[Journal of the Simplified Spelling Society, 1989-2 pp.22-24 later designated J11]
[David Stark: see Journals, Newsletters.]

2 - the Principal of Minimal Interference.

David Stark.

This article is the second in a series concerned with the implementation of spelling reform. The first appeared in Journal J10 1989/1, pp. 19-20.

Spelling Reform will be of no direct benefit to those who are at present fully literate. On the contrary, they will tend to see it as a threat. If society converts to a new standard which they do not have a proper command of, they could lose status and power.

A common tendency of reform proposals therefore has been for them to look as near as possible to traditional orthography by retaining existing alphabetic relationships, where they are acceptably consistent and simple, and the existing alphabetic symbols (graphemes). This reduces the risks of rejection by existing literates and makes it easier for them at least to read text in the new orthography. It would also facilitate reading of traditional orthography by those who have been educated in the new one.

There needs to be a balance between the present, complicated and inconsistent orthography and a completely 'rational' one. The ultimate latter solution could be said to be the one inspired by George Bernard Shaw, [see newsletters and journals] who left money in his will to fund a competition. The winner invented a completely new alphabet with spelling rules of mathematical consistency. However, it has never been considered as a practical solution.

An existing literate converting to a new orthography will learn it in a different way to a child. The existing literate will already possess all the comprehension skills and will be impatient of any obstacle which impedes quick understanding of the text. In contrast to the child who tends to rely heavily on alphabetic relationships for reading and spelling, the adult will not wish to regress to phoneme-grapheme units.

To do this would reduce reading to around 60 words per minute. This would be very boring to a brain which has developed to deal with comprehension of speech and hence script at about 200 words per minute. The experienced reader processes whole words and bits of words at a time to reach this speed. Writing is slower, largely for reasons of limited manual dexterity, but still relies heavily on non-alphabetic skills.

In reading, therefore, the existing literate would require to relearn a multitude of syllabic and word signs as well as alphabetic relationships. Writing is a different and more difficult problem. The existing literate, as well as having stored in his brain a number of syllabic and word signs, will have memorised a literary pronunciation for most words.

The need to make a good first impression on existing literates is also a reason to seek minimal interference with traditional orthography. It may be an emotional response for someone to dismiss a new spelling as 'peculiar' or 'ridiculous' merely by looking at it. However, it will be easier to make them look a second time, and convince them that the new spellings are merely a more rational advance on the old ones, if the link with traditional orthography is clear.

There are a number of guidelines to be followed when aiming at minimal interference.

1. Use only the existing letters of the alphabet.

There are only 26 letters to represent about 43 phonemes. In theory it might be sensible to invent the extra letters to suit. In practice one has to make use of digraphs at present in traditional orthography.

2. Maximise TO sound-symbol correspondences.

Ideally one should standardise the most common phoneme-grapheme relationships in traditional orthography. For the consonants this is possible, but when the most common way to spell /z/ is <s>, one is more inclined to adopt the less common <z>. When one confronts the vowel sound in see one finds that the graphemes <ee> and <ea> are about as common as each other. In choosing one of them, the reformer will be changing at least 50% of the spellings of the vowel.

One must resist using a redundant grapheme like <x> (which could be substituted by <ks>) for a phoneme with which it has not been associated in traditional orthography, say the <ch> sound in loch. It is also unfortunate when a set of proposals like New Spelling has to use <ae> for the vowel in take when this grapheme is not used commonly for this purpose in traditional orthography.

3. Do not make new phoneme contrasts.

Many reform proposals split the two <th> phonemes in this:thin. In so doing they require to invent a new grapheme (usually <dh>) and force existing literates to make th' phonemic distinction for the first time. It is unnecessary to introduce this complication, as traditional orthography has proved from experience not to require the phoneme contrast for any morphemic purpose.

The RP distinction in the vowels of the words lass-pass is also one which traditional orthography has survived happily without. The introduction of opposing graphemes <a, aa> is unnecessary.

4. Minimise the chances of semantic confusion.

Some words will translate into the new orthography in forms which suggest other words in traditional orthography. For example the word talk will change to tauk perhaps, while talc (short for talcum powder) might change to talk. Someone who learned TO first would hesitate at the new word talk in its context before making sense of it.

However, the most obvious area for semantic confusion for the new user of the revised orthography is the translation of homophones. Previously differentiated words like whole: hole will now normally be spelled the same.

5. Retain existing spellings in preference to phonetically precise pronunciations.

Even if one accent is used to base the revised spelling system, the poor interpretation by ordinary people of phonetic distinctions allows some latitude in the pronunciations suggested by the spellings. For example, if RP is used as the base dialect one might accept more as opposed to mor to retain the existing spelling.

6. Retain TO spelling rules as far as possible.

The retention of slightly more complicated alphabetic spelling rules will not make the acquisition of literacy significantly more difficult. For example, the TO rule for the vowel sound in coin is that it should be represented by <oi> before a consonant and <oy> before a vowel and terminally. The TO user will be familiar with this. (There are a few exceptions like gargoyle, oyster.)

7. Retain as many non-alphabetic TO rules as possible.

Traditional orthography contains a mixture of alphabetic and non-alphabetic patterns. For example, the <-tion> ending occurs in hundreds of words, and with its unclear vowel sound might be better not translated into alphabetic elements (<-shun, -shon, -shin>?). It would certainly suit existing literates if it were not.

To many reformers, the criteria set out above would, if fully implemented, produce a solution which diverged too much from the straightforward alphabetic strategy. That is, that there should be one grapheme to match each defined phoneme, with the minimum number of spelling rules to be learned.

Nevertheless, these reformers accept that their more radical proposals will have great difficulty being accepted by existing literates. Consequently, a variation on minimal interference has been developed over the last 20 years or so, where one advances towards the ultimate reform package in a series of minimal interference reforms.

