The conjecture about Power free-values of polynomials is: Let $F(X)$ be a polynomial with integer coefficients and no repeated roots. For any $\epsilon > 0$, there exists a constant $C_{\epsilon,F}$ such that for any integer $n$ $$ |n|^{\deg{(F)}-1-\epsilon} \le C_{\epsilon,F} \operatorname{rad}(F(n))$$

The conjecture implies this polynomial version. For $f(x) , g(x) \in \mathbb{Z}[x]$ and $f(x)$ squarefree,

$$\deg (\operatorname{rad}(f(g(x)))) > \deg(g(x)) (\deg(f(x))-1) \qquad (1)$$

The bound is tight because for Chebyshev polynomials $T_n,U_n$, $T_n(x)^2 - 1 = (x^2-1) U_{n-1}^2(x)$ with $f(x)=x^2-1$.

Is (1) proved for polynomials?

  • 1
    $\begingroup$ The very web page you link to tells you that this conjecture would follow from abc, and abc is known over polynomials (where it is called the Mason-Stothers theorem.) $\endgroup$
    – JSE
    Dec 11, 2012 at 21:20

1 Answer 1


Haven't seen that anywhere, but it's easy to prove: Write $f(g(X))=\prod(X-\beta_i)^{e_i}$ for distinct complex $\beta_i$, and $e_i\ge1$. Taking the derivative gives $f'(g(X))g'(X)=h(X)\prod(X-\beta_i)^{e_i-1}$ for some polynomial $h(X)$. Now $f(X)$ and $f'(X)$ are relatively prime, hence so are $f(g(X))$ and $f'(g(X))$. Thus $\prod(X-\beta_i)^{e_i-1}$ divides $g'(X)$, implying $$\text{deg}(f(g(X))-\text{deg(rad}(f(g(X)))\le\text{deg}(g'(X))=\text{deg}(g(X))-1,$$ and the claim follows.

  • $\begingroup$ "hence so (realtively prime) are $f(g(X))$ and $f'(g(X))$". Isn't $f(x)=x, g(x)=x^2$ a counterexample? $f(g(x))=x^2$ which is certainly not coprime to the derivative. There is no restriction on $g(x)$, it need not be squarefree. There is an infinite family of counterexamples to this. $\endgroup$
    – joro
    Dec 12, 2012 at 7:24
  • $\begingroup$ I didn't say that $f(g(X))$ and its derivative are relatively prime. Rather I say and use that $f(g(X))$ and $f'(g(X))$ are relatively prime, which is trivial: Let $\alpha$ be a common complex root of $f(g(X))$ and $f'(g(X))$. Then $g(\alpha)$ is a common root of $f(X)$ and $f'(X)$, contrary to $f(X)$ and $f'(X)$ being relatively prime. (Remark: I fixed a typo in the answer regarding the derivative of $f(g(X))$ which is $f'(g(X))g'(X)$.) $\endgroup$ Dec 12, 2012 at 8:29
  • $\begingroup$ I misunderstood, sorry. btw, (1) fails for $g(x)$ constant giving 0=0. $\endgroup$
    – joro
    Dec 12, 2012 at 8:44
  • $\begingroup$ Yes, it is unclear why you don't need $g(x)$ to be non-constant since for constant $g(x)$ my strict inequality doesn't hold (it becomes equality). $\endgroup$
    – joro
    Dec 12, 2012 at 12:02
  • 1
    $\begingroup$ My proof uses the fact if a polynomial $w(X)$ divides $g'(X)$, then $\text{deg}(w(X))\le\text{deg}(g'(X))$, a trivial fact which holds unless $g'(X)=0$, in which case $g(X)$ is a constant. But in this degenerate case, your inequality does not hold, so of course my argument cannot work then either. $\endgroup$ Dec 12, 2012 at 12:26

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.