MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

It is well known that under suitable conditions, a function which is:

  1. a polynomial in each variable separately is a polynomial in all its variables jointly.

  2. a rational function in each variable separately is a rational function.

  3. a holomorphic function in each variable separately is holomorphic in all its variables.

A complete analytic function can be single-valued or multiple-valued according as it does not have, or does have, branch points. The algebraic functions are examples of the latter.

Here is my question: is a complete analytic function, which is finitely multiple-valued in each variable separately, also finitely multiple-valued jointly?

share|cite|improve this question
Isn't the function x+y single-valued in each variable separately, but takes any value infinitely often? Or am I misunderstanding your definition? – user3953 Feb 11 '10 at 16:30
A multi-valued analytic function is a function like the logarithm which can take multiple values at a particular point depending on branching. x+y takes a single value at any particular point. – Qiaochu Yuan Feb 11 '10 at 16:56
x+y is single-valued as a function of (x,y) question asks if for each $(x,y_0)$, $f(x,y_0)\equiv g(x)$ is finitely multiple-valued as a function of $x$, and $f(x_0,y)\equiv h(y)$ is a finitely multiple-valued function of $y$, then is $f(x,y)$ a finitely multiple-valued function of the pair $(x,y)$? If, for example, one writes "algebraic" in place of "finitely multiple-valued", then under certain conditions the joint function is algebraic. – Mark B Villarino Feb 11 '10 at 17:01
I don't understand: if for each $x$, the function $f(x,-)$ is finitely-valued, then equivalently for each $(x,y)$, $f(x,y)$ takes only finitely many values, but this is the statement you want, right? – Theo Johnson-Freyd Feb 11 '10 at 17:08
Just to check my understanding, by finitely multivalued, you mean that the function takes only a finite number of values at each point? Then in one sense the answer is trivially “yes”. In another sense, a “no” answer would mean that, even though $(f(x_0,y_0))$ could take an infinite number of values, if you were to change only one variable in a continuous manner (as in moving around a branch point) and returning to $(x_0,y_0)$, you can reach only a finite number of the values of $(f(x_0,y_0))$, whereas you can reach an infinite number by moving both coordinates together? – Harald Hanche-Olsen Feb 11 '10 at 17:11

Engineer a function that has n>=2 values exactly in the set $S_n \backslash S_{n+1}$ where $S_n$={(x,y) | x>n, y>n, (x-n)*(y-n)>1}, and 1 value elsewehere. That should work as a counterexample, as far as $C^{\infty}$ functions go - analytic, not sure

share|cite|improve this answer

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.