Shavian (2/3): the development
How Kingsley Read built a writing system to match George Bernard Shaw’s alphabetic utopia
However we read or write, we are used to working with mature if not always well supported writing systems. Shaped and reshaped by the currents of politics, economics, ergonomics, writing implements, cultural change and material availability, they will have been tested by a thousand hands. Inspired by George Bernard Shaw’s preface to The Miraculous Birth of Language, which called for a new alphabet for the English language, Ronald Kingsley Read spent eighteen years constructing Shavian: how did he hope to make something that bettered the products of centuries?
This article follows an earlier one which was largely concerned with the intellectual framework of Shaw’s recommendation for a new orthography, why he thought English should have a new writing system and what it could look like. This article is concerned with the relatively practical matter of how Kingsley Read designed a new alphabet.
The main features of Shavian
Finalised in 1960, Shavian is a 48 character writing system. Written left to right, it has the same numerals and broadly the same punctuation as Latin in the hands of standard English writers. Those 48 characters are divided into forty letters representing phonemes (sounds like the /k/ in ‘cot’ or ‘skill’, sounds that phonetic alphabets with a seemingly infinite capacity for distinction will render differently) and eight compound characters for syllables like ‘air’ and ‘yew’. The alphabet is unicase, using a ‘naming dot’ instead of capitals to indicate proper nouns. In Shavian, Shaw becomes ·𐑖𐑷. Four of Shavian’s letters, 𐑞, 𐑝, 𐑯 and 𐑑, can also be used for common words: ‘the’, ‘of’, ‘and’ and ‘to’ respectively.
Where previous phonetic alphabets, like Robert Bridges’, had been content to adapt and expand the Latin alphabet until they could be used to represent individual phonemes, in Shavian how a letter sounds and how a letter looks are linked in a system of logical beauty. Its letters can be sorted into three groups: all the short letters are used to represent vowel, liquid (r, l) or nasal sounds, almost all the tall letters are used to represent voiceless consonants (the /th/ in thigh) and almost all the deep letters represent voiced consonants (the /z/ in zoo).
Shavian’s letters can also be sorted into pairs. All but two of the short characters can be matched with a mirror or a cousin with an extra stroke, and each tall character can be matched with a twin that’s been turned upside down. These pairings reflect phonemic relationships: the /p/ and /b/ of bop: 𐑐, 𐑚; the /m/ and /n/ of mine: 𐑤 𐑮.
The Symbols, Read’s first manual
Read’s first proposal for a new alphabet, the one which had met with so much enthusiasm from George Bernard Shaw in January 1942, shows how Read starts to grapple with this question. He did not start from scratch, the influence of Henry Sweet and Shaw’s preface can be seen clearly, but so too can Read’s experience as a printer and commercial artist.
Like Sweet’s Current Shorthand, the letters in The Symbols are designed to be written in a continuous stroke and there is a system of categorisation. Consonantal, liquid and other sounds protrude above the x-height, below the baseline, or both. Vowel characters are short, but where Sweet had reduced his vowels to the point of being slightly elaborated horizontal strokes, Read expanded them to fully occupy the space between the baseline and x-height. In his first letter to Shaw, he argued that Sweet’s vowels were ‘no solution for type printing, even when his letters are separated.’[1]
Both these groups are then further divided into groups of related sounds. As with the final version of Shavian, visual relationships echo phonemic relationships, but the simplicity and similarity of the characters are striking, even problematical. Many are distinguished only by a shift in their relationship to the baseline, or through being flipped and inverted. This issue of similarity is particularly acute in the consonants. In groups one and two (‘pops’ and ‘fizzes’) an uninterrupted stroke consisting of a straight line with a curve is shortened, lengthened, mirrored, rotated and mirrored and rotated to make a collection of twelve hockey sticks. In groups three and four, the curve closes into a loop and the operation repeated — a collection of golf clubs.
The Symbols includes a lengthy explanation of the sounds its characters represent and their relationship to each other, to neat effect, but it also takes account of typographic concerns. Inner pages introduce a display style that leaves vowel signs unchanged, but compresses all the consonants between the base and ascender line. To distinguish the consonants, the display style uses ‘positioning bars’ to indicate their former relationship to the baseline. Read intended the display style for display purposes, but he also saw them as a substitute for otherwise ambiguous isolated characters.
Read also introduced two new typographic symbols. The first would be printed in the top-left corner of a sheet of paper to indicate which way up the text had been written or printed, and the second — a short vertical bar that would evolve into a dot — would be used instead of capitals or the display style to indicate proper names.
