Philippe has kindly provided me with the English version of Tom's speech - as relevant now as it was back in 2001:
Speech by Tom Burkard at Enseignement et Liberté’s conference on January 16 2001, in Paris, at the Palais du Luxembourg
Good evening, ladies and gentlemen. I would like to thank M Gorre and M Pecheul for inviting me to speak tonight, and I would like to compliment Mme Wettstein-Badour for her excellent book, Lettre aux parents des futurs illettres. I regret that it is necessary for me to speak in English; although I studied French for three years in school, Anglophone schools are even worse at teaching foreign languages than at teaching children to read their own language.
As the founder and secretary of The Promethean Trust, my principal interest is finding practical ways of teaching dyslexic children to read and spell. Vast amounts have been written on the subject, and my job is to find out what works, and what doesn’t. Even in England, where it is officially recognised that children must be taught phonics, there is much dispute as to how this is best taught. Because we work with parents and teachers, we are aware of both the theoretical problems and the practical problems. Over the years, we have tried many ideas, and we have retained only the most efficient methods.
There are two problems I will address this evening. First, I would like to discuss the problem of automaticity. Many studies have identified slow letter-naming ability as being one of the strongest predictors of poor reading performance, and there is no doubt at all that poor readers cannot translate letters to sounds quickly and efficiently. In our clinical work we are pioneering a simple method which answers this problem very effectively and quickly. Secondly, I would like to comment on the problem of reading comprehension. Advocates of whole-language teaching routinely assert that teaching children the phonetic code of their language distracts them from the essential task of getting meaning from print.
Children with more severe reading difficulties almost invariably are slow at naming tasks. Their neurological apparatus does not admit for easy and automatic processing of visual stimuli to the appropriate spoken response. This applies equally to naming objects, naming letters, or translating graphemes to phonemes. A 1998 study at Newcastle University also demonstrates that ‘dyslexic’ children are poor at responding to dynamic stimuli. Not only do they find it difficult to respond to transitions in phonemes—as the work of Barbara Tallal at Rutgers in the US has demonstrated—but they find it difficult to scan print quickly and translate letter patterns to sound patterns. They will have great difficulty in synthesising sounds into words because they identify them so slowly that they forget the first one before they have got to the last. Teachers are so frustrated by these poor abilities that they normally encourage such children to use whole-word methods. Thus, the children most in need of phonological training are least likely to get it.
In our clinical work at The Promethean Trust, we are developing simple techniques to overcome this problem. We recognise that the problem is constitutional, but we also know that training in specific tasks will greatly improve performance. The children who come to us for assistance vary considerably in age, ability, and the level of phonological skills that they have already mastered. The following sequence of training can be entered at any point, but for the sake of clarity I assume here that the child comes to us with no skills at all, as many of them do.
In general, we use a modelling procedure to teach all skills. For instance, we use flash cards [a small square of cardboard with the letter or letters printed on it—I will bring along a few to demonstrate] to teach children to produce sounds for individual letter and letter digraphs. The card is presented, and the teacher articulates the appropriate phoneme, and the child repeats it. If the child has difficulty producing the phoneme accurately, we demonstrate how the tongue, lips and voice are used to produce it.
Depending upon the child’s level of difficulty, we may teach anything from 2 to 6 new sounds each week. For a period of up to 5 minutes, we will repeat the flash cards. If the child does not produce the correct sound in 2 seconds, the teacher repeats the modelling process. When the child succeeds without prompting, he is given the card and it is not repeated that day. As with all our training procedures, the parent is instructed to use the same techniques each day at home. Letters and digraphs learned in the previous weeks are practised until the response in instant and automatic. The child is also trained to write the correct letter in response to the sound. Because the process takes very little time, and because the child is never left to struggle, the child develops confidence in the method.
At the same time, the child is taught to blend sounds into words. Many children can only learn to do this in small stages, and it may be necessary to make each learning task as simple as possible. Often we must start with oral exercises, by simply saying the sounds of short words to the child as seamlessly as possible—articulating the consonants cleanly without the ‘uh’ sound—and then asking them what the word was. In this way, they begin to grasp the phonemic principle.
As soon a the child can do this, we have them begin to read from lists of single-syllable words that have regular spellings. The first list consists of only three-letter CVC (consonant-vowel-consonant) words, and the subsequent lists introduce digraphs and consonant blends. Because there is no context, children have no means of ‘predicting’ the word. By this time the child can easily recognise the common English digraphs which only have one pronunciation, such as ‘ch’, ‘or’ and ‘ar’. Often the child will still be unable to blend the sounds successfully, because the task is still too difficult. In this case, we point to the letters as we say the sounds, and of course they succeed. Gradually, we allow them to read the letters themselves.
Once they are able to do this reliably, we then set targets to improve their speed. The youngest children have lists with six words to a line, and they progress to sheets which have 11 words to a line. They read one line at a time, chosen at random so the pupil cannot memorise the words. We time them with a stop watch, and record the times. At first, the child may have a target of as long as 60 seconds to read each line, and it may take several weeks before all of the lines on the sheet have been crossed off. Next, they will have to read the same words in 45 seconds, then 30 seconds, 20 seconds and finally 12 seconds. The lists of words are long enough so that they cannot merely remember the words as ‘sight words’, and many of the words will not be in their speaking vocabulary. In this way, the child’s ability to process letters to sounds automatically is assured. As they progress through the programme, their rate of learning accelerates, and their confidence is improved enormously.
