|
Chapter summaryLanguage is perhaps the most sophisticated of all human abilities. For this reason language is also possibly one of the most difficult areas to understand, despite the fact that there are numerous methods available to research language. In this chapter we consider some of the main elements of language, including reading and text comprehension. We then go on to examine a number of theories of the use of language and its relation to other cognitive processes, such as thought. How children acquire language is also discussed and several proposed theories are evaluated. Elements of languageSpoken language can be broken down into a number of elements, such as phonology, syntax, semantics, morphology and pragmatics. These are important aspects of language that you need to study before being able to fully appreciate the issues discussed in Sections 2 and 3. Language is a means of using words to communicate. However, it is also used to gather information from our environment. Anyone who has spent time in a country that speaks a foreign language is keenly aware of how difficult it is to communicate and obtain information about local affairs. Often when we are deprived of something, we fully appreciate its function and value. In order to study language one needs to understand the types of physical and linguistic information involved, such as speech and the written word. We begin this section with an analysis of phonology, syntax, semantics and pragmatics before considering more general theories of language. PhonologyThe study of linguistic sounds is known as phonology. It concerns the production of sounds made by the mouth, tongue, vocal chords and so on, whose combined movements produce speech. The physics of the raw sounds produced in speech is known as acoustics. The basic unit of speech is the phoneme. One way in which languages differ from each other is in terms of the sorts of phonemes that are used. Phonemes can consist of consonants, such as /d/, /b/, or vowels, such as /i/, /a/. Phonemes can also be voiced (/b/, /k/, /g/, etc.) or voiceless (/s/, /t/, etc.), and are characterised as such based on the location at which the initial sound burst is made, whether it is in the vocal chords (voiced) or within the mouth (voiceless). SyntaxA sentence consists of a number of words. The way words are put together to form sentences is known as syntax and it concerns the rules of language use, e.g. its grammar. The order of words within a sentence most often determines its meaning. For example, if we transpose two words in the sentence the dog bit the man to form the man bit the dog, we create an entirely different meaning. As we will recall from our English lessons, words are categorised as nouns, verbs, adjectives, adverbs, and so forth. In English, the first noun is the subject of the sentence and the second noun is usually the object. Between them is usually a verb. Thus the previous exemplar sentences are of the form the (Subject) (Verb) the (Object), and this is known as an S-V-O word order. It has been estimated that 75 per cent of all languages use the S-V-O form (Ratner and Gleason, 1993). One exception is Japanese which uses an S-O-V form (the dog the man bit), and another is Welsh, which uses a V-S-O (bit the dog the man) word order (Ratner and Gleason, 1993). SemanticsSyntax alone in insufficient to abstract meaning from a sentence. The obvious example of this is that it is possible to create sentences that are syntactically correct in terms of the S-VO word order but have no meaning, as in the sentence the jumper milked the sideboard. The study of semantics is concerned with how we process the meaning of linguistic information. There are five aspects of language important for the study of semantics (Bierwisch, 1975): • Ambiguity Words can have more than one meaning and this can make a sentence ambiguous, as in go to the right table – does this instruction refer to the correct table or the table on the right-hand side? • Anomaly This is a sentence with correct syntax but one that has no meaning, as in tables fly long hairs in the air. • Entailment In our use of language we often say something that conveys more information than that contained in the sentence, as in Helen is my best friend, which also reveals, by inference, that my best friend is female. • Conflicting meanings A sentence may contain items that appear to contradict each other, as in my cat is not my pet. Ordinarily, my cat and my pet would be synonymous, except that here the word not creates a conflict and it may mean that the cat is something more than just a pet. • Implication While words can have more than one meaning, so too can sentences, as in Students hate annoying lecturers. This could have two different interpretations: that students dislike those lecturers who can be annoying or that students dislike annoying their lecturers. MorphologyWhile phonemes are the basic physical unit of speech, morphemes are the basic units of word meanings. Morphemes include prefixes (pre-, un-, mis-, dis-, and so on) and suffixes (such as -s, -ness, -d, and so on) as well as individual words. So, for example, the word skirts has two morphemes, skirt (article of clothing) and s (a suffix to denote more than one of these items). Morphemes such as most prefixes and suffixes are not words by themselves and hence are said to be bound. Morphemes that are words by themselves are said to be free. In contrast to theories that stress the importance of syntax in the meaning of a sentence, the case grammar approach emphasises the importance of our knowledge of words and their relationship with other words in attaching meaning to a sentence. So while the syntactic approach informs us of how a sentence is constructed in the terms of the order of different word forms, it tells us nothing of the subject of the sentence.