The second manual
That George Bernard Shaw’s response to Read’s 1942 manual was so brief and free from advice is striking. Even if he hadn’t spent the last four decades campaigning to reform English writing, few men have exceeded Shaw in their certainty that their opinion was not only welcome but vital.[2] By 1943, Read needed more. To no result, he had written to Otto Neurath, amongst others and now he hoped that Shaw might provide ‘a bare minimum of criticism’ regarding the script’s direction; and perhaps a collaborator?[3]
Although Shaw was sceptical of collaborators, he saw them as rivals, he provided the criticism. Arguing that all Read had to do was state his objectives and make his script ‘phonetic enough to enable every Englishman to understand what every other Englishman writes’[4], Shaw said the case for the alphabet should be self-evident if it was displayed with examples of its use. As a result, with the character set reduced and the forms slightly more diverse, Read’s manual of July 1943 shows only slight refinements. Instead of theory, tackled in a supplementary document, it is animated by a calligraphy guide, examples of lettering styles, and mock-ups of signs and advertisements.
Shortly after sending this to Shaw, Read would get what he wanted: a collaborator.
Enter Isaac James Pitman
The grandson of the inventor of Pitman’s shorthand, Isaac James Pitman was a printer, educationalist, alphabet and spelling reformer; a man Shaw would describe as ‘by far the best-equipped adventurer in the field’.[5] When he first introduced himself to Shaw in 1941, he was a Squadron Commander, months later he would be a director of the Bank of the England.[6] When it came to the question of alphabet reform, Shaw told Pitman, there was a man who had, ‘invented by far the best script I have seen so far.’[7] Pitman introduced himself to Read with a warm, digressive seven-page letter about the first Shavian manual, beginning a friendship.
Pitman had an essential question. Would Read’s writing system be improved if its forms were designed to take account of the frequency of the phonemes they represented?[8] After a three-way and sometimes confusing exchange of confused ideas on economy, readability and ergonomics, Pitman and Read agreed to meet in person on April Fool’s day 1944. It was there that Pitman lent Read his ‘Dewey’.
The Dewey Research
The son of the librarian, spelling reformer and mastermind of library classification, Melville Dewey, and the mind behind Lake Placid’s successful candidature for the 1932 Winter Olympics, Godfrey Dewey released his Relativ Frequency of English Speech Sounds in 1923. On a basis of 41 phonemes, Dewey had analysed 100,000 words of English prose ‘found in periodicals, books and other popular literature’[9] and had tabulated how frequently they occurred; what characters and what groups of characters were used to represent those phonemes and how frequently they occurred; and where and how frequently that phoneme occurred within a word sound, on an initial, medial, final or (if independent) syllabic basis.
From this it was possible to extrapolate how substituting the characters representing a phoneme with a single new character — as Shavian intended to — would reduce the total number of characters required to print the sample text. And, if the sample text was truly representative, English writing itself. With Dewey’s research, Read no longer had a theoretical idea of the efficiencies of a phonemic alphabet, but an empirical one. Furthermore, Dewey’s data about phonemic occurrence, combined with Read’s personal experience, gave him an improved model for understanding how the characters he designed would interact with each other. Relativ Frequency of English Speech Sounds became a key reference, its tables were copied out and annotated in pencil and pen,[10] its detail a running source of calculations between him and Pitman.
Read began to design his alphabet so that it not only reflected phonetic relationships but the pattern of English phonemes. Natural ligatures could be devised for frequent neighbours and the commonest characters could have the simplest forms.
A year later, Pitman’s secretary asked for the book to be returned[11], but by 1946 it had not:[12] Pitman’s Dewey was indispensable.
No example of Read’s third manual exists in a single bound form, but Dewey’s influence can be seen in the scraps, sample sheets and annotated pages that survive. The third version of Shavian was more heterogeneous. The golf clubs and hockey sticks gone and replaced by characters that more closely resemble the final Shavian alphabet, even if they are not attached to the same sounds. Although Read had rejected a suggestion from Shaw that the alphabet should be a joined script [13] on the basis that it should be suitable for printing [14], the characters no longer stand awkwardly apart. Instead, the characters of written Shavian begin to lead into each other and form natural ligatures. Read was no longer designing his system as an abstraction, but as a writing system tested through use.
