1

Let $k$ be a field and suppose $F,G$ do not have common factors in $k[X,Y]$.

Then $V(F,G) = \{P \in k^2\mid F(P) = 0 = G(P)\}$ is finite.

The proof my book provides does the following:

Because $F,G$ have no common factors in $k[X,Y] = k[X][Y]$, they don't have common factors in $k(X)[Y]$ (here $k(X)$ is the fraction field of $k[X]$). Because $k(X)[Y]$ is a PID, we have

$$(F,G) = 1$$ in $k(X)[Y]$

Questions:

(1) I can see that if $F,G$ have no non-constant common factors in $k[X,Y] = k[X][Y]$, they don't have common non-constant factors in $k(X)[Y]$. Shouldn't this be included in the proof?

(2) Why does it follow that $(F,G) = 1?$ I know that a gcd always exists in a PID.

2 Answers2

1

A constant in $k(X)[Y]$ is an element of $k(X)$ and hence a unit. Units are excluded from the definition of factor because they divide everything. Greatest Common Divisors and Least Common Multiples are only meaningfully defined up to unit multiples. With this in mind, knowing that $F,G$ have no common factors in $k(X)[Y]$, you know that their $GCD$ is $1$. In a PID, the GCD of two elements generates their ideal.

Badam Baplan
  • 8,688
0

The first claim (1) is Gauss' lemma. Second claim follows from the fact $(F,G)$ is an ideal, and it must be principal. It's generator is the gcd of $F$ and $G$, which is in this case $1$ because we assumed $F$ and $G$ have no common factors in $\mathbb k[X,Y]$ and by Gauss' lemma this transfers to $\mathbb k(X)[Y]$.

Pedro
  • 122,002