Dual Route reading model

Moderators: Debbie Hepplewhite, maizie, Lesley Drake, Susan Godsland

Hugo
Posts: 107
Joined: Mon Apr 07, 2008 7:42 pm

dual route reading model

Post by Hugo » Tue Sep 23, 2008 2:41 pm

Jenny, I agree with your demands. I also think Smith satisfies them!
www.hugokerr.info

chew8
Posts: 4152
Joined: Sun Nov 02, 2003 6:26 pm

Post by chew8 » Tue Sep 23, 2008 4:06 pm

I do not think Smith satisfies them but will flip through the books I have by him before saying more.

Jenny C.

JIM CURRAN
Posts: 3123
Joined: Fri Oct 31, 2003 7:18 am

Subject

Post by JIM CURRAN » Tue Sep 23, 2008 9:42 pm

“ That direct instruction in alphabetic coding facilitates early reading acquisition is one of the most established conclusions in all of behavioral science ( Adams 1990; Anderson et al., 1985; Chall 1983b, 1989; Perfetti, 1985; Stanovich, 1986b). Conversely , the idea that learning to read is just like learning to speak is accepted by no responsible linguist, psychologist, or cognitive scientist in the research community ( see Liberman & Liberman, 1990 ).” ( Progress In Understanding Reading, page 415, Professor Keith Stanovich )

chew8
Posts: 4152
Joined: Sun Nov 02, 2003 6:26 pm

Post by chew8 » Wed Sep 24, 2008 7:49 am

Hugo and I agreed that Frank Smith had probably not done empirical research of his own. I said that people like this should at least act responsibly by making sure that what they said was in line with the findings of empirical research. Hugo said he thought this was true of Frank Smith - I said that I thought otherwise but would re-read some Smith before saying more.

I have now re-read most of his 1978 book 'Reading' (second edition). Virtually no references to scientific research studies are made in the text. Just seven references are listed at the end - three have Smith himself as sole author and one has him as a joint author.

I'll give just one example of something he says which I think is irresponsible. On page 107, he criticises the view that 'a child should be taught letters of the alphabet one at a time, given lots of practice until a is learned and then moved on to b and so forth'. He goes on to propose his own theory: 'But the view I am outlining suggests that in order to learn how to identify any letter you must see what the alternatives are. Children cannot even begin to learn to recognize a until they can compare it with every other letter that is not a'.

He does not explain how children can learn all 26 letters at once and provides no references to research studies showing that it is possible. If his theory fitted in with common sense (and logic), research references might not have been necessary. But his theory is far removed from common sense, and he should not expect people to accept it unless he can point to research showing how the learning of all 26 letters at once was done and that there were statistically significant results in favour of doing it this way rather than the one-letter-at-a-time way.

Jenny C.

Hugo
Posts: 107
Joined: Mon Apr 07, 2008 7:42 pm

dual route reading model

Post by Hugo » Wed Sep 24, 2008 9:15 am

Jenny, I really don't want to get into a confrontational discussion, but "reading" is not Smith's best book and the edition you refer to is 30 years old. My edition of "understanding reading" which is his most important book I think has references from pages 333 - 360. This is a fairer sample I think.

I am also of the opinion (paradoxical, given my otherwise ruthlessly scientific bent) that absolutely slavish demand that every remark ever made be based on research means that often we run up blind alleys in thrall to very few pieces of research and get lost among flawed or conflicted findings. We have also to stand back and use our common sense. Not much research is truly "counter-intuitive" after all. Intuition is usually about right. So much research is very obviously flawed (I read a lot about dyslexia, for example, a field saturated with post-modernist credulity and biased or weak reasoning) and our innate good sense should more frequently be deployed rather than an absolute and invariant deference to "findings". As Jim has pointed out on this board, research shows that 50% or so of research is wrong! (I know that is a mad statement. In fact it probably proves my point for me!) In my family, if you want to indicate that you are saying something about which which people's sceptical antennae should vigorously twitch you begin by saying "studies show ...".
www.hugokerr.info

chew8
Posts: 4152
Joined: Sun Nov 02, 2003 6:26 pm

Post by chew8 » Wed Sep 24, 2008 2:34 pm

I've read 'Understanding Reading' in the past but I don't have a copy of my own and am reluctant to buy one at the new and second-hand prices I've seen, so I can't easily revisit it. I've managed to read the Preface and the first 31 pages on line, however. Some bits are more or less word for word as in 'Reading'.