McGurk showed participants a videotape of a person uttering simple consonant-vowel combinations, such as /ga/. However, the soundtrack was replaced with a voice uttering a different sound, such as /ba/. The soundtrack was synchronised with the lip movements on the videotape. Participants report not hearing /ba/, nor did they report hearing /ga/. Instead they reported hearing /da/, a sound that lies between the two. This phenomenon is know as the McGurk effect. The effect demonstrates that what we hear as speech is not just a simple combination of phonemes and lip movements of the speaker. One implication of the effect for film-makers is that comprehension of movies can be hampered when the voice-over is dubbed (as with films made in one language but dubbed with another language). PragmaticsLanguage is used differently in different contexts. Pragmatics concerns the social rules that determine how language is used in certain contexts. For example, you might describe your college differently to a future employer than you would to a close friend. In the former, you might emphasise the academic aspects of being a student at the college, but in the latter you might emphasise the social aspects of college life. According to Grice (1975) conversations in different contexts vary according to four maxims. Thus a discourse can be: • highly informative versus weakly informative; • completely truthful or untruthful; • most relevant or irrelevant; • perfectly clear or completely unclear.
People vary each aspect depending upon with whom they are conversing. So, for example, someone who is lying to their boss about why they didn't turn up to work might use a discourse that is weakly informative, untruthful, irrelevant and unclear. On the other hand, the same person talking to a friend might use a discourse that is informative, truthful, relevant and clear. Other rules of conversation have been identified that concern the type of utterance that might be used (Searle, 1979): • An assertive is a statement of opinion or belief, as in “I study more than my fellow students do”. • A commissive is a statement that commits the speaker to some action, as in “I am going to study for three hours this evening”. • A declaration is a factual statement, as in I spent two hours watching TV yesterday evening. • A directive is an instruction directed at the listener, as in “Carry these books for me”. • An expressive is a description of the speaker's internal state, as in “I'm so pleased I got an A+ for my cognitive psychology essay”.
In each case it is the listener's job to respond appropriately and some statements require more of a response than do others. Reading and text comprehension In this section you will be reading about the cognitive processes involved in reading and text comprehension, and also when these processes break down, as in developmental and acquired dyslexia. ReadingReading involves many cognitive processes. First, you need to be able to identify the printed characters as letters and the letters as words. Secondly, you will need to hold individual words in memory so that you can understand a complete sentence and relate it to previous sentences. You will also need to be able to comprehend the text and integrate new information conveyed in the sentence you are currently reading with information acquired from previous portions of the text. Hence reading involves object recognition, immediate memory, long-term memory, semantic memory and many other processes. Despite the involvement of so many complex cognitive operations, reading seems effortless and is usually very accurate. Reading differs from spoken language in several ways. First, reading is visual and spatial whereas spoken language is auditory and time-dependent, and while readers can speed up, slow down or pause, listeners cannot do this as listening is dependent on the speaker (although it is possible in some cases to ask someone to repeat themselves). Also, reading involves understanding word units that are separated by white spaces, but speech is continuous and many words are co-articulated. The meaning of the words can be augmented in speech through the use of stresses and accents, but this is not possible with printed words (except with the use of italics to emphasise certain words). Reading involves concerted attention and controlled eye movements and it is usually difficult to do something else while reading. Speech, on the other hand, can be processed relatively easily while the listener is engaging in some other task (such as driving). Eye movements and readingReading involves visual orienting to words across a page. We move our eyes most of the time and the most obvious reason is that there are a number of limitations in the visual system. For instance, we do not have a 360-degree visual field and most visual detail lies in a small region known as the fovea. Visual acuity gradually decays as we move from foveal to peripheral vision.
SACCADES Movements of the eyes are known as saccades. There appear to be two types: Voluntary saccades are slower eye movements under direct conscious control, while express saccades occur automatically and are involved in correcting errors when we try to locate a visual target. In an early study, Saslow (1967) reported that the saccade normally takes between 220 and 250 ms to perform, and that this can depend upon the properties of the stimulus being viewed (such as its brightness). This latency period of between 220 and 250 ms was found in similar studies and was reported in texts as the saccadic latency. However, Fischer and Boch (1983) found that under certain conditions monkeys frequently made very short saccades (80–100 ms). Soon after this Fischer and Ramsperger (1984) reported observing a similar rapid or express saccade but this time in humans, although it was slightly longer (110 to 130 ms). Therefore, we now distinguish between the voluntary, slower saccade and the express saccade, which is a very small involuntary movement of the eyes. While saccades can be rapid they can also be quite inaccurate. It has been observed that a small secondary saccade normally follows a voluntary saccade to correct for fixation error. It is understood that saccadic errors of about 10 per cent are normal (Kowler and Blaser, 1995).
PERCEPTUAL SPAN The region of vision between saccades is known as the perceptual span (Rayner, 1998). During reading the perceptual span has a bias for letters to the right of the current letter being looked at. It has been shown that the span ranges from 4 letters to the left and 15 letters to the right of the current letter. This right-sided bias occurs because reading involves looking ahead to determine important features of the sentence, such as the length of the word and punctuation marks. Words that are highly predictable, such as the word the, are skipped during reading, but unusual words or misspelled words are scanned for longer periods and involve small saccades (Rayner, 1988). In addition, letters and words outside foveal vision are recognised during reading (Norton and Stark, 1971) and it is likely that this occurs because readers use their knowledge of words and their understanding of the text. This is especially so with skilled readers and implies that they do not fixate on each and every letter and word of the text during reading. Saccades made during reading are also dependent upon reading ability. Good readers make larger jumps across the sentence, but poor readers make smaller saccades and often jump backwards towards earlier words in the sentence. Reading and working memorySince working memory is of limited capacity, individuals with a larger memory span tend to make better readers. This is because they have a greater capacity to utilise important cognitive resources during reading. For example, a large memory span aids processing ambiguous sentences (Miyake et al., 1994), the ability to infer the meaning of unusual words (Daneman and Green, 1986) and the ability to understand complex sentences (e.g., Carpenter et al. 1994). Routes involved in word recognitionAn important research question is how readers interpret the meaning of words and what the processes involved are.
