This paper is divided into two major portions; the first outlines a general theoretical framework in which to view human memory, and the second describes the results of a number of experiments designed to test specific models that can be derived from the overall theory.

The general theoretical framework, set forth in II Structural Features of the Memory System, III Control Processes in Memory, categorizes the memory system along two major dimensions. One categorization distinguishes permanent, structural features of the system from control processes that can be readily modified or reprogrammed at the will of the subject. Because we feel that this distinction helps clarify a number of results, we will take time to elaborate it at the outset. The permanent features of memory, which will be referred to as the memory structure, include both the physical system and the built-in processes that are unvarying and fixed from one situation to another. Control processes, on the other hand, are selected, constructed, and used at the option of the subject and may vary dramatically from one task to another even though superficially the tasks may appear very similar. The use of a particular control process in a given situation will depend upon such factors as the nature of the instructions, the meaningfulness of the material, and the individual subject’s history

A computer analogy might help illustrate the distinction between memory structure and control processes. If the memory system is viewed as a computer under the direction of a programmer at a remote console, then both the computer hardware and those programs built into the system that cannot be modified by the programmer are analogous to our structural features; those programs and instruction sequences which the programmer can write at his console and which determine the operation of the computer, are analogous to our control processes. In the sense that the computer’s method of processing a given batch of data depends on the operating program, so the way a stimulus input is processed depends on the particular control processes the subject brings into play. The structural components include the basic memory stores; examples of control processes are coding procedures, rehearsal operations, and search strategies.

Our second categorization divides memory into three structural components: the sensory register, the short-term store, and the long-term store. Incoming sensory information first enters the sensory register, where it resides for a very brief period of time, then decays and is lost. The short-term store is the subject’s working memory; it receives selected inputs from the sensory register and also from long-term store. Information in the short-term store decays completely and is lost within a period of about 30 seconds, but a control process called rehearsal can maintain a limited amount of information in this store as long as the subject desires. The long-term store is a fairly permanent repository for information, information which is transferred from the short-term store. Note that “transfer” is not meant to imply that information is removed from one store and placed in the next; we use transfer to mean the copying of selected information from one store into the next without removing this information from the original store.

In presenting our theoretical framework we will consider first the structural features of the system (Section II) and then some of the more generally used control processes (Section III). In both of these sections the discussion is organized first around the sensory register, then the short-term store, and finally the long-term store. Thus, the outline of II Structural Features of the Memory System, III Control Processes in Memory can be represented as follows:These first sections of the paper do not present a finished theory; instead they set forth a general framework within which specific models can be formulated. We attempt to demonstrate that a large number of results may be handled parsimoniously within this framework, even without coming to final decisions at many of the choice points that occur. At some of the choice points several hypotheses will be presented, and the evidence that is available to help make the choice will be reviewed. The primary goal of II Structural Features of the Memory System, III Control Processes in Memory is to justify our theoretical framework and to demonstrate that it is a useful way of viewing a wide variety of memory phenomena.

The remaining sections of the paper present a number of precise models that satisfy the conditions imposed by our general theoretical framework. These sections also present data from a series of experiments designed to evaluate the models. Section IV is concerned with an analysis of short-term memory; the model used to analyze the data emphasizes a control process based in the short-term store which we designate a rehearsal buffer. Section V presents several experiments that shed some light upon processes in the long-term store, especially subject- controlled search processes. Some of the experiments in IV Experiments Concerned with Short-Term Processes, V Experiments Concerned with Long-Term Search and Retrieval have been reported by us and our co-workers in previous publications, but the earlier treatments were primarily mathematical whereas the present emphasis is upon discussion and overall synthesis.

If the reader is willing to accept our overall framework on a provisional basis and wishes to proceed at once to the specific models and experiments, then he may begin with Section IV and as a prerequisite need only read that portion of Section III,B concerned with the rehearsal buffer.