Skip to main content
10 of 12
edited title
T. Amdeberhan
  • 43.2k
  • 5
  • 57
  • 217

modulus-related analytic functions

Let $D\subset\mathbb{C}$ be the open unit disk. Suppose $f,g,F,G:D\rightarrow\mathbb{C}$ are analytic functions related by $$\vert f(z)\vert^2+\vert g(z)\vert^2=\vert F(z)\vert^2+\vert G(z)\vert^2; \qquad \forall z\in D.$$

Question 1. If $f\neq \alpha g$ and $g\neq\beta f$ for any $\alpha, \beta\in\mathbb{C}$ then is the same true for $F$ and $G$?

Caveat. Not true for real analytic functions: take $f=\sin x, \,g=\cos x, \, F=\frac1{\sqrt{2}}=G$.

EDIT. Encouraged by Christian's positive answer, let's upgrade the problem.

Question 2. Suppose $f, g, h$ are linearly independent (over $\mathbb{C}$) such that $$\vert f\vert^2+\vert g\vert^2+\vert h\vert^2=\vert F\vert^2+\vert G\vert^2+\vert H\vert^2.$$ Should $F, G, H$ be linearly independent, too?

T. Amdeberhan
  • 43.2k
  • 5
  • 57
  • 217