# Tensor product of two finitely-generated modules over a local ring is zero

If $R$ is a local ring and $M$ and $N$ are finitely generated $R$-modules such that $M\otimes N=0$ then how does it follow from Nakayama's lemma that either $M=0$ or $N=0$?

This is an exercise in Atiyah and Macdonald. The part I could not show in the hints is $({M{\otimes}_R N)}_{k}=0$ implies $M_{k}{\otimes }_{k} N_{k}=0$, where $k=R/\mathfrak m$ and $\mathfrak m$ is the maximal ideal of $R$.

• Welcome to math.SE: since you are new, I wanted to let you know a few things about the site. In order to get the best possible answers, it is helpful if you say in what context you encountered the problem, and what your thoughts on it are; this will prevent people from telling you things you already know, and help them give their answers at the right level. Consider editing you question – Dennis Gulko Mar 5 '13 at 12:00
• @mrigankhs: please do not use (almost) all caps. It is considered to be shouting and usually deemed impolite. As to the question you asked: while it is considered a good thing for downvoters to leave constructive comments on how you can improve your question, it is by no means a requirement. Sometimes asking for an explanation for the downvote will get you an answer, but more often anyone who would give you comments would've done so already. – Willie Wong Mar 5 '13 at 16:27

From $M\otimes_RN=0$ we get $R/\mathfrak m\otimes_R (M\otimes_R N)=0$. Then it follows that $(R/\mathfrak m\otimes_R M)\otimes_{R/\mathfrak m}(R/\mathfrak m\otimes_R N)=0$. But $R/\mathfrak m\otimes_R M\simeq M/\mathfrak mM$ and similarly $R/\mathfrak m\otimes_R N\simeq N/\mathfrak mN$. Since $M/\mathfrak mM$ and $N/\mathfrak mN$ are $R/\mathfrak m$-vector spaces, it follows that $M/\mathfrak mM=0$ or $N/\mathfrak mN=0$, that is, $M=\mathfrak mM$ or $N=\mathfrak mN$, so by Nakayama we have $M=0$ or $N=0$.
Edit. For proving $R/\mathfrak m\otimes_R (M\otimes_R N)\simeq(R/\mathfrak m\otimes_R M)\otimes_{R/\mathfrak m}(R/\mathfrak m\otimes_R N)$ use the associativity and the following property of the tensor product: $L\otimes_SS\simeq L$, where in this case $S=R/\mathfrak m$ and $L=R/\mathfrak m\otimes_RM$.
• I'm looking for a counterexample (with $M$ or $N$ not finitely generated) where the implication $M = 0$ or $N = 0$ does not hold. Do you have any ideas? – mathcourse Jul 2 '18 at 21:26