1
$\begingroup$

Let $k$ be an algebraically closed field and $f: \mathbb{A}^2_k \rightarrow \mathbb{A}^2_k$, $(x,y) \mapsto (x,xy)$. I understand what the image looks like, i.e. $\mathbb{A^2_k} - (\mathbb{A}^1-0)$. By observing that: $\mathbb{A}^2_k = \textrm{Im} \cup \mathbb{A}^1_k$ and since $\mathbb{A}^2_k$ is irreducible, by taking the closure we get that $\overline{\textrm{Im}}$ is the whole plane and therefore that the image is not closed. If it was open, the intersection $\textrm{Im} \cap \{X \neq 0\} = 0$ would also be open, which is impossible because $\mathbb{A}^2_k$ is irreducible. Finally, since the image is dense and is not open in the whole set, it can't be locally closed either.

I find this reasoning to be a bit arbitrary. In particular, for the "locally closed" part, I initially wanted to look that at every neighbourhood of $\{(0,0)\}$, which would be the whole plane without some curves, but I feel like that is too tedious.

Is there a more intuitive (understand obvious) way to think about these things ? In general, is there an intuitive way to think about sets being closed or open in the Zariski topology?

$\endgroup$
4
  • $\begingroup$ Thank you for the quick answer. Perhaps I should have specified that this is exactly what I don't completely get: justifying that something CANNOT be the zero set of any polynomial. You say that in this a case, the polynomial is "clearly" zero but it is not so clear for me why exactly. $\endgroup$ Commented Oct 12, 2023 at 20:10
  • $\begingroup$ What I mean is that for the complement, though it is very natural to think that the polynomial should also vanish at the origin, I can't come up with a formal and direct argument that it should beyond the above proof. $\endgroup$ Commented Oct 12, 2023 at 20:15
  • $\begingroup$ We can write such a polynomial by separating its coefficients $ \sum k + \sum X + \sum Y + \sum XY$. By taking x=0, we're reducing the polynomial to the Y and constant parts, and hence to a polynomial on $k[Y]$, were it should only have finitely many roots. Hence, all coefficients of constants and the Y part are 0 and the polynomial can be factored by X, so it is zero everywhere. Is this it ? For the limit part, I'm unaware of what it refers to. $\endgroup$ Commented Oct 12, 2023 at 20:41
  • $\begingroup$ By zero everywhere, I meant zero at the origin as well :) $\endgroup$ Commented Oct 12, 2023 at 20:51

1 Answer 1

1
$\begingroup$

Here is a different method than suggested in the comments. If the image is closed (resp. open), it's intersection with any set must also be closed (resp. open). Now pick some good sets to check against: the two axes will be instructive.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .