> I am pretty certain “Master” in Git is not derived from master/slave duality.

It apparently comes from copying bitkeeper, which uses master/slave, so it is, indirectly, derived from master/slave.

It's easy enough to interpret (and retain) in the artistic sense of a master copy (which usage is not derived from master/slave), so even with it's etymology it is far from the most problematic use of “master”, but the git usage is not historically unconnected to master/slave.