On p. 3 he says that the 'best strategy for working out the identity of meaning of an unfamiliar word is to work it out from context....An equally good way in different circumstances is simply to ask someone what it is....A very poor strategy is to try to "sound it out"'. He is confusing word-identification and comprehension.

I'm not inspired by statements such as 'Allusions to scientific studies don't prove a thing...' (p. 4) and 'Reading print is as natural as reading faces' (p. 5). There is a general consensus among serious researchers that reading print is not natural.

What he says about 'The Alphabet' (p. 5) is very different from what Diane McGuinness says, and I find Diane much more convincing. He is right that people can learn to read non-alphabetical writing such as Chinese, but he skates over the fact that even proficient Chinese readers are able to read far fewer words than proficient English readers.

I agree with you, Hugo, that we can't expect to find research evidence for everything and that we often have to trust our common sense. I made a similar point when I said, in my earlier message, that Smith's failure to cite research about children learning alphabet letters wouldn't bother me if his theory were in line with common sense - but it isn't. The fact that the book from which I was quoting is 30 years old and not Smith's best is neither here nor there - even 30 years ago there was no justification for what he said.

Jenny C.

kenm
Posts: 1495
Joined: Sat Dec 17, 2005 5:19 pm
Location: Berkshire

Post by kenm » Wed Sep 24, 2008 8:04 pm

chew8 wrote:I[...]I'm not inspired by statements such as 'Allusions to scientific studies don't prove a thing...' (p. 4) and 'Reading print is as natural as reading faces' (p. 5).

They inspire me - with a strong feeling of distrust of any other statement made by someone as ignorant of evolutionary processes as he appears to be.
There is a general consensus among serious researchers that reading print is not natural.
How could it be? A rather small number of our ancestors over the last 40000 years could do it, so there has been very little for natural selection to work on.
"... the innovator has as enemies all those who have done well under the old regime, and only lukewarm allies among those who may do well under the new." Niccolo Macchiavelli, "The Prince", Chapter 6

JIM CURRAN
Posts: 3123
Joined: Fri Oct 31, 2003 7:18 am

Subject

Post by JIM CURRAN » Wed Sep 24, 2008 10:05 pm

'As Jim has pointed out on this board, research shows that 50% or so of research is wrong'

That’s not what I said Hugo. The National Reading Panel ( 2000 ) had to reject most of the research on reading since the 1960’s as it didn’t fulfill the criteria for scientific research.

A. Published in a peer- reviewed journal
B. Replicated
C. Part of a cumulative body of research

Educational research has generally been pretty shoddy and nowhere more so than with whole language type research. Good research is invaluable.


Here are just a few gems from Frank Smith courtesy of Martin Kozloff ‘A whole Language Catalogue of the Grotesque’

‘ Reading without guessing is not reading at all’ ( Psychology and Reading 1973 )

‘Reading by phonics is demonstrably impossible, ask any computer’ ( Essays Into Literacy 1986 )

‘Phonics, which means teaching a set of spelling to sound correspondence rules that permit the decoding of written language into speech, just does not work’ ( Reading without nonsense 1985 )

‘ The art of becoming a fluent reader lies in learning to rely less and less on information from the eyes’ ( Comprehension and Learning 1975 )


‘ The first alternative and preference is – to skip over the puzzling word. The second alternative is to guess what the unknown word might be. And the final and least preferred alternative is to sound the word out. Phonics, in other words comes last.’ ( Language Arts 1999 )

http://people.uncw.edu/kozloffm/wlquotes.html

User avatar
Debbie Hepplewhite
Administrator
Posts: 3648
Joined: Mon Aug 29, 2005 4:13 pm
Location: Berkshire
Contact:

