Let $K=F_p(X,Y)$, where $F_p$ is a finite field of characteristic $p$, and $F=F_p(X^p,Y^p)$. I have been given the following problem:

Determine the degree of extension $[K:F]$.

My experience with problems regarding the degree of field extensions is limited to the case where the field extension is generated by a finite number of elements, such as the prototypical example $\mathbb{Q}(\sqrt{2})/\mathbb{Q}$. In general, how does one proceed on a problem such as this, where one does not have a field extension generated by the adjunction of elements to a base field?

  • 2
    $\begingroup$ Consider the intermediate extension $F_p(X^p, Y)$. $\endgroup$ – user18119 Apr 18 '12 at 16:21
  • 1
    $\begingroup$ In this context, the base field is $F$, not $F_p$. $\endgroup$ – Lubin Apr 18 '12 at 16:47
  • $\begingroup$ In fact, there are infinitely many subextensions of degree $p$. $\endgroup$ – Sungjin Kim Jul 31 '15 at 23:27

Ok, so the basic idea whenever you have an extension which is the appendation of multiple elements is to take things one at a time. Namely, let's first find $n$ and $m$ in the following diagram

$$\begin{array}{c}F_p(X,Y)\\ \vert n\\ F_p(X^p,Y)\\ \vert m\\ F_p(X^p,Y^p)\end{array}$$

The basic idea now is that going from $F_p(X^p,Y^p)$ to $F_p(X^p,Y)$ is just appending one element since $F_p(X^p,Y)=F_p(X^p,Y^p)(Y)$. So, the question is what is the minimal polynomial of $Y$ over $F_p(X^p,Y^p)$? Well, we clearly have a candidate since $f(t)=t^p-Y^p\in F_p(X^p,Y^p)[t]$ and $f(Y)=0$. The question then is why $f(t)$ is irreducible in $F_p(X^p,Y^p)[t]$. The reason why would expect it to be is that any factorization should be one involving the fact that $t^p-Y^p=(t-Y)^p$--but of course this won't work since $t-Y\notin F_p(X^p,Y^p)[t]$. To make this more concrete suppose that $f(t)=p(t)q(t)\in F_p(X^p,Y^p)[t]$ is a factorization into irreducibles. Note then that we have factorized $f(t)$ in $F_p(X,Y)[t]$ and since $F_p(X,Y)$ is a field we know that $F_p(X,Y)[t]$ is a UFD and so use this to conclude that, up to constants, $p(t),q(t)$ are powers of $(t-Y)^k$. Conclude then that (from the primality of $p$ which tells you that $Y^k$ is not in $F_p(X^p,Y^p)$ for $k<p$) the one factor must be $(t-Y)^p$ and so the other is a unit. Conclude that $f$ is irreducible and so the minimal polynomial of $Y$ over $F_p(X^p,Y^p)$. Conclude that $m=p$.

Now you try for $n$ and then use the multiplicative property of field towers.


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.