Let $L = \mathbb{Q}(\sqrt[3]{5},\omega)$ be an extension of the field of rational numbers $\mathbb{Q}$, where $\omega = \frac{-1+\sqrt{3}}{2}$.

(i) Show that the Galois group $G(L/\mathbb{Q})$ is isomorphic to $S_3$.

(ii) Let $H\subset G$ be the unique Sylow 3-subgroup of $G$. Find an element $\alpha \in L$ such that $LH = \mathbb{Q}(\alpha)$ where $LH$ is the intermediate field $\mathbb{Q} \subset LH \subset L$ fixed by $H$.

For (i), I know that we have $$L = \mathbb{Q}(\sqrt[3]{5},\omega) = \mathbb{Q}(\sqrt[3]{5}, 1+\sqrt{-3}) = \mathbb{Q}(\sqrt[3]{5},\sqrt{-3}) = \mathbb{Q}(\sqrt{-3})(\sqrt[3]{5}).$$ And the minimal polynomials for $\sqrt[3]{5}$ and $\sqrt{-3}$ are $x^3-5$ and $x^2+3$, respectively. Now $[L : \mathbb{Q}]\leq 3\cdot2=6.$ Now To show we actually have equality, we just need to show that $\sqrt[3]{5}\notin\mathbb{Q}(\sqrt{-3}).$ Indeed, let $a,b\in\mathbb{Q}$, and suppose $\sqrt[3]{5} = a+b\sqrt{-3}. $ Then

\begin{align} (a+b\sqrt{-3})^3 &= 5\\ a^3-9ab^2+(3a^2b-3b^3)\sqrt{-3}&=5\\ \frac{5-a^3+9ab^2}{3a^2b-3b^3}&=\sqrt{-3}, \end{align} which implies that $\sqrt{-3}\in\mathbb{Q}$, a contradiction. Hence the extension has degree 6. I'm not sure how I would show the Galois group is nonabelian to conclude that $G\cong S_3.$

Alternatively, I know that $L$ is the splitting field for $(x^3-5)(x^2+3)$. But the Galois group for this polynomial appears to be $D_6$, the dihedral group of order 12. I determined this just by listing out all of the automorphisms of the roots. Where am I going wrong here?

  • $\begingroup$ Well $D_6$ and $S_3$ are isomorphic. Pick two automorphisms and evaluate them at an element to show that they don't commute. $\endgroup$ – Ravi Jul 25 '16 at 17:39
  • $\begingroup$ $D_6$ and $S_3$ have different orders, so they aren't isomorphic. $\endgroup$ – user346096 Jul 25 '16 at 17:48
  • $\begingroup$ I think $D_6$ here refers to the dihedral group of order $6$, not the symmetries of a hexagon. There's always some confusion about whether to use $D_n$ or $D_{2n}$ to refer to the Dihedral group of an $n$-gon. $\endgroup$ – Ravi Jul 25 '16 at 18:00

After you know the degree of the extension is 6, you don't need to compute anything else. $L$ is the splitting field of $x^3-5$ over $\mathbb{Q}$, hence is Galois and the Galois group permutes the root of the polynomial $x^3-5$. So it's an order 6 subgroup of $S_3$, therefore all of it.

Now the desired subgroup is generated by the automorphism which maps $\sqrt[3]{5} \mapsto \omega \sqrt[3]{5}$ and fixes $\omega$. Hence the associated fixed field is $\mathbb{Q}(\omega)$, which one verifies by checking degrees.


As you mention, we have the fields tower

$$\Bbb Q\subset\Bbb Q(\sqrt[3]5)\subset L:=\Bbb Q(\sqrt[3]5, \omega)\;,\;\;\;\omega=e^{2\pi i/3}$$

and as you pointed out, we in fact have that $\;L=\Bbb Q(\sqrt[3]5,\,\sqrt{3}\,i)\;$ (and in fact $\;L=\Bbb Q(\sqrt[3]5+\sqrt3\,i)$)

Now, every automorphism of $\;L/\Bbb Q\;$ maps roots of some irreducible polynomial to roots of this same polynomial, so you have here two automorphisms:


and now check that $\;\sigma\tau\neq\tau\sigma\;$ , by observing that



and then:



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.