Let $f(x) \in F[x]$ be an irreducible separable polynomial of degree p, with distinct roots $\alpha_1, \ldots, \alpha_p$. I want to show that if $F(\alpha_1) = F(\alpha_1, \alpha_2)$, then $F(\alpha_1) = F(\alpha_1, \ldots, \alpha_p)$, ie. $F(\alpha_1)$ is the splitting field of $f(x)$.

So far, what I know is that $F(\alpha_1) \simeq F(\alpha_i)$ for all $1 \leq i \leq p$, can I somehow extend this to show that $(F(\alpha_1))(\alpha_1) \simeq (F(\alpha_1))(\alpha_i)$? Thanks

  • $\begingroup$ The title is not quite consistent with the question $\endgroup$ – lhf Dec 1 '17 at 1:38

Let $G = \operatorname{Gal}(f) \subset S_p$ be the galois group of the splitting field. By a standard argument, $G$ contains a cycle of length $p$ (Since any element of order $p$ in $S_p$ is such a cycle).

Let $\sigma \in G$ be a cycle of length $p$. Replacing it by a suitable power, we can assume $\sigma(\alpha_1) = \alpha_2$.

Let $L = F(\alpha_1)$. In particular $\sigma(L) \subset L$ holds, by iterating this we get that $\sigma^n(L) \subset L$ holds for any $n \geq 1$. But since $\sigma$ is a cycle of length $p$, we get that all the $\alpha_i$ occur among $\sigma^n(\alpha_1), n \geq 1$. Hence $L$ contains all roots and is thus the splitting field.

Some more details on the "standard argument":

If $K$ is the splitting field, we have $F \subset F(\alpha_1) \subset K$, hence $$|G| = [K:F] = [K:F(\alpha_1)] \cdot [F(\alpha_1):F] = [K:F(\alpha_1)] \cdot p$$ is a multiple of $p$. By Cauchy's Theorem, $G$ has an element of order $p$.

  • $\begingroup$ Thanks. Since posting the question, I've actually figured out all the above except for the "By a standard argument" part. How can we be sure that $G$, which (turns out to be) a strict subgroup of $S_p$, contains an element of order $p$? I've thought about this for a long time and it is not obvious to me. $\endgroup$ – shmth Dec 3 '17 at 12:37
  • 1
    $\begingroup$ This is Cauchy's Theorem. The order of the group is divisible by $p$, hence there is an element of order $p$. $\endgroup$ – MooS Dec 3 '17 at 12:49
  • $\begingroup$ @ihmth I have added some details in my answer. $\endgroup$ – MooS Dec 3 '17 at 13:58

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.