(A) Literature Survey (Pt.2, 115pp.)
(B) Questionnaire (Pt.2, Vol.2, pp.1-4)
(C) Surveys of Space-saving and Frequency of Contractions (Pt.1, Vol.2, pp.5-80)
(D) Experiments with Modified Grade 2 Braille Codes to Determine their Effect on Reading Speed (Pt.1, Vol.4, p.39; Material Used, Pt.1, Vol.3, pp.1-58)
(E) "The Limitations of Braille as a Medium for Communication and the Possibility of Improving Reading Standards" (Lorimer, J., 1978) (Pt.1, Vol.4A, pp.1-21)
(F) Analysis of Symbols, Meanings and Rules of Standard English Braille (Pt.1, Vol.4B, pp.1-16)
(G) Analysis of Errors (Pt.1, Vol.1, pp.31-34)
(H) Outlines of a Short Course to Improve Braille Reading Efficiency (Pt.1, Vol.4C, pp.1-47)
7. THE WASHINGTON INTERNATIONAL CONFERENCE ON ENGLISH BRAILLE GRADE 2 (1982)
8. POST WASHINGTON
Capital Letters
Code Revisions
The BAUK Questionnaire
9. THE LONDON INTERNATIONAL CONFERENCE ON ENGLISH BRAILLE GRADE 2 (1988)
10. OBSERVATIONS ON CODE DESIGN AND CHANGES IN ENGLISH BRAILLE FROM C1870 TO 1960
Circa 1870
1895
1905 Revised Braille
1932
1960
After revised braille had been accepted at the international conference held in Edinburgh in 1905, a time of consolidation followed. The National Lending Library for the Blind, which supplied books in both braille and moon codes, had been in existence since 1862. The standard of transcription was high for "volume by volume the books come in, are corrected at the library, shellaced and bound to wait on a cellar shelf until completed. A report on each volume is sent to the writer, so that any mistakes are not repeated" (Outlook for the Blind, 1909, 2, p.173). "Shellac" refers to a type of varnish applied to the pages which made them stronger and the dots less likely to be depressed. The first braille copies were made by sighted transcribers then blind people made replicas either voluntarily or as a means of paid employment. To maintain a high standard of teaching, the College of Teachers of the Blind was inaugurated in 1907. Its examinations were recognised by the Board of Education and these included a test of a knowledge of braille.
The settled state of the code was demonstrated by the fact that no meetings of the British Braille Executive Committee were entered in its minute book after 1905 until May 1916, though records of sub-committee meetings to tidy up the 1905 decisions still exist.
In 1916 the National Institute for the Blind convened a meeting of representatives of educational societies, publishing houses, the home-teaching society and libraries, "to appoint a thoroughly representative committee (which shall in effect be a national committee) to deal with the whole subject of Uniform Type". It became known as the National Uniform Type Committee and the first work to be considered was the changes contained in "Revised Braille for reading and writing, 1915" at the request of the American Commission on Uniform Type (see the previous chapter). Besides the ensuing meetings and correspondence made necessary while the Americans were determining their code, a meeting was held in Britain in 1922 (Irwin, 1922, p.64) to the motion "that the present system known as Official Grade II does not adequately meet the requirements of the blind student, and that the present time is opportune for a more scientific adaptation of the Braille System". In fact, the committee considered the timing most inopportune, for many Americans already thought Grade 2 was far too complicated and therefore differences between the two grades should not be increased. It was decided that if so required, a scientific work could be prefaced with suitable symbols and their meanings. No further code problems arose until Irwin put some American suggestions to a meeting of the British Uniform Type Committee in November, 1929.
Once the Commission on Uniform Type in the United States had accepted Grade 1½ (1918) no time was lost in the production of text books for schools. It was hoped that Grades 1½ and 2 were sufficiently alike for exchanges of adult books to take place across the Atlantic. In fact, it gradually became apparent from opinions given by library users, that whereas British people rarely read Grade 1½, approximately 20% of Americans read Grade 2 and the trend seemed to be increasing (Irwin and Wilcox, 1929, p.1). The Commission on Uniform Type for the Blind (1917, p.14) had been instructed to "continue its labors with the British National Uniform Type Committee with the view of reaching a complete agreement regarding the whole question of uniform type". Having produced a number of text books for children the Americans were about to provide classics and more fiction for adult readers (Irwin, 1929, typescript, p.6), so it seemed the moment for the research department of the American Foundation for the Blind to undertake a more detailed study of the differences between the two codes.
