The Bernard Shaw Alphabet

by P. A. D. MacCarthy

[extracted from Haas, W., Alphabets for English, Manchester University Press (1969)]

The Bernard Shaw Alphabet (or Shavian, as it has come to be called), represents an attempt to alter the mode of writing English—and ultimately perhaps any language—that is almost diametrically opposed to all the proposals for spelling reform (in its widest sense) that have been discussed up till the present in this series,—in particular the complete oppo­site of the system called New Spelling that I myself spoke about here not so long ago.

New Spelling—let me remind you—involved (1) retaining the traditional roman alphabet intact (apart from the elimina­tion or non-use of a couple of redundant letters); (2) the rearrangement of the roman letters to the best advantage—fairly ruthlessly from the standpoint of the present-day user, yet disturbing current spellings as little as possible consistent with a thoroughgoing, systematic reform—; and (3) the liberal use of combinations of two letters to stand for single sounds, this being a necessary evil resulting from the deficiency in number of roman letters for writing any language having more than 26 distinctive sounds.

Both the other previous speakers in this series also put forward schemes based upon the roman alphabet:—Dr. Wijk's proposals ('Regularized English') involved even less departure from traditional usage than does New Spelling—they would leave the great bulk of current spellings intact. Sir James Pitman's Initial Teaching Alphabet, conceived as a method for speeding the process of learning to read, makes concessions to present spelling with the express aim of facilitating the transition back to it (once the reading skill as such has been acquired). Pitman's i.t.a. is, apart from that, based evidently on New Spelling, but, in addition, employs a number of newly-designed ligatured forms, resulting in most cases in strictly non-roman shapes (thus avoiding the New Spelling device of juxtaposing two familiar letters to form digraphs).

Now Shavian is different in kind from all of the foregoing: Shavian is a new alphabet. No one could read Shavian at sight simply for having learnt to read roman. Before one can read Shavian the value of each of the completely new symbols has to be learnt, and before reading from Shavian fluently the skill of scanning and interpreting the sequences of those symbols, and of recognizing the shapes of whole words from previous familiarity with those shapes, has to be acquired over a period of time. Shavian therefore represents a radical departure from tradition (though there have of course been other proposed radical departures in the past), and I should like first to deal with the motivation in the mind of its inspirer, George Bernard Shaw, so that we may more clearly understand what he was trying to bring about, and why.

As a professional writer, Shaw was struck by the wastage involved in English spelling: the waste of time spent writing by hand many silent or redundant letters (Shaw himself always wrote in Pitman's Shorthand); the waste of materials (paper, ink, etc.) used in writing those letters; the waste of time spent by every child learning to spell English words with their inconsistencies and irregularities, time spent by every teacher of children, and by every foreign learner, and teacher, of English anywhere in the world. He campaigned frequently and vigorously during his lifetime to draw people's attention to the significance of this state of affairs, and to break down their apathetic conservatism, if not their championing adherence to what he termed 'Johnsonese', for he ascribed (no doubt correctly) the standardization of our present spelling largely to the publication and subsequent wide use of Dr. Johnson's Dictionary.

Realizing that there were in any case not enough roman letters to go round, he set himself firmly against any attempt at rearrangement of those letters (this is what he called 'spelling reform'), and he had additional reasons for so doing: the impression of near-illiteracy that he claimed was made on educated people by rearranged, 'phonetic' spellings, and the likelihood that such spellings might have undesirable, possibly obscene, associations.

Towards the end of his life this campaign seems to have been intensified, in long letters to The Times newspaper and elsewhere, since he had formed the idea of leaving his fortune to promote a new alphabet; but many years before, in his Preface to Pygmalion—a play whose principal theme was the pronunciation of English—we find him alluding to English in its written form. I quote: 'The English have no respect for their language, and will not teach their children to speak it. They cannot write it, because they have nothing to write it with but an old foreign alphabet of which only the consonants, and not all of them, have any agreed speech value'.

I mentioned a moment ago that Shaw always wrote his works using Pitman's Shorthand (his secretary then trans­cribed this into longhand for revision, and the printer). It might perhaps be thought that here surely was the answer to Shaw's quest for a new script for English, in other words that a system of shorthand (not necessarily Pitman's) would achieve the economy—at least of time—that was wanted. But this is far from being the case. In the first place, the all-purpose utility of an orthography cannot be achieved by any system of writing based upon hand movements and designed with considerations of sheer speed primarily— almost exclusively—in mind. Technical problems arising for the printer would alone rule out sequences of shorthand strokes where individual letters have no separate identity.

It is notable that existing scripts based on a cursive style, such as Arabic, lend themselves extremely awkwardly to handling by the modern technologies of printing and typing, designed for handling separate, discrete letter shapes. Because the majority of Arabic letters have to be joined to a following one in the same word (a few may not be so joined), the typewriter and printing press are required to simulate this joining up of successive letters, by bringing the final stroke of one letter and the beginning stroke of the following one extremely close together, so that they appear to touch. And because no overlapping of forms can be obtained from mechanical devices that move forward a notch at a time for a letter of whatever shape to be placed in the given space, awkward distortions of letters are unavoidable, and there is a wastefully high proportion of white space to black letter on the printed page, particularly since the large number of essential diacritical dots above and below each line of writing makes it necessary to set successive lines a considerable distance apart in the interests of legibility. Shorthand too is notoriously wasteful of space, the spidery outlines having a vertical as well as a horizontal sprawl. Yet further objections to a shorthand can be made: that it tends to become illegible to all save the writer, and even to him or (more often) her after any lapse of time before the eventual intended retranscription into longhand; that the sought-after speed itself is not attained without such drastic abbreviations and omissions that ambiguities constantly arise or illegibility even results.

Shaw himself realized the fundamental unsuitability of a shorthand as an orthography, though he did on one occasion (Preface to The Miraculous Birth of Language by R. A. Wilson) hold up, as a model of what could be done, Henry Sweet's Current Shorthand. He went on to point out that the basic strokes devised by Sweet would have suited very well, if only Sweet had not himself gone on to complicate and bedevil matters by seeking to make his system suitable for verbatim reporting, with its inevitable contractions and innumerable conventions. (It is, however, interesting to realize that it was Shaw's racy description of Sweet's Current Shorthand contained in the above Preface—"A straight line, written with a single stroke of the pen, can represent four different consonants by varying its length and position. Put a hook at the top of it, and you have four more consonants. Put a hook at the lower end, and . . . etc."—that provided the inspiration for the designer of what was eventually to become the "Shavian" of today, as it may well have done also for many of the other competitors in the Bernard Shaw Alphabet Competition. But I anticipate.)

The events following upon Shaw's death in 1950 must be fairly well known to Shaw-lovers and spelling reformers alike, so I think I need only summarize them here, and recall how Shaw, in his Will, appointed the Public Trustee to be his executor and trustee, and imposed on him the duty of seeking and publishing an alphabet of at least forty letters, more efficient than the existing one of twenty-six letters, so as to enable 'the said language (English) to be written without indicating single sounds by groups of letters or by diacritical marks'; this alphabet he termed 'Proposed British Alpha­bet'; how Shaw's residuary estate was directed by him to be held for a period in certain trusts for this purpose, but this was declared by a Judge of the High Court to be invalid in law. The Public Trustee appealed from the decision, and by way of compromise the residuary legatees (the British Museum, the Royal Academy of Dramatic Art, and the National Gallery of Ireland) eventually agreed to pay a certain sum to the Public Trustee for the promotion of this project of a new alphabet. It was indeed a modest sum in relation to Shaw's estate as a whole at that time, and much more so since that estate began to be augmented by the massive royalties that have been coming in since the appear­ance of My Fair Lady. However, it was the best that could be managed—and it has been instrumental in the creation of Shavian. Not until 1957 were all the legal complications out of the way, and the field cleared for the Public Trustee to inaugurate the Alphabet Competition by offering a prize of £500 for the best design of a new alphabet complying with the provisions of Shaw's Will.

Technical advice was available to the Public Trustee from a panel of three outside specialists, who also acted as judges for the Competition. As one of those judges I had the interesting experience of examining each of the 467 entries that came in during 1958. By the end of the following year the work of judging was completed, and the Public Trustee announced that the prize money would be divided equally between four competitors, none of whom had submitted a design outstanding enough to be accepted as it stood, i.e. to be the 'Proposed British Alphabet', for use in the subsequent transliteration of Shaw's play Androcles and the Lion, as stipulated in his Will.

Now, clearly, that transliteration could not possibly be made until there was an alphabet to transliterate into. So a problem still remained. This the Public Trustee sought to resolve by inviting the four prizewinners to work further on their designs with a view to arriving at the best possible solution. This did not mean (as on the surface might have appeared to be the case) that the best features of each en­trant's work could be extracted and somehow fused into the ideal alphabet: a glance at the individual entries would be sufficient to make it plain that the basic conceptions of each were so utterly different that a combination of any features drawn from more than one of them could only have created a hotch-potch that would have been unsatisfactory from practically every point of view, could not possibly have produced an aesthetically pleasing and harmonious whole, and would have been almost certainly inferior to each of the prizewinning entries considered separately. What in fact happened was that each of the prizewinners was encouraged to do further work on his submitted design in the light of criticism and comments by the judges. I was asked to conduct any ensuing correspondence, and to begin by letting each entrant know what we thought in detail of his entry. Now we, in the course of our judging, had had the opportunity (and, I must say for myself, the extremely interesting experience) not only of inspecting several hundred newly-designed original alphabets, but of forming each his own opinion, and our joint opinions, based on the experience so gained as to the basic criteria for a script suitable for the transliteration—the minimum requirements, as it were, that any proposed alphabet should conform to for complete acceptability. Just because no individual competitor had met all these demands, the prize-money had not been awarded to one outstandingly successful design.

The response of the four competitors was varied; one dropped out immediately and took no further part in the proceedings; one submitted several completely fresh designs in a short space of time; one entered into very full discus­sion with me, in the light of which he set about modifying his design. This last was Mr. Kingsley Read, whose final re-cast version was eventually re-submitted to all the judges and approved for use in the transliteration, thus becoming 'Shavian', the Bernard Shaw Alphabet. The extent of the re-casting may be judged from the fact that only about ten characters out of the total of 48 reappeared unchanged, and of these not more than two or three retain their original value, i.e. stand for the sound they stood for previously. Yet the overall appearance of the design is so similar as to be un­mistakably the work of the same hand, and quite unlike all other entries. It may be of interest if I mention the main feature in Mr. Read's competition entry that the judges considered unsatisfactory and that he later eliminated success­fully: it was that certain letter shapes were used unchanged for different sounds except for being placed in different positions in relation to other letters; for instance a given shape might be placed high in relation to the line of writing to stand for one sound, and the same shape placed low in relation to the line to stand for some other sound. This would mean, of course, that the form itself could never be interpreted unambiguously in isolation, i.e. without other letters or at least some essential guide line or rule from which to judge its relative position. When this disadvantage was brought to the designer's attention, he found himself in­volved in the radical re-casting described above, but even­tually came up with the masterly solution we see today.

In executing the transliteration, I had a number of deci­sions of my own to make, notably as to the treatment of the letter R, the representation of syllabic consonants, of strong and weak forms, and of course the spelling of a number of doubtful words. This included the problem of many words having acceptable alternative pronunciations, even without going outside the general type of English that Shaw himself had stipulated, namely that spoken by His Majesty King George V. This the judges were able to interpret as indi­cating a pronunciation widely known to linguists and phone­ticians as RP. I decided too that though Americans might favour a certain number of spelling differences, to fit better their own distribution of sounds (their, or anyone's, different manner of pronouncing English sounds as such being irrelevant to the orthography), nevertheless the proper way to tran­scribe this, the first publication in Shavian, emanating from Britain and moreover conforming to Shaw's prescription, was to transliterate the words in accordance with the conventions of the new system and in line generally with the pronunci­ations likely to be heard on the English stage.

Here now is a specimen of Shavian printing—the second part of the Gettysburg Address by Abraham Lincoln. There follow reproductions of the two sides of the card in­serted in each copy of Androcles and the Lion, Shaw Alphabet Edition, (Penguin Books Ltd.), on which appear classified lists of the Shavian letters, with keywords that serve also as the actual names of those letters, by the help of which the connected text may be deciphered. Alternatively, the New Spelling version of the Address (p. 118) may be consulted for this purpose.

Let me draw your attention to some of the salient features of the script.

Shavian is written, and read, from left to right across the page or other writing surface, in lines which run from top to bottom of the page—just as they do in current roman usage. There is of course nothing inevitable or universal about our present way of doing these things,—and a certain number of the competition entries did propose a departure from tradition in this very matter; nevertheless the winning entry un­doubtedly commended itself to the judges on the grounds (among many others) that in it, established practice as to the direction of the writing was to be retained.

A line of writing consists of a sequence of separate letter shapes (apart from ligatured forms, see below), and these are grouped by words, i.e. with a blank space to mark off word division. In this respect too, as also in matters of punctu­ation generally, current usage remains unaltered.

Letters are of three sizes in the vertical dimension:— Shorts, corresponding to lower-case sorts of x—or ribbon— height; Tails, that ascend above the ribbon while standing on the same base; and Deeps, that fall from the top level of the ribbon to below its base. In other words, here again there is little departure from the present state of affairs in printed roman (some cursive roman forms, e.g. handwritten f may both ascend and descend from the ribbon). The proportion of ascenders and descenders to x-height letters has been arranged to be about the same as in roman, but ascenders and descenders are more equally distributed than now as between one another (there being in roman about six times as many ascending letters as descending letters on the printed page). These changes have been intentionally introduced in the knowledge that in normal rapid reading, words are recognized as wholes, and that it is visible differ­ences in the external outline or silhouette of words that are mainly instrumental in enabling the eye rapidly to dis­criminate between different words and so to grasp their meaning. Long sequences of x-height letters tend of course towards uniformity of both top and bottom outline, but sequences of letters standing on the base of the ribbon, i.e. those of x-height together with those having ascenders, (which would produce uniformity of the bottom outline) are as a result of this alteration less likely to occur in Shavian than in roman. It may however be noted that the Shavian Tails and Deeps do not depend for identification solely on the fact of their having ascenders and descenders respectively: they are already adequately differentiated by their intrinsic shapes. So it would be perfectly feasible, and might well be found desirable for certain purposes, e.g. display work, to use founts of letters all of one size—there being in any case no capitals.

This brings me to the first point to be mentioned thus far, embodying a notable divergence from current roman usage (aside, that is, from the forms and general appearance of the Shavian letters themselves); there is to be no difference of shape (or size or thickness, though these would be optional) between initial and other letters, thus eliminating one of the major objections to current roman usage: the fact that the 26 letters have so many alternative shapes—for capitals and lower-case, not to mention cursive forms. To meet the occasional need for an equivalent to capitalization, a raised 'namer' dot is available in Shavian, e.g. for proper names, at least at their first occurrence in a text, or where obscurity might otherwise result.

As regards handwritten Shavian, the letters have deliber­ately been designed so as to discourage the cursive writing of linking strokes not forming part of a letter, thus eliminating a source of much ambiguity and illegibility in connected writing. If adjacent letters lend themselves to joining with­out extra linking strokes, they may well be joined (as the printed Shavian ligatures are joined), but not otherwise, the inevitable upward movements of the hand between letters being made off the writing surface (i.e. in the air) instead of in contact with it, with, it is believed, little if any overall slowing down of hand movements, once the somewhat different technique of writing has been acquired. Moreover much thought was devoted to ensuring that the actual shapes chosen for letters should, in the interest of legibility, not lend themselves to confusion with one another when written at speed, or to distortion from their basic shape, leading in the long run to divergences arising between printed (or typed) and handwritten forms—which is of course what has happened in roman, thus giving rise to the multiplicity of forms referred to above. Even supposing it were found in practice (and this is by no means proved) that writers tended by the nature of the script to form slightly more deliberate strokes, this would be very much to the advantage of the Shavian reader in ease of deciphering.

