Peter Kirby, over at the Christian Origins blog, has done some work with word frequencies in the Pauline Epistles. There are two posts to review:
I haven't had time to fully digest Peter Kirby's posts yet, I hope to start on that over the next few days. He may just re-open the word-frequency can of worms for me. If I have further comments after looking through what he's done, I'll surely post them here.
As longer-term ricoblog readers know, I've flirted with this area as well. I think P.N. Harrison (Problem of the Pastoral Epistles, Oxford, 1921) did the most work with the problem of unique vocabulary in the Paulines, and I think Harrison's work can be used as fodder for both sides of the argument. Donald Guthrie, in his short monograph The Pastoral Epistles and the Mind of Paul (Tyndale House Publishers (UK): 1954(?)) uses Harrison's own numbers in an effective argument against Harrison's position. That said, there are interesting things possible with word frequencies, but the agreed-upon Pauline corpus is just too small a sample to make any conclusions about particular epistles. At least, that's my conclusion.
One thing I find even more interesting than word frequencies is the frequency or recurrence of three-word phrases in a given corpus. I've used James Tauber's data (thanks again, James ... though you really need to XML-itize it!). A colleague here at Logos coined the term "tri-log" for three-word-phrase (where the lexical form of the word is considered, not the inflected form), so I've used that term, at least for now. My initial forays in this area can be seen here:
I've only generated the data, I haven't analyzed it. My basic hypothesis is that these "tri-logs" (three-word-phrases) may be a better indicator of style or authorship than frequencies of individual words across documents in a given corpus.
Stephen C. Carlson, of Hypotyposeis fame, has done similar stuff (with analysis!) that he's posted to the web. Also recommended to me from at least two different sources (though I haven't obtained a copy or read the work) is A Stylometric Study of the New Testament by Kenny (Oxford: Clarendon Press, 1986). The whole area is interesting to me, but I think I should read Kenny's book and perhaps a few others on style and stylometry before going too much further down this road.
[updated; a few sentences added and a paragraph moved]