[Journal of the Simplified Spelling Society, J27, 2000/1 p16-18]
[See other journal articles by Edgar Gregersen.

Compromise Spellings and World English.

Edgar A Gregersen.

Edgar A Gregersen is professor of anthropology and linguistics at Queens College and the Graduate School, City University of New York. He has a PhD from Yale University and long-standing interests/specialties in African languages and Norwegian. He is at present doing a cross-linguistic study of insult topics, and has used/will use reformed spellings of obscenities in the published accounts.

0. Abstract.

Reformed spelling, it is argued, should not undermine traditional pronunciations and should try to accommodate as many varieties of English as is feasible with compromise forms for major classes of words. Some variant spellings will have to be admitted but they should, as much as possible, be kept to a minimum - more or less on the order of the variation found at present.

1. The need for compromise.

In a recent e-mail survey of SSS members' views on particular points of English spelling (Langscape, 1998), the question arose about the desirability of reducing AE and OE to E in words such as aeon, mediaeval, phoenix, amoeba, oenophile, oecology, and the like. Apparently, most members approved of such a change (which is already common in the United States). I did not. My argument then - as now - was that such a change would tend to destabilize the pronunciation.

The traditional pronunciation of such words is with a so-called 'long E' as in bee (/i:/ or /ij/). In the US the reduced spelling has led to many pronunciations with 'short E' and this may have influenced speakers elsewhere as well. The result is that we find at least two pronunciations in the English-speaking world for words such as (a)esthete, p(a)edopile, (o)ecumenical, and possibly also the variation in economics (formerly spelt oeconomics).

Chris Upward has suggested that because of the vacillation between 'long' and 'short' E's in such words, the spelling with E alone could be seen as a compromise, because speakers from various regions would interpret it in their own way. He has himself incorporated this reduction in Cut Spelling (Upward, 1996).

Even H W Fowler (1958), usually a champion of the most ridiculous traditional spellings, generally went along with the reduction:
It seems desirable that ... all words in common enough use to have begun to waver between the double letter and the simple E ... should be written with the E alone...
It must be remembered, however, that Fowler was opposed to a general spelling reform:
English [spelling] had better ... not be revolutionized but amended in detail, here a little & there a little as absurdities become intolerable ... [op. cit. p554]
In Fowler's scheme of things, vowel length would not be consistently shown because the present system doesn't do it. So a few more ambiguous E's would pose no great problem.

With Upward's Cut Spelling the situation is not so clear. Only if Cut Spelling in more or less its present shape is to be the final stage of spelling reform could the reduced forms be considered as compromise spellings. Upward has never conceded that he holds to this view. In thoro-going systems of reform they could not possibly be, because 'long E' and 'short E' would have to be distinguished everywhere: no spelling reformer to my knowledge has wanted to write both dead and deed as *ded. (I hope no one ever will.) But this is precisely what pushing for reductions of AE and OE would imply. This cannot function as a compromise in the final stage of the orthography and intermediate use will only obscure the goal of a basically phonemic system.

No matter how much one might wish it, two or more spellings for at least certain words are inevitable in a thoro-going reform of English, altho at the present one form suffices for them because of the great ambiguity of a large number of symbols - especially vowel symbols - have in the traditional orthography. Thus, tomato (Klasik Nue Speling, tomaatoe or US tomaetoe), either (iedher or eedher), - even consonants, as in greasy (greezy or greesy), nephew (nevue or nefue), as well as a considerable number of words ending in -sia, -sian, -sial, etc (more on these later).

2. Acceptable compromises.

What legitimate 'compromises' could be built into an ideal spelling? One that virtually all modern reformers agree on is the retention of R in words like far, start, port, ladder, etc, despite the fact that all Australians, New Zealanders, many New Englanders, and many people in England and Wales do not pronounce it. Some American reformers might believe that the R must be written because it is 'correct'. But all respectable dictionaries admit that (if they do in fact indicate such R's in their pronunciation guide) readers may or may not pronounce the R according to their own dialect.

Now, it is perfectly reasonable to argue that everyone should be able to write his own dialect. Henry Sweet, the famous nineteenth century phonetician and possibly the prototype of Bernard Shaw's Professor Higgins, apparently would have taken such a position. In fact he proposed that in England, unpronounced R's should not be written so that we find faam 'farm', wəək 'work' in his ideal orthography.

But at present, such an extreme approach is generally rejected because it would tend to break up the English-language community. Publishers would find it exorbitant to publish the same books in many versions. As a matter of fact, critics of spelling reform almost always play up this difficulty.

