Transcendence and measures via the refined Diophantine exponent
Quang-Khai Nguyen · 2026
Read Transcendence and measures via the refined Diophantine exponent by Quang-Khai Nguyen (2026) 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-2605-30606.txt.
Introduction
Nguyen introduces the refined Diophantine exponent, a measure of repeats in infinite words that detects weaker repeats than the classical exponent while keeping its strength. Adamczewski and Bugeaud introduced the classical exponent in 2007 to prove transcendence results for numbers whose base expansions are simple. The refined version applies where repetition is partly hidden by noise, joining earlier results of Corvaja and Zannier and of Kebis, Luca, Ouaknine, Scoones, and Worrell. Finer versions yield results on transcendence measures, spurred by recent work in algebraic dynamics.
Catalog record
| CREATOR | Quang-Khai Nguyen |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 132009 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.30606
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Transcendence and measures via the refined Diophantine exponent Quang-Khai Nguyen Abstract . In 2007, Adamczewski and Bugeaud introduced the notion of the Diophantine exponent of an infinite word as a quantitative measure of repetition, leading to new transcendence results for real numbers whose expansions in an integer base are sufficiently simple. In the present article, we introduce the refined Diophantine exponent , which detects weaker forms of repetition while preserving the full strength of the classical approach. This new exponent applies in situations where repetition is partially obscured by some form of noise . Related ideas already appear in the work of Corvaja and Zannier in 2002 and, more recently, in the works of Kebis, Luca, Ouaknine, Scoones, and Worrell. Our approach provides a unified framework that recovers and extends these results, as well as those of Adamczewski and Bugeaud. We also develop quantitative refinements of this method, leading to results about transcendence measures. The recent breakthrough of Bell, Diller, and Jonsson in the context of algebraic dynamics is partly based on a similar idea, which also served as a motivation for the present work. 1. Introduction The study of the Diophantine properties of values of analytic functions f_{\mathbf{a}}(z)=\sum_{i\geq 0}a_{i}z^{i}\in\overline{\mathbb{Q}}[[z]] at algebraic points is a central theme in transcendental number theory. A long-standing approach consists in exploiting combinatorial properties of the coefficients sequence \mathbf{a}=a_{0}a_{1}\cdots to investigate the transcendence of the values of f_{\mathbf{a}} , as well as more quantitative questions such as irrationality and transcendence measures. The underlying philosophy is that when the sequence \mathbf{a} is sufficiently simple , in the sense that it has a simple combinatorial or dynamical structure, it can often be well approximated by eventually periodic sequences. Consequently, the function f_{\mathbf{a}} can be well approximated by rational functions, which in turn implies that its values at algebraic points admit good algebraic approximations. From the perspective of Diophantine approximation, this typically leads to a dichotomy : such values are either transcendental, or algebraic of a very restricted form. This circle of ideas goes back to the pioneering work of Liouville, who proved the transcendence of numbers such as \sum_{i\geq 1}\frac{1}{10^{i!}}\,\cdot Later, deeper tools from Diophantine approximation led to further classes of transcendental numbers obtained in this way. Particularly significant are the works of Maillet [ 36 ] and Baker [ 12 ] on continued fractions, and, a few decades later, the work of Ferenczi and Mauduit [ 25 ] on Sturmian numbers. This approach acquired a new dimension in the works of Adamczewski and Bugeaud, and of Adamczewski, Bugeaud, and Luca, where the full strength of the p -adic Subspace Theorem [ 39 , 40 ] , which is a powerful extension of Schmidt’s Subspace Theorem [ 41 ] , was brought into the picture. The general strategy was formalized by Adamczewski and Bugeaud in [ 2 ] through the introduction of the notion of the Diophantine exponent , which measures the periodicity of an infinite word. The Diophantine exponent of a word \mathbf{a} , denoted {\bf Dio}(\mathbf{a}) , is defined as the supremum of the real numbers \rho for which there exist three sequences of integers (r_{n})_{n\geq 0} , (s_{n})_{n\geq 0} , and (t_{n})_{n\geq 0} satisfying, for all n\geq 0 , (i) -1\leq r_{n}<s_{n}<t_{n} ; (ii) the words \mathbf{a}[r_{n}+1,\,r_{n}+t_{n}-s_{n}] and \mathbf{a}[s_{n}+1,\,t_{n}] coincide; (iii) t_{n}\geq\rho\,s_{n} ; (iv) s_{n}-r_{n}\to\infty as n\to\infty . Here, \mathbf{a}[r,s] denotes the finite word a_{r}a_{r+1}\cdots a_{s} . The main result of [ 1 , 2 , 3 ] asserts that if \mathbf{a} takes values in a finite set of integers and if \beta is an algebraic number with |\beta|>1 such that {\bf Dio}(\mathbf{a})>\frac{\log M(\beta)}{\log|\beta|}\,, then the following dichotomy holds: • either f_{\mathbf{a}}(1/\beta) belongs to the number field \mathbb{Q}(\beta) , or it is transcendental. Here M(\beta) denotes the Mahler measure of \beta . The original definition of the Diophantine exponent is slightly different, but easily seen to be equivalent to the one given above. Setting U_{n}=\mathbf{a}[0,r_{n}] and V_{n}=\mathbf{a}[r_{n}+1,s_{n}] , Condition (i) implies that \mathbf{a} coincides with the eventually periodic sequence U_{n}V_{n}^{\infty} up to index t_{n} . When \beta is an integer, or more generally a Pisot or a Salem number, the above result only requires that {\bf Dio}(\mathbf{a})>1 . However, stronger repetitivity is needed to treat values of f_{\mathbf{a}} at more general algebraic points. In particular, in order to treat values at all algebraic numbers, one needs to assume that {\bf Dio}(\mathbf{a})=\infty , which is a very strong condition. In retrospect, a way to overcome this difficulty takes its source in a previous work by Corvaja and Zannier [ 23 ] on lacunary numbers, by applying the Subspace Theorem in higher dimensions. More recently, similar ideas have appeared, apparently independently, in several works [ 33 , 34 , 28 , 29 ] , where new combinatorial transcendence criteria are established for some class of coefficents sequences called echoing and stuterring over arbitrary algebraic bases. Such ideas also arise in the breakthrough work of Bell, Diller, and Jonsson [ 13 ] in the context of algebraic dynamics, where the authors construct the first example of a map with a transcendental dynamical degree. The present paper is motivated partly by these recent developments. We introduce a new exponent, called the refined Diophantine exponent and denoted by \mathbf{Rdio} , which is defined in the same spirit as the classical Diophantine exponent, but with Condition (ii) replaced by a weaker requirement (ii’): instead of asking the two words \mathbf{a}[r_{n}+1,r_{n}+t_{n}-s_{n}] and \mathbf{a}[s_{n}+1,t_{n}] to coincide, we only require them to be sufficiently close in a precise sense related to the Hamming distance as follows. To define the refined Diophantine exponent, we need to introduce the notion of (\epsilon,\delta) -closedness. Definition 1.1 . For two finite words U=u_{1}\cdots u_{L} and V=v_{1}\cdots v_{L} of the same length L>0 and for \epsilon>0 and integer \delta\geq 0 , we say that U and V are (\epsilon,\delta) -closed if there exists \delta subintervals I_{1},\ldots,I_{\delta} of \{1,\ldots,L\} such that (i) U and V agree outside \displaystyle\bigcup_{j=1}^{\delta}I_{j} , more precisely, \{1\leq i\leq L:u_{i}\neq v_{i}\}\subseteq\displaystyle\bigcup_{j=1}^{\delta}I_{j}; (ii) \Bigg|\displaystyle\bigcup_{j=1}^{\delta}I_{j}\Bigg|\leq\epsilon L . Here, we note that the intervals I_{j} can be empty. Definition 1.2 . Let \mathbf{a}=a_{0}a_{2}\cdots be an infinite word. Let \rho\geq 1 be a real number. We say that \mathbf{a} satisfies Condition (*)_{\rho} if for every \epsilon>0 , there exist three sequences of integers (r_{n})_{n\geq 0} , (s_{n})_{n\geq 0} , and (t_{n})_{n\geq 0} satisfying, for all n\geq 0 , (i) -1\leq r_{n}<s_{n}<t_{n} ; (ii’) there exists \delta\geq 0 such that for all n large enough, \mathbf{a}[r_{n}+1,r_{n}+t_{n}-s_{n}] is (\epsilon,\delta) -closed to \mathbf{a}[s_{n}+1,t_{n}] ; (iii) {t_{n}}\geq\rho{s_{n}} ; (iv) s_{n}-r_{n}\to\infty as n\to\infty . We define \mathbf{Rdio}(\mathbf{a}) to be the supremum of the real numbers \rho for which \mathbf{a} satisfies Condition (*)_{\rho} . The mismatches between these two words may be viewed as a form of noise that partially obscures the periodicity of {\bf a} . It follows from this definition that 1\leq\mathbf{Dio}(\mathbf{a})\leq\mathbf{Rdio}(\mathbf{a})\leq\infty . Our first main result, Theorem A , is the analogue, for this new exponent, of the main result of [ 2 ] . It also allows for more general coefficient sequences and for the use of arbitrary absolute values. In this sense, it expresses the fact that such noise can, to a certain extent, be ignored from the Diophantine point