Growth of generalized greatest common divisors along orbits of self-rational maps on projective varieties
Yohsuke Matsuzawa · 2025
Read Growth of generalized greatest common divisors along orbits of self-rational maps on projective varieties by Yohsuke Matsuzawa (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-2507-05027.txt.
Introduction
Yohsuke Matsuzawa proves a growth result in arithmetic geometry in this 2025 paper. The work concerns repeated application of a self-map on a smooth geometric space. It tracks two height measures, which are tools that gauge the size of points in such studies.
Matsuzawa shows that one height, tied to a fixed sub-object, grows far more slowly than the other along a typical orbit. In the limit, their ratio falls to zero. The proof needs set conditions: the orbit must be generic, and a growth rate for the point must beat a threshold drawn from the map. The result holds either when the map is defined everywhere or when the sub-object sits inside a safe region. It adds to the theory of how orbits spread in these spaces.
Catalog record
| CREATOR | Yohsuke Matsuzawa |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 69 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-05027
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Growth of generalized greatest common divisors along orbits of self-rational maps on projective varieties Yohsuke Matsuzawa Abstract . Consider a dominant rational self-map f f on a smooth projective variety X X defined over ℚ ¯ \overline{\mathbb{Q}} . We prove that (0.2) lim n → ∞ h Y ( f n ( x ) ) h H ( f n ( x ) ) = 0 , \displaystyle\lim_{n\to\infty}\frac{h_{Y}(f^{n}(x))}{h_{H}(f^{n}(x))}=0, where h Y h_{Y} is a height associated with a closed subscheme Y ⊂ X Y\subset X of codimension c c , h H h_{H} is any ample height on X X , and x ∈ X ( ℚ ¯ ) x\in X(\overline{\mathbb{Q}}) is a point with well-defined orbit, under the following assumptions: (1) either f f is a morphism, or Y Y is pure dimensional, regularly embedded in X X , and contained in the locus over which all iterates of f f are finite; (2) the orbit of x x is generic; (3) d c ( f ) 1 / c < α f ( x ) d_{c}(f)^{1/c}<\alpha_{f}(x) , where d c ( f ) d_{c}(f) is the c c -th dynamical degree of f f and α f ( x ) \alpha_{f}(x) is the arithmetic degree of x x . 2020 Mathematics Subject Classification Primary 37P15; Secondary 37P55 1. Introduction Let X be a smooth projective variety, and f\colon X\dashrightarrow X a dominant rational map, both defined over \overline{\mathbb{Q}} . For a proper closed subscheme Y\subset X , we can define associated global height function h_{Y} ( [ 28 ] ), which is nothing but a Weil height function when Y is a Cartier divisor. When Y has codimension at least two, the case we are interested in, h_{Y} is sometimes called generalized greatest common divisor. This name may be justified by the following example. Let Y=\{(0:0:1)\}\subset{\mathbb{P}}^{2}_{\overline{\mathbb{Q}}} and x=(a:b:c)\in{\mathbb{P}}^{2}(\overline{\mathbb{Q}}) with a,b,c\in{\mathbb{Z}} coprime. Then (1.1) \displaystyle h_{Y}(x)=\log\frac{\max\{|a|,|b|,|c|\}}{\max\{|a|,|b|\}}+\log\gcd(a,b). For a point x\in X(\overline{\mathbb{Q}}) with well-defined forward f -orbit, h_{Y}(f^{n}(x)) is expected to be very small. More precisely, given any Weil height function h_{H} associated with an ample divisor H on X , we expect (1.2) \displaystyle\lim_{n\to\infty}\frac{h_{Y}(f^{n}(x))}{h_{H}(f^{n}(x))}=0 unless there are any special (geometric) reasons (cf. [ 30 , section 19] ). Roughly speaking, this claims that the number of digits of greatest common divisors among coordinates of orbits are smaller than the number of digits of coordinates. There are works [ 22 , 12 ] that directly study this problem, but they all assume Vojta’s conjecture. We remark that Silverman pointed out [ 30 , Example 35] that a famous theorem by Bugeaud-Corvaja-Zannier [ 5 , Theorem 1] can be interpreted dynamically. For example, let a,b\geq 2 be multiplicatively independent integers, and f\colon{\mathbb{P}}^{2}_{\overline{\mathbb{Q}}}\longrightarrow{\mathbb{P}}^{2}_{\overline{\mathbb{Q}}},(x:y:z)\mapsto(ax:by:z) . Let Y=\{(1:1:1)\} . Then (1.3) \displaystyle\frac{h_{Y}(f^{n}(1:1:1))}{h_{H}(f^{n}(1:1:1))} (1.4) \displaystyle=\frac{1}{\log\max\{|a^{n}|,|b^{n}|,1\}}\bigg(\log\frac{\max\{|a^{n}|,|b^{n}|,1\}}{\max\{|a^{n}-1|,|b^{n}-1|\}}+\log\gcd(a^{n}-1,b^{n}-1)\bigg) and [ 5 , Theorem 1] is equivalent to say that this goes to 0 as n\to\infty . The key input to the proof of their theorem was Schmidt subspace theorem. In this paper, we prove a sufficient condition for eq. 1.2 . Our theorem is unconditional, i.e. does not assume Vojta’s conjecture. Our strategy is to take advantage of