Let's throw a bit of theory into this, shall we:
Computational processing of textual data
Interesting parts:
Part-of-Speech tagging (Brill, HMM)
Part-of-Speech tagging (HMM ctnd.)
Then I recommend "Parsing, formal grammars" and "Stochastic Parsing", as well as "Classification - Visualization". As long as we don't have the same theoretical background, there's no use discussing anything. Personally, I took computational processing of textual data and "natural language processing" (audio) and some opinions/beliefs in this thread amuse me.