Atkinson and Shiffrin (1971)

Copyright Notice: This material was written and published in Wales by Derek J. Smith (Chartered Engineer). It forms part of a multifile e-learning resource, and subject only to acknowledging Derek J. Smith's rights under international copyright law to be identified as author may be freely downloaded and printed off in single complete copies solely for the purposes of private study and/or review. Commercial exploitation rights are reserved. The remote hyperlinks have been selected for the academic appropriacy of their contents; they were free of offensive and litigious content when selected, and will be periodically checked to have remained so. Copyright © 2002, Derek J. Smith (Chartered Engineer).

First published online 08:30 30th July 2002

This version dated 08:30 30th July 2002

Memory Models Before Atkinson and Shiffrin (1971)

Atkinson and Shiffrin (1971) have been justifiably widely cited for their analysis of the nature and functional layout of the mind's memory systems, but before looking in detail at their work we need firstly to note some of the explanatory problems they faced at the time.

The idea that the brain contains a limited capacity short-term memory (STM) able to "consolidate" a selected portion of its throughput into a much larger capacity long-term memory (LTM) goes back to the end of the nineteenth century (Maudsley, 1876; James, 1890; Müller and Pilzecker, 1900). Such "two-store" models of memory were popular for many decades because they were consistent with "cell assembly" explanations of memory as a physiological process (Lorente de No, 1938; Hebb, 1949): STM could readily be seen as neural electrical activity, and LTM could readily be seen as consequent changes in neural microstructure. Nevertheless, the two-store model had to be forcibly upgraded in 1960, following George Sperling's research into visual iconic memory (Sperling, 1960), and it continued to demand further refinement throughout the 1960s in the light of data from a succession of cleverly designed verbal learning experiments (eg. Peterson and Peterson, 1959; Conrad, 1964; Baddeley, 1966). Here are some of the most influential lines of experimentation from that period:

• Very Short Term Memory (VSTM): Following Sperling's lead, analogs of the visual "Sperling store" were found in other sensory modalities (eg. Darwin, Turvey, and Crowder, 1972, for audition, and Bliss, Crane, Mansfield, and Townsend, 1966, for touch), implying that STM was preceded by VSTM in all input modalities.
• Working Memory (WM): This was one of several important concepts to be imported into cognitive psychology from computer science (relatively silently by Miller, Galanter, and Pribram, 1960, and Douglas, 1967, and then more famously by Baddeley and Hitch, 1974).
• Rehearsal: The idea that auditory material can be constantly refreshed by repeating it to oneself was a central part of Sperling's explanatory models of visual iconic memory (Sperling, 1963; 1967), and the nature of the underlying system was much discussed throughout the 1960s. The concept was finally modeled in some detail by Baddeley and Hitch (1974), in whose theory the process is allocated to a cognitive subsystem known as the "articulatory loop".
• Confusibility: A confusibility effect is a memory difficulty which emerges when the material to be retained consists of a number of items which are similar in some important respect. The standard explanation for this is that the corresponding engrams will also be confusible in that same respect, and therefore tend to get mixed up. Tests of differential confusibility are therefore tests of differential mental encoding, and, using them, two major confusibility effects were discovered during the 1960s:

The Phonological Similarity Effect: The phonological similarity effect refers to an STM impairment when presented with acoustically similar material, that is to say, items which sound alike. It was first detected by Conrad (1964), who found that misrecollections of target letters were more likely to be acoustically similar than not. Thus "D" would be more commonly an error for "B" (with which it rhymes) than for "R" (with which it does not rhyme). Where consonant sequences were to be memorised, Conrad and Hull (1964) found that acoustically similar sequences such as "B-G-V-P-T" were more prone to error than acoustically dissimilar sequences such as "Y-H-W-K-R". The same effect was found where word sequences were to be memorised, with "man-mad-cad-mat-cap" being more prone to error than "pit-day-cow-sup-bar" (Baddeley, 1966).

The Semantic Similarity Effect: The semantic similarity effect refers to an LTM impairment when presented with semantically similar material, that is to say, items which can be associated by their meaning. It was first detected by Baddeley (1966), who found that semantically similar sequences such as "large-great-huge-long-big" were more prone to recall error after 20 minutes than semantically disparate sequences such as "old-wet-strong-thin-deep". He also detected a weak semantic similarity effect in STM.

• Imagery: The idea that visual material can be represented visually in some sort of "mind's eye" was developed by workers such as Paivio (eg. Paivio and Csapo, 1969), and the concept is also to be seen in Baddeley and Hitch's (1974) visuo-spatial scratchpad.

Atkinson and Shiffrin's (1971) Memory Model

It was this rapidly expanding body of empirical data which Atkinson and Shiffrin set out to make sense of. They published both mathematical and diagrammatic models of memory in the mid-to-late 1960s (Atkinson and Shiffrin 1965, 1968), and then combined the strengths of both approaches in their 1971 model, a model which was so useful as a general purpose summary of what was going on during a memory task that it soon came to be called the "modal" (ie. "popular" or "standard") model (Murdock, 1971). Here is that model:

Atkinson and Shiffrin's (1971) Model of Memory: Here is the model as published. It represents the stages through which sensory information has to go in order to influence behaviour. Three stages are identified, ranged sequentially from left-to-right, as follows:

Sensory Registers: Sensory input is initially detected by an array of sensory registers. These contain memory resources with a lifespan of only a few hundred milliseconds. They include Sperling's iconic memory in the visual modality, plus similar VSTMs in the auditory and other sensory modalities. Each sensory register allows some rudimentary processing to take place before the input is passed to the next stage.

