[Journal of the Simplified Spelling Society, J32, 2003/1, pp4-8]
[Steve Bett: see Journals, Newsletters, Personal View, Web link.]

The Two Stage Approach to Spelling Reform.

Steve Bett.

Should the society endorse a parallel pronunciation guide writing system?

Instead of trying a frontal assault on traditional spelling, this proposal recommends an end run. First accustom the reader to a pronunciation guide spelling by 1. using it in a children's dictionary and 2. using it as an i.t.a. to introduce preschoolers to the basic code. Using this "road to the code" the reader will never forget how to sound-spell and as a result sound spellings will not look that odd. Sapéna will be just as familiar as subpoena and much easier to pronounce and spell.

1. The Quandary: The kind of minimal change that the public would find acceptable will not be sufficient to solve the literacy problem.

Pronunciation guide spelling systems have been rejected by minimalists on the grounds that they are too radical. The public will never accept such a "wholesale" drastic change in the representation of spoken English. On the other hand, few of the benefits observed in the teaching and learning of highly phonemic writing systems such as Italian and Spanish will be realized until the English writing system achieves a comparable level of consistency and phonemicity.

According to Columbia linguist, Mario Pei, anything short of complete phonetization is a complete waste of time [1968]. It is not enough to remove a few stumbling blocks.

If the minimalists were successful in changing the spelling of 100 or even 1000 words in a 72,000-word dictionary, this would not be enough to make a dent in the literacy problem. A minimal change ["patchwork reform"] may send a message and put a dent in the indifference to spelling reform but it will not make English easier to teach and learn.

2. The Relative Regularity of English [7%], French [23%] and Spanish [83%].

The regularity of English has been estimated at between 7% and 85%. It all depends on how you define the regular and how you measure it [Yule, 2003, Dewey 1978]. One measure of phonemicity or regularity is the number of phonemes divided by the number of spellings. Spanish has 24 phonemes and 29 ways to represent them. This makes the Spanish writing system 83% regular. If the regularity of English were measured the same way it would be only 7% regular. According to G. Dewey [1971], English has 41 phonemes and 561 ways to represent them. 41/561=7%. French: 30/130 = 23%.

If such regularity represented the relative ease with which an orthography is learned, then Spanish would be 11.9 times easier than English. Children would be able to learn to read and write in this orthography over 10 times as fast as they could learn English.

Can someone learn to encode and decode in Spanish ten times as fast as they can in English? So far, studies have been able to document that Spanish speaking children can acquire reading and writing skills in the phonemic Spanish writing system about five times faster than in English speaking children can learn to read and write English. [see Campbell's list of studies in JSSS30].

It is not unusual for a Spanish or Italian child to be able to read aloud a newspaper. This doesn't mean that they can fully understand what they have read. A study in the 1960's revealed that half of the high school graduates in Eastern Kentucky could not duplicate this feat. They could not read aloud enough words in a newspaper article to make any sense of it.

A study of German and English dyslexics isolated a similar problem. English students with reading disabilities usually could not sound out multisyllable words. German dyslexics usually could. Both were slow but the Germans were slow and accurate. [Wimmer & Landerl, 1997.]

3. The first step: Over-learn sound spelling and the basic code.

Before suggesting that sound spelling be a replacement for the traditional writing system, a way must be found to accustom the public to it, and to the radical changes in spelling required for a pronunciation guide system.

At the present time, there is little familiarity with the alternatives to traditional spelling. To a tradspel adept, substituting a sound spelling for a traditional spelling usually looks odd and out of place.

It is mostly a case of familiarity. Few would object to LITE for light because the variant spelling is ubiquitous in advertisements for beer.

Acceptance Requires Familiarity.
Although shorter and probably more readable, LITE and THRU continue to be low prestige variants. These variants, however, have achieved what needs to be achieved for all proposed respellings. Before the public will even consider a respelling, it has to be familiar.

The present proposal accomplishes two things: it accustoms the public to alternate spellings and it provides a rationale for why the variant should be used instead of the traditional form.

4. Are mini reforms the only way to minimize disruption?

To make English orthography as transparent as Spanish, a radical reform is required affecting as much as 60% of the words in the dictionary.

For the last 100 years, the proponents of phonemic spelling have thought that the only way to achieve their goal would be thru a series of 5% reforms. The first 5 reforms would fail to make a dent in the literacy problem so they could only be appreciated in terms of being more logical.

