Well, my question is essentially:

Let $R$ be a Factorial Ring (UFD, basically) and let $p$ be a prime element in $R$. Let $d$ be an integer larger than 2, and let

$f(t) = t^d + c_{d-1} t^{d-1} + ... + c_0$ be a polynomial belonging to $R[t]$. Let $n \ge 1$ be an integer.

Show that $g(t) = f(t) + \displaystyle \frac{p}{p^{nd}}$ is irreducible in $K[t]$, where $K$ is the quotient field of $R$.

Okay, well, I know that it is enough, from Gauss' Lemma, to show that (after multiplying through by $p^{nd}$) $g$ has no non-trivial factorizations in $R$ itself. EDIT: My apologies, I forgot that this holds specifically for primitive polynomials, and no such specification was given. Nevertheless, perhaps, I suppose the gcd of the coefficients could be factored out?

I also tried thinking of $g$ as essentially just a translation of the graph of $f$ slightly upward, and that $g$ converges to $f$ as $n \to \infty$, but I don't know exactly what to make of that, because, well, as $n$ gets larger, it starts to inch towards the roots of $f$, and would no longer be irreducible, but perhaps I've confused myself.

(As an auxilliary question, there was a previous exercise which asked us to prove that if $R = \mathbb{Z},$ then, if $f$ had $m$ real roots, then, "for sufficiently large $n$", $g$ also has $m$ EDIT: real roots. This should be fairly apparent, I should think, from the graph? But if not, then how would one rigorously establish this?)

Anyway, thanks a lot.


As $p$ is a unit in $K$, the irreducibility of $g(t)$ is equivalent to the irreduciblity of $h(t):=p^{nd}g(t)$. Write $$h(t)=(p^nt)^{d}+c_{d-1}p^n (p^nt)^{d-1}+\cdots + c_1p^{(d-1)n}(p^nt)+p(1+p^{nd-1}c_0).$$ Now apply Eisenstein criterion to the polynomial $$ H(T)=T^d+c_{d-1}p^n T^{d-1}+\cdots + c_1p^{(d-1)n}T+p(1+p^{nd-1}c_0)$$ and Gauss lemma to conclude that $H(T)\in K[T]$ is irreducible. As $h(t)$ is obtained from $H(T)$ by a change of variables, it is also irreducible.

Edit Forgot the auxillary question. It is false: consider $f(t)=t^2$ and $p$ a prime number.

  • $\begingroup$ Lovely. How did I not think of that!? $\endgroup$ – AlpArslan Sep 10 '13 at 14:55
  • $\begingroup$ @MWarsi: the second question has a positive answer for simple real roots. $\endgroup$ – Cantlog Sep 10 '13 at 15:53

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.