The field of memory and learning questions how we represent, store and retrieve information. It encompasses the domains of working memory (e.g. interference and decay of memories); long-term memory (e.g. retention and activation of memories); and skills acquisition (e.g. practice, procedural knowledge).
Both of these fields of cognitive psychology make considerable use of the experimental method in their research. Controlled laboratory experiments obtain samples of performance at a particular time and place. The techniques of these experiments, however, differs somewhat between the two areas and so allows us to compare and contrast them.
An example in memory, of an experimental laboratory study using ability to recall information as a means of analysis is that by Tulving, Schacter and Stark (1982, as cited in Eysenck and Keane, 1995). The aim of the experiment was to try to understand what is involved in implicit memory. Implicit memory can be seen at work when, for example, performance on a task is aided despite a lack of conscious recollection. Explicit memory also exists, and can be seen when performance on a task requires conscious recollection of previously learned information or experiences.
Subjects were asked to learn a list of multi-syllabled and fairly rare words (e.g. ‘toboggan’). Either one hour or one week later, they were asked to fill in the blanks in word fragments to make a word. Although subjects were unaware, the solutions to half of the fragments were words from the list they had previously been asked to learn. This is a test of implicit memory, as it did not require conscious recollection to be used on the word-fragment test. Evidence for implicit memory was found - subjects managed to correctly complete more of the fragments when the solutions matched words which appeared on the previously learned lists. This effect is known as ‘repetition priming’. It may be argued that this was merely a test of explicit memory, and repetition priming only occurred because subjects deliberately searched through the previously learned list. However, further evidence which rules out this possibility comes from Tulving (1982). He found that the repetition priming was no greater for list-words which were recognised than for those that were not recognised. This discovery suggests that two different forms of memory are involved in repetition priming and recognition memory.
A different kind of experimental laboratory study, with the focus on language processes, was carried out by Hardyck and Petrinovich (1970, as cited in Eysenck and Keane, 1995). The experiment looked at the role of inner speech in reading comprehension. Inner speech is the sounding-out of words and sentences silently in one’s head. Some argue that inner speech is of no value and is just a habit carried through from childhood (children learn to read out loud before they can read silently). Others believe it to be very important in aiding comprehension. Hardyck and Petrinovich took electromyographic (EMG) recordings of some of the muscles used in subvocal articulation (the aspect of inner speech involving movements in the speech tract). This is of importance as the level of EMG activity in the speech tract usually increases a great deal during reading. Subjects were asked to read easy and difficult texts whilst EMG recordings were being made. Each time the level of muscle activity in the speech tract rose above a certain predetermined ‘relaxation’ level, a tone sounded. Subjects were told to try and prevent the tone from being sounded. It was found that the reduction of EMG in the speech tract (i.e. the reduction of subvocal articulation) significantly impaired the comprehension for the difficult text. No impairment was found, however, in the comprehension of the easy text.
Differences between these two experimental studies can clearly be seen. Generally, memory experiments tend to focus on subjects’ reaction times and ability to recall information, whereas language experiments tend to present subjects with sentences, words, pictures and test understanding and recognition. Language-related studies (as with memory studies) also make use of reaction and processing times as a means of analysis.
A study in language, which attempts to make inferences from the analysis of error patterns, is that of McKoon and Ratcliff (1986, as cited in Eysenck and Keane, 1995). The minimalist hypothesis and the constructionist position are two opposing viewpoints in the area of language. Those who support the constructionist view claim that inferences are drawn in reading; those who support the minimalist hypothesis argue that there are constraints on the number of inferences that are generated automatically. Subjects read several short texts containing sentences such as “the actress fell from the fourteenth storey” (this would lead to the inference that she died from the constructionist viewpoint, but not the minimalist). They then completed a recognition memory test where they had to quickly decide whether or not certain words had been presented in any of the texts. Critical test words were presented that had not been seen before, but which represented inferences from the studied texts (e.g. ‘dead’ in the example of the actress falling). If subjects had made inferences, then they should incorrectly identify these words as having been in the studied texts. It was found that the number of errors on the critical test words was no higher than on control words when they were immediately preceded by the neutral word ‘ready’. When they were preceded by a word from the relevant sentence, however (e.g. ‘actress’), there was an increase in the number of errors to the critical test words. Therefore inferences were not fully generated (supportive of the minimalist hypothesis), but the fact that they were formed to a limited extent provides some support for the constructionist viewpoint.
Another widely used method in the area of memory is that of case studies. Case studies engage in the intensive study of single individuals. An example of a case study is that by Eslinger and Damasio (1985). They investigated a former accountant known as ‘EVR’ who had had a large cerebral tumour removed. The theory behind the study focuses on the ‘central executive’ and whether it exists as a unitary system. The central executive is the most important component of working memory; it is quite like an attentional system and is used for a variety of purposes (trouble-shooting, decision making, planning). It is also claimed to have a limited capacity. EVR has a high IQ and performed well on tests requiring reasoning, resistance to distraction and memory interference etc. This suggests that his central executive was still intact, however his judgement and decision-making skills were found to be very poor. Eslinger and Damasio interpreted these findings as showing that EVR’s central executive was partially intact and partially damaged. This implies that the central executive is not unitary, but consists of two or more component systems. Case studies are also implemented in the area of language, but perhaps to a lesser extent.
