A family of accumulation points of non-free rational numbers
Christopher (Bobby) Buyalos, Jayden Thadani, Xinbei Wang, Bradley Zykoski, Michael Zshornack · 2025
Read A family of accumulation points of non-free rational numbers by Christopher (Bobby) Buyalos, Jayden Thadani, Xinbei Wang, Bradley Zykoski, Michael Zshornack (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-2511-10422.txt.
Introduction
Christopher Buyalos and four co-authors study a question about groups built from two simple matrices, one fixed and one depending on a number q. Kim and Koberda conjectured that for every rational q between negative four and four, the two matrices never generate a free group. The paper builds families of rational values where the conjecture holds.
Earlier work by Smilga is extended through three different constructions. One uses Diophantine geometry on certain polynomials, one uses sums of geometric series, and one uses ratios of Pell-type sequences. The families pile up at countless points inside the interval, adding strong support for the conjecture.
Catalog record
| CREATOR | Christopher (Bobby) Buyalos, Jayden Thadani, Xinbei Wang, Bradley Zykoski, Michael Zshornack |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 30 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-10422
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A family of accumulation points of non-free rational numbers Christopher (Bobby) Buyalos, Jayden Thadani, Xinbei Wang, Bradley Zykoski, Michael Zshornack Abstract . For any q ∈ ℝ q\in\mathbb{R} , let A := ( 1 1 0 1 ) , B q := ( 1 0 q 1 ) A:=\big(\begin{smallmatrix}1&1\\ 0&1\end{smallmatrix}\big),B_{q}:=\big(\begin{smallmatrix}1&0\\ q&1\end{smallmatrix}\big) and let G q := ⟨ A , B q ⟩ ⩽ SL ( 2 , ℝ ) G_{q}:=\langle A,B_{q}\rangle\leqslant\operatorname{SL}(2,\mathbb{R}) . Kim and Koberda conjecture that for every q ∈ ℚ ∩ ( − 4 , 4 ) q\in\mathbb{Q}\cap(-4,4) , the group G q G_{q} is not freely generated by these two matrices. We generalize work of Smilga and construct families of q q satisfying the conjecture that accumulate at infinitely many different points in ( − 4 , 4 ) (-4,4) . We give different constructions of such families, the first coming from applying tools in Diophantine geometry to certain polynomials arising in Smilga’s work, the second from sums of geometric series and the last from ratios of Pell and Half-Companion Pell Numbers accumulating at 1 + 2 1+\sqrt{2} . 1. Introduction For any q\in\mathbb{R} , let A:=\left(\begin{smallmatrix}1&1\\ 0&1\end{smallmatrix}\right),B_{q}:=\left(\begin{smallmatrix}1&0\\ q&1\end{smallmatrix}\right) and let G_{q}:=\langle A,B_{q}\rangle\leqslant\operatorname{SL}(2,\mathbb{R}). In general, we are far from completely understanding the dependence of the group-theoretic properties of G_{q} on the choice of parameter q . When |q|\geq 4 , a simple ping-pong argument shows that G_{q} is both discrete and freely generated by A and B_{q} , but for |q|<4 , much less is known about the overall structure of these groups. For one, it is not known if such a freely generated example exists for |q|<4 , though such an example would necessarily be indiscrete. The freeness question for G_{q} was initially explored in [ 4 ] and it was later conjectured in [ 3 ] that for all q\in\mathbb{Q} with |q|<4 , G_{q} is never isomorphic to a free group of rank 2 . For concreteness, we define the set \mathcal{R}_{\mathbb{Q}}:=\left\{q\in\mathbb{Q}\,:\,G_{q}\textrm{ is not freely generated by }A\textrm{ and }B_{q}\right\}. We refer to elements q\in\mathcal{R}_{\mathbb{Q}} as non-free rational numbers. In [ 3 ] , the authors pose the following conjecture. Conjecture 1 . \mathcal{R}_{\mathbb{Q}}=\mathbb{Q}\cap(-4,4) . Kim and Koberda give evidence for this conjecture by establishing it for all q\in\mathbb{Q}\cap(-4,4) whose numerator, in least terms, is at most 27 and not 24 , along with exhibiting positive density subsets of rationals with arbitrary fixed numerator such that the conjecture holds. Despite this progress, a number of seemingly basic properties of the set \mathcal{R}_{\mathbb{Q}} remain mysterious. For instance, it is still unknown if \mathcal{R}_{\mathbb{Q}} is dense in [-4,4] , or even if \pm 4\in\overline{\mathcal{R}_{\mathbb{Q}}} . To the authors’ best understanding it is even unknown whether \overline{\mathcal{R}_{\mathbb{Q}}} contains any interval or has positive measure. Short of establishing this conjecture in full, it is fruitful to try and show \mathcal{R}_{\mathbb{Q}} contains larger and larger subsets