On first inspection the main differences seemed clear cut. Grade 1½ used 44 of the 189 contractions included in Grade 2; two-space contractions, double letter signs and shortforms were not included; and the American version used capital letter signs wherever capital letters occur in inkprint.
The American lexicographer, Noah Webster, devoted much of his life, particularly between 1789 and 1828 "to the zealous championing of the cause of American language, its spelling, its grammar and its punctuation" (McCrum, et.al., 1987, pp.240-241). Webster's maxim was, "A good articulation consists in giving every letter in a syllable its due proportion of sound, according to the most approved custom of pronouncing it, and in making such a distribution between syllables of which words are composed that the ear shall without difficulty acknowledge their number", and school children were often made to chant columns of his spelling lists (ibid., pp.241-242). It does not seem too far-fetched to suggest that his writings had an effect on American preference for syllabication in braille usage over a century later. Syllable divisions are not always easy to determine so a dictionary is a useful tool for the transcriber.
By contrast, in Britain contraction rules permitted the shortening of words so long as the clarity of meaning was not impaired. This sequencing often resulted in contractions crossing syllable boundaries and sometimes a choice of contractions had to be made involving the need for yet more rules of usage.
No exhaustive analysis of space-saving values had been carried out in the 1920s when Grade 1½ was being used in America. In the interests of uniformity, Irwin and Wilcox (1929) made a study of 91,564 words taken from four volumes, specifically to determine the space-saving effects of the use of Grade 2 over Grade 1½. They were particularly interested to know what the effect would be if Americans accepted the 44 contractions and shortforms not included in Grade 1½.
The 91,564 words were treated as if written in a continuous line to obviate the effect of varying lengths of lines in the publications consulted. When written in Grade 1½ they took up 427,776 cells and the Grade 2 version used 376,866 cells, that is, 11.9% less space than Grade 1½. This saving was caused not only by the differences between the space saving of the contractions, but because Grade 1½ had to omit some of the contractions because of the syllabication rule (Loomis, 1936, pp.4-5). For example, the following contractions were not permitted:
AR in library, military,
ER in serious, exterior,
AND in candle, abandon.
Because of these imposed restrictions in the use of Grade 1½ the code was less flexible and the uncontracted letter strings were likely to have had a deleterious effect on the rate of reading.
Irwin and Wilcox included tables showing in descending order the space saving value of each contraction and also its frequency of occurrence. Both these measures are necessary for determining the relative value of contractions for inclusion in the code. The extent to which they facilitate or hinder reading is more difficult to determine and more accurate results can only be provided by research. The results showed that the code was even less economic than forecast, for it was found that though TO saved 7% of all the space saved by Grade 2 over Grade 1½, fifty contractions at the lower end of the table taken together did not save as much space as some individual contractions at the top of the table, and seven contractions or shortforms did not appear in the literature selected. The study was sent to members of the (British) Uniform Type Committee with the hope that by a judicious addition of contractions to Grade 1½ and some deletions from Grade 2, a better measure of uniformity might ensue.
Irwin set out some general proposals for bringing the two grades closer together before the meeting of the British authorities held in July, 1929 (British NUTC minutes). Though not all members were enthusiastic, they agreed to set up a sub-committee to meet with the Americans. The American Printing Press were so concerned by Irwin's activities that the director sent a cable to England, the gist of which was "We are not sure what you are doing in your effort to bring about braille uniformity, but whatever it is, please desist" (Irwin, 1955, p.50). It did not reach him until his meeting with the British was concluded!
The AAIB and AAWB were asked to bring the study before their members to enable them to discuss what further steps should be taken. Publishers of braille magazines in America had helped readers to understand possible changes by gradually introducing Grade 2 signs in their publications before a conference was held in London in July, 1932, between representatives of the braille authorities of Britain and America (Irwin, 1932, p.138). By contrast with past disagreements, the atmosphere was cordial and it would seem that a great deal of this co-operation was due to the careful preparation carried out by Robert Irwin, himself a blind man.
The main decisions (British UTC Minutes, 1932) included the following:
1. The dot numbers within the cell were changed from:
1 · · 2 3 · · 4 5 · · 6 |
to |
1 · · 4 2 · · 5 3 · · 6 |
(This reversion to the French method was probably taken so as to be compatible with braille in mainland Europe. It is suggested by the writer that the interim method may have been used so as to be the same as dot numbering for music.)