The direct-access hypothesis is that readers directly translate the written word into meaning, i.e. perception of a particular pattern of letters directly activates a region in semantic memory (Baluch and Besner, 1991). The phonologically mediated hypothesis is that readers translate words into their speech form (internally, of course) before understanding their meaning (e.g. Wydell et al., 1993). Recall from the discussion of immediate memory, especially the discussion of working memory in Chapter 5, that there is good evidence that working memory operates primarily in an acoustic code. Errors made in working memory with similar sounding words also occur when the material is presented visually (Baddeley, 1966). This suggests that the phonemic quality of words might be important in reading. A third view is a combination of the first two, namely the dual-route hypothesis (Coltheart and Rastle, 1994). EvaluationEvidence that printed words are translated into speech sounds during reading comes from the finding that children who are more knowledgeable about phonemes are better readers (Seidenberg, 1995) and that children trained on phonemic skills become better readers than children trained on semantic skills (Byrne and Fielding-Barnsley, 1991). Evidence that reading might not involve internal phonemic expression comes from Bradshaw and Nettleton (1974) who presented words that were similar in the way they were spelt but were pronounced differently (such as horse and worse). When participants read the word pairs aloud, it took them less time to read the first word than it did to read the second word. The conflict between the visual similarity and the phonemic dissimilarity of the word pairs may have caused interference, and hence is likely to have slowed the reader. However, this was not so when readers were asked to read silently. As there was no interference, this suggests that words were not translated into speech sounds. Yet, McCuthcheon et al. (1991) have shown that adults read tongue twisters (e.g. She sells seashells on the sea shore) very slowly even when reading silently, and this indicates that printed words are translated into sounds.
Participants were presented with word pairs and were asked to state, as quickly as possible, whether the words were related or unrelated in meaning. The key manipulation was whether the second word sounded the same as (was phonemically similar to) a word that was semantically related. For example under one condition, the second word, e.g. BARE, which sounds the same as another word, BEAR, that was semantically related to the first word, LION, produced more reading errors than pairs that were neither semantically nor phonemically related, such as LION–BEAN. Since participants did not read the words aloud, this pattern of errors suggests that they were producing the words phonemically but silently. The direct-access hypothesis suggests that children should be taught to read through a whole-word approach (Crowder and Wagner, 1992), which is that children should not be encouraged to produce the sounds or focus on the sounds of words. There may be some value in this approach since in English, at least, there is not 100 per cent correspondence between the spelling of a word and its correct pronunciation. Conversely, the dual-route approach suggests a phonic approach to teaching children to read. This involves encouraging children to learn to read new words by attempting to produce the sounds of the individual letters (or sounds produced by combinations of letters, such as sh). Research shows that the phonics approach can produce better spellers than can no such training (Pressley et al., 1996). Text comprehensionAs an illustration of the complexity of reading and comprehending text, Medin et al, (2001) tell the story of how, during the 1960s, US computer scientists attempted to devise a computer program to translate Russian into English. To test their program, they translated English sentences into Russian and then back into English. The major problems involved in translating text become apparent by the following effort of this program: ‘The spirit is willing but the flesh is weak' was returned back into English as (after first being ‘translated' into Russian) ‘The vodka is good but the meat is rotten'. (I have just logged onto an online translation service (http://translation2.paralink.com/), and I typed in the same sentence, translating the text first into German and then from German back into English, with equally disastrous results: ‘The spirit is ready, but the meat is weak'). Such literal translations miss one of the beauties of language, which is the metaphor. The point is that the meaning of a sentence is more than just a particular combination of words.
