It is known that Birkhoff's pointwise ergodic theorem (unlike von Neumann's mean ergodic Theorem) fails to hold for general Folner sequences. The counter-example usually given is the Folner sequence $F_N=\{N^2,N^2+1...,N^2+N\}$, however I've only seen it referenced to the first such result by Akcoglu and del Junco in 1975, where the Folner sequence which fails is actually $G_N=\{N,N+1,...,N+\lfloor \sqrt{N}\rfloor\}$. Since $\{F_N\}$ is just a (quite sparse) subsequence of $\{G_N\}$ it could happen that the ergodic averages converge pointwise along $\{F_N\}$ even if they don't along $\{G_N\}$.

Moreover, Lindenstrauss showed that every Folner sequence has a subsequence along which convergence holds.

So I'm just curious if there is a published proof (or if it is easy and I'm just overlooking something) that the sequence $\{F_N\}$ is not good for pointwise ergodic theorem.


Bellow, Alexandra(1-NW); Jones, Roger(1-DPL); Rosenblatt, Joseph(1-OHS) Convergence for moving averages. Ergodic Theory Dynam. Systems 10 (1990), no. 1, 43–62.

  • $\begingroup$ Thanks! I couldn't find that paper online, but following citations I found what I was looking for in Butler, S. V.(1-IL); Rosenblatt, J. M.(1-IL) Moving averages. Colloq. Math. 113 (2008), no. 2, 251–266. $\endgroup$ – Joel Moreira Jan 15 '13 at 17:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.