2. The British signs for capital letters and italics were reversed, the new positions being:
Capital Letter Italics
· · · l
· l · l
(This decision helped recognition of two-spaced compound contractions. America continued to use capital letters wherever they occur in print, whereas in Britain the capital was not generally used.)
3. Britain agreed to drop the use of nine biblical words except in religious literature, and five new shortforms were added (see Appendix 1).
4. The Americans agreed to adopt all the contractions and shortforms not already included in Grade 1½.
The Americans had come a long way in accepting so many of the Grade 2 rulings in the interests of uniformity, but they still differed over some of the British practices with regard to sequencing within words. Rule 34 was inserted in "Standard English Braille, Grades 1 and 2" (1932, National Institute for the Blind, p.18) as a compromise: "The contractions forming parts of words should not be used when they are likely to lead to obscurity in recognition of pronunciation, and therefore they should not overlap well-defined syllable divisions. Word signs should be used sparingly in the middle of words unless they form distinct syllables. Special care should be taken to avoid words of relatively infrequent occurrence."
Lists of words showing preferred methods of contracting were appended, which is evidence of the problems involved in trying to use a code of comparatively few signs to represent the large vocabulary making up a language. Despite these difficulties, after a century of endeavour, the English-speaking peoples on both sides of the Atlantic were united in having a code similar enough for an increase in the exchange of books, and any further code alterations could be decided together in a spirit of harmony. Canada, Australia, and New Zealand would benefit as well as the peoples for whom English is a second language. The conference was a significant occasion in the history of the development of the braille code.
4. A SURVEY OF THE FREQUENCY AND SPACE-SAVING OF THE CONTRACTIONS OF GRADE 2
Although a close degree of uniformity had been reached in 1932, a sub-committee of the (British) UTC undertook to look to the future to determine whether yet more improvements could be made. A more detailed research was needed (British NUTC minutes, 28th January, 1950), so the first task undertaken was a wide ranging survey of the frequency and space saving values of all existing Grade 2 contractions to see how far it is an effective system. This was carried out by Lochhead and Lorimer, J. (1954).
To fulfil its purpose a survey not only has to be carried out with accuracy, but must cover as wide a range as possible in proportionate amounts within the confines of the subject studied. With these criteria in mind, the larger the number of words studied, the more accurate the study should prove to be.
The study was to cover general reading so learned and scientific works were not included. Twelve subject areas were covered ranging from nineteenth century fiction to modern day periodicals. No calculators or computers were available at that time so the processing of half a million words was carried out by volunteer helpers who were all experienced braillists. From this information the relative space saving of each contraction, and also the usefulness of "families" of contractions which have similar rules for usage, could be determined. The results again showed that Grade 2 is a far from efficient system. Nine contractions had been deleted and 5 new ones added in 1932 bringing the total to 185. Of these, the first 14 contractions do almost as much work as the remaining 171 and the last 45 contractions between them only save 500 spaces. The survey was a model of careful planning and efficient execution. A copy of the report was sent to the Joint Uniform Braille Committee in America (British NUTC minutes, 25th March, 1955).
With the publication of a detailed rule book by the British UTC (1953), and the Lochhead and Lorimer survey showing strengths and weaknesses of the code, the NUTC felt in a strong position to make suggestions for possible improvements. It was considered that "space saving was important, but that readability and ease of learning were more important" (British UTC minutes, 25th March, 1955). A sub-committee produced 10 recommendations (ibid., 3rd June, 1955), which were sent to the Americans, who in turn sent some of their own which were less far-reaching. They also suggested a joint working party for all the proposals to be discussed, so representatives from both countries met in London in 1956.
The joint decisions (ibid., 30th November, 1956) were considered by representative bodies in Britain and North America. Many of the suggestions were accepted in Britain, but in America, not only were the deletions suggested in 1956 approved, but further deletions had been approved since, thus breaking the "gentlemen's agreement" between the two countries. It had been a democratic decision by the American public. The main changes concerned the use of lower signs when used in sequence, the addition of four shortforms (bringing the total to 189), and the dropping of the use of a hyphen in three compound words. America had agreed to most of the contents of Grade 2 in 1932. The position was now reversed with Britain agreeing to American preference.
