Learning to Read Braille: Barriers, Challenges and Opportunities

Return to Blindness and Visual Impairment

Speech given at the Braille Symposium 2002

Date: 16/04/2002
Venue: Copenhagen

1. Introduction

Whenever I travel I carry braille with me, usually committee papers, magazines and print outs off my office embosser. As seeing people do, I leave a lot of this material on the train seat when I leave. Almost always people chase me along the train platform, sometimes at risk to themselves, sometimes a the risk of missing the train, in order to give me back what I have stupidly left. No doubt this is partly a reaction to the stereotype of blindness; I could not possibly have deliberately left this material behind; I couldn't see it so I didn't notice it sitting there; but it also reflects a very important attitude we have transmitted to the world about Braille. Not only is the skill in learning to read and write Braille precious, the material itself is very precious. Of course that was the case when I started to learn Braille, using a stylus to make symbols one dot at a time. This was the time when you either had to make single mechanical Braille writer copies or you had to make plates laboriously for hundreds of copies.

Since the advent of computer driven Braille printers in the late 1970s I don't think our attitude has changed rapidly enough. Many of us have simply transferred our production techniques from analogue to digital without thinking about what we are doing. I say this in the context of education because we have two supreme challenges to meet:

First, we just want a great deal more Braille, more cheaply
and secondly, we want it more quickly.

This means that the real standard of quality should be in the archive file which will improve as errors are spotted and improvements suggested. We should not expect the same production standard for adult leisure reading as we do for mathematics textbooks for children. The insistence on applying precisely the same production values to all Braille increases its cost catastrophically and, secondly, it slows down the process. There is not much point having a set of football fixtures in perfect Braille if you receive them a month after the season is over. Many organisations in the 'West' have been terrified into submission in this area by a tiny group of Braille Stalinists who call for quality but refuse to consider price and timeliness as components in a definition of quality. They have the privilege of all those through the ages who spend other people's money; they can set any standard they like because they are not having expected to pay. These matters are not the key elements in my presentation but they are vitally important. We do need a culture change in respect of the way we look at Braille and in the way we look at the flexibilities of production allowed by ICT.

2. Learning Barriers

In the first part of my presentation I want to talk about what I think are the five major barriers to the learning of Braille; these I summarise as follows:

Too difficult
Lacks Praxis
Taught in isolation
Conflates reading and writing
Not enough incentive.

Too difficult

If you went up to the average intellectual and said: "It's difficult learning Braille music because the notation gives no idea of the shape of the musical idea" he would be sad; but if you then said: "And we use the letter D to represent the note C, and so on, until we get to A which is represented by an I and B which is represented by a J" he might then raise his eyebrow. Why does it have to be so difficult? The answer to that question is very similar to the answer to the question: "Why is the layout of a computer keyboard so irrational?" The answer is that the computer keyboard is a direct descendant of the typewriter whose array was designed to avoid key clashes between the most frequently used letters; so one of the most frequently used letters, a is pressed with the weakest finger of all eight, the little finger on your left hand. So with Braille Louis didn't have a whole seminar on options and choose two vertical columns of three dots. And when he came to select the dots for letters he used a system of intrinsic logic rather than thinking about maximum differentiation between the shapes of letters. If we started again I would, for instance, find the most extravagantly different signs for the five vowels and move from there. There is no linguistic justification for the relationship between A, K and U or F, L and V; they are just abstractions of the vertical reading of the braille table.

So we have a problem with braille because it is intrinsically logical but not in any way designed for the reader. In print you may well suffer from dyslexic problems with lower case c and g and with b, d, p and q but in braille you have d, f, h j, i and e (two key vowels); and, to add to the absurdity, even Braille Grade One you have differentiated signs for open and close quotes but not for open and close brackets.

That is only the beginning. The system of contraction for the European languages that have it provides an extra layer of complexity. Let us look at a few salient facts:

Most potential braille users of school age are developmentally behind their sighted peers when they arrive at school
Perhaps half of these children have an additional handicap
In most countries these children are expected to fulfil the national curriculum, plus the additional curriculum (Braille, O&M) plus the 'hidden curriculum' (communications, body language, style, grooming etc)
Learning contracted braille severely complicates the additional curriculum and takes time from other curricular elements
The two major justifications for contracted braille, the analogue need to reduce the number of cells processed and the need to save space, no longer apply.

I would argue that as a fundamental educational reform Braille Grade Two ought to be completely abandoned as part of the curriculum and that if the brightest children want to learn Grade Two at secondary school or later they should have that option.

It is now technically possible to provide 'gates' allowing some contractions through and blocking others. This is useful both for learning and teaching and for providing hard copy which meets the needs of a variety of users. It is also now technologically possible with relative ease to provide users either with grade One or Two depending on choice. The key factor here is to separate the metadata and tools for page making from the body text.

There is no longer any justification in congenitally blind, highly talented, high school and often special school educated late middle aged adult males to make inflexible Braille legislation in an increasingly plastic environment. Nor should teachers of blind children be so inflexible, simply to protect their professional identity which is frequently built around the knowledge of braille.

The only argument, which may be worth considering, is the extent to which speed of reading is affected by the degree of contractedness in the code. The problem with this is that such arguments are usually based on testing fluent Grade Two users and seeing how they fare in Grade One; such tests do not, of course, include those who do not read braille at all because the current code is too difficult.

It was just unfortunate that the cassette tape recorder appeared almost simultaneously with the move towards integrated education in 'The West' and that Braille legislators were not alert. The gap between the tape recorder and the first computerised Braille printer and associated Braille translation software was about 15 years. We have a lot of ground to make up.

Lacks Praxis

One of the key findings of research into the way in which children and adults learn about ICT shows that the official line of politicians and educators just will not work. What policy makers want is for people to sit down at computers and learn techniques for word processing, working in spreadsheets and using databases. This is just about acceptable to boys if there is a games or competitive element and technical problems such as 'head crashes' always help to keep up interest but girls are totally indifferent to this self indulgent nonsense. They want outputs, they want to see what the benefits are of going through the process. In other words, people in general but females in particular are interested in learning through praxis, through doing, through content creation, through achieving an objective. They are not interested in learning for its own sake, on the off chance that they will need a technique a few years from now. The problem with braille is that it is much less praxical than print. You can only use it for a very few things. The two major outputs are to access the intellectual content of authors through documents like storybooks and to keep a record of your own thoughts and work or that of your peers. Considering the massive variety of activities which children undertake, this range is very narrow. Seeing children use their reading skills in particular for a wide variety of activities, some of which are sanctioned and many of which are frowned upon by adults but which adults can't control. Seeing children benefit from a wide degree of serendipity; they see shop signs, newspaper headlines, fashion logos, pop videos, captioned cartoons, food labels, magazines for their own age group and, more interesting, magazines for the next age group above them. There is a great incentive to read particularly if it broadens the opportunities for self-expression, non-conformance, and the effective guerrilla war against adults. Braille offers no such opportunities; it is dull, establishment, dictated by adults, narrow in opportunity. In other words, it is a conformance rather than a rebellion tool.

For children who have so little vision that braille is their sole reading and writing medium outside tape, we have to do something more fundamental than drop Grade Two. With whatever subtlety we can command we must teach our children how to rebel. That means serendipity, it means surfing, apparently wasting time, fiddling, idling, getting lost in cyberspace. Blind children already tend to be far too conformist but in a world where even Scandinavia contemplates periods of political rule by parties which rate individual achievement and choice higher than social solidarity, we have to be very careful that the gap between our children and society does not widen even further than it already has. The wider that gap becomes then the less likely our children will be to form effective human relations, undertake a job and be happy.

