CRITIQUE OF THE RETRIEVAL DEBLURRING ASSUMPTIONS OF THE THEORY OF DISTRIBUTED ASSOCIATIVE MEMORY/

This note critiques the retrieval assumptions proposed by Lewandowsky and Murdock (1989) in their application of Murdock's theory of distrib uted associative memory (TODAM) to problems in serial-order recall. Tw o different methods for ''deblurring'' the products of retrieval are d escribed, along with simulations appropriate for each. The authors dem onstrate that the deblurring assumptions used by Lewandowsky and Murdo ck ( 1989) in their simulation version of the model, although appropri ate for some aspects of serial-anticipation learning, provide an inade quate account of general serial recall and, in fact, predict several t rends that are inconsistent with known data. The authors also briefly consider the analytic version of the model and note some additional pr oblems. Overall, the deblurring assumptions critically affect the perf ormance of TODAM, and it is these assumptions rather than the model pe r se that determine a sizable proportion of its behavior.