10
$\begingroup$

Let $K$ be a field and $x$ be trnacendental over $K$. Compute $[K(x):K(\frac{x^5}{1+x})]$.

I've never came across questions like these. It's easy to see that this degree is at most $5$, since: $$x^5-\alpha(x+1)=0$$ ($\alpha$ being $\frac{x^5}{1+x}$). But how can I show the other direction? Why can't there be any $f\in K(\alpha)[X]$ with $\deg f<5$ and $f(x)=0$? Or can there be?

Help would be appriciated. Thank you!

$\endgroup$
  • 1
    $\begingroup$ Small nitpick: be careful using $x$ both as the transcendental element over $K$ and as the polynomial variable. It is a recipe for confusion. $\endgroup$ – Arthur Jul 22 '17 at 9:51
  • 1
    $\begingroup$ Oops, right. I changed the polynomial cariable to capital X. $\endgroup$ – 35T41 Jul 22 '17 at 10:39
  • $\begingroup$ Well, $\alpha$ is certainly transcendental over $K$, so that $K[\alpha]$ is a polynomial ring, so Principal Ideal Domain. The polynomial $X^5-\alpha X-\alpha$ is Eisenstein over the PID $K[\alpha]$, thus irreducible. I think that suffices. $\endgroup$ – Lubin Jul 22 '17 at 13:47
4
$\begingroup$

We only need to show that the polynomial $x^5-\alpha x - \alpha$ is irreducible over $K(\alpha)[x]$. Note that the polynomial is primitive, thus it suffices to show it is irreducible over $K[\alpha][x]$.

We use the idea of Eisenstein criterion, let $\varphi:K[\alpha]\to K$ be the evaluation map sending $\alpha \mapsto 0$. Extend this homomorphism to get $\pi: K[\alpha][x]\to K[x]$. If $x^5-\alpha x -\alpha = fg$, with $f,g \in K[\alpha][x]$, $f,g$ non-unit. Then $$\pi(f)\pi(g) = x^5$$ Note that $\ker \pi = (\alpha)$, the ideal generated by $\alpha$.

Since $K[x]$ is a UFD, $f = x^r + \alpha f_0, g = x^{5-r} + \alpha g_0$, where $f_0,g_0\in K[\alpha][x]$. But this would imply that the constant term of $fg$ is a multiple of $\alpha^2$. Hence $x^5-\alpha x -\alpha$ is irreducible over $K[\alpha][x]$.

$\endgroup$
  • $\begingroup$ Thanks a nice extension of these terms that I know from the $\mathbb{Q}$ and $\mathbb{Z}$ world to these rather abstract fields. Thank you for that! $\endgroup$ – 35T41 Jul 22 '17 at 13:48
  • $\begingroup$ Very nice, I didn't think of Eisenstein criterion, +1. Though, you could have just applied it directly since $(\alpha)$ is prime ideal (you basically wrote a proof of Eisenstein criterion). $\endgroup$ – Ennar Jul 22 '17 at 14:48
1
$\begingroup$

What you have to do is show that $f(X)=X^5-\alpha(X+1)$ is irreducible over $K(\alpha)$ and you are done.

This is because we would have $[K(\alpha)(x):K(x)][K(x):K(\alpha)] = [K(\alpha)(x):K(\alpha)] = 5$ and $K(x)\neq K(\alpha)$ would imply $[K(x):K(\alpha)]=5$ by $5$ being prime.

Now, $f$ is primitive polynomial in $K[\alpha][X]$, so by Gauss's lemma, it is irreducible over $K(\alpha)$ if and only if it is irreducible over $K[\alpha]$.

First of all, $f$ can't have a root in $K[\alpha]$ since it would mean that $$p(\alpha)^5-\alpha(p(\alpha)+1)=0$$ and that $\alpha$ is algebraic over $K$.

Thus, if $f$ weren't irreducible, it would have to allow factorization of the form

$$f(X) = (X^3 +a(\alpha)X^2+b(\alpha)X+c(\alpha))(X^2+d(\alpha)X+e(\alpha))$$ which would lead to system

\begin{align} a(\alpha)+d(\alpha)&=0\\ b(\alpha)+a(\alpha)d(\alpha)+e(\alpha)&= 0\\ c(\alpha)+b(\alpha)d(\alpha)+a(\alpha)e(\alpha) &= 0\\ c(\alpha)d(\alpha)+b(\alpha)e(\alpha)&= -\alpha\\ c(\alpha)e(\alpha)&=-\alpha \end{align}

and now you can use that $K[X]\cong K[\alpha]$ since $\alpha$ is transcedental and degree arguments to show that the above system is impossible to solve in $K[X]$.

$\endgroup$

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.