# Is the product rule for logarithms an if-and-only-if statement?

If a function $f(x)$ is proportional to $\ln x$, then we know $$f(xy) = f(x) + f(y).$$

My question is, is the converse true? If we know that, for an unknown function f, $$f(xy) = f(x) + f(y),$$ can we conclude that the function must be proportional to $\ln x$? Why?

• Put $g(x)=f(e^x)$ and read this.
– Pp..
Commented Jan 21, 2015 at 6:48
• Hi. Does the function $f$ continuous? Commented Jan 21, 2015 at 7:38
• @Pp.. Much obliged. Plus I learned about an interesting new class of pathological functions! Commented Jan 21, 2015 at 7:55
• @Farhad, no, it is not explicitly stated to be so. But I deduce from Pp..'s linked article that a sufficient condition on f(x) for the logarithmic solution to be unique is that it be continuous at a single point (or monotonic over any interval or bounded over any interval). Given the context in which the question arose, I feel alright excluding any functions pathological enough not to meet any of those conditions. Commented Jan 21, 2015 at 8:00
• In this post you can find a few links some related equations. This question asks about continuous solutions of your (or Cauchy's) equation. Commented Jan 21, 2015 at 8:10