Short Term Store (STS): This contains memory resources with a lifespan of only a few seconds. It is subdivided into a working memory component and a control processes component. The working memory component provides a general purpose processing resource, whilst the control processes have more specific purposes. Four control processes are formally marked on the diagram, but more are mentioned in Atkinson and Shiffrin's supporting text, and others in their 1968 paper. Together, they are responsible for cross-associating current sensory input with the contents of LTM, thus allowing past knowledge to modulate current and future behaviour. Here is the full list:

• Rehearsal: As described above. Atkinson and Shiffrin propose that rehearsal relies upon a subdivision of STS capable of holding a limited number of items pending further processing. They borrow Sperling's (1963) terminology, and call this small dedicated store a Rehearsal Buffer.
• Coding: Atkinson and Shiffrin do not discuss this topic in much detail, but define the process as "a class of control processes in which the information to be remembered is put in a context of additional, easily retrievable information, such as a mnemonic phrase or sentence" (p83).
• Decision Rules, Organisational Schemes, and Problem Solving Techniques: These are higher cognitive control processes which subjects can call upon at their discretion. The authors give little further detail in the 1971 paper, but Atkinson and Shiffrin (1968) had already mentioned grouping, chunking, and selection as useful organisational principles.
• Retrieval Strategies and Probe Selection: This is another optional control process, and it reflects the possibility that retrieval from LTS "is considerably more complicated" than retrieval from STS. The authors suggest that "probe information" is important when retrieving from LTS. It might be, for example, that you are trying to remember a film star's name and decide that it begins with the letter "A". This gives you the retrieval probe <film stars, A>, and this will work well providing your initial judgements were accurate. If, however, the film star's name happens to begin with "K", or if the face in question is not a film star at all, then your probe is inaccurate and the act of retrieval will initially fail.
• Imaging: As described above. This is a control process "in which verbal information is remembered through visual images" (p83). It is therefore the opposite process to the auditory coding of visual input which takes place when text is viewed.

Long Term Store (LTS): This is a high-volume memory with a potential lifespan of decades. Indeed, so much information can be stored that the main problem is finding the particular bit you are interested in. This is made considerably easier by dividing the content up into logical "subsets", accessed using the STS retrieval strategies described above. "This portrayal of the memory system almost entirely in terms of the operations of the short term store is quite intentional" (p84).

Redrawn from Atkinson and Shiffrin (1971). This graphic Copyright © 2002, Derek J. Smith.

It is important to note that the proposed model is a functional, or logical, model of the layout of the memory system, and not a physical one. It tries to describe what is going on, not where. The authors themselves point out that the physical separation of two mental processes on a model "does not require that the two stores necessarily be in different parts of the brain or involve different physiological structures[; one] might consider the short-term store simply as being a temporary activation of some portion of the long-term store" (p83). Specifically, STS can be seen as a state of excitation within LTS. For an attempt to quantify the flow of information from sensory input via VSTM and STM to LTM, and then out again via motor action, see Frank (1963).

References

Atkinson, R. C., & Shiffrin, R. M. (1965). Mathematical models for memory and learning. Technical Report 79, Institute for Mathematical Studies in the Social Sciences, Stanford University. In P. Kimble (Ed.), Proceedings of the Third Conference on Learning, Remembering, and Forgetting, New York: New York Academy of Sciences.

Atkinson, R.C. & Shiffrin, R.M. (1968). Human memory: A proposed system and its control processes. In Spence, K.W. & Spence, J.T. (Eds.), The Psychology of Learning and Motivation, New York: Academic Press.

Atkinson, R.C. & Shiffrin, R.M. (1971). The control of short term memory. Scientific American, August 1971, 225(2):82-90.

Baddeley, A.D. (1966). The influence of acoustic and semantic similarity on long term memory for word sequences. Quarterly Journal of Experimental Psychology, 18:302-309.

Baddeley, A.D. & Hitch, G.J. (1974). Working memory. In Bower, G.H. (ed.), Recent Advances in Learning and Motivation (Volume 8), New York: Academic Press.

Bliss, J.C., Crane, H.D., Mansfield, P.K., & Townsend, J.T. (1966). Information available in brief tactile presentations. Perception and Psychophysics, 1(8):273-283.

Conrad, R. (1964). Acoustic confusions in immediate memory. British Journal of Psychology, 55:75-84.

Conrad, R. & Hull, A.J. (1964). Information, acoustic confusion, and memory span. British Journal of Psychology, 55:429-432.

Darwin, C.J., Turvey, M.T., & Crowder, R.G. (1972). An auditory analogue of the Sperling partial report procedure: Evidence for brief auditory store. Cognitive Psychology, 3:255-267.

Douglas, R.J. (1967). The hippocampus and behaviour. Psychological Bulletin, 67(6):416-442.

Hebb, D.O. (1949). The Organisation of Behaviour. New York: Wiley.

James, W. (1890). The Principles of Psychology. New York: Holt.

Lorente de No, R. (1938). Analysis of the activity of the chains of internuncial neurons. Journal of Neurophysiology, 1:207-244.

Maudsley, H. (1876). The Physiology of Mind. London: MacMillan.

Miller, G.A., Galanter, E., & Pribram, K.H. (1960). Plans and the Structure of Behaviour. New York: Holt.

Müller, G.E. & Pilzecker, A. (1900). Experimentelle Beiträge zur Lehre vom Gedächtnis. Zeitschrift für Psychologie, Ergänzungsband, 1:1-300.

Murdock, B.B. (1971). Short term memory. In Bower, G. (Ed.), The Psychology of Learning and Motivation (Volume 5), New York: Academic Press.

Paivio, A. & Csapo, K. (1969). Concrete image and verbal memory codes. Journal of Experimental Psychology, 80:279-285.

Peterson, L.R. & Peterson, M.J. (1959). Short-term retention of individual verbal items. Journal of Experimental Psychology, 58:193-198.