So, at root, what we have to look at is the assumptions that we make behind the kind of teaching that we do. In an environment where much of special education has, for historical reasons, got a high moral and/or religious content I am asking a great deal but I have to ask it on behalf of the children. If the means of autonomous text reading, braille, becomes more of a hindrance than a help in making a child effective and happy then we will have to consider losing it altogether. There isn't much point in learning braille so that you can stick labels on food packets if the price is not forming a lasting relationship so that you don't need to stick braille labels on food packets; in a sad way the process can become self fulfilling.

I am going to be characterised as provocative, I know it; it happens so often when I look at the fundamentals of a problem; but what I am saying is not provocative; the analysis lies in the reality of the way children are steadily moving, with or without their teachers, away from prescriptive and towards a much freer kind of learning. We spent fifty years fighting for integration in many countries; we don't want to squander that victory by then making our braille learning children second-class citizens compared with their sighted peers.

Taught in isolation

Children who can see read in conjunction with a rich pictorial environment. They also read in a sound world of subtitles and headlines, relating what they see to what people are saying. So when they are walking down a street they will hear people reciting the names of the shops they can see. This is related to the serendipity point I have just made but now I want to go further.

We have all seen questionnaires - some of us might even have designed them - which ask a blind person what is their primary reading medium. This is a fair question when, for example, a bank wants to know whether to supply a braille, tape or large print bank statement but it is a pointless question in a multimedia environment. Most of us do not turn the sound off just because there are interesting pictures on the television; most of us do not turn the pictures off if the dialogue is particularly stimulating; we do not choose between pictures and text when both are provided. We live in a multimedia world and that is the world in which blind children should learn braille. After all, 95% of them will go on to read some print, most of them will be able to gain something from pictures and most of them will be able to hear. On that basis we should be teaching braille within as rich a multimedia environment as possible. This largely means teaching braille on a multimedia workstation with a screen full of good text and images and the voice synthesis or audio track switched on. The objection to this is that people are not truly reading symbolic language if it is supported by speech and/or pictures, but this argument is bogus. Children are not being taught to read and write so that they can operate like Montaigne in his retreat or Proust in his hermetically sealed apartment; we are talking about reading and writing as functions within an environment. Children may need to read and write while alone, without a computer but there is no reason why they should not learn with voice support and pictures just as their sighted peers do. We should not lock them into the braille text with no additional support. We have to think about the whole learning environment and how it supports braille reading and writing, instead of seeing this as a terribly serious, worthy, almost monastic activity.

In doing so we have to face one major political problem. Most legislators think that the ability to comprehend symbolic language is an end in itself; so they do not think, for example, that teaching people with literacy skills to read should involve the simultaneous use of a screen and voice synthesiser. I think that blind children have so much to contend with that the easier it is to grab symbolic language the better. I will be accused of depriving them of the ability to function autonomously in a text environment but my response is that we should make it possible for those who want to maximise that opportunity to do so but it does not mean that children who find this difficult and whose lives will render it irrelevant should be forced into this strategy. We need to be much clearer, particularly with those children who have difficulty learning, where we shift our strategy from educating in process towards training for output. Again, there isn't much point having highly literate blind people who can't function in their environment and can't find a job. Because being able to see widens choice, not being able to see should not artificially narrow choice; the very disability calls for policy makers to be as flexible as possible to broaden options.

Too often we think that the choice is between only using braille and never using braille. It is that kind of rigidity which gives the medium a bad name.

Conflates reading and writing

I now come to a much more controversial point. The development of braille translation software means that children should be able to adopt separate strategies for reading and writing braille. Personally I use a refreshable braille display and a Braille printer. So, when I am editing documents for a standard printer I edit all my text on the refreshable braille display in Grade One so that I know what the layout will look like but for embossing hard copy to read, like this speech, I emboss in Grade Two.

At a more fundamental level, apart from using a Perkins to punch labels for my CDs, I never write braille, I always use a qwerty keyboard. And yet I read Braille Grade Two as fast as anyone I know.

This leads to a very simple proposition which we have all experienced in learning new languages: it is much easier, on the whole, for most people, to read a language than to write it, to listen than to express, to choose from options than to nominate a solution. So it is with braille. Those who can read braille perfectly may not be proficient in writing it; and yet the two processes are taught so closely together that they can hardly be distinguished. In many countries a blind child is taught to read and write braille using braille devices and is then taught to type on a qwerty keyboard. My suggested method would be for the child to be taught to write on a qwerty keyboard, being taught to read its output in refreshable Grade One Braille with simultaneous speech output. We can then consider such options as reducing the voice output support, adding Grade Two contractions into the reading mix and, finally, teaching the child to write braille using a six key device. My problem in arguing for this in an extended form is that I find it very difficult to imagine the objections to it.

The argument for learning to write braille before print used to rest on the need to communicate with yourself before you need to communicate with people who use print as their primary medium but the note taker with a qwerty keyboard renders this argument obsolete. I accept without qualification the primacy of learning to read braille, starting with Grade One but the methodology for storing notes and communicating with others must lean heavily towards an output that is qwerty based, not least because keying braille, particularly Grade Two, gives no real impression of what a finished print document looks like. Children need to be taught a special, though hopefully not large, module in word processing for print output.

Not enough incentive

Finally, in this first section, I turn to the question of incentive, which will bridge us into the second part of the presentation.

Considered alongside the serendipity problem, the absence of casual visual data, the main problem with learning braille from a utilitarian point of view is that there isn't enough braille to provide the incentive to overcome what is a difficult code. This is a problem which teachers and pupils share and it goes back to my opening illustration. Why bother to learn braille when there is hardly any material available in it and where tape and audio are both cheaper and more abundant? I would argue that the act of reading alone, with the faculty of imagination to create the sound world of the text is absolutely vital in our culture; we don't want to hear an actor reading a novel because it is then no longer our novel. But that autonomy of internal expression, of creation, is bought at a very high economic and learning price; and, of course, it does not apply to a large sector of the learning process in science, mathematics, economics, sociology, geography and much of history. It is not so important in ephemera, particularly where the magazine or newspaper is only paraphrasing or expanding items that are ubiquitous in the electronic media. You don't need or want the autonomy of creation when the text reports actuality which you can get on the radio, television, the CD player or the on-line audio clip.

You may have an incentive if, like me, you want large novels to read on long train journeys, where your relationship is with the page but, still, in the world of PDA's and solid state reading devices, of eve more reliable wireless technology, how long will we want to put people through the braille learning process simply for what might soon be regarded as a luxury. The only way in which we can improve the chances of the survival of braille is to find its proper place in the new digital environment, so think it more comfortable than audio plugged into your ear, to think it more comfortable than a piece of hardware that delivers forty characters per line, to think it more pleasant in bed than dragging your PDA from under your pillow. We are going to have to argue for braille in the context of multimodal multimedia and it is with that which I will begin the second part of this presentation.

3. Challenges and Opportunities

In many ways this second part is a mirror of the first. It deals with the technological environment in which braille will be operating in the next decade and offers some solutions to the problems from the first part. I will deal with four topics:

Multimedia and multimodal
The Internet; XML, CSS and translators
Laws and Standards
Design for All and Niche markets

Multimedia and multimodal

If anybody asks me what I want information provision for blind people to be like in the age of rich multimedia I always answer that for all people, whatever their aptitudes and limitations, the law is simple; multimedia should be multimodal. This means that, whatever the initial composition of a piece of intellectual property, it should, as far as possible, be capable of rendition in any mode quite independent of the others no matter how much they reinforce each other in the original. So, if there is a drama on television we would like audio captioning for the blind, visual subtitling and a signing avatar for the deaf, recognition and retention cues for those with cognitive problems and easy switching for those with physical problems.

