$\def\truesubset{\subsetneq}$I took grad school algebra many eons ago and recently decided to re-learn it just because. I have been unable to figure out the following question that is the key to a different problem I was working on (that problem described at the bottom).

Let $R$ be a commutative ring with unity. Let $I_1, I_2$ be two ideals. Are the following two equivalent?

  1. If $I_1, I_2$ are co-prime (e.g. $I_1+I_2=R$), then there is no prime ideal $P$ such that $I_1+I_2 \truesubset P \truesubset R$.

  2. If there is no prime ideal $P$ such that $I_1+I_2 \truesubset P \truesubset R$, then $I_1, I_2$ are co-prime.

In other words, being co-prime and the sum not bring contained in any prime ideal are equivalent characterizations.

1 above is obvious from the definition of co-prime. Since $I_1 + I_2=R$, it is impossible to squeeze any SET between $I_1+I_2$ and $R$.

I can see why 2 would be true in a principal ideal ring, but I cannot prove it as stated. I cannot disprove either because I don't have fresh in my mind enough examples of non-principal ideal rings.

FYI, the characterization in 2 was used in this website some years ago to show that if two ideals are co-prime, so are the sums of any powers of the two ideals (e.g. $I_1+I_2=R$ implies $\forall n \in \mathbb N^+, I^n_1+I^n_2=R$).

  • $\begingroup$ equivalence is by definition, maybe notice that R/I+J is isomorphic to the tensor product of R/I and R/J... $\endgroup$ – Morph Mar 24 '17 at 9:18
  • $\begingroup$ 2. is basically equivalent to the statement that a ring has no maximal ideals. This would turn the whole algebra upside down :) $\endgroup$ – MooS Mar 24 '17 at 9:39

Your property 2 is not true -- a counterexample would be to let $I_1$ be some maximal ideal and $I_2$ be $\{0\}$.

Your property 1 is true, but for the trivial reason that you're assuming that $I_1+I_2=R$, then the conclusion is just that there is no prime ideal $P$ such that $R\subsetneq P\subsetneq R$. But of course there is nothing (prime or not, ideal or not) that sits strictly between $R$ and itself.

  • $\begingroup$ Henning, are you referring to 2 in rschwieb or 2 in my problem? It seems you are giving a counter example to 2 in his, not mine. I still do not have a counterexample to #2 above. I know the example cannot come from a principal ideal ring like the integers. $\endgroup$ – Pablo A Perez-Fernandez Mar 24 '17 at 14:54
  • 1
    $\begingroup$ @PabloAPerez-Fernandez: It's a counterexample to your number 2. With my counterexample $I_1+I_2$ is a maximal ideal (namely, it equals $I_1$), so there is no ideal (prime or not) that is strictly between $I_1+I_2$ and $R$. But $I_1$ and $I_2$ are not coprime. $\endgroup$ – Henning Makholm Mar 24 '17 at 15:17
  • $\begingroup$ @HenningMakholm---I understand your counter example for #2 in my question. It even works for the integers. Let $I_1 = 2 \mathbb Z$. Then, $I_1$ is maximal. Letting $I_2 = \{0\}$, we see that there is no proper prime ideal properly containing $I_1+I_2$. Yet, $I_1+I_2 \neq \mathbb Z$. Thanks much for the help. $\endgroup$ – Pablo A Perez-Fernandez Mar 24 '17 at 16:58

The whole problem stems from the $\subsetneq P$ in your proposed property $2$, as subsequent counterexamples in comments and solutions demonstrated.

But relaxing that to $\subseteq$, the following are obviously equivalent for any $I\lhd R$ (granted Zorn's lemma etc etc):

  1. There is no prime ideal $P$ such that $I\subseteq P$.
  2. There is no maximal ideal $M$ such that $I\subseteq M$.
  3. $I=R$

Applying this to $I=I_1+I_2$, you would indeed be able to conclude that $I_1+I_2=R$.


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.