8

We know $e^{2\pi i} = 1$, and that $(x^m)^n = x^{mn}$. This way, we can rewrite $e^{n}$ as some version of $(e^{2\pi i})^{\frac{n}{2\pi i}}$ for most n (right?).

But if this is true, then why isn't $e^3 = 1$, for example, if we can rewrite it as $(e^{2\pi i})^{\frac{3}{2\pi i}} = (1)^{\frac{3}{2\pi i}} = 1$ ? What am I missing here?

I just came upon this issue by accident while doing a problem, and I'm not sure how to best resolve it.

Hans Lundmark
  • 53,395
creative_name
  • 189
  • Have you seen a definition of exponentiation for complex numbers? There's a lot of insight in learning about that definition that a more brief answer is likely to miss. – Milo Brandt Apr 08 '19 at 03:40
  • @Milo Brandt I know Euler's formula, but that's about it... what are you referring to? – creative_name Apr 08 '19 at 03:43
  • Usually, if you have complex numbers $x$ and $y$, you define $$x^y=e^{y\log(x)}$$ where $e^z=\sum_{i=0}^{\infty}\frac{z^i}{i!}$ and $\log$ is the natural logarithm. The thing to understand about this is that it's not clear how to define $\log$ - for instance, is $\log(1)=0$ or is $\log(1)=2\pi i$ since $e^0=e^{2\pi i}=1$? This breaks most hope of things working nicely, but it sort of needs a deeper understanding of what's going on. – Milo Brandt Apr 08 '19 at 03:50
  • Similar: https://math.stackexchange.com/questions/281528/what-is-wrong-with-this-fake-proof-ei-1 – Hans Lundmark Apr 08 '19 at 05:01

1 Answers1

8

Good question! The answer is that although the rule ${(x^b)^c} = x^{bc}$ holds when $b$ and $c$ are integers, it does not hold in general when they are not integers.

Consider the following simpler example. As you know, $(-1)^2 = 1$. Raising both sides to the power $\frac12$, we get $${((-1)^2)}^{1/2} = 1^{1/2},$$ which is still correct, but we cannot then apply the ${(x^b)^c} = x^{bc}$ rule to the left side to obtain $$(-1)^{2\cdot(1/2)} = (-1)^1 = -1 = 1.$$

MJD
  • 65,394
  • 39
  • 298
  • 580
  • This is much better than my (hypothetical) answer, much more elementary; I was kind of going on this rant about complex numbers and multivalued functions lol. – PrincessEev Apr 08 '19 at 03:35
  • 4
    Also worth noting: It holds when $x$ is a non-negative real number and when $b$ and $c$ are both real. – Milo Brandt Apr 08 '19 at 03:35