measuring relative importance/social proximity of an addressee based on length of written explanation in letters with multiple addressees.

[Shuki Cohen]

Dear Colleagues, 

I have a corpus of personal letters in which writers explain a tough decision to multiple addressees, each in its own section/paragraph of the letter (for the most part).

I wonder if anybody has ever empirically measured the attention/production/verbosity associated with each addressee (e.g. number of words) as a measure of the importance to the writer of their understanding the reasoning behind the decision. so far I've seen vague and indirect references to that in the politeness 

literature, pragmatics (e.g. Grice's maxim of quantity), discourse analysis register/style shift, etc. but nothing too empirical, systematic and/or quantitative.

any thoughts are most welcome.

sc