[Simplified Spelling Society Newsletter Autumn 1985 pp18-24. Later designated Journal 1]
[Chris Upward: see Journals, Newsletters, Leaflets, Media, Pamflet, Book, Papers.]

Cut Spelling as a First-Stage Reform.

Christopher Upward.

[Christopher Upward lectures in the Modern Languages Department at Aston University, Birmingham, and is on the Society's working party revising New Spelling, as well as being editor of the Newsletter.]


Spelling reformers must consider the practical implications of reform. Total reform without stages would risk undermining literacy, and, to avoid this, two criteria for a first stage are proposed: 1) it must be easily readable by adults without instruction; 2) children who are taught the first stage must be able to read t.o. To ensure these criteria are met, ease of reading must have priority over ease of writing, and the disruptive visual impact of changes must be minimized. Types of reform are then analysed for their visual impact, and finally, Cut Spelling is claimed by its nature to satisfy the above requirements uniquely well, and is introduced with accompanying exercises.

1.1 Overall Strategy.

To introduce a reformed orthography for English implies several major planning stages. Firstly, orthographers would need to propose an attractive, practical scheme and get it accepted. Secondly, there would need to be international co-ordination, at least by the English-speaking countries, or perhaps through the U.N., to ensure a single world standard for written English. Thirdly, each government would need to plan implementation of the reform in its own country. This article is however only concerned with the first part of the first of these major stages: the design of a reformed orthography.

1.2 Design and Implementation.

When designing a reformed orthography, one cannot ignore the practicalities of its implementation. Design and implementation inevitably interact: design is subject to practical constraints, but at the same time the proposed reform may require practical measures to be taken which go beyond questions of spelling as such. Thus a reform needing whole populations to return to school is clearly impractical, while the introduction of even a single new character would entail changes to printing equipment.

1.3. Orthographic Stages.

Some reformers earlier this century concerned themselves exclusively with designing what they hoped was a complete, coherent orthographic system, but gave little thought to the impact of its sudden introduction to a society already inured to a very different system. More recently the awareness has spread that reform must be introduced gradually, to ensure a smooth transition, but a full programme of suitable orthographic stages has yet to be devised. This paper attempts to establish some criteria for devising such a programme and then proposes Cut Spelling as a specific, far-reaching first stage that accords with these criteria. But to begin with we need to examine more closely why stages are needed at all.

2.1 Reform without Stages?

Two kinds of reform without stages are conceivable: one would tinker with details, without regard to any overall master-plan; and the other would aim at systematic wholesale restructuring of the orthography

2.2 Unplanned tinkering.

English spelling could of course be improved by just adapting some details, regularizing a phoneme here, rationalizing an outrageous spelling there. But the chances of achieving the supreme orthographic quality of consistency by this method would be small. Almost certainly, opportunities for a profound rationalization would be missed, and later changes might conflict with earlier ones. This is the criticism Axel Wijk made of the 1920 proposals of the American Simplified Spelling Board, though it is perhaps also ultimately the weakness of Wijk's own work, albeit on a much more scholarly level.

2.3 Systematic Reform without Stages.

Could a comprehensive, systematic reform be introduced at a stroke, without stages? The Simplified Spelling Society's New Spelling is essentially such an attempt at the near-total regularization of English spelling, but its authors did not examine the likely effect of its introduction on different kinds of reader. Let us now do so.

2.4 Test-text for Reformed Orthographies.

The following verse (with title) contains at least one example of each letter of the alphabet and each RP phoneme, and so demonstrates in concise fashion the visual effect of the orthography it uses. In t.o. it reads:

Fuzzy-opaque Orthographical Visions.
There was a poor boy couldn't spell
Half the words in our language too well.
His teachers thought: "Brain-sick!"

Mum and Dad hoped: "Dyslexic?"

Yet the child rashly jeered: "What the hell!"

2.5 New Spelling (NS) Transcription.
Transcribed into New Spelling (1948) the verse reads:

Fuzy-oepaek Orthografikal Vizhonz
Dhaer woz a puur boi koodnt spet
Haaf dhe wurdz in our langgwej tuu wel.
Hiz teecherz thaut: "Braen-sik!"

Mum and Dad hoept: "Disieksik?"

Yet dhe chield rashly jeerd: "Whot dhe hel!"

