Strongly clean ring elements that are one-sided inverses

George M. Bergman · 2025

Read Strongly clean ring elements that are one-sided inverses by George M. Bergman (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-2508-14396.txt.

Introduction

George Bergman probes a longstanding algebra question in twenty twenty-five. A ring is strongly clean if every element splits into an idempotent and a commuting unit. A ring is Dedekind-finite if one-sided inverses are always two-sided. Whether every strongly clean ring is Dedekind-finite has been open for years.

Bergman gives a ring containing two strongly clean elements. Those elements are one-sided inverses of each other but not two-sided ones. The example suggests the open question's answer may be negative. He then discusses routes to strengthen the example into a full counterexample. Brief observations on uniquely strongly clean rings close the paper. Readers in ring theory get a concrete object pushing against a clean conjecture.

Catalog record

CREATORGeorge M. Bergman
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE17 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-14396

Browse the full library · About the Tower of Babel

Full text

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

Strongly clean ring elements that are one-sided inverses Thanks: Readable at http://math.berkeley.edu/~gbergman/papers/ . After publication, any updates, errata, related references, etc., found will be noted at http://math.berkeley.edu/~gbergman/papers/abstracts . George M. Bergman Abstract . A longstanding open question is whether every strongly clean ring is Dedekind-finite (definitions recalled below). We give an example of a ring with two strongly clean elements that are one-sided, but not two-sided inverses of one another, suggesting that the answer to that question may be negative. We discuss possible ways of strengthening this result to give a full negative answer. We end with some brief observations on related topics, in particular, uniquely strongly clean rings. Comments, corrections, and related references welcomed, as always! T e X ed August 10, 2026 2020 Mathematics Subject Classification Primary: 16U40, 16U60. Secondary: 16S50, 16U90. 1. The example Here is some standard usage that we will follow: Definition 1.1 . In this note, all rings are understood to be associative and unital. An element x of a ring R is called clean if it can be written as the sum e+u of an idempotent e and a unit u of R, and strongly clean if it can be so written with e and u commuting with each other, equivalently, commuting with x. A ring R is called clean, respectively strongly clean, if all of its elements have the property named ( [ 5 ] , [ 7 ] ) . A ring R is called Dedekind-finite if every one-sided-invertible element of R is invertible. The question of whether every strongly clean ring is Dedekind-finite was posed in 1999 by W. K. Nicholson [ 5 , Question 2, p. 3590] . For k a field, we construct below a \!k\! -algebra R having a pair of elements that are both strongly clean, and which are one-sided but not two-sided inverses of one another. Thus, if this ring can somehow be embedded in a ring where all elements are strongly clean, the resulting example would answer Nicholson’s question in the negative. Here is some notation that will be used in describing our example: Definition 1.2 . Throughout this section k will be a field, and V will denote the underlying k-vector-space of the subalgebra of the rational function field k(t) comprising elements having denominator not divisible by t; i.e., of the localization k[t]_{(t)} of the polynomial ring k[t] at the prime ideal (t). We denote by R the \!k\! -algebra \mathrm{End}_{k}(V) of all \!k\! -linear endomaps of the vector space V, written on the left of their arguments and composed accordingly. (We could equally well take for R the subalgebra of \mathrm{End}_{k}(V) generated by the five maps that will be denoted y, x, (y-1)^{-1}, e, and (x-e)^{-1} below. But it will be convenient to