We are still developing this technique. In the more advanced stages, pupils learn to recognise more complex and less common spelling patterns. It is important that these exercises be performed each day for five or ten minutes. In a clinical setting, it is absolutely necessary to teach parents to perform the same exercises with their children at home, as we only see them once a week. We have found that this technique produces remarkable results, and it can be used successfully even when children attend schools where they are encouraged to identify whole words through guesswork.
At the same time that we are doing this, we teach children spelling using the SRA Spelling Mastery programme developed by Dixon and Englemann in the US. This is a highly structured programme of direct instruction which progressively teaches children strategies which minimise the amount which must be learnt by rote. Indeed, once our pupils have learnt to decode shorter words efficiently, it is spelling that teaches them about the morphological structure of longer words. We show them how the word exceptionally breaks down into the morphemes ex-cept-ion-al-ly—this knowledge makes it far easier to both read and spell longer words. I think we may safely presume that this strategy is frequently used in German schools.
Now let us assume our pupil has learnt how to the identify words on the printed page accurately and fluently. This is no small achievement for children who have the neurological dysfunctions so ably described by Mme Wettstein-Badour, but it is perfectly achievable for all but the most severely handicapped children. Now how do we teach children to understand what they read? There is certainly very little point in teaching children to decode if they cannot understand what they read.
In 1917, Edward Thorndyke published research which demonstrated that children were unable to extract factual information from a short paragraph. Now this paragraph was written in a style so grammatically complex that it could only have been written by an British civil servant, and subsequently it was found that most adults couldn’t understand it either. Nonetheless, this sparked off a massive movement to teach reading comprehension, a veritable quest for the holy grail or the philosopher’s stone. If only we could teach children to understand what they read, then teachers would have a very easy job indeed. It would help educators realise the progressive ideal of the child as an independent learner, pursuing his own goals in his own way.
Literally thousands of studies and projects have been initiated to teach decontextualised reading comprehension skills, yet none one of them has demonstrated significant gains in reading comprehension as measured by standardised tests. It never seems to have occurred to these researchers to ask whether there is any significant difference between reading comprehension and listening comprehension. In 1972, Sticht conducted a study of US Army recruits and found very little difference in the understanding of what they read on one hand, or what they were told on the other. These findings were confirmed by studies three subsequent studies. The only exceptions are individuals with poor decoding ability, whose mental attention to the text may be distracted by the effort of identifying words.
So it follows that teaching reading comprehension is no different from teaching comprehension per se. And if this is so, why bother with printed text when the pupil is in the classroom with you? Speaking is a more efficient means of communication—unless we are in a very bad mood indeed, do we write notes to people in the same room?
Some researchers are now beginning to realise that reading comprehension is almost entirely a function of the reader’s prior knowledge of the subject. Any text is written on the assumption that the reader already has a vast store of information. Even the simplest children’s books must make this assumption. In writing this talk, it was necessary for me to assume that you already understand quite a lot about how children are taught to read—otherwise, my task would have been impossible. This theory originated with Gestalt psychology and the notion that we all possess various schemata of knowledge which enable us to make sense of new information.
In 1996, Perfetti concluded that “higher level [comprehension] skills…develop ordinarily in tandem with the gradual accumulation of knowledge.” In other words, reading comprehension—for children who can decode efficiently—is inseparable from a good education. And a good education must be properly structured, with basic knowledge and skills being learnt thoroughly before more complex material is introduced.
An important component of that education will be the acquisition of new vocabulary. Written text will normally have a more complex vocabulary than ordinary conversation, and that after the age of 9 or 10 most new words are learned through reading. Children who cannot read words without conscious effort are unlikely to read any more often than is strictly required. Keith Stanovich likened this to the ‘Matthew effect’—the child who can and does read will become richer in his store of information and vocabulary, whereas the child who cannot becomes progressively impoverished.
I would like to conclude with a few comments about the new consensus represented by the National Literacy Strategy in Britain, and by the work of Reid Lyon, the influential researcher at the National Institute of Child Health and Development in the United States. This new orthodoxy is that both phonics and ‘whole language’ are necessary to teach children to become fully literate. As a practical matter, we must realise that millions of teachers have been trained in whole language mythology, and it is no easy matter to induce them to abandon their beliefs.
But as a simple matter of fact, this new orthodoxy is only a slight improvement on the old. Children are taught phonics through ‘phonemic awareness’ training, using rhymes and alliteration to increase their awareness of the phonemic structure of words. Alas, this is unnecessary if children are taught to blend phonemes in the manner I have described. Our method is still universal in German schools, and Heinz Wimmer of Salzburg has demonstrated that phonemic awareness is a natural consequence of learning to read in this manner.
By the same token, whole language activities—teaching children to predict what comes next in a story, teaching them about different ‘genre’, and teaching them ‘comprehension strategies’—are all an utter waste of children’s time. The practical consequence of cluttering the curriculum with all these extraneous activities is to distract the teacher from the simple goal of teaching all children efficient decoding skills, and teaching children enough knowledge so that they can in fact become independent learners.
