The following is the text of a talk given at the ‘Battle of Ideas’ conference organised by the Institute of Ideas on 30 October 2005. The organisers had chosen the title for this session and had asked the Queen’s English Society to choose two speakers likely to represent different viewpoints. The other speaker was Prof. Kathy Hall, of the Open University, who is president-elect of the United Kingdom Literacy Association.
I want to use the handout to show that there are two dimensions to reading: one is decoding and the other is comprehending.
Which of Texts 1, 2 and 3 can you read?
Text 1 (from ‘Jabberwocky’, by Lewis Carroll):
’Twas brillig, and the slithy toves
Did gyre and gimble in the wabe:
All mimsy were the borogroves,
And the mome raths outgrabe.
Text 2 (Part of a letter written by a 50-year-old woman with severe learning difficulties):
I have been on boat trip harbour ocean. View for meal there Turkish apple tea for a nice meal out at the Café Restaurant….. I picked fruit by pushes railway track for Houses every Autumn Seasonal outside. View at the weekends area. I hope you have had a nice birthday greetings treat Happy days every year.
Text 3: (To be identified in due course)
ın δə laıt əv δə mu:n ə lıtl εg lεı ɒn ə li:f. wʌn sʌndεı mɔ:nıŋ δə wɔ:m sʌn kεım ʌp ænd pɒp! aut əv δi εg kεım ə taını ænd vεrı hʌŋgrı kætəpılə.
Phonics is concerned with the decoding dimension of reading: converting the black marks on the page into words spoken aloud or silently. The approach which teaches beginners to do this purely by converting letters into sounds and blending the sounds together has come to be called ‘synthetic phonics’: ‘synthetic’ because of the ‘synthesising’ element – once the separate sounds /c/-/a/-/t/ have been produced they need to be blended or synthesised into a seamless whole, ‘cat’. This was the approach used in an experiment in Clackmannanshire in Scotland which produced very good results and had a lot of media coverage. It was apparently The Scotsman newspaper which first called phonics the ‘holy grail of reading’ in reporting on this project. I’ll say more about this later.
Everyone agrees that comprehension is the ultimate goal of reading, but the battle of ideas continues about whether a strong early emphasis on phonic decoding makes children better or worse at comprehension in the long term. What is clear is that although decoding does not guarantee comprehension, comprehension is impossible without decoding. This view is supported by research and common sense, and you were adopting it if you felt that you could ‘read’ texts 1 and 2 but not text 3 – this was because you could decode texts 1 and 2 (translate the black marks into words spoken aloud or silently) but could not decode text 3 unless you knew the International Phonetic Alphabet. It’s the opening of a popular book for preschoolers: The Very Hungry Caterpillar by Eric Carle. The words are ‘In the light of the moon, a little egg lay on a leaf. One Sunday morning the warm sun came up and pop! Out of the egg came a tiny and very hungry caterpillar’. In terms of comprehension, Text 3 is much more straightforward than Texts 1 and 2, but one can’t know this without decoding it.
Most of us regard ourselves as ‘reading’ if we can decode, but not if we can’t. Of course we want to understand too if at all possible, but we accept that there are times when this may require extra effort. For now, I want to make the point that when children start school, decoding is the thing that most of them can’t do. They are in the position that some of you were in with text 3: they can understand it if it’s read aloud to them, but they don’t yet have the decoding key which would allow them to do the whole job themselves. The top priority of synthetic-phonics teachers is to bring the children’s decoding up to the level of their comprehension, and so they teach phonics first and fast to beginners. This does not mean that they neglect comprehension – it just means that they regard the children’s existing comprehension level as sufficient for any reading they do at this stage. They begin by teaching the children to recognise letter-shapes and to produce sounds for them (starting with the simplest letter-sound correspondences), and teach them to blend those sounds together into recognisable words. A common procedure is to teach a few letter-sound correspondences at the rate of about one per day and to have the children decoding simple words made up of those correspondences within a few days of starting school. Thus they realise from the start that one identifies printed words from the letters in them. They then go on learning a letter and a sound per day. This soon gets them on to the sounds represented by letter-combinations such as ‘sh’ and ‘ee’, but they would not be expected to read texts containing the word ‘sheep’ until they had been taught these digraphs, whereas children taught by the more typical British approach would be expected to read such words by recognising them as wholes or guessing them from pictures or context. One teacher who does teach synthetic phonics recently received a report from the local authority language consultant on one of her pupils. This report stated ‘He seems to be trying to sound out the words with some success, but without looking at the pictures for clues. He needs to be praised for giving a plausible word which fits in with the syntactic and picture clues without necessarily being the exact word’. This idea that a plausible word is as good as the exact word is very prevalent but surely wrong.
