Looking for papers on explanation for constraint satisfaction I ran across Meta-explanation in a Constraint Satisfaction Solver by the pioneering French AI scientist Jacques Pitrat. From the abstract:

Meta-explaining is useful to a human or an artificial researcher in order to learn how to use its knowledge efficiently. For instance, when the system RESEARCHER solves the same prob- lem using two different methods, it can meta-explain why it finds a more elegant solution with one of them. 

It might be interesting to compare Pitrat’s work here to the work I did with Chavalit Likitvivatanavong and Rick Wallace in Deriving Explanations and Implications for Constraint Satisfaction Problems.

Pitrat’s paper is related to a very ambitious — and very “meta” — research agenda that he was pursuing. See A Step toward an Artificial Artificial Intelligence Scientist, Artificial Beings: The Conscience of a Conscious Machine, and his blog. Unfortunately Pitrat passed away recently; the record of a day of tribute can be found here.