Post by Debbie Hepplewhite » Wed Sep 24, 2008 11:16 pm

How could anyone, EVER, have followed this man? :shock:

User avatar
maizie
Administrator
Posts: 3121
Joined: Sun Mar 14, 2004 10:38 pm
Location: N.E England

Post by maizie » Wed Sep 24, 2008 11:34 pm

I tried to be fair and consider that Kozloff's quotes might look different when seen in context, but I just couldn't imagine a context which would make this one look sensible:
‘ The art of becoming a fluent reader lies in learning to rely less and less on information from the eyes’ ( Comprehension and Learning 1975 )
Where do we get the information from then? Back to nasal learners, is it?

Does this explain why some children try to read without looking at the words?

Is there hope for me yet, as my eyesight succumbs to advancing years? As a fluent reader I won't actually need to see the words to be able to make meaning from them - Hooray! :grin:

g.carter
Posts: 1859
Joined: Wed Nov 05, 2003 7:41 pm

Post by g.carter » Thu Sep 25, 2008 8:28 am

Didn't Morag Stuart say that Frank Smith was very charismatic (in the session with you, Debbie, before the H of C Select Committee)? Maybe they were short on charisma at the Institute of Education when they fell for him, hook, line and sinker.

In case anyone hasn't read Martin Kosloff, leading expert on autism - aka Professor Plum - there is no-one who writes better on the absurdity of the Whole Language debacle.

What is absolutely mind-blowing is the fact that there are highly paid, intelligent people in the DCSF civil service who apparently don't understand the difference between Whole Language/Reading Recovery and Synthetic Phonics. Have they not read Kosloff, Hempenstall, Diane McGuinness, let alone the wealth of other material? We read this material - are unpaid - and yet DCSF can plead ignorance.

Hugo
Posts: 107
Joined: Mon Apr 07, 2008 7:42 pm

dual route reading model

Post by Hugo » Mon Sep 29, 2008 9:05 am

Going back over the correspondence which set off to be about the dual route model, I note that almost nobody, including myself, has addressed the original essay written by Susan ten days ago and seeking our comments. I will put my two penny worth in now, as I should have done then.

Susan questions that so many can “believe that there are two independent ‘routes’ or pathways the brain uses to read words”. This idea seems to me, though, to be self-evident, as well as being a widespread view among cognitive psychologists everywhere and in no way, or so it seems to me, any danger to any of our ideas about the teaching of literacy. I will use an example from my other world to step back and invoke intuition.

This morning I saw some lovely fat cattle in the field behind our house. “Ah, cattle!” I said to myself. Then I thought, “Of course, that’s the analogy I need for the dual-route model of reading!” (I don’t always think in exclamation marks, but I did this morning.)

When I see a bullock I know instantly, by sight I am confident, that it’s a bullock. I do not see an animal, scan the brain for a cross-reference to the sound it makes when it does, find a match with the lowing sound, take that sound into the semantic system, find a match there & thereby identify a bovine. I do it directly and visually. It would be crazy to suggest otherwise (in my opinion).

The lowing-of-cattle entry in my reference bank of sounds, however, does exist and will also, and inevitably, be somewhat activated by the sight of a bullock because my brain associates every aspect of everything immediately and comprehensively - but I will already have “bullock” in my semantic system, already have identified a bovine of a certain type. It will have got all the way to meaning without reference to sound.

If I wanted particularly to use the sound of animals I saw (perhaps playing a game to pass time on a car journey and shouting the sound of every kind of creature we saw) I would powerfully activate the ‘lowing’ reference, all the way to conscious, or user, level, otherwise it would be rather subliminally activated and detectable perhaps only with brain scan or myelograph (as subvocalisation is recognised even under SSR). If I saw a difficult example of an animal whose exact name I couldn’t immediately recall directly (a gibbon, say) I might have to get as far as the chattering howl they make before I could reach the name-of-animal entry, I might have to use the sound-of-animal entry to find the name-of-animal entry, but this would be an unusual event as I mostly recognise animals I see directly, which means visually.

