Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

Let $L:H\rightarrow H$ be a continuous linear operator and $R_n:H \rightarrow H$ a sequence of continuous linear operators, where $H$ is a Hilbert space. If the $\sum_{n=1}^{k} R_n$ converge pointwise to $L$, meaning for every $x\in H$ we have $$Lx=\sum_{k=1}^{\infty} R_k x=\lim_{k\rightarrow \infty}\sum_{n=1}^{k} R_nx,$$ is then do we know then, that the operator $T$ defined by $$ Tx=\sum_{k=1}^{\infty} R_kx,$$

1) equals L ? 2) is continuous ? 3) convergences in the operator norm to $L$ ?

( For 1) and 2) with the help of Asaf Karagila, I think the answer is "yes", since for 1) $L$ and $T$ agree for every value and for 2) we know that $L$ was continuous, so using 1), also $T$ has to be continuous, right? )

EDIT: Everything has been cleared except that I need a counterexample for 3)

share|cite|improve this question
You don't have to assume that $L$ is continuous, it follows automatically from the uniform boundedness principle. – Ahriman Aug 24 '12 at 9:20
What do you mean with you question 3) ? Do you want to know if $\sum_{k=1}^n R_k$ converges in operator norm to $L$ ? This in not necessarily true, there are surely counterexamples in textbooks. – Ahriman Aug 24 '12 at 9:22
@Ahriman Yes, that's what I meant. I just thought that would be true, because convergece in the operator norm would imply continuity of $L$, so I thought that maybe the "converse" would hold, i.e. if $L$ were continuous, it would converge to it in the operator norm. (cont) – user10324 Aug 24 '12 at 10:24
(cont) So 1) and 2) are true ? A counterexample by the would also be a (negative) answer... – user10324 Aug 24 '12 at 10:25
Yes 1) and 2) are true, I don't even see where there could be a problem. – Ahriman Aug 24 '12 at 10:49
up vote 1 down vote accepted

Here is a simple counterexample for 3) :

Suppose your Hilbert space $H$ is infinite dimensional and separable, meaning it has an infinite countable orthonormal complete system.

Let denote such a system by $(e_n)_{n \ge 0}$. We know, from Parseval's identity, that every $x \in H$ is equal to $\displaystyle \sum_{n \ge 0} \langle x, e_n \rangle e_n$.

Let $R_n$ defined by $R_n x = \langle x, e_n \rangle e_n$. Then $\sum_{k=0}^n R_k$ converge pointwise to the identity operator.

But if the convergence occurs for the operator norm, since every operator $\sum_{k=0}^n R_k$ is of finite rank, the limit would be a compact operator. But the identity is not compact by Riesz theorem, since $H$ is not finite dimensional.

The special case $H = \ell^2(\mathbb{N})$ with the complete orthonormal system given by $e_n(k) = \delta_{n,k}$ doesn't even require the use of Parseval's identity.

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.