INFERENCES DURING READING Comprehending text is an inferential process: it involves making many inferences and forming coherent representations. For example, consider the sentences The python caught the mouse and Jane caught the ball. The first sentence yields the inference that the python caught the mouse in its mouth, and the second that Jane caught the ball with her hands. So rich are the inferences that when people are asked to recall these sentences some time later, many report the first sentence as The python ate the mouse (Brewer, 1977). Carpenter et al. (1995) reveals that readers also create rich internal models of characters when reading a story. Kintsch (1994) developed a processing cycles theory of text comprehension, which is based on the idea that understanding a sentence involves integrating it into the context developed in the previous sentences. Each new sentence brings with it a new representation of the text, and this representation is constantly updated. There is much research to suggest that, when reading, we make inferences that may not have been intended by the writer. In other words, we draw on information that is not present in the text (van den Broek, 1994). One factor that encourages this is the possibility that semantic memory is schematically organised. We saw in Chapter 5 that when given partial information about something we can recall its whole. So, for example, if you read the text ‘Waiter!' your schematic memory of eating at a restaurant would become active instantly. Similarly, when reading a novel, the reader spontaneously makes inferences about the causes of events, such as the motivations of the characters and the likely outcome of a passage. This represents the constructivist view of reading: that readers actively construct meaning that goes beyond anything stated literally in the text (O'Brien and Myers, 1999). In contrast, the minimalist view is that readers only make inferences when it is necessary to understand adjacent sentences (McKoon and Ratcliff, 1992). One method they developed is to present a paragraph for comprehension and then give a word recognition test. Words that actually appeared in the text should be recognised quicker than words not presented in the text but are likely to have been inferred. For example, given ‘The director and the cameraman were ready to shoot close-ups when suddenly the actress fell from the fourteenth storey', a likely inference is death. McKoon and Ratcliff (1986) found that responses to inferred words were slow but only on an immediate test. Furthermore, they suggested that minimal encoding had taken place, such that participants infer something bad has happened. Minimal encoding may be useful for avoiding inappropriate inferences since further details emerging from the text might contradict an inference made from the earlier text. A considerable amount of research has been carried out in an attempt to resolve this debate, but it has yet to be resolved.
HUITEMA ET AL. (1993) INFERENCES MADE DURING READING Findings from Huitema et al. (1993) and Klin et al. (1999) have shown that readers build up inferences from a text, even when the inferences are at some distance apart, and this further suggests the use of long-term memory in reading. It also indicates that readers look to make inferences that are consistent with those already made from earlier passages of a text. Reading disordersDyslexia is a disorder that affects reading ability but other cognitive processes are normal. There are two main subgroups of dyslexia: developmental dyslexia and acquired dyslexia. The former develops during childhood, but its origin is not well understood, while acquired dyslexia results from brain damage. Developmental dyslexiaIt is not clear whether the disorder has a neurological origin or a psychosocial origin. However, it is known that in dyslexia there is poor development of the magnocellular pathway (Eden et al., 1996), a structure involved in the visual detection of movement, direction and depth. Furthermore, although in the normal brain language functions are specialised in the left hemisphere, in individuals with dyslexia, there is an unusual symmetry of function between both hemispheres (Bishop, 1990). Pollatsek and Rayner (1990) examined the eye movements of individuals with dyslexia and found that they make Text similar to the following was presented to participants: 1. Dick had a week's vacation due 2. and he wanted to go to a place 3. where he could swim and sunbathe. 4. He bought a book on travel. 5. Then he looked at the ads 6. in the travel section of the Sunday newspaper. 7. He went to his local travel agent 8. and booked a ticket to Alaska . Sentences 1 to 3 imply that Dick wants to go to a sunny beach. However, this inference is contradicted in line 8. Notice that the points where these inferences are made are separated by several lines. Huitema et al. presented four different types of stories: a far/inconsistent story (as above), a near/inconsistent story (where the contradiction occurs on the next line), a far/consistent story (where two non-contradictory inferences are separated by several lines), and a near/consistent story (where the two inferences occur in adjacent sentences). Participants tended to read the inconsistent stories slower than the consistent stories (which should be expected when inferences contradict each other), but the distance at which the two inferences appear in the story (whether near or far) did not influence reading speed. These results provide difficulty for the minimalist view, which predicts that reading speed should be influenced by the distance of two inferences. Instead they support the constructivist view. a larger number of smaller eye fixations, with much jumping back to earlier words in the sentence (which are known as regressions), and reading is generally slower than in normal readers. Dyslexia is also characterised by poor awareness of the phonological features of speech. For example, transposing the first sounds of two words (reading mustard-salad as sustard-malad) is difficult. There also appears to be a deficit of the perceptual span for words but not for other types of stimuli. Acquired dyslexiaMany insights can be gained by trying to understand cognitive disorders, most of which result from brain damage. The connectionist approach (discussed in greater detail in Chapter 9) is an extremely useful method for testing cognitive theories since brain damage can be simulated in a connectionist network (by removing a number of processing units or by reducing the values of connection weights). If the ‘lesioned' network behaves in a similar way to individuals with a specific disorder then the network may provide a plausible model of both the underlying process (reading) and the deficit. Acquired dyslexias have received a great a deal of attention from language researchers and connectionist modellers. The two main types of acquired dyslexias are peripheral dyslexia and central dyslexia. There are some similarities between developmental dyslexia and acquired dyslexia in terms of the deficits in reading. Peripheral dyslexiaAlexia is one form of peripheral dyslexia and is characterised by a weak ability to visualise words as single units, such that reading is achieved by attending to words letter by letter. A simple test of alexia should reveal abnormally long word-naming times for long words (Warrington and Shallice, 1980). The disorder may be due to an inability to process letters in parallel (Patterson and Kay, 1982) or the loss of stored visual memory for printed whole words (Warrington and Shallice, 1980). Neglect dyslexia is another form of peripheral dyslexia in which words located on one side of the visual field receive minimal processing (they tend to be ignored). Reading errors are characterised by the omission of words on