Images of polynomial maps and the Ax-Grothendieck theorem over algebraically closed division rings

Elad Paran 1,†, Tran Nam Son 2,3,‡,* · 2025

Read Images of polynomial maps and the Ax-Grothendieck theorem over algebraically closed division rings by Elad Paran 1,†, Tran Nam Son 2,3,‡,* (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-2505-06667.txt.

Introduction

Elad Paran and Tran Nam Son extend a beloved theorem of algebra into the noncommutative world, in this 2025 paper. The classical Ax-Grothendieck theorem says that a polynomial map from a space to itself that never sends two points to the same place must actually reach every point: one-to-one forces onto. It fails in general when multiplication does not commute, as numbers in a division ring can misbehave.

Working over division rings that are finite over their centers, the authors prove the theorem again for maps built from polynomials in non-commuting variables. They also settle simpler image questions with no size condition at all. A Waring-type result follows: every square matrix over such a ring breaks into differences of products of a certain special form. These are landmarks in the algebra of skewed number systems, cousins of the quaternions.

Catalog record

CREATORElad Paran 1,†, Tran Nam Son 2,3,‡,*
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE48 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-06667

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Images of polynomial maps and the Ax-Grothendieck theorem over algebraically closed division rings Elad Paran 1,†, Tran Nam Son 2,3,‡,* Abstract . We study the images of polynomial maps over algebraically closed division rings. Our first result generalizes the classical Ax-Grothendieck theorem: We show that if f 1 , … , f m f_{1},\ldots,f_{m} are elements of the free associative algebra D ⁡ ⟨ X 1 , … , X m ⟩ D\langle X_{1},\ldots,X_{m}\rangle generated by m ≥ 1 m\geq 1 variables over an algebraically closed division ring D D of finite dimension over its center F F , and if the induced map f = ( f 1 , … , f m ) : D m → D m f=(f_{1},\ldots,f_{m})\colon D^{m}\to D^{m} is injective, then f f must be surjective. With no condition on the dimension over the center, our second result is that p ⁡ ( D ) = D p(D)=D if p p is either an element in F ⁡ ⟨ X 1 , … , X m ⟩ F\langle X_{1},\ldots,X_{m}\rangle with zero constant term such that p ⁡ ( F ) ≠ { 0 } p(F)\neq\{0\} , or a nonconstant polynomial in F ⁡ [ x ] F[x] . Furthermore, we also establish some Waring type results. For instance, for any integer n > 1 n>1 , we prove that every matrix in M n ​ ( D ) \mathrm{M}_{n}(D) can be expressed as a difference of pairs of multiplicative commutators of elements from p ​ ( M n ​ ( D ) ) p(\mathrm{M}_{n}(D)) , provided again that D D is finite-dimensional over F F . 2020 Mathematics Subject Classification 12E15; 14R99; 15A23; 16K40; 16S50. *Corresponding author 1 Department of Mathematics and Computer Science, The Open University of Israel 2 Faculty of Mathematics and Computer Science, University of Science, Ho Chi Minh City, Vietnam 3 Vietnam National University, Ho Chi Minh City, Vietnam † [email protected], ‡ [email protected] 1 Introduction The classical Ax-Grothendieck theorem, proven in the 1960s independently by Ax [ 7 ] and Grothendieck [ 27 ] , states that any injective polynomial map from the m -dimensional complex space into itself must be surjective. This result holds more generally over any algebraically closed field. A striking noncommutative counterpart was established by Augat in [ 5 ] , yielding a free version of the theorem for polynomial maps in freely non-commuting variables. The proof involves a detailed study of the noncommutative Jacobian matrix. Further progress was made in [ 9 ] , where the authors considered polynomial evaluations on matrix algebras whose entries lie in an algebraically closed division ring in the sense of Makar-Limanov [ 35 ] . One of the interesting results of [ 9 ] states that for any nonconstant multilinear polynomial, the evaluation map on matrix algebra is surjective. Since injectivity of an evaluation map already implies that the corresponding polynomial must be nonconstant, these surjectivity results also apply in that context. Overall, these results suggest a common theme: Algebraic structures that share properties with algebraically closed fields, such as algebraically