I'm studying (basic) Galois theory and I'm stuck in front of this exercise:

Find the degree of $\mathbb{Q}(\zeta_{169},\sqrt[169]{34})$ over $\mathbb{Q}$.

It seems a very standard exercise, anyway I cannot find a solution..maybe because I'm used to solve this kind of exercises directly by finding the minimal polynomial and now this is a little harder. Please, give me some hints or a solution.

Thank you in advantage.

  • $\begingroup$ All I've managed to prove is that $169\le[\mathbb{Q}(\zeta_{169},\sqrt[169]{34}):\Bbb Q]\le 26364=156\cdot 169$ from the degree-$156$ and $169$ minimal polynomials for these elements. I too would like to see what approach one uses to attack this sort of question. $\endgroup$ – Mario Carneiro Feb 6 '15 at 19:03
  • $\begingroup$ This is the splitting field of the polynomial $x^{169}-34$ over $\mathbb{Q}$, in particular it is a Galois extension. Moreover that polynomial is irreducible by Eisenstein criterium. Since the degree of $[\mathbb{Q} (\zeta_{169}):\mathbb{Q}] = \phi (169) = 12 \cdot 13$, we have that the degree is at least $12 \cdot 13^2$, and divides $12 \cdot 13^3$. $\endgroup$ – Crostul Feb 6 '15 at 19:05

I have some more thoughts:

If $\mathbb Q(\sqrt[169]{34},\zeta_{169})/\mathbb Q(\zeta_{169})$ would be of degree $13$ (we know $13$ or $13^2$, we want to rule out $13$ in the following), consider the action of $Gal(\mathbb Q(\sqrt[169]{34},\zeta_{169})/\mathbb Q(\zeta_{169})) \cong \mathbb Z/13\mathbb Z$ on the roots of $X^{169}-34$. The action is transitive on each irreducible factor. The lenghts of the orbits of each root is $1$ or $13$. But there is definitely no linear factor because $\sqrt[169]{34} \notin \mathbb Q(\zeta_{169})$, hence length $1$ can be ruled out. So $X^{169}-34$ splits into 13 factors of degree $13$ over $\mathbb Q(\zeta_{169})$.

Now let us multiply $13$ distinct roots of $X^{169}-34$. We get a number of the form $\zeta_{169}^m \sqrt[13]{34}$. On the other hand this is the constant term of an irreducible factor over $\mathbb Q(\zeta_{169})$. This shows $\sqrt[13]{34} \in \mathbb Q(\zeta_{169})$. But this is definitely a contradiction, since $\mathbb Q(\zeta_{169})/\mathbb Q$ is an abelian extension, hence admits only normal intermediate fields.

All in all we deduce that the degree of $\mathbb Q(\sqrt[169]{34},\zeta_{169})/\mathbb Q$ is $169 \cdot 12 \cdot 13$. And the galois group is the semidirect product of $\mathbb Z/169\mathbb Z$ and $(\mathbb Z/169\mathbb Z)^*$

Edit: Of course we can abbreviate the argument a little bit: If the degree of $\sqrt[169]{34}$ over $\mathbb Q(\zeta_{169})$ would be $13$, the minimal polynomial would be the product of $13$ linear factors of the polynomial $X^{169}-34$. The constant term (up to sign) has the form $\zeta_{169}^m \sqrt[13]{34}$, so we can deduce $\sqrt[13]{34} \in \mathbb Q(\zeta_{169})$, getting the contradiction above.

But i think the longer version somehow shows what galois theory is capable of, when everything comes together.


This is a partial answer.

This is the splitting field of the polynomial $x^{169}-34$ over $\mathbb{Q}$, in particular it is a Galois extension. This means that the degree coincides with the order of the Galois group $G$.

The polynomial $x^{169}-34$ is irreducible by Eisenstein criterium, so $$[\mathbb{Q}(\sqrt[169]{34}) : \mathbb{Q} ] = 169 = 13^2$$

This means that $G$ has a subgroup of index $169$.

Moreover $$[\mathbb{Q} (\zeta_{169}):\mathbb{Q}] = \varphi (169) = 12 \cdot 13$$ so $G$ has a subgroup of index $12 \cdot 13$


$$[\mathbb{Q}(\sqrt[169]{34}, \zeta_{169}) : \mathbb{Q}] = [\mathbb{Q}(\sqrt[169]{34}, \zeta_{169}) : \mathbb{Q}(\zeta_{169})] \cdot [\mathbb{Q} (\zeta_{169}):\mathbb{Q}]$$ and clearly $[\mathbb{Q}(\zeta_{169})(\sqrt[169]{34}) : \mathbb{Q}(\zeta_{169})]$ divides $[\mathbb{Q}(\sqrt[169]{34}) : \mathbb{Q}] = 13^2$, so it is $13$ or $13^2$.

So there are two options: the answer is $12 \cdot 13^2$, or it is $12 \cdot 13^3$. But now I am not able to conclude.


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.