In the area of visual impairment and blindness there is the general sense in which you need description of static and moving graphics, the ability to arrest scrolling, additional contextual information and so on; but I also mean multimodal in a much narrower context here. I believe that the only real future for braille production, given the audio and modified print market, is if braille is simply seen as one of the outputs of a multimodal production system. When an agency is going to record a book on tape, the reader reads into two microphones simultaneously, one attached to a digital recording machine, the other to a voice recognition system that captures a text file The two files can be tagged so that an audio user can watch the print as she listens or can use print cues to skip to another chunk of audio. The basic text file is established so that it can be rendered in a variety of print sizes and in braille. Equally, if a user requests a piece of large print then it should be derived from a source file capable of producing braille. Instead of managerial structures depending on a medium we should use a model where people are allocated to title production, responsible for all formats: large print, braille, audio, tactile graphics, haptics, world Wide Web link management, bibliographic cross-referencing, indexing, synthetic speech output.

What this requires is the preparation of an absolutely metadata free body or basic text to which applications tools and metadata can be applied. To give you a simple example; if you integrate the page making metadata and tools into a file you are stuck with the page size in the specification, so if you want to alter the size of the braille page or you want to alter the size of the print the pages in the new size are a mess because their headers and other metadata are rigidly stuck in their allocated place in the body text. What we need is to attach tools to files so that they can be applied according to context.

This is not really a code question. We have spent far too much time in the past two decades having esoteric arguments about braille coding rules. What we need to look at is word processing for braille output and the management of files that need to be migrated for variable braille page sizes and variable print sizes.

We should begin to think of alternative format personnel as producers rather than as simple transcribers. They are there to realise a piece of intellectual property not simply to render it in the proper braille code. In any case the two most complex problems facing braille transcribers have nothing to do with code; they are layout and description. On layout there are no easy answers. Quite bluntly, the kind of layout that looks elegant and economical in braille such as the sub header set left with the text starting on the next line in column three just looks ugly in print; and most print layout that relies upon centring and large amounts of white space is not easy to reproduce economically in braille. This is why we need to code our heading equivalents in style sheets so that translators can render these without manual intervention.

As for description, there is no reason why people who work with blind children or who put print books into braille should be any good at describing anything. In the case of teachers, they ought to be taught and examined in description as if they were budding novelists, but a transcriber should be able to work backwards either to author description and indicators or should be able to point to authoritative descriptive devices. If, for example, a transcriber is faced with a picture from an art gallery where the caption is given, he should be able to write a URL for the gallery's catalogue with a page indicator for the actual entry the reader will need; there is no reason why the picture should be described by an amateur, no matter how good, when the cataloguer has already done some of the work.

There are, however, two very closely linked observations here with which I want to finish this sub-section: the first is that painters and photographers usually provide meta commentary on their work, properly believing that the picture must speak for itself; the second is that there are some graphical elements of a picture which cannot be described, even for a highly intelligent blind child. The smile of the Mona Lisa comes to mind. So even multimodular multimedia has its limitations. In spite of that, it should remain our central objective in information provision which therefore allows braille to be included in the alternative format archiving system and service delivery package.

The Internet; XML, CSS and translators

Obviously, the greatest source of information during this Century will be the Internet and its successors. At the moment let us try to think inside the World Wide Web. As you know, the language of the Web can vary but the World Wide Web Consortium, of which the Web Accessibility Initiative is a part, has set a Web page writing standard known as XML, or Extensible Mark Up Language. A mark up language simply attaches labels to pieces of information so you know what kind of information they are. So, for example, if you look at a simple piece of paper and there is a heading centred at the top and a chunk of text split into three by sub headings, then you might well mark the top line "Heading level one", the sub headings as "heading level Two" and the text as "body text". This allows you to post a Web page in a way which separates the content from the way it is rendered so that a variety of Web browsers can render the material with the hierarchy you specified but not necessarily with the precise physical appearance you started with. One browser might render "Heading Level One" as 14 point, the Heading Level 2 as 12-point and the body text in 10-point thus establishing hierarchy purely by the size of the print. Another might use different typefaces, boldness, capitalisation, centring.