During the 50s and 60s some research studies on the braille code were carried out by members of the research committee of the College of Teachers of the Blind. Meanwhile, a substantial amount of speculative and empirical braille research was accelerating in America into the perception of braille and methods of making the code easier to learn and to read. The time had come for the CTB research committee to be disbanded and instead, the full-time Research Centre for the Education of the Visually Handicapped was set up at Birmingham University in 1970. Warwick University also had a research centre.
Douce and Tobin (1976) considered that the time was ripe for a multi-variate approach "to generate the kind of information that would make it possible to formulate decisions about code modifications that would, in turn, make the learning of braille, and subsequent improvement in reading efficiency, easier". A joint three year project was therefore undertaken by Warwick and Birmingham universities (Lorimer, J., et al., 1982, vol.1, p.2) with two main aims: (1) to explore the possibility of producing an alternative Grade 2 code that would be shorter and simpler to learn, but could be read as fast as the existing Grade 2 code and would require no more space, and (2) to investigate ways of improving teaching methods and materials. Descriptions of some of the main studies follow.
An annotated bibliography was compiled by the writer during the first year of the study for the benefit of the research team. It included works on the development of the braille code, research and education, but because of the time factor, parameters had to be set; approximately 300 entries were included, most of them published between 1945 and 1980. An author index was included and also a subject index compiled from the descriptors which had been included with each entry.
The questionnaire survey was undertaken to help in identifying specific aspects of the braille contraction system which merited detailed study. Three hundred and one responses were received from blind, partially sighted, and sighted subjects whose age range was 20-93 years, all of whom lived in the United Kingdom. Information was received on such aspects as preferred grade of braille, opinions on space saving, specific problems in recognition of certain contractions, and opinions on punctuation and unit signs, and format. In conclusion (p.4) "the general view seems to be that there is no desire for any change unless it can be very clearly demonstrated that there are some very significant advantages to be gained by modifying the system".
(c) Surveys of Space-saving and Frequency of Contractions (Pt.1, Vol.2, pp.5-80)
The Gill corpus of 2,255,326 words was compiled at Warwick University using a selection of material already stored in the data bank. This consisted of short runs which had been put into braille at the request of individuals, and also an almost equal number of words entered from novels, short stories and non-fiction. Also included were the contents of the Kucera and Francis count (1967), often referred to as the Brown Corpus. Numbers of pairs, three-letter and four-letter strings of letters were tabulated indicating whether they occurred at the beginning, medial and final positions in words, and a table in descending order of space-saving was also included. Thus, by inspection, it was possible to determine which groups of letters would be the most economical to include in a projected improvement to Grade 2 braille.
The Lochhead and Lorimer study (1954) containing approximately 540,000 words, and the Kederis, Siems and Haynes (1965) count containing 291,000 words from children's literature, were already in existence, but it was felt that none of the four corpora was entirely satisfactory because of differences in the selection and proportion of the materials treated, the differences in the number of words analysed and also the variations in language and braille usage between Britain and America. To obtain a more accurate result each contraction was reduced to occurrences per million words in each of the four counts and then the mean frequency was computed and tabulated in descending order. A similar treatment was given to demonstrate the space-saving values. A further table showed the space saving contribution made by 'families' of contractions each containing contractions which have similar rules. This knowledge is useful because keeping a whole family or deleting the whole or part of it may have a considerable effect on the use of space and the number of necessary rules.
(d) Experiments with Modified Grade 2 Braille Codes to Determine Their Effect on Reading Speed (Pt.1, Vol.4, p.39; Material Used, Pt.1, Vol.3, pp.1-58)
In general, reductions in the contents of a code will decrease the number of necessary rules and make the learning task easier, whereas extensions will increase the learning load. However, before changes are made it is necessary to estimate the probable effect of these changes in terms of space-saving, ease of learning and on rate of reading. Findings of frequency and space-saving surveys had given information that might lead to improvements, but so far no measurements had been made of the effect of possible modifications based on this information in terms of words read per minute.
During the first year of the study while other investigations were proceeding, 10 code modifications were tried out under controlled conditions in order to determine their effect on rate of reading. It was realised that because of lack of experience with the experimental codes, subjects tended to read them more slowly than the Grade 2 version with which they were familiar, and therefore differences in the scores obtained should be regarded as maximal. However, they provided a useful indication of the effect of certain modifications on reading rate.