2.6 Criteria of Acceptability.

How can we judge if a reformed orthography is suitable for immediate introduction? We must consider the pattern of literacy after reform, when children would be taught the new spelling in school, but most adults, especially the less well educated, would not have received formal instruction. So the first criterion of feasibility has to be: could all adults read the new system without instruction? But equally important is surely the criterion: can children who have only been taught the new orthography read t.o.? For obvious reasons, failure to meet either of these conditions could mean a serious breakdown of written communication in society.

2.7 Does NS meet these Criteria?

Adults. Well-educated adults can struggle through an NS text, because (except for the digraphs DH/ZH/AE/UU) the sound-symbol correspondences are more or less familiar from t.o. and they are regular; and fluent reading would come with practice. Whether the less well-educated could or would master reading in NS must be more doubtful.

Children. More serious however is whether children educated in NS would still be able to read t.o. The irregularity of t.o. would present enormous difficulties to children who had been taught only the regular forms of NS (and to have to teach t.o. as well would defeat the object of the exercise), the letters Q and X would be unknown, and almost 70% of words (according to Wijk) would be spelled differently. Even simple words in the limerick like there/ was/ though/ hoped/ vision might prove obscure. NS thus appears not to meet the criteria for introduction proposed in paragraph 2.6.

3.1 Reform by Stages.

Stages are needed, therefore, because any total reform will be visually so different from t.o. as to be incompatible with it from the reader's point of view. An idea for stages to overcome this problem is put forward in Harry Lindgren's Spelling Reform, A New Approach. His final system, Phonetic B, is even more remote from t.o. than NS is, and he suggests there should be some 50 small stages leading up to it. The need for stages here will appear inescapable from the transcription of the limerick.

3.2 Phonetic A/B Transcription.

For lack of some of the diacritical characters used by Phonetic B, the following transcription also uses some forms from Harry Lindgren's intermediate system, Phonetic A, and is therefore somewhat less radical than full Phonetic B.

Fuzi-ôpék Orthografikl Vizhnz
Dher wz 'púr bó kudnt spel
Hàf dhwurdz in âr langgwj tû wel.
Hiz tìcrz thòt: "Brén-sik!"

Mum 'n Dad hôpt: "Disleksik?"

Yet dhcàld rashli jírd: "Wot dhhel!"

Without studying Fonetic A/B, even cryptologists might hesitate at this. It looks like a foreign language, uses apostrophe for shwa and diacritics instead of digraphs, spells vowels phonemically, merges articles with nouns, and only 4/33 words appear as in t.o. The system has great virtues, but could not be introduced without stages.

3.3 SR1, a Phonemic Stage.

As a first stage towards Phonetic B, Harry Lindgren suggests the sound /e/ should always be written plain E (head thus becomes hed) - a step he labels SR1. But SR1 can be criticized on two counts. Firstly its scope is very limited, as it only affects 1 out of 44 phonemes. Secondly it adopts an auditory, or phonemic, rather than a visual approach: though it writes /e/ consistently as E, SR1 does not mean the letter E is always pronounced /e/. SR1 tells you how to spell a sound, not how to pronounce a letter; in other words, it helps the writer rather than the reader.

4.1 Reading as a Prime Social Function.

This leads us to propose a central reform-principle that the needs of readers must have priority over those of writers. In a literate society, reading is a universal, indispensable activity: we are surrounded by the written word, on TV, on signs and hoardings, on public transport, in shops, in instructions, at work, and in our dealings with government and the law. Unlike what we write, we have no choice over much of what we read - nor over its spelling. We are obliged to read whatever text we are face, as best we can. It is therefore crucial that readers should not be defeated by strange spellings.

4.2 Reading as a Psychological Function.

Psychologically too reading and writing differ. Reading entails visually recognising groups of letters; while writing depends on recollecting the sounds of words (though there is also an auditory element in reading and a visual element in writing). If for a smooth transition between old and new we have to give priority to ease of reading, then the stages we propose for our reform must be based on visual rather than auditory criteria.

