A splitting field is a Galois extension

77 Views Asked by At

Please tell me if my logic is correct.

I know that the Galois group divides the order of the extension and will use this fact as given.

If $K$ is the splitting field for irreducible $f(x)$ over $F$, there are $n$ roots for $f$ in $K$. Then our Galois group could be made out of isomorphism that keep $F$ fixed by sensing roots to roots. So for instance if $a_1,a_2...a_3$ are roots of $f$, our first automorphism could send $a_1$ to $a_2$ and this gives us $n!$ as our group order at the very least. And $K:F$ is a extension of less than $n!$ so order Galois group is equal to the order of extension and we are done.

My book wrote 2 pages for this proof and this makes me feel like my proof is missing something/wrong Is this wrong? Please help. (I am new to this area so please forgive me)

1

There are 1 best solutions below

0
On BEST ANSWER

The elements of the Galois group is not merely "permutations of roots". They are automorphisms of the field.


To give an example: let $F$ be the rational field $\Bbb Q$, and let $f$ be the polynomial $x^4 - 10 x^2 + 1$.

Then $f$ has $4$ roots: $\sqrt 2 + \sqrt 3$, $\sqrt 2 - \sqrt 3$, $- \sqrt 2 + \sqrt 3$, $- \sqrt 2 - \sqrt 3$.

Not all the $24$ permutations of these $4$ roots are automorphisms of the splitting field. To see this, suppose that a field automorphism sends $\sqrt 2 + \sqrt 3$ to $\sqrt 2 - \sqrt 3$, then it must send $-\sqrt 2 - \sqrt 3 $ to $-\sqrt 2 + \sqrt 3$.

In other words, you don't have "total freedom" to permute these roots.

Does this answer your question?