0
$\begingroup$

Q is a rational field. Q[x] is polynomial ring over Q 。(x) is maximal ideal of Q[x]. Take Q[x]/(x) as a module over Q[x]. Then what is Q[x]-module Q[x]/(x) localize at 0??

I think the result is Q[x]/(x) \ otimes_{Q[x]}Q(x) but on the other hand, from another way, I know it should be Q[1/x]/Q.

But how can I prove they are isomorphism?

$\endgroup$
6
  • $\begingroup$ This looks like homework and is not appropriate for MO. Have you read the FAQ? $\endgroup$ Dec 13, 2009 at 16:13
  • $\begingroup$ Sorry. In fact, I am using very general framework to revise the constructing injectives in Grothendieck category. I want to test this machinary using this example. However, I suddenly found that I could not prove this isomorphism $\endgroup$
    – MAJIA
    Dec 13, 2009 at 16:30
  • 1
    $\begingroup$ Do you mean localized at the point x=0, or localized at the prime ideal (0)? $\endgroup$ Dec 13, 2009 at 17:07
  • $\begingroup$ I am sorry. The multiplicative set should be S=Q[x]-(0) $\endgroup$
    – MAJIA
    Dec 13, 2009 at 17:27
  • 2
    $\begingroup$ Connected to Scott's answer, you may also like to note that $\mathbb Q[x]/(x) \otimes_{\mathbb Q[x]}\mathbb Q(x) = 0$ since you can always take any basis element $r\otimes f$ and rewrite it as $r\otimes fx = rx\otimes f = 0$. Also, just by the recognition that $\mathbb Q[x]-(0)$ is your multiplicative set, you'll be able to see that the module vanishes; basically by the same argument again: x annihilates all of $\mathbb Q[x]/(x)$, and thus, if $x$ suddenly is invertible, we can take any $r\in\mathbb Q[x]/(x)$ and rewrite its image in the localization to $rx/x = 0/x = 0$. $\endgroup$ Dec 13, 2009 at 17:46

2 Answers 2

7
$\begingroup$

If you localize Q[x] at zero, you get the field of rational functions Q(x), so your module localizes to a vector space over this field (see Wikipedia). Your module is finitely generated, in fact by a single element, so you get a vector space of dimension zero or one. To decide which is correct, we can use Proposition 2.1 in Eisenbud's Commutative algebra (with a view toward algebraic geometry): If M is finitely generated, then $M[U^{-1}] = 0$ if and only if M is annihilated by an element of U. In this case, x is such an annihilating element, and your localized module is zero. Therefore, your first answer is correct, and your second answer is wrong.

$\endgroup$
0
$\begingroup$

Isn't $Q[x]/(x)$ just $Q$. In general $Q[x]/(x-\alpha)$ is just evaluation of all polynomials at alpha.

Then localization does nothing as we already have a full field of fractions.

$\endgroup$
1
  • $\begingroup$ yes it is. Q as a Q[x]-module $\endgroup$
    – MAJIA
    Dec 13, 2009 at 20:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.