[Journal of the Simplified Spelling Society, J28, 2000/2, pp3-11]
[See items by and about John Downing in Journals, Anthology, SPB and about ita.]

The Transfer of Skill in Language Functions.

John Downing.

We here present a re-edited version of John Downing's talk to the SSS on 6 September 1986. Then Professor of Educational Foundations in Reading at the U. of Victoria, British Columbia, in the 1960s he had directed the research into the effectiveness of the Initial Teaching Alphabet at the London Institute of Education. His copious writing on reading psychology culminated in Psychology of Reading (Downing & Che Kan Leong, 1982). Succeeding Sir James Pitman in 1972, he was President of the SSS until his death in 1987. The first version of this paper was published in JSSS 1987/2 J5 (pp5-12); for a tribute to John Downing's life and work see JSSS 1987/3, J6 p6.

0. Abstract.

§1 outlines the phases of skill acquisition, whether applied to literacy or any other skill.
§2 describes how easily literacy in one script is transferred to a different script.
§3 gives statistics for the improved literacy of children taught their first literacy skills in the regularly spelt Initial Teaching Alphabet (i.t.a.) compared to the irregularity of traditional English orthography (TO).
§4 shows how the same effect of transfer operates between different languages, notably the languages of New Guinea, where, however, the involvement of English seriously disrupts otherwise straightforward processes of literacy acquisition.
§5 sketches the different modalities of transfer in Russian, Chinese, Japanese.
§6 demonstrates the advantages of learning that proceeds from what is already known; literacy is thus best acquired first of all in the mother tongue, and using a regular orthography, which trains the mind in accurate, logical observation.
§7 discusses the risks of negative transfer arising from poorly thought-out learning sequences, and emphasizes the need for experiment to test innovations.
§8 Conclusion
§9 A range of further issues is touched on as discussion points at the end.

1. The acquisition of skills, including literacy.

My subject is the psychological question of transfer from one spelling system to another, and I have entitled my talk 'The Transfer of Skill in Language Functions'. I shall discuss it under two main headings, firstly the psychology of skill, and secondly, how to maximize positive transfer of skill. In Psychology of Reading we say a lot about skill, some of it relevant to orthography.

There are four main language skills, which overlap with each other: speaking, listening, reading and writing. (Other language skills are perhaps better regarded as subskills.) The four don't necessarily come in that particular order, but speaking and listening are early skills, and reading and writing are later skills.

A short definition of a skill, based on generally held views on skills in psychology and covering all the main aspects, is as follows: a skill is a complex integration of habits (in psychology we use the word habit not in its everyday meaning of bad or good habits, but meaning a small piece of learning), including cognitive, attitudinal, evaluative and motor behaviours. The most important word in the definition is integration: psychologically it is the most important feature of a skill, which is made up of numerous subskills, and the performance of the skill involves the integration of all of them by the brain.

We need to recognize the different kinds of behaviour involved in a skill. There is obviously the cognitive aspect, which means the acquisition and use of knowledge. Less obvious, though extremely important, is the attitudinal aspect, which we recognize perhaps more for instance in sporting skills: in playing soccer you need not only to know how to play, but also to have the right kinds of attitude towards the sport; and in reading and writing too you need the right kinds of attitude towards those skills in order to perform well. The evaluative aspect means you have to be able to evaluate your own performance, to know whether you're performing well or not. And then there are motor behaviors, which are obvious in some skills like playing soccer or driving a car, but which are also important in all the language skills I have mentioned: in speaking, perhaps rather less in listening, but certainly in reading and writing. Eye movements, for instance, are extremely important motor behaviors in reading.

This kind of definition relates to any skill. For instance Whiting at the Free University of Amsterdam would use the same definition in his area, which is physical skills like sports and physical education. For our book Psychology of Reading we first of all considered what psychologists have written about skills in general, and made a list of about 20 features. Then we examined the behavior that goes on in reading to see whether it has those features - and it does. We were able to say, if reading is a skill, we should be able to take all the knowledge and research results we have in psychology concerning skills and apply them to the teaching of reading and writing. That's quite an important part of the book: we were able to apply a lot of research in psychology which had not yet been applied to the teaching of reading and writing.

Something that was discovered a few years ago in experimental psychology is that any kind of skill passes through three learning phases. The cognitive phase is where learners figure out what they are supposed to do, how they are to understand the task - whether learning to jump, or learning to read or drive a car. The second phase is then to practise the skill until it can be performed perfectly. And the final phase in achieving highly skilled performance is becoming automatic, so you don't have to think about it any more, except when something unusual occurs.

