Consider the polynomial $f(x) = x^3 + \zeta x + \sqrt{3}$ in $\mathbb{C}[x],$ where $\zeta$ is a primitive third root of unity. Given a root $\alpha$ of $f(x)$ in $\mathbb{C},$ prove that $4 \leq [\mathbb{Q}(\alpha) : \mathbb{Q}] \leq 12.$

We note first that $[\mathbb{Q}(\sqrt{3}, \zeta) : \mathbb{Q}] = 4.$ Either $f(x)$ is irreducible in $\mathbb{Q}(\sqrt{3}, \zeta)[x]$ or not. Given that $f(x)$ is irreducible in $\mathbb{Q}(\sqrt{3}, \zeta)[x],$ we have that $[\mathbb{Q}(\sqrt{3}, \zeta, \alpha) : \mathbb{Q}] = 12.$ We note that $\mathbb{Q}(\alpha) \subseteq \mathbb{Q}(\sqrt{3}, \zeta, \alpha),$ from which we conclude that $[\mathbb{Q}(\alpha) : \mathbb{Q}] \leq 12.$ Getting the lower bound is proving to be more difficult than this. I would like to say that if $f(x)$ is reducible in $\mathbb{Q}(\sqrt{3}, \zeta)[x],$ then $[\mathbb{Q}(\alpha) : \mathbb{Q}] \geq 4,$ but I am not convinced. Can anyone provide a push in the right direction?


If $f$ is reducible in $K=\mathbb{Q}(\zeta,\sqrt{3})$, it either splits completely or it decomposes into a linear and a quadratic factors. In the first case $\alpha\in K$. In the second case, $\alpha$ is either the roots of the linear factor and again $\alpha\in K$, or is a root of the second factor and the extension requires a further quadratic extension of $K$. So, the lower bound is when $\alpha\in K$.

Now, if $\alpha\in K$, we need to show that $\alpha$ is not in an (strict) intermediate extension between $K\supset H\supset \mathbb{Q}$. If that is the case then $[H:\mathbb{Q}]=2$, the only divisor of $4=[K:\mathbb{Q}]$ that is not $1$ or $4$. Then $\alpha$ is a root of a quadratic polynomial with rational coefficients. This polynomial $x^2+ax+b$ must divide $f$. The condition that the remainder is $0$ gives you $\zeta$ and $\sqrt{3}$ expressed as a rational numbers:

$$\frac{x^3+\zeta x+\sqrt{3}}{x^2+ax+b}=(x-a)+\frac{(\zeta-b+a)x+(\sqrt{3}+ab)}{x^2+ax+b}$$

Therefore $$\begin{align}\zeta&=b-a\\\sqrt{3}&=-ab\end{align}$$

which contradicts that $\zeta$ and $\sqrt{3}$ are not rational. Therefore, $\alpha$ cannot be in a field properly contained in $K=\mathbb{Q}(\zeta,\sqrt{3})$.


The minimal polynomial of $\alpha$ turns out to be $x^{12}-2x^{10}+ 3 x^{8}-8x^6+7x^4+3x^2+9$. Therefore, actually $[\mathbb{Q}(\alpha):\mathbb{Q}]=12$

  • $\begingroup$ Excellent. During my original attempt of the problem, I had worked out exactly what you stated in the first paragraph, and I had thought about whether or not $\mathbb{Q}(\alpha) = \mathbb{Q}(\sqrt{3}, \zeta),$ but as you correctly pointed out, we simply need to show that $\alpha$ is not contained in a degree two extension of $\mathbb{Q}.$ Thank you very much for your time. $\endgroup$ – Dylan_Carlo_Beck Aug 2 '18 at 6:06
  • $\begingroup$ Can you elaborate on how you obtain the minimal polynomial? I am fascinated. $\endgroup$ – Dylan_Carlo_Beck Aug 2 '18 at 6:08
  • $\begingroup$ @Dylan_Carlo_Beck That was done by a computer. $\zeta$ and $\sqrt{3}$ are expressed in radicals, and so can be the roots of a cubic. So, you can compute the three values of $\alpha$. Then the minimal polynomial can also be found algorithmically by searching for rational combinations between its powers. No something worth doing by hand. $\endgroup$ – user580373 Aug 2 '18 at 10:30
  • $\begingroup$ I just typed up my solution to the problem, and I noticed that we should have $\zeta = b - a^2.$ Of course, the contradiction is reached all the same, but I just figured I'd point that out. $\endgroup$ – Dylan_Carlo_Beck Aug 2 '18 at 18:35

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.