Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I am working on a specific problem and I've almost got it solved. To solve it, however, I need to prove one last claim (if it is even true):

Consider an integral domain $R$ that is not a field. Let $a$ be an element of $R$ that is not a unit or $0$. I want to prove that there cannot exist nonzero elements $r_n$, $n \in \mathbb{N}$, of $R$ such that

$$ r_1 a=r_2a^2=r_3a^3=\cdots= r_na^n=\cdots$$

First of all, is this even true? My intuition says it is. And am I missing something obvious? This is the last step in a long proof, so I hope it is true!

EDIT: It turns out my proof goes through without this being true (which is good since it's false), but definitely thanks for the answers. It has improved my understanding.

share|improve this question
Well, if you have something like that then you can manufacture a strictly increasing sequence of ideals, so if you know you have a noetherian ring you would have a contradiction. –  Zhen Lin Dec 6 '12 at 18:49
I have a feeling this might fail in a ring which contains an element which cannot be factored into irreducible elements. –  Tobias Kildetoft Dec 6 '12 at 18:53
So this does hold in, a UFD for instance. –  JSchlather Dec 6 '12 at 18:58
The "not a field" requirement seems artificially added to get rid of the obvious finite-field examples like $2 \in \mathbb{Z}_3$. But you could easily extend to, say, $2 \in \mathbb{Z}_3[x]$, and then it's no longer a field but still a counterexample. –  MartianInvader Dec 6 '12 at 19:41
@MartianInvader But 2 is a unit in that ring. –  Three Dec 6 '12 at 19:59

4 Answers 4

up vote 15 down vote accepted

This is false. As Zhen Lin points out, $R$ can't be Noetherian, so let $R$ be my favourite non-Noetherian domain, $\{f(x)\in \Bbb{Q}[x]\mid f(0)\in \Bbb{Z}\}$, and consider $a=2$. Then $$x=2\frac{x}{2}=4\frac{x}{4}=8\frac{x}{8}=\ldots$$ but $2$ is not a unit.

share|improve this answer
Aha! I was trying to extract this from non-standard integers example, but I hadn't realized the ring had such a simple description! –  Hurkyl Dec 6 '12 at 18:58
@Hurkyl, are you saying that the ring provided by Chris Eagle is isomorphic to nonstandard integers? –  PrimeRibeyeDeal Jun 6 '13 at 21:24
@PrimeRibeyeDeal: No, but it is isomorphic to a subring of the nonstandard integers (send $x \to 2^H$). ("subring" is given its standard meaning) –  Hurkyl Jun 7 '13 at 8:00
@Hurkyl, what is $2^H$ in this context? Is there a "nice" description of the nonstandard integers as a ring? –  PrimeRibeyeDeal Jun 9 '13 at 17:23
@PrimeRibeyeDeal: My comments were referring to my answer in this thread. Re your other question, the only "nice" description of the nonstandard integers I know of is that it is a nonstandard model of the integers. –  Hurkyl Jun 9 '13 at 21:30

Let $R = {}^\star \mathbb{Z}$ be the hyper-integers; that is, the ring of integers in a non-standard model of real analysis.

Let $H$ be an infinite hyper-integer.

Then, we can choose $a = 2$ and $r_i = 2^{H - i}$ to get a counter-example.

share|improve this answer
Hum. If I construct the nonstandard reals via an ultraproduct, does this denote the subring of the nonstandard reals which are a.e. equal to an integer? –  Qiaochu Yuan Dec 6 '12 at 19:08
@Qiaochu Yuan: Yes. And in that case, we might as well make an explicit choice of $H$ and set $r_i = ( 2^{-i}, 2^{1-i}, 2^{2-i}, \cdots )$. –  Hurkyl Dec 6 '12 at 19:09

No. Let $R = \mathbb{Z}[\frac{1}{x}, 2x, 2x^2, 2x^3, ...]$, let $r_i = 2x^i$, and let $a = \frac{1}{x}$.

share|improve this answer
$a$ was specified not to be a unit. –  Greg Muller Dec 6 '12 at 18:49
@Greg: fixed. ${}$ –  Qiaochu Yuan Dec 6 '12 at 18:55

Let $R$ be the ring of functions $[0,1]\to \mathbb R$ that are analytic in $(0,1]$ and continuous on $[0,1]$. Let $a$ be the function $x\mapsto x$. It is not a unit because $a(0)=0$.

For each $n\in \mathbb N$, let $$r_n(x)=\begin{cases}x^{-n}e^{-\frac1x}&\text{if }x\ne 0\\0&\text{if }x=0\end{cases}$$ Note that $r_n$ is analytic in $(0,1]$ and is continuos also at $0$. We have $$(a^nr_n)(x)=\begin{cases}e^{-\frac1x}&\text{if }x\ne 0\\0&\text{if }x=0\end{cases}$$

share|improve this answer
That's not a domain. –  anonymous Dec 6 '12 at 18:57
@anonymous D'oh, you are right. I started off at polynomials and needed more and more. But continuous is of course too big a ring. I had to tighten the conditions to (almost) analytic, but now it's an ugly example. –  Hagen von Eitzen Dec 6 '12 at 19:04

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.