ABSTRACT
The relationship between recall and recognition remains one of the most fundamental and long standing theoretical problems in the scientific study of memory. Over the years, much evidence about this relationship has accumulated and many theoretical ideas have been proposed to explain it. In the history of this research three basic empirical strategies have been used to explore the problem. The first strategy has been to search for variables that influence one or the other or both of the two sorts of test—a search for patterns of functional dissociation and correspondence (for reviews see, e.g., Kintsch, 1970; McCormack, 1972; Tulving, 1976). The second strategy has been to try to make “direct” comparisons between recall and recognition performance under conditions in which all factors except that of test may be presumed to be equated (e.g., Broadbent & Broadbent, 1975; Watkins & Todres, 1978). The third strategy has been to give subjects successive tests of each sort and to analyze the relation between recognition and recall in terms of individual subject-item recognition-recall contingencies. In particular, this third strategy has been concerned with the phenomenon of recognition failure of recallable words, that is, the finding that subjects fail to recognize items that they can subsequently recall (Tulving & Thomson, 1973).