There is another way to introduce phonemic spelling that does not involve 50 years of turmoil. It still takes about 50 years but the new spelling is learned as a parallel pronunciation guide spelling. When a majority of users are familiar with two systems, then the switch can be made without much disruption. The change would be that traditional spelling would not be taught until college. Those that wanted to access our heritage of print could do so on their own. After the transition, reading old books would not be as difficult as reading Old English.

The benefit would be more universal literacy. Instead of up to 40% not being able to read the newspaper we would have 40% who would have no access to our "heritage of print". Over 95% would be able to read the newspaper.

We would be using "dumbed down" spelling but why not? The power of the insults would not have the same impact on a populace that was familiar with the pros and cons of the two writing systems.

The two stage proposal involves starting out children with a transparent orthography that reflects the basic code. After three months of writing with the phonemic writing system, the children would learn several other highly likely ways to spell the same 42 sounds.

The phonemic writing system would coexist with the traditional writing system as the pronunciation guide spelling in a dictionary. If successful, it would become the dominant pronunciation guide spelling in all new dictionaries.

To complete the two stage approach will take at least twenty years and as many as 60 years. The new PG spelling and associated dictionaries would have to be used in 60% of the schools for at least ten years. At least 100 million people would have to become familiar with the new spelling and be convinced that it represented how words should be spelled before a radical revision would have a chance of success.

Some would argue that general familiarity and acceptance of the proper way to represent spoken English has to be accepted before any new spelling would be accepted. People can read brite lite thru and altho as easily as they can the traditional equivalents. This has not increased their use outside the advertising world. Currently Lite and thru are not generally considered to be "proper".

How do you account for the preference for the longer obscure spelling over the shorter sound spelling? It is easy enough to list over 300 similar spellings that are well known enough to be listed as variant spellings in at least one dictionary. [see the American Literacy Council list] People continue to use such spellings as through because they are considered proper, learned, and correct.

How do you change these opinions?
Frank Laubach in one of his books, Forty Years with the Silent Billions, defines literacy as the ability to read a newspaper. This is the level of literacy that nearly 40% of those attending the schools in English speaking countries fail to achieve. Many can read 4th grade children's books and have a sight word vocabulary of several hundred words. This is not sufficient to read a newspaper aloud and with understanding because they are tripped up by almost every multi-syllable word.
"English has a booby trap in almost every word." It takes the prize as the "World's worst spelt language" - Frank C. Laubach 1970.
Laubach taught classes in over 300 languages. None of them, he said, were as difficult to teach as English. French, he remarked, requires more lessons than most languages because there are so many groups of letters that stand for the same sound, like <eau> pronounced "o" meaning water [p.293f]. It is necessary to teach every one of the letter clusters. French has over 130 phonograms for about 30 phonemes. And yet French is not as difficult as English because the spellings have only one interpretation. There are virtually no exceptions. <eau> has but one interpretation: /o/. "English, on the other hand, has a booby trap in almost every word." There are six different ways, for instance, to pronounce the string <ough>.

5. The Proposal.

The proposal is to start with a writing system that is more regular than Spanish and get it accepted as the pronunciation guide spelling in dictionaries and taught as an initial teaching alphabet in the classroom. While mini-reforms could still be pursued, the purpose of the standardizing on a pronunciation guide spelling optimized for writing and typing would be to accustom the public to the kind of drastic changes that would be required to upgrade the writing system to the level of Spanish.

The minimalists are right, if such drastic changes were attempted in one step, the proposal would be rejected. However, if it was not introduced until over half of the public was already familiar with it, the situation would be different. The new spelling would no longer look that odd.

1. Invent a PG spelling system that looks good in print and can be rapidly typed.

2. Use the notation as the PG [pronunciation guide] in a children's dictionary. Eventually get all dictionaries to adopt the notation.

3. Use it as an i.t.a. in a writing to read program to teach basic code.

4. Although used initially, the PG code continues as a parallel writing system. It is not scaffolding to be discarded as with the original i.t.a.'s.

5. Moving the parallel writing system into the mainstream may take 50 years. Wholesale spelling reform has to wait until 50% of the population is familiar with it as a parallel system.

No attempt should be made to replace the traditional writing system until the public is ready: The public can be said to be ready when [a] they are almost as familiar with the PG spelling as with the traditional spelling and [b] when a majority can fully grasp the superiority of a less complex alphabetical writing system. Until that time, the PG writing system is simply a parallel writing system that doubles as a dictionary pronunciation guide and is used to introduce the basic code.

