Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let $V$ be a connected set in $\mathbb{C}$, and $f$ holomorphic on $V$ such that $$f(z)^2=\overline{f(z)}$$ on $V$.

I want to show that $f$ must be constant on $V$.

My attempt: As $$|f(z)|^2=f(z)\overline{f(z)}=|f(z)||\overline{f(z)}|=|\overline{f(z)}|,$$ it follows that the $\overline{f(z)}\equiv 1$ on $V$. Then it would follow that the imaginary part of $f$ must also be constant since it must be harmonically conjugate to $1$.

Is this ok?

share|improve this question
My \overline's are acting funny. –  Thelonius Aug 29 '12 at 1:14
Here's a hint: can you prove that a function which is both holomorphic and anti-holomorphic (meaning that the conjugate is holomorphic) is constant? –  Akhil Mathew Aug 29 '12 at 1:17
Your proof is essentially right. You should note the possibility that $f\equiv 0$ in $V$ as well. Also, alternatively, you could argue that $|f|\equiv 1$ implies $f$ is constant by the open mapping theorem. –  froggie Aug 29 '12 at 1:18
I don't see how you got $\overline{f(z)}=1$. What I would do first is ask myself: which complex numbers have the property that the square of the number equals its conjugate? –  Gerry Myerson Aug 29 '12 at 1:27
I meant rather that $|f(z)|\equiv 1$. Did I make an error? –  Thelonius Aug 29 '12 at 1:39

3 Answers 3

up vote 5 down vote accepted

Why isn't anyone taking my hint from the comments?

  1. If $z^2=\overline z$ then $z=0,1,e^{2\pi i/3}{\rm\ or\ }e^{4\pi i/3}$.

  2. If a continuous function takes on only finitely many values on a connected set, it's constant.

share|improve this answer
Thanks, this is a good answer. I would have been happier if someone pointed out the flaw in my own argument (modulo my typo noted in the comments above), but this is certainly a nicer argument in any case. –  Thelonius Aug 29 '12 at 13:14

Since $f(z)$ is holomorphic, if $f=u+iv$, its real and imaginary parts satisfy Cauchy-Riemann equation, i.e. $$\tag{1}u_x=v_y\mbox{ and }u_y=-v_x.$$ Since $f(z)$ is holomorphic, $\overline{f(z)}=f(z)^2$ is also holomorphic. Since $\overline{f}=u-iv$, Cauchy-Riemann equation implies that $$\tag{2}u_x=-v_y\mbox{ and }u_y=v_x.$$ Combining $(1)$ and $(2)$, we have $v_y=0$ and $v_x=0$ on the connected set $V$, which implies that $v$ is constant on $V$. Since $v_y=0$ and $v_x=0$, we have $u_y=0$ and $u_x=0$ on $V$, which implies that $u$ is constant on $V$. Therefore, $f=u+iv$ is constant.

share|improve this answer
It is not immediate to me that you can claim that $\overline{f(z)}$ is holomorphic on $V$ unless you know something about $V$ (such as being open, for example). For example, $f(z) = |z|$ equals the holomorphic function $z \mapsto 1$ on the unit circle, but is clearly not holomorphic. –  copper.hat Aug 29 '12 at 4:33
Since $f(z)$ is holomorphic on $V$ by assumption, $f(z)^2$ is also holomorhpic on $V$. By assumption, $\overline{f(z)}=f(z)^2$. So $\overline{f(z)}$ is holomorphic on $V$. Am I correct? –  Paul Aug 29 '12 at 4:53
Well, I don't think it follows just from equality on $V$, unless $V$ is, for example, open. See my example above of a holomorphic function equaling a non-holomorphic function on a particular set. –  copper.hat Aug 29 '12 at 4:55
This is certainly a good idea if we assume $V$ is open, though. –  Potato Aug 29 '12 at 6:24

Here is a proof that only requires $f$ to be continuous.

Since $|f(z)^2 |= |f(z) |^2=|f(z)|$ it follows that $|f(z)| \in \{0,1\}$ for all $z \in V$. Since $V$ is connected and $f$ is continuous, it follows that either $|f(z)| = 0$ on $V$ or $|f(z)| = 1$ on $V$.

If $|f(z)| = 0$ on $V$, then clearly $f$ is constant on $V$, so assume that $|f(z)| = 1$ on $V$.

Consider $f(z)^3 = |f(z)|^2$. Since $|f(z)| = 1$, we have $f(z)^3 = 1$ for $z \in V$. Since the cube roots of unity form a disconnected set, it follows that $f(z)$ is constant (equal to one of the cube roots of unity, of course) for $z \in V$.

share|improve this answer
It occurs to me that this was essentially a long-winded way of following @Gerry Myerson's suggestion above. –  copper.hat Aug 29 '12 at 4:29

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.