Of course a skill like reading or flying a plane is not just three phases. Those phases occur over and over again, because a complex skill has many parts, which are called subskills, and you have to go through these phases every time you face up to a new subskill. Also there is no neat frontier or boundary between the phases, you can't just draw a line between phases in terms of actual human behavior, they blend into each other. In other words, a phase is an abstraction. So in the cognitive phase you may think you've figured out what to do, and start practising away, and then you discover you've not got quite the right idea, and you have to modify your idea of what to do.

Now from our point of view in the Simplified Spelling Society another very important thing is that we know from research that once a skill has become automatic, it's never lost - unless there is brain damage or something like that. Once normal humans have acquired the skill at the automaticity level, it's permanent. We can confirm that from our own experience: I became a skilled bicycle rider as a child and went on bicycling for many years as an adult, but I stopped doing so long before I emigrated to Canada in the late 1960s; but after I'd been there about 10 years, I bought a bicycle, and just got on and rode. I never had to think about the skill because it was so automatic that I hadn't forgotten anything. Similarly, I shall be working in France now for six months, and all my French language behavior will come back, although I haven't used it for some time. There's no problem once a skill is automatic.

Another psychologically important feature of skill is that, once it is automatic, it readily transfers from one particular subject or situation to another. Consider the skill of driving a car: I travel a lot, and very often I land at an airport and rent a car, but I don't know what sort of car I'm going to get; in that situation, all you have to do is find out where the controls are, which is sometimes a little difficult, but once you've located them, it doesn't take very long before your automaticity returns. That's very important for us because of what it tells us about the timing of transition. If you're going to simplify English spelling, there must be automaticity if you're going to get ready transfer from one spelling to another.

2. Transfer of reading skills.

We found the same in the research into children's use of the Initial Teaching Alphabet (i.t.a.). Fergus McBride did a special study in Edinburgh of the transition from i.t.a. to traditional orthography (TO) and found that transition only took place with ease if the child was already reading very fluently in the i.t.a. The details are given in my book Evaluating the Initial Teaching Alphabet (Downing, 1967), which contains a lot of information relevant to the Simplified Spelling Society's work. If you look up Fergus McBride in the index, you'll find out what page the study is on. Unfortunately the book is difficult to find, but there will be some in libraries.

If we look at the first folio edition of Hamlet (1623), we don't have much difficulty reading it, though of course we are helped if we know the text well. It uses an older spelling of English, and if you have no experience of reading it, what slows you down at first is the S that looks like an f, and there is the confusion of U and V. When this was printed there wasn't consistent spelling, and there was a lot of freedom.

Or consider the test which Sir James Pitman made up, where you have to read words written with the spelling reversed and each letter in its mirror-image form. Once you know there's one rule underlying the text, you can read it. Most people can read words written backwards or upside-down.

The following appeared in Punch (Herbert, 1920):


To-day I am MAKing aN inno6£vation. as you mayalready have gassed, I am typlng this article myself Zz1/2 Instead of writing it, The idea is to save time and exvBKpense, also to demonstyap demonBTrike= =damn, to demonstratO that I can type /ust as well as any blessedgirl if I give my mInd to iT'" Typlng while you compose is realy extraoraordinarrily easy, though composing whilr you typE is more difficult. I rather think my typing style is going to be different froM my u6sual style, but Idaresay noone will mind that much. Iooking back i see that we made rather a hash of that awfuul wurd extraorordinnaryk? in the middle of a woRd like thaton N-e gets quite lost? 2hy do I keep putting questionmarks instead of fulstopSI wonder. Now you see i have put a fulllstop instead Of a question mark it nevvvver reins but it loours.
These examples show that despite the strange orthography none of us has difficulty in transferring because we've reached the automatic level of skill which allows ready transfer.

3 Statistics on i.t.a. - TO transfer

Between 1961-67 we examined how well children transferred from i.t.a. to TO (see chart of i.t.a. symbols). The table below compares children who had started learning to read with i.t.a. and children who had started in the usual way with TO. It gives the proportion of 'failures' in post-transition tests - 'failures' in quotes because we didn't really think of the children as being failures; we meant just the number who fell below a certain score in the test.


Schonell GWRT
Neale Accuracy

Neale Rate

Neale Comprehension

NFER Sentence
NFER Comprehension
after 26 mth
after 19 mth
after 34 mth
after 19 mth
after 34 mth
after 19 mth
after 34 mth
after 34 mth
after 54 mth
after 60 mth
i.t.a. %
TO %

All the tests were printed in TO, whether the children were in the i.t.a. group or the TO group. The TO children had learnt to read entirely in TO, whereas the i.t.a. children had usually learnt for 18 months-2 years in i.t.a. (the average was 18 months before the changeover). Some of the i.t.a. children tested may not have transferred, but probably the majority would have done so when the tests were given.

