root_lattice_form_det
det(I_k − 11ᵀ/(k+1)) = 1/(k+1).
Statement
theorem root_lattice_form_det (k : ℕ) : (1 - (1 / ((k : ℝ) + 1)) • Matrix.of (fun _ _ : Fin k => (1 : ℝ))).det = 1 / ((k : ℝ) + 1)
- Source
- arXiv:2610.26970 (Gaussian Lower Bounds for Distinct Subset Sums in Higher Dimensions), Proposition 3.2: the determinant claim
- Verified
- 21 Oct 2026
- Axioms
Classical.choiceQuot.soundpropext
The proof, explained
The determinant of I_k − 1/(k+1) 11ᵀ is evaluated using the matrix determinant lemma for rank-one perturbations of the identity matrix. The matrix is rewritten as I_k + u vᵀ, where u = −1/(k+1) 1 and v = 1, so its determinant equals the scalar 1 + v ⬝ᵥ u. Evaluating the dot product as −k/(k+1) and simplifying 1 − k/(k+1) yields the result 1/(k+1).
The matrix I_k − 1/(k+1) J_k is equal to the rank-one perturbation I_k + u vᵀ, where u is the column vector with entries −1/(k+1) and v is the row vector of all ones.
We verify the matrix equality entry by entry for any indices i, j. By definition of matrix multiplication (Matrix.mul_apply), the (i, j)-th entry of the product of the single-column matrix u and the single-row matrix v is the single term (−1/(k+1)) * 1 = −1/(k+1). Adding this to the (i, j)-th entry of the identity matrix yields the entry of I_k − 1/(k+1) J_k, using the fact that subtraction is addition of the negation (sub_eq_add_neg).
uses
Matrix.mul_apply,sub_eq_add_neghave h : (1 - (1 / ((k : ℝ) + 1)) • Matrix.of (fun _ _ : Fin k => (1 : ℝ))) = 1 + Matrix.replicateCol (Fin 1) (fun _ : Fin k => -(1 / ((k : ℝ) + 1))) * Matrix.replicateRow (Fin 1) (fun _ : Fin k => (1 : ℝ)) := by ext i j simp [Matrix.mul_apply] exact sub_eq_add_neg _ _
The dot product v ⬝ᵥ u equals −k / (k + 1).
By definition of dotProduct, the dot product of the all-ones vector v and the constant vector u with entries −1/(k+1) is the sum over Fin k of the constant −1/(k+1). Since Fin k has cardinality k (Fintype.card_fin), summing this constant over all k indices gives k * (−1/(k+1)) by Finset.sum_const.
uses
dotProduct,Finset.sum_const,Fintype.card_finhave hdot : (fun _ : Fin k => (1 : ℝ)) ⬝ᵥ (fun _ : Fin k => -(1 / ((k : ℝ) + 1))) = (k : ℝ) * (-(1 / ((k : ℝ) + 1))) := by simp [dotProduct, Finset.sum_const, Fintype.card_fin]
The determinant of I_k − 1/(k+1) J_k equals 1 + v ⬝ᵥ u.
Rewriting the matrix as I_k + u vᵀ using the decomposition established in the first step, Sylvester's determinant formula for rank-one perturbations of the identity (Matrix.det_one_add_replicateCol_mul_replicateRow) reduces the determinant of the k × k matrix to the scalar 1 + v ⬝ᵥ u.
uses
Matrix.det_one_add_replicateCol_mul_replicateRowcalc (1 - (1 / ((k : ℝ) + 1)) • Matrix.of (fun _ _ : Fin k => (1 : ℝ))).det = (1 + Matrix.replicateCol (Fin 1) (fun _ : Fin k => -(1 / ((k : ℝ) + 1))) * Matrix.replicateRow (Fin 1) (fun _ : Fin k => (1 : ℝ))).det := by rw [h] _ = 1 + (fun _ : Fin k => (1 : ℝ)) ⬝ᵥ (fun _ : Fin k => -(1 / ((k : ℝ) + 1))) := Matrix.det_one_add_replicateCol_mul_replicateRow _ _
The scalar expression 1 + v ⬝ᵥ u simplifies to 1 / (k + 1).
Substituting the dot product v ⬝ᵥ u = −k / (k+1) gives 1 − k/(k+1). Since k ≥ 0, the denominator k + 1 is strictly positive and therefore nonzero (hk, via positivity). Clearing denominators with field_simp and simplifying the resulting polynomial (k + 1) − k with ring yields 1 / (k + 1).
_ = 1 / ((k : ℝ) + 1) := by rw [hdot] have hk : ((k : ℝ) + 1) ≠ 0 := by positivity field_simp ring
Matrix.mul_apply- The (i, k)-th entry of a matrix product MN is the sum ∑_j M_ij N_jk.
sub_eq_add_neg- Subtraction equals addition of the additive inverse: a - b = a + (-b).
dotProduct- The dot product of two vectors indexed by a finite type is the sum of their pointwise products, ∑_i v_i u_i.
Finset.sum_const- The sum of a constant term b over a finite set s is equal to |s| • b.
Fintype.card_fin- The standard finite type Fin n has cardinality n.
Matrix.det_one_add_replicateCol_mul_replicateRow- For vectors u and v, the determinant of the rank-one perturbation I + uvᵀ is 1 + v ⬝ᵥ u.
- In Lean, the rank-one matrix uvᵀ is represented as `Matrix.replicateCol (Fin 1) u * Matrix.replicateRow (Fin 1) v`, which models vectors of length k as k × 1 and 1 × k matrices indexed by the singleton type `Fin 1`.
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 root_lattice_form_det (k : ℕ) : (1 - (1 / ((k : ℝ) + 1)) • Matrix.of (fun _ _ : Fin k => (1 : ℝ))).det = 1 / ((k : ℝ) + 1) := by have h : (1 - (1 / ((k : ℝ) + 1)) • Matrix.of (fun _ _ : Fin k => (1 : ℝ))) = 1 + Matrix.replicateCol (Fin 1) (fun _ : Fin k => -(1 / ((k : ℝ) + 1))) * Matrix.replicateRow (Fin 1) (fun _ : Fin k => (1 : ℝ)) := by ext i j simp [Matrix.mul_apply] exact sub_eq_add_neg _ _ have hdot : (fun _ : Fin k => (1 : ℝ)) ⬝ᵥ (fun _ : Fin k => -(1 / ((k : ℝ) + 1))) = (k : ℝ) * (-(1 / ((k : ℝ) + 1))) := by simp [dotProduct, Finset.sum_const, Fintype.card_fin] calc (1 - (1 / ((k : ℝ) + 1)) • Matrix.of (fun _ _ : Fin k => (1 : ℝ))).det = (1 + Matrix.replicateCol (Fin 1) (fun _ : Fin k => -(1 / ((k : ℝ) + 1))) * Matrix.replicateRow (Fin 1) (fun _ : Fin k => (1 : ℝ))).det := by rw [h] _ = 1 + (fun _ : Fin k => (1 : ℝ)) ⬝ᵥ (fun _ : Fin k => -(1 / ((k : ℝ) + 1))) := Matrix.det_one_add_replicateCol_mul_replicateRow _ _ _ = 1 / ((k : ℝ) + 1) := by rw [hdot] have hk : ((k : ℝ) + 1) ≠ 0 := by positivity field_simp ring