A review of Memory Models from Global to Bayesian

Abstract

In this paper, the theoretical and empirical foundations of some of the most important memory models are explored. An inportant stage in memory research was the development of "Global Memory Models". In these models, judgements about single events are based on comparing all memory traces in memory to the probe. The three most important global memory models are SAM (Gillund and Shiffrin, 1984), TODAM (Murdok, 1982), and MINERVA2 (Hintzman, 1986, 1988). This paper is organized chronologically: first data and models before the advent of global memory models are reviewed; then global memory models are discussed; and finally some of the data and models that are inconsistent with global memory models are discussed along with contemporary directions in memory modeling like Baysian Memory Models.