In practice, the synthetic-phonics approach enables children to make excellent progress, as has been shown in Clackmannanshire, where all the local authority primary schools participated in the project. Three groups of children were taught to read in different ways for just 16 weeks at the very beginning of their schooling. In the experimental group, the children were taught to read all words that they encountered by the synthetic-phonics approach – producing sounds for the letters and letter-groups and blending the sounds. The other groups in the project (the control groups) used an approach for reading in which they learnt words as wholes first and only then learnt to break them down into smaller units. This type of phonic analysis after words have been identified in some other way is common throughout the UK. After the 16-week programme (that’s all it takes to lay these foundations), the synthetic-phonics group was so far ahead in reading and spelling that it was considered unethical to withhold the programme from the other children. They were then given the programme and the results were reported for the group as a whole from then on. No further special teaching took place, but by the time the children left primary school nearly seven years later, their word-reading ability was three and a half years above the average for their age – their comprehension was only three and half months above average, but even that was very good, as I hope to show. Their spelling was 20 months above average. I’m not supposed to be talking about spelling today, but it is a pet subject of mine and it’s closely related to reading. The Clackmannanshire experiment suggests that when beginners are taught word-reading and spelling as mirror-images of each other, each powerfully reinforces the other and both benefit.
Despite its good results, this kind of approach for beginners is still treated with some suspicion in British schools. One reason is that there is a widespread belief that beginners naturally start by recognising words as wholes. A leading teacher-trainer has written, ‘Initially, whatever we try to teach them, young children recognise words as unanalysed wholes, making no attempt to map the component letters into speech sounds’ (Henrietta Dombey, Professor of Language and Literacy at Brighton University, in an article in Literacy Today in 1999). The fact is, however, that beginners can map letters on to speech sounds if taught to do so – this is how reading is routinely taught in many non-English-speaking countries. 16 weeks of such teaching in Clackmannanshire put the experimental group significantly ahead of the children taught to recognise words as wholes – this should not have happened if whole-word recognition is more natural for beginners. The problem is that most teachers are not trained to teach synthetic phonics – rather, they are trained to teach children to identify words as wholes, often by looking for clues in the pictures and context.
So the belief that beginners can process written words only as unanalysed wholes is one reason why phonics has not been favoured – but it’s a mistaken belief. Another reason is that an early emphasis on phonics is suspected of making children think that reading stops with decoding, and that understanding what they read is not important. This brings us on to the matter of comprehension. As we saw from text 3, if we can’t decode we can’t understand, so decoding is an important first step. There can be two main outcomes of competent decoding: one is that we understand as we decode, which is what happens with simple texts, and the other is that the decoding makes us realise that we have not understood, or not understood fully. As proficient readers, we do encounter texts that we don’t fully understand as we first decode them. But it is the decoding which gives us hints about what to do in order to increase our understanding. For example, we may need to look up unfamiliar words in a dictionary, or tussle with complicated sentence-structure. Or we may realise that however hard we try, a particular text will remain beyond our comprehension. But all this arises only with difficult texts, and these are not the texts we are talking about in connection with young children. Rather, we are talking about texts such as The Very Hungry Caterpillar: if children understand this as an adult reads it aloud to them, they will understand it if they can decode well enough to read it to themselves. It is only later that they will start encountering texts where decoding may not yield instant comprehension – but if they have not actually decoded the words accurately, how will they know whether they need to consult a dictionary or to grapple with syntax?
I said I would return to the matter of the Clackmannanshire children comprehending, by the end of primary school, only three and a half months above the average for their age whereas they were three and a half years above average in word-reading. This seems a big discrepancy. But bear in mind that decoding is a simple mechanical skill which can be managed even by people with severe learning difficulties, though they often do not understand what they read – this point was made by Prof. Margaret Snowling, of York University, in a recent Channel 4 Dispatches programme on dyslexia. By contrast, good comprehension depends on vocabulary, general knowledge and intelligence. As a group, the Clackmannanshire children were from the most deprived 10% of the population, a factor likely to depress vocabulary and general knowledge to some extent. Children from this sort of background would typically score not just below but far below the national average in comprehension, but the Clackmannanshire children scored significantly above the national average. This means that they comprehended better than children who had been taught more whole-word recognition and guessing from pictures and context, as the test had been standardised on a representative sample of children taught by this typical British approach.
How did children from the most deprived 10% of the population manage to comprehend significantly better than average children whose teachers had emphasised reading for meaning more and decoding less? Bear in mind that reading for meaning is commonly taught in the way suggested by the language consultant quoted above: children are taught to guess at words from pictures and context rather than to decode them accurately. This is surely not the best way of fostering comprehension. The Clackmannanshire teachers started by putting much more emphasis on accurate decoding, and found that as the children grew older, they could manage more advanced work than previous cohorts had done – the strong decoding start had evidently helped rather than hindered their comprehension.
To sum up: phonics is the holy grail of reading in the sense that it teaches decoding very systematically and decoding opens the way for comprehension. Decoding allows instant comprehension when texts are easy, and it allows readers to decide what other steps might assist comprehension when texts are difficult. So if comprehension is the ultimate goal of reading, teaching children to decode efficiently is a very good first step, according to both research and common sense.