I tried solving this using the definition of $cos(z)=\frac{e^{iz}+e^{-iz}}{2}$ and equating it to $\frac{3}{4}+\frac{i}{4}$ and converting it to a complex quadratic equation through a substitution $t=e^{iz}$ and finding roots via the complex quadratic formula but it didn't seem to work. I would prefer solutions via elementary methods.

Here is my attempt:

By definition we have $\frac{e^{iz}+e^{-iz}}{2}=\frac{3}{4}+\frac{i}{4} \implies e^{iz}+e^{-iz}=\frac{3}{2}+\frac{i}{2}$. Let $t=e^{iz}$ so then we have $t+\frac{1}{t}=\frac{3}{2}+\frac{i}{2}$ and if we multiply both sides by $t$ we have $t^2+1=(\frac{3}{2}+\frac{i}{2})t$ and hence $t^2+(\frac{3}{2}+\frac{i}{2})t+1=0$ By the quadratic formula for complex numbers we have, $a=1, b=\frac{3}{2}+\frac{i}{2}, c=1 \implies z=\frac{-(\frac{3}{2}+\frac{i}{2}) \pm \sqrt{(\frac{3}{2}+\frac{i}{2})^2-4(1)(1)}}{2(1)}$. Simplyifing we have $z=\frac{-\frac{3}{2}-(\frac{1}{2})i \pm \sqrt{-2+(\frac{3}{2})i}}{2}$ We wish to express $-2+(\frac{3}{3})i$ in polar form so we have $|-2+(\frac{3}{2})i|=\frac{5}{2}$. Now equating the real and imaginary parts we have $\frac{5}{2}\cos(\theta)=-2 \implies \cos(\theta)=-\frac{4}{5}$ and $\frac{5}{2}\sin(\theta)=\frac{3}{2} \implies \sin(\theta)=\frac{3}{5}$. From this we have $\tan(\theta)=-\frac{3}{4} \implies \theta=\arctan(-\frac{3}{4}) \approx -.6435$ rad. So we have $w=-2+(\frac{3}{2})i=\frac{5}{4}(\cos(\theta)+i\sin(\theta))=\frac{5}{4}e^{i\theta}$. By Proposition 1.3.12 we have $\sqrt{w}=\sqrt{\frac{5}{4}}e^{\frac{i\theta}{2}}=\frac{\sqrt{5}}{2}e^{\frac{i\theta}{2}}$. Similarily for $-\frac{3}{2}-(\frac{1}{2})=\frac{\sqrt{10}}{2}e^{i\varphi}$ Where $\varphi=\arctan(\frac{1}{3})$. So finally we have $z=\frac{-(\frac{3}{2}+\frac{i}{2}) \pm \sqrt{(\frac{3}{2}+\frac{i}{2})^2-4(1)(1)}}{2(1)}=\frac{\sqrt{10}e^{i\varphi} \pm \sqrt{5}e^{\frac{i\theta}{2}}}{4}$ as solutions to $\cos(z)=\frac{3}{4}+\frac{i}{4}$.

  • 3
    $\begingroup$ Your method is the standard technique. $\endgroup$ Jan 26, 2014 at 21:27
  • 1
    $\begingroup$ How can it be more elementary? $\endgroup$
    – J.R.
    Jan 26, 2014 at 21:30
  • $\begingroup$ I wasn't aware that was the "standard" technique. One thing I did differently was instead of using the form $a+bi$ for complex numbers, I opted to use $re^{i\theta}$ as the representation so I could avoid taking $\sqrt{a+bi}$ in the discriminant since $\sqrt{re^{i\theta}}=\sqrt{r}e^{\frac{i\theta}{2}}$ seemed easier to work with. $\endgroup$
    – 1028
    Jan 26, 2014 at 21:37
  • $\begingroup$ The only other reasonable technique I know of would be to compute $\arccos$. However, IMO, the formula for $\arccos$ is pretty annoying. $\endgroup$
    – user14972
    Jan 26, 2014 at 21:40
  • 1
    $\begingroup$ If you show the details of the computation that "didn't seem to work", it would be much easier for someone to explain what you did wrong. $\endgroup$ Jan 26, 2014 at 21:42

