My response to some of Maggie’s most recent points:
Maggie: Indeed, he [Smith] was echoing much earlier theorists, such as Huey, in this belief and, of course, by the time he was writing many readers may have been using such strategies because of being taught by Word methods (I’m sticking to my hypothesis!). I can’t find that he has any evidence for his assertion and, as I pointed out, Stanovich and West disproved his theory.
Me: The first five chapters of Snowling & Hume’s book The Science of Reading are devoted to reviews of work on word recognition processes in reading. Most of the research looks at the ways in which adult, expert readers read. What emerges from these five chapters is that:
• expert readers do not use one single method for reading words; they tend to use rapid whole-word recognition for familiar words and slower, stepwise decoding for unfamiliar words;
• the speed with which they respond to target words increases in response to different types of priming;
• the jury is still out on how reading mechanisms actually work.
It was the fact that expert readers use two strategies that resulted in a plethora of ‘dual route’ models of reading; the first was proposed in the 1920s, but studies of brain-damaged patients had noted this in the 19th century. This is exactly what West and Stanovich found. What they ‘disproved’ was that the use of contextual information by children increased with age and reading ability.
There was a great deal of work on priming effects in reading during the 1970s, so although Smith might have been wrong, he wasn’t just ‘echoing earlier theorists’. He had a PhD in psycholinguistics/cognitive psychology from Harvard, so would have been very familiar with the direction of travel in contemporary reading research.
Your hypothesis that expert readers were using mixed methods because that’s how they’d been taught to read, might be right. But a more likely explanation is that recognition of complex sensory stimuli (e.g. words) becomes automated and fast if they are encountered frequently, but requires step-by-step analysis if they’re not. That’s how human brains deal with complex sensory stimuli.
There is no question that expert readers use more than one strategy when reading. The question is whether explicitly learning those strategies is the best way for children to learn to read.
the rejection of the alphabetic principle
Me: Maggie says my statement that the alphabetic principle and analytic phonics had been abandoned because they hadn’t been effective for all children ‘makes no sense at all’. If I’m wrong, why were these methods abandoned?
Maggie: I still don’t think it makes any sense. For a start, you give no time scale. When did this abandonment take place? And you are conflating Alphabetic with Analytic which I don’t think is correct (see my earlier comment).
Me: They were abandoned gradually. My PGCE reading tutor, who trained in the 1930s, was keen on analytic phonics but not on ‘flashcards’. I remember spending hours preparing phonics reading activities. Several teachers of her generation that I’ve spoken to, took a similar view. They didn’t advocate using analytic phonics ‘systematically, first and only’, but as a support strategy if children were struggling to decode a word. Clearly, the teachers I’ve encountered don’t form a representative sample, but some of them were using analytic phonics until they retired and at least one teacher training college in the UK was teaching students to use it until at least the late 1970s. And this definitely wasn’t ‘alphabetic’, it was phonetic. According to my reading tutor, the alphabetic method was widely perceived as flawed by the 1930s. The consensus amongst these teachers was:
• children use a range of strategies when learning to read
• whatever method of teaching reading is used, some children will learn with little effort and others will struggle
• no one method of teaching reading will be effective for all children, but some methods are more effective than others (which is why they still used analytic phonics).
I’m not saying they are right, but that’s what they thought.
Maggie: Another point is that you are crediting educationists and teachers with a degree of rationality which I don’t think is justified. The widespread acceptance of the Word method, which had no evidence to back it but strong appeals to ‘emotion’ with the language of its denigration of Phonic methods, is a case in point. Boring, laborious, ‘drill & kill’, barren, mechanical, uncomprehending, the list is long (and very familiar). It is a technique promoted today as ‘framing’ (though I might acquit its original users of deliberate use of it). Very easy to be persuaded by the language without really considering the validity of the method it purports to describe.
Me: I think you are not crediting them with enough rationality. The ‘drill and kill’ they were referring to was an approach many teachers resorted to in the early days of state education. Those teachers were often untrained, had to teach large numbers of children of different ages, had few books, were on performance related pay, used corporal punishment and had been taught themselves through rote learning entire lessons. Complaints about children being able to recite but having no understanding were commonplace in those early days. What has happened over time is that denigrating rote learning everything (justified in my view) has morphed into denigrating rote learning anything (not justified).
Prior to the 1980s, teachers in the UK were left to their own devices about how they did things, and some at least, took a keen interest in developing their own methods; they didn’t all slavishly follow fashion by any means. I agree that the ‘Word’ method might have been framed emotively, but it’s not true to say there was no evidence to back it.
The evidence was in the form of adult reading strategies. If you’re a teacher who’s seen ‘drill and kill’ not working for all children, then alphabetic and analytic phonics not working for all children, and someone comes along and tells you that scientific research has shown that adults use a range of strategies when reading (and you check out the research and find that indeed it has shown just that) so it would make sense to teach children to use a range of strategies to learn to read, what would you, as a rational person, do?
I think you are seeing claims that adults use a range of reading strategies through the spectacles of the ‘teaching reading’ literature, not through the spectacles of the ‘reading mechanisms’ literature. The body of evidence that supports the idea that adults use a range of strategies in reading is vast. And every teacher will have witnessed children attacking words using a range of strategies. Putting the two ideas together is not unreasonable. It just happens to be wrong, but it wasn’t clear that it was wrong for a very long time.
Maggie: I would also suggest that the discourse of ‘science’, ‘research’, ‘progressive’ would be enough to convince many without them delving too deeply into the evidence. Brain Gym, anybody?
Me: You’re quite right. The point I’m making is that there was robust evidence to support the Word method. But it was robust in respect of people who had learned to read, not those who hadn’t. The way the brain functions after learning something (in adults) doesn’t reflect the way it learns it (in children). But that was by no means clear in the 1970s. There is still a dispute going on about this amongst cognitive scientists.
using a range of cues
Me: The cues I listed are those identified in skilled adult readers in studies carried out predominantly in the post-war period. Maggie’s hypothesis is that the range of cues is an outcome of the way the participants in experiments (often college students) had been taught to read. It’s an interesting hypothesis; it would be great to test it.
Maggie: I stand by it! I have worked with too many children who read exactly as taught by the Searchlights!
I thought I would revisit these ‘cues’ which are supposed to have offered sufficient exposure to auditory and visual patterns to develop automated, fast recognition. They are ‘recognising words by their shape, using key letters, grammar, context and pictures, recognising words by their shape’.
Confounded at once by the fact that many words have the same shape: sack, sick, sock, suck, lack, lick, luck, lock, pock, pick, puck, pack,
using key letters, Would those be the ones that differentiate each word in the above word list?
grammar, Well, I can see how you might ‘predict’ a particular grammatical word form, noun, verb, adjective etc. but the specific word? By what repeated pattern would you develop automatic recognition of it?
context I think the same might apply as for grammar. You need a mechanism for recognising the actual word.
pictures, Hm. Very useful for words like oxygen, air, the, gritty, bang, etc.
Me: Again, you are confusing the strategies adults use when reading with the most effective way of teaching children to read. They are two different things. Your examples illustrate very clearly why using multiple cues isn’t a good way of teaching reading. But those inconsistencies don’t stop adults using these cues in their reading. If you don’t have a copy of Snowling and Hume’s book, get one and read it.
Maggie: In view of Stanovich & West’s findings I would be interested to see any studies which show that skilled adult readers did use the ‘cues’ you listed. (as above)
Me: There’s a vast literature on this. Summarised very well in Snowling and Hume, which is why I’ve recommended it. Incidentally, a ‘cue’ isn’t a term invented by proponents of the Word method, it’s a perfectly respectable word denoting a signal detected in incoming information; it can affect subsequent information.
Me: In chapter 2 of Stanovich’s book, West and Stanovich report fluent readers’ performance being facilitated by two automated processes; sentence context (essentially semantic priming) and word recognition.
Maggie: I appreciate that but this is described as a feature of fluent, skilled reading. To assume that beginning readers do this spontaneously might be to fall into the same trap as ‘assuming that children could learn by mimicking the behaviour of experts’
Me: In your original post, you said “Stanovich and West showed, in the 70s that these were strategies used by unskilled readers and that skilled readers used decoding strategies for word recognition (this is an extreme simplification of the research Stanovich outlines in ‘Progress in Understanding Reading’) and this has been the conclusion of cognitive scientists over the subsequent decades the validity of these strategies is seriously challenged.”
I think you’ve misunderstood what Stanovich and West (and other cognitive scientists) have shown. The literature shows, pretty conclusively, that fluent readers use word recognition first and decoding if word recognition fails. Sentence context isn’t used as a conscious strategy, it’s subconscious, because the content of the sentence increases access to words are semantically related. It’s not safe to assume that because experts do something, novices learn by copying them. Nor is it safe to assume that experts use the same strategies they did when learning as novices.
Me: According to chapter 3, fluent readers use phonological recoding if automated word recognition fails.
Maggie: Isn’t that the whole point. Fluent readers didn’t use context, or other ‘cues’, to identify unfamiliar words, they used phonological recoding.
Me: No. The point is that they used it if automated word recognition failed.
Maggie: It is also moot that they use context to predict upcoming words (although I do understand about priming effects). There is also the possibility that rapid, automatic and unconscious decoding is the mechanism of automatic word recognition (Dehaene). Possibly with context confirming that the word is correct? A reading sequence of ‘predicting’, then, presumably, checking for correctness of form and meaning (how? by decoding and blending?) seems like a strange use of processing when decoding gets the form of the word correctly straight away and immediately activates meaning.
Me: It’s possible that rapid, automatic and unconscious decoding is the mechanism of automatic word recognition but work on masking and priming suggests that readers are picking up the visual features of letters and words as well as their auditory features and semantic features. In other words, there are things going on in addition to decoding.
Whether readers use context to predict upcoming words depends on what you mean by ‘predict’. Priming results in some words being more likely than others to occur in a sentence; this isn’t a conscious process of ‘prediction’ but it is a subconscious process of narrowing down the possibilities for what comes next. But in some sentences you could consciously predict what comes next with a high degree of accuracy.