Short-Term Memory for Temporal and Spatial Order Information

The present investigation includes theoretical and experimental comparisons of short-term memory for temporal and spatial order information. Two questions are of central interest: whether short-term memory for letter sequences is limited to a verbal-acoustic coding process and whether the loss of information from short-term memory results from interference or from a mere decay with time. The basic experimental paradigm employed in the present study was the Brown-Peterson design. The items to be recalled were sequences of four consonants presented visually in varying temporal and spatial arrangements. On each trial the to-be-remembered sequence of consonants was followed by an intervening task which in general consisted of digit shadowing. At the end of the retention interval the subject recalled the consonants in the order in which they were presented on the given trial. Two experiments were conducted which dealt exclusively with memory for temporal order information. These studies provided an experimental separation of item and order information. In one experiment, the subject was required to learn only order information, and in the other, he was required to learn only item information. Two Markov models were proposed to describe behavior in these experiments as well as in an experiment by Bjork and Healy (1970) where both item and order information had to be learned. The models provided good fits to the data of the Bjork and Healy experiment and of the experiment where the subject had to learn only order information. This result supports the notions that in these situations acoustic coding is employed, item and order information are lost independently, and each item is independently represented in memory. In five additional experiments, temporal and spatial order information were varied independently. These studies revealed a considerable difference between short-term memory for temporal and spatial order information. There was no evidence for acoustic coding in the spatial v order recall case as there had been in the case of temporal order recall. In addition, the time course of forgetting in spatial order recall was much flatter than in temporal order recall. The latter finding was explained in terms of the intervening task. Similarity of the processes involved in the intervening and recall tasks, rather than similarity of the items involved in the two tasks, was found to be the essential determinant of intervening task effectiveness. A probabilistic model was proposed to account for the spatial order recall situations and provided a good fit to the data. The model incorporates the notion that in spatial order recall the basic memory unit is the temporal spatial pattern of item presentations, rather than a representation of each separate item as in temporal order recall. These studies ruled out the notion that short-term memory for letter sequences is limited to a verbal-acoustic coding system. Evidence was found instead for a flexible coding system. Although the present results could not settle the question of whether forgetting is due to interference or decay, they do suffice to put restrictions on the form of an acceptable decay or interference theory. It is concluded that an acceptable decay theory must postulate different types of rehearsal, and that an acceptable interference theory must postulate process similarity as well as item similarity as a source of interference