2

How do I show $\arccos(4)$ is ${2\pi k}$ $\pm$ $i\operatorname{arcosh}(4)$?

I am getting $\pm$ $iln(4+\sqrt{15}$)

When I just use $\cos(z)$ =$\cosh(iz)$, I get: $-i\operatorname{arcosh}(4)$.

I’m lost! If someone could type out the full proof, it'd be greatly appreciated. I'm not seeing it from the answers below.

Thanks.

Edit: I follow the answers on the whole, but cannot seem to derive why arcosh has a prefix of $\pm$. If someone could show a derivation for this I'd be grateful. From function theory I'm expecting only the one positive root, not the -ve one. That said, I get the $\pm$ for the $\ln(x)$ version so I believe it should be there, but need a proof! :) Thanks.

Noobcoder
  • 227

4 Answers4

2

$\arccos 4 = a+bi$

$4 = \cos(a+bi) = \cos a \cos bi - \sin a \sin bi = \cos a \cosh b - i\sin a \sinh b$

By comparing real and imaginary parts,

$\cos a \cosh b = 4$ and $\sin a \sinh b = 0$

Where $a, b$ are both real.

One possible solution to the second equation is $\sinh b =0 \implies b=0$, but that would make the first equation $\cos a = 4$, which has no real solution. So disregard this, i.e. $b \neq 0$.

Which leaves $\sin a =0 \implies a = n\pi, n \in \mathbb{Z}$.

In the first equation that makes $\pm \cosh b = 4$. Since the hyperbolic cosine for real values is non-negative only accept the positive case.

So $a = 2k\pi, k \in \mathbb{Z}$ (only even multiples of $\pi$ give a positive cosine) and $\cosh b = 4 \implies b = \pm \cosh^{-1} 4$ (since the hyperbolic cosine is an even function).

Thus we get the solution $\arccos 4 = 2k\pi \pm i\cosh^{-1} 4, k \in \mathbb{Z}$.

Note that you can also express the $\cosh^{-1} 4$ part in terms of logarithms if you wish. I believe this is what you did. To be clear $\cosh^{-1} 4 = \ln (4 + \sqrt {4^2 - 1}) = \ln (4 + \sqrt{15})$ , which means you can also write $\arccos 4 = 2k\pi \pm i \ln (4 + \sqrt{15}), k \in \mathbb{Z}$. This is an equivalent answer.

Deepak
  • 26,801
  • 1
    (+1) $2k\pi\pm i\log\left(4+\sqrt{15}\right)$ is what I had come up with also. While $\operatorname{arccosh}$ is not a bad function, $\log$ is more comfortable to me. – robjohn Mar 24 '20 at 07:23
  • i get that cosh is even, but how does that result in arcosh having a +-? Surely arcosh has a 1-1 mapping? – Noobcoder Mar 24 '20 at 10:12
  • $f(x) = \cosh x$ is an even function which means it satisfies $f(-x) = f(x)$. To solve $f(x) = c$, you will get two values of $x$ (for nonzero $x$). Think of what happens when you solve $x^2 = 4$. Two solutions right? $-2$ and $2$. Same deal here. The inverse function is the positive square root function and that is one to one, so you need to express the solution as $\pm \sqrt 4$. Hope that's clearer. – Deepak Mar 24 '20 at 10:22
  • @robjohn How did you get mathjax to give you a nice arccosh? I simply couldn't, seems the function is not in the standard library. So I used the inverse "exponent" notation, which I'm not a big fan of. – Deepak Mar 24 '20 at 10:25
  • @Deepak so cosh x = 4 is analagous to x$^2 = 4$. x = $\pm \sqrt4$, and if I take arcosh of both sides i'd get: x = $\pm\operatorname{arcosh} (4)$? so the arcosh 4 would itself be the positive root, but the plus and minus would be for the original function? Kind of got it but still missing something... – Noobcoder Mar 24 '20 at 10:49
  • @Deepak $\cosh(x)$ doesn't have a square, so how is it analagous to taking a square root? Forgive my simpleness! – Noobcoder Mar 24 '20 at 10:55
  • It's an analogy. When solving x squared = something, you need to consider two values, both plus and minus. Similarly, when solving cosh x = something (that's what I'm doing in my solution when I solve $\cosh b = 4$),you also need to consider plus and minus. I've kept things as simple as possible I think. It's basically what you have to do when you solve equations involving even functions. – Deepak Mar 24 '20 at 11:03
  • Just in case, this is the part that's confusing you, I'm not saying arccosh has two values. It has only one value. That's precisely why you need the plus minus because the solution of cosh x = something has two possibilities. – Deepak Mar 24 '20 at 11:04
  • I still don't get it. if you know of a website with examples of this property of even functions please can you link me. I've gone through my textbooks and I can't find any example of a +- in front of the hyperbolic function (or even the even cosine function for that matter) – Noobcoder Mar 24 '20 at 11:28
  • See this link, look at the first graph: https://www.mathsisfun.com/algebra/functions-odd-even.html See graphically how the "mirrored" x values represent the two solutions to $x^2 +1 = $(something). If you solve $x^2 + 1 =4$, you'll get two solutions, same magnitude, one pos, one neg. The graph of cosh x is very similar looking to that graph, so you'll observe the same thing when solving cosh x = 4. – Deepak Mar 24 '20 at 11:59
  • I'm sorry I don't have the time for an extended chat now. Perhaps someone else can help. You accepted an answer - that also has the plus/minus. Perhaps you can ask that user, maybe the explanation will be easier for you to understand. Sorry. – Deepak Mar 24 '20 at 12:32
  • @Noobcoder: $\cosh(x)$ is an even function and therefore, if $\cosh(x)=y$, then we also have $\cosh(-x)=y$. Thus, if we want to find values of $x$ so that $\cosh(x)=y$, we have at least two solutions, $x=\pm\operatorname{arccosh}(y)$. – robjohn Mar 24 '20 at 16:36
  • Hi guys, I finally get it. Just needed to think about it graphically! Cosh x =4 has two solutions so it follows that x =+- arcosh(4) as you say. For some reason I was getting wrapped up with the fact that arcosh’s domain needs to be restricted to +ve to satisfy it's need to be an inverse function! (Basically over complicating it!) – Noobcoder Mar 25 '20 at 11:38
  • @Noobcoder I'm glad you got it. I thought the graph in the link I gave would be helpful, maybe it was, but anyway, good thsr you have. Now here's a bit of complexity (no pun intended) - working in the real numbers, $y =\arccos x$ has just one value. There are an infinite number of values of $y$ that satisfy $\cos y = x$ but you select just a single one by limiting the range so that you get a function. (continued) – Deepak Mar 25 '20 at 12:07
  • (continued) But when working with complex numbers it's not so simple. There is a way to choose a single value systematically by a particular convention but it involves considering "branch cuts", and the theory is complicated. In any case, it is likely this question is asking for the evaluation of the so-called multi-valued complex arccos "function" (a misnomer because it's not really a true function at all) which basically returns all possible values that have a cosine of $4$. I hope this addendum didn't confuse you more. – Deepak Mar 25 '20 at 12:07
1

If we use $e^{i\theta}=\cos{(\theta)}+i\sin{(\theta)}$, we can show that $\cos{(\theta)}=\frac{e^{i\theta}+e^{-i\theta}}{2}$ or $\cos{(\theta)}=\cosh{(i\theta)}$.

$$ \begin{aligned} \cosh{(\pm\cosh^{-1}{(4)})}&=4\\ \cosh{(i(2\pi n\pm i\cosh^{-1}{(4))})}&=4\\ \\ \cos{(2\pi n\pm i\cosh^{-1}{(4)})}&=4 \end{aligned} $$

however... i do not know the range of $\cos^{-1}{(x)}$ in complex plane

acat3
  • 11,897
1

The point is that

$$\cosh(x +i 2 \pi k) = \cosh(x) = \cosh(-x) $$

for $k$ integer.

lcv
  • 2,506
1

Let $z = \arccos(4)$ and use the identity $\cos(iw)=\cosh(w)$ to express,

$$\cos(z) = \cos(2\pi n\pm z ) =\cos[i(-i2\pi n \mp iz )]=\cosh(-i2\pi n \mp iz ) = 4$$

or,

$$-i2n\pi \mp iz = \text{arccosh}(4)$$

which leads to,

$$z = \arccos(4) = 2\pi k \pm i\>\text{arccosh}(4) $$

Quanto
  • 97,352
  • I follow most of this but am stumped by the +- bit. Can you elaborate why you have +- on the iz and not the 2pin to start please, and what is the reasoning you moved it across to the arcosh? Thanks – Noobcoder Mar 24 '20 at 09:41
  • i figured out why the +- against the iz. Still not sure why it was moved to the arcosh though.if you could add this reasoning to the solution to make it clearer it would be helpful. Thanks! – Noobcoder Mar 24 '20 at 10:19