3. Attempts at World English.

To keep the English-language community together, then, is a serious goal reformers must keep in mind. To do so, with as few regional variations as possible showing up, several compromises would have to be made - and, I think, could be made. The end result, an 'ideal' English spoken by no one natively, I call World English.

In a little known article by the American linguist Martin Joos (1960) a few interesting compromises are suggested. They deserve consideration by all reformers.

He starts out by establishing the necessity for compromise: 'there is no single standard of English pronunciation... Instead of ... a single ideal [more or less as in French], the English-speaking world has at least half a dozen' (p256). A decent orthography for English, Joos rightly insists, must do what the traditional spelling does: 'serve[ ] as an automatic translating machine between standards of English speech' (p256) - something like what traditional Chinese writing does for dialects so different they are probably separate languages.

Some proposals of reform do not meet this criterion, or do so only in part. For example, altho the great majority of native speakers of English (in Scotland, Ireland, Canada, and America) differentiate between which and witch, whale and wail, and the distinction is recognized in the OED, Cut Spelling ignores the existence of the WH/W contrast and permits only W (except in wher). Practically all schemes preserve a distinction between balm and bomb, but one American proposal (made by Abraham F Citron) lumps them together as bom and even writes fother for father.

The American Literacy Council (ALC) has proposed a scheme (formerly called American Spelling) that restricts itself to representing one variety of American English only. For example, it requires showing the palatalization of T, D, S, Z in unstressed syllables: -TU- and -DU- are to be written as -CHOO- (or -CHUR) and -JOO- (or -JUR). Thus, actual becomes akchooal; obituary, obichooerry; gradual, grajooal; individual, indivijooal; literature, literachur. Similarly sexual, issue become sexshooal, ishoo. No variation is allowed. But it is conceded that 'reformed British and Australian orthographies will [likely] differ from some of the above' (Rondthaler & Lias, 1986, p296).

In a way, the adoption of these forms goes against the ALC American Spelling strategy of keeping the appearance of words as close to their traditional shape as possible. For this reason, the TH/DH distinction is unfortunately ignored (thigh and thy are thi, wreath and wreathe both become reeth). Furthermore, (hard) C, Q and X are all preserved. And glory, historian keep their present spelling despite the fact that the majority of American dictionaries indicate the most common pronunciation that would be expected as gloery, histoerian.

Klasik Nue Speling (Ripman & Archer, 1948) the SSS's proposed orthography of 1948 [slightly adapted from Archer's version used on p3ff of this issue - Ed.], is more tolerant - it permits both glory and gloery - and also provides forms that could serve as compromise spellings. For example, it writes aktueal, obitueary, gradueal, individueal, literatuer, seksueal, isue (but also ishuu). If it incorporated a rule that unstressed -tue-, -due-, -sue-, etc, are pronounced either with a 'real' T, D, S, etc, or with 'palatalized' correspondences, CH, J, SH, we would have genuine compromise spellings that could be used for all dialects of English.

Walter Ripman, the formulator of Klasik Nue Speling, did not develop this compromise consistently. But in an even more thoro-going way, Joos did. He believed that the most elegant solution to a variety of problems would be to use J to mean a Y sound (as in German, Skandinavian, all Slavic languages that use the Roman alphabet, and the International Phonetic Alphabet). Then, CH, J, SH, ZH could be reinterpreted as TJ, DJ, SJ, ZJ, in line with the assimilations usually heard in meet you, did you, miss you, please you. Joos suggests that in unstressed or weakly stressed syllables, TI, DI, SI, ZI could be interpreted as either variants of CH, J, etc, or TY, DY, or TI, DI - depending on dialect. For example, Christian would be rewritten as Kristian (this is not his example) but interpreted variously as Kristian, Kristyan or Krischan (all of these forms are reported from modern RP). The example Joos does give is bestial which would not be rewritten but could be interpreted as bestial, bestyal, or even as beschal.

Joos goes further. He proposes that 'long U' as in use (ie, /ju:/) both be ius or iuz. But statue, mutual, individual, gradual would similarly become: statiu, miutiual, individiual, gradiual - with the understanding that a number of different pronunciations would be accommodated.

Doing something of the sort is necessary if only because of the great variety some classes of words exhibit. In Daniel Jones' English pronouncing Dictionary, negotiate, amnesia have 4 possible pronunciations each; and an(a)esthesia, 5; mutual 6; Asian 8; Polynesia(n) 10.

