4
$\begingroup$

When talking about a number to a rational exponent, there are as many answers as the denominator of the exponent. Like the question: Is $9^{1/2}$ equal to $3$ or $-3$.

However when we have an irrational exponent like $2^{\sqrt2}$, I cant immediatly see more than one candidate solution: $e^{\sqrt2\log2}$. Are there any ways to interpret irrational exponents such that they give several candidate solutions?

Another way to state this question would be:

are there more than one complex solution to $x^{1/p}=y$ for irrational $p$ and positive real $y$.

$\endgroup$
  • $\begingroup$ Yes but you really should be thinking of these things as functions of a complex variable. What you want to look up is "branches" of analytic functions. $\endgroup$ – Gregory Grant May 12 '15 at 19:11
  • $\begingroup$ @GregoryGrant okay, so what would another candidate value of $2^{\sqrt2}$ be? It should obviously satisfy $x^{\sqrt{1/2}}=2$. $\endgroup$ – Alice Ryhl May 12 '15 at 19:15
  • $\begingroup$ @GregoryGrant What I mean is, if a number is given one irrational number as exponent, can we get several possible values like we did with $\sqrt2$? $\endgroup$ – Alice Ryhl May 12 '15 at 19:42
  • $\begingroup$ I suppose my question is equivalent to: are there more than one complex solution to $x^{1/p}=y$ for irrational $p$ and positive real $y$. $\endgroup$ – Alice Ryhl May 12 '15 at 19:51
  • $\begingroup$ Try the complex number $z$ that has argument $2\pi/\sqrt{2}$ and modulus equal to $2$. It lives somewhere in the 3rd quadrant of the complex plane. $\endgroup$ – Gregory Grant May 12 '15 at 19:59
1
$\begingroup$

In the real numbers field we have no ambiguity if we define (as usual) $a^{\frac{1}{2}}$ as the positive root of $a$ and $a^\alpha=e^{\alpha \log a}$ (for $a\le 0$) .

The problems come out when we go to complex numbers, as a consequence of the fact that $e^{i 2 k \pi}=1 \;\forall k \in \mathbb {Z}$. So $y=e^{i x}$ is a periodic function ( also for $x \in \mathbb{R}$) and is not invertible.

So, in the complex numbers field we have: $$ y=e^x \iff y=e^x\cdot 1 = e^xe^{i2k\pi}=e^{x+i2k\pi} $$ and the inverse '' function'' is a multivalued function : $$ \log y= x+i2k\pi $$ we can find $x$ writing the complex number $y$ in polar form as $y=|y|e^{i\theta}$ so, from $|y|e^{i\theta}=e^x$ we have $x=\log |y| +i\theta$ and: $$ \log y= \log |y| +i\theta+i2k\pi $$

As a consequences of this fact for $y,x,\alpha \in \mathbb{R}$ and $y,x >0$ (this imples $\theta=0$), we have: $$ y=x^\alpha \iff y=e^{\alpha\log x}=e^{\alpha(\log |x|+i2k\pi)} $$

For $\alpha=1/2$ this gives the two square roots and for $\alpha$ irrational gives, in general, many values.

$\endgroup$

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.