My professional relationship with Bill Estes began as a graduate student at Stanford in the years 1964- 1968. I spent my first year with Gordon Bower and the next three years with Dick Atkinson as research advisors, but nevertheless learned much from Bill in classes, research seminars, and discussions. Certainly a central impetus toward the development of the SAM retrieval model took place during a qualifications examination in which Bill asked me how I knew whether the memory effects I was describing arose during memory storage or memory retrieval. Much of the subsequent development of the SAM model and its attendant research has revolved around attempts to deal with this question. Subsequently I spent two different years on leave with Bill at Rockefeller University, and we developed a close friendship as well as a fruitful professional relationship. I have learned much from Bill, about research, science, humor, and life.
This article describes the evolution of the SAM model for memory, summarizing the model as described in Raaijmakers and Shiffrin (1980, 1981) and Gillund and Shiffrin (1984), and its precursors, but focusing primarily upon developments since 1984. It is particularly appropriate for this volume since the second author has recently utilized the idea of stimulus fluctuation theory (Estes, 1955a,b) in an extension of SAM to deal with classical interference phenomena (Mensink & Raaijmakers, 1988, 1989).