9
$\begingroup$

I have this problem in a exercise list:

"Prove that $K=\mathbb{Q}(2^{\frac{1}{2}},2^{\frac{1}{3}}, 2^{\frac{1}{4}}, \ldots)$ is an algebraic extension, but not a finite extension of $\mathbb{Q}$."

What I did:

Let $\alpha \in K$, then $\alpha = a_0 + a_12^{\frac{1}{n_1}}+a_22^{\frac{1}{n_2}}+\cdots+a_m2^{\frac{1}{n_m}}$.

Now, since each factor of the sum above is algebraic over $\mathbb{Q}$, it follows that $\alpha$ is indeed algebraic over $\mathbb{Q}$ (because the set of algebraic numbers is a field).

Suppose now that $K$ is a finite extension of $\mathbb{Q}$. Then, by Steinitz's theorem, there is $u \in K$ such that $K=\mathbb{Q}(u)$. Let $p(x)$ be the minimal polynomial of $u$, and let $n$ be $p$'s degree. Then, K is an extension os degree $n$. However, the minimal polynomial of $2^{\frac{1}{n+1}}$ has degree $n+1$, so K is an extension of degree at least $n+1$, contradicting the hypothesis that $K$ is an extension of degree $n$. So, $K$ is an infinite extension of $\mathbb{Q}$.

My doubts: Can $\alpha$ be an infinite sum? If yes, can I use the same argument?

$\endgroup$
  • $\begingroup$ No, it cant.... $\endgroup$ – Gaston Burrull Sep 10 '14 at 3:38
  • $\begingroup$ No, unless you add some topology in which the sequence converges. $\endgroup$ – user40276 Sep 10 '14 at 3:41
  • $\begingroup$ @GastónBurrull, so is what I did correct? Or did I make some mistake somewhere? (: $\endgroup$ – Anna Sep 10 '14 at 3:42
  • 1
    $\begingroup$ Yes but Steinitz's theorem is unnecesary, if the extension is finite of degree $n$, by tower law you have that $k|n$ for every $k\in\mathbb{N}$ a contradiction. $\endgroup$ – Gaston Burrull Sep 10 '14 at 3:45
  • $\begingroup$ @GastónBurrull what would be this k? The degree of the extension $\mathbb{Q}(2^{\frac{1}{k}})$? $\endgroup$ – Anna Sep 10 '14 at 3:47
10
$\begingroup$

As has been mentioned, the result is a direct consequence of the fact that $[\mathbf{Q}(2^{1/n}):\mathbf{Q}]$ = n. This is because we then have, for every $n$, the inequality $$n = [\mathbf{Q}(2^{1/n}):\mathbf{Q}] \leq [K:\mathbf{Q}].$$ Since $[K:\mathbf{Q}]$ is larger than every natural number, it must be infinite.

The trickiest point is why $2^{1/n}$ has degree $n$ over $\mathbf{Q}$. For this it is enough to establish that the polynomial $X^n - 2$ is irreducible. This follows immediately from Eisenstein's criterion for the prime $p = 2$.

There is an error in your proof that every element of $K$ is algebraic over $\mathbf{Q}$. It is not clear at all why you claim any $\alpha \in K$ can be written in the form of a linear combination of various roots of $2$. The only thing that is obvious is that $\alpha$ can be written as a polynomial expression of these numbers, not a linear combination.

The proof that $K$ is algebraic is as follows. The field $K$ is generated by the set $S = \mathbf{Q} \cup \{2^{1/2}, 2^{1/3}, \dots\} \subseteq \overline{\mathbf{Q}}$, where $\overline{\mathbf{Q}}$ is the field of algebraic numbers over $\mathbf{Q}$. It follows that $K$ itself is contained in $\overline{\mathbf{Q}}$. Thus $K$ is algebraic over $\mathbf{Q}$.

$\endgroup$

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.