I've been working on this problem for a couple days and I haven't heard back from any of my support team (TAs/Prof) at Uni. Not looking for an answer, but help and hints would be greatly appreciated. Thank you!


Let $f \in \mathbb{Q}[x]$ be an irreducible quintic. Let K be it's splitting field. $f$ has Galois group $D_{10}$ (Dihedral of order 10). Let $\mathbb{Q}(\sqrt{d})$ be the unique subfield of K of degree 2 over $\mathbb{Q}$.

Let $\zeta = e^\frac{2\pi}{5}$. Show that the Galois group of the extension $K(\zeta):\mathbb{Q}(\zeta)$ is also Dihedral of order 10 unless $d = 5$.

My thoughts so far:

So, I know that $\mathbb{Q}(\zeta)$ is a degree four extension over $\mathbb{Q}$. I also know that since $f$ has Galois group $D_{10}$, K must have degree 10 over $\mathbb{Q}$.

I know that the subfield structure for K must mirror the subfield structure for $D_{10}$. So, first, $\mathbb{Q}(\sqrt{d})$ corresponds to $C_5$. Also, given this information about $d$, we know that $f$ has at least one real root.

Basically, to solve the problem, we only need to show that $|K(\zeta):\mathbb{Q}(\zeta)| = 10$, since there is only one option for Galois groups of order 10.

Now, our professor told us to look at the subfields of $K(\zeta)$, K and $\mathbb{Q}(\zeta)$ and to look at $K \cap \mathbb{Q}(\zeta)$. This is where I am stuck. I don't know how to think about this intersection because I have no clue how to find what else is in K given the information we have.


Let us fix notations: $K/\mathbf Q$ is Galois with group $\cong D_{10}$ and $\mathbf Q(\sqrt d)$ is its unique quadratic subfield. Denoting by $\zeta$ a primitive $5$-th root of unity, you want to show that $K(\zeta)/\mathbf Q(\zeta)$ is Galois with group $\cong D_{10}$ unless $d=5$ (more precisely, unless $5d^{-1}$ is a square in $\mathbf Q$).

It is well known that the cyclotomic extension $\mathbf Q(\zeta)/\mathbf Q$ is cyclic of degree $4$, hence admits a unique quadratic subfield, which is here its maximal real subfield $\mathbf Q(\zeta)^+=\mathbf Q(\zeta+{\zeta}^{-1})=\mathbf Q(\sqrt 5)$. Because $gcd (4,10)=2$, it follows that, unless $d=5$, the extensions $\mathbf Q(\zeta)$ and $K$ are linearly disjoint over $\mathbf Q$, hence $Gal(K/\mathbf Q) \cong Gal(K(\zeta)/\mathbf Q(\zeta))$, and we are done.

NB: For any odd prime $p$, it is known from the computation of the discriminant of $\mathbf Q(\zeta_p)/\mathbf Q$ that the unique quadratic subfield of $\mathbf Q(\zeta_p)$ is $\mathbf Q(\sqrt {p^*})$, where $p^*={(-1)}^{\frac {p-1}{2}}$, so your problem (and answer) can be generalized when replacing $5$ by $p$, and the condition $d\neq 5$ by $d\neq p^*$.


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.