Continuation to this previous question.

According to Lemke-Oliver, an irreducible polynomial $G$ of degree $g$ with positive leading coefficient and $\Gamma_G\neq0$ (with $\Gamma_G$ a certain factor defined in the reference) is conjectured to represent primes with density $\Gamma_G/\log x$. (What is the name of this conjecture? Is it Bouniakowsky? Or Hardy-Littlewood's F?).

This conjecture is wide-open. In fact, according to MathWorld, it seems that it is not even known whether the density is non-zero: such polynomials, while expected to represent primes infinitely often, are not proven to represent primes at least once.

If we strengthen the hypotheses, such as $g\equiv2$ (and, say, negative discriminant), is there any known result concerning the density of primes? Is it known whether quadratic polynomials of the form above represent primes at least once?

  • 1
    $\begingroup$ No. Typically we expect that if we can show a polynomial that doesn't obviously represent a prime represents a prime at all then it represents infinitely many primes. At least, that's a corollary of all successful proofs so far. $\endgroup$ – Stanley Yao Xiao Nov 13 '18 at 16:49
  • $\begingroup$ @StanleyYaoXiao Thank you for your comment, but I must obviously missing something. "It represents a prime" cannot be equivalent to "it represents infinitely many primes". For one thing, $x^2+1$ clearly represents $2^2+1=5$, one prime, but it is not known if it represents infinitely many. So "representing one prime" is a much weaker statement than "it represents infinitely many". Or did I misunderstand your comment? $\endgroup$ – Delmastro Nov 13 '18 at 17:48
  • $\begingroup$ $x^2 + 1$ obviously represents 5, because you can find an explicit $x$ such that $x^2 + 1$ is equal to 5. This is not doable for a generic quadratic polynomial, say $ax^2 + bx + c$. So outside of obvious 'eyeballing', one can't prove that a quadratic polynomial represents a prime at all. If one can in fact show that a generic quadratic polynomial represents a prime, then very likely the argument will in fact produce infinitely many primes that it can represent. $\endgroup$ – Stanley Yao Xiao Nov 13 '18 at 17:54
  • 2
    $\begingroup$ To say the same thing another way: the strategies that number theorists have come up with for trying to show that a given polynomial must represent at least one prime (other than brute-force computation) are all actually strategies for showing that it must represent infinitely many primes. It seems hard to think of a theoretical approach to proving "one prime" that doesn't also prove "infinitely many primes". $\endgroup$ – Greg Martin Nov 13 '18 at 18:26
  • $\begingroup$ @StanleyYaoXiao Oh, I now I see what you meant. Kind of anticlimactic, but I guess that's the way it goes. Anyway, thank you very much for your insight! $\endgroup$ – Delmastro Nov 13 '18 at 22:18

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.