have a ring that we can refer to when defining these elements.) Let y be the endomorphism of V given by multiplication by t in k[t]_{(t)} : (1.1) y(f(t))\ =\ t\,f(t) for f(t)\in V, and x the endomorphism of V given by multiplying by t^{-1} and, in power series notation, dropping the t^{-1} term if any. In rational function notation, this is (1.2) x(f(t))\ =\ t^{-1}(f(t)-f(0)) for f(t)\in V. It is clear that (1.3) x\,y\ =\ 1. On the other hand, y\,x has the effect of dropping the t^{0} term from the power series representing an element of V, so (1.4) y\,x\ \neq\ 1. We will now prove x and y strongly clean in R. Note that y-1 is invertible in R : it takes f(t) to (t-1)f(t), and t-1 is invertible in k[t]_{(t)}. Thus y is the sum of the invertible element y-1 and the idempotent element 1, which commute, showing that (1.5) y is strongly clean in R. In proving x strongly clean, we can’t take for e one of the idempotents 1 or 0 : x-1 is not invertible since x fixes (1-t)^{-1} (easily seen by looking at power series expansions); and x-0 is not invertible since x annihilates 1. So to get a decomposition x=e+u, we will have to use a nontrivial idempotent e. To construct e, note that (1.6) V\,=\ V_{0}\oplus V_{1}, where (1.7) V_{0} is the subspace k[t]\,\subseteq\,V, and (1.8) V_{1}\subseteq V is the space of rational functions whose numerators have lower degrees than their denominators. Indeed, given an element p(t)/q(t)\in V, where p(t),\,q(t)\in k[t], its decomposition as in ( 1.6 ) arises from the decomposition of p(t) as a multiple of q(t) in k[t], plus a remainder of degree less than that of q(t). (Intuitively, V_{1} is the set of f(t)\in V such that f(\infty)=0.) We now (1.9) Let e\in R be the projection of V onto V_{0} under the decomposition ( 1.6 ). Clearly, x, defined in ( 1.2 ), carries V_{0} into itself. A little thought shows that it also carries V_{1} into itself: Although when ( 1.2 ) is applied to an element f\in V_{1}, the expression f(t)-f(0) has in general lost the property that f had, of having value 0 at \infty, it still has a finite value there, namely -f(0); so multiplying by t^{-1} again brings the value at \infty to 0. (The reader can translate all this into reasoning about degrees and values at 0 of numerators and denominators.) Hence, (1.10) x commutes with e. Now since V_{1} contains no nonzero constants, we see from ( 1.2 ) that x has trivial kernel on V_{1}. It is also surjective on V_{1} : Given f(t)\in V_{1}, if we let c\in k be the value of tf(t) at \infty, we find that tf(t)-c\in V_{1} and x carries that element to f(t). (This is like the construction of the preceding paragraph, with the roles of 0 and \infty interchanged.) Thus x acts invertibly on V_{1}. Since e annihilates V_{1}, this tells us that (1.11) x-e acts invertibly on V_{1}. How does x-e behave on V_{0} ? For every n\geq 0, we see from ( 1.2 ) that the action of x on polynomials in t of degree \leq n is nilpotent, while by ( 1.9 ), e acts as the identity endomorphism on V_{0}; so (1.12) x-e acts invertibly on V_{0}. Together, ( 1.11 ) and ( 1.12 ) say that x-e acts invertibly on V; so x is the sum of the idempotent e and the invertible element x-e, which commute by ( 1.10 ). So (1.13) x is strongly clean. We have thus proved Proposition 1.3 . In the k-algebra R of all \!k\! -vector-space endomorphisms of the vector space V defined in Definition 1.2 , the elements x and y described by ( 1.1 ) and ( 1.2 ) are strongly clean, and satisfy x\,y=1 but y\,x\neq~1. ∎ 2. Thoughts on possible stronger results I wonder whether one can push the approach of the above construction further, and get an example actually answering Nicholson’s question: Question 2.1 . As in Definition 1.2 , let k be a field and V the underlying \!k\! -vector-space of k[t]_{(t)}; and as in ( 1.6 )-( 1.8 ), let V=V_{0}\oplus V_{1}, where V_{0}=k[t] and V_{1} is the space of elements of k[t]_{(t)} whose numerators have lower degrees than their denominators. Further, (2.1) Let R be the algebra of \!k\! -linear

Read the complete text