6. Teaching two codes: Using the simpler consistent code to introduce the regularities of English spelling.

Although those exposed to Pitman's i.t.a. achieved reading proficiency levels twice as fast as those who had to contend with the irregular code, it was not sold as an alternative way to build phonemic awareness. PA is the awareness of how an alphabet is supposed to work including the awareness of the sounds associated with certain letters and letter combinations.

When up to 40% of the students encountered difficulties transitioning to traditional spelling, those difficulties were often attributed to the i.t.a. rather than to the complexities of the traditional writing system. Had it been explained that the i.t.a. teaches only 41 of the 300+ overlapping sound symbol relationships needed to be a proficient reader and writer in the traditional system, some of the backlash might have been avoided. The i.t.a. student was not prepared for the onslaught of irregularity but since the 41 relationships they had over-learned represented the dominant spelling patterns in English, they were better prepared than most of those who had taken the traditional route.

About 10% of the children in the i.t.a. class were transitioned to TO before they had fully mastered the i.t.a. [Downing ...] Using a different method, Malone was able to insure that all students had mastered Unifon before they transitioned. He found that they were ready to transition after 3 months with Unifon and saw no benefits in prolonging exposure to the consistent code if the goal was proficiency in the traditional code. In contrast to the 3 months, Pitman's i.t.a. basal reader program took 2 years to complete.

Malone [1962] provided no explicit training in the common ways that the 40 sounds of Unifon are spelled in English. This seems like a good idea for those who might not pick this up on their own. By learning the 5 most frequent ways to spell each of the "long" vowels, students would have mastered 85% of the ways that these sounds are spelled in the dictionary. See www.unifon.org/uu-29ways.html.

Although the i.t.a. was considered by some to be pronunciation guide spelling, Pitman never made such a claim. He seemed to go out of his way to emphasize that it was not a pronunciation guide. Its potential as a "writing to read" program was rarely tested. As initially deployed, it was a basal reader series with a "funny font" and an augmented alphabet. Stories were not written 100% phonemically but in a manner that mimicked the traditional writing system. The 5 transcribed books in the basal reader series usually took 2 years to complete. After book 5, children were expected to transition to the traditional orthography.

The real point of the approach was never fully explained. It was simply a way of introducing the basic code by postponing exposure to irregular spellings. Had the basic code been established as the goal, then children could have transitioned to traditional print as soon as they had learned it.

Later studies have shown that 3 months exposure to a phonemic writing system is sufficient for 95% to over-learn the basic code in a writing to read program. 40% of the words in English are already written in the basic code and another 40% introduce only one irregularity.

The basic code is obvious in a phonemic writing system because it is simply a consistent way of writing the 41 basic sounds of English speech. Memorize 41 symbol-sound correspondences and you have it. Some people can memorize 41 paired associates in 15 minutes. Even the most memory challenged student can memorize the sound-symbol correspondences in 3 weeks.

The problem with prolonging the exposure to phonemic spellings is that students will start to memorize spellings as word-signs. Giv, ryt and hav will start looking rite.

The goal of the original i.t.a. research was to see if changing the medium [the alphabet] would make a difference so no attempt was made to find the optimal method of instruction. It did make a difference in the sense that children progressed through the basal reader almost twice as fast as those in the control group. They probably did not progress as fast as children learning how to read Italian and Spanish because of the lack of parental support and assistance.

Because of the idea that the teaching method didn't matter, this research program is unlike many pilot programs where every effort is made to maximize the impact of the new feature.

More recent studies [Flynn, 2002] have indicated that the i.t.a. advocates were on the right track. When taught using a different method, it has been shown to be five times as effective in remedial reading programs as the next best phonics program in teaching phonemic awareness. The i.t.a. was also found to be much more effective in teaching reading and writing as measured by the traditional performance measures.

7. Difficulties in Producing Pronunciation Guide Spelling Systems.

Although New Spelling was very close to a pronunciation guide spelling system, one can also argue that a writable PG system has never been advocated in quite the same way as in this proposal. The proposal endorsed by a half dozen people on Saundspel and the SSS lists is to promote a standardized parallel PG [pronunciation guide] writing system. This does not conflict with the desire to make a dent in the indifference to reform by proposing and promoting a mini-reform such as f for /f/. It is simply a second front. An alternative strategy.

Linguists such as Wells might object on the grounds that everyone should learn and use IPA. IPA has made great strides in the 1990's. It has become a much more popular PG for dictionaries. ESL teachers now use this code more than any other to represent English speech. The basic problem with IPA is that only extended Unicode fonts support some of the turned or rotated characters it uses.

