10
$\begingroup$

Why is the log of $x$ equal to the log of the absolute value of $x$ plus $i$ times $\pi$?

$$\log(x)=\log(|x|)+i\pi\text{, for }x < 0$$

Where does the $\pi$ come from? Is it from a logarithmic identity?

I know it sounds silly, but I was not able to find an answer from existing solutions: here, here, and here.

$\endgroup$
  • 2
    $\begingroup$ It comes from the identity $e^{i\pi} +1 = 0$, we rearrange this to be $\log(-1) = i \pi$, then for any negative number, $\log(-x) = \log( (-1) \cdot (x) ) = \log(-1) + \log(x) = i \pi + \log(x)$ $\endgroup$ – Rob Bland Jan 9 '17 at 1:09
  • 2
    $\begingroup$ Are you familiar with the complex logarithm? $\endgroup$ – pjs36 Jan 9 '17 at 1:10
  • $\begingroup$ Probably it comes from $e^{i\pi}=-1$ $\endgroup$ – drzbir Jan 9 '17 at 1:10
  • $\begingroup$ someday, when you study complex analysis, you will understand that what you said is only approximately true... $\endgroup$ – GEdgar Jan 9 '17 at 3:09
  • 1
    $\begingroup$ You asked "why is this not always true" but you linked to a page that says the formula is always true. $\endgroup$ – David K Jan 9 '17 at 3:28
18
$\begingroup$

$e^{πi} + 1 = 0$.

$e^{πi} = -1$.

$πi = \log(-1)$.

$\log(-1) + \log(|a|) = \log(-|a|)$.

$\log(-|a|) = \log(|a|)+iπ$.

$\endgroup$
5
$\begingroup$

The standard explanation would go like this:

Any complex number $z$ can be written in polar form: $$z = |z|e^{i\theta}.$$ So, if we want to define the complex logarithm, we do so as follows: $$\log (z) = \log(|z|e^{i\theta}) = \log(|z|) + \log(e^{i\theta}) = \log(|z|) + i\theta.$$ In particular, the logarithm of a negative real number $x$ can then be calculated as $$\log (x) = \log(|x|e^{i\pi}) = \log(|x|) + \log(e^{i\pi}) = \log(|x|) + i\pi.$$

However, this explanation is not sufficient and the logarithm as presented is NOT a well-defined function. The angle $\theta$ is, well, an angle, and hence only defined up to multiples of $2\pi$: $$|z|e^{i\theta} = |z|e^{i(\theta+k2\pi)}$$ for all $k \in \mathbb{Z}$. Therefore, the complex logarithm is only defined up to multiples of $2\pi i$ !

For example: $$\log (x) = \log(|x|e^{i\frac{3}{2}\pi}) = \log(|x|) + \log(e^{i\frac{3}{2}\pi}) = \log(|x|) + i\frac{3}{2}\pi,$$ or maybe $$\log (x) = \log(|x|e^{-i\frac{177}{2}\pi}) = \log(|x|) + \log(e^{-i\frac{177}{2}\pi}) = \log(|x|) - i\frac{177}{2}\pi.$$ But clearly $\pi \neq \frac{3}{2}\pi \neq -\frac{177}{2}\pi$. The point is: the complex logarithm is not a function, but what we call a multivalued function. To turn it into a proper function, we must restrict what $\theta$ is allowed to be, for example $\theta \in (-\pi,\pi]$. This is called the principal complex logarithm and is usually denoted by $\operatorname{Log}$ (capital L).

Technically, it doesn't matter to what range you restrict $\theta$, as long as the resulting logarithm is a proper function (not a multivalued function) and you are consistent in your restriction of $\theta$. For example, the following "proof" can be obtained if you're sloppy: \begin{align} e^{\pi i} = -1 & \implies (e^{\pi i})^2 = (-1)^2 & \text{ (square both sides)}\\ & \implies e^{2\pi i} = 1 & \text{ (calculate the squares)}\\ & \implies \log (e^{2\pi i}) = \log(1) & \text{ (take the logarithm)}\\ & \implies 2\pi i = 0 & \text{ (calculate the logarithms)} \end{align} Clearly this is wrong!

$\endgroup$
4
$\begingroup$

Slightly tongue in cheek but partially seriously I'm going to say "magic".

What I mean is before we talk about $\ln -k$ for a negative number we have to define what on earth it could possibly mean to have $e^{m} = -k$. There is no "natural" meaning to that and mathematicians, in one sense, must contrive a meaning for it. If our universe is the real numbers then this is .... impossible. If $b > 0$ then $b^x > 1$ and $e^x$ being negative is simply impossible.

But what if our universe is the complex numbers. Then if $z = a + bi$ ($i^2 = -1$) then what could $e^{a+bi}$ possibly mean? After all we can't just "multiply $e$ by itself the square root of negative one times".

Well, we know for real numbers, $e^{x + y} = e^xe^y$ and $\frac {de^x}{dx} = e^x$. For this to still be true for complex numbers as well as for real numbers the only way to define $e^{z}$ where $z$ is a complex number, so that $e^{z+w} = e^ze^w$ and so that $\frac {d^z}{dz}$ the only possible way for that to stay true we have to define $e^{a+bi} = e^a(\cos b + i \sin b)$.

Okay. That was ... a lot of hand waving. But it works and if you study complex analysis it will be derived in great detail. BUT this means....

$e^{i\pi} = e^{0 + i\pi} = e^0(\cos \pi + i \sin \pi) = 1*(-1 + i*0) = -1$.

This is Euler's Formula, one of the most famous mathematical formula in history.

Suddenly $e^z$ being a negative number is not impossible. But if $e^z$ is negative then we need to have $e^{z} = e^{a + bi} = e^a(\cos b + i \sin b)$ so $\cos b + i \sin b$ is a negative number. That means $b = \pi$ . So $z = i \pi$.

So this means $\ln -1*x = \ln -1 + \ln x = i\pi + \ln x$.

And that's where $\pi$ comes from. When we define $e^{a+bi} = e^a(\cos b + i \sin b)$, exponents become thoroughly linked with trigonometric functions. As such, $\pi$ is an essential part of the inverse.

For the real numbers only and logs of positive real numbers only, we don't have to worry as $e^x = e^{x + 0*i} = e^x(\cos 0 + i \sin 0) = e^x(1 + i*0) = e^x= k$ and $\ln k = x$ and $\pi$ is not relevant.

$\endgroup$

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.