The following is a known result in algebraic geometry:

Let $k$ be an algebraically closed field of characteristic zero (for example, $k=\mathbb{C}$). Let $L$ be a field such that $k \subset L \subset k(x,y)$ and $L$ is of transcendence degree two over $k$. Then there exist $h_1,h_2 \in k(x,y)$ such that $L=k(h_1,h_2)$.

Is it possible to find $g_1,g_2 \in k[x,y]$ such that $L=k(g_1,g_2)$?

The motivation is the following result: If $k \subset L \subset k(x,y)$ is of transcendence degree one over $k$, then $L=k(h)$, where $h \in k[x,y]$; see this answer. Perhaps the arguments in that answer are also applicable here?

I have asked the above question in MSE, with no comments. See also this question.

Thank you very much!

  • $\begingroup$ @YCor, thank you very much for adding the highly relevant tags. $\endgroup$ – user237522 Mar 3 at 19:02

No. $M\mathrel{:=}k(x,y)$ has a $k$-automorphism $\sigma:x\mapsto 1/x,\,y\mapsto 1/y$, of order 2. Let $G\mathrel{:=}\langle\sigma\rangle$, and put $L\mathrel{:=}M^{G}$, the fixed field. The elements $x+1/x$ and $y+1/y$ of $L$ are algebraically independent over $k$, hence $L$ has transcendency degree 2. However, no non-constant polynomial $g\in k[x,y]$ can be in $L$ (that is, be invariant under $\sigma$).

By the same token, the automorphism $\tau:x\mapsto 1/y,\,y\mapsto 1/x$, fixes the field $k(x/y)$, which has transcendency degree 1 over $k$, but it cannot fix any non-constant polynomial. So $k(x/y)\neq k(h)$ for every $h\in k[x,y]$. (Here, $M$ is again quadratic over the fixed field of $\tau$, so that the latter is of tr. deg. 2 over $k$ again. A transcendental element independent of $x/y$ is $x+1/y$, for instance).

| cite | improve this answer | |
  • $\begingroup$ Thank you for the nice answer! $\endgroup$ – user237522 Apr 11 at 21:27

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.