4.3 Visual Disruption in SR1.

When we look at the SR1 changes, rather than listening to them, we find that the simple regularization of the phoneme /e/ has some far from simple visual consequences. Admittedly, most words in the SR1 list drop A (head --> hed) when it misleads as to pronunciation, and since no unaccustomed letters are inserted, visual disruption with these words is minimal (SR1 ignores the same pattern in the 14 words of the earn type, though). However SR1 also affects some common words whose forms differ jarringly from t.o. They all substitute letters, either one-for-one (as A/U = /e/ = E : Any/ bUry = Eny/ bEry ), or one-for-two (as Al/AY/A ... E= /e/= E: agAIn/ sAYs/ AtE = agEn/ sEs/ Et ). One also notes that the /e/ pronunciation is not necessarily standard in these words, especially not in ate, which Americans pronounce /eit/.

4.4 Cutting to Reduce Visual Disruption.

If the visual, rather than the phonemic, effect is the essential criterion for a first stage, clearly the regular pattern of cutting EA to E is visually far less disturbing than the sundry substitutions made in et/ agenst/ ses etc. Indeed in general we can say that omission of superfluous letters is less disruptive to the familiar look of words than is inserting or substituting letters: cutting leaves the essential phonic skeleton of words untouched, and indeed in longer words may even pass unnoticed.

4.5 Arbitrariness of Phonemic Stages.

Another aspect of SR1 is the seeming arbitrariness of the foneme chosen. Almost all RP phonemes require attention, so why pick /e/? It may seem simple to deal with, but we have seen the visual disruption it causes.

4.6 Disruptive Long Vowels.

If one criticism of SR1 is its limited effect, regularizing long vowels transforms spellings. The problem here arises from precisely that fact: because so many words are changed, a visual upheaval results. Again, the choice of the long values of AEIOU is rather arbitrary; and we have to ask whether children taught an orthography with regularized long vowels could then read t.o.

5.1 Degree of Change from T.O.

We have arrived at the idea of minimal visual disruption as a criterion for first stage reform. Let us apply now this concept to different orthographic systems, and examine how radically each diverges from t.o.

5.2 Computer-readable Codes.

Computer-readable codes are perhaps the most remote from t.o. The opening words in a t.o. dictionary might be:
a aardvark aardwolf aback abacus abandon;

for the computer these could be compacted as
a 1ardvark 4wolf 1back 3cus 3ndon.

Or the word itself could be spelled itse because no other word starts with these 4 letters and no ambiguity arises. Such forms may be efficient for data-processng, but they are not readily assimilated by the human reader, since the brain cannot store for ready recall all the information required to interpret such codes.

5.3 Redesigned Alphabets.

One step nearer t.o. might be the most efficient conceivable human-readable system, perhaps not based on the Roman alphabet at all, but on characters designed specifically for easy reading and writing. Such systems may interest the typographical designer, but the upheaval their introduction would entail rules out their consideration for first stage reform.

5.4 Shorthands.

Closer to t.o. again are shorthands designed for fast writing rather than easy reading. Some, like Pitmans, do not use the Roman alphabet, but others, like Speedwriting, do. Secretarial experience suggests shorthands are inherently reader- unfriendly, but a systematic examination of their potential for normal use might be illuminating.

5.5 Non-phonographic Abbreviation.

The same doubt about reader-friendliness arises with systems of abbreviation whose forms do not represent phonemes, but have to be learnt individually. T.o. uses many such, with acronyms like USA and symbols like %. Dutton Speedwords, though using the Roman alphabet, builds up the vocabulary from 493 abbreviations for common root-words; but its abbreviations do not always derive from English - the personal pronoun 'you' for instance is written 'v', from the French 'vous'. Professor Citron's 100 Speed Words on the other hand all have some affinity to t.o., but many nevertheless abandon the phonographic principle and so constitute a code.

5.6 Phonographic Systems with Diacritics.

The systems listed below do adhere to the phonographic principle, but the closer to t.o. they get the more erratic become the sound-symbol correspondences. Phonetic B, which replaces the clumsy, ambiguous device of digraphs with acute, grave and circumflex accents, is highly efficient and economical (the great advantage of diacritics) but their disruptive visual impact rules them out as a first stage: they make English look more like Czech.

5.7 Phonographic Systems with Digraphs.

NS on th other hand is neither phonetic nor economical. It uses digraphs for long vowels, diphthongs and consonants where the alphabet lacks an appropriate single symbol, and even expands X to KS, which is uneconomical as well as visually disruptive. Though less revolutionary than Harry Lindgren's diacritics, NS is still disturbingly different from t.o., and clumsy to boot.

