1
$\begingroup$

Baby Rudin proves in Theorem 8.1 the differentiability of real power series in their interval of convergence by leveraging his Theorem 7.17:

7.17 Theorem
Suppose $\{f_n\}$ sequence of real-valued functions, differentiable on $[a, b]$ and such that $\{f_n(x_0)\}$ converges for some $x_0$ in $[a, b]$. If $\{f_n'\}$ converges uniformly on $[a, b]$, then $\{f_n\}$ converges uniformly to a function $f$ and $$f' = \lim_{n \to \infty}\{f_n'\}$$ on $[a, b]$.

Examining his proof of Theorem 7.17, it seems that the real-valued Mean Value Theorem is used only to get a bound on differences, and so it seems that Theorem 7.17 should generalize to complex differentiable functions on convex open sets in $\mathbb{C}$.

Motivation:
The proofs that I've seen (Papa Rudin, Ahlfors, Spivak calculus book, and Bak/Newman) that complex power series are termwise differentiable seem to be based one way or another rewriting terms $\frac{z^n - w^n}{z-w}$ as sums and then passing to a limit. I was just wondering if there is another path that avoids the direct calculation.

Question:
Am I correctly understanding that this provides a path to showing that complex power series are termwise differentiable inside their disk of convergence?

Thanks

$\endgroup$
3
  • $\begingroup$ This theorem in the complex case is much less interesting as it is very weak since complex differentiability is quite a strong property; you can go that way sure since power series preserve their convergence disk under differentiability of any order, but the main result, namely complex differentiability (just the usual limit definition without any other assumptions) on an open set equivalent to complex analyticity (existence of power series) there needs different proof $\endgroup$
    – Conrad
    Commented May 3, 2019 at 14:24
  • $\begingroup$ @Conrad - Thanks, yeah I know it's not very interesting in the complex case for the reason you mentioned. I've added a hopefully clarifying motivation to the question. $\endgroup$
    – bryanj
    Commented May 4, 2019 at 14:59
  • 1
    $\begingroup$ Yes definitely you can use the result for derivatives, noting that the radius of convergence stays the same. For power series you can always show they satisfy Cauchy by uniform integration on smaller discs and show differentiabillity that way too. $\endgroup$
    – Conrad
    Commented May 4, 2019 at 17:52

1 Answer 1

1
$\begingroup$

Baby Rudin said that they should restrict themselves to real values of x.

enter image description here

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .