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

Can you give an example of a sequence of continuous functions $f_n:[0,1]\to [0,1]$, such that $f_n\to 0$ pointwise and there is no subsequence $(f_{n_k})$ for which $\frac 1 m\sum_{k=1}^{m}f_{n_k}$ tends to zero uniformly?

I think it's the same as asking whether the Banach space of continuous real valued functions has the "weak Banach-Saks property", but I was unable to find out the answer.

share|cite|improve this question
@JesseMadnick: not much. I was hoping someone would just know the answer... – KotelKanim Feb 9 '12 at 17:56
Did you try some triangular wave which increases frequency but such that $f_{n + 1}$ has the same peaks as $f_n$ but more of them? – Jonas Teuwen Feb 9 '12 at 18:27
[What have you tried so far?] not much. I was hoping someone would just know the answer... – Did Feb 9 '12 at 20:03
It was somewhat rhetoric, @Didier. (if you are replying on my comment) – Jonas Teuwen Feb 9 '12 at 21:20
@Jonas: Your comment is standard. – Did Feb 9 '12 at 21:48

That $C[0,1]$ fails the weak Banach Saks property appears as problem 17 in chapter VII in Joseph Diestel's Sequences and Series in Banach Spaces.

The problem gives an outline:

1) For $k$ a fixed positive integer, construct a nonnegative sequence $(g_n^k)_n$ in $B(C[0,1])$ satisfying:

  • $g_n^k(t)=0$ if $t\notin ((k-1)/k, k/(k+1))$.
  • $(g_n^k)$ converges pointwise to 0 on $[0,1]$.
  • If $n_1<n_2<\cdots<n_k$, then there is an $a\in[0,1]$ with $g_{n_1}^k(a)=g_{n_2}^k(a)=\cdots=g_{n_k}^k(a)=1$.

2) Define $f_n=g_n^1+g_n^2+\cdots+g_n^n$. Show that:

  • $(f_n)$ is weakly null in $C[0,1]$ (which is equivalent to saying $(f_n)$ converges pointwise to 0 and is bounded).
  • If $n_1<n_2<\cdots<n_m<n_{m+1}<\cdots< n_{2m}$, then $(f_{n_1}+\cdots f_{n_{2m}})(t)\ge {1\over 2}$ for all $t\in[0,1]$ (I believe Diestel has a typo here, it should be $\ge m/2$ for some $t\in[0,1]$).

This was first proved by J. Schreier in Ein Gegenbeispiel zur Theorie der schwachen Konvergence, Studia Math. 2 (1930), 58–62. H. P. Rosenthal's article in volume 2 of Handbook of the Geometry of Banach Spaces (proposition 4.21, page 1585) may also be helpful.

share|cite|improve this answer
The question did say $f_n : [0,1] \to [0,1]$ so they are bounded in $C[0,1]$. (I missed this at first also.) – Nate Eldredge Feb 10 '12 at 1:18
@Nate Ah! I should have seen that... – David Mitra Feb 10 '12 at 1:22
How is $(\frac12, n)$ in $[0, 1]$ for $n \geqslant 1$? – Jonas Teuwen Feb 10 '12 at 8:44
@Jonas I initially misread the question, and thought the OP wanted only pointwise convergence. – David Mitra Feb 10 '12 at 13:30

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.