5.8 Tidying-up Systems.

Axel Wijk's Regularized Spelling essentially constitutes a tidying-up system. It changes the t.o. spelling of far fewer words than NS (29% instead of 69%); and to that extent it is visually less disruptive. It tries to ensure that a given spelling always represents the same phoneme; but it does not ensure the reverse, that a given phoneme always has the same spelling. Even so, the system requires a plethora of basically arbitrary rules, and one must ask: if one is going to go to such trouble, why not devise a fully consistent system? But if the system is unconvincing, the book itself is a treasury of t.o. sound-spelling correspondences of great value to orthographers for reference.

5.9 Phoneme-based Reforms.

We have analysed two such proposals, SR1 and long vowels, in paragraphs 3.3 and 4.3-4.6 above. Their problem lies in the paradox that the more comprehensive they are, the greater the visual disruption they cause, and vice versa.

5.10 Reform by Visual Pattern.

Lastly, before t.o. itself, we come to Cut Spelling, which is based on phonemically-defined visual patterns. It has just 3 main patterns (cf. the 50 SRs, or Axel Wijk's innumerable case-studies), and by omitting rather than actually changing letters, it ensures that the essential phonemic features of t.o. spelling are preserved, while visual disruption is minimized. Yet the gain in both regularity and economy is very substantial indeed.

6.1 Cut Spelling: its Rationale.

Cut Spelling (CS) aims at maximum advantage by way of regularity and economy, combined with minimum disadvantage by way of visual disruption. A basic feature is that no letter that is pronounced, even if wrongly, is omitted; in this way the fundamental, recognizable framework or phonemic skeleton of each word in t.o. is retained. The word any (which SR1 changes to eny ), for instance, is unchanged: because the A is pronounced, it is kept. If strictly applied, CS does not actually change any letters at all, but only omits them. Nevertheless it is worth considering whether some letter-changes can perhaps be profitably combined with CS. CS should meet the two key criteria previously set out: adults can read it uninstructed, and children taught only CS should still be able to read t.o., since it contains all the CS letters.

6.2 Rules of Cutting.

CS has 3 main rules. Rules 1 and 2 (omit silent letters, simplify doubled consonants) are commonplace in most reform proposals. Rule 3 (non-spelling of post-accentual shwa) has a broad application to an intractable t.o. problem. CS cuts only post-accentual shwa because in fluent reading it is primarily the early letters in words that trigger recognition, while later letters (except for the final one) have the lowest visual prominence. Not spelling post-accentual shwa has the additional advantage of indicating the stress-pattern in words more clearly: if the noun present is cut to presnt, the second syllable cannot possibly be stressed, whereas the spelling of the verb present shows the second syllable may be stressed.

6.3 Application of CS to T.o.

The 3 rules may be formally stated as follows:
1. omit letters having no bearing on pronunciation;
2. simplify doubled consonants;
3.1 do not spell post-accentual shwa before L/M/N/R & -BLE suffixes (on the model of syllabic L/M in appl(e)/ spasm/ rhythm); and
3.2 do not spell the unstressed vowel in -D/-S/-ST inflections (on the model of hate+D, hale+S, late+ST ).

These rules produce forms as follows:

Rule 1
A hed ern coco
C sience
B lam det
D wensday
E edg ampl com shon imagin ar loos delegat liv valu ew; hart siv
G naw foren eit
H eir onest scool gost rythm thru wen wy
K nave nife noledg
N condem
P seudo sycology receit
T hasen wisl cach
W rite windo
NB No cut in: comB, to delegatE, siGn, acHe, shoWn.

I frend juce receve
L shud samn haf
O peple choclat tuch colnl
S iland
U bild sholder tho

Rule 2.
BB eb pebl abreviat
DD ad padl adict
GG eg gigl agravate
LL bel alow filet aleviate
NN in winr anul
(QQ=)CQU aqit aqire
SS fus tasl asembl
(XX=)XC exept
CC/CK pik pikl aclaim
FF snif bafl aford
(JJ=)DJ ajust ajectiv
MM hamr imediat
PP apl aply
RR er wory iritate
TT batl atemt
ZZ buz puzl
NB - No cut in aCCept, etc as both Cs are pronounced.
- Ambiguity in: hoping/hopping, duly/dully,etc.

