Recognition memory is the type of memory in which people can recognize whether or not they have seen an item or experienced an event before. It is a basic form of memory, essential for normal cognitive function; its absence is a sign of very severe pathology. Recent work has shown that normal recognition memory is characterized by strong regularities that permit the analysis of its underlying psychological mechanisms. One such regularity is that anything that affects performance on old items, items seen before, such as increased study time or forgetting, also affects performance on new items, those not seen before, in exactly the same way. For example, if forgetting is produced by increasing the time from presentation of old items to test, an individual performs more poorly on the old items, as one might expect, but also performs more poorly on new items; this is unexpected, since the new items will never have been seen before the test and might be expected to be unaffected by a delay preceding the test. The symmetry of the effect indicates that the individual performs (though almost certainly not consciously) a complex analysis of the state of items, an analysis of the sort expected by a particular set of decision theories, including signal detection theory. This research will explore this regularity experimentally, as well as related regularities, and develop a mathematical description of the underlying decision process that produces the regularities.
