A sum-bracket theorem for simple Lie algebras (2204.02018v2)
Abstract: Let $\mathfrak{g}$ be an algebra over $K$ with a bilinear operation $[\cdot,\cdot]:\mathfrak{g}\times\mathfrak{g}\rightarrow\mathfrak{g}$ not necessarily associative. For $A\subseteq\mathfrak{g}$, let $A{k}$ be the set of elements of $\mathfrak{g}$ written combining $k$ elements of $A$ via $+$ and $[\cdot,\cdot]$. We show a "sum-bracket theorem" for simple Lie algebras over $K$ of the form $\mathfrak{g}=\mathfrak{sl}{n},\mathfrak{so}{n},\mathfrak{sp}{2n},\mathfrak{e}{6},\mathfrak{e}{7},\mathfrak{e}{8},\mathfrak{f}{4},\mathfrak{g}{2}$: if $\mathrm{char}(K)$ is not too small, we have growth of the form $|A{k}|\geq|A|{1+\varepsilon}$ for all generating symmetric sets $A$ away from subfields of $K$. Over $\mathbb{F}_{p}$ in particular, we have a diameter bound matching the best analogous bounds for groups of Lie type [BDH21]. As an independent intermediate result, we prove also an estimate of the form $|A\cap V|\leq|A{k}|{\dim(V)/\dim(\mathfrak{g})}$ for linear affine subspaces $V$ of $\mathfrak{g}$. This estimate is valid for all simple algebras, and $k$ is especially small for a large class of them including associative, Lie, and Mal'cev algebras, and Lie superalgebras.