The criteria for the experiments included the fiats that space-saving and reading rate should be at least equal to those for Grade 2. The results of the experiments showed that about 70 of the useful contractions could be deleted, but to fulfil the terms of the criteria, about 10 new space-saving contractions of greater value could be added. Statistical data supplied by other investigations in the study supplied further predictions (Vol.2, Ch.5, p.79).
(e) "The Limitations of Braille as a Medium for Communication and the Possibility of Improving Reading Standards" (Lorimer, J., 1978) (Pt.1, Vol4A, pp.1-21)
This paper, written during the early stages of the project, was read at a meeting organised by the British Psychological Society. Comparisons are made between accuracy, comprehension, and reading rates of blind and sighted children (Ashcroft, 1960; Williams, 1971; Lorimer, 1977). Information is also included concerning intelligence and braille reading, the limitations of touch perception, and the application of rapid reading techniques for inkprint to braille reading in an attempt to improve braille reading standards.
(f) Analysis of Symbols, Meanings and Rules of Standard English Braille (Pt.1, Vol.4B, pp.1-16)
It has long been realised that one of the main difficulties in reading braille has been the number and complexity of the rules. Fewer contractions with attendant rules might make the code easier for some people at the learning stage, particularly adults who can already read inkprint yet have to learn a new set of symbols and need training in their perception. Computer programmers would also welcome a simpler code. An analysis was carried out to determine:
The Lorimer Braille Recognition Test (1962) was administered to 91 subjects in the United Kingdom, of whom 64 were under the age of 20 and 27 aged 20 and over. The test included 178 of the 189 contractions of Grade 2. No errors were recorded on 23 contractions and of the 23 contractions most associated with error, 311 mistakes were recorded by subjects under the age of 20 and 111 errors by those who were 20 years and above. This test had been standardised on a sample of 332 children but there is no standardised braille reading test yet available in Britain for use with adults. These facts, though gained on a relatively small sample, can be taken as an indication of some of the reading difficulties of Grade 2 braille.
(h) Outlines of a Short Course to Improve Braille Reading Efficiency (Pt.1, Vol.4C, pp.1-47).
Based partly on rapid reading techniques involved for sighted readers (Fry, 1963; de Leeuw, M. and E., 1965) and research on rates of braille reading (McBride, 1974; Crandall and Wallace, 1974), Lorimer, J. planned a course intended to improve the rates of reading of pupils in the lower grades of senior schools. The materials used are included in the report and also a table for converting reading time to reading rate so that replication of the experiment is possible.
The Birmingham/Warwick study was the most extensive investigation of braille ever attempted and was bound to accelerate further studies of the code. The report was sent to the Department of Health and Social Security, and members of the Braille Authority of North America (BANA). The writer was present at the Toronto conference held in 1981 in preparation for the Conference of English Braille, Grade 2, held a year later. The chairman praised the work of the university investigators and remarked (pers. comm.) that the enquiry was at least five years ahead of any other enquiry at that time.
7. THE WASHINGTON INTERNATIONAL CONFERENCE ON ENGLISH BRAILLE GRADE 2 (1982)
The basic assumption of the universities' study was that there should be one official code in general use. Members of BAUK recognised the advantage of reducing the number of contractions to be learned without sacrificing space and it was hoped that such changes would induce more adults to want to learn the code. However, a few members of the Braille Authority of the United Kingdom (BAUK, formerly British UTC) wanted an advanced grade as well, with many more contractions for the use of fluent, extensive braille users even though such people would be in the minority (Milligan, 1982, p.93).
In 1978, the Braille Authority of North America (BANA) submitted the following recommendation for consideration by BAUK (proceedings of the Washington conference, 1982): "The Braille Authority of North America wishes to join with the National Uniform Type Committee of the United Kingdom to explore the possibility of devising a common Literary Braille Code for the English language. The exploration should encompass both readability and computer implementation."
At the Washington Conference (1982) delegates recognised that computer use of braille would make some code changes inevitable, but such changes should be regarded as "temporary, experimental guidelines, rather than amendments to the braille code" (Resolution 6, p.248).
Amongst the research suggestions (Resolution 7, p.248) were: the development and field-testing of a contracted literary code based on the experience of users and teachers as well as on theoretical consideration and research findings; the achievement of uniformity in the form, placement, and spacing of unit abbreviations for coinage, weights, and measures; and attempts to persuade properly qualified persons to provide information about the structure and functions of the touch sense (with reference to braille reading). These suggestions seem wide ranging and it was encouraging to the writer, who was present, to note the attempted welding together of expert theoretical knowledge with user experience.