In other words I identified the bullock visually all the way to meaning using but one route. Immediately behind this, though, I also activated the sound bovines make, at least to some degree. (I will also incidentally have activated a lot of other bovine-related stuff, smell, behaviours, marketing potential etc, but probably only subliminally.) Thus it is with reading, it seems to me perfectly non-controversially. I don’t think we have any problem here.

A visual signal (letter, letter pattern, morpheme, word – probably variable units according to circumstances) is seen, analysed inevitably by its visual characteristics (it is a silent signal) until recognised as a representation of a visual symbol and this representation used to match up with a semantic representation (and reach meaning). Direct, cheap, simple and likely to be foolproof. However, because the brain so feverishly, mandatorily and comprehensively associates everything with every aspect of itself, as soon as any representation of anything appears in the mind it is associated with all other representations of itself in other modes, a very important one being, in the case of word identification, the association of a visual representation with appropriate sounds. This association, however, inevitably demands mental effort. It requires a mental translation from one mode of representation (visual symbol stuff) into another (sound stuff) so it inescapably takes a fragment of time (many milliseconds) so it will be many milliseconds behind the original task of visual recognition, the cascade of representations in mode one (symbolic stuff). While the visual representation is proceeding to its conclusion in a semantic match the translation into sound representation is following behind. This all seems to me to be a logical inevitability but also not the least dangerous or controversial.

At least in the case of fairly easily visually recognised symbol stuff which needs no assistance from its sounds to enable recognition, then, the semantic store (meaning) will inevitably be reached by the visual stuff many milliseconds before it is reached by the sound stuff bringing up the rear. This is not to say that the arrival of the sound stuff has no role, merely that it didn’t get there as fast as the visual stuff did so is, by definition almost, likely to be used as a source of good information as to what is taking place in the world but a secondary source. It will enormously enrich my appreciation of the world, such as even to enable higher level activities like comprehension and aesthetic appreciation, but it will not have performed the low level identification trick.

So “independent” is a little strong as although, in circumstances where text is easily read, sound is not strictly necessary it will arrive and may well, almost certainly does, have a role (and I have suggested at least two). All the dual route theory says is that text is almost, but not quite, simultaneously visually and phonologically analysed but that the sound-based analysis usually arrives at “meaning” second and so usually has only secondary roles (secondary to “reading” which we are here taking to mean simply word recognition).

Susan, you write that the evidence for the theory that sound representation may be (among other things) used as a ‘back-up’ system comes only from a group of acquired dyslexics (in which you and I might believe) and another of children with developmental dyslexia (when you and I would be sceptical). There is, though, a lot more evidence around than this and many writers accept it rather absolutely. Taft (1991), Ellis (alone and in company), Adams (1990), all reference other research. Much of this is from a while ago but has not been seriously challenged and is regarded by people like Stanovich as valid foundation to the debate (Morton, Murrel, Gibson etc in the 70s & 80s for examples).

There was a period when people did this very basic cognitive research (often using priming techniques) but just now not much seems to be current. Perhaps most people feel the question has been settled? Nobody I read ever claims that the secondary arrival of sound means it hasn’t an important role to play, especially in the wider definition of reading (comprehension & enjoyment, for instance), merely that it is not how we primarily identify most words in text.

Glushko’s work is well discussed in Taft (1991) where he debates the dual route theory (and comes down in agreement with it). Your statement that processing of each and every word occurs in parallel is perfectly consistent with what I have been saying. The visual and the phonological analysis of every word probably do proceed in parallel, but the visual analysis must be chronologically slightly ahead (it had a headstart) and so, in usual circumstances, it will be the visual analysis which enables the initial word recognition to occur. (In parallel implies at least two routes, after all.)

The fact (and it is a fact) that phonological processes occur when reading, even if it is silent reading, does not damage the dual route theory of word identification. (I hope my stance on this will be clear from the above.) And the duality of the dual route means we need phonic ability and instruction, just not exclusive phonics. We need to learn and appreciate visual attack too.

