Just some random thoughts here based on the paper I'm wrapping up for the SBL meeting.

How does one go about quantifying and measuring style? Style seems to be a very organic thing, hard to pin down. Efforts to quantify it normally fail, on the whole, because there are no real rules to follow.

Some detect style based on word usage patterns (common or infrequent words). Others look into other grammatical features such as morphology. Still others look at sentence lengths and sentence length distribution. Some have posited authorship traits dealing with the part-of-speech of the last word in a sentence.

A trend of some sort is noticed, that trend is tested against a larger corpus, data on the trend is analyzed, and that analysis becomes the basis of some posited rule or even conclusions about authorship of a document.

The bottom line is that trends do not a rule make. Correlation does not prove causation. But that isn't to say trends aren't useful. It only means we need to consider an expanding amount of such trends in evaluation of style. (Note well: I didn't say authorship, I said style.)

Also, I think that, at least as regards the Pauline epistles and perhaps the New Testament, stylistic studies have unfortunately been conflated with authorship studies. That is, the goal seems to be to isolate an author's style for purposes of authorship attribution—not necessarily to better understand the document content and structure so as to better comprehend the document(s) in question.

Why is this? Why can't style be examined outside of the bounds of the authorship question?

Why can't we be content to analyze what is being communicated and how it is being communicated; why do we (myself included!) get bound in the larger and likely unresolvable authorship question?