News8. (underlined words and letters are presented as headings or in italics here.)
On other pages: part 1, part 2, part 3, part 4.

[David Stark: see Journals, Newsletters.]

Newsletter Summer 1985. part 5.

David Stark: Alphabetic Consistency in Reading.

To someone interested in spelling reform, it may seem axiomatic that the inconsistent and complicated relationship between the written and spoken word in English is the main cause of literacy difficulty among English speaking people. However, there have been many studies into literacy problems in English, and in the books produced, little or no mention is made of the orthography as a factor, nor of reform being considered as a solution to the problem. Where previous attempts at reform are mentioned, they are quickly dismissed as irrelevant.

Certainly, the mental processes involved in reading and writing are far more complex than most spelling reformers realise. Reading is not just "barking at print", and writing involves far more than matching graphemes to phonemes. Spelling reformers, if they are to be taken seriously, must not regard reform as an abstract theoretical puzzle, but must understand the reading and writing processes and the part which more consistent phonographic correspondencies could play in making literacy easier to acquire.

Let me leave the writing process for another article and first of all consider reading. A large number of studies have been carried out into reading problems, but unfortunately the results of these are seldom encouraging for spelling reformers. Probably because linguists do not consider the orthography as a variable factor, they tend to look to other areas of the reading process to seek solutions to reading problems. Also, many studies seem to infer that the English orthography does not pose much greater difficulty than alternative orthographies in other languages. Figures are often quoted to show that about 4-5% of the people in certain non-English speaking European countries, with more alphabetically consistent orthographies than English, have serious reading difficulties. The Adult Literacy and Basic Skills Unit in Britain estimated that 6% of the adult population in England and Wales is functionally illiterate.

Assuming that one can be satisfied that the criteria used in measurements of literacy in different languages are comparable, one is inclined to conclude that English speaking people suffer little more, if at all, from reading problems than our foreign counterparts.

Furthermore, one study suggests that only 1% of Japanese have severe reading problems. If this figure is true, it challenges the belief that an alphabetic orthography is automatically better than an ideographic one, While it may take an entire school life to master the minimum of 2000 characters necessary for functional literacy in Japan, results appear to indicate that the efficiency of being able to translate directly from visual symbol to meaning is worth all the effort. One American study substituted English words by Chinese ideographs, and found that many retarded American readers mastered these in a matter of weeks, whereas they had struggled for years to absorb alphabetic relationships.

Let me try to find an explanation for these apparent anomalies, anomalies that is for advocates of spelling reform, by exploring the part which phonographic correspondence plays in the reading process. Spelling reformers may be surprised to find out how small a part this is.

Reading is all about drawing meaning from written or printed text, meaning which is being conveyed by the author. The whole exercise is a detective hunt, a process of using clues in the text to build up hypotheses of intended meaning, and testing these to confirm if they make sense, based on one's own experience and knowledge. One could also use the analogy of a jigsaw puzzle, where one is trying to build up an overall picture of a subject using individual pieces which can be separately identified but which have limited or variable meaning on their own, out of context.

One might expect alphabetic relationships to play an important part in the identification of the individual jigsaw pieces, that is, the individual words in text. One might use sequences of individual letters or digraphs to build up pronunciations which one would recognise as meaningful words. While there is some scope for such an alphabetic analysis, in practice we only partly use alphabetic clues in reading, preferring instead to identify whole words, parts of words and groups of words as visually meaningful elements. We usually go straight from printed word to meaning without trying to determine pronunciation.

To understand why we do this, we must understand what "short term memory" is, how it works, and bow it holds information until a search has been made in the "long-term memory" for clues to meaning.

Short-term memory is capable of holding small amounts of information for a short period. It is ideal for remembering telephone numbers between looking them up in a phone book and dialing them. A short telephone number of 6 or 7 digits can be remembered in this way, but if an area code of perhaps 4 or 5 more digits is added one will usually have to write the number down, unless of course the area code is held in one's long term memory.

Short term memory is also necessary for mental arithmetic. Imagine one is adding up a string of numbers, say 8+7+3+5. One would start by adding 7 to 8 to give a sub-total of 15. The 8 and 7 would then disappear from the short term memory, leaving the sub-total to be added to the 3. A new sub-total of 18 is remembered for as long as it takes to add this to the 5 to achieve the result of 23.

While individual numbers will usually be processed one at a time using relationships stored in one's long-term memory (for example 8+7=15), to save time one might take pairs of numbers which one knows add up to 10 and process these as individual units. For example, one might start with the 8 and notice the (7+3) as a combination which means 10, and go straight from 8 to a sub-total of 18.

The combination of letters to form a word uses short term memory. Supposing a child learning to read comes up against the word "consult" and does not recognise it, that is, the word is not familiar enough to him for its visual image to be stored in his memory. He would then try to determine the pronunciation of the word, since he will know many more words by their pronunciation than by their visual appearance.