Computational models have been used to investigate the area of language. Patients with brain lesions in specific areas have been found to suffer from ‘semantic dementia’, having difficulty assigning objects to semantic categories (semantic refers to the meaning of a word). Some patients have trouble naming animals, foods etc whilst others cannot name objects. Correlations have been observed between the type of semantic deficit and the area of lesion, with it being suggested by Warrington et al (as cited in Gazzaniga et al, 2002) that patients’ problems are reflections of the types of information stored with different words in the semantic network. Biological categories are said to rely more on physical properties/visual features, man-made objects are thought to be identified by their functional properties. Farah and McClelland (1991, as cited in Gazzaniga et al, 2002) devised a model to test this idea of a modality-specific organisation of the semantic network. They modelled semantic memory as being comprised of separate visual and functional sub-systems, and also included the aforementioned idea that living things must include representations based on visual attributes whereas non-living things include information about functional attributes. The findings were consistent with the hypothesis put forward by Warrington et al; if visual properties in the model were ‘lesioned’ (by removal of nodes in their network in the computer simulation), the model was particularly impaired in dealing with living things. Similarly, ‘lesioning’ of functional properties led to impairments with non-living things.
Cognitive neuroscience has been increasingly used to study the area of memory and learning. Squire et al (1992, as cited in Eysenck and Keane, 1995) used PET scans (positron emission tomography) to investigate the brain structures involved in explicit memory (encountered earlier in the essay). They found that blood flow in the right hippocampus was considerably higher when subjects were performing an explicit memory task rather than an implicit memory task. This provides support for the view that the hippocampus plays an important role in explicit memory.
Paller, Kutas et al (1995, as cited in Gazzaniga et al, 2002) took electrical recordings (event-related potentials, or ERPs) of brain activity to assess whether performance on implicit tests is correlated with explicit recollections about test items. The aim was to reveal how much recollective experience patients might have for previously viewed information, even though they are not required to recollect anything about that information. The method used involved varying the depth of encoding words by having subjects generate a mental image corresponding to the word, or simply indicate whether the word contained more than one syllable. The subjects were later asked to make a lexical decision (is it a word?) about a list of words and nonword letter strings. The words consisted of those used in the previous image task list, the previous syllable task list, and new words not on any previously seen lists. In a second experiment, subjects were asked to perform the lexical decision task and give a recognition judgement about whether they had seen the words before. Each experiment involved administering separate recognition post-tests, in order to assess the extent of subjects’ recollection of words. Behavioural measures showed priming (as discussed earlier) in both conditions (image vs syllable). Post-test recognition scores however, differed significantly in that those who generated an image when learning a word were better at recognising previously seen words. It was discovered that the voltage of ERPs elicited by words was more positive 500 to 900 msec after the onset of words that had been seen previously in the image condition, as compared with the ones in the syllable condition – a physiological sign of recognition.
McKee and Squire (1992, as cited in Eysenck and Keane, 1995) found that amnesiac patients with medial temporal lobe lesions showed similar forgetting rates to amnesiac patients with diencephalic lesions at retention intervals of between 10 minutes and one day. These findings are the basis of the argument put forward by Squire et al (1993) – that the diencephalon and medial lobe structures are of equal importance to explicit memory.
The area of language, however, also makes use of the techniques of cognitive neuroscience - PET scans, ERPs, lesion studies and fMRI scans. An example of cognitive neuroscience as used in language but not memory, however, can be seen by looking at split–brain research. This involves severing the corpus callosum – the structure that joins the left and right hemispheres of the brain. Studies aim to determine which skills are confined to which hemisphere. Sperry (1968, as cited in Carroll, 1994) presented a patient with a picture of a spoon in her left visual field. She reported to see nothing. When asked to select the spoon with her left hand from an array of common objects that were out of sight she did so correctly. When asked what she was holding, she responded ‘nothing’. When asked to reach for the object with her right hand she performed at chance level, and was equally as likely to pick up a pencil as a spoon. This shows us that the left hemisphere predominantly controls speech, whereas the right hemisphere is able to communicate awareness of a stimulus in nonverbal ways (since information from the right visual field projects to the left hemisphere and vice versa).
In conclusion, it can be seen that both the areas of language processes and memory and learning share numerous research methods in common. This can perhaps be explained when it is noted that the area of language also delves into the area of memory in its research – for example studies have been carried out testing whether language can affect memory for colours etc. However, the extent to which these methods are utilised in these areas differs. Language uses mainly experimental studies, where memory makes wider use of cognitive neuroscience techniques.
References:
Best, J.B. (1995). Cognitive Psychology. U.S.A. West Publishing Company.
Carroll, D.W. (1994). Psychology of Language. U.S.A. Brookes/Cole Publishing Company.
Eysenck, M.W, and Keane, M.T. (1995). Cognitive Psychology: A Student’s Handbook. East Sussex, UK. Psychology Press.
Gazzaniga, M.S, Ivry, R.B and Mangun, G.R. (2002). Cognitive Neuroscience: The Biology of The Mind. U.S.A. W.W. Norton & Company, Inc.
Groome, D. (1999). An Introduction to Cognitive Psychology: Processes and Disorders. Hove. Psychology Press.
Sekuler, R, and Blake, R. (1994). Perception. New York. McGraw-Hill.
Sternberg, R.J. (1996). Cognitive Psychology. London. Harcourt Brace College Publishers.