The popularity of IPA as a parallel notation would be increased if the code could be easily handwritten and keyboarded. A rapid handwritten phonemic code that could be typeset was Shaw's dream. With the decline in handwriting skills, the dream today is more for a rapidly typed phonemic code.

Phonemic writing involves the manipulation of at least 40 symbols representing the 36 pure or uncombined phonemes and a few commonly used combinations. Phonemic keyboards and scripts have a symbol for every sound that has the power to change the meaning of a word.

There is almost an infinite number of vowel sounds that the human ear can discriminate. Spoken English, however, has only 14 categories of vowel sounds capable of changing the meaning of a word. 14 phonemes but only 5 phonograms.

It is difficult to represent 14 phonemes with only 5 phonograms but other Germanic languages are faced with the same problem and have evolved better solutions. For example, a first year student of German makes fewer spelling errors in German than when the same essay is written in English. [Upward, 1992].

IPA's solution in addition to importing a few Greek symbols and a few new symbols that Isaac Pitman and others had developed was to rotate the letters. This was a clever idea in the era of handset type. It is not such a great idea in the digital type era.

Those who developed the first computing standards [ASCII] were not aware of these needs and so neglected to augment the alphabet. Some augmentation was added in the mid 1990's. The Latin 1 character set contains enough characters to fully represent all European writing systems. The IPA character set, however, was not popular enough at the time to be included.

The IPA has become an increasing popular ponemic notation for representing English. Henry Sweet, was one of the few that ever used it for representing large blocks of text. He called his system "broad romic" to distinguish his transcription from one that would be used to describe a dialect.

Examples of Pronunciation Guide Systems
Yule argues that a purely fonemic spelling for English even if it were practical may not be the best solution. She prefers a solution that considers the needs and abilities of end users, particularly those with learning disabilities. The basic idea is that morphemic regularity may be easier than phonemic regularity.

Various attempts have been made to find a way to write phonemic English using what is available on the keyboard. The proposal below uses the Latin 1 character set which is supported by your Internet browser but not by all email systems.

Xi: wi cánt icspéct güvørnmnts tu txéndj spéling radicli sevrl taimz in ø sentxøri! it wud bi øbsœrd! EVRÍKING wud hav tu bi ri-printid! <Roly from Oxford, UK>

ALX: Wí kant ikspékt guvernménts tu chaenj spéling radikelí séverel týmz in a sénceri! It wød bí ebsurd! EVERÍÞING wød hav tu bí ríprinted. <Alex Walker from CA> áéíóúôûý

The notations above use Latin 1 characters. The ones below use only ASCII & the QWERTY keyboard. They would be more compatible with older email readers that strip out HTML code.

Unigraphic solution-ENgliS
EN: .wE kqnt expect governments tw chaenj speliN radikaly sevRal tYmz in a senCary. It wvd bE absurd! .evRyTiN wvd hav tu bE rEprintad. <Steve Bett from Louisiana>

Digraphic solution-Spanglish
SS: Wi cant expeckt guvernments tu cheinj speling radicly sevral taimz in a cenchery! It wvd bi absurd! EVRYTHING wvd hav tu bi reprinted! <Steve Bett from Louisiana>

A few exception rules help the Spanglish spelling system to be more like tradspel [traditional spelling].

The rules for Spanglish include:
1. Stress on the first syllable unless marked.
2. u before a consonant is always a stressed /ˆ/ the sound in up. [the up-u]. ur = 3' as in her /h3'/
3. a is always unstressed in multisyllable words as in ago /@/. There is no relative stress in a monosyllabic words so about the only one syllable schwas are the lone <a> and the terminal <e> in <the>. Before this rule was added, there were always two ways to write a monosyllable [unstressed or stressed]: e.g. her or hur. Now <her> has to be interpreted as in IPA <hair/heir> when it appears in a one syllable word. As in tradspel, there is a sometimes confusing switch when we move to multisyllable words: The letter string her in <other> would be spelled <uther> and pronounced /udhar/ where the a represents schwa.

Publishers just want a standardized writing system. Since the publication of Johnson's Dictionary in 1755, they have had what they wanted and they would just as soon keep it that way.

Few publishers are going to be interested in changing the spelling of a hundred or more words. The fact that Webster was able to change the spelling of over 1000 words is truly amazing when you think about the possible resistance.

It was not possible to extend Webster's reforms when it was tried again in 1906. The conservatives were able to get widespread support after the newspapers published a series of popular editorials against what President Roosevelt had proposed in his executive order.