The most common pronunciation of each letter is placed together on a string, /k+ɔ+n+s+ʌ+l+t/. It does not usually matter if this is only an approximate pronunciation of the word, as the person will have heard the word pronounced in varying dialects and in various conditions, and will be able to jump from unknown pronunciations to a close, familiar one.

Short term memory is required to process the combination of the six digits, remember the combination long enough to explore the meaning of the resultant word, and fit the word into context before the context is forgotten. This will be difficult for a child who has recently started school and has not developed the necessary mental agility.

A slightly more experienced reader will have learned a shortcut to reduce the burden on the short term memory. He will realise that words consist of syllables, and will process the word "consult" as /kɔn+sʌlt/ leaving two syllable units to be retained in the short term memory while the search for meaning and context is made. The chances are that he will have come up against the syllable "con" several times before, and have its pronunciation stored in his long term memory. This would be useful for longer words like "consonant" which would be much easier to process as three syllable units instead of nine graphemes.

In a single glance, one can only process four or five letters every second. If a child had to rely on processing every individual letter, he would plod through a text at a rate of about 60 words a minute. Any speed of less than 100 words a minute places too large a burden on the short term memory, and by the time he reached the end of each sentence, its overall meaning would have been forgotten.

Our brains are accustomed to making sense of dialogue spoken at 100-200 words per minute. Speech slower than this will burden the short term memory by asking it to hold individual words and phrases while the brain is waiting for meaningful chunks of dialogue to be conveyed. In such circumstances, the listener becomes bored and sleepy. A reading speed of 200 words per minute is a good target for students of reading to aim at, and an experienced reader will achieve about 300 words per minute, largely by learning to ignore auxiliary and unnecessary words.

A child must quickly learn to process whole words at a time so that only the occasional word has to be broken down into smaller units; and it would be beneficial if some of these units were syllables rather than individual graphemes. Use of "look and say" techniques might continue to be relevant, even if the orthography were revised so that it became alphabetically regular and simple.

If the reading process is primarily morphographic rather than alphabetic, opponents of reform argue that alteration of the orthography to become consistently alphabetic would render it a less efficient tool for reading.

Reformers must concede that, for example, the eradication of homophones would increase the semantic searching process in reading, and that the memorising of common homophones does not significantly increase the learning process; nor does the representation of inflections by morphographic means rather than alphabetic, for example, -s to represent the plural inflection whether the pronunciation is /s/or/z/ (there are also good arguments for using the alternative -z in a revised orthography).

A spelling reformer would naturally seek to make the orthography as completely consistent alphabetically as possible, but should realise that the inclusion of very common words which did not comply with the defined alphabetic rules would not make it significantly more difficult to learn. Nor would the inclusion of some consistent syllabic elements, like -tion for /ʃən/. While whole word and syllable recognition can be compatible with a strictly alphabetic standard, it may be necessary in practice to appease the protests of existing literates by allowing the retention of some features of traditional orthography.

There may be some dubiety about the need for alphabetic consistency in common words, or for phonographic accuracy in all letter/phoneme relationships in long words. However, the facility for children to work out at least the approximate pronunciation of visually unknown or unfamiliar words plays a vital part in the learning process. Children can learn how to read words without constant help from their teacher; and alphabetic clues will always be necessary for identifying less common words and for separating visually similar or confusing pairs of words, like inert/invert or was/saw. Mute letters and letters which can have three, four or even more possible pronunciations will cause confusion and slow the identification of individual words, so placing great strain on the short term memory.

If the percentages of people in different countries with reading difficulties quoted earlier are correct, showing English to produce a slightly larger number of poor readers, many people would argue that the fairly small gap (about 2%) could be made up with appropriate remedial action and extra teaching. The educational establishment is loath to consider the orthography as anything but unimpeachable, even if they do recognise that it has shortcomings; and many people will not even admit this.

Certainly, the present educational establishment tends to ignore the very positive benefits which greater alphabetic consistency would afford, especially in the learning process. Unfortunately, since so few educational psychologists regard spelling reform as a serious alternative, there have been few studies to show how these theoretical positive benefits would translate into practice.

In one of the books mentioned below, the author ponders at one point on how interesting it would be if one could compare English with other languages which have different kinds of orthographies. But unfortunately it is found that practically all the research into reading difficulties has taken place in English speaking countries. The author questions this no further, at least in the book. Spelling reformers know the answer. There have been so many studies into reading difficulties in English because the English orthography is so difficult.


"Dyslexia or Illiteracy?" Peter Young and Colin Tyre. The Open University Press, Milton Keynes 1983.

"The Psychology of Reading and Spelling Disabilities" AF Jorm. Routledge and Kegan Paul 1983.

Back to the top.
On other pages: part 1, part 2, part 3, part 4.