On the Schonell graded-word reading test (GWRT) after 26 months of schooling, there isn't a significant difference between the i.t.a. and TO children, nor is there with the Neale accuracy test after 19 months. However, after nearly 3 years, there is a big difference on the same test. There are only about half as many i.t.a. children failing in TO as TO children failing in TO, and the other results show the same thing. When you look at the very late tests, at the bottom of the table, after about 4½-5 years there is a very big difference in favor of the i.t.a. group.

It's clear that the i.t.a. children were much more likely to achieve automaticity in the reading skill and therefore to transfer readily. The TO group on the other hand hadn't reached the same level of automaticity, though in their case perhaps we shouldn't talk about 'transfer' at all, as the children were performing a task they had been taught from the start. The point is, the TO children had less reading skill to apply to TO reading than the i.t.a. children had, although the i.t.a. children were at a disadvantage in facing a less familiar script. If you had reversed the test and tested all the children in i.t.a., the superiority of the i.t.a. children would have been even more marked. Figures showing what happened a few weeks before the transfer reveal a big difference in favor of the i.t.a. group. The transfer was definitely a setback for the i.t.a. children in the first test after they changed over: they weren't reading as well in TO as they had read in i.t.a., but they were reading just as well as the TO children. A few months later they had recovered from that setback and were reading better than the TO children. The test on reading speed, we found, was not very valid, nor were the differences in number of words read per minute very significant. The comprehension test on the other hand showed slightly better results after 19 months for the i.t.a. group (only 50.8% 'failures', compared with 56% failures for the TO children), but the next test, after 34 months, shows a very significant advantage for the i.t.a. group with only 14% failures compared with nearly 25% in the TO group.

We also tested spelling, and the results are interesting here too.


Schonell Spelling

Capital Letters
Sentence completion
after 30 mth
after 42 mth

after 5 yrs
after 5 yrs
after 5 yrs
after 5 yrs
after 5 yrs
after 5 yrs
i.t.a. %

TO %



The Schonell spelling tests after 30 months show fewer failures among the i.t.a. children, although they suffered several disadvantages: the spelling test was in TO, not in i.t.a.; no marks were given for the correct i.t.a. spelling; and no allowance was made for the fact that some of the i.t.a. children had not yet transferred to TO. For all that, the difference in achievement is not significant statistically, despite first appearances. But the results of the tests given after 42 months show a very significant difference between the groups in favor of i.t.a.

Another test, the NFER (National Foundation for Educational Research) test, comprised several different subtests, and after 5 years on every one of them the i.t.a. group was significantly superior to the TO group even with capital letters, which in i.t.a. are just majuscule versions of the small letters. The performance of the i.t.a. children was superior with regard to tenses, sentence completion, and abbreviations. So there is a lot of spillover from starting out with something that works well and fits in with the child's level of ability.

4. Skill-transfer between languages.

This same phenomenon of ready transfer from one situation to another of skills that have been learnt only once and have become automatic is seen not only between i.t.a. and TO, but across languages. We can read the sentence I say that you are right, the father works while the son does nothing and we don't have to re-learn to read it in another language. To understand je dis que vous avez raison, le pére travaille tandis que le fils ne fait rien, you have got to learn French, but you don't have to learn to read all over again. That skill is transferred from one language to another. This applies even to non-alphabetic languages like Japanese and Chinese. If you analyze what you're doing psychologically when you read English, and what you're doing psychologically when you read French, the skills that transfer are all the skills you already have in English. There are certain things in French that are quite different, and those will slow you down for a time, but the skill itself of reading will transfer from one language to another.

A few months ago I went out to New Guinea for a second fieldwork study comparing children in different environments. We were trying to find out more about how children think about language. We compared people living in a village where the people had never seen any writing, with children living in villages where they didn't have writing in their own village but saw writing (not in their own language but in other languages) when they went to market, and then with children living in villages where they had schools and churches, and where they learned to read and write at school. There is also a local pidgin called Tok Pisin, which is a lingua franca; it has an Australian flavor and a German admixture, but is influenced by the 750 languages of the country (not counting dialects).

I was in a village whose language is Hako, and where an experiment had been in progress for three or four years. All of the children there are now taught to read and write in their mother tongue, which linguists had analyzed years ago and for which prayer books and the like had been produced. The children have their first lessons in their own language two years before being introduced to English, after which everything in the school is in English. But there is also the other language, Tok Pisin, in this village, which is never taught in the school, although its written form is seen everywhere, in the prayer books, on the health posters, on the police posters, and so on. In fact it's not taught anywhere in Papua New Guinea - it's just assumed that if you have learnt to read in one language, you can read Tok Pisin too (although there was a movement in the Lutheran missions a few years ago to teach Tok Pisin, and the idea still being thought about). A pamphlet in Tok Pisin which I picked up from a religious bookshop in the capital of the province where I was doing the research has the following sentence in it, particularly striking on account of the rude word in it: Strongpela Dring i bagarapim yu. It is taken for granted everyone can read that if they have learnt to read in another language.

