3
$\begingroup$

If $F=\mathbb{F}_p(x)$ is the rational function field in one variable over $\mathbb{F}_p$, find $[F:F^p]$.*

Here's where I am at with this question: Using the multinomial theorem (and some tears), I can prove that $$\left(\sum_{i=1}^na_ix^i\right)^p=\sum_{i=1}^na_ix^{ip}.$$

This at least shows that $\{1,x,x^2,...x^{p-1}\}$ is a basis for $\mathbb{F}_p[x]$ over $\mathbb{F}_p[x]^p$. But I am not sure how to get from here to the field of fractions.

My other thought was to try to find the minimal polynomial of $x$ over $F^p$, which I guess would be $t^p-x\in \mathbb{F}(x)^p[t]$, but I am not certain if $\mathbb{F}_p(x)$ is generated over $\mathbb{F}_p(x)^p$ by $x$.

Are any of these approaches sensible? How do I deal with the nasty quotients?

*Patrick Morandi: Field and Galois Theory

$\endgroup$
  • 2
    $\begingroup$ Note that $a(x)/b(x) = a(x)b(x)^{p-1} / b(x)^p$, and try applying your basis argument to the numerator. $\endgroup$ – user125932 Jul 6 at 23:59
  • 2
    $\begingroup$ For $a,b \in \Bbb{F}_p[x]$ then $(a(x)/b(x))^p = a(x^p)/b(x^p)$ thus $\Bbb{F}_p(x)^p = \Bbb{F}_p(x^p)$. In general $K(u_1,\ldots,u_n)^p = K^p(u_1^p,\ldots,u_n^p)$. $\endgroup$ – reuns Jul 7 at 1:20
  • 1
    $\begingroup$ Note it is enough to show (in characteristic $p$ commutative ring) $(u+v)^p = u^p+v^p$ to obtain that $p$ is a ring homomorphism, that here it is injective and $\left(\sum_{i=1}^na_ix^{i-1}\right)^p=\sum_{i=1}^na_ix^{(i-1)p}$, $(a(x)/b(x))^p = a(x^p)/b(x^p)$ $\endgroup$ – reuns Jul 7 at 2:06
1
$\begingroup$

First we will show that $$\left(\sum_{i=1}^na_ix^{i-1}\right)^p=\sum_{i=1}^na_ix^{(i-1)p}.$$ We will use (without proof) the multinomial theorem: $$\left(\sum_{i=1}^na_ix^{(i-1)}\right)^p=\sum_{k_1,k_2,...k_n}\binom{p}{k_1,k_2,...k_n}\prod_{t=1}^na_t^{k_t}x^{(i-1)k_t}$$ where the sum runs over all combinations of $n$ non-negative integers $k_i$ which sum to $p$. Now, if $k_i\neq p$ $$\binom{p}{k_1,k_2,...k_n}=\frac{p!}{k_1!k_2!...k_n!}, ~~~~~~~~\text{so}~~~~~~~~p\Big|\binom{p}{k_1,k_2,...k_n}$$ because the denominator cannot have $p$ as a prime factor since all the $k_i<p$. So, in a finite field $\mathbb{F}_p$, we are left with only the terms where one $k_i$ if $p$ and the others are zero, and there are $n$ such terms which look like: $$\sum_{0,0,...k_i,0,0...}\binom{p}{0,0,...k_i,0,0...}\prod_{t=1}^na_t^{k_t}x^{(i-1)k_t}=a_i^{k_i}x^{(i-1)k_i}=a_i^px^{(i-1)p}$$ $$\text{Therefore, }~~~~~~\left(\sum_{i=1}^na_ix^{i-1}\right)^p=\sum_{i=1}^na_ix^{(i-1)p}.$$ By this formula, $\mathbb{F}_p[x]$ is generated over $\mathbb{F}_p[x]^p$ by the basis $\{1,x,x^2...x^{p-1}\}$. Now, by user125932's observation, if $\frac{f(x)}{g(x)}\in F$, then $\frac{f(x)}{g(x)}=\frac{f(x)g(x)^{p-1}}{g(x)^p}$. By the above argument $f(x)g(x)^{p-1}\in \mathbb{F}_p(x)$, so it is generated over $F^p$ by a basis of p elements, and clearly $g(x)^p \in F^p$.

Therefore, $[F:F^p]=p$.

$\endgroup$

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.