Rule 3.1
L principl principly hovl fosl petrl usefl dificlt
M madm systm victm fathm conundrm autm
N hoolign importnt importnce beatn dependnt dependnce dependncy raisn curtn suspicn cushn pasn informatn lemn
R burglr standrd boundry teachr bitrly modrn lotry amatr authr histry vigr murmur figr martr
BL pasbl posbl
NB - Repeated consonants arise in such cases as: maximm eminnt wandrr Febrry probbl

- Possible non-omission after vowels, L and R.

Rule 3.2
(Regularize tense and number inflections of verbs and nouns, and the superlative inflection of adjectives, by only ever adding -D/-S/-ST)
+S archs teachs hedgs dodgs bushs pushs axs fixs breezs freezs
+D heard
+ST richst lushst
NB Forms containing doubled and/or repeated consonants such as: masses gases losses fusses buses added needed become: mass gass loss fuss buss add needd.

6.4 Test-text transcribed into CS.

Fuzy-Opaqe Orthograficl Visns
Ther was a poor boy cudnt spel
Haf th words in our language too wel.
His teachrs thot: "Brain-sik!"

Mum and Dad hoped: "Dyslexic?"

Yet th child rashly jeerd: "Wat th hel!"

(This CS transcription contains 152 letters; t.o. has 13.6% and NS 9% more, but Phonetic B 16% less.)

6.5 Degree of Visual Disruption.

Just as Section 5 arranged various orthographies in order of remoteness from t.o., so the following hierarchy of cutting-patterns in CS shows the degree of visual shock each occasions, relative to t.o., the first being the least disruptive and the last the most disruptive.

1. Post-accentual shwa in polysyllabic words (Rule 3): meteorologicl vice-presidnt
2. Simplified consonants in polysyllabic words (Rule 2): acomodating paralel comitee
3. Silent letters in mid-word (Rule 1): honymoon iland hauty cardbord gingerbred
4. Post-accentual shwa in short words (Rule 3): chapl randm pistn presnt undr
5. Silent final letters (Rule 1): tho handl imagin theatr goos minut elusiv ew wido
6. Silent letter and post-accentual shwa omitted (Rules 1 & 3): autm
7. Consonant simplified, silent letter omitted (Rules 1 & 2): rubd lakd snifd lagd feld jamd pland dropd rord hisd buzd buzd folo wino aro
8. Consonant simplified, post-accentual shwa omitted (Rules 2 & 3): batl botm bitn butr
9. Omission of post-accentual shwa highlights misleading consonant-value: suspicn visn permisn informatn crucifixn
10 Silent initial letter omitted (Rule 1): naw neel seudonym rong
11 Various combined cuts: y
If some straightforward letter-substitutions are incorporated in CS, then the following patterns might arise:
12 GH/PH = /f/ = F: laf cof fonetic autograf nefew
13 DGE/GE = J: ej juj bajr larj jinjr.

6.5 Simplicity of CS.

As well as satisfying the two criteria of first stage feasibility, a great attraction of CS is its simplicity. Firstly, it is simple to use, because it regularizes some of the most perplexing inconsistencies of t.o., such as single/double consonants and the spelling of post-accentual shwa, and it shortens words - a benefit that should not be underestimated. Secondly, it is relatively simple to apply to t.o., since cutting redundant letters is simpler than inserting new letters (insertion requires an additional set of rules).

6.6 Some difficulties of CS.

Cutting does not however always just involve mechanically applying 3 rules, since snags are encountered in some orthographic or phonotactic contexts. Consider for instance the effect of omitting the capitalized letter in the following words: for Rule 1, siGn (becomes sin); for Rule 2, hopPing (becomes hoping); for Rule 3, fatAl (appears to rhyme with cattle which is cut to catl). In these cases omission is problematic, as will also be seen in the exercises at the end of this paper. These cases are however only snags for users of CS proficient in t.o.; children taught CS instead of t.o. would not perceive them as snags.

6.7 Implementation scenario.

How could the implementation of CS be envisaged? The two key steps would be its introduction in schools in place of t.o., and a few years later its use by the publishing industry, particularly newspapers. Adults would receive guidelines for its use, but the only adults who would need actually to become proficient in it would be teachers and those professionally concerned with the production of printed text for public consumption. Such a relatively straightforward scenario depends on that characteristic of CS which other reform schemes do not share to anything like the same degree: the mutual intelligibility of CS and t.o. - which is another way of saying that CS meets the two essential criteria of feasibility.

See Part 2, Reading practice, Exercises and Key.
Back to the top.