one side of the page. The disorder may be due to an attentional deficit (Riddoch et al., 1990) or to damage to brain areas responsible for representing information on the neglected side of the visual field. Central dyslexiaIn phonological dyslexia the ability to name words is unaffected, but naming nonwords is severely impaired. In surface dyslexia, reading of words and nonwords is normal but the reading of exception words (words that have an irregular correspondence between the way the word it pronounced and its spelling, e.g. subtle) is impaired (Patterson et al., 1985). The dual-route model of reading can account for central dyslexia. Phonological dyslexia may be the result of an impaired phonological encoding route, while surface dyslexia may be the result of an impaired direct reading route. The existence of these two distinct disorders implies that they result from two independent processes in reading. In deep dyslexia, a variety of reading errors are made, such as those based on the meaning of words (sword read as dagger), those based on visual and semantic similarity (shirt read as skirt), and the poor ability to read nonwords. A possible cause of deep dyslexia is that there may be damage to both routes of the dual-route model. An alternative account is that individuals with deep dyslexia attempt to read with their right hemisphere rather than the left hemisphere which is specialised for reading (Saffran et al., 1980). Connectionist models of reading and dyslexiaIn Chapter 9 we present a description of the connectionist approach to modelling cognition. Several connectionist models have been designed to simulate language functions, especially reading. Furthermore, these models can be tested by making ‘lesions' to the network. The subsequent errors made by the network may (or may not) adequately simulate brain damage (for example, reading errors in acquired dyslexia).tion 3 Language and thoughtIn this section you will be reading about theories of whether language is determined by the way we think or vice versa. Surprisingly, some behaviourists had views on the relationship between language and thought. For example, one theory was that thought could be regarded as inner speech. One researcher allowed himself to be given a curare derivative that caused paralysis and hence prevented any subvocal speech. If thinking was merely a form of inner speech and if speech were prevented, then thinking should stop. Despite almost total paralysis, he reported still being able to think. A number of other studies have revealed that thinking can occur independently of language. For example, participants shown a repeating sequence of cards gradually learned to correctly identify the next card, and they did so long before being able to articulate how they arrived at their decision (Humphrey, 1951). There are many everyday examples we could think of that demonstrate decision-making in the absence of any linguistic activity, internal or otherwise. For example, think of a physical skill that you have (playing tennis, sprinting, playing cricket, juggling, and so on) and try to articulate exactly how you make decisions while you are carrying out that skill. It is quite difficult in most cases. Therefore, it is clear that cognitive processes can occur independently of language. The Whorfian hypothesisWhorf (1956) developed a theory known as the linguistic relativity hypothesis. He argued that the linguistic system is ‘not merely a reproducing instrument for viewing ideas but rather is itself the shaper of ideas, the program and guide for the individual's mental activity†We dissect nature along lines laid down by our native language' (pp. 212–13). In other words, language shapes the way people make sense of and understand their world. The theory has been modified in several ways, and there are strong and weak forms of the theory (Miller and McNeill, 1969). Each version varies as to the extent of the relationship between language and other cognitive processes: • The strong hypothesis: language determines thinking. According to this view, a concept cannot be available to the speakers of a language if no word exists for the concept. On the other hand if a culture places great importance on a particular concept then it will have more words to reference the concept than other languages whose culture placed less importance on it. Supporting evidence most often cited for this is that one Eskimo language has 27 different words to describe snow (Boas, 1911). Snow words can refer to whether the snow is falling or has settled, whether it is lightly falling or is heavy, where the snow has fallen, and so on. Clearly, it would seem that the large number of concepts Eskimos have about snow has had an influence on their language. However, it could just as easily be argued that since snow is an important aspect of their environment, it is the environment that has influenced both their language and their thinking. However, Martin (1986) has since suggested that Eskimos have no more words to describe snow than any other language of people who live in warmer climates.
• The weak hypothesis: language affects perception. Farb (1974) showed one way in which language is able to influence thinking. Japanese women living in America were interviewed in both English and Japanese. When asked to complete the sentence (asked in Japanese): ‘When my wishes conflict with my family's…', they responded with ‘…it is a time of great unhappiness', but when asked the same question in English they replied with ‘…I do what I want'. Hence different attitudes were revealed depending upon which language was used. According to Farb, this is evidence that language expresses cultural ideas and attitudes. EvaluationThe main problem with these studies is that it can be misleading to infer cognitive processes from language (Greene, 1975). As an example of this, while Hopi Indians use the same word for ‘insect', ‘air-pilot' and ‘aeroplane', this does not mean that they categorise these as being the same kind of object. English speakers do not categorise a pause, a decent snooker score and a fracture as part of the same concept when using the word break. The verbal deprivation hypothesisAccording to Bernstein (1961), most sophisticated forms of thinking are dependent upon language. Bernstein distinguished between universalistic meanings and particularistic meanings. • Universalistic meanings refer to knowledge that exists independently of any specific context. Universalistic meanings require an elaborated linguistic code. Perhaps one of the strongest sources of evidence that language can influence perception comes from Miyawaki et al. (1975), who studied infants' perception of phonemes. In order to determine how well infants can discriminate between sounds, Miyawaki et al. used the sucking technique. When an infant is presented with something novel its rate of sucking on a dummy increases. If the stimulus continues then after awhile the sucking rate drops back to baseline. The sucking rate can be directly measured and recorded. Japanese infants were presented with the phoneme /l/,which caused an instant increase in the sucking rate. The phoneme was sounded continuously until the sucking rate returned to baseline. Then the sound was changed to a repeating / r/ and the sucking rate increased, which indicated that the babies were able to detect the change and hence were able to discriminate between the /l/ sound and the /r/ sound. Remarkably, however, Japanese adults are unable to make the same discrimination. Thus, although at birth Japanese people can discriminate between these two sounds, they gradually lose this ability as adults. The reason why this ability is unlearned, of course, is because the Japanese language itself does not discriminate between them. Both sounds can be used in the production of the same word.
