The distribution of intersections in SL ( 3 , ℤ ) \ SL ( 3 , ℝ ) \mathrm{SL}(3,\mathbb{Z})\backslash\mathrm{SL}(3,\mathbb{R}) and lattices related to roots of cubic congruences
Matthew Welsh · 2025
Read The distribution of intersections in SL ( 3 , ℤ ) \ SL ( 3 , ℝ ) \mathrm{SL}(3,\mathbb{Z})\backslash\mathrm{SL}(3,\mathbb{R}) and lattices related to roots of cubic congruences by Matthew Welsh (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-00898.txt.
Introduction
Matthew Welsh, writing alone, studies intersections in a quotient space built from three-by-three matrices. The matrices have determinant one, over integers and reals. He looks at translates of closed orbits of the positive diagonal subgroup. Those translates meet a maximal parabolic subgroup. The pattern of those meetings is his main object.
The meetings connect closely to roots of cubic congruences. The polynomials are monic and irreducible, with integer coefficients. Welsh shows the meetings are jointly equidistributed. That means they spread evenly across both subgroups. It follows that certain affine lattices are also jointly equidistributed. Those lattices are set by pairs of congruence roots. They correspond to ideals in the ring of integers. The methods parallel work on multidimensional Farey sequences.
Catalog record
| CREATOR | Matthew Welsh |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-00898
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
The distribution of intersections in SL ( 3 , ℤ ) \ SL ( 3 , ℝ ) \mathrm{SL}(3,\mathbb{Z})\backslash\mathrm{SL}(3,\mathbb{R}) and lattices related to roots of cubic congruences Matthew Welsh Abstract . In this note we study the distribution of the intersections between certain translates of closed orbits of the positive diagonal subgroup in SL ( 3 , ℤ ) \ SL ( 3 , ℝ ) \mathrm{SL}(3,\mathbb{Z})\backslash\mathrm{SL}(3,\mathbb{R}) with a maximal parabolic subgroup. These intersections are closely connected to roots of congruences for certain monic, irreducible cubic polynomials F ( X ) ∈ ℤ [ X ] F(X)\in\mathbb{Z}[X] . The the main result is that the intersections, considered as a sequences in the diagonal subgroup and the parabolic subgroup, are jointly equidistributed. This implies that certain affine lattices determined by pairs of roots of the cubic congruences are jointly equidistributed with corresponding ideals in the associated ring of integers. We note that the techniques here roughly parallel those which has been developed to study the multidimensional Farey sequence, and one hopes that techniques to study roots of congruences will continue to develop. 1. Introduction Let G=\mathrm{SL}(3,\mathbb{R}) and \Gamma=\mathrm{SL}(3,\mathbb{Z}) . The positive diagonal subgroup A\subset G , which we parametrize by (1) a(t)=\begin{pmatrix}\mathrm{e}^{-2t}&0&0\\ 0&\mathrm{e}^{t}&0\\ 0&0&\mathrm{e}^{t}\end{pmatrix},\ a_{1}(t_{1})=\begin{pmatrix}1&0&0\\ 0&\mathrm{e}^{-t_{1}}&0\\ 0&0&\mathrm{e}^{t_{1}}\end{pmatrix} has closed orbits in \Gamma\backslash G through special points \Gamma g_{l} associated to ideal classes [I_{l}] in totally real cubic orders, see eg Linnik’s classic book [ 13 , Chapter VII] , or, for a more modern treatment, works by Einsiedler, Lindenstrauss, Michel, and Venkatesh [ 5 , 6 ] . In this short note, our interest in these closed orbits of A comes from their connection to roots of cubic congruences, solutions \mu\bmod m to the congruence F(\mu)\equiv 0\bmod m with F an integer, cubic polynomial. Theorem 1.1 below connects these