There are various kinds of acquired dyslexia. In deep dyslexia, it is common for patients, when asked to read a word out loud, to make errors such as `kindness' for SYMPATHY or `orchestra' for SYMPHONY. They see one word, but produce another that is related in meaning. Sometimes, they make errors based on visual similarity instead-they might say `symphony' when shown SYMPATHY. Things get really complex when they make both kinds of error simultaneously-'orchestra' for SYMPATHY. It looks as if something has gone very wrong with the associations that link the visual form of each word with its meaning.
Patients suffering from this deficit tend to exhibit a range of symptoms, including an inability to read aloud nonsense words, and for there to be fewer problems with words that are easy to visualize (e.g. CLOCK) than with words that are harder to visualize (e.g. CLICK). Perhaps as a consequence of this, deep dyslexics tend to do better with nouns than with verbs, with adjectives coming in between, and grammatical function words and inflections being particularly problematic. But could a single underlying deficit lead to such a wide range of symptoms? Currently, there is no single agreed theory as to the causes of the symptomatology of deep dyslexia. Generally, these patients have very extensive damage to the left hemisphere. One idea is that the kinds of stroke that lead to deep dyslexia disrupt a blood supply that branches off into different areas of the left hemisphere. When cells in those areas die, the associations to and from those areas (and the information they encode) may also be disrupted. There is not just one underlying deficit-rather there are several, each caused, originally, by the same physical damage. Interestingly, networks of artificial neurons (see Chapter 13) can be built which simulate some of the different kinds of association implicated in normal reading. When `damaged', they exhibit deficits that can be similar to those found in deep dyslexia.
A subset of the symptoms found in deep dyslexia are often found in another kind of dyslexia phonological dyslexia. Here, the defining characteristic is that patients are unable to read aloud nonsense words (e.g. GLIM), although they have problems understanding certain real words as well. As we shall see, there is reason to care about a deficit in which patients cannot read words which do not exist. One widely accepted explanation of phonological dyslexia is that these patients have somehow lost information about the letter-to-sound correspondences that would allow them to read aloud nonsense words. it is called a phonological dyslexia because the deficit has affected their knowledge of how letters translate into the sounds, or phonology, of their language. In the previous chapter we saw that skilled adults tend to access the meaning of a word directly from its visual form, and not by sounding out the word according to letter-to-sound correspondences and subsequently retrieving its meaning on the basis of sound-to-meaning correspondences. However, adults do resort to such correspondences for infrequently occurring words. And these are exactly the real words that phonological dyslexics have problems understanding. We shall return shortly to why such dyslexics have lost those letter-to-sound correspondences. But first, it is useful to contrast these dyslexics with yet another kind-surface dyslexics.
The defining characteristic of patients with surface dyslexia is that they are very bad at reading irregular words (such as YACHT, PINT, and REIGN). But they are good at reading nonsense words. The explanation is that these dyslexics rely heavily on their knowledge of letter-to-sound correspondences, and have lost the direct association between the visual form of a word and its meaning. This is the converse of the phonological dyslexics, who seem to have lost their knowledge of the letterto-sound correspondences and so rely more on that direct route to meaning.
How do these deficits come about? The simplest explanation is that each route involves a different set of associations. One route is between letters and the phonemes they represent, then between those phoneme sequences and the meanings associated with them. The other is between entire letter sequences and meaning (embodying the direct association between the visual form of a word and its meaning) Each set of associations is likely to involve different areas of the brain-the associations with meaning will involve the different areas associated with the different aspects of meaning, and the letter-to-sound associations will involve the areas associated with hearing and producing sounds. Localized damage to one area of the brain could in principle destroy associative connections leading to one (other) part of the brain or another, giving rise to one kind of dyslexia or another.
Patients rarely exhibit a pure dyslexia. It is more common for them either to retain some abilities more than others (a partial deficit), or to exhibit a blend of different dyslexias and/or aphasias. Some phonological dyslexics seem to have a more general phonological impairment-they are bad at separating spoken words into their component sounds, or knowing that /k/ and /at/, for instance, make `cat'. In these cases, the deficit is not specific to letter-to-sound correspondences.
Often, dyslexia is accompanied by a deficit in spelling. but this does not always happen, and sometimes, a deficit in spelling can be found in the absence of a deficit in reading. The acquired spelling disorders, or dysgraphias, can be quite similar to the acquired dyslexias. Deep dysgraphics make semantic errors when writing words, although once they have written the word, they may notice their own mistake. Phonological dysgraphics cannot spell nonwords, and surface dysgraphics cannot spell many irregular words (they regularize their spellings). One patient had both a dysgraphia and a naming deficit-when shown, for example, a picture of a lion he said `leopard' but wrote down TIGER. The words he produced were related in meaning, more or less. Visual similarity was not the problem, as he would make the same kinds of mistakes with, for example, very different looking vegetables. Some dysgraphias, though, have nothing to do with associations between words and their spellings, and have instead to do with the physical side of writing, and lead to problems in choosing the appropriate letters (letters may be omitted or added), or physically writing the letters (individual strokes may be omitted or added).