of \mathbb{Q}\cap(-4,4) than previously known. For instance, in [ 8 ] , the author produces novel sequences in \mathcal{R}_{\mathbb{Q}} which limit to “large” elements of [-4,4] . Note that, for instance, while the results in [ 3 ] provide a wealth of new examples of non-free rational numbers, a priori, these all accumulate at 0 . In practice, it is much harder to find examples of non-free rationals which accumulate at large elements of [-4,4] , such as \pm 4 . Smilga’s examples show that both \pm(2+\sqrt{2}) and \pm\frac{3+\sqrt{5}}{2} are such accumulation points and, at the time of writing, are among the largest elements of [-4,4] known to be contained in \overline{\mathcal{R}_{\mathbb{Q}}} . This paper extends the methods of [ 8 ] to construct new families of elements in \mathcal{R}_{\mathbb{Q}} which accumulate at new points in [-4,4] . None of the examples we produce are larger than those in [ 8 ] (our largest example is 2 ) but the main result constructs sequences of non-free rational numbers accumulating at a family of rational numbers naturally arising in the work of Kim–Koberda. See Theorem 9 for a more precise statement. Theorem 2 . For nonzero integers r,s,t , we have that \frac{r+t}{rst} is a limit of a non-constant sequence of non-free rational numbers. 1.1. Sketch of the proof In [ 8 ] , the author shows that a relation of a particular form in A and B_{q} holding is equivalent to q being a root of a polynomial in a certain family. The roots of the quadratic polynomials in this family can be explicitly determined, but for these roots to be rational, a certain discriminant must be a perfect square. To construct non-free rational numbers accumulating at many different points in [-4,4] , we use work of [ 7 ] to show that for some of the quadratic polynomials in this family, this Diophantine obstruction is described by an algebraic curve containing infinitely many integral points. Consequently, we produce many q\in\mathbb{Q} satisfying half-relations of the form in [ 8 ] . The non-free rational numbers produced depend on a tuple of parameters, (a_{1},a_{2},a_{3},a_{4},a_{5}) which, when sent to infinity appropriately, produce the various accumulation points in Theorem 2 . 1.2. Groups generated by opposite parabolic matrices A subgroup of \operatorname{SL}(2,\mathbb{Q}) generated by a pair of parabolic matrices stabilizing distinct lines in \mathbb{R}\mathbb{P}^{1} is always conjugate over \operatorname{SL}(2,\mathbb{R}) to a group of the form G_{q} for some choice of q . Such a conjugacy need not preserve the group being contained in \operatorname{SL}(2,\mathbb{Q}) and so, more generally, one can study similar questions explored here about subgroups generated by any pair of distinct parabolic matrices with rational entries. Recent work on questions asked by Greenberg and Shalom on commensurators of discrete subgroups of semisimple Lie groups bears a number of consequences on the structure of such groups. In [ 1 ] , the authors consider the question (dubbed the Greenberg–Shalom Hypothesis ) of whether Zariski-dense discrete subgroups of semisimple Lie groups with almost dense commensurator are necessarily arithmetic lattices. This question generalizes a characterization of arithmeticity among lattices originally due to Margulis and in particular, if true, would imply that Conjecture 1 holds. More specifically (c.f. Theorem 4.8 in [ 1 ] ), if q=\frac{r}{s}\in\mathbb{Q}\cap(-4,4) is written in least terms, then the Greenberg–Shalom Hypothesis implies that G_{q} is a finite-index subgroup of the S -arithmetic group \operatorname{SL}(2,\mathbb{Z}[\frac{1}{s}]) , and in particular, is not isomorphic to a free group. Further evidence for both the Greenberg–Shalom Hypothesis and Conjecture 1 is given in [ 5 ] where the author shows, among other results, that G_{q} is indeed finite index in the corresponding S -arithmetic group for q\in\mathbb{Q}\cap(-4,4) whose numerator, in least terms, is at most 4 . 1.3. Some further questions We close with a number of questions that have not been recorded in previous literature on this topic, but we feel are worth further exploration. First, the rational numbers in the statement of Theorem 2 are of note because they arise as some of the simplest types of relation non-free within a hierarchy introduced in [ 3 ] . See § 2.1 for a precise definition but, in short, these 1 -step relation numbers are among the elements of \mathcal{R}_{\mathbb{Q}} for which it is particularly easy to exhibit a relation in the corresponding group. It is worth