Recognition and retrieval processes in free recall.

Abstract

A model of free recall is described which identifies two processes in free recall: a retrieval process by which the subject accesses the words, and a recognition process by which the subject decides whether an implicitly retrieved word is a to-be-recalled word. Submodels for the recognition process and the retrieval process are described. The recognition model assumes that during the study phase, the subject associates to the to-be-recalled words. The establishment of such associates is postulated to be an all-or-none stochastic process. In the test phase, the subject recognizes to-be-recalled words by deciding which words have relevant list markers as associates. A signal detectability model is developed for this decision process. The retrieval model is introduced as a computer program that tags associative paths between list words. In several experiments, subjects studied and were tested on a sequence of overlapping sublists sampled from a master set of common nouns. The twoprocess model predicts that the subject's ability to retrieve the words should increase as more overlapping sublists are studied, but his ability to differentiate the words on the most recent list should deteriorate. Experiments confirmed this predicted dissociation of recognition and retrieval. Further predictions derived from the free recall model were also supported. This paper has several aims: First, we offer a critique of two popular theories, one which relates recall and recognition to the strength of one and the same memory trace, and another which relates only recognition memory to a similar strength measure; second, we develop a particular conceptualization about recognition memory which we believe satisfies the criticisms of the traditional strength theory; third, we illustrate how that recognition mechanism could be interfaced with a retrieval mechanism so as to yield a viable theory about multilist free recall. Along with reviewing published data relevant to these points, we also shall present new 1 This research was supported by Grant MH