Nonsense and droodles

Nonsense and droodles
Meaning has a major influence on memory. Ebbinghaus (1964/ 1885) recognized that the study of material which already had meaning for the learner would be influenced by that meaning. So it seemed to Ebbinghaus that if he was to discover the fundamental principles of memory, then he would need to study the learning of simple, systematically constructed materials. He created syllables by stringing together a consonant sound, a vowel sound and a consonant sound. Some of these were words or meaningful parts of words but most were simply syllables. He made lists of these syllables and learned them in order – often requiring many trials to learn them perfectly. In contrast to his experience learning poetry, learning these syllables was slow. A demonstration of the importance of meaning for the recall of very different material was provided by Bower, Karlin and Dueck (1975). They studied memory for droodles – simple line drawings of nonsense pictures. Some participants were given a meaning for each droodle (e.g. a mi get playing a trombone in a telephone booth; an early bird who caught a very strong worm). These individuals were able to sketch the pictures from memory far better (70 per cent correct) than participants who were not given these meanings (51 per cent correct).

IMPORTANCE OF MEANING

THE IMPORTANCE OF MEANING
Remembering names Meaning plays a major role in determining what we can remember. Consider the case of remembering (or rather forgetting) names. People who feel they have a bad memory commonly complain that they find names especially difficult to remember. In fact, people are generally poor at dealing with a new name. When introduced to a new person, our minds are usually occupied and so we fail to attend to their name. Then we most likely do not use or try to think of the name until much later, by which time memory often fails. But there is more to the problem of remembering names than merely not paying attention and not using the names until much later. Cohen and Faulkner (1986) presented participants with information about fictitious people: their names, the places they came from, their occupations and hobbies. The participants remembered all of the other attributes better than the names. Why? Not merely because names are unfamiliar words – many names are also common nouns (e.g. Potter, Baker, Weaver, Cook). McWeeny, Young, Hay and Ellis (1987) tested people who studied the same set of words; sometimes the words were presented as names, sometimes as occupations. The same words were remembered much better when presented s occupations than as names. It is apparently easier to learn that someone is a carpenter than that they are named Mr Carpenter! Nevertheless, names that are also real words do have an advantage over ‘non-word’ names. Cohen (1990) showed that meaningful words presented as names (e.g. Baker) are better remembered than meaningless words presented as occupations (e.g. ryman). Even so, names are often treated as being meaningless – think for a second how it sometimes comes as a surprise when we recognize that they are also occupations (for example, the names of the former British prime ministers Thatcher and Major). We know that attending to the meanings of names can improve memory for them, especially when combined with practice in recalling them (Morris & Fritz, 2002, 2003). One aspect of what makes a word meaningful is the associations that it has with other terms (Noble, 1952). Words that trigger more associated words (e.g. ‘kitchen’) certainly seem more meaningful than unusual words (e.g. ‘rostrum’) and the e, in turn, seem more meaningful than non-words (e.g. ‘gojey’). The lack of associations to some names may be one of the main reasons they are hard to learn. Cohen and Burke (1993) point out that many names lack semantic associations, while occupations have many semantic associations.


[Hermann Ebbinghaus (1850–1909), a German philosopher, read Fechner’s work on the study of sensation and perception in the late 1870s and decided to adapt these methods to the study of memory. He devised a systematic way of simplifying memory tasks so that aspects of memory could be manipulated and measured. Ebbinghaus invented syllables made up of two consonant sounds separated by a vowel (e.g. ‘tir’, ‘kam’, ‘dol’) in an attempt to avoid the contaminating effects of prior familiarity, and then measured the number of repetitions required to learn them. He also devised a clever way of measuring forgetting. He counted the number of repetitions required to relearn the material and found that it usually took fewer repetitions to re-learn something than to learn it in the first place. Ebbinghaus’s experimental method for the study of memory established a major field of psychology and continues to influence our understanding of memory today.]

MEMORY AND THE BRAIN

MEMORY AND THE BRAIN
Psychologists’ study of memory has focused, appropriately, on what people do, say, feel and imagine as a result of their past experiences. But how are these activities of remembering reflected in our brainThe study of amnesia has been important in recent years, not only as a way of discriminating between certain types of memory processes, but also in linking deficits in remembering with localized brain damage in patients who have sustained injury. In addition, the development of techniques such as functional magnetic resonance imaging (fMRI) has added significant new information by allowing us to study the parts of the brain that are active when ordinary people remember. For an excellent review of this research, see Parkin (1997). Making generalizations about memory and the brain is difficult because remembering is a complex process, involving most other cognitive and emotional aspects of a person. So many parts of the brain will be active when someone is remembering. We cannot just remember something without also feeling and thinking, so it is very hard to isolate any neural activity that might be unique to remembering. But certain parts of the brain do seem to be important to memory in particular. For example, damage to the hippocampus and the thalamus can prevent new episodic memories being formed (Squire, 1992). Patients with hippocampal damage can learn new skills without forming episodic memories. So the patient who had had his hippocampus surgically removed, was eventually able to solve a complicated puzzle that he attempted over many days. Yet each time he was given the puzzle, he denied having ever seen it before (Cohen & Corkin, 1981). This tells us that the hippocampus appears to play an important part in the formation of episodic memories.

