Jacobson's theorem states that If $R$ is a ring, and for every $x\in R$, there exists $n(x)\geq 2$ such that $x^{n(x)}=x$. Then $R$ is commutative.

I wonder if the following stronger assertion(in case $R$ has unity) is true.

Let $R$ be a ring with unity. For every $x$ in $R$, there exists $n(x)\geq 2$ such that $x^{n(x)} = x$. Then, $R$ is embedded in a product(possibly infinite) of fields $F_i$, where each $F_i$ is an algebraic extension of $F_{p_i}$ (prime field of $p_i$ elements).

If this is not true, then I am also interested in counterexample.

Thanks in advance.

  • 2
    $\begingroup$ Also note that $R$ is a comm. von Neumann regular ring (i.e. for each $r$ there is $s$ s.t. $r = r^2s$) and such rings are known to embedd into a product of fields. $\endgroup$ – Ralph Apr 5 '12 at 3:28

Yes, this is true.

By Jacobson's theorem, $R$ is commutative. Now the radical of $R$ is the intersection of all primes $P$ of $R$. Hence we have an embedding $$\phi: R/rad(R) \to \prod_P R/P.$$ For each $x \in R/P, x \neq 0$ there is $n \ge 1$ such that $x(x^n-1)=0$ and since $R/P$ is a domain, $x^n =1$, i.e. $x$ is a unit. Thus $R/P$ is a field. Since elements $\neq \pm 1$ of $\mathbb Q$ aren't roots of unity, $R/P$ has prime characteristic and each element of $R/P$ is algebraic. Therefore $R/P$ is an algebraic extension of some $\mathbb{F}_p$.

To finish the proof, we have to show $rad(R) = 0$. Let $x \in rad(R)$. Let $k>0$ be minimal with $x^k=0$ and let $n \ge 2$ with $x^n=x$. Suppose $k > n$. Write $k=qn+r$ with $q > 0$, $0 \le r < n$. Then $0 = x^k = (x^n)^q x^r = x^{q+r}$. Minimality of $k$ implies $k = q+r$, i.e. $n=1$, in contradiction to the assertion $n \ge 2$. Hence $k\le n$ and $x = x^n = x^k x^{n-k} = 0$.

  • 1
    $\begingroup$ Thanks, I got the idea. How about starting from the Jacobson radical of $R$? I think it is easier. Also, showing the Jacobson radical is zero, is easy. If $x\in J(R)$, then $x^n-1$ is unit for any $n\geq 1$. So, $x=0$. The Jacobson radical is intersection of maximal ideals, so we get $R/M$ is field for each maximal ideals $M$. $\endgroup$ – Sungjin Kim Apr 5 '12 at 2:20
  • $\begingroup$ Looks good. In particular, you get $R/M$ a field for free. In effect, since $R/P$ is a field, each $P$ is maximal and $J(R) = rad(R)$. But using $J(R)$ in the proof is certainly better. $\endgroup$ – Ralph Apr 5 '12 at 2:29

This is true. Let $R$ be a ring satisfying your property. Then $R$ has no nilpotent element since if $x^a=0$, then for $b$ an integer such that $n(x)^b > a$ we have $x=x^{n(x)^b}=0$. So the radical of $R$ is $0$, and since the radical is the intersection of all prime ideals of $R$, we see that the natural map $R \mapsto \prod_{P} R/P$ is an injection. It therefore suffices to prove the result for $R/P$, that is for a domain.

Assuming no that $R$ is a domain, the equation $x^{n(x)}-x=0$ factors as $x=0$ or $x^{n(x)-1}=0$. So, in $R$, every non-zero element is a root of unity. This is also true of the fraction field $K$ of $R$. No $K$ is of characteristic $p>0$, since otherwise it would contain $\mathbb{Q}$ which contains many non-roots-of-unity such as $2$. And clearly, $K$ is algebraic over the prime field. So we're done.

  • $\begingroup$ It is not necessary to consider a fraction field $K$ of $R$ because you already know that "each non-zero element [of $R$] is a root of unity". Hence $R$ ist alreday a field. $\endgroup$ – Ralph Apr 5 '12 at 2:13
  • $\begingroup$ Ralph, you beat me by 6 seconds ! Mathoverflow is really becoming too competitive :-) $\endgroup$ – Joël Apr 5 '12 at 2:23
  • $\begingroup$ Thank you for you too Joel. Both answers are basically the same. So, I accepted the first one. $\endgroup$ – Sungjin Kim Apr 5 '12 at 2:26
  • $\begingroup$ Yes, your answer showed up directly after mine. Even more amazing is that we not just had the same idea of proof but used nearly the same words: "this is true", "is the intersection of all primes". In any case I wished it were possible in MO to accept more than one answer. $\endgroup$ – Ralph Apr 5 '12 at 2:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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