Consider the finite field $\mathbb{F}_3$ and define the polynomial $f(x)=x^4+x-1$ over $\mathbb{F}_3$.

I want to find its Galois Group.

I observe that $f$ has no root over $\mathbb{F}_3$, so if it is reducible, it must be as quadratic factors. Is there any way to show that this isn't possible?

Assuming it isn't (which is the case), we adjoin a root of $f$, $\alpha$ say to form $F=\mathbb{F}_3(\alpha)$.

This is a degree $4$ extension (since $f$ is irreducible) and so $\mathbb{F}_3(\alpha)\cong \mathbb{F}_{81}$.

Now we need to determine if the other roots of $f$ are in this field - if so then we know our Galois Group has size $4$ and we can work out if it's $C_4$ or $V_4$.

How can I work out whether these roots are in $\mathbb{F}_3 (\alpha)$?


To show that $f$ is irreducible over $\mathbb{F}_{3}$, you may try dividing it by the monic irreducible polynomials of degree $2$. Actually, you need check for all but one, as $f$ has no multiple roots, and thus it is not a square.

So check that $f(x)$ is not divisible by $x^{2} + 1$ and $x^{2} + x - 1$, and you are done with the irreducibility.

You should have seen if $E$ is a finite field, which is an extension of the (necessarily finite) field $F$, then $E/F$ is Galois. Thus $\mathbb{F}_{81}/\mathbb{F}_{3}$ is Galois, and its Galois group is cyclic, generated by the Frobenius automorphism $\sigma : z \mapsto z^ {3}$.

So if $\alpha$ is a root, the four roots are $$ \alpha, \alpha^{3}, \alpha^{9}, \alpha^{27}. $$ You can use the fact that $f(\alpha) = 0$ to express these roots in the form $$a_{0} + a_{1} \alpha + a_{2} \alpha^{2} + a_{3} \alpha^{3} \in \mathbb{F}_{3}[\alpha],$$ with $a_{i} \in \mathbb{F}_{3}$.

  • $\begingroup$ Thanks, how about determining whether $\mathbb{F}_3(\alpha)$ is in fact the splitting field? $\endgroup$
    – Mathmo
    May 14 '14 at 11:17
  • $\begingroup$ You're welcome! I should have addressed your question with the most recent edit, please feel free to write again in case. $\endgroup$ May 14 '14 at 11:20
  • $\begingroup$ 'every extension of finite fields is Galois'.... this is absolutely not true! $\endgroup$
    – Dan
    Nov 18 '14 at 9:43
  • $\begingroup$ @reinbot, quite right, fixed. $\endgroup$ Nov 18 '14 at 12:15

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.