Inclusion of affine algebraic sets - converse direction

60 Views Asked by At

Let $I, J \subseteq K[x_0, \dotsc, x_n ]$ with $I \subseteq J$, then I have $V(J) \subseteq V(I)$ (looking at it from the affine space).

Now I was wondering whether also the converse direction holds, i.e. for $V(J) \subseteq V(I)$, I can deduce that $I \subseteq J$?

So far, I've always only seen it stated in one direction.

It would be great if anybody could tell whether or under which circumstances also the other directions holds. Thank you!

1

There are 1 best solutions below

7
On BEST ANSWER

The converse is not necessarily true. Consider the ideals $$J=((x-1)(x-2)^2),~I=((x-1)(x-2)(x-3))\subseteq K[x]$$

Then $$V(J)=\{1,2\}\subset V(I)=\{1,2,3\}$$ but $$I\not\subset J$$

I think you can also see why this is a problem, it's because of $(x-2)^2$, the square. It doesn't change the set of roots, but it makes the ideal smaller.

What is true, is that if $K$ is algrebraically closed, then $$V(J)\subseteq V(I)\rightarrow \sqrt I\subseteq\sqrt J$$ because then we have no square factors. This can be proven using Hilbert's Nullstellensatz.

Indeed, $\sqrt J=((x-1)(x-2))$ and $\sqrt I=((x-1)(x-2)(x-3))$, so $$\sqrt I\subseteq \sqrt J$$


An even simpler counter-example:

Consider the special case in which $V(I)=V(J)$ (that is $V(I)\subseteq V(J)$ and $V(J)\subseteq V(I)$), if the conserve was true this would imply $J\subseteq I$ and $I\subseteq J$, that is, $I=J$. Now take $$I=(x-1)^2,~J=(x-1)\subseteq K[x]$$

$V(I)=V(J)$, but $I\neq J$.