How Valid is the Modal Model of Memory?

Memory can be thought of as the ability to process and retrieve information. In one hundred years of scientific research on this subject, there is still no dominant theory which explains human memory. One of the most prominent models in this area over the past thirty years has been the modal model of memory, which was originally devised by Atkinson & Shiffrin in 1968. This model theorises a distinction between sensory, primary and secondary stores. Although this model continues to be highly influential, it has given rise to much discussion pertaining to its general validity and the need for the distinct memory stores. This essay will examine both supporting and opposing research and theories in order to establish how valid the modal model is and what its limitations are.

Historically, the first distinction to be made between primary and secondary stores was made by William James in 1890. Primary memory was outlined by James as being "that which is held momentarily in consciousness." Secondary memory he described as being "unconscious but permanent" (cited by Healy & McNamara, 1996).

A more contemporary description of the modal model was postulated by Glanzer & Cunitz, (1966, cited in Gross, 1992). Results of their extensive laboratory research into the existence of the dichotomy of memory stores has generally been presented using the serial position curve. When participants are given a list of items to commit to memory, their recall is usually better for items which appeared early in the list (the primacy effect) and late in the list (the recency effect), than it is for items in the middle of the list (the asymptote). This is represented diagrammatically shown below.

The primacy effect has been explained as being caused by the operation of the long term memory (LTM). In theory, the earlier items of the list will have been rehearsed and processed in the short term memory (STM) long enough to allow for transfer into the LTM (approximately 30 seconds). The recency effect is said to be characteristic of the operation of STM, as after the presentation of the final item, less than 30 seconds will have elapsed since presentation of the later items and hence these items will still be operating in STM. Recall will typically be poor on the middle items because they have been displaced following rehearsal in STM and will not yet have been consolidated into LTM. (Gleitman, 1995).

Further compelling evidence for the existence of separate stores was discovered by Brown (1958), and Peterson & Peterson (1959), (cited in Gross, 1992), who used a similar methodology which has now become known as the Brown-Peterson technique. They both investigated the capacity of the short term memory (STM) using a technique which involved presenting subjects with trigrams, which are three letter nonsense syllables, for a short time and testing recall after a specific retention interval. Results showed that items are rapidly forgotten when the process of rehearsal is prevented by a task such as counting backwards. Almost 70% is lost after a delay of 9 seconds, and 90% is lost after a delay of 18 seconds (Gross, 1992). These results have been put forward as evidence of the existence of a STM. The loss of information has been explained in terms of the decay of the STM trace due to the presence of the distracter task (Peterson & Peterson, 1959, cited in Gross, 1992).

Other findings which also made distinctions between primary and secondary memory stores came from evidence from patients with brain damage. If the STM and LTM are indeed separate stores, there should exist evidence of certain kinds of brain damage which effects one store without impairing the other. Milner et al (1978) studied a patient which he referred to as "HM" who had a defective LTM although his STM appeared normal. Another amnesic patient - Clive Wearing, was almost completely incapable of transferring information from the STM to the LTM. The problems which the amnesic patients suffered were not "some general ...

This is a preview of the whole essay

It was Atkinson & Shiffrin (1968) who presented the most abundant and perhaps the most widely recognised version of the modal model when they added a further dimension to the two store dichotomy by introducing the concept of sensory memory. Only a tiny fraction of the enormous amount of sensory information received from the environment is thought to be selected for further processing. Visual information is thought to be retained in the iconic store, first postulated by Sperling (1960, in Gross, 1992), and auditory information is thought to be retained in the echoic store (Neisser, 1967, in Gross, 1992). The sensory, STM and LTM can be thought of as being the structure of memory whilst the control processes (mainly, rehearsal, but also attention and coding) constitute the processes which run within that structure.

According to Atkinson & Shiffrin, the sensory memory acts as a "buffer" in which rehearsal can take place in order to transfer information into LTM. The amount of information transferred into the LTM is contingent upon the length of time it has been rehearsed in the buffer. This model does go some way to explaining the serial position curve. The recency effect can be explained by the fact that those items are still in the buffer at the time of recall. The primacy effect is explained by the fact that the items at the beginning of the list stay in the buffer longer than latter items. As the buffer is empty at the beginning of the recall experiment, the initial items are not displaced by latter items until the buffer has reached its limited capacity. Once information has reached LTM it is thought to be relatively stable, with an essentially unlimited capacity for information.

However, evidence from Shallice & Warrington’s (1970) case study of "KF" created problems for the unitary STM proposed by the modal model. KF suffered damage to the left parieto-occipital region of his brain. This left him with an impaired STM. He did however, demonstrate normal LTM retention since his accident. If it is correct that all information which eventually arrives in the LTM must first pass through the STM, then surely if the STM is damaged then access to the LTM would be rather arduous? (Shallice & Warrington, 1970).

An early alternative to the modal model was proposed by Craik & Lockhart in their levels of processing model, (1972, in Gross, 1992), who argued that access to LTM is determined by the process of encoding. Shallow processing at the level of physical and phonetic features creates weaker memory traces than deeper processing at the level of semantic encoding. They argue that it is the level or depth of processing which determines the rate of forgetting, not the limited capacity of internal structures.

The controversial dichotomy between the separate stores still exists today with several authors claiming evidence against their existence. Robert G. Crowder (1993) in his article Short-term memory: where do we stand? stated that "the popularity of short-term stores grew during a time when we were busy inventing such storage receptacles. Nowadays that attitude seems archaic and, to some of us, even downright quaint" (Crowder, 1993). Baddeley and Hitch (1977), found evidence which refutes the claim that STM is responsible for the recency effect as they discovered recency effects in LTM. They discovered a recency effect amongst members of a rugby team trying to retrieve the names of teams whom they had recently played against. (Baddeley & Hitch, 1977).

Despite the fact that the modal model has been described as containing a unitary STM, Atkinson & Shiffrin (1968), did draw a distinction between the control processes, (mainly rehearsal) in the STM. This distinction lead the way for ensuing theories which have divided the STM into independent systems.

Perhaps the most influential model of separate primary memory systems is Baddeley’s working model of memory (Baddeley & Hitch ,1974). Similarly to Craik & Lockhart, Baddeley & Hitch also argued that the structural approach was an inadequate way of describing the diverse processing involved in STM. The working model of memory places emphasis on active processing rather than passive storage, and the single STM store is replaced by a number of separate processors. This system comprised an attentional system of limited capacity which was named the primary acoustic store, a central governing system, called the central executive, which was supported by two further systems: the articulatory loop and the visuo-spatial sketchpad. Like the STM in the modal model, the articulatory loop retains information via a process of rehearsal. The visuo-spatial sketchpad deals with visual and spatial material. Baddeley has shown how his system of a compartmentalised STM can highlight how certain cognitive functions are affected by various brain defects. Baddeley’s research suggested that patients with Alzheimer’s disease suffered because of an impairment within the central executive rather than the articulatory loop or the visuo-spatial sketchpad, (Healy & McNamara, 1996).

However, Baddeley’s model has also been criticised for its limitations in explaining STM. The main problem is that the component we know the least about (the central executive), is the most important. Allport (1980) argues that the idea of a central processor should be replaced with a system containing specific processing mechanisms, (cited in Gross, 1992). However, Eysenck (1986) argues that a central system is a necessity, as a system of specific processors would be "chaotic", (Eysenck & Keane, 1995).

The emphasis which Atkinson & Shiffrin place upon rehearsal in their modal model has been criticised as being non-essential and over generalised (Craik & Watkins ,1973, cited in Gross, 1992). Craik & Watkins proposed that retention of information over long periods was not necessarily related to the length of time in the STM buffer or the number of rehearsals. It was Craik and Watkins (1973, cited in Gross, 1992), who first outlined the necessity for different types of rehearsal. Elaborative rehearsal; which creates associations with existing knowledge to add meaning, they discovered was more effective than the simple maintenance rehearsal put forward by the modal model.

The modal model has also been attacked by Gruneberg (1970, cited in Gross), who argued that the features which distinguish the separate memory stores could in fact be the features of any memory system, this does not actually prove the existence of the separate stores.

A further model offering a different approach to that of the modal model is the hierarchical network theory, Collins and Quillian (1969, cited in Gross), proposed that knowledge is stored in a hierarchical network of semantic links. This network comprises a structure of pieces of information with interconnecting pathways. The links are representative of semantic associations between the connections of nodes and pathways. According to Collins & Quillian, retrieval involves searching through this semantic hierarchical network. Consequently, the more levels of the hierarchy which need to be searched, the longer it should take to access the required information. The authors tested their ideas by using sentence verification tasks. Unfortunately, it seems that yet again inadequacies have been found in the proposed model. Rips, Shoben & Smith (1973) demonstrated instances which prove the model’s theories incorrect.

Evidence has also been presented which emphasises an additional STM store for conceptual memory. Potter (1993, cited in Healy & McNamara, 1996) proposed that this memory system is a separate distinction and is used for very short term retention of visual conceptual information, for example a line of text. This conceptual STM has a larger capacity than the simple STM but the information rapidly decays unless it can be associated with a concepts already stored in LTM.

Many extensions and revisions have been made to the modal model of memory over recent years. One particular extension of the original model is the Search of Associative Memory (SAM) model, (Raaijmakers & Shiffrin 1981, 1982, cited by Healy & McNamara, 1996). The SAM model expands the modal model in its description of the procedures involved in searching for and retrieving information from LTM. In the SAM model information is stored in LTM in image form. LTM is accessed by a series of retrieval cues, which comprise the process of remembering.

The potency of a retrieval cue is contingent upon its association to a given image in LTM which is determined by prior knowledge and by the coding and rehearsal procedures operating in STM.

The modal model assumes the storage in the LTM is more or less permanent. Evidence supporting these theories of durability come in the form of several longitudinal studies. One particular study was conducted by Bahrick (1984, cited in Healy & McNamara, 1996) who studied the retention of knowledge over a 50 year period. Bahrick studied over 700 subjects who had learnt Spanish. Bahrick found that there was a rapid decline of information in the first 6 years, followed by a period of stability of around 30 years, followed by an eventual decline which may have been attributable to ageing. As a result of his studies Bahrick introduced the idea of a permanent LTM which he termed "permastore".

In summary, much research has been undertaken, mainly in the area of free recall, examining the primacy, asymptote and recency effects in order to explore whether or not there is a dichotomy of memory stores. Much of this research and also the research into patients with brain damage has supported the view of separate stores. However, counter evidence has been presented by various authors which refute these ideas. Several alternative models have been proposed which both revise and expand the original thinking, however, none of these totally refute the modal model in its entirety. The biggest criticism of the modal model is perhaps that it is too simple an explanation for such a complex structure such as human memory.

So, is the modal model of memory therefore valid? This essay has examined the validity of the modal model in terms of its contribution to thought as a scientific paradigm upon which much further research has been based. In extending and revising the original model many of the alternative paradigms have consequently answered more questions and in greater detail. But, none of these alternative theories have come anywhere near answering all that science wishes to know about memory either. As a result of examining the evidence in this essay it remains that the modal model has established itself as the basis for much current scientific thought and for that reason alone, the model does hold a considerable amount of validity.

References

Baddeley, A.D. and Hitch G.J. (1977) Recency Re-examined. In Attention and Performance, ed. S Dornic, 6:647-67. Hillsdale, NJ: Erlbaum.

Crowder, R.G. (1993) Short Term Memory, Where Do We Stand? Memory and Cognition. 21: 142-45.

Eysenck, M. W. & Keane, M. T. (1995) Cognitive Psychology: A Student’s Handbook. (3rd ed), Hove: Erlbaum.

Gleitman, H. (1995) Psychology. (4th ed). Norton.

Gross, R. (1992) Psychology The Science of Mind and Behaviour. (2nd ed). London: Hodder and Stoughton.

Healy, A.F., D.S. McNamara. (1996) Verbal Learning and Memory: Does the Modal Model Still Work? Annual Review of Psychology. 47: 143-72.

Parkin, A. (1987) Memory and Amnesia, An Introduction. Oxford: Blackwell.

Rips, L.J., E.S. Shoben and E.E.Smith (1973) Semantic distance and the verification of semantic relations. Journal of Verbal Learning and Verbal Behaviour. 12, 1-20

Shallice, T. and E.K. Warrington (1970) Independent Functioning of Verbal Memory Stores: A Neurophysiological Study. Quarterly Journal of Experimental Psychology. 22. 261-73