A primary ideal (in a commutative ring with unity) is an ideal $J$ for which if $ab\in J$, then either $a\in J$ or $b^n\in J$ for some integer $n\geq 1$. So it also implies (due to commutativity) that if $ab\in J$, then $a^m,b^n\in J$ for some integers $m,n\geq 1$.

Wouldn't the latter be a nicer definition for a primary ideal? Why is it not used?


Your "it implies" is wrong, it is possible that $ab \in J$ and $a \in J$ but $b \notin J$.

For example any prime ideal is primary, but it doesn't necesary satisfy your condition....

Added: P.S. I think the equivalent condition you seek is the following: whenever $ab \in J$ we either ( have one of $a$ or $b$ in J ) OR ( $a^m,b^n\in J $ for some integers $m,n$ .)

But then, the standard definition is much cleaner.


Your criterion is not correct as stated: if $a \in J$, it need not be the case that some power of $b$ is in $J$.

Another definition is that zerodivisors in $R/J$ are nilpotent. If $R$ is Noetherian, then this is equivalent to saying that $J$ has exactly one associated prime. This is what is really important, and actually used when primary hypotheses are needed.

  • $\begingroup$ @rschwieb: You're right, this is weaker. I've just decided to remove it altogether $\endgroup$ – zcn Jan 15 '14 at 4:04
  • $\begingroup$ I went ahead and did the same with the comment :) I remember following the same line of thought a long time ago! $\endgroup$ – rschwieb Jan 15 '14 at 4:06

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.