For those with a strong stomach, the article that Reagle refers to is here
http://www.thoughtcrumbs.com/publications/328-burke.pdf It begins with the astounding and thought-provoking observation that people like to present and identity with 'positive social value' and want to have that identity validated by others. But they may be presented with face-threatening criticisms, so engage in "face-work" when communicating to help maintain each other's identitities. Thus instead of saying 'take out the trash' (which might cause the person to lose face) you say 'Would you please take out the trash'. What an insight!! And the authors, in that brief paragraph, manage to refer to two other papers, one of which is Goffman's Theory of Politeness. That is the reality of social sciences generally, a bunch of stupid and vacuous people writing papers for crappy journals, and citing other equally vacuous papers, which in turn cite other .... You never get to anything real. I'm speaking as one who has a higher degree in the subject (I realised halfway through that it was complete nonsense but it was too late by then and I had to complete, although I did successfully refuse to submit one piece of required coursework on the grounds of its utter banality and stupidity - the examining committee surprisingly agreed and I was passed).
Then there is a 'typology of linguistic politeness strategies'. Papers like this always have a 'typology' despite the fact there is no such word in English. I was asked last year to deliver a lecture which had the word in the title and I made them change it to 'types'.
Then the authors build a 'model of linguistic politeness' using messages from some forum. To measure the 'impact' of politeness they performed 'a negative binomial regression' on the number of replies each message received. Well that is a sort of impact but it is utterly meaningless. As everyone who uses these forums knows (a) the intellectual level at which they are conducted is practically zero anyway, so it is not worth writing a paper about in the first place (b) the number of replies in no way reflects the quality of the message, quite the reverse, the more dim-witted and moronic and stupid the message, the more replies. The replies will be a mixture of the more intelligent members pointing out how stupid the message is, with those equally stupid or even more stupid members agreeing wholeheartedly with the idiotic sentiments expressed in the message. Anyone who follows AN/I will certainly know this.
The real question is whether politeness or 'civility' has any impact, using an appropriate measure of 'impact'. In Wikipedia's case, the only sensible and appropriate measure would be whether the message contributes to the goal of a universal, comprehensive and reliable reference work. But how is anyone going to build a computer model to assess a universal, comprehensive and reliable reference work? Anyone able to do that, would have solved the problem of humans getting involved in the first place.
This post has been edited by Peter Damian: