Phonics is not the dirty word it once was, but despite its reappearance on the literacy agenda, policy on the teaching of reading remains unbalanced. There is an over-emphasis on the analysis of spoken words into smaller units of sound and a serious under-emphasis on the synthesis of spoken words from smaller units of sound. In other words, analytic phonics is stressed at the expense of synthetic phonics.
So what's the problem? Either way, surely, children are learning about phonemes (the smallest units of sound in spoken words) and the way they are represented by letters and letter-groups. Well, the idea: of teaching children to analyse spoken words into phonemes is not, in itself, the problem: the problem is that analysis into phonemes is now assumed to play a direct part in reading (as well as in spelling, where it has traditionally belonged) and, most worryingly, is assumed to be possible only after children have gone through an initial phase of reading words 'logographically' (as wholes). A clear statement of this position comes from Henrietta Dombey, Professor of Literacy in Primary Education at the University of Brighton:
'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. [Frith] terms this the logographic phase, stating that towards the end of this phase children may well notice some at least of the letters involved. But they only start to make systematic use of this knowledge when they enter the next phase, what she terms the alphabetic or analytic phase. Here they are learning to relate letters and groups of letters to phonemes and groups of phonemes. In other words, they are learning phonics' (Literacy Today No. 20, September 1999).
According to Prof. Dombey, then, children learn phonics only after an initial phonics-free logographic stage, and teaching is powerless to change this sequence. Such views are unacceptable from a synthetic phonics perspective.
The reading of words as logographs or 'unanalysed wholes' is not universal among young children, as Wimmer and Hummer showed in an article published in 1990: 'How German-speaking first graders read and spell: doubts on the importance of the logographic stage' (Applied Psycholinguistics Vol. 11, pp. 349-368). The fact that English-speaking children manifest logographic tendencies while other children do not is often attributed to the unusual complexity of English letter-sound correspondences, which is thought to dictate a whole-word approach. Of course the English writing-system is complex, but this does not inevitably mean that children read logographically 'whatever we try to teach them'. There is ample evidence that they do not read logographically if we teach them synthetic phonics. This involves teaching them to produce phonemes in response to letters and then to blend or synthesise (hence 'synthetic phonics') these phonemes into a word-pronunciation. Phonics-first teachers will welcome the recent study by Dr Karin Landerl (University of Salzburg) showing that this can be very effective, even in English, and suggesting that it may be rendered more necessary, not less so, by our complex writing-system (European Journal of Psychology of Education 2000, Vol. 15 No 3, pp. 239-257).
Landerl's study arose out of an earlier study by Wimmer and Goswami, who thought that teaching methods might account for the differences they found between English and German children, but were more inclined to blame the complexity of our letter-sound correspondences. Prof. Goswami has in the past suggested that this complexity, combined with preschool rhyme awareness, makes it more natural for English children to use rhyme units than to use phonemic units in their early reading. In a recent article, however, she has stated that the picture is 'dramatically affected by the type of literacy instruction that the child is receiving' (Goswami and East: 'Rhyme and analogy in beginning reading: conceptual and methodological issues', Applied Psycholinguistics 21, 2000, p. 82). Her growing recognition of the major role played by teaching is most welcome. But problems remain: the teaching provided by Goswami and East certainly improved the children's rhyming skills, but did it also have a dramatic effect on their reading? We are told the average reading age before this teaching but not after. What useful information can we glean from the 'before' picture? At first sight, the children's average reading age of 5 years 8 months at an average chronological age of 5 years 6 months seems quite healthy — but we have to ask how much the average reading age was inflated by the decision to exclude children with 'documented educational or hearing difficulties' (p. 73). The teaching approach which produced the 'before' picture was 'a mixed program of reading instruction that included the establishment of a sight-word vocabulary (using reading schemes based on a whole language approach) and some systematic tuition in letter-sound relationships' (p. 73). Elsewhere, this type of approach is said to be 'characteristic of early reading instruction in England' (p. 71). True — but note the difference when this 'characteristic' approach is replaced, from the start, by synthetic phonics, as in the Clackmannanshire study by Watson and Johnston (Interchange 57, published in 1998 by the Scottish Office). The whole-class teaching of synthetic phonics, with no children excluded because of their difficulties, clearly produced better results than the 'mixed program' accepted by Goswami and East as a foundation enabling most (but not all) children to benefit from onset-rime training. The same reading test (British Ability Scales) was used in both studies:
Goswami and East subjects
Watson and Johnston subjects
sight words, whole language, some phonics
5 years 6 months
5 years 5 months
5 years 8 months
(children with documented difficulties excluded)
6 years 0 months
(no children excluded)
The Goswami and East subjects were then given onset-rime training in groups of three or four for about five hours spread over eight weeks. This may have raised their average reading age dramatically, but we simply do not know: although reading was re-tested after the training, the results are not given. Why not? Onset-rime training is not an end in itself, but a means to an end, where the 'end' is good reading attainment. And we still do not know whether onset-rime training works with problem readers or even with genuine beginners — the Goswami and East subjects were not genuine beginners, as they had already had several months of a 'mixed program' with, it seems, a good deal of emphasis on sight-words. The belief in sight-words as a first step is found everywhere.
