To be fair, when you're getting into concurrent structures, you have a fundamental conflict with the OP's criteria. (Scala and haskell will teach you a lot about language design.)
"less of an emphasis on, e.g., doing things "the right way", programming language theory, strict enforcement of rules"
-------------
Some other language i find interesting, tho i haven't written much or any code:
factor, ioke, objective-C, J, common lisp, scheme
------------
sounds like you shd go to emerging languages:
Have you seen how complete Ocaml's coverage on PLEAC is? http://pleac.sourceforge.net/pleac_ocaml/index.html
It's very FP but you can mutate state explicitly if you insist. It runs on the JVM so performance is very good (much faster than Ruby and Groovy) and Java interop is excellent, even better than with Scala. So you have access to the whole wealth of Java library code.
A big advantage is the very strong concurrency support. Laziness and FP give you some multi-core goodness with no effort, and the STM functionality gives you cleaner and more foolproof inter-activity synchronization.
In other words, recommended.
All that said, given a free choice, I use Ruby for most tasks.
It doesn't get as much press for tweakability as Ruby, but you can definitely muddle and muck around in Haskell for the sheer sake of creating something.
Check out #haskell on Freenode. Between the category-theoretical arguments, you'll find plenty of hacking for the sake of hacking.
edit: check out http://www.willamette.edu/~fruehr/haskell/evolution.html and tell me you don't think such a flexible language requires a sense of play.