Despite these advances, in March 1946 he wrote to Shaw that for months he had been unable to imagine further development. He was stuck; [15] although Read kept nagging at his alphabet design, the examples he had enclosed were to be the last examples sent to Shaw. More than four years later, on Sunday 10th of September 1950, Shaw slipped while taking his secateurs to a projecting branch and fell suddenly into illness. In the early hours of the first of November, he declared, ‘I am going to die.’
Shaw’s will and a competition for a new British alphabet
In the summer of 1944, Shaw remarked that ‘drafting [my] will has been more trouble than ten plays.’ [16] What measure he might have applied to the execution of it can only be speculated. Shaw had attempted to establish charitable trusts, ‘alphabet trusts’, that would use part of his estate to select and promote a new ‘British Alphabet.’ An annuity would go to his chauffeur-gardener, and the rest would go to the Royal Academy of Dramatic Art (RADA), the British Museum and the National Gallery of Ireland. Sniffing opportunity, The British Museum and RADA contested the will: were the alphabet trusts really ‘charitable’?
In January 1957, with the estate fattened from the royalties from My Fair Lady and now worth some £716,000, the case finally reached the courts. Michael Holroyd, Shaw’s biographer, contrasting the proceedings with Dickens’s Jarndyce & Jarndyce, provides a summary account that runs to only eight pages. The result was unhelpful: the alphabet trusts were invalid. Two years of campaigning by Pitman and the indefatigable Barbara Smoker, then Secretary of the Shaw Society, secured an appeal and a compromise. There would be a competition run by the Public Trustee and some promotion, but the budget would be a shadow of Shaw’s intentions.
Naturally, Read made a submission. By the end of the 1958, it had been joined by 466 other entries from around the world; of the world’s continents, only Antarctica was unrepresented.[17] The task before the two judges, Read’s friend James Pitman and Peter MacCarthy, head of the University of Leeds’ Department of Phonetics, was enormous.
Read’s submission consisted of two parts. A visual presentation and a 14 page document titled Notes. Notes explained the phonemes the writing system was built around, and with reference to Shaw’s arguments for a new orthography, expands on its logic. The visual presentation, three sheets of graph paper, each labelled as a plate, was intended to have impact. The first presented the alphabet and its basic principles, both of which were, apart from the reassignment of some consonantal sounds to different glyphs, unchanged from the previous manual.
Taped to the first plate, a second showed how the Shavian alphabet could be adapted for printing. The typographic considerations were not limited to a question of how Shavian’s printed letters might be modulated, however. This version of the alphabet is grouped into eight sets of different sizes and these groups are predicated not on the basis of shared phonemic values, but on shared uniform widths: the ‘set widths’ of a Monotype matrix case.
Turning to the third plate, Pitman and MacCarthy could see yet more typographic considerations. Along with an italic and an angular style, Read showed three widths of a Shavian typeface in his ‘modified styles of letterpress’. Read’s display style, this time labelled as a ‘provision seldom required’, lives on as an adaptation for particular circumstances: isolated characters or for being tightly set ‘between parallel lines.’
Although Pitman asked Read how he might judge legibility [18] — an eternal question — the two judges were to rely on their own experience. To Barbara Smoker, this was problematical: they both had experience of phonetics and spelling reform but neither had calligraphic or educational psychology expertise.[19] In their attempts to determine a single winner, Pitman and MacCarthy contracted a calligrapher to transcribe a sentence of their choosing in four of the submitted alphabets[20] but an exact choice escaped them. On New Year’s Eve 1959, four joint winners were announced.
Initially, the four winners and their design were to be brought together to make a final alphabet. Although a similar operation had not been for Dr Frankenstein, Barbara Smoker viewed this as ‘impossible’. Pitman was more optimistic, but it is unclear how much of an attempt was made; or at least how well-managed the attempt was. First the designers were told to only communicate with each other through the Public Trustee, then only through Peter MacCarthy and then — with the incumbent complexities of such discussions compounded by the time and distance — it was decided that there should be only one designer, Kingsley Read, and no monster.
The final version of Read’s submission
In August 1960, Read submitted the final version of the proposed alphabet on a single sheet of card. Although concerned with defining the skeletons of the characters and their associated phonemes, it proposed some of the most significant changes to the alphabet of all the specimens. The alphabet has been expanded and its characters are divided into pairs of flipped/rotated characters which belong to groups of four which share shapes. Beyond the introduction of three new compound characters (array, ear, ian), there are two characters for setting the Welsh /ll/ and Scottish hard /ch/, neither of which would appear in the guide printed in Androcles and the Lion.
In their correspondence Read and MacCarthy had evaluated every character with an eye towards efficiency and ergonomics. Although the discussion was complicated by the question of how they might evaluate new characters when they were only newly habituated to the old, it lead to new characters, to new shapes. It also lead to choices. The character pair hung/haha (𐑣 𐑙), for example, reverses the usual tall/deep voiceless/voiced consonant relationship, attracting puzzlement. In response to a curious letter from Ivor Darreg,[21] Read explained why he opted for ease of writing over systematic consistency:
H is mostly an initial letter and (in Shavian) connects as often on the upper parallel as on the writing line, while NG is by far more likely to end a word or syllable following short-i (ing) on the writing line. It was one of the unfortunate features about devising the Shaw Alphabet — in my considered view — that it was done, for want of practical experience, with too much attention to purely academic consistency.[22]
In his submission to the competition, Read’s manuscript notes had ended :
This Alphabet’s practicability by one commonly used printing process has been confirmed by The Monotype Corporation. After very detailed consideration, they find no exceptional difficulties. Initial production of suitable matrix dies and key-bar frame is found comparable to that of redesigned Roman or Eastern alphabets. Small Monotype machines will accommodate some 57 to 60 signs and ligatures, each in three styles such as medium, bold and italic, with necessary punctuation marks, numerals etc. Other common processes seem equally applicable.[23]
This was to be tested. The Public Trustee adopted the proposed alphabet and the task was now to prepare type for the printing of the Shavian edition of Androcles and the Lion.
Next:
Shavian (3/3): the typographic implementation
Endnotes:
[1] K. Read to G. B. Shaw (Sutton Coldfield, Birmingham: 20 January 1942), sh6/63/f1–2.
[2] At a lunch in honour of Henri Bergson, Shaw not only explained Bergson’s ideas to him but when the philosopher attempted to correct him butted in: ‘Oh my dear fellow, I understand your philosophy much better than you do.’ While Bergson fizzed, Shaw blithely returned to his exposition.
[3] K. Read to G. B. Shaw (Sutton Coldfield, Birmingham: 20 January 1942), SH6/16/f1–2.
[4] G. B. Shaw to K. Read (Ayot St. Lawrence, Hertfordshire: 3 August, 1943), SH6/63.
[5] I. J. Pitman and J. St John, ‘Alphabets and Reading: Initial Teaching Alphabet’ (Bath: Pitman Press, 1969), p. 111.
[6] D. H. Laurence, ‘Bernard Shaw: Collected Letters, 1926–1950’ (London: Max Reinhardt, 1988), p. 615.
[7] G. B. Shaw to I. J. Pitman (Ayot St Lawrence, Hertfordshire, 27 September, 1943), SH9/13/f20.
[8] I. J. Pitman to K. Read, (London, 4 October, 1943), SH1/KR22.
[9] H. B. Bisgard, ‘Book Review (Relativ Frequency, etc)’ in ‘Journal of the Simplified Spelling Society, vol. XVIII’ (Unknown, Simplified Spelling Society, 1978), pp. 18–19.
[10] SH1/KR48, SH1/KR36.
[11] E. O. Blissett to K. Read (London, 21 April, 1944) SH1/KR52.
[12] I. J. Pitman to K. Read (London, 18 February, 1946), SH1/KR68.
[13] G. B. Shaw to K. Read (Ayot St Lawrence, Hertfordshire: 3 August, 1943), SH6/63f1–3.
[14] K. Read to G. B. Shaw (Sutton Coldfield, Birmingham: 7 September, 1943), SH5/33/ f1–5.
[15] K. Read to G. B. Shaw (Sutton Coldfield, Birmingham, 9 March, 1946), SH6/12/f2.
[16] D. H. Laurence, ‘Bernard Shaw: Collected Letters, 1926–1950’ (London, Max Reinhardt, 1988), p. 713.
[17] M. Holroyd, Bernard Shaw (London: Pimlico, 2011), p. 803.
[18] I. J. Pitman to K. Read (London: 1 December, 1958), SH1/KR281.
[19] M. Holroyd, Bernard Shaw (London: Pimlico, 2011), p803.
[20] The item is listed in the Museum of English Rural Life’s catalogue of the archive as SH6/10/f1–15 (p.100) but its location is unknown or the object unrecognised.
[21] I. Darreg to K. Read (Los Angeles: n.d. (c. December 1972–January 1973), SH9/2/f4.
[22] K. Read to I. Darreg (Abbots Morton: 26 January, 1973), SH9/2/f5.
[23] K. Read, Competition Submission to the Public Trustee (13 September 1958), SH6/10/f15.
