Monday, October 22, 2012

One spelling, different sounds

I wrote this post and then thought that it would be easier to listen to than to read, so here's the link to the audio file.

As I was asked to post the transcript, here it is below:

In this audio file I’m going to talk about how spellings in English can represent different sounds.
In written English spellings can represent different sounds and this is something that people can find troublesome to teach.
Fluent readers rarely even notice this feature, until, that is, they come up against a word they’ve never seen before, visit a place whose name they don’t recognise, or come across someone’s name they are unsure how to pronounce. However, teachers and parents notice it all the time and, if the forums on MumsNet or the TES are to be believed, are utterly baffled about how to go about teaching it.
Amusingly, this particularity of the language could be witnessed at first hand  yesterday on the ‘Today’ programme on Radio 4, when Evan Davis was introducing a guest by the name of Niall Crowley. Davis wanted to call him /n/ /ie/ /l/   /k/ /r/ /ow/ /l/ /ee/; whereas, in fact, his name is /n/ e/e/ /l/  /k/ /r/ /oe/ /l/ /ee/.
Another example of this is the place name ‘Broughton’. Broughton is a place in Northamptonshire and is pronounced /b/ /r/ /ow/ /t/ /ǝ/ /n/. In Milton Keynes, however, there is also a place of the same spelling that is pronounced /b/ /r/ /or/ /t/ /ǝ/ /n/. So, the spelling ough can be /ow/ and it can also be /or/.
The question is how would you know this? To which the answer is that you wouldn’t unless someone in the know told you. Now, here’s the thing: this is only likely to happen in the case of fluent readers with examples such as the ones I’ve given above. In virtually all other cases where this happens in everyday reading, as I’ve already indicated, fluent readers don’t even notice this aspect of the language. For example, when reading the sentence ‘Last night I ate a tasty steak.’ a fluent reader would have no problem with the spelling in the word ‘steak’. Of course, in this word, the ea spelling represents the sound /ae/. In other words, though, it can be the sound /ee/ (‘seat’) or /e/ (‘head’). So, how do we know that in ‘steak’ it’s /ae/. Well, firstly, as fluent readers, we process all the information in the word so fast (in milliseconds) that our brain isn’t aware of what we’re doing above the level of consciousness. We are also simultaneously processing meaning: ‘Last night I ate a tasty ‘stek’ or a ‘steek’ just doesn’t make sense and our brains ‘know’ this before we ‘know’ it (are consciously aware of it). There are also orthographical patterns in the language that are so common (and insistent!), they ‘demand’ a certain response. For example, the spelling a most usually represents the sound /a/ but when we see it positioned in front of the spelling ll for /l/, we immediately read it as /or/ (‘ball’, ‘tall’, etc.).
The problem for beginning readers and for people who are having trouble learning to read is that they are rarely explicitly taught that, in most cases, a spelling can represent/stand for/be more than one sound. Why this should be, goodness only knows because this is by no means a difficult concept to grasp. We live in a world of symbolic tools: signs, symbols, graphs, musical notation, etc. Spellings are no different. Even quite young children have no difficulty in understanding that a circle can represent different things: a ball, a pizza, a moon. Exactly the same is the case for, say, the spelling o: it can be /o/ in ‘hot’, and it can also be /oe/ in ‘no’.
Neither are many beginning readers or poor readers given the skills and the systematic code knowledge to enable them to engage successfully with this aspect of the code. If they read the name ‘Gemma’ as /g/ /e/ /m/ /a/, it won’t be recognisable as a name they are familiar with and, in this case, they should be being taught to ask themselves what the potential problem is. If we teach them that the spelling g can be /g/ but in some words it can also be /j/, then /j/ /e //m/ /a/, which in turn becomes normalised as /g/ /e//m/ /ǝ/, is likely to make sense.
Good quality phonics teaching does NOT involve teaching code knowledge alone. It also requires that learners understand how the code works – how the spelling system relates to the sounds of the language – and that they are given the skills necessary for them to use the knowledge they have.
Of course, as always, to teach children these elements of conceptual understanding, of code knowledge and the skills needed, teachers themselves need to know exactly how the system of spellings and sounds works, or there will be huge potential for confusing children.


Anonymous said...

I doubt if many teachers challenge the idea that all children need to learn the sounds represented by the graphemes in our complex orthography but I do believe that the question of how this knowledge is learned is still a matter for legitimate debate. In the swinging sixties when phonics teaching virtually vanished from the curriculum, about eighty percent of the population still managed to become literate individuals including people like JK Rowling, Laura Albert, Stephen Fry, Paul Adam, Simon Armitage and many others.

My own five children were brought up on the Ladybird series Peter and Jane books and are now all high achieving professionals. One of my sons read the Catch 22 tome at the age of eleven which I suppose must constitute proof positive of his complete mastery of the grapheme/phoneme correspondences without having had the benefit of any ritual phonics teaching.

The tragedy of the ‘whole word’ fashion was that in common with any ‘one-size-fits-all’ philosophy, it ignored the needs of 20% or so of the population whose brains appear to be wired in such a way that they respond more productively to perceptual rather than ritual learning. I am involved in a number of research programmes which examine the effect of perceptual learning on the acquisition, not of reading skills in isolation but holistically of all literacy skills. My work focuses exclusively on children who for whatever reason, are failing to acquire these skills in classrooms where synthetic phonics teaching is the norm and where the majority of children are responding well to this norm.

Phil Kellman’s work on Perceptual Learning at UCLA is well known as is the world-wide success in language teaching of the Rosetta Stone organisation which eschews ritual teaching entirely in favour of PL. The Charleston Academy in Inverness leads the field internationally in developing their own highly successful strategy for using PL to boost the literacy skills of those who have been failed by ritual teaching methods and have now accumulated four years statistical evidence of their success. They are currently providing consultative support to a Highland Area sponsored project to examine the potential impact of perceptual learning on younger children who are not responding well to ritual teaching methods.

I believe that reading is defined as the retrieval and assimilation of the meaning that is encoded in text and that since meaning is encoded in words and not in graphemes, the reading process involves a greater reliance on sight vocabulary than your views would suggest. A sight vocabulary of two or three thousand words would suffice to enable anyone to read fluently not only these few thousand words but all unfamiliar but similarly constructed words. Anyone who has internalised the commonly occurring word ‘found’ also has instant access to the means of decoding ‘ground, sound, bound, hound, pound, round, astound etc etc. Similarly, internalisation of ‘at’ provides access to cat, sat, mat, fat, hat and so on ad infinitum. A five thousand word sight word vocabulary is common among those who use Chinese orthography.

I don’t believe that good spellers spell well because they have learned all the spelling rules. I believe that they know when a word ‘looks right’ and this of course is a perceptual rather than a cognitive skill.

John said...

Thank you Anon for your comment.
From the start, I have to ask
what evidence you have for saying that 'eighty percent of the population still managed to become literate', and what evidence you have for maintaining that phonics 'virtually vanished from the curriculum'? There have been numerous studies, amongst which are and OECD (1997) study, which found that only 48% of adults were properly literate, and the USA’s National assessment Board’s Education Testing Service in conjunction with the National Centre for Education Statistics (1992), which found that forty-seven percent of adults were 'barely literate' and, even more surprisingly that only 3% of adults scored at Level 5 (highly competent) on the test. In fact, a vast array of independent studies in English-speaking countries has found the same thing over and over again and what they point to is that whole language approaches fail somewhere arpound fifty percent of the school population.
Without wishing to offend you, the information you provide about your own children is statistically meaningless.
I don't know what you mean by referring to people whose 'brains appear to be wired in such a way that they respond more productively to perceptual rather than ritual learning' and I don’t believe that there is any scientific evidence to support your hypothesis. You also need to concretise exactly what you mean by 'perceptual' and 'ritual learning'.
Unlike you, I don’t believe that 'reading is defined as the retrieval and assimilation of the meaning that is encoded in text'. First and foremost, reading is decoding the written symbols on the page into sounds which combine into words. Understanding what the text (the words on the page) means is comprehending. No-one can comprehend anything if they can't decode.
There is pretty good evidence to suggest that adults don't have anything like a sight-recognition memory for 3,000 words. In actual fact, linking an abstract symbol (a word, taken as a whole unit) to a meaning is extremely difficult for anyone to do. Try doing it for Russian or Greek words and see how you get on. It is called paired-associate learning and humans find it very hard indeed. In addition to that, there are over a million words in English and, if we want to give all children the opportunity to reach their potential (whatever that may be) we need to teach them to be able to decode/read all words.
On the question of analogising words, I’m sorry to say that the evidence is not at all strong. The same is true for the teaching of 'onset and rime'.
Your claim about Chinese orthography is simply not the case. All writing systems are based on the sounds of the languages they represent. There are no exceptions and all the world's experts on this subject are in agreement on this. I suggest you read Peter Daniels and Bright's The World’s Writing Systems (OUP, 1996).
Again, on spelling, what you believe or don't believe is neither here nor there. I think there is good evidence that teaching spelling 'rules' is not effective because there are too many exceptions and I've never yet met a teacher who knows all the rules, never mind the exceptions. However, spelling can be taught very successfully, as our study of more than one thousand five hundred children in Key Stage 1 showed:
I'm sure that Phil Kellman is a delightful chap and is dedicated to making the lives of young children better but I don't see in his research anything that convinces me that he knows anything about teaching young children to read and spell.

Anonymous said...