Synthetic phonics may sound abstract (and even artificial, as that can be one of the connotations of 'synthetic') but it is actually very practical: synthesising gives the children an immediate practical use for their knowledge of letter-sounds. With logographic or whole-word reading, by contrast, children have to leave any letter-sound knowledge they may have on a back burner until the later analytic phase. Alternatively, perhaps, they are just not taught anything about letters and sounds in the earlier phase — which makes it entirely unsurprising that they make no use of phonics in reading, but leaves one wondering how on earth they tackle spelling
Most phonics-first teachers accept that a logographic start is possible, but would stress that it is not inevitable, and, crucially, that it does not make it easy for children to progress to more mature strategies. This last point is conceded even by Prof. Dombey:
'Children who have acquired quite a wide reading vocabulary in the earlier logographic phase may well need cajoling, repeated prompting and considerable support to tackle words analytically' (Literacy Today No. 20, September 1999).
So why not spare children the 'earlier logographic phase' and simply teach them from the start to sound out the letters in words and then blend the sounds? Granted, there will come a point at which we have to tell them that this will not work for every part of every word they encounter, but phonics-first teachers find that children negotiate this hurdle without 'cajoling, repeated prompting and considerable support'. In other words, it seems a much easier step than the step from logographic to analytic reading, which evidently does require cajoling etc.
Galvanized by Prof. Dombey's Literacy Today article and by the conviction that the logographic-reading-comes-first view was at the root of many of our problems, I wrote an article on this subject and submitted it to a different teacher-friendly publication on literacy. The article was rejected and returned to me with the comments of the reviewers who had recommended rejection. One of them had written the following:
'Most adults who have been around most children know that most children do perceive logographically at first.'
I had hit precisely the barrier that I was trying to break through! This reviewer was simply not going to allow any challenge to the view that the first stage of reading is logographic: how could I be so stupid as to question something that is believed by pretty well everyone who knows anything about children? Well, I do question it, and will continue to do so. Teachers must not be brainwashed into believing that logographic reading is natural if in fact it is the result of teaching, as the evidence suggests to be the case.
Unfortunately, most of the work on phonemes in the National Literacy Strategy materials leans towards the analytic end of the spectrum, frequently with the assumption that the word is first recognised as an unanalysed whole. Take, for example, the activity starting on p. 29 in the Additional Literacy Support: Phonics and Spelling, Reading (Guided and Supported) book. The object of this activity is given as being 'to practise reading words', and yet the procedure involves the child 'reading the word, saying the phonemes, and the word again — thump, th-u-m-p, thump'. This is back to front: to use phonics in a practical way for reading, the child should first say the phonemes in response to the graphemes (letters and letter-groups) and then blend (synthesise) these phonemes into a pronunciation for the whole word. Saying the phonemes after the word has been read is a type of analysis which contributes nothing to the reading of the word.
Another example comes in the Additional Literacy Support video. We see a Year 3 child struggling to read the word 'soft'. We actually hear her correctly sounding out each letter under her breath. Instead of helping her to blend the sounds, however, the classroom assistant says 'Look at the first bit. What would make sense?' — in come the whole-language strategies and bang goes a perfect opportunity to demonstrate the way synthetic phonics works for reading. One is left feeling that the people who have overseen the production of these materials simply do not understand the role of synthesising.
Teachers who do understand that reading is about producing phonemes in response to graphemes and then blending (synthesising) the phonemes ensure that their pupils learn graphemes, phonemes and blending first and fast. They know that they will not be teaching words as wholes and that the children may be left unable to read any words at all unless they are good at sounding out and blending. Teachers who use whole-word methods, however, tend to take their time over teaching graphemes and phonemes (as does the National Literacy Strategy), and not to put much emphasis on blending. The contrasting results are nicely illustrated in two inspection reports (available at www.ofsted.gov.uk), one on Barkerend School in Bradford, which is featured in the Additional Literacy Support video, and the other on Kobi Nazrul School in Tower Hamlets, whose headteacher at the time, Ruth Miskin, was a leading proponent of the sounding-out-and-blending approach for reading. The intakes of the schools appear to be comparable in the sense that many children have limited English on entry. Here are some relevant comments from the OFSTED reports:
Barkerend: 'By the end of Key Stage 1.... some [pupils] know the sounds of letters but have difficulty in transferring their knowledge to the reading of texts'.
Kobi Nazrul: [children who have been through the school's nursery] 'start reception knowing all the letter names and sounds and how to write them'.
Barkerend [at the end of Key Stage 1] 'standards in reading are below the national expectation. Many pupils lack confidence, are unable to read with appropriate fluency and are just beginning to master key words to give them access to a range of books'.
Kobi Nazrul: 'Key Stage 1 test results match the best in the country....By the age of seven...most pupils read accurately and fluently. They read aloud with very good expression....'.
Why, why, WHY are the 'experts' not more open to the possibility that synthetic phonics first and fast produces much better results than phonics which is still dragging on at the end of Key Stage 1 and is mixed up with the learning of 'key words' (presumably as sight words)?
There are parallels between events in the BSE crisis and the field of literacy. In both cases, government ministers and the public have been misled by 'experts' (civil servants and others — some, at least, well meaning). People are rightly horrified at the eighty or so deaths which BSE has caused. What about the many thousands of innocent children who have been crippled intellectually, and/or psychologically and/or in their job prospects, by blunders and cover-ups over the teaching of reading?