The paper is an interesting attempt at quantifying politeness when it comes to requests. They used MTurk to annotate requests on talk pages on Wikipedia and Stack Overflow questions and answers. From the annotated data, they were able to build a classifier that could annotate new request texts with close to human level accuracy.
The analysis on Wikipedia that editors who are more polite are more likely to be promoted to admins. But the question now is, what can be done in order to make sure someone continues to be polite, even after gaining power? More generally, what incentive system can be built to prevent the someone’s power from getting to their head? We already have the obvious checks and balances like banning someone even if they are an admin if they become too disruptive. But what about preventing even the small devolution that was observed? Simply stripping the privileges one gained at the slightest sign of impoliteness would be a bad idea, for sure. We could think about implementing the trained politeness classifier as a browser extension (like Grammarly [1] but for politeness) to tell you what your politeness levels are for what you are typing. But this might end up being suffocating to the user who has to deal with seeing this application that they are not as polite as they should be. And also, as Lindah [2] pointed out, the classifier is far from perfect, so this might not be a good idea either.
This isn’t a very shrewd insight or anything but I may as well point out that the classifier is obviously skewed to the U.S. understanding of what politeness is. The authors were upfront about how they chose their annotators but this does mean that the final results end up with marking a request as impolite or polite based on politeness in the context of the USA. Though the biggest parts of Stack Overflow and Wikipedia are in English, people from all over the world do end up contributing to it. What someone from the US may consider impolite (like the ‘Direct question’ strategy) would seem perfectly polite to a Scandinavian, as their culture is one of directness in their speech. Any future work that builds on this one must keep this in mind.
All in all, I believe this was a pretty solid paper. They set out to do something and they did it, documenting the process. Potential future work would be to take this idea and redo it in a different English speaking culture, to identify how their ideas of politeness differ from the US perspective.
[1] https://www.grammarly.com/
[2] https://wordpress.cs.vt.edu/cs5984/2018/10/07/conversational-behavior/
