# On the existence of field morphisms

Let $K$ and $L$ be two fields, does the existence of two field morphisms $f\colon K\rightarrow L,\ g\colon L\rightarrow K$ imply that, as abstract fields, $K\cong L$ (not necessarily via $f$ or $g$)?

• One can rephrase the question as 'does $F\le K\le L$ and $F\cong L$ implies $K\cong L$?' Jun 10, 2015 at 9:37
• What if we choose $f\equiv g\equiv 0$? Jun 10, 2015 at 9:40
• $0$ is not a field homomorphism unless the codomain is the trivial field with $0=1$. Jun 10, 2015 at 9:41
• @Berci Which isn't a field anyways. Jun 10, 2015 at 10:05
• @Hayden: Well, that's only a matter of taste / definition.. similar to the question whether $0$ is natural or not. Jun 10, 2015 at 10:59

I think I found a counterexample:

Let $F:=\overline{\Bbb Q(x_1,x_2,\dots)}$, i.e. the algebraic closure of the extension of $\Bbb Q$ by infinitely many independent transcendent elements.

Let $K:=F(x_0)$ a simple transcendent extension. This field is not algebraically closed.

Finally, let $L:=\overline K$, its algebraic closure.

Unless I am mistaken, $L=\overline{F(x_0)}=\overline{\Bbb Q(x_0,x_1,x_2,\dots)}\cong F$.

Thus we gain field morphisms $K\to L$ and $L\cong F\to K$, though $L\not\cong K$.

• This is a very nice example, thank you! Jun 10, 2015 at 13:48