Notes on the History of Ideas Behind AGM

Abstract

The purpose of this paper is to present a part of the history of ideas that led up to the formulation of the AGM theory [1]. David Makinson [26] has already written an excellent survey of what was new in the theory. Some historical details are also included in his obituary of Carlos Alchourron [25]. Here I will present the story from my own perspective. During the academic year 1973-74, my last period as a Ph.D. student was spent in Princeton. Among other things, I participated in Carl Hempel's seminar on explanation. He made it to a fascinating topic and after my dissertation, I started working in the area. I soon became dissatisfied with the purely logical analyses that were used in the covering-law accounts of explanation. Bengt Hansson [16] wrote a mimeographed paper with the title "Explanations of what?", which focused on the idea that explanations must be judged relative to a particular knowledge situation. I continued along the same lines and it resulted in my paper "A pragmatic approach to explanations" [7]. The key idea of that paper was that an explanandum E that was known in a knowledge situation К should be evaluated with respect to a knowledge situation K' in which E was not known.1 In retrospect, it can be said that this condition presumes that one can identify contractions of belief sets. I then began looking for a way of characterizing such contractions. In Princeton I also worked with David Lewis. In particular, I became interested in the logic of conditionals. Since I am not a modal realist, I was dissatisfied with a semantics of conditionals that is based on possible worlds. Again, I wanted to evaluate the validity of a conditional in relation to the beliefs of a person asserting