Possibly the most important resolution passed was the setting up of an international co-ordinating committee whose members were to be one representative from each of the participating nations. Their mandate included the preparation of proposals for the establishment of an international authority on English literary braille. This was a commendable step, for so far the only authorities making decisions had been those of Britain and America. The activities of the proposed international authority will be further considered in Chapter 9.
Another conference was planned to take place in London when definite proposals for changes to the code would be considered and braille users were to be fully consulted before any such changes were implemented. The exception was the immediate acceptance by Britain of the American practice of using the AR contraction in preference to EA in the letter group "ear".
A joint investigation was carried out between the USA, Canada, and Britain to determine what effect the capital letter sign had on the comprehension and rate of reading. The results were insignificant and at present each country is following its own preference regarding the use of capital letters. Americans still prefer to use them where they occur in print and the British use them very rarely. An ongoing enquiry under the auspices of BAUK is at present (1996) being carried out concerning "Capitalization in British Braille" (see Chapter 9).
A large proportion of inter-conference time was spent in Britain on evolving a variety of new versions of the literary braille code system. These may be consulted in "Essays in Code Design" (RNIB, 1986) where (p.5) the report is regarded as "the culmination of the research" carried out by Birmingham and Warwick universities. They include:
These codes, together with the experimental ones tried out in the universities' study, demonstrate very clearly that there can be no one universal code suitable for all braille readers. Varying abilities, interests, and circumstances need to be considered, preferably the codes should have the same meanings for the symbols that are common, and it would be impractical in terms of expense to have books in too many code versions. The trials were useful exercises in demonstrating the difficulties for code design is a very complex exercise.
A questionnaire on braille code reform was circulated by BAUK in 1986 to which there were in all 1627 replies, 1217 of whom were touch readers. Braille is an emotive subject for blind people. It is something they regard as their own and conservative feelings about change were expected, especially among the older members of the population. On the whole, advanced readers and young readers were the more ready for reform. About two-thirds of the braille using respondants wished the contents of Grade 2 to remain substantially unaltered but with some further contractions added (Lorimer, 1988).
9. THE LONDON INTERNATIONAL CONFERENCE ON ENGLISH BRAILLE GRADE 2 (1988)
The BAUK questionnaire (1986) had revealed that the majority of braille users who sent in replies did not want any major changes to the literary braille code. The Washington braille conference (1982) had recommended the setting up of an international council on English braille which would be ratified at the London Braille Conference of 1988. These two decisions led to a conference decision that any further improvements to the code should be delayed until the International Council on English Braille was set up. Arrangements for the latter were to begin immediately following the conference. However, a few changes were agreed involving the use of quotation marks, fractions, and unit abbreviations.
10. OBSERVATIONS ON CODE DESIGN AND CHANGES IN ENGLISH BRAILLE FROM C1870 TO 1960
Because touch reading is such a slow medium, the choice of contractions to reduce the amount fingers must travel is of great importance. In theory, the more contractions that are included, the greater the rate of reading, but there are other qualifying aspects. Too many different symbols increase the learning load and number of rules, and of even greater importance is an understanding of the problems presented by embossed material. These perceptual difficulties are complex and will be considered at greater length in Chapter 8. More easily measured is the amount of space-saving contributed by each contraction leading to an increased rate of reading.
Because "the Grade 2 code is not as efficient a communication medium as it might be" (Lorimer, J., Tobin, Douce, and Gill, 1982, p.121) it is obvious that some contractions save more space than others. For example, "immediate" IMM saves 6 spaces whereas "your" YR saves 2, so it might be thought that "immediate" would be of better value in space-saving, but it is also necessary to know how often the contraction is likely to occur. Even after calculating the mean values of four counts in order to gain greater accuracy of values of space-saving, Lorimer, J. added ("Study of Braille Contractions", Vol.2, p.49) that these "must be regarded as approximations, for there can be no absolute frequency or space-saving value for any contractions". Using Table 17 (ibid., Vol.2, p.58) it may be seen that in a million words "immediate" saved only 865 spaces whereas "your" saved 3,134, the reverse of what might have been previously expected before frequency was considered.
