The nilradical of a commutative ring $R$, denoted $Nil(R)$, is the set of all nilpotent elements in $R$. It can be shown that $Nil(R)$ is an ideal of $R$. Every ideal is equal to the kernel of some ring homomorphism: just consider the projection $\pi : R \to R/I$.

I'm wondering: is there a ring homomorphism whose kernel is $Nil(R)$ that is more explicit than the projection map $\pi : R \to R/Nil(R)$?

  • 2
    $\begingroup$ What do you mean by “more explicit”? What you have given seems like the most obvious solution, so it doesn’t seem to mean “obvious” $\endgroup$ – rschwieb Jul 2 at 2:36
  • 1
    $\begingroup$ I could have been more clear. I had something like the following in mind: suppose instead I was looking for a map whose kernel is $A_n$, the alternating group on $n$ elements. Of course, the projection $\pi : S_n \to S_n / A_n$ would be such a map, but I would say that the signature map $sgn : S_n \to \{1,-1\}$ (where $\{1,-1\}$ is viewed as a group under integer multiplication) is a more explicit map. $\endgroup$ – Charles Hudgins Jul 2 at 3:26

You're not going to get anywhere near something as crisp as the quotient $S_n/A_n$, because that is a highly special case: it's always order $2$ (for $n >1$) and cam be defined from the parity of the structure of the elements themselves.

In contrast, $R/Nil(R)$ is going to be quite wild by comparison. The quotient doesn't have a fixed order or any particularly special property, other than the fact it is a reduced ring.

But maybe the following is along the lines of what you're looking for. Let $\mathcal C$ be a family of prime ideals such that $\bigcap \mathcal C=Nil(R)$. Of course, it could be all prime ideals, but sometimes fewer will do.

You always have a homomorphism $R\to \prod_{P\in\mathcal C}R/P$, whose kernel is $Nil(R)$. By the first isomorphism theorem, $R/Nil(R)$ is a subring of the thing on the right, which is a product of domains. You could additionally even embed those domains in their fraction fields, and have a map of $R/Nil(R)$ into a product of fields.

Under special conditions, $R$ may permit selection of a finite $\mathcal C$ with the properties above, getting you an embedding into a finite product of fields: that's quite special.

If you know that there are finitely many maximal ideals that intersect to $Nil(R)$, then the Chinese Remainder Theorem actually says this injection is an isomorphism. In that case $R$ would be called a semilocal ring.


So anyway, now I hope you can see the diversity of structure of $R/Nil(R)$. I don't know if it is "more explicit," but it does tell you more about the structure of the quotient.


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.