2

Let E/K be a field extension, $1+1\neq0$ in K and $\alpha$, $\beta \in K^x=\{a \in K |\, \exists \, b: ab = ba =1\}$.

Proof with Galois theory: $K(\sqrt \alpha) = K(\sqrt \beta) \Leftrightarrow \exists \, \gamma \in K^x : \alpha = \gamma^2\beta$


Got so far that $\sqrt \alpha, \sqrt \beta \notin K \Rightarrow [K(\sqrt \alpha):K]= 2 =[K(\sqrt \beta):K] \overset{char(K) \neq 2}\Rightarrow K(\sqrt \alpha), K(\sqrt \beta)$ are galois.

Zorro_C
  • 43

1 Answers1

2

$\implies$: Assuming $K(\sqrt\alpha) = K(\sqrt\beta)$.

If $\sqrt\alpha = x\in K$, then $K(\sqrt\alpha) = K$, so we have $\sqrt\beta\in K$. Set $\gamma = \frac{\sqrt\alpha}{\sqrt\beta}$ and we're done.

Otherwise, we have $\sqrt\alpha\in K(\sqrt\beta)$. In other words, there are $x, y\in K$ such that $$ \sqrt\alpha = x + y\sqrt\beta\\ \alpha = x^2 + \beta y^2 + 2xy\sqrt\beta $$ From this we can deduce that $2xy = 0$, which because $1+1\neq 0$ gives us $xy = 0$, meaning either $x = 0$ or $y = 0$. However, $\sqrt\alpha\notin K$ implies $y\neq 0$, so we must have $x = 0$. Set $\gamma = y$, and we're done.

$\Longleftarrow\,$: Assuming $\alpha = \gamma^2\beta$.

We have $\sqrt\alpha = \gamma\sqrt\beta \in K(\sqrt\beta)$ and $\sqrt\beta = \sqrt\alpha/\gamma\in K(\sqrt\alpha)$, meaning the two fields must be the same.

Arthur
  • 199,419
  • 1
    The solution is absolutely right and I've solved the problem already that way. But I wanted to come up with a solution in which Galois theory would be applied. – Zorro_C Dec 13 '18 at 10:24
  • @Zorro_C If you've done this already, that's the kind of thing you would want to put in the question post so that people don't spend time and effort duplicating your work. – Arthur Dec 13 '18 at 10:28
  • Sorry, but I have mentioned in every line, also in the title, that I like to get a solution with application of Galois theory – Zorro_C Dec 13 '18 at 10:39
  • What do you mean by "application of Galois theory".? If you admit that Kummer theory (see any text book) is contained in Galois theory, then you can apply it directly. For a summary of Kummer theory, see e.g. https://math.stackexchange.com/a/1609061/300700 – nguyen quang do Dec 13 '18 at 17:33