It is well known that there's no conclusion now whether $\pi+e$ is rational or not. What would happen if we knew that $\pi+e$ is rational? Specifically, are there related open problems that would be settled?

  • 11
    $\begingroup$ Probably nothing; usually, its the techniques that lead to a new result that are the exciting part, rather than the result itself. $\endgroup$ – goblin Mar 28 '14 at 10:51
  • 33
    $\begingroup$ Human sacrifice, dogs and cats living together... mass hysteria! $\endgroup$ – user98602 Mar 28 '14 at 11:07
  • 15
    $\begingroup$ Depends. If $\pi+e$ turned out to be rational, I would be inclined to believe that God is joking with us. $\endgroup$ – Jyrki Lahtonen Mar 28 '14 at 11:43
  • 9
    $\begingroup$ If $\pi + e$ turned out to be rational then there will be a much more simpler and natural relation between $e$ and $\pi $ than the traditional $e^{i\pi} + 1 = 0$. $\endgroup$ – Paramanand Singh Mar 28 '14 at 11:47
  • 8
    $\begingroup$ Restating the fact given by @ParamanandSingh, if $\pi + e$ was rational, or even algebraic, Schanuel's hypothesis (currently an open problem in transcendence theory) would fall off, since $\pi + e = k$ for some $k \in \Bbb Q$ and $e^{i\pi} + 1 = 0$ are two different expressions in the exponential ring, and Schanuel's conjecture essentially implies that Euler's identity is the unique $\mathcal E$-algebraic relation between $e$, $\pi$ and $i$. $\endgroup$ – Balarka Sen Mar 28 '14 at 13:39

This is a comment that's too long for the usual format. If $\alpha=e+\pi$ can be written as a fraction of integers, both numerator and denominator must be $\geq 2 \times 10^{32}$. To see this, let $A,B,C,D$ be integers defined by

$$ \begin{array}{lcl} A &=& 3063742572717320569341511991159738 \\ B &=& 522834163445445988434458010516405 \\ C &=& 9765222175513935643148512770417523 \\ D &=& 1666455861030599542832067804101203 \\ \end{array} $$

Then any good formal computing system will confirm to you that $\frac{A}{B} < \alpha < \frac{C}{D}$ and $BC-AD=1$. If $\alpha$ is rational, $\alpha=\frac{p}{q}$ with $p,q \in {\mathbb N}_{>0}$, then $u=pB-qA$ and $v=qC-pD$ must be positive integers. But then $p=Cu+Av\geq A+C \geq 2\times 10^{32}$ and similarly $q=Du+Bv\geq B+D \geq 2\times 10^{32}$.

| cite | improve this answer | |

It would imply in particular that $e$ is a period, which it conjecturally isn't. There are deep reasons why $e$ is conjectured not to be a period, stemming (I believe) from Deligne's theory of motivic weights. I recommend taking a look at Kontsevitch and Zagier's fantastic article Periods.

| cite | improve this answer | |
  • $\begingroup$ Interesting ref, thanks! +1 $\endgroup$ – Markus Scheuer Jan 15 '15 at 13:38
  • $\begingroup$ @Markus My pleasure! $\endgroup$ – Bruno Joyal Jan 15 '15 at 13:45

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.