Then there are two statements you make which seem to me to be unlikely. Maybe I have read you wrong? One is that deaf people over 9 cannot learn to spell correctly and the other is that our sight word memory is limited to 2000 items. I only have McGuiness & McGuiness on the shelf, does she repeat this there? Neither of these conclusions seems to me to be probable.

Likewise, I do not believe that there is good research showing that we do not read words as wholes – did you ever get that email, for example (which did not come from Cmabirdge Uvinretsiy as claimed!) but wchih sohwed taht we can raed wrods in wchih the ltetres are all jubemld up rahter ealisy, so lnog as the fsirt and lsat leretts are in plcae. We cannot read all words as a whole, of course, but we do read smaller, more common words as wholes. We read a variety of ‘units’, but almost never single letters I think.

I am absolutely satisfied that we do not read every letter when we read. We do fixate almost every word, but we do not examine all of every word to letter level. The unit we ‘read’ varies. Often, probably, it is morpheme or letter pattern. Sometimes, though, it is a whole word. The research I have read (and I have read a lot) also does not clearly show that we read by sound – as I hope I have described, we do not “eschew” sound, indeed we probably use it well, but we do not mainly identify words phonologically either. What I have read, in the main accepts that we mostly read (identify words) visually. We do lots of other things immediately following word identification, and we may enthusiastically use sound as a lead skill in some of these, but I believe we identify words visually from the reading I have done.

Ellis, Rayner & Pollatsek, Underwood and Batt, Stanovich, Pumfrey & Reason, Taft etc are the sorts of place to look for this research.

Finally, you write that struggling readers rely mostly on visual information to decode words, and that this is the cause of their difficulty. Here I am on home ground. Almost every struggling reader I have ever come across deploys just one word identification strategy and this is to sound out. They have an absolute belief that text is sound written down and only sound written down. They clearly do not use visual attack. They do not even understand that they could. They improve dramatically when induced to do this, but they absolutely do not do it unless so induced!
www.hugokerr.info

User avatar
Susan Godsland
Administrator
Posts: 4973
Joined: Thu Oct 30, 2003 11:10 pm
Location: Exeter UK
Contact:

Post by Susan Godsland » Mon Sep 29, 2008 11:25 am

Hugo,
Goodness, I feel overwhelmed with words when I see your posting, but thank you for your extensive comments anyway!

I only recently 'discovered' and have since been exploring the theory –triggered by your book, so I haven’t sorted everything out in my mind yet.
I’ve lots more reading to do – the Ellis book 'Reading Writing and Dyslexia' which I have out of the library, Snowling and Hulme’s 'The Science of Reading' which has a chapter on the subject…. but, from what I’ve read so far on both sides, I’m not happy with the Dual Route theory however lovely it sounds. As you say in your own book, ‘sloppy science never did anyone any good’ –I agree with knobs on!

chew8
Posts: 4152
Joined: Sun Nov 02, 2003 6:26 pm

Post by chew8 » Tue Sep 30, 2008 9:11 am

I don't have any great problem with the dual route theory if it's simply an account of what seems to be happening when people can already read quite fluently - i.e. they seem to take in familiar words as wholes and to resort to phonic decoding only when they come to unfamiliar words. I do have a problem, though, if it's interpreted as meaning that young children should be taught not just phonic decoding but also whole words as 'sight' words so that they look as if they are using both of the 'dual routes' from the start.

If children are real beginners, then surely all written words are unfamiliar to them, and they should therefore be using phonic decoding for 100% of words. As they make some progress, the percentage of words which they can read apparently without sounding out gradually increases (any phonics teacher knows this), but it will be some time before the balance is even 50-50, let alone before almost all words are read instantly and only an occasional one needs to be sounded out. The evidence that we have so far is that the best results are achieved in schools where beginners are taught to rely heavily on sounding out. They don't remain stuck in this mode for very long - they soon start looking as though they are 'dual route' readers, but this has happened naturally rather than because they have been taught 'sight' words as global wholes.

