5
$\begingroup$

Let the ring be $A=k[x^2,x^3]$ where $k$ is a field. Let $I$ be the ideal generated by $x^4$. I want to prove that there's only one ideal in $A/I$. For that I considered that in $k[x]$ the ideals that contain $x^4$ are exactly the ideal of lower power of $x$. So that leaves me the choices of $(x^3), (x^2), (x)$ but only the ideal of $(x^2)$ contains $x^4$ in A. So the only ideal in $A/I$ is exactly $(\bar{x^2})$.

Is this reasoning/result correct? Thanks!

$\endgroup$
  • $\begingroup$ What about the ideal $(x^3,x^4)$? $\endgroup$ – Mohan Jul 6 '18 at 17:39
  • $\begingroup$ @Arthur How so? There's something I'm not quite understanding. $\endgroup$ – Leo Lerena Jul 6 '18 at 17:42
  • 1
    $\begingroup$ Yes, but what have you tried? Any ideal of $A/I$ corresponds to an ideal of $A$ containing $I$. There are only finitely many of them and easy to write down with a little thought. $\endgroup$ – Mohan Jul 6 '18 at 18:46
  • 1
    $\begingroup$ @Mohan Finitely many? It seems like that depends on $k$. $\endgroup$ – rschwieb Jul 6 '18 at 18:54
  • 1
    $\begingroup$ @rschwieb No, I did not mean finitely many and was a mistake. $\endgroup$ – Mohan Jul 6 '18 at 21:50
1
$\begingroup$

I am doing more than just determining the ideals of $A/I$.

Note that $A\cong k[y,z]/\langle y^3-z^2\rangle=:B$ via the ring isomorphism $\varphi:B\to A$ sending $y\mapsto x^2$ and $z\mapsto x^3$. Under $\varphi$, the ideal $I$ of $A$ is associated to the ideal $\langle y^2,y^3-z^2\rangle/\langle y^3-z^2\rangle$ of $B$. That is, $$A/I \cong k[y,z]/\langle y^2,y^3-z^2\rangle=:C\,,$$ which is a $4$-dimensional vector space over $k$ with basis $\{1,\bar{y},\bar{z},\bar{y}\bar{z}\}$, where $\bar{y}$ and $\bar{z}$ are the images of $y$ and $z$, respectively, under the canonical projection $k[y,z]\to k[y,z]/\langle y^2,y^3-z^2\rangle$.

Observe that $\bar{y}^2=0$ and $\bar{z}^2=0$ in $C$. Hence, $C$ is in fact isomorphic as a ring to $$k[u,v]/\langle u^2,v^2\rangle\cong k[u]/\langle u^2\rangle \otimes k[v]/\langle v^2\rangle\,,$$ where $\otimes$ is the tensor product of algebras over $k$. Note that $C$ has many ideals: the zero ideal $0$, the whole ring $C$ itself, the primitive ideals $\langle \bar{y}+\alpha\bar{z}+\beta \bar{y}\bar{z}\rangle$ with $\alpha,\beta \in k$, the primitive ideals $\langle \bar{z}+\gamma \bar{y}\bar{z}\rangle$ with $\gamma\in k$, the primitive ideal $\langle \bar{y}\bar{z}\rangle$, and the unique maximal ideal $\langle \bar{y},\bar{z}\rangle$. If we translate this list back to ideals of $A/I$, then all the ideals of $A/I$ are:

  • trivial ideals: $0$ and $A/I$;
  • primitive ideals: $\langle \overline{x^2}+\alpha\overline{x^3}+\beta\overline{x^5}\rangle$, $\langle \overline{x^3}+\gamma\overline{x^5}\rangle$, and $\langle \overline{x^5}\rangle$, where $\alpha,\beta,\gamma \in k$;
  • maximal ideal: $\langle \overline{x^2},\overline{x^3}\rangle$.

Here, $\overline{x^2}$, $\overline{x^3}$, and $\overline{x^5}$ are the images of $x^2$, $x^3$, and $x^5$ under the canonical projection $A\to A/I$. You can say that $A/I$ is a local ring.

$\endgroup$
  • 1
    $\begingroup$ Maybe it's useful to notice that $\langle y^2,y^3-z^2\rangle=\langle y^2,z^2\rangle$. This way there is no need to introduce two more variables $u,v$. $\endgroup$ – user26857 Jul 6 '18 at 22:16

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.