Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Suppose $F/L$, $F'/L$, $L/K$ finite extensions of fields. If $F$, $F'$ isomorphic over $K$ then does it follow that they are isomorphic over $L$? I think probably not, but I can't come up with a counterexample. I've tried thinking about splitting fields of quartics or about function fields and neither has given me any joy. Obviously finite fields are no good. Could someone give me a hint? Thanks!

share|improve this question
    
You should be able to come up with an example with $K = \mathbb{Q}$. What kind of things have you tried for $L$? –  Qiaochu Yuan May 18 '12 at 15:36
    
I've tried quadratic and cubic extensions, but can't see anything that works... –  Edward Hughes May 18 '12 at 15:41
    
Well, here's kind of a big hint. I don't claim that this is the only way to do the problem. Choose $L$ so that it embeds into $\mathbb{R}$, and choose $F, F'$ so that if you fix an embedding of $L$ into $\mathbb{R}$, one of $F, F'$ is also embeddable into $\mathbb{R}$ and the other isn't. –  Qiaochu Yuan May 18 '12 at 15:43
1  
Is it $L=\mathbb{Q}(\sqrt{2})$,$F=\mathbb{Q}(\sqrt[4]{2})$, $F'=\mathbb{Q}(i\sqrt[4]{2})$? –  Edward Hughes May 18 '12 at 16:01
3  
@hgbreton: yep! –  Qiaochu Yuan May 18 '12 at 16:05

2 Answers 2

up vote 6 down vote accepted

This is indeed false. An easy counterexample is $L=\mathbb{Q}(\sqrt{2})$, $F=\mathbb{Q}(\sqrt[4]{2})$, $F'=\mathbb{Q}(i\sqrt[4]{2})$.

Clearly $F$, $F'$ isomorphic over $\mathbb{Q}$ by $\sqrt[4]{2}\mapsto i\sqrt[4]{2}$. Now suppose that $F$, $F'$ isomorphic over $L$ by $\tau$ say. Then $\tau(\sqrt[4]{2})=a+ib\sqrt[4]{2}$ some $a,b\in L$. We must have

$\sqrt{2}=\tau(\sqrt[4]{2})^2=(a+ib\sqrt[4]{2})^2=a^2-2ab\sqrt[4]{2}-b^2\sqrt{2}$

This forces $a=0$ so $b^2=-1$ and contradiction.

share|improve this answer
    
Alternatively, $\tau$ must map a root of $X^4 - 2$ to another root, so it pretty much has to send $\sqrt[4]{2}$ to $\pm i\sqrt[4]{2}$, but then $\sqrt{2}$ maps to $-\sqrt{2}$ so we can't fix $L$. (PS. 2011 paper 1 question 18H, right? :P) –  Ben Millwood Jun 2 '12 at 17:48
    
Yep - that's right! –  Edward Hughes Jun 2 '12 at 18:29

Both $F = \Bbb{Q}(\sqrt[4]{2})$ and $F' = \Bbb{Q}(i\sqrt[4]{2})$ are isomorphic to $\Bbb{Q}[x]/(x^4 - 2)$, hence $\Bbb{Q}$-isomorphic, ($x^4 - 2$ is irreducible over $\Bbb{Q}$, so these are field extensions. The actual isomorphism is given by $\sqrt[4]{2} \mapsto i\sqrt[4]{2}$). However, they are not isomorphic over $L = \Bbb{Q}(\sqrt{2})$, as $x^2 + \sqrt{2} \in L[x]$ splits in $F'$, but not in $F$.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.