3

If $f\in\mathbb Z[X]$ and there are $x_1,\dots, x_n\in\mathbb Z_+$, where $n>\deg f$, such that $f(x_i)\in\mathbb P$, $i=1,\dots n$, then $f$ is irreducible over $\mathbb Z$. Because, if $\,f=g\cdot h$ then either $g(x_i)=1$ or $h(x_i)=1$ which can't happen $n>\deg g+\deg h$ times.

But what about the opposite? Are there non constant, irreducible polynomials $p\in\mathbb Z[X]$ such that

$x_1,\dots, x_n\in\mathbb Z_+$ and $p(x_i)\in\mathbb P$ for all $i=1,\dots,n$ $\:\implies\:$ $n\leq \deg p$?

Lehs
  • 13,791
  • 4
  • 25
  • 77
  • 1
    I think this question is pleny interesting even without the $n \leq \deg p$ limitation. Just that there are only finitely many inputs that yield primes. – Arthur Dec 17 '18 at 09:13
  • @Arthur: the limitation would be nice for an eventual computational test method. Else I agree. – Lehs Dec 17 '18 at 09:18
  • 1
    See also https://math.stackexchange.com/a/2942032/589. – lhf Dec 17 '18 at 10:55

1 Answers1

3

Your claim seems to be negative.

For example, if we take $f(x)=x^3-x+3$. It is an irreducible polynomial of degree $3$, and for each positive integer $n$, $f(n)$ is divided by $3$ ( By Fermat's little theorem). Since $f(n)>3$ when $n>2$, the only $n$ such that $f(n)\in \mathbb{P}$ is just $1$.

Bonbon
  • 891
  • 1
    I'm not sure, but it seems that the irreducibility of $f$ can be deduced from the method in https://en.wikipedia.org/wiki/Irreducible_polynomial#Over_the_integers as $f$ is irreducible over $\mathbb Z_2$. Is that correct? – Lehs Dec 17 '18 at 11:17
  • Yes, you are correct! – Bonbon Dec 17 '18 at 11:24
  • Thanks a lot! Now I know how to proceed from your example. – Lehs Dec 17 '18 at 11:26