• Particularistic meanings refer to immediate, specific ideas or examples, which are often tightly dependent on the context in which they occur. Particularistic meanings require only a restricted linguistic code. Bernstein argued that socialisation and the class system expose individuals to either code: restricted code for the working class and elaborated code for the middle class. Because of the differential exposure to linguistic use, middle-class children begin school with a distinct advantage over working-class children who are verbally deprived. EvaluationThemain critic of this approach is Labov (1970) who argued that working-class children are exposed to a linguistically rich environment, albeit in a distinct dialect rather than a standard form. Labov argued that differences between elaborated and restricted codes is less significant than Bernstein thought. Labov (1972) studied black children's use of English and concluded that the black linguistic code used was of comparable complexity and sophistication to standard English. The developmental viewPiaget's view was that knowledge is the precursor to language and that language is only one of several ways in which knowledge can be represented. As the child experiences the world, he or she uses language to represent this experience both in terms of knowledge and in terms of social interaction. For Piaget then, thought can influence language development much more than language development can influence thought. Vygotsky's view: language and thought are separateVygotsky argued that language and thought are entirely separate and have different origins. Thought, he argued, is a cognitive activity that occurs as the child experiences its environment. The child needs to store memories about the world in order to understand it and to adapt to it. Language on the other hand is a social activity that arises from hearing it from others and from communicating with others. Much of everyday language is affiliative (involved in acting socially and developing social relations) – it commonly exists as ‘small talk'. Language is not always cognitive in the sense that is not always about transmitting knowledge in an elaborate way. The cognitive functions of language, he argued, are served by inner speech, which acts as a way of monitoring and structuring mental activity. Young children often produce egocentric speech, in which the child's speech reflects what is in its own mind without reference to any listener. This is because the child is unable to separate the two main functions of language, which are cognitive and social. Accents and dialectsLyons (1981) distinguished between three variations of a language: • an accent, which is a regional or social variation in the pronunciation of words; • a dialect, which is a variation of a language that has its own grammatical rules and vocabulary; • an idiolect, which is a personal style of language use that includes speech patterns and speech habits. Most individuals have more than one idiolect which they use depending on the social context.
Haugen (1966) argued that there is little qualitative difference between a dialect and a language and it depends on the number of people who speak it. For example, Flemish, which is used by Belgians who have a Dutch origin, is now considered as a language in its own right, but French-speaking Belgians once considered it a dialect of Dutch. Giles (1973) found that the social status of an accent could influence the perceived credibility of an argument. Participants were exposed to arguments for and against capital punishment in different accents: standard pronunciation, a Welsh accent, a Somerset accent and a Birmingham accent. They were then asked to rate the quality of the argument presented to them. Arguments in standard pronunciation were rated the highest, even though the type of argument was counterbalanced between participants. Research shows that this effect is more pervasive than one might imagine.
TEACHERS' PERCEPTION OF ACCENTS Edwards (1979) conducted a study that examined student teachers' perceptions of the reading abilities of children with either working-class or middle-class accents. They listened to tape recordings of children reading the same passage, and were then asked to rate the reading on a number of scales, such as ‘intelligence', ‘enthusiasm', ‘happiness', and so on. The student teachers rated the middle-class children more positively than the working-class children, despite the fact that the only difference between the two samples of children was their accent. This implies that people infer intellectual and other personal attributes on hearing a particular accent. What is of great concern is that teachers should draw such incorrect inferences. Other research shows that teachers' expectations of pupils can have an effect on children's learning: expectations can become self-fulfilling prophecies. If a teacher has an incorrect belief about a child's intellectual capacity then it can adversely influence how that child is treated, which will in turn influence the child's educational performance. Language acquisitionIn this section you will be studying theories and research about how children acquire language, the stages they go through, and whether language acquisition is innate or requires appropriate input from the environment. You will also be studying research on the level at which animals can communicate with humans. Earlier, we discussed four features of language, such as the basic unit of sounds and the uses of language in different contexts. Shaffer (1993) claimed that these four aspects of language were acquired by children in the same order: • phonology: the sounds produced in a language; • semantics: the meanings of words and utterances; • syntax: the rules that govern how words may be combined to make sentences (grammar); pragmatics: the ways in which language use is modified according to the context (e.g. alking to a child versus talking to a college tutor).
