Show that $\mathbb{Q}(\sqrt{3},\sqrt[4]{3}, \sqrt[8]{3},...)$ is algebraic over $\mathbb{Q}$ but not a finite extension.

I think for the algebraic part, since for every simple extension, each of those elements can be adjoined and each of these simple extensions has a minimal polynomial that cannot be reduced in $\mathbb{Q}$. For example, the simple extension $\mathbb{Q}(\sqrt{3}, \sqrt[4]{3})(\sqrt[8]{3})$ has minimal polynomial $x^{8}-3$. And since each simple extension has an increasingly large degree, the degree of the simple extensions over the previous extension gets larger for each attachment. But I am not sure how to express this formally...

For the infinite degree part, I was thinking because the set $\left \{\sqrt{3},\sqrt[4]{3}, \sqrt[8]{3},...\right \}$ is linearly independent?


Any element $\alpha$ of $F$ is a rational expression in the numbers adjoined. As such, it can involve only finitely many of the $\sqrt[2^k]3$. If in such an expression, $\sqrt[2^n]3$ is one with maximal $k$, then all other $\sqrt[2^k]3$ are powers of $\sqrt[2 k]3$. It follows that $\alpha\in\Bbb Q(\sqrt[2^n]3)$ and $\alpha $ is algebraic.

A different approach for the infinity part: By Eisenstein, the polynomial $X^{2^n}-3$ is irreducible. Hence $[F:\Bbb Q]\ge[\Bbb Q(\sqrt[2^n]3):\Bbb Q]\ge 2^n$, where $n$ is arbitrariy. It follows that $[F:\Bbb Q]$ is inifnite.

  • $\begingroup$ Thanks for the insight! I didn't realize I could use Eisenstein's criterion. $\endgroup$ – numericalorange Dec 7 '18 at 18:07

Let $L = \mathbb Q(3^{\frac 1{2^n}})$.

Why is $L$ algebraic over $\mathbb Q$? It is because the generating set of $L$ is $\mathbb Q \cup \{3^{\frac 1{2^n}}\}$, which are all contained in $\bar{\mathbb Q}$, the algebraic closure of $\mathbb Q$. Therefore, $L \subset \bar {\mathbb Q}$, which by definition of the algebraic closure implies it is algebraic over $\mathbb Q$.

However, $L$ contains subfields of the form $K_n$, where $K_n = \mathbb Q(3^{\frac 1{2^n}})$. One can conclude that $[K_n : \mathbb Q] = 2^n$, since $x^{2^n} - 3$ is irreducible via the Eisenstein criterion(shifted).

Then, suppose $L$ were finite, then $[L : \mathbb Q] < \infty $, but by the tower property it is equal to $[L : K_n][K_n : \mathbb Q]$ for each $n$. In particular, $[L : \mathbb Q]$ is a multiple of $2^n$ for all $n$, clearly impossible if it is finite.


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.