Some Bounds Related to the 2 2 -adic Littlewood Conjecture
Dinis Vitorino, Ingrid Vukusic · 2025
Read Some Bounds Related to the 2 2 -adic Littlewood Conjecture by Dinis Vitorino, Ingrid Vukusic (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-2506-04110.txt.
Introduction
Dinis Vitorino and Ingrid Vukusic sharpen bounds tied to a number theory conjecture in a 2025 paper. The two-adic Littlewood conjecture says no irrational number can keep bounded continued fraction digits under all doublings. A past result showed any counterexample must allow a bound of at least eight. The authors raise that floor.
Their proof pushes the needed bound to fifteen for the main form of the conjecture. For a variant based on limit behavior, they prove a floor of five. A classic algorithm for doubling continued fractions carries the argument. Families of quadratic irrationals with long matching chains also appear along the way.
Catalog record
| CREATOR | Dinis Vitorino, Ingrid Vukusic |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 61 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-04110
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Some Bounds Related to the 2 2 -adic Littlewood Conjecture Thanks: Research of the second author supported by NSERC grant 2024-03725. Dinis Vitorino, Ingrid Vukusic Abstract . For every irrational real α \alpha , let M ( α ) = sup n ≥ 1 a n ( α ) M(\alpha)=\sup_{n\geq 1}a_{n}(\alpha) denote the largest partial quotient in its continued fraction expansion (or ∞ \infty , if unbounded). The 2 2 -adic Littlewood conjecture (2LC) can be stated as follows: There exists no irrational α \alpha such that M ( 2 k α ) M(2^{k}\alpha) is uniformly bounded by a constant C C for all k ≥ 0 k\geq 0 . Badziahin proved (considering a different formulation of 2LC) that if a counterexample exists, then the bound C C is at least 8 8 . We improve this bound to 15 15 . Then we focus on a “B-variant” of 2LC, where we replace M ( α ) M(\alpha) by B ( α ) = lim sup n → ∞ a n ( α ) B(\alpha)=\limsup_{n\to\infty}a_{n}(\alpha) . In this setting, we prove that if B ( 2 k α ) ≤ C B(2^{k}\alpha)\leq C for all k ≥ 0 k\geq 0 , then C ≥ 5 C\geq 5 . For the proof we use Hurwitz’s algorithm for multiplication of continued fractions by 2. Along the way, we find families of quadratic irrationals α \alpha with the property that for arbitrarily large K K there exist β , 2 β , 4 β , … , 2 K β \beta,2\beta,4\beta,\ldots,2^{K}\beta all equivalent to α \alpha . 2020 Mathematics Subject Classification 11A55, 10A30, 11J13 1. Introduction Every irrational real number \alpha can be uniquely represented by its infinite simple continued fraction expansion \alpha=[a_{0};a_{1},a_{2},\ldots]=a_{0}+\frac{1}{a_{1}+\frac{1}{a_{2}+\frac{1}{\ddots}}}, where a_{0} is an integer and a_{1},a_{2},\ldots are positive integers. Continued fractions have many applications in various fields of mathematics, and they have been studied extensively for centuries, see, for example, [ 14 ] for a reference book. For \alpha=[a_{0};a_{1},a_{2},\ldots] , the integers a_{n} are called the partial quotients of \alpha , and we denote the n -th partial quotient of any real irrational x by a_{n}(x) . For every bound C\geq 1 , it is easy to construct irrational real numbers \alpha with a_{n}(\alpha)\leq C for all n\geq 1 . Such numbers are called badly approximable . If we define M(\alpha):=\sup_{n\geq 1}a_{n}(\alpha), then \alpha being badly approximable is equivalent to M(\alpha)<\infty . Badly approximable numbers play an important role in the field of Diophantine approximation, and while they are easy to construct, it is in general a very hard problem to decide whether a given real number is badly approximable or not, and in fact “most real numbers” are not. When multiplying a badly approximable number \alpha by an integer k , the partial quotients might increase, but k\alpha will be badly approximable as well. For example, if M(\alpha)=C , then M(2\alpha)\leq 2C+3 . This follows, e.g., from [ 15 , Cor. 4] . From 1 in Section 5 one can see that actually M(2\alpha)\leq 2M(\alpha)+2 . Sometimes, we can even multiply a number by 2 repeatedly without increasing the maximal partial quotients at all. For example, if we set \alpha=(\sqrt{17}-1)/8 , we have \alpha=[0;2,\overline{1,1,3}] , 2\alpha=[\overline{1;3,1}] , and 4\alpha=[1;\overline{1,1,3}] , where the bar indicates that the digits repeat periodically. Now the following might