Children's Thinking

What Develops?

Robert Siegler

• Publisher: Psychology Press
• Year: 2013
• ISBN:
  9781134993161
• Edition: 1
• Topic: Psychology
• Subtopic: History & Theory in Psychology
• Index: Psychology

First published in 1978. In 1963, John Flavell posed one of the truly basic questions underlying the study of children's thinking; his question was simply "What develops?" This volume holds the papers from the 13th Annual Carnegie Cognition Symposium, held in May 1977, that considering what progress had been made toward answering this question in the past 15 years.

|

MEMORY DEVELOPMENT

1 Skills, Plans, and Self-Regulation

Ann L. Brown
Judy S. DeLoache
University of Illinois

I. INTRODUCTION

In thinking about memory development, we have rarely questioned the essential similarity of the processes studied under the rubrics “problem-solving skills” and “memory strategies” (Brown, 1975a, 1977, in press a). A general class of Information-Processing models, with their emphasis on routines controlled and regulated by an executive, seems suitable for describing the major psychological processes of interest in both domains. However, because our charge is to function under the memory development heading, we have decided to refocus our thinking from our usual position of regarding the problem-solving and memory people as those who study the same processes on different tasks. Instead, we have begun by looking for any interesting differences between the major emphases and accomplishments in one field that could intelligently aid the development of the other. There do appear to be some psychologically interesting differences, not only in the tasks and skills studied but in the depth of the analyses of those tasks and skills and in the commitment to addressing instructional goals. In the first part of this chapter we highlight some of these differences between the two approaches and try to illustrate a weakness in the current mainstream of memory-development research. In the second part we concentrate on an area of concern to both the problem-solving and memory-development literatures: self-regulation and control, our candidate for the most fundamental difference between the experienced and the naive. In the final section we indicate new problem areas and new ways of considering what it is that develops with age and experience.

II. DIFFERENCES BETWEEN THE MEMORY DEVELOPMENT AND PROBLEM-SOLVING APPROACHES TO COGNITIVE DEVELOPMENT

Because our task is to consider what memory theorists have to say about development, we approach the issue from the perspective of the memory-development literature. Studies of a few aspects of memory dominate the field at this time, at least in number if not in content. Before we address the issues central to such research, a brief history of the way developmental psychologists interested in memory have approached the question of “What develops?” is illustrative.

A. Early Studies

Prior to the 1960s, the question “What develops?” would not have been raised. Obviously, memory develops. Lacking a fine-grained analysis of memory processes, early researchers selected tasks and age groups somewhat randomly. They found that on most tasks older children remembered more than younger ones, and slow learners had more difficulty remembering than those of average ability! The predominant explanation, when one was offered at all, was that immature learners have a limited memory “capacity.” As they mature, this capacity increases, allowing them to retain more. The underlying metaphor, whether implicitly or explicitly stated, was of the mind as a container: Little people have little boxes or jars in their heads, and bigger people have bigger ones. Any demonstration of inferior performance on the part of the smaller person proved the capacity limitation “theory,” not surprisingly since such a theory was merely a restatement of the data (Chi, 1976). The same general state of affairs also characterized the problem-solving literature, where early studies also showed poor performance by young children on a variety of tasks. Explanations of why the young did poorly were either not forthcoming or involved a circular argument: Little people have little problem-solving capacity, a restatement of the data masquerading as a theoretical explanation.

More sophisticated, or simply more adventurous, theorists subdivided the metaphorical containers. They attributed the deficits in memory or problem-solving performance to a limitation in the space available in one of the main architectural structures of the information-processing system, with space defined in terms of the number of slots, spaces, or buffer units available to the system at any one time. It was thought that as a child matured, his available space increased. The correlation of digit span with age, intelligence, and general problem-solving efficiency was taken as firm support for this notion of increasing space with increasing age, and short-term memory was cited as the most likely culprit in the young child's mental overpopulation problem. It should be noted, however, that most developmental psychologists avoided the issue of architectural systems and capacity limitations, thanks to two important influences on the field: the pioneering work of John Flavell (1970) on memory strategies in the young and the widespread dissemination of levels-of-processing approaches Craik & Lockhart, 1972) with their de-emphasis on capacity, coding, and flow in and between containers.

B. The Production-Deficiency Hypothesis

The guiding hypothesis of developmental memory research, initiated by Flavell in the 1960s and still popular today, is that the main difference between young children and mature memorizers is the tendency to employ a variety of mnemonic strategies whenever feasible. Borrowing from mediational theories of learning, Flavell introduced the terminology of production and mediational deficiencies to describe this difference. A production deficiency is said to exist when the child does not spontaneously produce a task-suitable mnemonic but if trained to do so can use the mnemonic to improve performance. A mediational deficiency exists when a child produces a necessary mnemonic, either spontaneously or under instruction, but the mnemonic fails to enhance performance. Probably owing to the paucity of strategies selected for study, mediational deficiencies have rarely been documented, and therefore the central issue in the memory-development research has been the spontaneous production of appropriate mediation.

Simply stated, the theory consists of three propositions: (1) Young children do poorly on a variety of memory tasks because they fail to introduce the necessary mnemonic intervention; (2) if they are trained to use a suitable strategy, their performance improves, at least temporarily; (3) if the memory task does not demand mnemonic intervention, developmental differences will be minimal. To prove or disprove one or the other of these hypotheses is still the goal of the majority of developmental memory studies.

