0
$\begingroup$

In Riemann's publication about Abelian functions 'Theorie der Abelschen Functionen' (Here the original paper in german) at the end of Chapter 4, part 2 is clamed that for every Riemann surface $T$ and $ a \in T$ every meromorphic multivalued (?) function $f: T \to \hat{\mathbb{C}}$ with $f(a)= \infty$ can be locally expanded via

$$g + A \cdot \log \ r +B r^{-1}+ C r^{-2}+ \ldots $$

where $g$ is finite & continuous in $a$ (so 'nice' enough), the values $A, B, C,\ldots$ are appropriate complex coefficients and the function $r: T \to \mathbb{C}$ is the so called infinitely small function of first order. In modern language of complex analysis this $r$ is nothing but the local parameter in $a$, that is locally near a point $a \in T$ the function $r$ behaves like $z$ near $0$ on the plane $\mathbb{C}$. So for sake of simplicity say $T= \mathbb{C}$ and $f: \mathbb{C} \to \hat{\mathbb{C}}$ is multivalued with $f(0)= \infty$.

Why locally $f$ can be always writen as

$$ g(z) + A \cdot \log \ z +B z^{-1}+ C z^{-2}+ \ldots $$

with $g(z)$ finite in $0$ and continuous?

Riemann wrote that this ensues from known theorems on expansions of functions in power series due to Cauchy.

Unfortunately I nowhere in literature found a result on this kind of expansion. Does anybody know it? Especially the presence of $\log z$ confuses me. In this expansion $\log z$ is the only term which is multivalued and the existence of such expansion near $0$ suggests that every kind of multivaluedness of a function with $f(0)= \infty$ can be 'absorbed' by $\log z$, that's strange and I never saw such result in any book.

From modern complex analysis it is known that there are for meromorphic functions three kinds of singularities: the removable one, pole of finite order and the essential singularity. And in any modern book on complex analysis no one of these three kinds of singularities is associated with an expansion containing the $\log z $ term. Does anybody know the ancient result by Cauchy to which Riemann refer here and has it a modern analogue?

Here the relevant translated excerpt from Riemann's paper:

enter image description here

$\endgroup$
8
  • 6
    $\begingroup$ I repeat my advise: learn the basic notions from some modern Complex variables book, and only AFTER that read Riemann. His terminology and the manner of exposition are out of date. Once you have some experience with the subject, it is easy to translate what he says to the modern language. Riemann did not write for the beginners. And his terminology is outdated. It is not an easy reading for a modern student. $\endgroup$ Commented Mar 20, 2021 at 1:31
  • $\begingroup$ @AlexandreEremenko: maybe I'm running to overestimate my knowledge on this topic, but I think that the most principles which one can look up in a basic modern complex variables book are at least roughly familar to me. So I know at least basic notions. About the advise you gave me in my last question here mathoverflow.net/questions/386700/… you are right, that was almost basic. A local parameter on a RS is exactly what Riemanns calls infinitely small function of first order. $\endgroup$
    – user267839
    Commented Mar 20, 2021 at 1:48
  • 1
    $\begingroup$ But concerning this question I don't agree with you. I looked in several books on this topic and I nowhere found a treating of the problem I described above. Honestly, I don't think that this treating of the singularity using expansion containing this log term & it's meaning there is standard and something 'that can be easy translated in modern language'. If you think that I wrong then I would appreciate if you could give a hint where I could look up it. $\endgroup$
    – user267839
    Commented Mar 20, 2021 at 1:49
  • 1
    $\begingroup$ Perhaps, in context, he has reason to assume the derivative of the function is single-valued (and has no essential singularity)? Then the derivative would have a Laurent series expansion, which, integrating, would give an expansion of this form. $\endgroup$
    – Will Sawin
    Commented Mar 21, 2021 at 15:23
  • $\begingroup$ @Will Sawin: That's a good suggestion, possibly I somewhere overlooked that Riemann maybe somewhere explicitely assumed that he deals only with multi-valued functions with single-valued derivative. I don't know the historical background but possibly Riemann studied during his lifetime when he 'invented' Riemann surfaces only multi-valued functions with single-valued derivative, I don't known, but of course the expansion above make is always given only with multi-valued functions with single-valued derivative. $\endgroup$
    – user267839
    Commented Mar 23, 2021 at 5:12

0

You must log in to answer this question.