If $x$ is a rational number and $y$ is an irrational number then $x^y$ is irrational.

I have tried to prove it,

let $x= 2$ and $y=\sqrt{2}$ then $x^y = 2^{\sqrt{2}}$,

if it is an rational number then let $x=2^{\sqrt{2}}$ and $y=\sqrt{2}$, then, $x^y= 2^{\sqrt{2} *\sqrt{2}}=4$ which is rational,

therefore $x^y$ is rational. Is it correct?

  • $\begingroup$ Tidy up your proof a little, but I'm pretty sure that's all there is. $\endgroup$ – ConMan Sep 27 '16 at 0:23
  • 2
    $\begingroup$ You have proved (almost, if you tidy up it will be proved) that there is a rational $x$ and an irrational $y$ such that $x^y$ is rational. Not what your title asks for. $\endgroup$ – Jean-Claude Arbaut Sep 27 '16 at 0:25
  • 1
    $\begingroup$ Not following...are you trying to prove that $2^{\sqrt 2 }$ is rational? $\endgroup$ – lulu Sep 27 '16 at 0:25
  • 4
    $\begingroup$ Simple counting shows that there must be an irrational $y$ such that $2^y$ is irrational (there are uncountably many irrationals, but only countably many rationals). $\endgroup$ – lulu Sep 27 '16 at 0:27
  • $\begingroup$ Didnvt you say you wanted to prove it was irational? $\endgroup$ – fleablood Sep 27 '16 at 0:28

very close!

$\color{white}{\text{I like the try though +1}}$

Let $x=2$ and $y=\sqrt{1/2}$.

Either $2^{\sqrt{1/2}}$ is irrational, or $\left(2^{\sqrt{1/2}}\right)^{\sqrt{1/2}}$ is irrational

  • $\begingroup$ Classic annoying example by Hilbert? Can't remember. $\endgroup$ – Jacky Chong Sep 27 '16 at 0:38
  • $\begingroup$ @JackyChong I don't know, but I decided to play along with the OP's attempted method. $\endgroup$ – Simply Beautiful Art Sep 27 '16 at 0:38

As a concrete example of $(\text{rational})^{\text{irrational}}=\text{rational}$: $$10^{\log_{10}3}=3$$ To see that $\log_{10} 3$ is irrational, suppose it equaled $\frac ab$ with $a,b\in \mathbb N$. We'd then have $$3=10^{\frac ab}\implies 3^b=10^a$$ But this contradicts Unique Factorization.

As a concrete example of $(\text{rational})^{\text{irrational}}=\text{irrational}$: $$10^{\log_{10}\pi}=\pi$$ To see that $\log_{10}\pi$ is irrational, suppose it equaled $\frac ab$ $a,b\in \mathbb N$. We'd then have $$\pi=10^{\frac ab}\implies \pi^b=10^a$$ Whence $\pi $ would be a root of the equation $X^b-10^a=0$. But $\pi$ is transcendental, hence is not the root of any polynomial with rational coefficients.

  • $\begingroup$ A simpler example of the former is $1^{\sqrt{2}} = 1 = \mbox{rational}$. $\endgroup$ – Michael Sep 28 '16 at 3:10
  • $\begingroup$ @Michael True, of course...but in this sort of question, it's usual to exclude $0,1$ from the base (as the examples aren't very interesting). For example, we have the powerful Gelfond Schneider Theorem which tells us that, if $a,b$ are algebraic with $b\notin \mathbb Q$ then $a^b$ is transcendental. But of course this must exclude $a=0,1$. $\endgroup$ – lulu Sep 28 '16 at 10:49
  • $\begingroup$ Given the countability argument and the Gelfond Schneider theorem you mention, there are abundant examples of $(\mbox{rational})^{\mbox{irrational}}= \mbox{irrational}$. So, it surprises me that it seems difficult to come up with one concrete example. Yours uses the transcendental nature of $\pi$ which is not a basic fact (i.e. proof is advanced). Are there simpler examples? Of course, I suppose you can replace $\pi$ with any transcendental, but then you have to prove existence of such, which perhaps goes back to countability arguments? $\endgroup$ – Michael Sep 28 '16 at 14:28
  • 2
    $\begingroup$ liouville's theorem let's you produce some "simple" transcendental numbers. Like $\sum 10^{-n!}$. $\endgroup$ – lulu Sep 28 '16 at 22:49

Changing the letters for convenience, let's show that there's a rational number $x$ such that $2^x$ is irrational. For each rational number $r,$ the curve $y=2^x$ meets the vertical line $x=r$ in just one point, and it meets the horizontal line $y=r$ in just one point. Since the set of rational numbers is countable, there are only countably many points on the curve $y=2^x$ with at least one rational coordinate. Since there are uncountably many points on the curve, there are uncountably many points with both coordinates irrational; i.e., there are uncountably many (in fact continuum many) irrational values of $x$ such that $2^x$ is also irrational. The same goes with $2$ replaced by any positive rational number different from $1.$


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.