dynamical complexity, and so only applicable to maps f with non-trivial first dynamical degree. Thus we cannot deduce Bugeaud-Corvaja-Zannier’s theorem from our one; the map f(x:y:z)=(ax:by:z) has trivial first dynamical degree 1 . To state our theorem, let us introduce some notions. Definition 1.1 (Orbits under rational maps) . Let f\colon X\dashrightarrow X be a dominant rational map on a projective variety (variety is irreducible and reduced in this paper) defined over an algebraically closed field k . The indeterminacy locus of f is denoted by I_{f} . For a point x\in X(k) , we say the forward f -orbit is well-defined if f^{n}(x)\not\in I_{f} for all n\geq 0 . The set of x\in X(k) with well-defined forward f -orbit is denoted by X_{f}(k) . For x\in X_{f}(k) , O_{f}(x)=\{x,f(x),f^{2}(x),\dots\} is called the orbit. We say the orbit is generic if O_{f}(x) is infinite and O_{f}(x)\cap Z is finite for any proper Zariski closed subset Z\subset X . We recall the definition of dynamical degrees. There are number of papers that study properties of dynamical degrees. We refer [ 31 , 6 ] for basic properties of dynamical degrees. (Note also that the recursive inequalities proven in [ 33 ] imply important properties of dynamical degrees, such as convergence, log concavity, and product formulas.) Definition 1.2 (Dynamical degrees) . Let f\colon X\dashrightarrow X be a dominant rational map on a projective variety defined over an algebraically closed field k . Let H be an ample divisor on X . For i\in\{0,\dots,\dim X\} , the i -th dynamical degree of f is (1.5) \displaystyle d_{i}(f)=\lim_{n\to\infty}(\pi_{2,n}^{*}H^{i}\cdot\pi_{1,n}^{*}H^{\dim X-i})^{1/n} where \pi_{1,n},\pi_{2,n} are the projections from the graph of f^{n} : (1.6) {\lx@inpgf@ignorespaces\Gamma_{f^{n}}} {\lx@inpgf@ignorespaces X} {\lx@inpgf@ignorespaces X.} \scriptstyle{\lx@inpgf@ignorespaces\pi_{1,n}} \scriptstyle{\lx@inpgf@ignorespaces\pi_{2,n}} \scriptstyle{\lx@inpgf@ignorespaces f^{n}} Note that d_{\dim X}(f) is simply the degree of the function field extension induced by f . In [ 24 , Appendix B.1] , for a dominant rational map f\colon X\dashrightarrow X on a projective variety, we introduced a subset of X over which all iterates of f are “finite”. We include here the definition, but refer [ 24 , Appendix B.1] for basic properties of this set. Definition 1.3 (Iteratively finite locus) . Let f\colon X\dashrightarrow X be a dominant rational map on a projective variety defined over a field k . (1) Let \Gamma_{f} be the graph of f and p_{i}\colon\Gamma_{f}\longrightarrow X for i=1,2 be the projections: (1.7) {\lx@inpgf@ignorespaces\Gamma_{f}} {\lx@inpgf@ignorespaces X} {\lx@inpgf@ignorespaces X.} \scriptstyle{\lx@inpgf@ignorespaces p_{1}} \scriptstyle{\lx@inpgf@ignorespaces p_{2}} \scriptstyle{\lx@inpgf@ignorespaces f} We write X_{f}^{\rm fin} the largest open subset of X such that (1.8) p_{2}\colon p_{2}^{-1}(X_{f}^{\rm fin})\longrightarrow X_{f}^{\rm fin} is finite and p_{2}^{-1}(X_{f}^{\rm fin})\cap p_{1}^{-1}(I_{f})=\emptyset . Note that f|_{X\setminus I_{f}}^{-1}(X_{f}^{\rm fin}) is isomorphic to p_{2}^{-1}(X_{f}^{\rm fin}) via p_{1} . (2) We inductively define X_{f}^{{\rm fin},n} for n\geq 0 as follows. We set X_{f}^{{\rm fin},0}=X . For n\geq 1 , we set (1.9) \displaystyle X_{f}^{{\rm fin},n}=X_{f}^{{\rm fin},n-1}\setminus f^{n-1}(f^{-(n-1)}(X_{f}^{{\rm fin},n-1})\setminus X_{f}^{{\rm fin}}). Here f^{-(n-1)}(X_{f}^{{\rm fin},n-1}) is f|_{X\setminus I_{f^{n-1}}}^{-(n-1)}(X_{f}^{{\rm fin},n-1}) by definition. By the construction, this is finite over X_{f}^{{\rm fin},n-1} . Note also that X_{f}^{{\rm fin},1}=X_{f}^{{\rm fin}} . (3) We define (1.10) \displaystyle X_{f}^{\rm back}=\bigcap_{n\geq 0}X_{f}^{{\rm fin},n}. When the dynamical system is defined over \overline{\mathbb{Q}} , the growth rate of Weil height function along the orbit of a given point is called arithmetic degree. Definition 1.4 (Arithmetic degree) . Let f\colon X\dashrightarrow X be a dominant rational map on a projective variety defined over \overline{\mathbb{Q}} . Let H be an ample divisor on X and h_{H} be an associated Weil height function on X(\overline{\mathbb{Q}}) . For a point x\in X_{f}(\overline{\mathbb{Q}}) , (1.11) \displaystyle\alpha_{f}(x)=\lim_{n\to\infty}\max\{1,h_{H}(f^{n}(x))\}^{1/n} is called the arithmetic degree of x provided the limit exists. This is independent of the choice of H and h_{H} (even the limsup and liminf are independent). If f is a morphism [ 13 ] or O_{f}(x) is generic [ 24 ] , then it is known that the limit exists. For the basic properties of arithmetic degree, see [ 23 ] and references therein. Now we state our main theorem. Theorem 1.5 (Main theorem) . Let X be a smooth projective variety