3
$\begingroup$

Recall that in a commutative ring $A$ an ordered pair of elements (a,b) is said to form a regular sequence if the ideal $\langle a,b\rangle $ is strictly included in $A$ ,if $a$ is not a zero-divisor in $A$ and if the class of $b$ is not a zero-divisor of $A/\langle a\rangle$.
A friend of mine has asked me if in that case we can conclude that $\langle b,a\rangle $ is also a regular sequence under the assumption that $A$ is a noetherian domain.
The answer is known to be yes for a local noetherian ring $A$, even it is not a domain

[Since I couldn't answer his question, I suggested to my friend that he ask here but he prefers that I do that]

Improve this question
$\endgroup$

1 Answer 1

Reset to default
8
$\begingroup$

The only part to be shown is that $a$ is not a zero-divisor on $A/(b)$. Consider some $s\in A$ such that $as\in(b)$, say $as=bt$. Since $b$ is not a zero-divisor on $A/(a)$, we conclude that $t$ maps to zero in $A/(a)$, i.e., $t=au$. Since $A$ is a domain, it follows that $s=bu$, so $s$ maps to zero in $A/(b)$.

Improve this answer
$\endgroup$
8
  • $\begingroup$ Strange, my friend ( a great algebraist who definitely knows what a quotient ring is!) told me he knew a counter-example due to Dieudonné in the case $A$ was a non-noetherian domain. He must have misremembered. Anyway your answer is obviously correct and optimal: thanks a lot. $\endgroup$
    – Georges Elencwajg
    Commented Oct 27, 2011 at 16:24
  • 1
    $\begingroup$ According to Matsumura, there is an example by Dieudonne' in nagoya math j. 1966, pp. 355-356, where A is non noetherian and local but not a domain. $\endgroup$
    – roy smith
    Commented Oct 27, 2011 at 16:26
  • $\begingroup$ Why is this the only part to be shown? We also need to show that $b$ is not a zero-divisor on $A$. That, it seems to me, is the hard part. $\endgroup$
    – Steven Landsburg
    Commented Oct 27, 2011 at 16:28
  • 3
    $\begingroup$ Steven, that bothered me too, but it seems A is a domain by hypothesis, so there is no zero divisor. $\endgroup$
    – roy smith
    Commented Oct 27, 2011 at 16:30
  • 3
    $\begingroup$ Update: I just phoned my friend to tell him about a-fortiori's solution. He is very happy and told me he would stop looking for counterexamples... $\endgroup$
    – Georges Elencwajg
    Commented Oct 27, 2011 at 16:33

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .