One’s memory is crucial in one’s understanding of one’s self, of the environment, one’s understanding of knowledge, with past and present experiences and with one’s identity; according to Blackmore; “without memory, we would be servants of the moment, with nothing but our innate reflexes to help us deal with the world.” He furthered this statement, commenting that without memory there would be no language, science, art, or culture, “civilisation itself is the distillation of human memory.” Gross (2009) Pg282. This essay will explain Atkinson and Shiffrin’s multi store model of memory, which is regarded as an influential model, and some of the further important research which followed it. The strengths and weaknesses of the model will be explored as well as the variety of responses to its findings, which have consequently arisen.
Atkinson and Shiffrin (1968-71) developed the multi store model to explain their theory as to how we process information and is sometimes called the duel model of memory due to its focus on short and long term memory stores. Their theory was that one’s memory involves a sequence of three stages; sensory memory (SM), short term memory (STM) and long term memory (LTM), and that information can pass through one stage to the other respectively.
the mult-store model
Sensory memory; this is the first stage in the storage, processing of information. Information comes directly from ones senses; sight (iconic), sound (acoustic), smell, touch (tactile), or tastes, and Sensory memory stores a literal copy of a model of the environment as experienced by ones sensory system. Sensory memory has a storage system for each of our senses and can hold a literal copy of this information momentarily, visual (iconic) information can be stored for up to 0.5 seconds, and sounds (acoustic) information up to 2 seconds, before either being forgotten, over written by new information, or passed on to STM. Their theory was that unless attention is given to information in the sensory memory store (SM), it will be forgotten. For example on your way into work you will see many different cars but unless anything is unusual about them in any way they will not be remembered. The memory system is bombarded with information from our senses and sensory memory (SM) acts as a filter, only passing on important, threatening, meaningful or unusual information, Lloyd (1984) remarked that “less than one-hundredth of all the sensory information that impinges on the human senses every one second reaches consciousness. Of this, only about five percent is stored permanently” Gross (2009) Pg284. STM and LTM can be assessed in terms of capacity, duration and coding: capacity, how much information can be retained, duration, the amount of time the information can be stored, and coding, how sensory information is incorporated within the memory system. The multi store theory was that information in STM is encoded mainly acoustically; visual coding also takes place to a lesser extent. Once transferred into STM; information can be stored for only a short amount of time, up to thirty seconds, although through rehearsal of information, it can be stored for longer. Rehearsal is a control process which serves crucial functions with regards to the multi store model. Firstly, it acts as an intermediate station between sensory memory and LTM by maintaining incoming information within STM, secondly, through rehearsal, information can be transferred into LTM. STM is seen as a work space which deals with conscious thought; solving problems and manipulating ideas for example. Incoming sensory information is scanned for matching information in LTM. If a match is found the information can be passed on to STM with a verbal tag from LTM. LTM is thought of as a being able to potentially hold a limitless amount of information for an indefinite period of time. The multi store model puts forward the idea that LTM holds information until it is needed. LTM can be explained as like a vast storage area which holds many different types of information which are not in use, but which are potentially retrievable. Bower (1975) remarked that different information stored within LTM includes; “a spatial model of the world around us, knowledge of the physical world, physical laws, properties of objects, beliefs, ..,norms, values.., goals”, motor skills, problem solving skills” Gross , Pg285. Other information held within LTM also includes the understanding of language, our plans for achieving, and interpretations of music. LTM stores verbal information mainly semantically (information given meaning) and has the ability to store information visually (iconic).
The multi store model generated a great deal of interest, and further research. In 1960 Sperling carried out experiments to measure the duration of iconic (visual) information in sensory memory. Participants were flashed a grid of letters, four long three down, for half a second. A different sound was made to indicate which row of letters was to be recalled, a clap for row one, a whistle for row two and so on. When the sound was made, participants recalled eighty per cent of the letters correctly. This was used as proof that as participants successfully recalled a large percentage of the letters, they must have all been held visually within SM; SM is therefore a large storage area but the information decays quickly, 250-500 milliseconds.
In his article ‘The magical number seven, plus or minus two’, published in 1956, Miller presented his findings concerning STM capacity. Miller found that the method of ‘chunking’ could be used in order to expand STM capacity. If one has the ability to remember seven letters from a list, each letter would occupy one of seven available storage areas within STM, but if the letters could be collectively coded, into words for example, then instead of one letter occupying each storage area, one could fit a whole word. To enable the chunking of information there needs to be a code. If one was asked to remember this sequence of numbers; 010678, each number would take a storage area within STM, but if this sequence could be given meaning, some ones birth date for instance, then the whole sequence would only take one storage area within STM. Giving meaning to information can only be done by using formally established storage areas (from LTM). Although this method can expand the amount of information that can be stored within STM, the same, seven plus or minus two rule still applies. But in this way one could fit an unlimited amount of related information in each storage area.
The Brown and Peterson Technique; Brown (1958) and Peterson (1959) developed a technique in order to research ‘pure’ STM; the idea behind the technique was that as information could be stored in STM indefinitely with rehearsal; to see how long information could be stored in STM without this aid.
the primacy recency effect
This was done by giving verbal trigrams, nonsense syllables, such as ‘nkf’ to participants, some of which were given distracter tasks; they were asked to count backwards from a particular number, for a specific amount of time, between three and eighteen seconds, to hinder rehearsal. They found that when asked to recall the trigram, participants would forget more information, the longer rehearsal was prevented. For example ninety percent of information was forgotten after eighteen seconds counting backwards, but the percentage of information recalled was high with a short distracter task. Without rehearsal, the information stored within STM, acoustic, and visual information, can only remain there for 15-30 seconds.
There are also case studies on brain damaged patients which lend support to the concept that humans have separate memory stores. Clive Wearing, a world expert on Renaissance music, chorus master of the Sinfonietta, and BBC radio producer, developed Herpes Encephalitis on March 29th 1985, a virus which left him with sustained damage to his Hippocampus and partial damage to his frontal and temporal lobes. As a result he developed total amnesia (anterograde amnesia and retrograde amnesia). Due to the damage to the Hippocampus, Wearing is unable to encode or store information in LTM, leaving him trapped in a snap shot of time. His memory lasts no longer than a few minutes at a time and he describes his condition as ‘hell on earth’. Mr Wearing remembers little about his life before his illness. For example, he does remember that he has children, but cannot remember their names. Interestingly, he does have memories associated with LTM. He has not forgotten how to play the piano, although he has no recollection of being taught how; nor has he forgotten the love he has for his wife Deborah, but each time he sees her, he greets her as if he has not seen her for years even though she may have left the room only for a moment. He keeps a diary in which he writes the same line over and over; “I am awake for the first time” every few minutes, he crosses the lines out and writes the it again, next to the correct time, being unable to contemplate that he could have written the previous lines. This pattern has been repeating itself for decades. The Clive Wearing case study has been used as evidence to support the multi store model. This case study seems to show the STM and LTM working as separate memory units. It shows Mr Wearing’s STM working normally whilst his LTM is severely damaged. One can see from these findings that one strength of the multi store model is that it has generated a great deal of interest and further research which appears to support and substantiate it. The multi store model of memory also followed on from previous works, it incorporated earlier ideas and developed them. There is evidence to suggest that there is separate STM and LTM components, which store and encode information differently within the brain.
Fundamental criticisms and evidence to dispute Atkinson and Shiffrin’s multi store model have now been put forward; some of which will now be explored. If one looks at the Clive Wearing case study it is apparent that some aspects of his LTM are still functioning. This would suggest that LTM is more complex than merely a unitary memory store. In reaction to the multi store model, Craik and Lockhart have asserted that the idea of a single LTM is not adequate. Remembering something for an hour is very different from remembering something for years, or decades. They argued that there must be more memory stores than just a single STM and a LTM. It is now thought that stored information within LTM must be coded in more ways than just semantically and visually. It is now widely thought that information must also be encoded by smells and tastes as well for example. In 1975 Tulving asserted that the fact that we can remember smells, tastes, sounds such as songs, and visual information, such as places visited, can be seen as proof that LTM has the ability to store information in all ways; it is a flexible, vast, infinitely long lasting processing facility. Other types of LTM include: Episodic (EM), memories of events such as what one ate for dinner yesterday, as well as semantic (SM*), information which holds meaning, which Tulving has identified as declarative information, ‘knowing that’, and procedural (PM), ‘knowing how’, knowledge of how to drive a car for instance. The multi store model is also been accused of being too passive, rigid, and simplistic, and the idea that information flows in a linear fashion, from one storage facility to another in one direction has subsequently been challenged; Atkinson and Shiffrin represented the LTM store as a passive unit, just storing information until required. This idea is considered to basic by many; LTM and STM seem to actually work together. If one looks at Millers idea of ‘chunking’ information in STM for example, it is apparent that in order to give these letters meaning, one requires prior knowledge of the words. LTM must be working with, or in fact directing STM in order to give information meaning. Another criticism of the multi store model is the importance given to rehearsal; the model proposes that rehearsal as the crucial factor it the retention of information within STM, and of the duration of time that information remains in STM, as well as the idea that the amount of the rehearsal of information is the trigger for the subsequent transfer of information into the LTM store. Eysenck and Keane (1995) remarked that unrehearsed information is constantly entering LTM, there have now been a number of tests to back up their concerns regarding this oversight in the multi store model. Tulving conducted an experiment in 1967, in which participants first rehearsed a list of words, by reading them over repeatedly. These words were then incorporated within a larger list. The participants recalled these rehearsed words no more frequently than the new unrehearsed words. This experiment seemed to directly contradict the idea that rehearsal had any importance in the likelihood of recollection. The multi store model has been falsified by the levels of processing theory. Craik and Lockhart devised a theory of how memory is processed, levels of processing, which contradicts the multi store model in a number of fundamental areas. They showed that it was how deeply information was processed rather than rehearsal, which was the important factor in the retention of information. They showed participants groups of words for short periods of time and asked them questions about the words which would ensure that the words would be processed internally on different levels. For example, the word ‘DOG’ was shown; participants were then asked, was the word writen in capital letters. This question insured only shallow processing. The word, ‘castle’ was shown, the following question asked was, does this word rhyme with tin? The question encouraged intermediate mental processing. To prompt deep mental processing, participants were shown the word, ‘tiger’, then asked, was the word the name of an animal, this question insured that the word was processed semantically. The words were twice as likely to be remembered if the word had been processed deeply rather than processed in a shallow manner. Craik and Tulving expanded on this in 1975 when they asserted that the more complex the processing, the more likely it is that information will be held in LTM. The multi store model has also been attacked for not taking every day, real life memory, into account, it has been remarked that learning unimportant information, like recalling meaningless letters in a laboratory, does not reflect how the memory system would react with real life, more meaningful information. The model has also been criticised for paying attention to the structure of the memory system, how information is transferred, and how much, but not enough on how it is encoded or processed. Atkinson and Shiffrin remarked that information such as the memory of a smell for instance, would be stored acoustically in STM; this appears to be an inadequate explanation of how information is encoded and has since been disputed.
There have been case studies on brain damaged patients that challenged many of Atkinson and Shiffrin’s ideas regarding memory. The case study of KF, conducted by Warrington and Shallice between 1969 and 1972 raise a number of questions. After a motor bike accident KF was left with brain damage (inferior parietal lobe damage). He was left with anterograde amnesia. His memory of events before his accident was not impaired; however, his memory of events after his accident was seriously impaired. Tests showed that KF had damage to his STM. Further tests showed that KF’s STM was greatly unaffected in relation to visual information and of semantic sounds. It was auditary; spoken information such as words which was impaired. These findings contradicted Atkinson and Shriffrin’s model of memory in a number of crucial ways. The fact that KF could recall past events despite damage to his STM compromises Atkinson and Shiffrin’s multi store model. They asserted that recalled information from LTM is passed through to STM for use; if this were the case, KF would not be able to use information from LTM or STM. The KF case study could also be used as an example for the argument that there is more than one type of STM, contrary to Atkinson and Schiffrin’s assertions of a unitary STM. In 1997 Baddeley and Hitch put forward their own, more complex model of memory, the working memory model (WM). They argued that STM is a multi component system which is run by a ‘central executive’; this is involved with conscious thought, it also controls sub systems which work independently of each other and hold information in different ways. It is now widely believed that STM is much more complex than Atkinson and Shiffrin’s multi store model explanation.
Although many aspects of the multi store model have been discredited, the idea of a separate short and long term memory store remains fundamentally acknowledged. The multi store model is understood to be the first widely accepted model of memory. There are now more complex models, such as Craik and Lockhart’s levels of processing and Baddeley and Hitch’s working model of memory; which have been put forward in continuation, or in reaction to Atkinson and Shiffrin’s work. The continued research into the area of memory has contributed to the greater understanding of the complex processes involved within the human memory system, which has, and continues to have a great impact within the modern world.
Cardwell M, Clark L, Meldrum C. (2003) Psychology for A Level, Harper Collins.
Gross R. (2009) Psychology The Science of Mind and Behaviour 5th Edition, Hodder Arnold.
Gross R, Rolls G. (2003) Essential AS Psychology, Hodder & Stoughton.
(2010) www.google.com/http://psychology.wikia.com/wiki/Atkinson/Shiffrin/memory/model (accessed 26th April 2010)
(2008) www.google.com/http://scienceaid.co.uk/psychology/cognition/multistore.html (accessed 26th April 2010)