We tested children for reading in three languages, their mother tongue, English and Tok Pisin, and what we found was this: if the children had begun to learn to read in the mother tongue, they had excellent transfer from mother tongue to both Tok Pisin and English; but if they'd started out in a strange language, English, the transfer was very poor. This was so marked that both in Tok Pisin and in English their scores were often near zero. And it wasn't only the scores: we also observed the children subjectively, and we saw the children who had started out in a foreign language were afraid to try in English, although they had been taught it. They were afraid to try any word that they hadn't been taught, any word that wasn't in their textbooks. And of course they transferred that negative attitude from the English experience into both Tok Pisin and their own mother tongue from their village.

A very important point that has to be taken into account here is that the English spelling is bound to cause them to feel uncertain, because of the uncertainty of what the letters represent. Both in the mother tongue and in Tok Pisin there is a very straightforward relationship between the letters and the sounds. Usually the people who have made up the spellings for the languages - half of them have been written now - have tried to ensure a good connection between Tok Pisin and the mother tongue.

5. Transfer in Russia, China, Japan.

Another example of this transfer is from the Soviet Union. I've just finished the manuscript of a new book, to be published by the North Holland Company in Amsterdam, and called Cognitive Psychology and Reading in the USSR (Downing, 1988). It's a survey of all the major works in Russian on the teaching of literacy, and one of the things I learnt from making that study of what goes on in the Soviet Union in reading is the surprising fact that in the Soviet Union from 1917 the people of that vast area have had the right to learn to read and write in their mother tongue rather than Russian, as they had to before the Revolution. And so it's quite a normal thing for people in the Soviet Union to learn to read and write in one language and then switch over to Russian later because all the higher levels of education and the materials need to be in the more common language. So the Russians have had a lot of experience with transfer. Lenin decreed that it was the duty of every citizen to learn to read and write, and in his decree he said they could learn in the mother tongue or Russian, as preferred, and that has been the policy ever since. Perhaps it was because Lenin was a schoolteacher (as was his wife Krupskaya too) and had very good intuition about how children learn to read and write; the decision would seem to have been a correct one.

There are also examples of transition, transfer of training, transfer of skill across orthographies. In China nowadays children begin with the roman alphabet, learning the phonemic Pinyin spelling of Chinese, and then switch to the classical characters, which though they've been reformed are quite unlike the roman alphabet. I don't know whether Pinyin will finally prove to have been a success: when I was in China in 1982 there was some research being done by a Chinese professor in Shanghai, and it wasn't very favorable to Pinyin. But politically it had strong support: both Chairman Mao, whose influence is still there, and Chou En-lai believed that a phonemic spelling for Chinese would result in a universal pronunciation for Chinese across China. I don't know if they were right, but that was their hope. The difficulty is that Pinyin cannot take account of all the dialects. In the school in Shanghai I observed the teachers trying to get the children to pronounce Pinyin according to the Pekinese dialect, a lot of which is meaningless to the children. The task isn't being tackled properly, and I don't know whether it can be. The research being done in Shanghai by this professor certainly did call into question the value of Pinyin.

Japan has a similar situation. Most Japanese books are printed in a combination of two types of characters, Kana and Kanji. Each Kana character represents a syllable, and the Kanji represent morphemes or words. Young children start out with everything printed in syllabic characters, in Kana, and then gradually they switch over more and more, and the Kana characters are dropped in favor of the Kanji characters, which came from Chinese originally. In Japan this works very well because as far as I know there isn't the problem with dialects. But the kana syllabary has a perfect one-to-one relationship between written syllables and spoken syllables, at least in standard Japanese. There is however a problem with homophones, which the syllabary cannot adequately distinguish. There are a lot of Japanese jokes about homophones, especially involving misunderstood telegrams, these being written in the syllabary and not in Kanji.

6. Maximizing the transfer of skill.

Our interest in the Simplified Spelling Society of course is in how to maximize the positive transfer of skill. What can we do to make sure that we get the most positive transfer between the reformed spelling and TO? Here the concept of readiness is important. Although the term reading-readiness has been subjected to quite a lot of criticism among teachers in the last few years, it is rather meaningless criticism, because you can't criticize something that's psychologically real. We know that, regardless of what the skill is, whether it's reading or playing tennis, you can't acquire it unless you're ready to learn it. You have to have the physical maturity to manipulate the tennis racket, and you have to have learnt something about the game of tennis which you can then use in the new learning that you're going to try to acquire. So when we think about how to maximize positive transfer of skill, we can use that concept of readiness.

