5
$\begingroup$

What is the conceptual error in the following derivation? $$ e^{i\theta} = e^{i\frac{2\pi}{2\pi}\theta} = (e^{i2\pi})^{\frac{\theta}{2\pi}} = 1^{\frac{\theta}{2\pi}} = 1 $$ It is clear to me that the second move is illegal, I just don't know why.

$\endgroup$
  • $\begingroup$ I suspect that $(a^b)^c\neq a^{bc}$ in general in $\mathbb{C}$. $\endgroup$ – Justin Benfield Apr 12 '16 at 7:22
  • $\begingroup$ I bet it's something like (-8)^(1/3) = -2, (-8)^(1/3) != (-8)^(2/6) = 64^(1/6) = 2 $\endgroup$ – S.Klumpers Apr 12 '16 at 7:23
  • $\begingroup$ I've got another head scratcher for you : $$1=\sqrt{1}=\sqrt{(-1)^2}=\sqrt{(-1)(-1)}=\sqrt{-1}\sqrt{-1}=i\cdot i = i^2=-1.$$ $\endgroup$ – Mathematician 42 Apr 12 '16 at 7:27
1
$\begingroup$

$a^{x y} \not = (a^x)^y$ in general.

For example, $[(-1)^2]^{1/2} = 1$ but $(-1)^1 = -1$.

You may wish to look up "branch cuts" to understand conceptually why this is true.

$\endgroup$
  • $\begingroup$ So only if $b,c$ are integers, then $(a^b)^c = a^{bc}$ always? $\endgroup$ – Joshhh Apr 12 '16 at 7:26
  • $\begingroup$ @Joshhh Not "only if". If $a, b, c$ are all positive, for instance, then that rule holds. $\endgroup$ – Patrick Stevens Apr 12 '16 at 8:01
1
$\begingroup$

First, the formula $$ e^{i\frac{2\pi}{2\pi}\theta}=(e^{i2\pi})^{\theta/2\pi} $$ is not correct, because $z^{\alpha\beta}=(z^{\alpha})^{\beta}$ does not hold, in general.

Second, the formula $$ 1^{\theta/2\pi}=1 $$ is not correct.

Complex power is defined as $$ a^b = e^{b\ln a} $$ and can have (in)finitely many values if $b$ is not a real integer.

If $z^{\alpha\beta}=(z^{\alpha})^{\beta}$ holds, then $\beta$ is an integer. For detailed information, see here, page 12.

$\endgroup$
  • $\begingroup$ But $1,\theta,2\pi$ are all real numbers. $\endgroup$ – Joshhh Apr 12 '16 at 7:29
  • $\begingroup$ They are not necessarily integers, except $1$. In fact, $2\pi$ is not an integer clearly. $\endgroup$ – choco_addicted Apr 12 '16 at 7:30
  • $\begingroup$ $1^x$ is a constant function, so $1^{\frac{\theta}{2\pi}}$ is necessarily $1$. Why is the second move I made not true in general? $\endgroup$ – Joshhh Apr 12 '16 at 7:32
  • $\begingroup$ In real, your statement is true. But in complex analysis it is different. Compute $1^{1/2}$:$$1^{1/2}=e^{\frac{1}{2}\ln 1} =e^{\frac{1}{2} 2k\pi i} = e^{k\pi i}=(-1)^k,$$ where $k$ is an integer. $\endgroup$ – choco_addicted Apr 12 '16 at 7:36
  • $\begingroup$ $ {i\theta} = {i\frac{2\pi}{2\pi}\theta} = \log_e (e^{i2\pi})^{\frac{\theta}{2\pi}} = \log_e 1^{\frac{\theta}{2\pi}} = 0 $ ... at log level $\endgroup$ – Narasimham Apr 12 '16 at 7:37
-1
$\begingroup$

$\theta$ is the angle from the real axis so what you did that while converting $e^{i\theta}$ to $e^{i\theta \frac{2\pi}{2\pi}}$ you change the complex number because if $\theta \ne 2\pi$ then $e^{i\theta \frac{2\pi}{2\pi}} \ne e^{i2\pi \left(\frac{\theta}{2\pi}\right)}$ because the argument of the $e^{i\theta }$ is $\theta$ and the argument of $e^{i2\pi}$ is $2\pi$

$\endgroup$

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.