John B. Black

Scripts in Memory for Text

These experiments investigate people’s knowledge of routine activities (e.g., eating in a restaurant, visiting a dentist) and how that knowledge is organized and used to understand and remember narrative texts. We use the term script to refer to these action stereotypes. Two studies collected script norms: people described what goes on in detail during familiar activities. They largely agreed on the nature of the characters, props, actions, and the order of the actions. They also agreed on how to segment the low-level action sequences into constituent “scenes,” suggesting a hierarchical organization in memory of the activity. Other studies investigated memory for a text narrating actions from a script. Subjects tended to confuse in memory actions that were stated with unstated actions implied by the script. This tendency increased as more related script instances were studied. Subjects also preferred to recall script actions in their familiar order; a scrambled text that presented some script actions out of order tended to be recalled in canonical order. We also investigated whether the reading time for adjacent statements in a text varied with their distance apart in the underlying script. A statement at a one-step distance was read faster than one at a two- or three-step distance; statements in the second half of a script were read faster than those in the first half. A final experiment found that goal-relevant deviations from a script were remembered better than script actions. The role of script knowledge in text memory was discussed, as was the relation of scripts to schema memory in general.

The Development of Cognitive Skills To Support Inquiry Learning

Establishing the value of inquiry learning as an educational method, it is argued, rests on thorough, detailed knowledge of the cognitive skills it is intended to promote. Mental models, as representations of the reality being investigated in inquiry learning, stand to influence strategies applied to the task. In the research described here, the hypothesis is investigated that students at the middle school level, and sometimes well beyond, may have an incorrect mental model of multivariable causality (one in which effects of individual features on an outcome are neither consistent nor additive) that impedes the causal analysis involved in most forms of inquiry learning. An extended intervention with 6th to 8th graders was targeted to promote (a) at the metalevel, a correct mental model based on additive effects of individual features (indicated by identification of effects of individual features as the task objective); (b) also at the metalevel, metastrategic understanding of the need to control the influences of other features; and (c) at the performance level, consistent use of the controlled comparison strategy. Both metalevel advancements were observed, in addition to transfer to a new task at the performance level, among many (though not all) students. Findings support the claim that a developmental hierarchy of skills and understanding underlies, and should be identified as an objective of, inquiry learning.

Shuttling Between Depictive Models and Abstract Rules: Induction and Fallback

A productive way to think about imagistic mental models of physical systems is as though they were sources of quasi-empirical evidence. People depict or imagine events at those points in time when they would experiment with the world if possible. Moreover, just as they would do when observing the world, people induce patterns of behavior from the results depicted in their imaginations. These resulting patterns of behavior can then be cast into symbolic rules to simplify thinking about future problems and to reveal higher order relationships. Using simple gear problems, three experiments explored the occasions of use for, and the inductive transitions between, depictive models and number-based rules. The first two experiments used the convergent evidence of problem-solving latencies, hand motions, referential language and error data to document the initial use of a model, the induction of rules from the modeling results, and the fallback to a model when a rule fails. The third experiment explored the intermediate representations that facilitate the induction of rules from depictive models. The strengths and weaknesses of depictive modeling and more analytic systems of reasoning are delineated to motivate the reasons for these transitions.

Point of View in Narrative Comprehension, Memory, and Production

A preference for a consistent point of view pervades narrative comprehension, memory, and production. Subjects read statements exhibiting a consistent point of view faster than statements exhibiting a change in point of view, and they rated the consistent statements as more comprehensible than change statements. Futhermore, subjects tended to misrecall change statements as consistent statements; and when asked to edit statements to make ,them more comprehensible, they rewrote change statements to be consistent. Merely making a character the subject of the narrative statement sufficed to establish his as the dominant point of view.

Knowledge structures in the organization and retrieval of autobiographical memories

In this paper, the role of knowledge structures in organizing and retrieving autobiographical experiences is investigated. It is proposed that autobiographical events are organized in memory by the knowledge structures that guided comprehension and planning during the experience. Individual experiences are retrieved from memory by first accessing the knowledge structures used to encode the event, then using information in those structures to predict features of the target event, thus directing search to paths likely to lead to that event. Two types of structures are investigated as candidates for these organizing contexts. Activities are sequences of actions performed to achieve a goal, while general actions are situation-free components occurring as part of several activities and represent what is common to that action across those activities. It is predicted that activities are more important in retrieving experiences, because (1) these structures constitute the principal contexts used to store experiences and (2) information contained within these structures is more useful for predicting features of target events. The greater utility of activities in retrieving experiences was demonstrated in two autobiographical memory retrieval time experiments. First, retrieval of a personal experience matching an activity and action combination was faster when subjects were given an activity cue before a general action cue, because processing could get a “head start” when the activity context was presented first. Second, specifying an activity and action led to faster memory retrievals than specifying only the action, while no such facilitation occurred when an activity was augmented by a general action. In both experiments, retrieval was slowed when more processing was required to infer probable features of the target experience, as predicted by the directed nature of the search process. These experiments and this model provide a general framework for studying the organization of events in autobiographical memory.