In fact, of course, overall writing time for the equivalent text would be reduced (given comparable familiarity with the script and its writing technique) on a number of counts: firstly, the 'silent' letters of present roman would naturally not be written at all—about one letter in six is now redun­dant; secondly, fewer strokes per letter are used in Shavian, so more letters can be written in a given time, or at the maximum speed at which an individual hand can form successive strokes. It may be noticed how simple are the actual shapes of the letters, a high proportion consisting of just a single stroke, and only one requiring as many as two down-strokes and more than one abrupt change of direction of pen on paper. Thirdly, much attention has been given to the allocation of the simplest available shapes to the most frequently occurring sounds. Fourthly, single-letter abbrevi­ation for the four commonest English words (the, of, and, to) are standard Shavian—and a number of other word-signs or reduced forms may well be used besides.

Eight of the printer's shorts shown in Specimen 2 are in fact compounds, i.e. ligatures, six of these being vowel + r in some commonly occurring cases, including one vowel that is itself already made up of a combination of two others. These compounds are to be written whether or not they be indicative of an individual writer's or speaker's use of r as distributed in his natural speech, thus enabling the Shavian spellings used by Londoners, Scots, Americans, etc., to resemble one another in respect of r—as they do in the standard orthography today. As to whether any alternative spellings would be available to speakers of different kinds of English (involving different distribution of their sounds), the answer is that they would, but I will not here attempt to go into details, and will only say that it would clearly be in the interest of a new world orthography for English that variation between writer and writer should be kept to a minimum. The two remaining compound sorts are ligatures of two other juxtaposed letters, which, as with the vowel + r liga­tures, may be looked upon as instances in print of the permissible easy linking of handwritten forms referred to above. However, on the Shavian typewriter keyboard (which is now available to the public) these ligatured forms are not found each on their own key, but are to be typed with the separate keys for the two letters concerned. This is partly to economize keys, but also because a closer approximation to equal width of all letters, on the equal-space typewriter, gives a better appearance to the page of typescript than would otherwise be the case. Copies of Shaw-Script, the quarterly periodical entirely produced in Shavian, from the Shavian typewriter keyboard, may be inspected in order to judge the effect.

Further study of the Shavian alphabet and of the relation­ship of letters to the sounds they stand for will make it plain that this is a coherent system, on a carefully considered plan embodying many distinct (and sometimes mutually conflic­ting) points, which has at times been obliged by the nature of the case to compromise, though without, I believe, abandoning any important theoretical principle or upsetting the harmonious working or appearance of the whole. It will become apparent that there are in the letter shapes many visual parallels to the phonetic or other relationships between sounds: thus, for instance, all vowels are Shorts; sixteen consonants are paired on a voiced | voiceless basis where this contrast functions within the language, the voiced Deep letter being the turned counterpart of the voiceless Tall in each case; the four other pairs, similarly related visually, record other types of relationship.

In conclusion, I do not feel it incumbent upon me to enlarge upon the conjectural future of Shavian: my interest in it has been largely professional, but this interest is still very much alive. At this moment I am concerned with the statistical estimate, stipulated in Shaw's Will, of the number of people writing English in the world today, on which it is intended to base an estimate of the economic savings to be achieved by its use. When this work has been completed, and perhaps some report on it made public, the last wishes of the late George Bernard Shaw will have been fully carried out. Whether or not the Shavian alphabet becomes more widely known or eventually adopted (or something like it) depends on its intrinsic merits and on the extent to which the purpose it is intended to serve kindles a spark in the minds of the present or of some future generation.