Let us apply this to the cognitive phase, that beginning phase where you'e trying to figure out the nature of the task. The i.t.a. experiment provides some examples, and the book Evaluating the Initial Teaching Alphabet contains quite a lot of material it would be useful to look at, if you are concerned to maximize positive transfer of skill from one spelling to another. Before we look at a helpful example, let us remember that Professor Magdalene Vernon in 1957 was the first psychologist to put forward the idea that the main cause of reading disability is cognitive confusion, by which she meant that the child didn't understand the task to be performed. For instance, the child may not understand why road is spelt R-O-A-D, and begins to doubt its ability to grasp the system rationally: though the teacher says R-O-A-D is right, it doesn't seem right. The advantage of i.t.a. was that it cut down this cognitive confusion: if you spell said S‑E-D, it isn't confusing any more. The TO children learning S-A-I-D are much more likely to be confused cognitively by the spelling form in question, but the i.t.a. children had no problem with S-E-D.

Following up Professor Vernon's theory, I wrote in Evaluating the Initial Teaching Alphabet (pp233-34):

The reduction in 'general cognitive confusion' may also have been effective in two experiments conducted by Sister John... In her first experiment Sister John's subjects were an experimental group of twenty five-and-a-half-year-olds who had been learning to read i.t.a. and a control group of twenty five-and-a-half-year-olds who had been learning with TO. Specially devised nonsense-symbols similar to characters of the Greek alphabet were used for two tasks: matching and recognition of the symbols. The experimental i.t.a. group achieved superior scores for both tasks.
Again you see transfer at work with the strange Greek-like characters.

On recognition the i.t.a. group's scores were significantly better...

In her second experiment Sister John used new subjects in the same schools. Two groups (one i.t.a., one TO) of twenty-five four-and-a-half-year-old pupils were studied during their first six months at school. The same nonsense symbols were used, but in this experiment the testing procedure was carried out twice: (i) soon after the subjects entered the school for the first time, and (ii) six months later. On the initial test the two groups were comparable,
(there wasn't any difference because they hadn't learnt anything yet)

but six months later the same significant difference had emerged as had been discovered in the earlier study, i.e. the children who had been learning i.t.a. for six months were better at recognizing unrelated nonsense-symbols than were children who had spent the same period learning TO.
In other words, the i.t.a. children had learnt something about the task of recognizing symbols which they could transfer over into something completely strange like these semi-Greek symbols.

Sister John's results indicate that in learning to read i.t.a. her subjects had learned appropriate orienting responses, so that their attention was directed towards the general features of the task of discriminating and remembering abstract symbols... which were then relevant in the new tasks with the nonsense-symbols. Conversely, the greater irregularity of symbol-sound relations in TO seems to have inhibited the development of appropriate orienting responses to the task of remembering differences between graphic symbols. It seems likely that this comparative failure of the TO group represents a greater degree of 'cognitive confusion' surrounding the task of recognizing graphic symbols.
Similarly in the Papua New Guinea work, we found that if you learn to read in the mother tongue first, you get better transfer to other languages than if you start with an unfamiliar language. So it's important to start off where the child is. What is the child bringing to the situation? And in the case of learning to read, it is clearly better to start, as all the research showed, with the mother tongue, because that's what children have got. If they have thought at all about their speech and other people's speech, it will be in terms of their own language, the sound of their language, not of course the sound of some strange language which they have hardly ever met.

The same principle applies to the letters of the alphabet. In the case of New Spelling (Ripman & Archer, 1948) and what we did with i.t.a.: the task was taken closer to the child's capability by the age of 4, 5 or 6 by having a simple relationship between the writing and the speech. The problems the child has to solve are much simpler in that case than with TO.

7. Avoiding negative transfer.

In thinking about maximizing positive transfer of skill, we also have to avoid negative transfer. That's an area of psychology that we haven't time to go into in detail, but we talk about negative pro-active interference. That means we've got to avoid teaching something at the beginning which is going to get in the way of learning something else later on. It is a very important principle that hasn't been applied sufficiently in education. One of the things I found in my Russian studies is that the Russian education authorities are very keen on this particular concept. You should never teach reading and writing for example in Year 1 in such a way that it is going to get in the way of some other aspect of linguistic achievement at a later stage. Not only in school but even at university level you should never teach something in a way at the beginning which is going to make it harder to learn something else later on. You should always learn from the beginning right through.

There was in fact some negative transfer from i.t.a. I.t.a. did a good job, as we saw from those results, but it could have been better. I said so in the book, though Sir James Pitman, God rest his soul, was not very pleased about it; but it had to be stated, because I was doing a scientific study of his work. As I said earlier, when children were tested in TO just after the transfer, they did less well in TO than they had in i.t.a., and it is possible that i.t.a. could have been designed better, to cause less of a setback when the children switched from one to the other. That is a point to bear in mind with spelling reform.