The representation of scripts in memory

Abstract Previous research has shown that people possess stereotyped knowledge about common events; that is, people have scripts containing information about the actions that comprise these events and about the temporal order of these actions. They use this knowledge in making inferences that help them to fill in gaps found in narratives or predict information to follow. The memory representation of common events was investigated by studying the pattern of inferences people make when they read descriptions of these events. People always made inferences that generalized the information presented. When a detail was presented, a more general concept of which it was a part was inferred. Readers often inferred items at the level of abstraction corresponding to scene headers when a sentence embodying a more abstract concept was stated. These results indicated that scripted events are represented in memory as hierarchically and temporally organized information packets. The connections between packets in the network are arranged in such a way that useful generalizations and predictions are available to aid people in understanding events.

The "soap opera" effect in story recall

When people learn a series of actions of a character, their memory is strongly influenced, we suggest, by beliefs about that character’s motives. Motives serve as schema for deciding the meaning of the actions, their importance, and their interconnections. In two experiments, subjects read and later remembered some dull activities of a character. Experimental subjects knew what this character was worrying about (e.g., an unwanted pregnancy); control subjects did not. Recall and recognition showed that motive subjects distorted many of the colorless events to be motive relevant. Although the motive schema helped connect the disparate actions, it interfered with accurate recording and recall of the details.

Causal coherence and memory for events in narratives

Causally related events in narratives were remembered better than events that were not causally related. In Experiment 1, subjects recalled sentences from stories when given the sentences that immediately preceded them as cues. Cued recall was better when the two sentences were causally related than when they were not. In Experiment 2, subjects free recalled the same stories. Again, recall was better for the sentences when they were part of a causally related pair. Also subjects were more likely to combine two sentences into one during recall when they were causally related.

Analog Imagery in Mental Model Reasoning: Depictive Models

We investigated whether people can use analog imagery to model the behavior of a simple mechanical interaction. Subjects saw a static computer display of two touching gears that had different diameters. Their task was to determine whether marks on each gear would meet if the gears rotated inward. This task added a problem of coordination to the typical analog rotation task in that the gears had a physical interdependency; the angular velocity of one gear depended on the angular velocity of the other gear. In the first experiment, we found the linear relationship between response time and angular disparity that indicates analog imagery. In the second experiment, we found that people can also solve the problem through a non-analog, visual comparison. We also found that people of varying spatial ability could switch between analog and non-analog solutions if instructed to do so. In the third experiment, we examined whether the elicitation of physical knowledge would influence solution strategies. To do so, we manipulated the visual realism of the gear display. Subjects who saw the most realistic gears coordinated their transformations by using the surfaces of the gears, as though they were relying on the friction connecting the surfaces. Subjects who saw more schematic displays relied on analytic strategies, such as comparing the ratios made by the angles and/or diameters of the two gears. To explain the relationship between spatial and physical knowledge found in the experiments, we constructed a computer simulation of what we call depictive modeling. In a depictive model, general spatial knowledge and context-sensitive physical knowledge have the same ontology. This is different from prior simulations in which a non-analog representation would be needed to coordinate the analog behaviors of physical objects. In our simulation, the inference that coordinates the gear motions emerges from the analog rotations themselves. We suggest that mental depictions create a bridge between imagery and mental model research by positing the referent as the primary conceptual entity.

Types of Inferences Generated during Reading

Which kinds of inferences people make when reading a narrative and whether those inferences become an integral part of the representation of the narrative information in memory were investigated in two experiments. In particular, the effects on reading time and memory were examined when the narrative text suggested goal, plan, action, and state inferences that could be made by readers. Increases in reading time were found when goal, plan, and action inferences were suggested, but not when state inferences were suggested. Similarly, a recognition memory test yielded high false alarm rates for goal, plan, and action inferences, but not for states. This shows that the suggested inferences were incorporated in memory with information explicitly stated in the tests. Thus the results indicate that goal, plan, and action inferences are likely to be made during reading and become an indistinguishable part of the memory representation, while little evidence is found for state inferences being made during reading or becoming part of memory for narrative texts.