I don't think FPLs by themselves are (Turing) tar pits. On the other hand, there are lambda-calculus tar pits - for example SKI combinators. This shows that eve

I don't think FPLs by themselves are (Turing) tar pits. On the other hand, there are lambda-calculus tar pits - for example SKI combinators. This shows that even a language that has referential transparency and lazy evaluation can be a tar pit; so these qualities by themselves are not a way to avoid complexity. I hope one day we will be able to mathematically formulate what is "unnecessary complexity".. but so far, no idea.