Two experiments were designed to investigate and integrate research from organizational and levels-of-processing perspectives. In the first experiment, subjects were given incidental learning instructions that directed attention either to relational aspects of items in a list or to the individual items themselves or to both. This processing was either semantic or nonsemantic in nature. The typical levels-of-processing effect was observed with the semantic and nonsemantic tasks for both relational and individual-item processing. Free-recall and recognition performance were found to be determined by the additive combination of relational processing and individual-item processing. Also, the two kinds of processing had functionally different effects on memory when clustering, recognition scores, and false alarms were examined. Relational processing apparently enhanced the formation of retrieval schemes, whereas individualitem processing seemed to facilitate discriminative processes. In the second experiment, the relative importance of relational and individual-item processing was shown to be dependent on the saliency of the structure of the learning materials. It is argued that the distinction between these two types of processing is a useful one and should be considered in relation to the structure of the learning material and to the type of memory task.As recently as 1970, the study of memory was experiencing a revolution, with "the insurgents marching under the banner of cognitive organization" (Bower, 1970, p. 18). Organization, a process of grouping information into new and larger units, was considered to be absolutely necessary for optimal retention (Mandler, 1967). Indeed, a considerable amount of empirical support was generated for this position. In spite of

The role of semantic memory activation in accounting for generation effects and reality monitoring was investigated in young and old normal adults and in patients with dementia of the Alzheimer's type (DAT). Both young and old normal adults demonstrated higher recall for internally generated information than for externally presented information, whereas the DAT patients failed to demonstrate a generation effect. Similarly, reality monitoring scores (discrimination between internal and external items) were high for both age groups of normals, but near chance levels for the DAT group. These results implicate semantic memory as an important factor in generation effects.

The generation of words from word fragments, both with and without context cues, was used as an encoding operation for investigating contributions to word ("perceptual") identification. Experiment 1, with isolated words, showed that a generation effect may obtain in word identification if test items are presented in fragmented form. Experiment 2, with context cues, also showed a generation effect, but only when an identical study cue was presented shortly before the identification trial; use of a different (but associatively related) cue produced equivalent identification performance between read and generated items. In identification "catch trials" of Experiment 3, nonstudied items were tested after a previously studied cue was presented. Results showed depressed but equivalent levels of performance between read and generate groups; however, incorrect identifications were associated with a tendency to respond with previously generated items.

We focus on the issue of whether cognitive effort is causally related to memory. We begin with a discussion of the concept of cognitive effort as derived from capacity models of attention. We then suggest that the theoretical analysis of memory may involve concepts from different levels of psychological analysis, and we draw a distinction between concepts that represent boundary conditions and sufficient cause. When applied to memory phenomena, attentional concepts serve only as a boundary-or limiting-function in memory theory. In contrast, concepts that represent memorial processes serve as a sufficient cause function. In some instances, cognitive effort appears to have been used as a sufficient cause concept, resulting in some confusion. A review of the literature reveals a haphazard correlation between indexes of cognitive effort and of memory performance. Alternatively, the application of cognitive effort or capacity to the memory performance of certain populations (clinical, children, and elderly) illustrates a potentially more appropriate use of the concept.Whatever future conclusion we may reach as to this, we cannot denythat an object once attended to will remain in the memory, whilst one inattentively allowed to pass will leave no traces behind. (James, 1890, p. 427) The conceptsof attentionand memory have been perennially interwoven in the fabric of cognitive theory. Any adequate theory of intellectual functioning must eventually describe the interaction between these concepts, as well as provide a coherent account of the two concepts separately. Research and theory on these concepts logically precede description of their interaction, but the inextricable intertwining of attention and memory sometimes encourages the transfer of concepts from attention to explanations of memory. Since theories of attentionand theories of memory, as separate conceptual systems, are designed to address different questions, borrowing concepts from attention to explain questions about memory can lead to confusion and, occasionally, to illogical propoThis article is based in part on a symposium at the 91st Annual Convention of the American Psychological Association in Anaheim, CA (Hunt, 1983 sitions. It is our position that some conceptual confusion has arisen over the application of the attentional concept of cognitive effort as an explanatory mechanism of memory phenomena.The goal of a theory of attention, in large part, is to describe the selective and limited character of cognitive functioning. The empiricalbasis for this work has derived historically from the question, "How many things can be done at one time?" Dependentmeasures of memory have often been used to answer this empirical question. Memory measures are prevalent in empirical studies of attention because theories of attention typically specify a mechanism for selection, and memory performance can be predicted from the hypotheticaloperation of this selection mechanism. For example, Broadbent's (1958)model of attention assumed that the processes responsible f...

The general problem under investigation in this study is the extent to which an orienting activity exerts control over the encoding process. Two experiments are reported in which associative meaningfulness was varied under conditions of semantic and nonsemantic processing. Contrary to assumptions of exclusive encoding control by the orienting task, both experiments showed effects of meaningfulness following both semantic and nonsemantic processing tasks. The results are consistent with previous reports of nonsemantic structural effects following semantic processing. These data imply that the memory trace may contain more than just those features congruent with the orienting task and, consequently, that encoding must be conceptualized in terms of both process and structure.Recent theoretical analyses of the input stage of memory have seen a dramatic shift from the traditional focus on the contribution of stimulus structure toward an emphasis on processes (Craik & Tulving, 1975;Kolers, 1976). Memory performance is analyzed in terms of what was done to the stimulus rather than in terms of the properties of that stimulus. The contemporary emphasis follows axiomatically from the informationprocessing framework. Fundamental to the information-processing approach is the assumption that mere specification of the at-