Page 1 of 1
TES: Synthetic phonics? Not alone thanks.
Posted: Fri Mar 10, 2006 7:37 pm
TES: Synthetic phonics? Not alone thanks.
Kathy Hall questions whether it is a good idea to teach five-year-olds to
read using a single method
Before young learners know how sounds and letters map on to each other,
they see the word as a whole unit - they are not yet able to discriminate
at the level of individual letters and sounds.
As they get better at noticing and discriminating (supported by good
teaching), they pick up on familiar elements of words - perhaps their own
name within a word, the first letter, or a letter string such as "ing" at
the end. With good teaching and plenty of exposure to print in many forms,
they learn to discriminate more finely at the level of the letter, and
eventually, the process of discriminating (decoding) becomes automatic.
The important point is that beginner readers typically move from
discriminating among large units, such as whole words, to smaller units,
such as parts of words and individual letters. Put more technically, they
move from onset-rime awareness to phoneme awareness (see box).
There is good evidence available now to show that the sequence of
development is from sensitivity to large units (onset and rime) to
sensitivity to small units (phonemes).
Good teaching respects this developmental sequence. Therefore it is
misguided to say five-year-olds are best served only by synthetic phonics,
which is an approach which goes straight to the smallest unit (the
phoneme). It does not follow from this that teaching should not involve
synthetic phonics, but it does follow that analytic phonics also has an
Some children may not be able to benefit from an approach based only on
synthetic phonics and may therefore be disadvantaged. That's the first
reason why it's not a good idea to give such privileged status to synthetic
phonics in teaching reading to five-year-olds.
Synthetic phonics is crucial since the beginner reader needs to get to the
phoneme, or individual sound (the focus of a synthetic approach) in order
to become a skilled reader. But sensitivity to onset and rime (the initial
sound followed by the final group, as in c-at and b-at), which is the focus
of an analytic approach, comes first developmentally.
It's a mistake to try to persuade teachers to choose between synthetic and
other approaches to developing decoding, because decoding can be supported
by a range of processes. Because these processes support each other, it is
unhelpful to focus only on one or two at age five, although individual
lessons might do just that from time to time. The other supporting
processes for decoding are:
* Visual perception of letters and symbols (graphics).
* Meaning of words (semantics).
* Structure of phrases and sentences (grammar and syntax).
* Assumptions and beliefs about the task in hand (cultural).
Posted: Fri Mar 10, 2006 8:30 pm
I have now e-mailed Kathy Hall about this. She and I were on the same platform at the end of Oct. (see my 'Holy Grail' article in Newsletter 57) and although she didn't agree with me she seemed to take me reasonably seriously. My main argument this time has been that the arguments she is now using against synthetic phonics would not be taken seriously in non-English speaking countries with alphabetic writing-systems (e.g. Germany and Spain) and that this is not because of their more transparent orthographies. Rather, it's to do with the fact that they don't regard large-to-small-unit phonological development and a tendency (if it exists) to start by seeing words as whole as reasons for not teaching synthetic phonics to beginners.
Posted: Sat Mar 11, 2006 12:25 pm
Ironically traditional 'dyslexia' teaching programmes, although they differ from SP in various ways - teaching letter names, consonant blends, rules etc - always started at the phoneme level. It was only with the influence of Goswami's view of phonological development (alongside Frith's stages, from logographic, through alphabetic to orthographic, which justified initial 'sight words' alongside phonics), that this new layer was added on. It frustrates me to think that in adding these extra activities, I was only taking longer to get to the nitty-gritty of what my pupils needed to learn. Still, at least it gives me experience of other approaches to compare with the much faster progress of my pupils now I am using the Sound Reading System.
Unfortunately Goswami's ideas are very much in the mainstream now, enshrined in the literacy strategy along with all the materials which work on onset-rime. It is a shame her change of view isn't so well publicised. Hopefully with the publication of the final Rose report, it will become clear that an onset-rime approach isn't compatible with true SP programmes.
I think my group will be open to SP in most ways but I'm sure this issue will come up. I agree that a document outlining the principles of SP would be very useful. I am going to use Diane McGuinness's protocol as a basis too. I suspect that many SpLD colleagues may well think that what they are doing is SP -as it is systematic, thorough etc.
Posted: Sat Mar 11, 2006 3:00 pm
In case you and others find it difficult to get hold of the whole article, here are some relevant extracts:
'...the small units (phonemes) usually correspond to single letters, which are clearly separable in the orthography, and most words used in the early reading curriculum have a 1:1 correspondence between letters and sounds...Hence pedagogically it may well be easier to begin teaching them about letters, which many of them may already know about, and then to proceed to instruction about larger units such as rimes' (Journal of Experimental Child Psychology 82, 2002, pp. 49-50). [I think she probably has an over-rosy picture of the extent to which the 'early reading curriculum' currently focuses on words with '1:1 correspondence between letters and sounds', but at least she has realised that if this WERE the case, a phonemic approach would be far easier than she had hitherto thought.]
I wrote a pamphlet challenging the onset-rime theories in 1994. Goswami responded in an article in Journal of Research in Reading in 1995. I wrote to her periodically from then on pointing out that even if children naturally learn to distinguish larger phonological units before smaller phonological units, learning to read involves print units as well as sound units and print units are easiest to deal with at the single-letter level at first. She was saying that beginners, having learnt 'cat' as a sight-word, would then see it as 'c - at' before they saw it as a sequence of 3 letters, which seemed very illogical. For one thing, there is no obvious visual break between the onset and rime (whereas there ARE visual breaks between the letters) and for another thing, she was assuming that children would not notice the individual letters in the rime but would nevertheless notice enough detail about it to make use of it in reading analogous words such as 'bat' and 'hat'. If they could register visually on the individual letters at the BEGINNINGS (onsets) of these words, why couldn't they go on registering visually on the remaining individual letters? She eventually conceded the point in print, but not until 2002, unless I missed something earlier.
In the same article, she writes 'This potential conflict between what is easy phonologically versus what is easy orthographically has led Brown and Ellis (1994) to argue that beginning readers are faced with the difficult task of establishing a mapping between incompatible levels in the orthographic and phonological domains. The "easy" orthographic units are phonologically harder (and more inconsistent) while the phonologically easy units may be orthographically harder (although less inconsistent'. Of course in thinking that phonemic units are phonologically hard, she is thinking in terms of children having to work them out themselves - she hasn't fully realised that there is nothing hard about phonemes if one teaches them explicitly and she hasn't quite joined her thinking here up with the thinking in the earlier extract to realise that using letters as visual symbols for phonemes and sticking to cvc words at first makes things easier still.
I think that all this is good ammunition to use against people who are still stuck in the onset-rime time-warp.
Posted: Sat Mar 11, 2006 6:52 pm
I often see it said that 'research shows' that whole words, or onset & rime, are easier for children to process initially.
Am I right in assuming that the 'research' emanates from Goswami, or was she drawing on someone else's research for her assertion? Is the assertion actually based on any 'research'?
I'm asking for clarification because the last phrase in this quote seems to imply that Goswami's assertion was not actually based on research, but on her own theorey. Or, it might be just the way you've phrased it. :-)
'...the small units (phonemes) usually correspond to single letters, which are clearly separable in the orthography, and most words used in the early reading curriculum have a 1:1 correspondence between letters and sounds...Hence pedagogically it may well be easier to begin teaching them about letters, which many of them may already know about, and then to proceed to instruction about larger units such as rimes' (Journal of Experimental Child Psychology 82, 2002, pp. 49-50). [I think she probably has an over-rosy picture of the extent to which the 'early reading curriculum' currently focuses on words with '1:1 correspondence between letters and sounds', but at least she has realised that if this WERE the case, a phonemic approach would be far easier than she had hitherto thought.
She has been very influential in the teacher training and NLS world, hasn't she? I find references to her work all over the place. Student teachers must get the impression that there are only about 6 academics involved in reading research. All the books on teaching reading that I have browsed through in Waterstones contain the same references to the same handful of studies!
(They also contain the same assertions that make me hop up and down squeaking with rage!)
Posted: Sat Mar 11, 2006 10:33 pm
I will try to answer your questions, but as is so often the case, answers are not dead simple.
In a sense the research on onset-rime can be said to emanate from Goswami, though she was actually preceded in the field in the late 1970s and early 1980s by her Oxford professor (Peter Bryant) and an earlier doctoral student of his (Lynette Bradley). An often-cited paper of Bryant and Bradley’s was published in ‘Nature’ in 1983, and a book written by the two of them appeared in 1985. The 3 of them are sometimes referred to as ‘the Oxford group. Goswami and Bryant published an influential book in 1990. Goswami, however, seems to have ended up making the onset-rime field more her own than the others – e.g. she produced a ‘Rhyme and Analogy Teacher’s Guide’ for OUP (published in 1996).
