RMS used to publish academically when he was at MIT. I remember reading a paper about Prolog backjumping written by him which had nothing to do with software licensing.