It might be an unpopular position to take here, but I really think that what I think of as "language snob languages" (Lisp, Haskell, Erlang, OCaml . . . not Ruby or Python, though) can't ever be mainstream because the barrier to entry is just too high even for competent programmers. There's a network effect problem to be sure, where people don't use it because the docs/frameworks/examples/toolkits aren't there, but part of the reason those things aren't there is that the language doesn't appeal in the same way to people who want to build those things and the languages themselves have enough of an impedance mismatch with how people think about things to keep a lot of otherwise-competent engineers from investing their time in them.
In other words, it's not just about tools or marketing or docs: some languages, while certainly more useful in the hands of a skilled expert, are so complicated and require so much dedication to perfecting their use that they can't possibly ever become mainstream.
The high barrier to entry for those languages is not just because of docs, tools, lack of standards, etc. It's just an inherent part of those languages, and what makes them so powerful is exactly what makes them difficult to get started with and what keeps them marginalized.