Transfer appropriate processing

Transfer appropriate processing
To achieve the best recall, the type of processing involved when studying needs to be appropriately matched to the type of processing that will be required for the test. Morris, Bransford and Franks (1977) demonstrated the effect of transfer appropriate processing in an extension of the Craik and Tulving (1975) ‘levels of processing’ experiments. In the original Craik and Tulving studies, participants were encouraged during encoding to focus on the physical, phonological (e.g. rhyming) or semantic aspects of the to-be-remembered word. Under typical testing conditions, semantic processing during encoding led to the best level of recall during testing. But in the Morris et al. study, a condition was added in the test phase: participants had to identify words that rhymed with the words presented earlier during encoding. In this new condition there was a closer match between the task carried out in the learning phase (identifying words that rhymed) and the task carried out in the test phase (identifying the word that rhymed with words presented in the learning phase). Recall for rhyming words was best when rhyming had been the focus of the learning task.

LINK BETWEEN STUDY AND TEST

LINK BETWEEN STUDY AND TEST
The encoding specificity principle Tulving (1983) developed the encoding specificity principle, (encoding specificity principle states that what is remembered later depends on the similarity of the retrieval situation to the original encoding conditions) which emphasizes the relationship between what occurs at study time (encoding) and what occurs at test time (retrieval). What is encoded in any particular situation is selective – it is determined by the demands on the individual at study time. According to the encoding specificity principle, what will be remembered later depends on the similarity between the memory test conditions and the original study conditions.


An experiment by Barclay et al. (1974) nicely illustrates encoding specificity. They required participants to study a series of sentences with key words embedded in the sentences. So, for example, the word ‘PIANO’ was presented in one of two sentences: ‘The man tuned the PIANO’ or ‘The man lifted the PIANO.’ Recall of the sentences was cued by phrases that were either appropriate or inappropriate to the particular attributes of the named object (the piano). Cued with the phrase ‘something melodious’, participants who had received the sentence about tuning the piano remembered ‘PIANO’. Participants who had studied the sentence about the piano being lifted were less likely to recall ‘PIANO’ after the ‘something melodious’ cue, because the melodious aspect of the piano had not been emphasized in their sentence. Conversely, participants who had studied the sentence about lifting the piano were more effectively cued at test by the phrase ‘something heavy’ rather than the cue ‘something melodious’. This experiment demonstrates two important aspects of encoding specificity:
1. Only those aspects of our experience that are specifically activated by the study situation are certain to be encoded.
2. For information to be optimally recalled, test cues need to target the particular aspects of the information that were originally encoded. In other words, remembering depends on the match between what is encoded and what is cued.


[Endel Tulving (1927– ) has been a dominant figure in research on memory for several generations and a pivotal figure in the late twentieth century. His work on subjective organization demonstrated that participants in memory studies are not passive but impose their own organization and expectations upon the material they study. He drew attention to a distinction, originally made by Plato, between the availability of items in memory and their accessibility. Tulving is even better known for his work on the relationship between what is encoded and what can be retrieved. He developed the encoding specificity principle and collaborated with Craik in exploring the ‘levels of processing’ framework. He was also the first psychologist to suggest that episodic and semantic memories were two separate memory systems. More recently, with Schacter, he has been involved in a considerable body of research and theorizing on the distinction between implicit and explicit memories and in research on the neuropsychological correlates of memory.]