Dyslexia as a childhood disorder
If different disorders have the same name, are they in fact the same? Confusingly, the answer is `no'. There have been various attempts to draw comparisons between the acquired and the developmental dyslexias, but these have proved extremely controversial. We shall come to some of the comparisons in a while. But first, what is this other kind of dyslexia?
Developmental dyslexia is a disorder of development and, specifically, the development of certain skills required for reading. As such, it is perhaps best viewed in terms of how the acquisition of these skills has broken down, and against a background of what we know about the normal acquisition of such skills. In fact, children are classified as dyslexic through comparison with other children whose development has proceeded more normally: if a child is at least 18 months to two years older than children with comparable reading skills, and there are no other obvious factors involved (the child is of normal intelligence, normal educational background, has normal eyesight, normal speaking skills, and so on), he or she is classified as dyslexic.
If dyslexia is diagnosed relative to the reading skills of other children, perhaps the only `deficit' is that dyslexics have progressed more slowly than expected, and simply need more time. But it is not just a matter of time. If it were, the only worry would be the knock-on effect of poor reading ability inside and outside the classroom. But such knock-on effects aside, what is the prognosis for dyslexics if no remedial help is given? Come to that, is remedial help successful? And what kind of remedial help is required?
Before answering any of these questions, we need first to know what the underlying deficit is. Although this has proved a controversial topic, the general consensus is that dyslexic children fail to attain some of the basic phonological skills that underlie reading. Phonology refers, simply, to the sounds of the language, and the way in which these sounds are put together to form meaningful units. Typically, dyslexic children have problems playing word games which involve stripping the phonemes off spoken words to cre
ate new words (e.g. stripping the /s/ from `speech'), blending sequences of phonemes to create new words (e.g. adding the /s/ onto `peach'), or spotting which words rhyme and which do not (e.g. spotting the odd one out in the spoken versions of `speech, grouch, bleach'). They also find letter-to-sound correspondences hard to master and have to rely heavily on developing a sight vocabulary of words they can recognize directly, without needing any such correspondences. This means they find new words extremely difficult, and consequently that reading alone, without help with those new words (or words that are still not familiar enough), is an almost impossible task. And whereas non-dyslexic children who have acquired a good knowledge of letterto-sound correspondences will often regularize their spellings, and spell words phonetically (e.g. YOT instead of YACHT), dyslexic children's spelling errors will often be phonetically unrelated to what they are aiming for, and may include completely inappropriate letter-to-sound mappings.
Does the deficit simply impair the letter-to-sound mappings that are required in an alphabetic language like English? One reason for the controversy regarding the nature of the deficit is that some dyslexic children (probably less than one in five) are unable to gain a sight vocabulary but instead use letter-to-sound mappings to spell out each word laboriously. These children seem to have the opposite problem from most dyslexic children, who have poor letter-to-sound mappings and rely on a sight vocabulary. So there is not, after all, a single deficit common to all children classified as dyslexic. There are superficial similarities between the symptoms of these two kinds of developmental dyslexia and the acquired dyslexias (specifically, surface dyslexia-poor sight vocabulary-and phonological dyslexia-poor letter-to-sound correspondences). But it is very unclear whether these similarities reflect the same underlying problems in acquired and developmental dyslexias, or whether they just look similar because some children rely more on a sight vocabulary, and others more on letter-to-sound correspondences. In any case, by the time these children are assessed, many will have developed different strategies for coping with their problem, so complicating their symptomatology.
The majority of dyslexics do none the less exhibit a fairly stable pattern of symptoms. This is borne out also by studies which have determined which factors can predict whether a particular child will have dyslexia. One of these predictors, assessed when the child is still pre-literate and has not yet started to read, is the child's awareness and sensitivity to the internal sound structure of spoken words (assessed using the word games mentioned earlier). Clearly, sensitivity to this structure is important when learning an alphabetic script. Another predictor is the ability to name the letters of the alphabet when shown them. When assessed in five year-olds, this is possibly the best predictor of which children will be diagnosed as dyslexic by the age of eight. Five is a crucial age in many of these studies, because five year-olds are only beginning to read, and so any evidence of dyslexia can only set in later. Why letter-naming should be such a good predictor is something we shall return to. But letter-naming and phonological skill are also good predictors of reading ability in non-dyslexic children. So some researchers have questioned whether dyslexics are just at one end of a continuum of reading ability. We know that individual readers vary considerably in their reading skill. Some children (and adults) place more reliance on letter-to-sound correspondences, whilst others place more reliance on whole-word reading, mapping the visual forin directly onto meaning (see Chapter 11 for further details). So are the children who have difficulty acquiring the letter-to-sound mappings doing anything that is qualitatively different from what other children do? To the extent that they rely more on a sight vocabulary, yes. But to the extent that having a sight vocabulary is normal, no.
