Douglas L. Hintzman

Schema abstraction in a multiple-trace memory model

A simulation model of episodic memory, MINERVA 2, is applied to the learning of concepts, as represented by the schema-abstraction task. The model assumes that each experience produces a separate memory trace and that knowledge of abstract concepts is derived from the pool of episodic traces at the time of retrieval. A retrieval cue contacts all traces simultaneously, activating each according to its similarity to the cue, and the information retrieved from memory reflects the summed content of all activated traces responding in parallel. The MINERVA 2 model is able to retrieve an abstracted prototype of the category when cued with the category name and to retrieve and disambiguate a category name when cued with a category exemplar. The model successfully predicts basic findings from the schema-abstraction literature (e.g., differential forgetting of prototypes and old instances, typicality, and category size effects), including some that have been cited as evidence against exemplar theories of concepts. The model is compared to other classification models, and its implications regarding the abstraction problem are discussed. How is abstract knowledge related to specific experience? In present-day terms, this question concerns the relationship between episodic and generic memories. This article explores the possibility that there is only one memory system, which stores episodic traces, and that abstract knowledge as such does not have to be stored but can be derived from the pool of traces of specific experiences at the time of retrieval. I demonstrate how this might work by applying a simulation model of a multipletrace memory theory to the sehema-abstraction experimental paradigm, which is widely believed to capture in the laboratory the processes by which generic or abstract ideas are formed. Multiple-trace theories assume that each event to which one attends gives rise to its own memory trace. Thus, repetition of an item such as a word in a list does not strengthen a prior representation (i.e., one predating the experiment or one laid down by the items first experimental occurrence); rather, it produces a new trace that coexists in memory with traces of other occurrences of the same item. Experiments supporting the multiple-trace assumption have been primarily concerned with the ability of subjects to remember an items presentation frequency, list membership, presentation modality, exposure duration, serial position, and so forth (e.g., Hintzman, 1976; Hintzman & Block, 1971; Hintzman, Block, & Summers, 1973;

Judgments of frequency and recognition memory in a multiple-trace memory model.

The multiple-trace simulation model, MINERVA 2, was applied to a number of phenomena found in experiments on relative and absolute judgments of frequency, and forced-choice and yes-no recognition memory. How the basic model deals with effects of repetition, forgetting, list length, orientation task, selective retrieval, and similarity and how a slightly modified version accounts for effects of contextual variability on frequency judgments were shown. Two new experiments on similarity and recognition memory were presented, together with appropriate simulations; attempts to modify the model to deal with additional phenomena were also described. Questions related to the representation of frequency are addressed, and the model is evaluated and compared with related models of frequency judgments and recognition memory.

MINERVA 2: A simulation model of human memory

An overview of a simulation model of human memory is presented. The model assumes: (1) that only episodic traces are stored in memory, (2) that repetition produces multiple traces of an item, (3) that a retrieval cue contacts all memory traces simultaneously, (4) that each trace is activated according to its similarity to the retrieval cue, and (5) that all traces respond in parallel, the retrieved information reflecting their summed output. The model has been applied with success to a variety of phenomena found with human subjects in frequency and recognition judgment tasks, the schema-abstraction task, and paired-associate learning. Application of the model to these tasks is briefly summarized.

Repetition and Memory1

Publisher Summary This chapter reviews the research on the effects of repetition conducted in the laboratory. Repetition is one of the most powerful variables affecting memory. The chapter discusses experiments called the “method of memory judgments.” In this technique, a list of items is presented, and the subject is then presented with each test item and asked to judge from memory some aspect of its presentation in the list. Judgments of recency, frequency, exposure duration, list membership, input modality, spacing, and serial position have all been used, singly and in combination. Judgments reveal much more about the richness of information encoded in memory than can be inferred from more traditional recall and recognition measures. Memory judgments have been particularly revealing where effects of repetition on memory are concerned. The chapter examines how frequency or number of occurrences is represented in memory. It reviews research on effects on memory of the spacing of repetitions and presents evidence on the role of repetitions as retrieval cues.