To date, the field has been remarkably limited in the tasks selected for examination. When seeking to prove or disprove proposition 1 or 2, investigators almost invariably choose some rote memorization situation, such as list learning, in which rehearsal or taxonomic categorization is the strategy of choice. When seeking to prove or disprove proposition 3, they select some sort of recognition memory task. We have objections to both approaches. We believe that the strategy–no-strategy distinction served a valuable function in its time by organizing a chaotic field and by attempting to distinguish when and where the limitations of youth, lack of experience, or low IQ would be most debilitating. However, the two main lines of research now following this tradition have such severe built-in limitations that future proliferation should be discouraged. The main problems center around the study of tasks rather than processes and the paucity of developmental information provided by the particular tasks selected

1. Nonstrategic Tasks. The first of these two lines of research currently generating a spate of studies is one for which we feel personally responsible. These studies focus on proposition 3, which we were rash enough to make explicit (Brown, 1973a, 1974, 1975a) rather than implicit as Flavell had done (1970). The proposition asserts that if a situation exists in which deliberate mnemonic intervention is not a prerequisite for efficient performance, developmental differences will be minimized. Obviously, it would be futile to seek tasks where no developmental differences occur. Not only must such tasks be impervious to mnemonic efforts, but they must also be uncontaminated by any other developmentally sensitive factor. The point of the original statement was not to prove the absence of developmental trends but merely to demonstrate that the magnitude of any developmental effect is sensitive to the degree that sophisticated plans and strategies can be used. In general, the hypothesis is well supported, whether the comparison involves intentional versus incidental learning instructions in adults or cross-age comparisons (Brown & Smiley, 1977b). Situations do differ in the degree to which intentional mnemonics can enhance performance, and some recognition memory and recency tasks are less sensitive to strategic intervention than are many other memory tasks that require rote recall (Brown, 1975a).

This does not seem to be the point that the current set of studies seeks to prove. Interest has shifted from processes to tasks per se, and the game has become one of trying to show developmental differences in recognition memory tasks. In general, such attempts are successful, but their success is not surprising. Recognition memory as a task is clearly not impervious to developmental differences. True, with distinct target and distractor items, excellent levels of performance have been found for very young children as well as for adults, a ceiling effect that often clouds interpretation of age effects. However, with careful choice of distractors, one could easily produce a floor effect across all ages. Matching-to-sample tests have been devised so that choice of the correct alternative is extremely difficult even without any memory load. Floor or ceiling effects can completely obscure developmental differences, and it was for this reason that we selected variants of a recency problem for our earlier studies of the strategy–no-strategy distinction (Brown, 1973a, 1973b, 1973c).

If the question of interest is whether or not age differences will be found in a recognition memory task, the distractor items are crucial. Even the simple manipulation of increasing the number of distractors on a choice trial increases the likelihood of finding a developmental trend because young children's performance is disrupted by this manipulation (Brown, 1975b). If, however, one would like to show that young children perform better than older children, then a more subtle manipulation might be needed. For example, one could vary the similarity of the distractor and target items along some scale of physical or semantic similarity not yet salient to the young but distracting to the old. The less mature child would not be snared by the “related” distractor and should outperform the confused older participant. If we knew enough about the development of conceptual systems, we would be able to produce any possible pattern of age effects in recognition tasks by varying the target-distractor similarity on dimensions differentially salient to the ages under investigation.

Such an endeavor, however useful for testing hypotheses concerning conceptual development, is not relevant to the original discussion of whether some situations exist in which young children perform well on memory tasks. Our position is that in order to understand what memory development is, it is essential to identify areas of strength as well as areas of weakness (Brown, 1974, 1975a). Furthermore, if we wish to devise remedial help for the inexperienced, we need to capitalize on naturally occurring strengths as well as to identify major areas of weakness. Finally, as Chi (1976) has argued persuasively, it is only by eliminating candidates for what develops that we can identify the true areas of developmental change.

2. The Modal Memory Strategy Experiment The second line of research currently dominating the field is a proliferation of replication studies demonstrating the developmental sensitivity of strategy-susceptible tasks. The typical experiment in this area consists of crude assessments of the presence or absence of strategic intervention. Children are then divided into those who produce and those who do not; those who do produce almost invariably outperform those who do not. We rarely have evidence of intermediate stages of production.

One problem with these studies is overkill – they long ago provided ample documentation of young children's mnemonic ineptitude. Another problem stems from undue concentration on a limited subset of tasks and strategies. Almost all studies concerned with the mnemonic production deficiency hypothesis have centered on list-learning tasks and the strategies of taxonomic categorization and rehearsal. Apart from the obvious undesirability of such a restricted focus and the oft-lamented lack of ecological validity of such tasks (Brown, in press a), there are some interesting limitations imposed by this particular approach. First, these two strategies tend to emerge between the ages of five and seven years and, under the conditions usually studied, do not undergo much refinement after the grade school years. Thus, we are left with an almost total lack of information concerning what develops before 5 and after 8 or 9 years of age. The second problem is that we lack detailed models of the gradual emergence of even these simple strategies; indeed, they may not be susceptible to detailed task analyses.

Probably the most important deficiency of this approach, though, is that the tasks are set up in such a way that we cannot say anything about nonproducers. If children are not rehearsing on our task, we have no way of knowing what they are doing. They perform poorly, and therefore they highlight the improvement with age that we wish to demonstrate. However, we know nothing about their state of understanding. They are characterized as not being at a certain level, as not having a certain attribute; they are nonproducers, nonmediators; they are not strategic or not planful. They are sometimes described as passive, even though the tasks are designed so that the only way to be characterized as active is to produce the desired strategy. All these descriptions are based on what young children do not do compared with older children, for we have no way of observing what they can do in the confines of these tasks....

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Contents
• Preface
• Dedication
• Part I Memory Development
• Part II Problem Solving
• Part III Representational Processes
• Author Index
• Subject Index