congruence-roots (under certain assumptions on F ) with the points in the intersection (2) \Gamma g_{l}Ag_{0}^{-1}\cap\Gamma P, where g_{0} is a basis matrix of the cubic order associated to the polynomial F and (3) P=\{\begin{pmatrix}*&*&*\\ 0&*&*\\ 0&*&*\end{pmatrix}\in G\}. We roughly order the points in ( 2 ) by considering (4) \Gamma g_{l}Ag_{0}^{-1}\cap\bigcup_{t\leq T}\Gamma P_{0}a(t) where (5) P_{0}=\{\begin{pmatrix}1&*&*\\ 0&*&*\\ 0&*&*\end{pmatrix}\in G\}. Our main theorem, theorem 1.2 below, is that the coordinates of the points in ( 4 ), considered as a sequence in A\times P_{0} are jointly equidistributed as T\to\infty . In terms of the congruence-roots, this implies that certain affine lattices in \mathrm{ASL}(2,\mathbb{Z})\backslash\mathrm{ASL}(2,\mathbb{R}) related to pairs of roots become equidistributed. While this may be of some interest, it falls short of a primary goal: a dynamical understanding of the distribution of the congruence-roots themselves. Nevertheless, theorem 1.2 draws interesting parallels to work done on the two-dimensional Farey sequence, which we review in the following section. 1.1. Background on the Farey sequence 1.1.1. \mathrm{SL}(2) setting Our methods and results are analogous to the study of the multidimensional Farey sequence, or equivalently, the rational points on closed horospheres. The analogy is most clear in the \mathrm{SL}(2) setting due to the identification of \mathrm{SL}(2,\mathbb{R})/\{\pm I\} with the unit tangent bundle of the upper half plane \mathbb{H} . In Iwasawa coordinates, we have (6) \pm\begin{pmatrix}1&x\\ 0&1\end{pmatrix}\begin{pmatrix}y^{\frac{1}{2}}&0\\ 0&y^{-\frac{1}{2}}\end{pmatrix}\begin{pmatrix}\cos\theta&-\sin\theta\\ \sin\theta&\cos\theta\end{pmatrix}\mapsto(x+\mathrm{i}y,\mathrm{i}y\mathrm{e}^{-2\mathrm{i}\theta}), where on the right we have written the unit tangent bundle of \mathbb{H} as the set of all pairs (z,w)\in\mathbb{C}^{2} with z\in\mathbb{H} and |w|=\mathrm{Im}(z) , so that w has unit length in the hyperbolic metric. Under this identification, the set of matrices of the form \begin{pmatrix}1&*\\ 0&1\end{pmatrix} corresponds to a horizontal horocycle with tangent vectors pointing vertically upwards. The image of this set under \begin{pmatrix}a&b\\ c&d\end{pmatrix}\in\Gamma is another horocycle, which will not be horizontal if c\neq 0 . In this case, the top of the horocycle is the point with tangent vector pointing vertically downwards. The calculation (7) \begin{pmatrix}a&b\\ c&d\end{pmatrix}\begin{pmatrix}1&-\frac{d}{c}\\ 0&1\end{pmatrix}=\begin{pmatrix}1&\frac{a}{c}\\ 0&1\end{pmatrix}\begin{pmatrix}\frac{1}{c}&0\\ 0&c\end{pmatrix}\begin{pmatrix}0&-1\\ 1&0\end{pmatrix} then shows that this top is \tfrac{a}{c}+\tfrac{i}{c^{2}} . The sequence of Farey fractions \tfrac{a}{c} may therefore be studied as the tops of the images of a horocycle under \Gamma or as the points of intersection between the sets \Gamma\begin{pmatrix}1&*\\ 0&1\end{pmatrix} and \Gamma\begin{pmatrix}*&*\\ 0&*\end{pmatrix}\begin{pmatrix}0&-1\\ 1&0\end{pmatrix} . This latter perspective was successfully used by Athreya and Cheung [ 2 ] to study gap distribution of Farey fractions and the BCZ map [ 3 ] , and the method can be extended to lattices other than \mathrm{SL}(2,\mathbb{Z}) , see [ 1 , 18 ] for example. If instead one starts with a the geodesic in \mathbb{H} from -\sqrt{2} to \sqrt{2} , then the tops of the images of this geodesic are points \frac{\mu}{m}+\mathrm{i}\frac{\sqrt{2}}{m} with \mu^{2}\equiv 2\bmod