The first of these step-by-step reforms was published in 1969 by Harry Lindgren in Australia. In 1984, the Simplified Spelling Society in Britain formally accepted Lindgren's first stage or SR1 along with 4 other elements as its Stage 1. In 1985, Chris Upward proposed Cut Spelling as a first stage reform.

While the criteria for normal minimal interference will apply at least in the early stages of step-by-step reform, stage reform has a number of additional factors to be considered. These were first set down in the Simplified Spelling Society's Newsletter of October 1984, after I had explored the problems involved in stage reform by designing a theoretical stage one proposal called NUFASE (New User Friendly Alternative Spellings for English).

1. Each stage must stand on its own.

It must be recognised that any stage might be the last one to be adopted. As the orthography tends towards regularity, society might decide that the benefits to be gained by progressing further with the simplifying process are not worth the bother of reform.

2. Each stage must be significant.

Each stage must be significant enough to be worthwhile bothering about in the first place. It will cause people bother to adopt any reform, and its benefits must be obvious, especially in the initial stages. It must be a marketable product.

3. The route to the ultimate proposal must be planned.

Since any stage reform is part of a package, the following stages must be planned, at least roughly, before the first stage is launched. A present stage must not determine or restrict future ones unless planned to do so.

Stage reform is not meant to be a substitute for more comprehensive reform but a means of reaching it. It may be possible to form non-controversial initial stages, but phoneme-grapheme correspondence choices, dialect definition, phoneme definition, indistinct vowels, non-phonic features to be retained or lost, and all the other difficult reform decisions require to be made as early as possible in the process.

4. The starting point for stage reform must be defined.

Stage reform will, in effect, be a process of rounding up the large number of rules and patterns in traditional orthography into smaller and simpler groups of rules based on alphabetic principles. In the early stages there will be more traditional orthography rules and patterns, and these must be identified and accepted - for example, a final <e> changing the sound of a preceding vowel as in cap/cape. It must also be realised that traditional orthography is a mixture of graphic elements - alphabetic, syllabic and morphographic (units of meaning, i.e. whole words learned at a time). In general, the latter two would reduce and the alphabetic rules become fewer and simpler as the process progressed.

5. There should be as few stages as possible.

Too many stages in the process of arriving at the final reform creates the following problems:

i Small scale reform stages offer little or no benefit in themselves.

ii It would be a cumbersome task to control and manage the presentation and introduction of several stages.

iii It would take time to introduce each stage and have it accepted. By the time a long series of reforms had been presented, they would probably be out of date owing to changes in pronunciation.

iv Several stage reforms might become isolated from each other, and the overall structure and plan of reform might be lost.

v There is the possibility that, if any stages are accepted, they will only be the first ones.

vi The Norwegian state reforms of 1907, 1917 and 1938 were on the whole each only accepted by the younger sections of society, resulting in one being able to work out roughly how old a person was by how out of date their spellings were. A series of reforms is likely to lead to people of different age groups spelling the same words in different ways.

6. Words should change as seldom as possible.

Stage reform will involve some words having their spelling changed more than once before a final spelling is fixed, assuming each stage is the result of consistently applied rules. For example, the word phase may go through the stages phase-fase-faze-faez. The scope for confusion in this process is enormous, where the public become unclear as to which is the current spelling.

7. Visually cued reforms are easier to identify.

In the first stages of reform, visually cued reforms are preferable, as these will be more easily spotted amongst the proliferation of phonographic relationships in traditional orthography. The likes of <ph-f> will be easier to notice in text and to remember than a phonetically cued reform like Lindgren's SR1 (/ε/= <e>). Here, each affected word would require to be learned individually.

Lindgren proposed 50 or more annual reforms, contending that each would hardly be noticed. This would not be an advantage, for as well as the large number of orthography changes proving a colossal administrative problem, each stage would be so insignificant as to prove difficult to sell on its own. SR1 has not yet been sold after 20 years. If it took 20 years to have each stage adopted, the reform process would take 1,000 years.

At its AGM in April 1984, the Simplified Spelling Society formally adopted what was generally considered as a candidate for a first stage reform. This consisted of the following elements:

SR: 1
SR: ough
SR: augh
SR: d.u.e.
SR: ph
Spell short /e/ with an <e>.
Respell all <ough> words.
Respell all <augh> words.
Drop useless <e> after short vowel syllables.
Respell <ph> words with a single <f>.

This package arose from a desire to agree at least a particular reform within the Society, but despite its limited size, it could have been justified as a first stage reform, or at least part of one, if the following stages and final reform had been defined. A working party set up at the time to devise a revised New Spelling never completed its task.

The Simplified Spelling Society Newsletter of July 1984 is significant, since as well as reporting on the results of the AGM that year, it contained some of John Beech's conclusions on minimal interference, as well as Valerie Yule's, Chris Upward's and my own ideas on cutting out superfluous letters from TO. (My own proposals in NUFASE also sought to remove some of the worst horrors in the present spelling in an attempt to win the sympathy of existing literates who had trouble with spelling).

Valerie Yule reckoned that a spelling system with the removal of the 5% surplus letters she defined would be faster to read than traditional orthography. Chris Upward took this further, aiming at about a 15% cut (although this has since been reduced to about 10%). Cut Spelling has been the most widely publicised reform since, but he has still to convince many people that the large degree of cutting is a good selling point, that it can be part of a viable long term reform package, and that its rules and limits can be mastered by existing literates.

However, I believe that a main failing of stage reforms to date has been an inability to take on board traditional orthography rules and patterns at the start. They are so anxious to find the end result that they ignore the start, where we all stand at present. Without defining a start as well as a finish, it is impossible to show how far along the road one has travelled. My next article will address this.

Back to the top.