Search

Let $\textrm{Mat}_2(\mathbb{R})$ be the set of $2 \times 2$ matrices with real entries. For any $\varepsilon>0$ and any finitely--supported probability measure $\mu$ on $\textrm{Mat}_2(\mathbb{R})$, we prove that either \[ T(\mu) = \sum_{X, Y \in {\rm supp}(\mu), XY = YX} \mu(X) \mu(Y) < \varepsilon \] or there exists some finite set $S$ contained in a $2$-dimensional subspace of $\textrm{Mat}_2(\mathbb{R})$ such that $\mu(S) \gg \varepsilon^{3}$. We prove quantitatively stronger results when \[ \mu ( (a_{i,j})_{1\leq i,j \leq 2}) = \nu(a_{1,1}) \dots \nu(a_{2,2}) \ \ \text{for every} \ a_{1,1}, \dots, a_{2,2} \in \mathbb{R}, \] with $\nu$ being some finitely--supported probability measure on $\mathbb{R}$. For instance, when $A \subset \mathbb{R}$ is a proper generalised arithmetic progression of dimension $d$, and $\nu = 1_{A}/|A|$, our methods imply that $T(\mu) \ll_d |A|^{-3}$. This is optimal up to the $O_d(1)$ factor. Our proofs involve incidence geometric techniques as well as ideas surrounding the recent progress on Bourgain--Chang type sum-product estimates over $\mathbb{R}$. The latter includes Schmidt's subspace theorem and the resolution of the weak polynomial Freiman--Ruzsa conjecture over integers., Comment: 17 pages