This aspect of space-saving was not always recognised in the past so Table 3 (below) was compiled by the writer to provide the information in a compact form. It is realised that vocabularies change slightly over the years but even so the values shown in the table give some indication of the success of the added contractions. A few of the contractions were given different symbols in ensuing revisions so, by adjustment, the contractions here show space-saving at the time when they were first included in the code.
Nine religious contractions and five others were included in the British Literacy braille code at various times between c1870 and 1932. These were not included after 1932 so there are no space saving values included for them in the 1956 and 1978 counts, and therefore do not appear in the table. The table should therefore be regarded as indicating main trends in the year before 1932.
Table 3
Table to show the number of new contractions introduced into the British literary braille code between c1870 and 1960, which saved at least one space over Grade 1 per million words.
(The space saving shown here does not include alphabetic, punctuation, composition or mathematical signs.)
Total Number of Spaces Saved by Each |
Number of Contractions Introduced |
Contraction |
c1870 |
1895 |
1905 |
1932 |
1960 |
Over 50 Thousand |
4 |
- |
- |
- |
- |
40,000 - 49,999 30,000 - 39,999 20,000 - 29,999 10,000 - 19,999 |
3 4 3 17 |
- - - - |
- - 2 3 |
- - 2 3 |
- - - - |
9,000 - 9,999 8,000 - 8,999 7,000 - 7,999 6,000 - 6,999 5,000 - 5,999 4,000 - 4,999 3,000 - 3,999 2,000 - 2,999 1,000 - 1,999 |
2 1 1 2 3 5 1 7 7 |
- 1 - 1 1 1 2 4 4 |
1 1 2 4 6 5 9 11 22 |
- - - - - - - - - |
- - - - - - 1 2 - |
900 - 999 800 - 899 700 - 799 600 - 699 500 - 599 400 - 499 300 - 399 200 - 299 100 - 199 Under 100 |
- - - 1 - - 1 - - - |
- 1 - - - - 2 - - 2 |
7 - 1 6 1 1 5 1 3 8 |
- - - - - - - - - - |
1 - - - - - - - - - |
Total Number of New Signs |
62 |
19 |
99 |
5 |
4 |
At once it becomes apparent that the choice of contractions included in the first version of the English braille code was excellent, for at least half of them saved over 10,000 spaces each per million words and very few must have saved less than a thousand spaces. However, it has to be remembered that there was a certain freedom of individual choice concerning the use of shortforms so these cannot be included. It is known that Armitage (1874) analysed passages from four works to calculate space saving, so it is surmised that he may have already used these values when the code was being adapted from the French version.
The additions had been decided in committee over the intervening years since 1870 (BFBA minutes) so do not represent a full-scale revision of the code. 1870 had had first choice for the most space-saving contractions so a concentration on middle-range contractions was to be expected. The worst feature was the inclusion of more contractions in the lowest space-saving range, which suggests that the choice was made without reference to frequency.
After revision in 1905 the new code became just over double the size. This was partly caused by the addition of specific shortforms instead of the writer being left with the choice. The decision made for less guessing by the reader and presumably speedier reading. The transference of some meanings from simple to compound symbols to make room for new words is to be commended. Altogether 13 simple upper wordsigns were added. IT saves 10,515 spaces per million words, 9 words were added in the middle range, and 3 each saving less than a thousand spaces. KNOWLEDGE occurs very infrequently and GO and US are so short that they were not worth being included, but other short words earn their place because of their frequency value. Most simple wordsigns were planned to be represented by their first letter to help lessen the learning load. It is not always easy to find a suitable replacement if this mnemonic help is to be retained. On the other hand braille is limited by the number of signs available. 1895 had included such words as RECEIVE, DECLARE, REJOICE, and the 1905 version continued the trend, sometimes seeming to prefer a collection of similar words instead of finding ones of more value. However, the committee members did not have the use of the extensive frequency lists that are available today. 1905 presented a great opportunity and some good decisions were made but insufficient knowledge and some subjective choices led to inclusion of some uneconomic contractions which remain in the code today.
The Americans were hoping that in accepting Grade 2 there would be a reduction of some of the contractions (Irwin, 1931, p.33). Only 9 religious words were eliminated.
Only four new contractions were introduced (British UTC, 1960).
The BAUK questionnaire had indicated that a majority of braille readers preferred no further changes to contractions in words for the time being. Vocabularies, codes and technical advance never stay still so the last chapter will include more information about the possible future of the English literary braille code.