You say, Hugo, that all the adult strugglers you have dealt with know only the sounding-out strategy. I am not disputing that this is what you have seen, but it is not what I have seen in working voluntarily with Year 3 children (aged 7-8) over the past 8 years, and it's not what most RRFers who work with children have seen. Our experience is of strugglers who cannot sound out when they get stuck on a word. This is not surprising in view of the multi-cueing approach which was official until the summer of 2007 - there was quite a lot of emphasis on phonics-for-spelling (starting with a whole spoken word, segmenting it into phonemes, and representing the phonemes by letters) but there was little emphasis on sounding out for reading. The new government guidance should mean that things are now changing, but it will be a while before the effects work their way up the system.

If the situation has been that most younger strugglers can't sound out and most adult strugglers can't do anything except sound out, one possible explanation is that strugglers are learning to sound out somewhere along the way between childhood and adulthood, but not in time for it to become automatic by the time they reach you, Hugo.

Early in this thread, I wrote 'As far as early reading instruction is concerned, I don't think it matters what researchers think is happening in the minds of proficient readers as long as they don't say that because this is what proficient readers do it's what beginners ought to do'. I stick with this view - I believe that the best way to ensure that children don't become adults who are still stuck in laborious sounding-out mode is to emphasise sounding out right at the beginning when it's needed for all written words because all these words are unfamiliar. Many words then become familiar by this route and children are able to read them without overt sounding out, but they resort automatically to sounding out when words are unfamiliar. That, I believe, is the best way of ensuring that they become adult readers who look as if they are illustrating the 'dual route' theory.

Jenny C.

kenm
Posts: 1495
Joined: Sat Dec 17, 2005 5:19 pm
Location: Berkshire

Post by kenm » Tue Sep 30, 2008 11:55 am

Classroom observation reported on this site confirms that both routes to reading occur. I agree with Jenny C. that what matters is how reading is taught initially, and while I have no experience relevant to this question, I find her recommendation to teach the aural route and let the visual route develop naturally (if, indeed, the latter is what happens) entirely consistent with what I have read about our evolutionary history.

The brain has general mechanisms for recognising both visual and aural signals. It also has specialist mechanisms for recognising patterns of importance, such as faces and facial expressions. These include memory stores of some sort (probably implicit in the connections between neurons). The important point to be considered in comparing the two routes to word recognition is the nature of the word memory store with which they terminate.

Some promoters of guessing methods have described reading as a natural capability, comparable to the acquisition of spoken language. This contradicts what we know of the last c. 40000 years of the history of the human race. This being a rough estimate of the period during which spoken language has been important. During that time, speech has been an important means of ensuring cooperation in collective activities, such as hunting, warfare and teaching, the latter being important in passing on knowledge and skills to the young. In contrast, written language started only 6000 years ago, and became widespread only within the last 2000. The majority of the world population became literate some time within the last 200 years.

Inherent brain mechanisms, like other physical characteristics, are developed by trial and the ruthless weeding of evolution. The implication of the previous paragraph is that those of our ancestors who could talk and understand speech outnumber those who could read by many orders of magnitude, and their contribution to our genes by a factor of about 160. It is unsurprising to find evidence of a specialist language mechanism that could have developed since lack of language became an impediment to reaching maturity and producing offspring; it would be amazing if a specialist reading mechanism or written word memory store could have developed in the brief time and number of generations since literacy began to have a comparable influence.

2000 words seems to be typical limit to a reader's visual memory store for words rather than a maximum. ISTR reading here about a young woman who acquired a reading vocabulary of 10000 words by pure visual recognition and found that that inhibited her from post-graduate work in philosophy, so she transferred to mathematics, but she was uncommonly talented.
"... the innovator has as enemies all those who have done well under the old regime, and only lukewarm allies among those who may do well under the new." Niccolo Macchiavelli, "The Prince", Chapter 6

Post Reply

Who is online

Users browsing this forum: No registered users and 3 guests