Pages 163-67, 241-44, and 285-87 in Evaluating the Initial Teaching Alphabet deal with the test words in the Schonell test and the Neale test. We examined them under various skill modes, and some words caused much more difficulty than others. One was the i.t.a. form wurk which changed to work. We found that it wasn't so much the configuration of the whole word, but details within the words which caused the errors; and in work the change from U to O produced a lot of errors. On the other hand a word like church did not cause problems. Another example of a gross negative effect from something for which i.t.a. didn't plan ahead was the word show. In i.t.a. it looked like shoe, so when the children saw show, they didn't read it right, and then when they saw shoe, they would probably have read it as show, because that is how they learnt it in i.t.a. We didn't analyze whether vowels or consonants caused more trouble, but that analysis could be done from this book. On ppl64 and l67 there are tables of words showing the number of errors, and that kind of analysis could be done from them, indeed it would be important in formulating a first stage of reform. I think that is definitely one of the advantages of Cut Spelling: there is much less likelihood that that kind of difficulty could arise.

Also relevant to the design of a reformed spelling system is this passage on p245:

...one conclusion seems clear, and that is that in the future development of this transitional alphabet approach to the teaching of reading, a series of experiments should be carried out to ensure that every element in the design of such an alphabet and its spelling rules has been established empirically as the best possible solution in the total complex of problems involved in making the needs of the beginner compatible with maximal transfer efficiency at a later stage.
In other words, before you decide between several alternatives, you need to do some experiments to find out which will produce the results that you want. They wouldn't have to be big experiments like the i.t.a. ones, but could be quite small, such as Valerie Yule does. And then on p247 I say:

... it seems possible that improvements in the design of a simplified alphabet, improvements in teaching material, and improvements in the methods of teaching with a transitional alphabet could reduce the extent to which TO sets back pupils' progress at the transition stage.
So besides the design of the alphabet, one needs to think about how it's going to be presented to children for teaching. When we set up the experiment, we went straight into it with very little idea of what was going to happen. But many of our teachers did develop of their own accord very good ways of making sure that children got over from i.t.a. to TO. We published some of their findings in a magazine that we then had for teachers. A lot could be done in preparation for transition, however.

8. Conclusion.

What I've tried to say is that reading, writing, speaking and listening are each what we call in psychology a skill. And in each case the many subskills that go into those major skills pass through those three phases: the cognitive phase (figuring out what you have to do), then mastery through practice, and then automaticity. If you want to maximize transfer, it's that automaticity that you have to try to achieve in the beginning stage of teaching children in school, because once you've achieved automaticity, you never lose the skill, and you're going to have the easiest transfer of that skill to other orthographies and other languages.

9. Points raised in discussion.

9.1 Experiments before the i.t.a.
Chapter 2 of Evaluating the Initial Teaching Alphabet is entitled 'A Review of Previous Investigations', and reports on Miss McCallum of Cowdenbeath, Scotland, getting good transfer to TO after using the International Phonetic Alphabet in a small experiment. Maurice Harrison's history of the Simplified Spelling Society (Harrison, 1971) refers to the Society's pamphlet (Simplified Spelling Society Pamphlet No.7, 2nd edition 1942) outlining earlier experiments carried out in 1915-24 in 16 schools, mainly in Northern England, but also in Scotland, London and elsewhere in England, where children had been taught to read using a reformed orthography; the headteachers wrote very graphically about it. What we notice is that good ideas like this have not lasted, but have disappeared.

9.2 I.t.a. Federation & Foundation.
I.t.a. hasn't disappeared yet [It was still taught in a few schools into the 1990s. - Ed.] because the teachers formed the i.t.a. Federation to support each other. It started in Australia, and at my suggestion it was set up in Britain too and gave the teachers the strength to carry on. Then there is the i.t.a. Foundation financed by money left by a millionaire in the United States. It mainly exists as a funding organization, giving grants to people for research. They've got quite a good service going where they go round and give workshops and so on, but it no longer operates in Britain.

9.3 Why i.t.a. declined. Godfrey Dewey had a furious argument with Sir James Pitman, trying to persuade him not to use the strange characters he had designed for i.t.a., and to use something like New Spelling instead. And i.t.a. lost popularity because parents objected to the unfamiliar letters, the augmentations. They didn't like them because they hadn't had them at school themselves, and they couldn't help their children because they didn't know how to use them, although we published a pamphlet for parents which was given away free. Parental objections were much stronger in the United States than in Britain, and i.t.a. has nearly disappeared in the United States, because in the United States the schools have to do what the parents tell them, and the parents have succeeded in squashing nearly all innovations that caused them any disturbance. But in Britain by the summer of 1986 less than 100 schools were still using i.t.a.