So far I have only been able to dig up one editorial that appeared in the Louisville Courier Journal and one editorial cartoon. There must of been hundreds of these. Roosevelt was forced to rescind his executive order after Congress voted 140 to 23 not to appropriate funds for its implementation. The only Newspaper to adopt the reform was the Chicago Tribune. They justified their spellings with the Funk & Wagnalls Dictionary, the only dictionary to support the reformed spellings. Funk was a member of the Simplified Spelling Board.

Given the potential for conservative backlash, there is a much better chance for getting dictionary publishers to change their PG writing systems than to get the general public to drop even one of their familiar spellings.

About the only people that would not be annoyed by spelling change would be those who had not acquired the habit. There is probably little chance that you can democratically change the spelling of half of the 1000 most common words unless the changes were already accepted as variant spellings.

For instance, it might be possible to change from light to LITE because people are quite familiar with the alternate spelling from beer ads. If this is true, then the task is one of getting more phonemic spellings into the dictionary as alternative or variant spellings.

One of the most productive ways to do this would be to change the spellings used in the dictionary pronunciation guide.

If we had a writable pronunciation guide spelling, this would tend to dictate the ideal spelling. The word is pronounced /eny/ but we spell it <any>. The word is pronounced /worant/ but we spell it <warrant>. The word is pronounced /giv/ but we spell it <give> as if it rhymed with hive.

I think there is a much better chance of changing the practices of 3,000 publishers than it is to change the practices of over 300 million readers.

The problem with the multiplicity of pronunciation guides is that few people have any ability to use it for spelling. Few of those who are taught or figure out the symbol-sound correspondences on their own are adept at sound spelling. Skill in sound spelling is lower today than it was 100 years ago when thousands had to learn it in order to master shorthand.

8. Summary



1. Invent a PG spelling that looks good in print and can be rapidly typed.
2. Use it as the PG [pronunciation guide] in a dictionary.
3. Use it as an i.t.a. to teach children how to read and write.
4. Do not try to replace the traditional writing system until the public is ready: They can said to be ready when [a] they are almost as familiar with the PG spelling as with the traditional spelling and [b] when a majority can fully grasp the superiority of the less complex alphabetical writing system.

In Minnesota, it has already been proposed that by the 6th grade, children should be able to use a dictionary. This implies that they would be familiar with at least one PG spelling as well as the spelling that matches tradition. Learning two writing systems is not that much of a stretch.

For most common words, they will have two images. In addition there will be 1000's of words they will be able to spell phonemically that they will not be able to spell traditionally without checking the dictionary.

When over half the population finds the traditional spelling to be stranger than PG spelling, the time will be ripe for moving from one parallel writing system to the other. Attempts at reforming before the public is familiar with the alternative and before the majority finds that they can spell it much easier than they can spell traditionally are not likely to be any more successful than in the past.

With a level playing field with respect to familiarity, implementing reform will be much easier.


Campbell, Allan. 2002. Recent Research on the difficulties of literacy learning. JSSS30, p.13.

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

Downing, John 2000. John Downings i.t.a. evaluation. JSSS28 2000/2 p.12-15. Twice as many students in the i.t.a. group completed book 5 as in the control group. After 2 years, 78% completed vs. 39% in the non-ita group. p.13. Performance on Schonell word recognition tests was also over twice as high for the i.t.a. group.

Flynn, Jane. 1994. The use of the initial teaching alphabet for remediation of dyslexia. Roslyn Heights, NY: The ITA Foundation. jflynn@smumn.edu www.unifon.org/ita.flynn.html

Laubach, Frank C. 1970. Forty Years with the Silent Billion. Revell, N.J.

Malone, John R. 1962. The larger aspects of spelling reform? Elementary English, 39:435-445, May, '62,

Upward, C. 1991 A role for dictionaries in spelling reform. Newsletter N1, April, p.3-4

Upward, C. Journal of Reading Research.

Upward, C. 1992. German and English Speling Difficulty Compared" research report. JSSS11 p.22

Upward, C. 2000. John Downings i.t.a. Evaluation: JSSS28 2000/2, pp12-15.

Wells, John. 2003. JSSS03 1987 and JSSS32 2003. [See Links.]

Wimmer, Heinz & Karin Landerl. 1997. How learning to spell german differs from learning to spell English. in Perfetti, C, Rieben, L, Fayol, M. 1997. Learning to spell: Reserch, theory & practice across languages. NJ: Lawrence Erlbaum Associates. p.81f. [See Dyslexia books.]

Yule, Valerie. 2003. Could English spelling be made regular without drastic change? JSSS32, April, 2003.

Back to the top.