I just encountered a piece of reasoning that I assume is standard but it's unfamiliar to me. Do you know a proof or a standard reference?
Setting: Suppose $R\rightarrow S$ is a map of commutative, unital rings, and $M,U$ are $R$-modules and $N$ an $S$-module. We have a natural map $$ \operatorname{Hom}_R(U,M)\otimes N \rightarrow \operatorname{Hom}_S(U\otimes S, M\otimes N)$$ given by $f\otimes n \mapsto f\otimes (1\mapsto n)$, with $f\in \operatorname{Hom}_R(U,M), n\in N$. (All tensor products are taken over $R$.)
Question: if $N$ is flat as an $R$-module, is this map always an isomorphism. If so, why?
This comes up on p. 13 of Bruns and Herzog's book Cohen-Macaulay Rings from 1993. In this context there are several other assumptions in place that I was assuming are extraneous to this particular point, but I would be happy to know otherwise. They are: (1) $R,S$ are Noetherian local rings; (2) $M,N$ are finite modules over $R,S$ respectively; and (3) $U$ is the residue field of $R$.