I'm studying chapter 7.6 Splitting field and Algebraic closure in Abstract Algebra by S. Lovett.

The above is one of its exercises and I'm not sure that I can prove it only with the materials that I have read.

I know that the above exercise is easy if we consider transcendence degree, but we didn't learn anything about it.

Here is one of my attempts.

Suppose that such $F$ exists.

Then there exists a transcendental element $\beta \in F$ over $\mathbb{Q}$, and $\mathbb{Q}(\beta) \subseteq F$.

Now we have an isomorphism $\mathbb{Q}(\pi) \rightarrow \mathbb{Q}(\beta)$.

Considering their algebraic closures, we have $\overline{\mathbb{Q}(\pi)}$ an algebraic closure of $\mathbb{Q}(\pi)$ and an embedding $f:\mathbb{Q}(\pi) \rightarrow \overline{\mathbb{Q}(\beta)}$ given by the above isomorphism.

So there exists an embedding $\lambda : \overline{\mathbb{Q}(\pi)} \rightarrow \overline{\mathbb{Q}(\beta)}$ extending $f$.

Since $\overline{\mathbb{Q}(\pi)}$ is algebraically closed and $\overline{\mathbb{Q}(\beta)}$ is algebraic over $f(\mathbb{Q}(\pi))={\mathbb{Q}(\beta)}$,

$\lambda (\overline{\mathbb{Q}(\pi)})$ is algebraically closed and $\overline{\mathbb{Q}(\beta)}$ is algebaic over $\lambda (\bar{\mathbb{Q}(\pi)})$.

So as an algebraic extension of an algebraically closed field $\overline{\mathbb{Q}(\beta)}=\lambda (\overline{\mathbb{Q}(\pi)})$.

Therefore $\lambda$ is actually an isomorphism.

I once thought that I have found the reverse inclusion $\overline{\mathbb{Q}(\pi)} \subseteq \bar{F}=F$, but I just found that $\overline{\mathbb{Q}(\beta)}$ and $\overline{\mathbb{Q}(\pi)}$ are isomorphic, not equal, so it does not mean anything.

I know that not all of you can access to the textbook that I have, so I just want to learn various ways to prove this without using transcendence degree, or a proof that is accessible for an undergraduate student who just started to learn about field extensions.

  • $\begingroup$ I think you need to show that $\overline{\mathbb{Q}(\pi)}$ isn't isomorphic to $k \subset \overline{\mathbb{Q}(\pi)}$ $\endgroup$ – reuns Nov 5 '16 at 11:29

What you are doing (showing $\overline{\Bbb Q(\pi)}$ is isomorphic to a subfield of $F$) is not what the question is asking.

Suppose $\overline {\Bbb Q} \subset F \subset \overline{\Bbb Q(\pi)}$. You want to show that $\overline{\Bbb Q(\pi)} \subset F$ or $F \subset \overline {\Bbb Q}$. Since $F$ is algebraically closed, to show $\overline{\Bbb Q(\pi)} \subset F$ it is enough to show $\pi \in F$.

Suppose $F \neq \overline{\Bbb Q}$. Then there is an element $x \in \overline{\Bbb Q(\pi)}$ such that $x \in F$ and $x \notin \overline {\Bbb Q}$. And since $F$ might as well just be $\overline{\Bbb Q(x)}$, to show that $\pi \in F$ is the same as showing that $\pi \in \overline{\Bbb Q(x)}$.

However, given that $\pi$ and $x$ aren't in $\overline {\Bbb Q}$, $\pi \in \overline{\Bbb Q(x)}$ and $x \in \overline {\Bbb Q(\pi)}$ are both equivalent to the existence of a polynomial $P$ in two variables with rational coefficients such that $P(x,\pi)= 0$, and so they are equivalent.


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.