Whether dyslexics are qualitatively different or not does not really matter. One way or another, they have a reading problem (and often, a writing problem too). Which means that, one way or another, something has to be done to help. What has to be done depends in part on when it is done. If problems are identified at around the age of five or six, preventative measures, such as training the child to become more phonologically aware by playing word games can be effective. But even later on, when the child is around eight or nine years-old, intervention can certainly improve matters. Children's ability to generalize to new, regularly spelled words can be improved by focusing on their phonological skills using those word games, and on their letter-to-sound correspondences. Their ability to develop a sight vocabulary which can generalize to new, similar, words can be improved by highlighting similarities between different words, and focusing on subparts of words-contrasting, for example, BEACH and TEACH. So intervention does improve matters (just as it does with non-dyslexic children, in fact). But that is not to say that it will be equally effective for all dyslexics.
The effectiveness of an intervention depends in part on the severity of the impairment. In one study, by Richard Olson and colleagues in Boulder, Colorado, children were given extra reading instruction by computer. Each time the children encountered a word that they could not recognize, the computer would either break the word up into its component syllables (e.g. CHIL-DREN) and speak each syllable individually using a speech synthesizer, or it would break the word up into smaller chunks than syllables, corresponding to everything up until the vowel (the onset-the `sP' in `SPEECH') and everything from the vowel onwards (called the rime-the `EECH' in `SPEECH'). For CHILDREN, this would result in the spoken sequence corresponding to ca 1-IL-DR-EN. The reason for these two kinds of segmentation was that a number of studies, primarily by Rebecca Treiman at the Wayne State University in Detroit, have shown that children do not naturally split the words they hear into their component phonemes, but find it easier to split them into syllables, and to split syllables into onsets and rimes. It turned out that children who were very impaired responded better to the syllable-based segmentation, whereas children who were only mildly impaired responded better to the onset/rime segmentation. Not surprisingly, the more aware a child is of the fine internal structure of a spoken word, the more he or she will respond to a remedial programme that assumes some sensitivity to that structure. Broadly speaking, children will benefit most from programmes that focus on the finest detail of internal structure that they are sensitive to. It is exactly this kind of sensitivity that early intervention can improve.
A possible cause of childhood dyslexia
The fact that dyslexic children have poor phonological skills, but benefit when given phonologically related remediation, supports the view that their deficit is phonological. Presumably, the reason such children fail to develop good letter-to-sound mappings is that their neural encoding of the sound patterns of their language has not developed in quite the same way as it does in non-dyslexic children. And this means that associations to the corresponding neural circuitry (from the letter circuitry, for instance) cannot themselves develop in the same way.
It turns out, though, that there is another factor that contributes to dyslexia, which is related to that earlier observation that one of the best predictors of dyslexia (and of reading skill generally) is the ability of the child to name letters-children who are better at this will make better readers and are less likely to be dyslexic. Of course, better letter-naming reflects a greater awareness of the alphabet, which in turn means a greater awareness of what an alphabet is for. But there is another possibility. Alphabet shapes are just arbitrary symbols. The names given to those symbols are, to all intents and purposes, equally arbitrary. Being able to name the letters means having been able to associate an arbitrary symbol with an arbitrary sound. And that is exactly what letter-to-sound correspondences are about.
A variety of studies have explored the ability of dyslexic children to associate symbols with sounds. Typically, dyslexic children perform more poorly than non-dyslexic ones when they have to learn nonsense names for arbitrary shapes. But this effect is specific to verbal stimuli. If you were to ask children to learn which figures go together, for instance, and taught them t
o associate one figure with another, then you would find that dyslexics would perform no worse than other children. So associations themselves are not the problem, it is associations to sounds that are the problem. But if this is the case, why do these children not have problems learning the names of the everyday things around them? After all, these objects are also just arbitrary shapes, even if they are not drawn out on paper. Evidently dyslexic children do somehow manage to learn those names, and are not so delayed that they stand out as having a general naming problem. On the other hand, they tend to be slower than non-dyslexics when they have to rapidly name objects. This difference can be found even in adult dyslexics. None the less, dyslexics are able to name objects, and do somehow overcome the problems they have associating them with spoken sounds.
We experience the objects around us in many different ways, ensuring that the meaning of each object is encoded over numerous neural circuits that each encode different aspects of that meaning (this was discussed earlier in relation to the naming deficits). So there are multiple associations between these different aspects of meaning and the same single name. But the correspondence, for instance, between a letter and a sound does not involve such multiple associations-a letter has no other meaning than its name or its sound. When learning the names of everyday objects perhaps the association problem is overcome by the sheer variety of associations between the object (or rather, the experience of that object) and its name. Support for this idea comes from the finding that children's ability to name letters improves if they also have to trace the letters, which presumably adds another level of experience.
The Ascent of Babel: An Exploration of Language, Mind, and Understanding Page 24