seem like an innocent question: Problem 1 . Does there exist a constant C and a real irrational \alpha such that M(2^{k}\alpha)\leq C for all k\geq 0 ? It turns out that this question is not innocent at all. Any reader familiar with the mixed Littlewood conjecture will know that 1 is in fact equivalent to the 2 -adic Littlewood conjecture; see the next section for more background. We do not prove or disprove the 2 -adic Littlewood conjecture in this paper, but pose and answer questions inspired by 1 . In the next section, we give more background on the relation between the 2 -adic Littlewood conjecture and other problems. We also introduce a variant of 1 , where M(\alpha) is replaced with B(\alpha):=\limsup_{n\to\infty}a_{n}(\alpha). The smallest number C that can satisfy the property in 1 has been studied in [ 2 ] . In Section 3 , we slightly improve the best known lower bound on it [ 2 , Theorem 1.2] . Moreover, motivated by our proof, we give an equivalent formulation of 1 . In Section 4 , we recall some basic facts about continued fractions. In Section 5 , we precisely describe an algorithm for multiplying continued fractions by 2 . This algorithm is well known; however, we present it in a way which we find more intuitive. This enables us to prove lower bounds for the B -variant of 1 in Section 7 . Along the way, we find out why digit patterns like \overline{3,1,1} from the example \alpha=(\sqrt{17}-1)/8 mentioned above are special. In fact, in Section 6 we find an infinite family of digit patterns with the same special property. Our journey also leads to a characterization of all numbers with the property B(\alpha)\leq 2 and B(2\alpha)\leq 2 ( Section 7 ). Finally, in Section 8 , we state some open problems. 2. The p -adic Littlewood conjecture and related problems Let \left\|x\right\|=\min_{k\in\mathbb{Z}}\lvert k-x\rvert denote the distance of x to the nearest integer. As mentioned in the introduction, numbers with the property M(\alpha)<\infty are called badly approximable. Equivalently, one can define m(\alpha):=\inf_{q\geq 1}q\left\|q\alpha\right\|, and say that \alpha is badly approximable if and only if m(\alpha)>0 . Since this is related to 1 and the 2 -adic Littlewood conjecture, let us briefly check the equivalence. For \alpha=[a_{0};a_{1},a_{2},\ldots] we denote the truncation of \alpha at its n-th partial quotient, [a_{0};a_{1},\ldots,a_{n}] , by p_{n}/q_{n} . This is called the n -th convergent to \alpha . Convergents are famously particularly good approximations to \alpha . In fact, they are “best approximations” to \alpha in the sense that if q_{n}\leq q<q_{n+1} then \left\|q\alpha\right\|\geq\left\|q_{n}\alpha\right\|=\lvert q_{n}\alpha-p_{n}\rvert . Moreover, the following estimates are well known for n\geq 0 : (1) \frac{1}{(a_{n+1}+2)q_{n}}<\left\|q_{n}\alpha\right\|<\frac{1}{a_{n+1}q_{n}}. Multiplying the inequalities by q_{n} , using the best approximation property, taking the infimum, and recalling that M(\alpha)=\sup_{n\geq 1}a_{n}(\alpha) , we obtain (2) \frac{1}{M(\alpha)+2}\leq m(\alpha)\leq\frac{1}{M(\alpha)}. In particular, m(\alpha)>0 if and only if M(\alpha)<\infty , as desired. The p -adic Littlewood conjecture ( p LC) is a special case of the mixed Littlewood conjecture, which was introduced by de Mathan and Teuliè in 2004 [ 5 ] . Let \lvert n\rvert_{p}=p^{-v_{p}(n)} denote the p -adic absolute value of n , where v_{p}(n)=\max\{k\colon p^{k}\mid n\} is the p -adic valuation of the integer n . Conjecture 1 ( p LC) . For every prime p and irrational real \alpha , we have \inf_{q\geq 1}q\cdot\lvert q\rvert_{p}\cdot\left\|q\alpha\right\|=0. We briefly check that deciding whether 1 is true for p=2 is equivalent to solving 1 . We remark that of course we could also formulate 1 for general p ; however, in this paper we choose to focus on the case p=2 throughout. First, note that \displaystyle\inf_{q\geq 0}q\cdot\lvert q\rvert_{2}\cdot\left\|q\alpha\right\| \displaystyle=\inf_{\begin{subarray}{c}k\geq 0,\ q^{\prime}\geq 0,\\ \gcd(q^{\prime},2)=1\end{subarray}}2^{k}q^{\prime}\cdot\lvert 2^{k}q^{\prime}\rvert_{2}\cdot\left\|2^{k}q^{\prime}\alpha\right\| \displaystyle=\inf_{\begin{subarray}{c}k\geq 0,\ q^{\prime}\geq 0,\\ \gcd(q^{\prime},2)=1\end{subarray}}q^{\prime}\cdot\left\|2^{k}q^{\prime}\alpha\right\|=\inf_{\begin{subarray}{c}k\geq 0\\ q\geq 1\end{subarray}}q\cdot\left\|q\cdot 2^{k}\alpha\right\|=\inf_{k\geq 0}m(2^{k}\alpha). Then we see from inequality