The reason why this is important for us is that it allows us, in effect, to lift the mark up language and make rules for the way it is rendered in braille. We could, for example, take a heading level One and say wherever this occurs it should be shown as "Centred text" with a full blank line between it and the next string of text. We might define Heading level two as set left with a blank line separating it from body text. We might then define body text either as being indented at the beginning of paragraphs or all set left with paragraphs shown by a blank line.

The vital fact here is that in doing this we are doing what Web users do in HTML, we are creating what are called Style Sheets, pieces of legislation for a document which say; when you see this, do that. They are sets of rules. We need these so that we can lift chunks of data off the Internet and render it in respectable braille.

Of course, all our problems are not over at this point. The Internet is in many ways seamless so that you will meet all kinds of different documents by going through links but at least your browser and some style sheet construction will manage some of the major headaches but there will be problems with very busy documents that employ, say six different kinds of heading. That makes rendering in reasonable braille a good deal more difficult and we might even need to go into the braille file to make some adjustments but, still, the majority of the text should be all right.

What we need, however, are much more sophisticated translator tools to put documents from the Internet into relatively simple braille that is unambiguously readable and elegant. Let me give you some examples of translator features which would be very helpful:

a column de-scrambler
A device to detect hard carriage returns that produce one and one third braille lines per print line
A status change detector which stops body text running directly on from sub headers
An over-ride device which stops the translator flipping between braille grades according to the origin of pieces of a now integrated document.

They are just four but there are many others but the poor translators are under permanent siege from the code fanatics.

If we don't get hold of this problem we will download stuff from the Internet and produce bulky and very ugly braille that will not help our cause. This should be an international effort. It is about time that the braille users and their intermediaries broke the stranglehold of the code fanatics and allowed us to establish some good conventions on tools for rendering Internet content in good braille. If we are not careful, however, all our braille and its rules will be settled by the United States, which knows not the tilde or the circumflex. There is a good core of work going on in the World Wide Web Consortium but it needs some expertise in the establishment of braille style sheets to make elegant translation easy.

Laws and Standards

Which leads neatly enough to laws and standards. In the context of economic globalisation the world is moving away from the use of legislation and regulation to maintain certain standards of service. Instead, the need for products to be globally usable is leading towards ever more global standards and standard setting bodies. So what we may lose in terms of legislation from parliaments we may gain from international standard setting bodies.

In the field of English braille there has been a wrangle going on for more than ten years now about the establishment of a common braille code for English using countries which spans literary braille, maths and basic science and computing. Global standards bodies are slow but not that slow and they will not wait. Sooner or late one of the major software companies, pushed by the ADA or 508 in the USA will include a braille style sheet in a software package and it will become the world market default; goodbye to all the self important braille legislators!

If educators of blind children who want braille to remain a vital reading medium do not become involved in these issues then they will lose out by default. I am staggered at the very tiny impact educators have had on braille coding questions in my own country. I cannot understand how self-perpetuating oligarchies can go on claiming to speak for us all. This picture varies from country to country but I think that although our braille codes may be different because of tradition, the use of accents, the occurrence of letter groups, we ought to be able to reach an international agreement on Braille Grade One code and also on some basic style sheet rules for rendering Internet content with economical and intelligible layout. If we do not do this, then one of my major educational objectives, the ability of children to surf the Internet, will be seriously in danger. Anybody who has worked with a braille display will know the frustration of acres of useless metadata and the presentation of the data in almost unintelligible fragments like tiny islands within lakes of empty space.

Design for All and Niche Markets

Before I draw conclusions I just want to say a few words about the concept of Design For All, or Universal Design and its relationship to niche markets.