Reading between the lines of what they have written, I deduce that all the research they conducted was with children taught by whole-word and whole-language methods, with little or no phonics. I believe that they DID find what they said they found (that children who had better pre-school rhyming skills became better readers), so in that sense the findings were research-based. But they didn’t take into account the likelihood that the issue might be affected by the way these children had been taught – if children are not taught systematically about the alphabetic code, those who are the most sensitive to sounds at onset-rime level will probably be the most successful at working some of it out by themselves. So this WAS research, and it DID show a connection between early rhyming ability and later reading ability, but only among children taught in a certain way - but as that was the way virtually all children were taught at the time, it seems not to have occurred to the researchers that different teaching might change the picture, or, indeed, that different teaching was even possible. Mona and Sue L. were writing to them way back in the 1980s saying this, but they got no response. Although phonics has come a bit more to the fore since then, we are basically still in a similar position in that most people still think in terms of children being taught sight-words and whole-language strategies (e.g. guessing from context etc.). A bit of onset-rime teaching thrown into that mix probably WILL raise standards a bit, and that’s good enough for many people – what they don’t realise is that good phonics teaching would raise standards far more.
I think that the whole-word theories and the onset-rime theories have actually fed off each other. As far as I know, the idea that children start by recognising words as wholes was not based on any serious research – in fact I think it started with people thinking that children got more quickly into reading meaningful sentences by being taught whole words, and then gradually transmogrified into the theory that they were programmed by nature to do this. The unquestioning acceptance of whole-word teaching by Oxford academics in the 1980s, however, probably gave it a new lease of life, making others even less likely to question it than they had previously been. The Oxford group were also saying very firmly that children found it very hard to analyse words into phonemes, which was grist to the whole-word mill. In 1994, Goswami said the following in print in the UKRA’s (now UKLA) ‘Reading’: ‘If we think in terms of the phonological knowledge that the average child brings into the classroom on the first day of school, then we can see that at least one method of teaching reading would pose unnecessary difficulties for this child, a method relying on tradition “phonics”. In traditional “phonics”, children are taught the spelling-sound correspondences for all of the phonemes in spoken English. The rationale is that once children know the “rules” for sounding out the different letters in words, then they can “build up” the pronunciation of any word that they meet by applying these rules. This logic is impeccable. The problem is that for most children, learning about individual spelling-sound correspondences is a rather difficult way into reading, as they cannot hear these individual sounds in the words that they are trying to decode’ (p. 33). I found this quite breathtaking: Goswami gave a pretty good account of what would now be called ‘synthetic phonics’ but at the end of it failed to realise that no hearing of the individual sounds in the words they were trying to decode was necessary apart from that which occurred as children sounded out the letters – unaccountably, she seemed to be thinking in terms of the children at that point hearing the whole spoken word and having to analyse it into phonemes. But she went from saying this, in 1994, to saying, in 2002, that a letter-sound approach was probably the easiest for beginners, and this WAS a change. That’s why I said that she now seemed to realise that a phonemic approach was easier than she had hitherto thought.
Anyway, the 1990s were a period in which onset-rime theories became very prominent: people were convinced that once children had started by recognising words as wholes, the most realistic way for them to start breaking them down was into onset and rime. In 1999, Goswami stated in print that children needed a good sight-vocabulary in order to use onset-rime strategies effectively (‘Reading development and the Teaching of Reading’, eds. Oakhill and Beard, p. 184).
I could go on and on, but had better not. It must be hard for relative newcomers on the scene to understand the way that past history has affected the picture as it now is. I think, though, that we still face the problem of people being sufficiently wedded to whole-word and whole-language teaching, with a bit of breaking-down of words after they have been identified, for phonics first, fast and only to seem quite a foreign concept. And as I see it, there is no simple way of dismissing the onset-rime approach as not based on research - it IS based on research, but research making too many wrong assumptions and incorporating some logical fallacies. Incidentally, even someone as high-powered as Marilyn Jager Adams has been impressed by it and has gone down the related track of believing that one has to train children in phonemic awareness before teaching them to decode phonemically. She may be excellent on the three-cueing system, but she doesn't get everything right. This can be a huge problem with research - people can get some things right and some things wrong, and it's very difficult to unpick it in a way that everyone can understand.
Sorry this is so long - it's been a hobbyhorse of mine since the early 1990s.
Posted: Sat Mar 11, 2006 11:47 pm
I'm glad that it was helpful - I was worried that I was going on too long, but the detail is important. Note, though, that a point I was trying to make was that it is not always easy to refute the 'research shows' mantra. The onset-rime research does, in a sense, show what it purports to show, but only if you start from its starting-point (whole-word reading), and we don't. I sometimes think that it's a bit like that joke where a traveller asks an Irishman 'Is this the way to x?' and the Irishman replies 'If I were going to x I wouldn't start from here'. In the case of the joke, obviously the traveller has to start from where he is, but in our case starting from somewhere else is the only way that the final destination will be reached. It's very difficult, though, for people standing where they are to see that another starting-place might be better.