K gives |Im(1 - z)| ≤ Re(1 - z). Simplifying using Im(1 - z) = -Im(z) and Re(1 - z) = 1 - Re(z) yields |z.im| ≤ 1 - z.re, from which linear arithmetic concludes z.re + |z.im| ≤ 1.
uses
standardized_laplacian_order_four_im_le_re,matrix_order_four_mem_spectrum_of_eigenpair,Complex.sub_im,Complex.one_im,zero_sub,abs_neg,Complex.sub_re,Complex.one_rehave hright := standardized_laplacian_order_four_im_le_re K hK (1 - z) (matrix_order_four_mem_spectrum_of_eigenpair K w hw0 _ hw) simp only [Complex.sub_im, Complex.one_im, zero_sub, abs_neg, Complex.sub_re, Complex.one_re] at hright linarith
abs_neg- The imaginary absolute value is bounded by the negation: |-a| = |a|.
Complex.one_im- The imaginary part of the complex number 1 is 0.
Complex.one_re- The real part of the complex number 1 is 1.
Complex.ofReal_one- The real number 1 coerced to the complex numbers is the complex number 1.
Complex.sub_im- The imaginary part of z - w is Im(z) - Im(w).
Complex.sub_re- The real part of z - w is Re(z) - Re(w).
congrArg- Function application respects equality: if a = b, then f(a) = f(b).
dotProduct- The dot product of two vectors, ∑ i, u i * v i.
exists_hasEigenvector- An eigenvalue of an endomorphism has an associated non-zero eigenvector.
Finset.sum_neg_distrib- Negation distributes across finite sums: ∑ x, -f(x) = -(∑ x, f(x)).
if_neg- If a condition c is false, then (if c then t else e) reduces to e.
matrix_order_four_centered_shift- The centered shift of a 4x4 matrix L with parameter b, defined as L_ij + (if i = j then b else 0) - b/4.
matrix_order_four_centered_shift_mulVec- Action of the centered shift of L on a vector v: K v = L v + b v - (b/4)(∑ j, v j) 1.
matrix_order_four_centered_shift_row_sum- If L has zero row sums, then its centered shift with any shift parameter b also has zero row sums.
matrix_order_four_mem_spectrum_of_eigenpair- If a 4x4 matrix has a non-zero eigenvector with eigenvalue z, then z belongs to its spectrum.
matrix_row_sum_zero_shifted_eigenvector- If L and K have zero row sums, z ≠ 0, λ ≠ 0, L v = z v, and K v = λ v + c 1, then K has an eigenvector with eigenvalue λ.
Matrix.mulVec- Matrix-vector multiplication: (A v)_i = ∑ j, A_ij v_j.
Matrix.neg_apply- The (i, j)-entry of -A is -(A_ij).
Matrix.spectrum_toLin'- The spectrum of a square matrix equals the spectrum of its associated linear endomorphism.
Module.End.HasEigenvalue.of_mem_spectrum- Over a finite-dimensional space over a field, every element of the spectrum of an endomorphism is an eigenvalue.
Module.End.mem_eigenspace_iff.mp- A vector v is in the eigenspace of f for μ if and only if f(v) = μ v.
Ne.symm- Inequality is symmetric: if a ≠ b, then b ≠ a.
Neg.neg- The negation function x ↦ -x.
neg_zero- In any additive group, -0 = 0.
Pi.neg_apply- Pointwise negation of a function: (-f)(i) = -f(i).
Pi.smul_apply- Pointwise scalar multiplication of a function: (a • f)(i) = a • f(i).
smul_eq_mul- Scalar multiplication in a ring agrees with ring multiplication: a • b = a * b.
standardized_laplacian_order_four_im_le_re- For any order-four standardized Laplacian L and eigenvalue z, |Im(z)| ≤ Re(z).
sub_ne_zero.mpr- If a ≠ b, then a - b ≠ 0.
zero_sub- In any subtractive group, 0 - a = -a.
- The proof transfers between matrix eigenvalues and linear endomorphism eigenvalues (via Matrix.toLin') to extract an eigenvector, and relies on the technical helper matrix_row_sum_zero_shifted_eigenvector to turn the affine relation Kv = (1 - z)v + c 1 into a genuine eigenpair Kw = (1 - z)w.
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.
theorem standardized_laplacian_order_four_sector_bounds (L : Matrix (Fin 4) (Fin 4) ℝ) (hL : is_standardized_laplacian 4 L) (z : ℂ) (hz : z ∈ spectrum ℂ (L.map (algebraMap ℝ ℂ))) : |z.im| ≤ z.re ∧ z.re + |z.im| ≤ 1 := by have hleft := standardized_laplacian_order_four_im_le_re L hL z hz refine ⟨hleft, ?_⟩ by_cases hz0 : z = 0 · simp [hz0] by_cases hz1 : z = 1 · simp [hz1] let K := matrix_order_four_centered_shift (-L) 1 have hrneg : ∀ i, ∑ j, (-L) i j = 0 := by intro i simpa only [Matrix.neg_apply, Finset.sum_neg_distrib, neg_zero] using congrArg Neg.neg (hL.1 i) have hrK : ∀ i, ∑ j, K i j = 0 := matrix_order_four_centered_shift_row_sum (-L) hrneg 1 have hK : is_standardized_laplacian 4 K := by refine ⟨hrK, ?_⟩ intro i j hij dsimp only [K, matrix_order_four_centered_shift] rw [if_neg hij] simp only [Matrix.neg_apply] have hb := hL.2 i j hij constructor <;> linarith [hb.1, hb.2] have he : z ∈ spectrum ℂ (Matrix.toLin' (L.map (algebraMap ℝ ℂ))) := by rwa [Matrix.spectrum_toLin'] obtain ⟨v, hve, hv0⟩ := (Module.End.HasEigenvalue.of_mem_spectrum he).exists_hasEigenvector have hv : (L.map (algebraMap ℝ ℂ)).mulVec v = z • v := Module.End.mem_eigenspace_iff.mp hve have hneg : ((-L).map (algebraMap ℝ ℂ)).mulVec v = -(z • v) := by rw [← hv] ext i simp [Matrix.mulVec, dotProduct] have hKv : (K.map (algebraMap ℝ ℂ)).mulVec v = fun i => (1 - z) * v i - (∑ j, v j) / 4 := by rw [matrix_order_four_centered_shift_mulVec, hneg] ext i simp only [Pi.neg_apply, Pi.smul_apply, smul_eq_mul, Complex.ofReal_one] ring obtain ⟨w, hw0, hw⟩ := matrix_row_sum_zero_shifted_eigenvector L K hL.1 hrK v hv0 z hz0 hv (1 - z) (-(∑ j, v j) / 4) (sub_ne_zero.mpr (Ne.symm hz1)) (by rw [hKv] ext i ring) have hright := standardized_laplacian_order_four_im_le_re K hK (1 - z) (matrix_order_four_mem_spectrum_of_eigenpair K w hw0 _ hw) simp only [Complex.sub_im, Complex.one_im, zero_sub, abs_neg, Complex.sub_re, Complex.one_re] at hright linarith