As mentioned before, I am self-studying some commutative algebra out of "A Course in Commutative Algebra" by Kemper. In the text, Kemper has the following problem:

Let $X$ and $Y$ be affine varieties over a field $ K$, and let $f: X → Y$ be a morphism with induced homomorphism $ϕ: K[Y ] → K[X].$ We say that $f$ is dominant if the image $f(X)$ is dense in $Y$ , i.e., $\overline{f(X)} = Y$ .

Show that $f$ is dominant if and only if ϕ is injective.

He provides a proof of the forward implication and the reverse implication. I have come terms in understanding the forwards implication, but I have two questions regarding the reverse implication. Here is the proof given:

Assume that $ϕ$ is injective. Write $K[X] = K[x_1, . . . , x_m]/I, K[Y ] = K[y_1, . . . , y_n]/J,$ and let $f$ be given by $f_1, . . . , f_n ∈ K[x_1, . . . , x_m].$ Take $g ∈ \mathcal{I}_{K[y_1,...,y_n]} (f(X)).$ Then the polynomial $g(f_1, . . . , f_n) ∈ K[x_1, . . . , x_m]$ vanishes on $X$, so $ϕ(g + J) = 0.$ This implies $g ∈ J.$ We have shown that $\mathcal{I}_{K[y_1,...,y_n]} (f(X)) ⊆ J$, so

$$Y ⊆ \mathcal{V}_{K^n} (J) ⊆ \mathcal{V}_{K^n} (\mathcal{I}_{K[y_1,...,y_n]}(f(X))) = \overline{f(X)} ⊆ Y$$$ \square$

Now form here, I can see why $f(X)$ is dominant.

Before I ask my questions let me clarify some of the author's notations:

$\mathcal{I}_{K[y_1,...,y_n]} (f(X))$ is the vanishing ideal of $f(X)$.

$\mathcal{K^n}(J)$ is the affine variety given by $J$.

For the most part, I am okay with the proof, my questions are as follows:

  1. What justifies the containment $Y ⊆ \mathcal{V}_{K^n} (J)$?

  2. What justifies the equality $\mathcal{V}_{K^n} (\mathcal{I}_{K[y_1,...,y_n]}(f(X))) = \overline{f(X)}$?

Any help would be appreciated.

  1. We have $k[Y]=K[y_1,\ldots,y_n]/J$, this means that $Y\subset K^n$ is exactly the vanishing set of $J$. Or we can say that the class of any function $a\in J$ is equal to zero in $K[y_1,\ldots,y_n]/J$, hence for any maximal (or prime) ideal $\mathfrak m\subset K[y_1,\ldots,y_n]/J$ the element $a$ is equal to zero in $(K[y_1,\ldots,y_n]/J)/\mathfrak m$, i.e. it vanishes on the corresponding point.
  2. $\mathcal V(\mathcal I_{K[y_1,\ldots,y_n]}(f(X)))$ is a closed set containing $f(X)$, hence $\overline{f(X)}\subset \mathcal V(\mathcal I_{K[y_1,\ldots,y_n]}(f(X)))$. Conversely, if $p\in K^n$ does not lie in the closure, then we can choose a distinguished open set $D(b)$ containing $p$ and not intersecting with $\overline{f(X)}$. This $b$ vanishes on $f(X)$, and does not vanish on $p$, hence $p\notin\mathcal V(\mathcal I_{K[y_1,\ldots,y_n]}(f(X)))$.

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.