The fact that you regard ‘reading’ and decoding as one and the same means that we have such very different views of the reading process that we could never agree about anything concerning literacy. I feel there will very few people that do not consider the retrieval of meaning as being part of the reading process. I will nevertheless address some of your points.
I was at one time, a professional linguist and was employed as an interpreter for four years and feel that I can speak from a basis of some experience. All language is sounds and all writing systems are indeed based on the sounds of the language. In Chinese are represented by ideograms rather than phonograms which have to be individually internalised. Most Chinese are familiar with about 5000 of these ideograms. (My daughter-inlay teaches in China) Although I am not a Russian speaker I am familiar with the Cyrillic alphabet and can therefore vocalise Cyrillic test. By your definition of reading, I am able to ‘read’ Russian! I can assure you that I were to apply for a position of interpreter or translator of Russian on the basis that I could ‘read’ Russian because I could decode Cyrillic, the employer would definitely not be amused.
I carried out a one term project last year involving 61 non or near-non Year 2 readers in six schools. The following is one of six similar comments from the schools involved. This is a Surrey primary school which uses conventional synthetic phonics (Jolly Phonics) “The Year 2 teacher benchmarked the children who took part in the project, today. They all have made an incredible amount of progress and read benchmarked level 16 books, which are 1a books; they started off being unable to read a benchmarked level 1 text. Without a doubt, I feel that every child completing the project is now likely to become a confident reader.”

You commented “I don't know what you mean by referring to people whose 'brains appear to be wired in such a way that they respond more productively to perceptual rather than ritual learning” When a small number of Y6 children fail to become confident in reading and writing after six or seven years of conventional teaching which was successful with the vast majority of their peers in fact become confident after a one term intervention using a non-ritual strategy, I think it is reasonable to claim that these children ‘learn differently’ which suggests to me a difference in the way their brains are wired.
You asked me to define ‘ritual’ and ‘perceptual’ learning.
Ritual teaching is where the teacher has knowledge which he or she passes on to children and which they are required to remember. Perceptual Learning is exactly the opposite; it routinely immerses children with literacy skills deficits in the meaningful exercise of those skills. No ritual teaching is involved. The children have nothing to forget because they have nothing to remember. This is precisely the technique used by Rosetta Stone which routinely immerses people with (for example) with no Japanese language skills, in the meaningful exercise of Japanese language.
I am currently running three separate one-term, perceptual learning projects which are focused mainly on Year 6 children who have in their teacher’s views, very poor literacy skills. The English project involves some thirty schools who should report on or about the 18th December. The Northern Ireland project involves similar children in nine schools which is being organised by a specialist advisory teacher will also report in December. The Scottish project is being funded by the Highland Area council consequent on the highly successful work which has been carried out in the Charleston Academy in Inverness over the past four years. The Charleston is providing consultative support for the Highland Area primary schools which are taking part and these schools will report their conclusions in March 2013 on one of the popular teacher forums. Charleston Academy is producing a video of their PL work which hopefully will also become available at the end of December.

John said...

Once again, thank you for taking the trouble to reply, Anon.
Well, firstly, I didn’t say that I don't consider the retrieval of meaning as being important. I said that before one can understand (retrieve meaning), first and foremost one must be able to decode successfully. As a person who used to teach literature at several universities, I think I have an adequate appreciation of the broader interpretation of the word 'reading'. But, again, as I insist, if one cannot decode, itself a complex process, one cannot possibly understand or give a 'reading' to the text one is working with.
You say that you can vocalise Cyrillic text. To me, that means you are able to read it. However, from what I gather, you are not able to understand (presumably) all of it and therefore comprehension is impeded by your lack of understanding or limited understanding of the meaning of the words once you've read them. Of course, there are other factors that impinge, such a grammar, discourse and so on. There are also well-known phrases or sayings (to an L1 speaker), which, when translated literally by a non-L1, make no sense. Still, one has to be able to decode in the first place
Your engagement with children learning through the medium of a synthetic phonics project sounds great. Surely, you are supporting the proposition that non-readers can learn to read successfully through phonics instruction. Or am I missing something here? Perhaps their brains have been re-wired. I’m not taking the mickey here either. We do know and recognise much more than previously that there is a degree of plasticity in how the brain can learn to perform new functions. Doidge et al are very interesting on this.
I think that when a certain proportion of children fail to read by the age of 10-11 years, there might be a variety of explanations, ranging from all of them possessing a large index finger to a very poor level of instruction in the early years. My money's on the latter, by the way.
'Ritual' teaching, then, I take to mean the kinds of traditional approaches which portray the child as a passive recipient of pre-packaged knowledge. I can tell you that both theoretically and empirically, it is not a model I feel has anything to do with what I see as a successful phonics programme.
Your claim that children, or anyone for that matter, can learn something without having to remember anything seems to me completely implausible.
I'm sure we all look forward to the report. I hope it's as transparent as the one we have published.