m . This connection between the tops of the \Gamma -images of a geodesic and roots of the congruence is used to study the distribution of these roots in [ 17 ] , [ 12 ] and can be interpreted in terms of the intersections between (8) \Gamma\frac{1}{(2\sqrt{2})^{\frac{1}{2}}}\begin{pmatrix}\sqrt{2}&-\sqrt{2}\\ 1&1\end{pmatrix}\begin{pmatrix}*&0\\ 0&*\end{pmatrix}\mathrm{\ and\ }\Gamma\begin{pmatrix}*&*\\ 0&*\end{pmatrix}\begin{pmatrix}\cos\frac{\pi}{4}&\pm\sin\frac{\pi}{4}\\ \mp\sin\frac{\pi}{4}&\cos\frac{\pi}{4}\end{pmatrix}, the latter corresponding to points with horizontal tangent vectors. With some modification to account for nontrivial narrow ideal classes of \mathbb{Z}[\sqrt{D}] , this extends to the roots of the quadratic congruence \mu^{2}\equiv D\bmod m with D>1 squarefree. These roots are therefore naturally analogous to the Farey fractions with tops of horocycles being replaced with tops of geodesics, and they can be studied by similar techniques. This analogy extends, in a somewhat simpler manner, to quadratic congruences with negative D as well, but with geodesics/horocycles being replaced with \mathrm{SO}(2) cosets of \mathrm{SL}(2,\mathbb{R}) , ie points in \mathbb{H} . 1.1.2. SL(3) setting The analogy requires more significant modification to extend in \mathrm{SL}(3,\mathbb{R}) to cubic congruences and the two-dimensional Farey sequence. Instead of roots of cubic congruences themselves, our primary result concerns certain lattices associated to the roots, see theorem 1.1 . The two-dimensional Farey fractions are pairs of rational numbers (\frac{r_{1}}{q},\frac{r_{2}}{q}) with q>0 and \gcd(q,r_{1},r_{2})=1 , ordered by the denominator q . Considered modulo \mathbb{Z}^{2} , these points, like the classic 1-dimensional Farey sequence, are easily shown to equidistribute on the torus \mathbb{T}^{2}=\mathbb{R}^{2}/\mathbb{Z}^{2} . Marklof [ 15 ] studies the fine-scale distribution of these points and their higher-dimensional counterparts using dynamics on \mathrm{SL}(3,\mathbb{Z})\backslash\mathrm{SL}(3,\mathbb{R}) and higher-rank homogeneous spaces. A key observation in the method is that the points in the intersections (9) \Gamma N\cap\bigcup_{t\leq T}\Gamma P_{0}a(t), where (10) N=\{n(x_{1},x_{2})=\begin{pmatrix}1&0&0\\ x_{1}&1&0\\ x_{2}&0&1\end{pmatrix}:x_{1},x_{2}\in\mathbb{R}\}, occur at n(\bm{x})\in N with \bm{x} a Farey fraction with denominator q\leq\mathrm{e}^{2T} . Moreover, the \Gamma\backslash\Gamma P_{0}\cong\mathrm{ASL}(2,\mathbb{Z})\backslash\mathrm{ASL}(2,\mathbb{R}) coordinates corresponding to the Farey fraction through this intersection provide information on the fine-scale distribution of the of the Farey points, see [ 15 ] . Limit theorems on the fine-scale distribution follow from the joint equidistribution of the Farey points and these points in \mathrm{ASL}(2,\mathbb{Z})\backslash\mathrm{ASL}(2,\mathbb{R}) as T\to\infty . Several subsequent works have studied these points using techniques from dynamics [ 7 ] as well as Fourier-analytic methods [ 11 , 8 ] . In addition to studying the problem in any dimension, these works are able to prove this joint equidistribution for the points with a single denominator q\to\infty as opposed to all denominators q\leq\mathrm{e}^{2T} . For the problem with an individual q , the relevant intersection, ( 9 ) with a fixed t , is between two-dimensional and five-dimensional homogeneous subspaces of an eight-dimensional space. These intersections are therefore considered “unlikely” and require additional