The concept of Design For All is a good anthem but, of course, nobody quite means it. We know that you can design goods and services for almost all people but that there will always be a small percentage whose needs can't be met from one design. The idea of Design For All is made much more possible, however, because of built in customisation features. There is, for example, a difference between a park bench and an adjustable office chair, or between a printed book and an on-line file where you can alter the size and font of the print. So the better the design and the better the customisation tools, the higher the likelihood that design will meet a greater number of needs.

In our sector we are accustomed to dividing equipment and materials between those which can be met by general design and those that we can obtain from small, specialist suppliers in the niche market of visual impairment. So, for example, we can buy some audio books and standard tape recorders from general suppliers but we buy DAISY books and special players from the niche market.

At first glance braille would appear to be a particularly special piece of the niche market. This is not a simple argument. It is perfectly possible, I think, to put braille Grade One into a general word processing package, as a character set or font; and it would certainly be possible to produce Grade One as a style sheet.

On the other hand, we do need to preserve specialist code writers. We will still need Grade Two; we will also need new codes as we reach international agreements; and there are, of course, specialist codes for mathematics, science, computing, chess and music. As commercial organisations in the niche market, in the supply of embossers, for example, produce free translation software, it is likely that the coding of braille will depart from the monopoly of braille authorities. This has its dangers as well as its opportunities. We have to start thinking about braille as a standard which is derived and changed according to international rules; the process must be much more open.

4. Conclusions

In conclusion I want to put what I have said into a slightly broader context so that braille is seen as part of the whole learning process. I have already said that I am passionately committed to learning braille within a rich, stimulating environment. I see our child with her hands on a braille display, looking at a screen of large sized letters and lively pictures. I hear a piece of real audio that is tagged in to the braille and the screen image; I see the child smiling.

At least we must be consistent when we think about blind children. Our fundraisers are forever reminding the public how difficult it is to be a blind child and how brave they are to achieve what they do, against the odds. Now whether or not you like this kind of rhetoric, whether you want to characterise what children do as brave or simply a response to the inevitable, the fact is that being a blind child is irritatingly, chronically difficult. It is also, for the most part, very dull, almost Calvinist. There is so little room for experiment and improvisation, there is so much pressure to get on equal terms which means always doing more than your sighted peers; it means not seeing pop idols and footballers at all, or not seeing them very well. For many children who will learn braille it means never being sure what you will see and not see. I remember when I was a braille learning child who could see a little that as the game went away from the end of the field where I was standing it got dimmer and dimmer; first I lost the ball, then finally lost all the players; it was like watching in fog. I suppose I was stoical about it then but we lived in an era when being stoical was admired. Since then, on both sides of the Atlantic, and even here in Scandinavia, we have been educated to believe that the primary purpose of life is self-fulfilment and self-expression, the exercise of maximum choice, a downgrading of solidarity. That is a particularly difficult agenda for people who must always to a certain, measured and visible effect, be dependent. Of course we are all dependent but to argue that we are all equally dependent would be false. Only those who know that they are in a strange place and that they cannot go to the toilet without asking somebody else for help know what it's really like to be blind.

And so, I want to leave you with two thoughts. The first is simple, it is that if we are going to get the balance wrong in the way we teach blind children let us get it wrong in favour of fun and pleasure. If you are not quite sure whether to do another page of the exercise or play a game; play a game. One of the really important metrics of the education of blind children is the degree of bodily relaxation and the number of minutes per day that they smile.

The second point is much more complex. If we are to treat blind children as equals in everything, as equal competing individuals, braille has to be liberating rather than conforming; it has to set children free, not tie them to whatever textual canon the politicians and technocrats determine. In the analogue age this was technically difficult and almost economically impossible. It is now achievable with a considerable degree of inventiveness but it will not be expensive. I doubt that it will be the technology that is the great barrier; it will be ourselves.

We have spent too much time concentrating on the process, on the code, on the sacredness of the braille text; now it is time to concentrate on the incentive, on the output, on the irreverence of the accidentally discovered.