What is the evidence for this claim?Evidence of early vocalisationsAt three weeks the infant makes ‘fake cries'. These occur in the absence of any distress and most likely reflect the infant's enjoyment at making sounds. Between three and five weeks the infant begins to coo, and by four to six weeks babbling begins to occur. Babbling sounds like random combinations of vowel and consonant sounds. After a while the infant shows echolalia, which is a repetitive sound, such as mamama. Up to eight months, babbling sounds are universal in the sense that they sound the same regardless of the language of the infant's environment (recall that in the previous section we saw that Japanese infants but not Japanese adults can discriminate between consonants). After eight months, babbling begins to be shaped by the language they hear in their environment. By 18 months, the child has entered the one-word stage, and begins to utter single meaningful words, such as mine. From18 months onwards the infant is in the telegraphic period, where words begin to be combined in way that is reminiscent of the way telegrams were once written (a sort of verbal shorthand), such as Daddy chair. Brown (1973) identified early rules that the infant uses in this stage, such as the basic word order rule: a sentence consists of subject-verb-object (e.g. Daddy eats dinner). Two-word utterances tend to be of the form subject-verb (e.g. Daddy walk) rather than verb-subject (walk Daddy), and reflects the same use of word order by adult English speakers. Grammatical morphemesAs language develops, the child acquires the use of grammatical morphemes, such as prepositions, prefixes and suffixes (in, on, the use of s to denote the plural, the, and so on). All children appear to learn grammatical morphemes in the same order (de Villiers and de Villiers, 1973), starting with simple forms followed by more complex forms (e.g. they are).
OVER-REGULARISATION Children's acquisition of linguistic rules can be evidenced by the mistakes they make, and one of these much studied is over-regularisation. The child notices that the past tense of a verb can be created by adding the suffix ed, which is the case for regular verbs (pass/passed, like/liked, pick/picked, and so on), but over-uses the rule for irregular verbs (e.g. The dog runned away). Irregular verbs do not conform to this rule (e.g., go/went, come/came, see/saw, and so on). Children's over-regularisation appears to follow several stages. PragmaticsChildren also begin to alter their use of language according to the context. For example, Shatz and Gelman (1973) observed that four-year-old children used complex sentences when talking to adults about a toy but simplified their speech when talking to younger children. Theories of language developmentNativist theories of child languageChomsky (1965) presented a theory of language acquisition based on the idea that it is innate and hard-wired into the system. He argued for a language acquisition device (LAD) that enables children to acquire language.
The issue of whether some abilities or characteristics (such as intelligence, language, personality, mental disorders, and so on) are inherited at birth (the naturist view) or acquired as a result of being shaped by the environment we live in (the nurturist view) arises in many areas of psychology. In terms of language acquisition, the naturist view is that we are born with specific brain structures that enable us to acquire language quite easily. The nurturist view is that to acquire language and to reach a high level of competence in its use requires exposure to a rich linguistic environment, especially in the early years. Chomsky distinguished between surface structure, which are the actual phrases used in a sentence, and deep structure, which are the meanings of sentences. Transformational grammar allows us to transform the deep structure of a sentence into the surface structure of the sentence. According to Chomsky transformational grammar is innate. He further argued that there are linguistic universals, which are features found in every language. Thus we are all born with a universal grammar, and this determines the surface structure of sentences. For example, Hocket (1966) has identified 13 general properties that are common to all human languages. Five of these are listed below: • Arbitrariness. Intrinsic connections between the symbols used in all languages and their meaning are not necessary. For example, the symbol or utterance grape has no obvious connection to the object grape. In other words, the symbols in language are abstract. • Cultural transmission. Languages are learned by direct contact with their cultural origin. • Discreteness. Languages each have a discrete set of phonemes. • Duality of structure. All languages have a finite set of phonemes than can be arranged to form meaningful units in a virtually limitless number of ways. • Productivity. In all languages it is possible to create a completely new sentence, which is unlikely ever to have been used before. An example I'm just inventing right now is ‘Eamon, the person writing this book on cognitive psychology, is the most intelligent, good-looking, easy-going, and most generous man you could ever meet.' Now, I can be pretty sure that this sentence has never been uttered by anyone (except, perhaps my mother)! Critical period hypothesisLenneberg (1967) argued that although language is innate, it does depend on a biological maturation process. As evidence of this Lenneberg (1967) reported that children who become aphasic before puberty, especially before the age of five, recover most or all of their lost language functions, whereas such recovery after puberty is slow and partial. Findings like these suggest that early opportunities to acquire language are critical for the development of competency in the language. EvaluationStudies of second language learning, such as that of Newport (1994) on Asian immigrants to the US , show that the younger the individual when entering the country the better they learned the language, which supports the critical period hypothesis. Other evidence has been taken from studies of deprived children, such as the case of the ‘Wild Boy of Aveyron' and the case of ‘Genie'. These studies show that if a child is not exposed to language, up to a certain age, then they never subsequently acquire it. However, other evidence is less clear. For example, Harley (1995) found that recovery rates from aphasia were not predicted by age. In terms of acquiring a second language, more recent studies suggest that when the same learning criteria are used