1

Let $R$ be a not necessarily commutative unital ring, and $R \hookrightarrow S$ a unital ring extension such that $S$ is flat as a (right) module over $S$ (i.e. a flat extension). See here for a discussion of flat extensions.

I am trying to learn about flatness in general, and began to wonder how flat extensions relate to flatness of modules. So here is my question: Is it then true in general that any (faithfully) flat $S$-module $M$ is (faithfully) flat as an $R$-module?

I guess that the answer to my question is well a known-known standard fact - and I would appreciate a standard reference. For example, the fact I am looking for appears to be discussed in the answer of Georges Elencwajg to this question in the second point of Formal properties.

  • Welcome to MSE. Your question is phrased as an isolated problem, without any further information or context. This does not match many users' quality standards, so it may attract downvotes, or closed. To prevent that, please [edit] the question. This will help you recognise and resolve the issues. Concretely: please provide context, and include your work and thoughts on the problem. These changes can help in formulating more appropriate answers. – José Carlos Santos Sep 20 '21 at 06:59
  • I am trying to learn abut flatness in general, not realy looking at specific examples. I guessed that the answer to my question is well known and that I am really looking for a reference. I have edited. – Martim Pereir Sep 20 '21 at 07:13

1 Answers1

2

It is true that if $R\to S$ is flat and $M$ is a flat $S$-module, then $M$ is flat as an $R$-module, but it is not true that if $R\to S$ is flat and $M$ is a faithfully flat $S$-module, then $M$ is faithfully flat as an $R$-module.

First claim: Let $M$ be a flat $S$-module, and let $$N'\to N\to N''$$ be an exact sequence of $R$-modules. We want to show that $$N'\otimes_R M\to N\otimes_R M\to N''\otimes M$$ is exact.

Since $R\to S$ is flat, we have that $$N'\otimes_R S\to N\otimes_R S\to N''\otimes_R S$$ is exact. Since $M$ is flat over $S,$ the sequence $$N'\otimes_R S\otimes_S M\to N\otimes_R S\otimes_S M\to N''\otimes_R S\otimes_S M$$ is exact as well. However, for any $R$-module $N,$ we have $N\otimes_R S\otimes_S M\cong N\otimes_R M,$ so $$N'\otimes_R M\to N\otimes_R M\to N''\otimes M$$ is exact, as desired.

Second claim: Now, suppose that $M$ is faithfully flat over $S.$ We already know that $M$ is flat over $R.$ Suppose that $N$ is an $R$-module such that $N\otimes_R M\cong 0.$ To prove $M$ is faithfully flat over $R,$ we would need to show that $N\cong 0.$ We have \begin{align*} 0&\cong N\otimes_R M\\ &\cong N\otimes_R S\otimes_S M, \end{align*} and since $M$ is faithfully flat over $S,$ it follows that $N\otimes_R S\cong 0.$ However, unless we assume that $S$ is faithfully flat over $R,$ it may not be the case that $N\cong 0.$ We can use this to construct a counterexample to the claim that if $R\to S$ is flat and $M$ is faithfully flat over $S,$ then $M$ is faithfully flat over $R.$

Let $R = \Bbb{Z},$ $S = M = \Bbb{Q}.$ Then $M$ is clearly faithfully flat over $S,$ but we have $\Bbb{F}_p\otimes_R M\cong \Bbb{F}_p\otimes_\Bbb{Z}\Bbb{Q}\cong 0,$ although $\Bbb{F}_p$ is not $0.$ This implies that $M$ is not faithfully flat over $R.$ More generally, letting $S$ be any flat extension of $R$ which is not faithfully flat over $R$ and taking $M = S$ will provide a counterexample to this second claim.

Stahl
  • 23,855