This is certainly an interesting exercise (at least for me) but before getting into the meat of my remarks, I'd like to question you about your methods and sources. First of all, where did you get these words from? How big was the text source? What were your criteria for defining what "word" means? (I see "won't" and "don't" in the list, which most people would say are two words :) )

I've noticed that most of the keywords in the list are action verbs. That seems to follow since programming is about giving a machine instructions to do something. And considering that these instructions are supposed to be on a fairly basic level, it's also not surpising that the action verbs used as Perl keywords are highly frequent.

Actually, this makes me recall a thread started by liz about the Natural Semantic Metalanguage (NSM), which proposes that word meaning can be decomposed into atomic units of meaning called primes. Research bears out that the primes tend to have a lot in common with very frequent words in most languages. The semantic primes, however, have a special meaning that may not exactly correspond to the meaning of the word in common use, even though it may have the same name for the convenience of the linguists who use NSM to describe meaning. This is a direct parallel of programming language designers' choices about which actions, relationships, or evaluations need to be expressed as a word, and which should be expressed as a sigil, or via syntax.