LEVELS OF PROCESSING

LEVELS OF PROCESSING
Another alternative to the continuing development of structural models has been to emphasize the importance of processing in memory, rather than structure and capacity. Craik and Lockhart (1972; Craik, 2002) argued that how well we remember depends on how we process information. They described different levels of processing, from ‘superficial’ levels that deal only with the physical properties of what is to be remembered, through ‘deeper’ processes involving phonological properties, down to yet deeper processes that involve semantic processing of the material (i.e. perhaps involving elaboration of the material). So, for example, if we see the word ‘SHEEP’, we might simply process it shallowly by noting that it is written in upper case. On the other hand, we might process it phonologically by registering that its sound rhymes with ‘leap’ and ‘deep’. Alternatively, we could think about the meaning of the word: ‘sheep’ refers to domesticated, woolly, grazing animals. Further semantic processing – elaboration bas d on the meaning of the word – is deeper processing, and should lead to better memory (for example, we might think about the grazing of sheep, the uses of sheep – for example, in providing food and material for clothing – and the large number of sheep in some parts of the world, such as Australia and New Zealand). Demonstrating the power of this approach, Craik and Tulving (1975) showed that the probability of the same word being recognized in a memory experiment varies from 20 per cent to 70 per cent, depending on the type of processing that is carried out on the word. When the initial processing involves only decisions about the case in which the word is printed, correct recognition occurs at the 20 per cent level. Performance is better following the rhyming (i.e. phonological) decisions, and far better (almost 70 per cent correct recognition) when processing involves decisions about whether the word fits meaningfully into a given sentence. Although many studies support the model, the details of the original ‘levels of processing’ model have been criticized (e.g. Baddeley, 1978). For example, it has been argued that a level of processing cannot be identified independently of the memory performance that it produces (in other words, it has been suggested that the definition of what constitutes ‘deep’ and ‘shallow’ processing is circular). More recently, though, Craik (2002) has pointed to physiological and neurological methods that may provide an independent measure of depth. Thoughtful discussion about the viability of the model continues. Wherever it leads, it is clear that a ‘levels of processing’ approach draws attention to important memory-related issues including the type of processing, elaboration of materials, and the appropriateness of this processing (in terms of ‘transfer’ to the later task). A key message from this research is that what we remember depends on what we ourselves do when we encounter a thing or an event, as well as the properties of the thing or event itself.

The phonological loop

The phonological loop
Much research has been concentrated on the phonological loop. By using a technique known as articulatory suppression, in which research participants repeat aloud (or silently) a simple sound or word, such as ‘la la la’ or ‘the the the’, the phonological loop can be prevented temporarily from retaining any further information. So contrasting performance with and without articulatory suppression demonstrates the contribution of the phonological loop. Like any loop, the phonological loop has a finite length. That length could be specified as a number of items or as a length of time. Baddeley, Thomson and Buchanan (1975) investigated this question. They showed that memory span – the number of words that you can hear and then repeat back without error – is a function of the length of time that it takes to say the words. A word list like ‘mumps, stoat, Greece, Maine, zinc’ is much easier to remember in a short-term memory test than ‘tuberculosis, hippopotamus, Yugoslavia, Louisiana, titanium’, even though the two l sts are matched in terms of the number of words and the meaning. This word length effect is eliminated if the participants have to carry out articulatory suppression while they study the list.


Another example comes from the varying speed with which the digits 1 to 10 can be pronounced in different languages. The size of the memory span for people who speak each language is highly correlated with the speed with which the digits can be spoken in that language (Naveh-Benjamin & Ayres, 1986). These and other observations demonstrate that the phonological loop must be time-limited. The central executive and the sketch pad More recently, Baddeley and his associates have turned to studying the central executive. Their technique is to ask people to perform two tasks at the same time. One of the tasks (the first task) is designed to keep the central executive busy, while the second task is being evaluated for whether the central executive is involved in its performance. When performance on the second task suffers due to the presence of the first task, they conclude that the central executive is involved in performing the second task. One task used to engage the central executive is the generation of random etter sequences. Participants generate letter sequences taking care to avoid sequences of letters that fall into meaningful orders, such as (T, V), (B, B, C) or (U, S, A). Participants must attend carefully to their letter choice, and this monitoring occupies the central executive. Robbins et al. (1996) showed that the memory of expert chess players for positions taken from actual chess games was impaired by the letter generation task but not by articulatory suppression, indicating that the central executive was involved in remembering the chess positions. These researchers also found that another task which is believed to interfere with the visuo-spatial sketch pad also reduced chess performance, reflecting the contribution of spatial short-term memory in the reproduction of the chess layouts. The episodic buffer Information that is retrieved from long-term memory often needs to be integrated to be appropriate for the current demands upon working memory. This is an important function of the episodic buffer p oposed by Baddeley (2001). Baddeley gives the example of imagining an elephant who plays ice-hockey. We can easily go beyond the information about elephants and ice-hockey our longterm memory supplies us to imagine how the elephant holds the hockey stick and what position it might play. The episodic buffer allows us to go beyond what already exists in long-term memory, to combine it in different ways, and to use it to create novel situations on which future action can be based.