𝔤𝔩𝑛𝕂

Lie algebra of nilpotent endomorphisms

Let 𝔤 𝔤𝔩𝑛𝕂 be a linear Lie algebra for which every 𝑥 𝔤 is a nilpotent endomorphism of 𝕂𝑛. Then there exists some nonzero 𝑣 𝕂𝑛 for which 𝔤𝑣 =0.1 #m/thm/lie

Proof

Let 𝔤 𝔤𝔩𝑛𝕂 be a linear Lie algebra. Assume that for dim𝔤 <𝑚, every 𝑥 𝔤 being nilpotent implies the existence of some nonzero 𝑣 𝕂𝑛 such that 𝔤𝑣 =0. Note that this clearly holds for 𝑚 =2.

Now take dim𝔤 =𝑚, and let 𝔥 <𝔤 be a strict subalgebra, so that dim𝔥 <𝑚. Then by Nilpotent transformations have nilpotent adjoint representations, 𝔤 acts on 𝔤 under the adjoint representation nilpotently, as does 𝔥 on 𝔤/𝔥: Thus we have a Lie algebra homomorphism

𝜋:𝔥𝔤𝔩(𝔤/𝔥)

such that 𝜋(𝔥) contains only nilpotent endomorphisms. Since dim𝜋(𝔥) <𝑚 satisfies the induction hypothesis, there exists a nonzero 𝑥 +𝔥 𝔤/𝔥 such that 𝜋(𝔥)(𝑥 +𝔥) =𝔥, or equivalently, the normalizer 𝔫𝔤(𝔥) is a strict superset of 𝔥.

Taking 𝔥 to be a maximal strict subalgebra, it follows that 𝔫𝔤(𝔥) =𝔥, thus 𝔥 is an ideal of codimension one: Hence 𝔤 =𝔥 +𝕂𝑧 for any 𝑧 𝔤 𝔥. Let 𝑊 ={𝑣 𝑉 :𝔥𝑣 =0} be the subspace of vectors annihilated by 𝔥. Since 𝔥 is an ideal, this is invariant under 𝔤 =𝔥 +𝕂𝑧, and since 𝑧 is nilpotent, it has an eigenvector 𝑣 𝑊 such that 𝑧𝑣 =0, and therefore 𝔤𝑣 =0 as required.


#state/tidy | #lang/en | #SemBr

Footnotes

  1. 1972. Introduction to Lie Algebras and Representation Theory