Nilpotent skew-symmetric tridiagonal Bohemian matrices: a proof of Corless's conjecture
Marcel Blattner
math.CO
Nov 18, 2026 · v1
math.RA
TL;DR
The main theorem (Corless's nilpotency conjecture) and its counting corollary are formalised in Lean 4 with Mathlib, in about 1,150 lines.
Abstract
Corless studied the skew-symmetric tridiagonal matrices $T_m(u)$ with superdiagonal $u \in \{1, i\}^{m-1}$. He showed that $T_m(u)$ can be nilpotent only if $m = 2^K - 1$ and that the $2^{K-1}$ vectors $u$ obtained from the empty vector by repeated application of $u \mapsto (u, 1, i, \rev u)$ and $u \mapsto (u, i, 1, \rev u)$ give nilpotent matrices, and he conjectured that there are no others. We prove this conjecture. Reducing the characteristic polynomial modulo $4$ gives a necessary condition for nilpotency that is linear over $\F$. For $m > 1$ it forces $u = (v, 1, i, \rev v)$ or $u = (v, i, 1, \rev v)$, and an identity implicit in Corless's proof then reduces the problem to $v$. Since $T_m(u)$ is diagonally similar to a finite section of the Feinberg–Zee hopping matrix, the theorem also determines the nilpotent tridiagonal sign matrices. We further show that the $(1,1)$ entry of the resolvent of a nilpotent $T_m(u)$ is $λ^{-1} \pm λ^{-3} \pm λ^{-7} \pm \dots \pm λ^{-(2^K - 1)}$, with signs that determine $u$. The proof of the main theorem has been formalised in Lean 4.
Problem
Corless conjectured that the skew-symmetric tridiagonal matrices T_m(u) with superdiagonal entries in {1, i} are nilpotent only for the 2^{K-1} vectors produced by his recursive doubling construction, with m = 2^K - 1. The conjecture had been checked by exhaustive computation only for m ≤ 31.
Approach
The characteristic polynomial is expressed through a signed matching polynomial of a path. Reducing it modulo 4 gives a necessary condition for nilpotency that is linear over F_2. This condition forces a folded structure on u, and an identity from Corless's proof then reduces the problem to the shorter vector v. The proof of the main theorem was formalised in Lean 4 with Mathlib, with the theorem defined through an inductive predicate CorlessVec.
Results
The conjecture is proved, and the result also classifies the nilpotent tridiagonal sign matrices. The Lean theorems corless_conjecture and card_nilpotent depend only on Lean's standard axioms. An exhaustive search confirms the case m = 63, and the (1,1) resolvent entry is shown to determine u.