Is there an analytic function with zeroes only at:

  • every $-2n$,
  • $\frac12\pm it$, and
  • at least one at $\frac12\pm\epsilon\pm it$ where $0<\epsilon<\frac12, t\neq0$ (and these zeroes observing the known reflective symmetries within the critical strip)?

A answer assuming the Riemann hypothesis true would be fine.

To my mind the uniqueness of any analytic continuation of $\zeta(s)$ is suggestive of the existence of such a function being incompatible with the Riemann Hypothesis.

If not, uniqueness of $\zeta$ is of course a nice simple sufficiency for the Riemann hypothesis.


Take $\zeta(z)\left(z-\frac14-i\right)\left(z-\frac14+i\right)\left(z-\frac34-i\right)\left(z-\frac34+i\right)$.

Every condition that you mentioned holds.

  • $\begingroup$ @RobertFrost Can you please explain that? I have no doubt that the zeros of my function are the zeros of the $\zeta$ functions plus $\frac14$. $\endgroup$ – José Carlos Santos Sep 5 '18 at 9:34
  • $\begingroup$ @RobertFrost I've edited my answer. What do you think now? $\endgroup$ – José Carlos Santos Sep 5 '18 at 10:28
  • $\begingroup$ @RobertFrost What about$$\zeta(z)\left(z-\frac14-i\right)\left(z-\frac14+i\right)\left(z-\frac34-i\right)\left(z-\frac34+i\right)?$$But note that the extra zeros of my previous answer ($\frac14$ and $\frac34$) were already symmetric with respect to the real axis. $\endgroup$ – José Carlos Santos Sep 5 '18 at 10:36
  • $\begingroup$ Yes, I saw they were already symmetric; I should have required they have a distinct reflection or $t\neq0$. If you put this in your answer I'll accept, thank-you for your help... I may inquire in another question to what degree we can generalise these examples. If your example is $\zeta(z)g(a+bi)$ then I am immediately struck by the question; suppose $a+bi$ is a counterexample to RH then what do $\zeta(z)g(a+bi)$ and $\zeta(z)/g(a+bi)$ look like? $\endgroup$ – user334732 Sep 5 '18 at 11:05
  • $\begingroup$ @RobertFrost Perhaps that you could post that question as a separate question. And if my answer was useful, perhaps that you could mark it as the accepted one. $\endgroup$ – José Carlos Santos Sep 5 '18 at 11:24

I think this is an open question. As long as your function is automorphic, the Grand Riemann Hypothesis asks this exact question.

  • 1
    $\begingroup$ Thanks, I didn't know that. Does the Grand Riemann Hypothesis ask if this is the case, and permit one to assume the Riemann hypothesis? Because if not, it would seem a potentially important result, if it were true, that to prove GRH assuming RH, would prove RH. $\endgroup$ – user334732 Sep 5 '18 at 9:39
  • $\begingroup$ That doesn't sound right. $\endgroup$ – Uri George Peterzil Sep 5 '18 at 15:34

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.