This is a follow-up to Is tensor product of local algebras local?.
Let $A, B$ and $C$ be local rings (commutative and noetherian). Suppose that we have local ring maps $C \to A$ and $C \to B$.
What assumptions guarantee that the tensor product $A \otimes_C B$ of $A$ and $B$ over $C$ is a local ring?

For example, $A$ is the completion of $C$? What if $A, B$ and $C$ are dvrs? Are there general criteria? (If I remember correctly, I have seen one criterion, when $C$ was a field. What about general $C$?)

  • 1
    $\begingroup$ Are you sure this holds when C is a field? $\endgroup$ – J.C. Ottem Feb 2 '11 at 18:28
  • $\begingroup$ Not, of course (many trivial counter-examples). I said that there is known (if I remember correctly) criterion for when such tensor is local, if we assume that C is a field. $\endgroup$ – unknown Feb 2 '11 at 18:37
  • 3
    $\begingroup$ Have you seen jstor.org/stable/2040682?seq=1? $\endgroup$ – J.C. Ottem Feb 2 '11 at 19:01
  • 1
    $\begingroup$ Yes. But it considers the case, when C is a field. $\endgroup$ – unknown Feb 2 '11 at 19:03

Here is a partial answer related to connected rings (which is more general than local rings) in the case where rings are not necessarily commutative.

If $K$ is a field, let $A$ be an artinian and connected $K$-algebra (in this situation it's equivalent to being local and artinian). Assume that $A$ is endowed with a morphism of algebras $A\longrightarrow K$. Then for any connected $K$-algebra $B$, $A\otimes_K B$ is connected.

An trivial counterexample : if $K$ is $\mathbb{F}_2$ and $L=\mathbb{F}_4$, $L\otimes_K L$ is not connected (sorry to write it this way but there seems to be a bug with latex).

edit : this results is also true if $K$ is only supposed to be a domain and with some assumption on $B$ (flatness).


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.