3 Answers 3


I agree with your approach for the most part, but I think you've messed up on calculating $b$ in the quadratic formula: $$e^{iz} + e^{-iz} -\frac{3}{2} - \frac{i}{2} = 0$$

Using $t = e^{iz}$: $$t + \frac{1}{t} -\frac{3}{2} - \frac{i}{2} = 0$$ Multiplying by $t$: $$t^2\color{red}{-}\left(\frac{3}{2} + \frac{i}{2}\right)t + 1 = 0$$ This is where your solution starts to go wrong. Those minus signs are just lying in wait for the innocent mathematician! :P

Thus: $$\begin{align} x &= \frac{-b \pm \sqrt{b^2 - 4ac}}{2a}\\ &= \frac{\left(\frac{3}{2} + \frac{i}{2}\right) \pm \sqrt{\left(\frac{3}{2} + \frac{i}{2}\right)^2 - 4}}{2}\\ &= \frac{\left(\frac{3}{2} + \frac{i}{2}\right) \pm \sqrt{\left(\frac{9}{4} - \frac{1}{4}+\frac{3}{2}i\right) - 4}}{2}\\ &= \frac{\left(\frac{3}{2} + \frac{i}{2}\right) \pm \frac{1}{2}\sqrt{-8+6i}}{2}\\ &= \frac{3+i \pm (3i+1)}{4}\\ &= \cdots \end{align}$$

Going from line three to line four: $$\begin{align} \frac{\left(\frac{3}{2} + \frac{i}{2}\right) \pm \sqrt{\left(\frac{9}{4} - \frac{1}{4}+\frac{3}{2}i\right) - 4}}{2} &= \frac{\left(\frac{3}{2} + \frac{i}{2}\right) \pm \sqrt{\left(\frac{8}{4}+\frac{3}{2}i\right) - 4}}{2}\\ &= \frac{\left(\frac{3}{2} + \frac{i}{2}\right) \pm \sqrt{\left(2+\frac{3}{2}i\right) - 4}}{2}\\ &= \frac{\left(\frac{3}{2} + \frac{i}{2}\right) \pm \sqrt{-2+\frac{3}{2}i}}{2}\\ &= \frac{\left(\frac{3}{2} + \frac{i}{2}\right) \pm \sqrt{-\frac{8}{4}+\frac{6}{4}i}}{2}\\ &= \frac{\left(\frac{3}{2} + \frac{i}{2}\right) \pm \frac{1}{2}\sqrt{-8+6i}}{2}\\ \end{align}$$

  • $\begingroup$ Why can you replace $t$ with $z$ on the LHS of that quadratic formula, when $t = e^{iz}$? $\endgroup$
    – hello all
    Jul 27, 2015 at 9:33

enter image description here

Two solutions of the system above: (Pi /4 , - ln(2)/2) and (- Pi/4 , ln(2)/2), among others.


One possible "elementary method" might start:

$$\cos(x+iy) = \cos(x)\cos(iy)-\sin(x)\sin(iy),$$

which simplifies to


Now let

$$\cos(x)\cosh(y)-i\sin(x)\sinh(y) \equiv \frac{3}{4}+\frac{i}{4},$$

which gives $$\cos(x)\cosh(y)=3/4$$ and $$\sin(x)\sinh(y)=-1/4.$$

You may be able to solve for $x$ and $y$.

  • $\begingroup$ I tried this approach but I got bogged down since I went off on a tangent using this. Thank you! $\endgroup$
    – 1028
    Jan 26, 2014 at 22:14
  • 1
    $\begingroup$ I highly doubt that a human would be able to solve those nonlinear equations in a reasonable amount of time: see W|A's solutions $\endgroup$
    – apnorton
    Jan 26, 2014 at 22:15

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.