Orientation in cognitive maps

Abstract In 14 experiments, subjects had to “point to” surrounding environmental locations (targets) while imagining themselves in a particular spot facing in various directions (orientations). The spatial information was either committed to memory (cognitive maps) or directly presented on each trial in the visual or tactile modality. Reaction times (RT) indicated that orientation shifts were achieved through mental rotation in the visual task, but not in the cognitive map or tactile tasks. Further, in the latter two tasks targets were located most quickly when they were adjacent to or opposite the imagined orientation. Several explanations of this finding were tested. Various aspects of the data suggest that cognitive maps are not strictly holistic, but consist of orientation-specific representations, and—at least in part—of relational propositions specific to object pairs.

VIOLATIONS OF THE INDEPENDENCE ASSUMPTION IN PROCESS DISSOCIATION

L. L. Jacoby, J. P. Toth, and A. P. Yonelinas (1993) advocated a process-dissociation procedure for estimating the contributions to task performance of consciously controlled (R) versus automatic (A) memory processes. The procedure relies on the strong assumption that memory-guided performance attributable to R is stochastically independent of that attributable to A. Violations of this independence assumption can produce artifactual dissociations between estimates of R and A. Such artifactual dissociations were obtained in a series of word-stem completion experiments: R increased with presentation duration, whereas A, paradoxically, decreased. Direct evidence for nonindependence was obtained from correlations between R and A in each of the experiments. These results suggest that the independence assumption was violated, and other applications of process dissociation should not be taken at face value without a thorough evaluation of independence.

Memory for Mode of Input.

Three experiments examined the retention of information about mode of input of familiar words. In Expt. I, input modes were vision vs. audition; in Expt. II, block vs. script letters (all visual); in Expt. III, male vs. female voices (all auditory). Eight 18-word lists were presented. Half the words in each list occurred in one input mode and half in the other, and each list was followed by immediate free recall. After the last recall test, an unexpected recognition test was administered in which the subject was to indicate the input mode of each recognized word. Although the between-modality distinction was more effective than the within-modality distinctions, all three studies produced similar results: free recall revealed clustering by mode of input, and the final test showed that input-mode information was retained for several minutes after presentation.

Retrieval dynamics in recognition and list discrimination: further evidence of separate processes of familiarity and recall.

Two experiments tested the hypothesis that the time course of retrieval from memory is different for familiarity and recall. The response-signal method was used to compare memory retrieval dynamics in yes-no recognition memory, as a measure of familiarity, with those of list discrimination, as a measure of contextual recall. Responses were always made with regard to membership in two previous study lists. In Experiment 1 an exclusion task requiring positive responses to words from one list and negative responses to new words and words from the nontarget list was used. In Experiment 2, recognition and list discrimination were separate tasks. Retrieval curves from both experiments were consistent, showing that the minimal retrieval time for recognition was about 100 msec faster than that for list discrimination. Repetition affected asymptotic performance but had no reliable effects on retrieval dynamics in either the recognition or the list-discrimination task.

Differential forgetting of prototypes and old instances: simulation by an exemplar-based classification model.

A common finding in studies of classification learning is that ability to classify the prototype of a category declines much less over a retention interval than does the ability to classify the previously seen exemplars themselves. We demonstrate here that this finding does not necessarily indicate the existence, in memory, of a representation of the prototype. MINERVA, a computer-simulation model that encodes memory traces only of presented exemplars, was tested on an appropriate task. Differential forgetting of prototypes and old instances was shown by a version of the model that assumed that (1) classification is based on the exemplar trace most similar to the test stimulus and (2) individual properties are lost from the traces over time in an all-or-none fashion. It is suggested that, in general, the key to the prediction of differential forgetting may be the concomitance of forgetting and generalization.

Spacing judgments as an index of study-phase retrieval.

It is hypothesized that the ability of subjects to judge how far apart two presentations of a word were in a list reflects study-phase retrieval of the trace of the first presentation of the word by its second presentation. Experiment 1 supported this hypothesis by demonstrating that the accuracy of spacing judgments for associatively related pairs of words, like that for repeated words, was high compared to that for unrelated words. Experiment 2 used spacing judgments to measure retrieval upon repetition of a homograph. In three conditions, context words accompanying a homograph on its two presentations were either the same, biased the same meaning, or biased different meanings. In all three conditions, later spacing judgments were more accurate than in an unrelated-word control. Accuracy did not depend on whether the two context words biased the same meaning or different meanings of the homograph.