9.4 Effect of i.t.a. on different abilities.
It is sometimes said that the brightest children were helped least by i.t.a., the least able children more, and those of average ability most of all. From the graphs in Evaluating the Initial Teaching Alphabet, on the other hand, it looks as though the brightest children gained most, and the slow ones didn't gain at all. But Professor Vernon of Reading University took me to task about that and said I should have looked at the results in another way, which is why afterwards I published a short article i.t.a. and Slow Learners - a reappraisal (Downing, 1969). There I analyzed the data as she suggested, and it showed that the slower children did gain from i.t.a. I'm very glad she drew my attention to that, because the graphs in the book are misleading in suggesting that slower children were not helped. The problem was that the slow achievers probably didn't have enough time with i.t.a. to show on the graphs. Vera Southgate and her colleague Professor Warburton (Warburton & Southgate, 1969) went round the country interviewing and did a very good nonstatistical study which showed the teachers really felt the slow learners had got a lot out of i.t.a., and I think they were right.

9.5 Testing for the Hawthorne effect.
It is often asked how far the Hawthorne effect may have helped the success of i.t.a., in other words, whether those involved may have been additionally motivated by the sense of being given extra attention rather than by the advantages of i.t.a. itself. We did an experiment deliberately to create a Hawthorne effect in the control group that was using TO. Whenever we did something for the i.t.a. teachers, we did something for the TO teachers too. We had two days of workshops for teachers in i.t.a. so they could learn how to use the new characters. Of course, we couldn't teach the TO teachers how to use the TO characters, but we did put on workshops for them too, for which we brought in well-known educational speakers and so on. They were given hints and special treatment. Then we compared the children in their classes with the children in the previous year, where the teachers hadn't had any special care, and there was no difference with them. I don't think there was any evidence for the Hawthorne effect with i.t.a. It would be the teachers rather than the five-year-olds who would be susceptible to it, and they would pass it on to the children; but the children seemed unconcerned by us coming into the classroom and soon got used to us.

9.6 Why teachers adopted i.t.a. We should perhaps have done a social psychological study on whether the teachers involved might have been selfselected as the most enthusiastic. I did talk informally to headmasters and headmistresses about why they were in the experiment, and there was a wide range of reasons. In Wolverhampton we had two extremes. One was a school in a very poor neighborhood, and the headmaster said, "Oh, there are so many problems of reading here. We'll try anything, and this is something we haven't tried". Then at the other end of the town, in an upper-middle-class suburb, the headmaster, who had a very good reputation with the local authority for his school, said, "Well, when I saw the people who were volunteering to do i.t.a., I thought you ought to have one person who had his head on his shoulders." Then in another school, in Stoke-on-Trent, the headmistress, who had the best reading results in the city, had looked at i.t.a. and decided it would produce better results - she didn't want to lose her place being top in reading in town. So there were all kinds of different reasons for them being in the experiment. It was the headteachers who decided, some more democratically than others - some decided and then told the staff, and some consulted the staff first.

9.7 Teachers' objections to i.t.a. I.t.a. was designed as a transitional system for children, and was never intended for general use. One of the reasons why i.t.a. didn't succeed as Sir James Pitman hoped it would, and one of the reasons that was given by many teachers and headteachers for not trying it, was that it was only transitional, and children couldn't continue to use it for the rest of their lives. They said, if you had suggested spelling reform, and hadn't brought in those new letters, but just used the alphabet in a more sensible way, then we would have accepted it, but not something like i.t.a. that was only going to be used in the first years of school.

9.8 Advantages of Cut Spelling. The idea behind Cut Spelling (Upward, 1996) on the other hand is that it could serve both functions: it would offer an improved system to help children learning to read and write, which they could continue to use indefinitely. Insofar as it is compatible with TO, children would be able to read TO but would never need to learn to write it. In this way TO and CS would exist side by side. Some people talked about children having to forget i.t.a. after the first two years of school, but CS would go on being useful for ever. And people can see it's got practical uses.

It is asked why New Spelling is less suited to immediate introduction and continued use than Cut Spelling, when for instance for come New Spelling proposes the fully phonographic form kum, while Cut Spelling com only removes the final -E from the TO spelling. One reason is the need for compatibility with TO, as was already demonstrated above over the difficulty of transfer from i.t.a. wurk to TO work. Another reason is that although you might say to parents that their child could continue to write kum for the rest of its life, they would fear the child would attract ridicule in the adult world by using such forms. On the other hand with Cut Spelling you could establish a parallel system of writing which had its own practical use for ever, just as we have arabic numerals in parallel with roman numerals, and both have their uses. This would appeal to the ordinary parent and the ordinary teacher as something that would help children to learn to read TO and then afterwards they could use it as an alternative system.