for adults and children, adults are better on language learning tasks than children are. As McLaughlin (1984) points out, these findings pose problems for the critical period hypothesis. Pinker (1994) refers to a study of Nicaraguan children of various ages who attended a new school for deaf children created by the government in 1979. Although they were not taught sign language directly, they developed their own system quite spontaneously. The sign language they created later became widely used (known as the Lenguaje de Signos Nicaragüense or LSN). Interestingly, younger children who were then taught LSN developed it further and in their own way. The sign language developed by these younger children (which became known as the Idioma de Signos Nicaragüense) included grammar. It has been argued that the inclusion of grammar occurred because these children were younger than those who created LSN. These spontaneous developments in the use of sign language support the critical period hypothesis. Environmental theoriesSkinner (1957) argued that language was acquired through operant conditioning. Parents reinforce the required utterances and so these become more likely to be repeated. As the linguistic abilities of the child become progressively more sophisticated so reinforcement become more selective. This process of reinforcement is known as shaping. He contended that children also learn through imitation: the child often tries to repeat what the parent says (called the echoic response). EvaluationAlthough children often learn language through imitation, the evidence for learning by reinforcement is generally not supportive. Brown et al. (1969) found that parents tend to reward children's speech on the basis of its truth rather than on the basis of the grammar used. The way parents influence their child's language development is through the way they speak to the children and through expansion, rather than through direct praise. Many parents, for example, use ‘motherese' when talking to very young children. Motherese consists of sentences that are very short and simple, which gradually become longer and more complex as the child's own use of language develops (Shatz and Gelman, 1973). In addition, parents tend to expand the sentences uttered by their children. So, for example, the child might say biscuit please and the parent might respond with Would you like a biscuit? This expansion instructs the child on how grammatical sentences are constructed. The interactionist approachSocial interactionists suggest that language acquisition depends on social stimulation. For example, Sachs et al. (1981) studied a young boy whose parents were both deaf and did not speak. By the time he was four his speech was below age level. Although subsequent speech therapy led to quick improvements, his impoverished speech earlier was due to exposure to an impoverished linguistic environment. Bruner (1983) proposed that instead of an LAD, children have an LASS (language acquisition support system), whose maturation depends on social and verbal interaction, especially with adults. Animal languageWhile animals are able to communicate with each other through various signals, animals are not generally considered to use language. This may be because language is an innate ability that animals do not have. However, if an animal could be taught to use a form of language and communicate in moderately sophisticated ways this might provide evidence that would sit uncomfortably with the nativist position. The problem with communicating with animals is that they are unable to vocalise the phonemes used in human languages because they do not have the necessary vocal systems. However, and for that reason, psychologists have attempted to teach animals, and especially apes, forms of sign language. Gardner and Gardner (1969) taught an ape Washoe 132 signs of the American Sign Language (ASL) over a four-year period. She could use signs that referred to objects and the state of objects (such as open). In addition, despite the fact that the researchers did not use ASL in the presence of a second ape, Loulis, the latter ape learned to use 50 signs by being in the presence of Washoe, her teacher.
Another remarkable study is that of Savage-Rumbaugh. (1986). Kanzi, a bonobo chimp, learned to communicate with humans through a keyboard containing numerous symbols. However, Kanzi acquired the ability by watching his mother, Matala, being trained to use the symbols. Kanzi could use the keyboard to request particular foods, and so on. For example, Kanzi could press the symbol for strawberry before going to the spot whether the strawberries were grown. In addition, Kanzi would go to the strawberry patch when the word strawberry was mentioned. He could also produce lexical combinations such as more drink and apple me eat. Savage-Raumbaugh et al. (1993) claimed that Kanzi's linguistic ability was equivalent to an 18 to 24 month-old human. This claim was based on comparing Kanzi and a daughter of one of the researchers, Alia (who took part in similar games and activities as Kanzi), on carefully controlled tests. Some critics have argued that these animals were merely conditioned to use signs by being given rewards and that the animals did not acquire an understanding of the meaning of the symbols in the same ways that humans understand them. However, in many of these examples, the animals used combinations of symbols that they were not directly taught. For example, one ape referred to a zebra as a white tiger, which suggests more understanding than the critics have argued. Typical Exam Questions1. What are the main elements of spoken language? 2. What evidence is there that readers make use of the pronunciation of words during reading? 3. ‘Thought is itself shaped by our use of language.' Discuss this view with respect to the relationship between language and thought. 4. How has the study of animal language helped resolve the nature–nurture debate in language acquisition? Further readingHarley, T. (2001) The Psychology of Language: From Data to Theory. Hove: Psychology Press. Clark , H. H. (1996) Using Language. Cambridge : Cambridge University Press. Ellis, R. and Humphreys, G. (1999) Connectionist Psychology: A Text with Readings . Hove: Psychology Press. |
||
This book was first published in 2003 by Crucial, a division of Learning Matters Ltd [ISBN 1 903337 13 5] © 2003 Eamon Fulcher; © 2009 GEFT Consultance Services (geft.co.uk). All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior permission in writing from Geft Consultancy Services, who may be contacted via www.geft.co.uk.
|
|||