Suppose we have an irreducible quintic polynomial $f(x)\in \mathbb{Q}[x]$ with 4 complex solutions, say for e.g. $x^5+x^4+x^3-2x^2-2x+5$. It is easy to see that the Galois group $Gal(E/\mathbb{Q})$ has a $5$- cycle.

Now if we label the complex roots as $a_1, a'_1, a_2, a'_2$ where $a'_i$ is the complex conjugate of $a_i$ and look at the automorphisms, $f:a_1 \to a'_1$ (fixing the rest)and $g:a_2\to a'_2$ (fixing the rest). But this does not give any information whether the $Gal(E/\mathbb{Q})\subset S_n$ or $A_n$.

So, what can we do further to get more information about how the Galois group of the polynomial looks like?

  • $\begingroup$ Why does it contain the two automorphisms $f$ and $g$ you describe? $\endgroup$ – Prahlad Vaidyanathan Mar 23 '15 at 2:25
  • $\begingroup$ One way to check if $G \subseteq A_n$ is to compute the discriminant. This paper discusses the topic, see p. 5. $\endgroup$ – André 3000 Mar 23 '15 at 2:25
  • $\begingroup$ @PrahladVaidyanathan: I am trying to permute the roots and try to get some information on the cycle structure $\endgroup$ – Rutherford Mark Mar 23 '15 at 4:02
  • $\begingroup$ @RutherfordMark: What I mean is, how do you know you can conjugate one pair of roots and fix the rest? As I see it, a priori, you can only assume that conjugation induces an element of the Galois group. This would give you the element $(12)(34) \in S_5$ $\endgroup$ – Prahlad Vaidyanathan Mar 23 '15 at 10:11
  • $\begingroup$ @PrahladVaidyanathan: That would mean that it has a $4$- cycle but it is not enough to conclude that the Galois group is $S_5$ $\endgroup$ – Rutherford Mark Mar 24 '15 at 2:15

Check all the transitive subgroups of $\mathfrak{S}_5$ you will have five groups (up to conjugation).

You have first the subgroup generated by one 5-cycle. Then its normalizer N in $\mathfrak{A}_5$, in $\mathfrak{S}_5$, $\mathfrak{A}_5$ and $\mathfrak{S}_5$.

For instance, if you have a 3-cycle in your Galois group then you are the whole symmetric group or just the antosymmetric one (then compute the discriminant of your polynomial).

To compute Galois group without a computer, I think you should do number theory (that's how you go further in the field).


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.