0

I'm struggling to make any progress on the following problem.

Let $K = \mathbb{Q}(\alpha)$ be an algebraic extension of $\mathbb{Q}$ with $[\mathbb{Q}(\alpha):\mathbb{Q}]=m$, and suppose that $f \in \mathbb{Q}[x]$ is irreducible.

a. Suppose that $m=2$. Prove that either $f \in K[x]$ is still irreducible, or that $f=gh$ for $1 < \deg(g),\deg(h)$ and $\deg(g)=\deg(h)$.

b. Suppose now that $m > 2$. Does it follow that $f \in K[x] $ remains irreducible or factors into $m$ irreducible polynomials?

So I believe the answer to both is yes, but I'm not really sure how. For part a, I start by supposing that $f(x) = p_1(x)p_2(x)\dots p_r(x)$ for $p_i \in K[x]$. I want to show that the only possibility is that $r=2$, and that $\deg(p_1)=\deg(p_2)$. I'm aware of the identification $K \cong \mathbb{Q}[x] / \langle \text{Irr}_{\mathbb{Q}}(\alpha,x) \rangle$, and so one can think of the coefficients on each $p_i$ as being degree $1$ polynomials in $\mathbb{Q}[x] / \langle \text{Irr}_{\mathbb{Q}}(\alpha,x)\rangle$, but I have no idea what else to try.

For part b, I'm assuming that If one figures out part a properly, then part b follows by some basic inductive argument.

Any help would be great. Thanks!

Isochron
  • 1,870
  • 1
    Check out for example this for possible ideas in 1) how to prove part (a) and 2) find a counterexample to part (b). – Jyrki Lahtonen May 01 '23 at 04:54
  • 1
    Mind you that thread is not an ideal duplicate target for this. Some others may serve better. Combined with the coincidence that I happened to answer that version I'm reluctant to vote to close. – Jyrki Lahtonen May 01 '23 at 04:56
  • 1
    Sorting my search results by scores lead to this older variant. – Jyrki Lahtonen May 01 '23 at 04:58
  • 1
    Anyway, what's the simplest non-Galois extension of $\Bbb{Q}$ you have seen? Use that to construct a counterexample to (b). – Jyrki Lahtonen May 01 '23 at 04:59
  • @JyrkiLahtonen So part b was indeed easy to figure out, however I'm still having trouble with proving part a. I see how you did a similar problem from the link you sent, but I was told in order to solve the problem, I should use that each $\sigma \in \text{Gal}(E_K)$ (for $E$ a splitting field of $f$ over $\mathbb{Q}$) fixes each $p_i$, but permutes its roots, and then to consider now what elements of $\text{Gal}(E_{/\mathbb{Q}})$, but I still don't see what I'm supposed to do. The only thing I know is that $\text{Gal}(K/\mathbb{Q})$ only has $1$ nontrivial automorphism. – Isochron May 04 '23 at 02:32
  • I also know that the one nontrivial automorphism $\phi$ in $\text{Gal}(K_{/\mathbb{Q}})$ can't fix every $p_i$, and I can extend $\phi$ to all of $E$, but I again don't see how to proceed. – Isochron May 04 '23 at 02:38
  • I have found an answer: https://math.stackexchange.com/questions/3254652/let-alpha-be-algebraic-over-mathbbq-with-mathbbq-alpha-mathbb?rq=1.

    Free bounty for anyone who wants it.

    – Isochron May 04 '23 at 02:56

0 Answers0