ABSTRACT
Ben Murdock's approach to the scientific study of human memory is founded on the belief that solid replicable data must provide the foundations for the theoretical superstructure. Throughout his career he has provided the research community with such empirical cornerstones, and his interests have now turned to providing us with a comprehensive theory of memory—TOD AM—which is widely respected and highly successful. Recently, Lewandowsky and Murdock (1989) have shown how the model can account for a wide range of phenomena connected with memory for serial order. Taking as their benchmarks the four paradigms of serial learning, memory span, partial report, and the Brown-Peterson paradigm, the authors, in an article of extraordinary lucidity (meaning that, to my amazement, I understood most of it!) show how TOD AM can generate excellent fits to the empirical data with minimal juggling of parameters. A main point of the article is that the model fits serial order data regardless of whether the paradigm involves immediate reproduction or longer-term cumulative learning. However, most of the data cited fall into the traditional realm of short-term memory studies, and indeed one could argue that Murdock's major contributions have been in that area. As many of my own earlier experiments addressed issues of short-term verbal memory, it was a great pleasure both to remind myself of Ben's current views and to read the four chapters on short-term memory issues in the present section.