closed division rings, often admit similar surjectivity phenomena for polynomial evaluation maps. The notion of an algebraically closed field can be generalized to division rings in various ways. A natural approach is introduced in [ 29 , Chapter 5, \lx@sectionsign 16, Page 255] : A division ring is said to be algebraically closed if every nonconstant polynomial with one-sided coefficients in a single variable over the ring has a right root within the ring. In this work, we adopt this definition; It is well known (see [ 38 , Theorem 1] or [ 29 , (16.14) Theorem (Niven, Jacobson)] ) that the real quaternions algebra \mathbb{H} is the first example of a noncommutative algebraically closed division ring. For further details on the characterization of algebraically closed division rings, we refer the reader to [ 29 , pp. 255–256] . Other generalizations have been introduced by Makar-Limanov [ 35 ] , Dauns [ 20 ] , and Smith [ 42 ] . In this work, we establish a noncommutative analogue of the Ax-Grothendieck theorem: For an arbitrary division ring D , let D\langle X_{1},\ldots,X_{m}\rangle denote the free associative algebra generated by m variables over D . Every element of this algebra induces a polynomial map from D^{n} to D . We prove the following generalization of the Ax-Grothendieck theorem (Theorem 2.3 ): If an algebraically closed division ring D is of finite dimension over its center, and for f_{1},\ldots,f_{m}\in{D}\langle X_{1},\ldots,X_{m}\rangle , if the map \displaystyle f \displaystyle= \displaystyle(f_{1},\ldots,f_{m})\colon D^{m}\to D^{m}, \displaystyle(a_{1},\ldots,a_{m})\mapsto\left(f_{1}(a_{1},\ldots,a_{m}),\ldots,f_{m}(a_{1},\ldots,a_{m})\right) is injective, then it is surjective. In order to prove this result, we first note that by a theorem of Baer (see [ 29 , (16.15) Theorem (Baer)] ), if D is not a field then it is a quaternion algebra over a real-closed field R . We then interpret f as a polynomial map from R^{4m} to R^{4m} using a theorem of Wilczynski, and then apply a version of the Ax-Grothendieck theorem for real-closed fields, due to Bialynicki-Birula and Rosenlicht. Furthermore, we take a moment to explore polynomials in central variables, with particular attention to the single-variable case, while the behavior in several variables remains an open question. As a direct consequence of the definition of algebraically closed division rings, we establish a straightforward analogue in this setting (see Proposition 2.1 ). This naturally raises the question of how large the image of a polynomial can be when D is not assumed to be algebraically closed. To address this, we prove in Theorem 2.2 that the image of a non-constant polynomial over an infinite division ring D must itself be infinite, provided that D contains infinitely many conjugacy classes. Additionally, we discuss the existence of such division rings and noncommutative division rings with only finitely many conjugacy classes. Our second main result is Theorem 3.1 , which states that if D is an algebraically closed division ring with center F , and p\in F\langle X_{1},\ldots,X_{m}\rangle is a polynomial with zero constant term satisfying p(F)\neq\{0\} , then p(D)=D . Building on this result, we show that the image of p contains all diagonalizable matrices. Furthermore, every matrix over D can be expressed as a product of two diagonalizable matrices. As a direct consequence, for any n>1 , every matrix in \mathrm{M}_{n}(D) can be written as a product of two elements from p(\mathrm{M}_{n}(D)) (see Theorems 4.2 and 4.3 ). In addition, when D is finite-dimensional over F , we establish that every matrix in \mathrm{SL}_{n}(D) can be expressed as a product of two multiplicative commutators from p(\mathrm{M}_{n}(D)) (see Theorem 4.6 ). This leads to the further conclusion, via Theorem 4.9 , that every matrix in \mathrm{M}_{n}(D) can be written as a difference of pairs of multiplicative commutators of elements in p(\mathrm{M}_{n}(D)) , provided that the characteristic of D is different from 2 . Moreover, when D has characteristic 0 , we show that any matrix A\in\mathrm{M}_{n}(D) can be represented as a product of two additive commutators of elements in p(\mathrm{M}_{n}(D)) , except in the case where A is invertible and n is odd. In this setting, we also establish that any nilpotent matrix A\in\mathrm{M}_{n}(D) can be written as an

Read the complete text