Any new spelling scheme, Cut Spelling or New Spelling, would have to be sold to the public, and CS has several selling points. Judging from my earlier background in public opinion research, Cut Spelling has a lot to offer from the point of view of public acceptance. There's a lot in it which would appeal to people lacking specialist knowledge of the subject. Some aspects of it are things I've done for myself: I've never learnt any particular shorthand, but I've made up my own as many people do, and a lot of things that are done in Cut Spelling are very like shorthand. I think a lot of people would be attracted by its economy, and it wouldn't look as strange as i.t.a. If I was a member of the general public, I'd look at CS and say, "Ah, that looks useful to me, I could use that for making my notes, because it's so short, much shorter - that'll save me a lot of writing, that will". You've got the economy, quite a big percentage, and then if everybody used it in business and so on, it would save a lot of time and paper, etc. You could sell CS on those lines, as well as on educational lines. It might also become popular because of computing.

What you'd have to do to sell New Spelling would be something that Chris Jolly and I have been talking about for quite a long time: you might be able to sell it on the grounds that this would be a good way of being able to write down the way people pronounce things. But you can't assume in my opinion that the public is going to buy spelling reform. My feeling as a psychologist is that Cut Spelling would be much easier to sell to the public.

The way in which CS had been seen as relating to New Spelling is as a first-stage reform eventually leading on to a thoroughgoing reform like New Spelling, which would be the ultimate destination after a series of intermediate reform stages. The Simplified Spelling Society has been considering reform from two opposite directions at once, discussing a revised version of New Spelling at the far end of the process, and Cut Spelling at this end as the first stage. As yet the CS Rule for omitting letters representing post-accentual schwa before L, M, N, R constitutes a discrepancy between the two ends, which should be brought into line with each other. Both a first stage reform and a long-term proposal are needed.

9.9 Japanese.
In Japan the children transfer from the Kana syllabaries to Kanji because the education system requires them to learn nearly 2,000 Kanji characters by the time they leave school. I believe the Kanji characters are needed to avoid confusion of homophones, such as can occur in telegrams, for which only Kana is used. In normal adult writing you see a combination of Kana and Kanji, the Kana words being used for the less important words, while the nouns will be in Kanji characters, which are complicated ones derived from Chinese. The difference between them hits you in the eye, with the important Kanji characters standing out in the sentence, and Kana in between. There are two kinds of Kana, Hiragana and Katakana, one of which is used tor foreign words. They print words in the roman alphabet, or Romaji, for foreigners. But if you hire a car, the maps are not in Romaji, and it's an interesting exercise to transfer your reading skills from English to Japanese in map-reading. The important items will be in Kanji, not Kana, and for instance we had to look out for the sign of the man hanging on the gallows!

9.10 English in India.
There's a tendency in India for reasons of snobbery to send children to convent schools where English is the medium of instruction. That is spreading to smaller towns now, where there is not the English atmosphere which can obtain in a city like Bombay (Mumbai), where quite a lot of people use Indian English. Instead there's a total Marathi atmosphere, in the midst of which there's an English medium class, taught by Marathi speakers. There must be a danger the children there would be much better off speaking their own mother tongue first and then transferring to English - unless they're extremely bright.

9.11 Age of optimum transfer.
There have been a lot of studies comparing different age levels of introduction to a second language. The popular idea is, the younger you start, the easier it is (that's the reason for the current Canadian fashion for French immersion with very young children). But it's not true: studies show that in general adults are better at learning a second language, and that the older the child, the better the result. There is one important exception: younger children pick up the phonology, the pronunciation better. So perhaps children should learn to speak a foreign language when they are young, but leave formal study of it until later.


Downing, John (1967) Evaluating the Initial Teaching Alphabet, London: Cassell.

- (1969) i.t.a. and Slow Learners: A reappraisal. Educational Research, Vol.11, pp229-31.

ed. Downing, John (1988) Cognitive Psychology and Reading in the USSR, Amsterdam: North-Holland.

Downing, John & Che Kan Leong (1982) Psychology of Reading, New York: Macmillan.

Harrison, Maurice (1971) A Short Account of Simplified Spelling and the Simplified Spelling Society, Simplified Spelling Society Pamphlet No.11, republished in Simplified Spelling Society Newsletter, Summer 1986, pp22-25.

Herbert, A P (1920): 'A Criminal Type', Punch; Vol. 159, 28 July 1920, p62, London.

Ripman, Walter & Archer, William (6th edition 1948, revised by Daniel Jones and Harold Orton) New Spelling, London: Sir Isaac Pitman Ltd.

Simplified Spelling Society Pamphlet No.7 (2nd edition, 1942), The Best Method of Teaching Children to Read and Write,London: Sir Isaac Pitman & Sons Ltd.

Upward, Christopher (1996, 2nd revised and expanded edition) Cut Spelling: a handbook to the simplification of written English by omission of redundant letters, Birmingham, UK: Simplified Spelling Society.

Warburton, F W & Southgate, Vera (1969) i.t.a. An Independent Evaluation, London: John Murray.

Back to the top.
See ita alfabet and sample text from an ita reader.