[EDIT]: Following Qiaochu Yuan's comment, it is better to clarify that I do not know what the right definition of a fractal in the following question should be. But a nice answer might contain such a definition. Moreover, a nice comment by Joël correctly points out that choosing which theory of rigid analytic spaces one wants to consider is crucial for my question which is otherwise ill-posed (but, again, I am more than happy if someone has an answer using whichever theory she prefers).

The infinite fern of Gouvêa-Mazur, introduced (I guess) in the paper by Barry Mazur "An "infinite fern" in the universal deformation space of Galois representations" is a subset of the space attached to a universal deformation ring to a certain $\mod{p}$ Galois representation and which consists of infinitely many paths crossing at infinitely many points (there is a picture on page 36 of Mazur's paper) and given that this infinitely many crossing are dense in each path (they roughly correspond to integers in a $p$-adic disk) it seems to me that the above set could naturally be regarded as having some fractal behaviour. So, my questions are: is this the case? If yes, are there any results or conjectures on its fractal dimension, or connections between its fractal structure and some arithmetic of $p$-adic families of modular forms?

  • 6
    $\begingroup$ What does "is a fractal" mean here? $\endgroup$ Jun 1, 2013 at 5:44
  • 1
    $\begingroup$ @Qiaochu: Well, good point. I do not know and it seems to me - as non-expert!- that in general there is no such a uniquely agreed definition. Most probably, I am thinking at something like "if you cut a piece of the infinite fern into $n$ equal pieces, each looks like $1/m$-th of the original piece with $n/m\notin\mathbb{Z}$", but I am more than happy if the answer contains a precise definition ;-) $\endgroup$ Jun 1, 2013 at 7:22
  • $\begingroup$ A possibility for a definition of "fractal" is "metric space with fractionary Hausdorff dimension". $\endgroup$
    – Qfwfq
    Jun 1, 2013 at 12:04
  • $\begingroup$ @Qfwfq - I don't think that's a good definition of fractal, there are metric spaces, even subspaces of $\mathbb{R}^2$, which have integer Hausdorff dimension and are quintessentially fractal. See my comment here for an example: mathoverflow.net/questions/56677/… (there's a typo, should be side length 1/4 rather than 1/2). $\endgroup$ Jun 1, 2013 at 14:08
  • 1
    $\begingroup$ I think the question is interesting, but somehow is still not well-posed. There is the problem of what is a fractal, discussed in earlier comments, but also the problem of what is the "infinite fern". I mean, it is a rigid analytic varieties, but there are at least three ways to formalize this notion, Tate's and Berkovich's and Huber's. Each leads to a different space.. $\endgroup$
    – Joël
    Jun 3, 2013 at 13:51

1 Answer 1


I post this as an auto-answer mainly not to leave the question open.

After googling a bit better, I discovered two recent works by M. Lapidus and L. Hung (both available on Lapidus' webpage )

  • “Nonarchimedean Cantor Set and String”, Journal of Fixed Point Theory and Applications 3 (2008), pp. 181-190, (Special issue dedicated to Vladimir Arnold on the occasion of his Jubilee. Vol. I.)

  • “Self-Similar $p$-Adic Fractal Strings and Their Complex Dimensions”, $p$-Adic Numbers, Ultrametric Analysis and Applications (Russian Academy of Sciences, Moscow, and Springer-Verlag), No. 2, 1 (2009), pp. 167-180.

which seem to pose a "good" definition for a $p$-adic fractal. The definition follows the usual self-similarity one, attaching to each family $\{\Phi_1,\dots,\Phi_n\}$ of similarity contractions $$ \Phi_j:\mathbb{Z}_p\longrightarrow\mathbb{Z}_p $$ the unique non-empy, compact, fixed subset $\mathcal{S}\subseteq \mathbb{Z}_p$ such that $\mathcal{S}=\Phi_j(\mathcal{S})$ for each $1\leq j\leq n$. They develop some theory for such objects and define its Minkowsy dimension, mainly following the box-counting dimension definition . In a closing remark of the second paper, they also say that "it would be interesting to generalize this theory from subspaces of $\mathbb{Q}_p$ to Berkovich spaces" but I was unable to find anything more on the subject.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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