Given that $n$ is rational, when is $\sqrt[n]{n}$ rational?

We can make a polynomial $x^{n}-n$ whose root is $\sqrt[n]{n}$ and using RRT we can show that there are no rational roots, but in the process we are assuming that $n$ is an integer (coefficients must be integers for RRT to work).

How can we solve this problem?

  • $\begingroup$ I think your title should read "If $n$ is rational, when is $n^{1/n}$ rational?" $\endgroup$ – Charles Hudgins Sep 11 at 16:56
  • $\begingroup$ $n^{\frac 1k}$ is rational if and only if $n$ is a $k$powered integer and $n^{\frac 1k}$ is an integer. $\endgroup$ – fleablood Sep 11 at 16:59
  • $\begingroup$ Oh, $n$ is rational. Okay, a little more work but same idea. $\endgroup$ – fleablood Sep 11 at 17:07

If $1/n$ is an integer, it works... conversely, if $(p/q)^{q/p}$ is a rational $a/b$, where $p,q$ are coprime positive (and $a,b$ as well), then $p^q/q^q=a^p/b^p$, both as quotients of coprime positive integers, so $p^q=a^p$.

So if $\pi$ is a prime divisor of $p$ with multiplicity $\nu$ (ie $\pi^{\nu}|p$, $\pi^{\nu+1}$ doesn't divide $p$), then the $\pi$-adic valuation of $p^q$ is $\nu q$. But as $p^q$ is a $p$-th power, said valuation is a multiple of $p$, and thus $p|\nu q$. But as $p$ and $q$ are coprime, $p|\nu$, thus $\nu \geq p$, and hence $p \geq \pi^{\nu} > \nu$, a contradiction. Thus $p=1$ (because $p$ has no prime divisors), whence the conclusion.

| cite | improve this answer | |
  • $\begingroup$ I'm not sure I understand your last sentence. $\endgroup$ – Timothy James Sep 11 at 18:00
  • $\begingroup$ Could you elaborate more and go through it a bit more thoroughly? $\endgroup$ – Timothy James Sep 11 at 18:00
  • $\begingroup$ I edited, is it clearer now? $\endgroup$ – Mindlack Sep 11 at 20:16
  • $\begingroup$ "But as $p^q$ is a $p$-th power, said valuation is a multiple of $p$, and thus $p|νq$.". How come it implies that the valuation is a multiple of $p$? $\endgroup$ – Timothy James Sep 11 at 20:36
  • $\begingroup$ Let $c=p^q=a^p$. Let $\nu$ be the $\pi$-adic valuation of $p$, $\mu$ that of $a$. Then the $\pi$-adic valuation of $c=p^q$ is $\nu q$, but the $\pi$-adic valuation of $c=a^p$ is $p \mu$. So $p\mu=q\nu$ thus $p|q\nu$. $\endgroup$ – Mindlack Sep 11 at 20:42

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.