The case for compromise in these instances is strong but details of Joos's proposal are so unlikely to be accepted that they must be somewhat recast. The major problems are the introduction of J to mean Y (no matter how laudable), and the way palatalization is handled. We will probably have to stick with CH, J, SH, ZH. But with somewhat less elegance Klasik Nue Speling forms like bestial, Kristian, muetueal, individueal can preserve the compromise. This means isue, not ishuu, negoesiaet not negoshyaet - the Klasik Nue Speling variants cannot be followed strictly if we are to be consistent.

4. Maximum differentiation.

A few other areas of compromise can be suggested. Following the strategy of writing R's everywhere with the understanding that some speakers will drop them, we can set up the general strategy of writing the maximally differentiated form as the compromise. That means keeping the WH in when, where (which Klasik Nue Speling does, but not Cut Spelling). As well as (1) showing a difference between or versus oer in words like for vs four, horse vs hoarse (which Klasik Nue Speling permits as an option but ALC's American Spelling does not); (2) differentiating the vowels in fir, fern, fur, as is often done in the Standard English spoken in Scotland (one pamphlet issued by the SSS. - Braeking dhe spel [1942] - permitted an optional two-way distinction as indicated by the first edition of the OED, separating fur from fir/fern, both written with ER; Cut Spelling leaves the traditional spelling unchanged in these instances); (3) differentiating the vowels in aunt and ant, class and classic. This is no longer a popular distinction to make and recent SSS proposals have dropped it - along with too many other distinctions. Unfortunately, no consistent way of indicating this distinction occurs in the traditional spelling, but if we follow the general strategy outlined above, the distinction must be made, as Klasik Nue Speling does: aant vs ant, klaas vs klasik. Speakers who do not observe such a distinction would simply learn that there are two ways to write an A sound (as speakers who drop R's have to learn two spellings for AA, say, as in father and farther).

The most vexing problem of all involves the so-called 'shwi', not a separate sound comparable to the 'shwa', but a cover term for different phonemes used in various dialects: (4) for example, the final vowel variously written -Y, -EY, -IE, -I in words such as lady, money, hippie, taxi. In some dialects this is always a short I (the traditional pronunciation listed in pre-1961 dictionaries); in others it is a 'long E'. Some English people, speaking a democratized RP, use 'long E' at the end of a word but 'short I' before endings and in compounds such as ladies, taxis, anything. Aristocratic RP uses I always (and sometimes a very much lowered variety, approaching the E in pet). Americans (except in the south) tend always to use 'long E'. Klasik Nue Speling writes -Y at the end of words, but -I elsewhere - leaving the Americans out: an unfortunate omission to say the least. Rondthaler's American Spelling generally writes Y everywhere for 'shwi', which is a possible compromise but abandons international values for Y. Furthermore, its use is inconsistent for at least some words: beauty, beauties, beautify, where Y, I(E) represent a 'long E', are all written with Y (buety, buetyz, buetyfi) but so is beautiful (buetyful) where such a pronunciation is non-standard. Joos suggests another solution: where American English has 'long E' in unstressed syllables, write I; where it has 'short I', write E. This means that candied becomes kandid, but candid becomes kanded (and so also meret for merit, hored for horrid). Rondthaler's American Spelling would write candyd and candid. Klassic Nue Speling has kandid for both. I think another solution is called for: kandi'd or kandi.d vs kandid - a position I have previously proposed (see SSS Newsletter 1986 Spring J2 pp14-17).

But this paper has become too long for me to repeat my argument for solving the shwi problem. I can only hope that a case has been made for various compromises in an acceptable orthography for English as a world language.


Fowler, H W (1958) A dictionary of modern English usage, Oxford: Clarendon, p11.

Jones, D (14th edition, revised by a C Gimson) English Pronouncing Dictionary, London: J M Dent & Sons.

Joos, Martin (1960) Review of Axel Wijk Regularised English, in Language, 1960, pp250-62.

Langscape (1998): This survey was conducted in preparation of the SSS's response to Questionnaire 1 of the Langscape survey of contemporary English usage (based at Macquarie University, Sydney, Australia). The questionnaire appeared in English Today 53 (vol.14, no.1, Jan. 1998, p5), and the results in English Today 56 (vol.14, no.4 October 1998, pp6-12). [See J25 Langscape Survey.]

Ripman, W & Archer W (6th edition 1948, revised by Orton, H, et al. 1940) New Spelling, London: Sir Isaac Pitman & Sons, Ltd.

Rondthaler, E & Lias, E J (1986) Dictionary Of American Spelling, New York: The American Language Academy (now The American Literacy Council).

Upward, C (1996) Cut Spelling: a handbook to the simplification of written English by omission of redundant letters, Birmingham, UK: Simplified Spelling Society, p61, §A.1.1, and p91, §O.2.

Back to the top.