0
$\begingroup$

According to Wolfram Alpha, $1^z=1$ for $z\in\mathbb{C}$. If this is true, then what is wrong with the following argument that $1^z$ has infinitely many values?

Let $z=x+iy$. Then, \begin{align} 1^z &= 1^{x+iy} \\ &= 1^x \cdot 1^{iy} \\ &= 1^{iy} \\ &= e^{iy\log(1)} \\ \log(1) &= \{0,2i \pi, 4i \pi,6i \pi, \ldots\} \\ 1^z &= \{e^{iy \cdot 0},e^{iy \cdot 2 i \pi},e^{iy \cdot 4 i \pi},e^{iy \cdot 6 i \pi},\ldots\} \\ &= \{e^0,e^{-2y\pi},e^{-4y\pi},e^{-6y\pi},\ldots\} \end{align}

Only one of these values is equal to $1$.

$\endgroup$
6
  • 6
    $\begingroup$ Wolfram alpha is not God so take its answers in context, not as eternal truths of mathematics; in particular, you are right that $1^z$ is (in general) an infinite set of which $1$ is a distinguished value $\endgroup$ – Conrad Dec 31 '20 at 19:06
  • 1
    $\begingroup$ @Jakobian because for some people Wolfram Alpha is God whose answers are eternally true $\endgroup$ – Conrad Dec 31 '20 at 19:07
  • 1
    $\begingroup$ @Conrad Thou shall not question the infinite wisdom of WolframAlpha ;) $\endgroup$ – Severin Schraven Dec 31 '20 at 19:09
  • 2
    $\begingroup$ @SeverinSchraven I am still waiting for the Beta version... :) $\endgroup$ – Raffaele Dec 31 '20 at 19:10
  • 1
    $\begingroup$ If you define $\log$ as a function, we have to pick a single value of $\log 1,$ and we’ll usually choose the value $0.$ If you have $\log$ as a multi valued function, then you are correct, $w^z$ is multi-valued, too. $\endgroup$ – Thomas Andrews Dec 31 '20 at 19:11
3
$\begingroup$

If, when $a>0$, you define $a^z$ as $\exp\bigl(z\log(a)\bigr)$, where $\log(a)$ is the only real logarithm of $a$, then indeed we always have $1^z=1$.

Otherwise, it's up to you to tell us how you are defining $1^z$. However, note that the set of all logarithms of $1$ is not $\{0,2\pi i,4\pi,i,6\pi i,\ldots\}$; it's $\{0,\pm2\pi i,\pm4\pi i,\pm6\pi i,\ldots\}$.

$\endgroup$
9
  • $\begingroup$ Thank you. I noticed that you once answered a question about a fake proof about why $re^{i\theta}=r$: $$re^{i\theta} = re^{(2 \pi i \theta)/2\pi} = r(e^{2\pi i})^{\theta /2\pi} = r(1)^{\theta/2\pi} = r \, .$$ Is the problem the same as the one here: $e^z$ is a function; but $(e^{2\pi i})^z=1^z$ is not a function if we define $z^w$ as $\exp(w\log(z))$? $\endgroup$ – Joe Dec 31 '20 at 19:23
  • $\begingroup$ No; it has something to do with it, but it's not the same thing. Here, we have a problem of defining the meaning of the expressions that we are using. I provided in my answer the more usual meaning of the expression $a^z$, when $z\in\Bbb C$ and $a\in(0,\infty)$. According to that definition, $1^z$ is indeed $1$, for every $z\in\Bbb C$. If you are using some other definition of $1^z$, then you should tell us which definition you have in mind. $\endgroup$ – José Carlos Santos Dec 31 '20 at 19:29
  • $\begingroup$ I wasn't aware that $a^z$ could have more than one definition if $a$ is a real number. Thank you for clearing up my confusion. In your experience, when people write $e^z$, is this generally just a shorthand for $\exp(z)$? Because if we interpret $e^z$ as $\exp(z\log(e))$, where $\log$ is the complex logarithm, then it would seem that $e^z$ is a multi-valued function. $\endgroup$ – Joe Dec 31 '20 at 19:46
  • $\begingroup$ In my experience, every single time that someone writes $e^z$, what that means is $\exp(z)$. However, concerning $\exp(z\log(e))$, note that I wrote in my answer that I was dealing with the only real logarithm of $a$. In this case, $a=e$, and its only real logarithm is $1$. So, $\exp(z\log(e))=\exp(z)$. $\endgroup$ – José Carlos Santos Dec 31 '20 at 19:49
  • $\begingroup$ Going back to the fake proof: you say that the problem is that in general $e^{ab} \neq (e^a)^b$. In the context of the fake proof, you appear to be saying that $re^{(2 \pi i \theta)/2\pi} \neq r(e^{2\pi i})^{\theta /2\pi}$. Is it possible if you elaborate on what the problem is here? $\endgroup$ – Joe Dec 31 '20 at 20:25

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.