1. COMPARISON OF VISUAL AND TACTUAL METHODS OF READING
2. QUALITY PROVISION
Quality of Materials
Size and Shape of Dots
Spacing Variables
Layout
3. TECHNIQUES OF READING
Use of Hands and Fingers
Types of Hand Movement
Characteristics of Movement by the Fingers
Can the Use of Hands and Fingers be Taught?
4. PERCEPTUAL FACTORS
Analysis of Errors Within Words
The Word Method of Learning to Read
Recognition of Single-cell Braille Characters
Effect of the Number of Dots in a Cell
Position of Dots Within a Cell
The Effect of the Use of Contractions
The Effect of Word-length, Familiarity, and Orthography on Recognition Thresholds for Braille Words
Unit of Recognition
5. DEVELOPMENTAL FACTORS AND THEIR EFFECT ON BRAILLE READING
General Development and its Effect on Braille Reading
Short-term Memory
Strategy Choices by Young Braille Readers
Strategy Choices by Fluent Braillists
6. STRATEGIES FOR IMPROVEMENT OF THE RATE OF READING
Changing the Code
Diagnostic Tests
Training in Rapid Reading
7. SUGGESTIONS FOR FUTURE RESEARCH
8. CONCLUSION
When comparing visual with tactual methods of reading Gibson (1962, 488-489) wrote, "Succession enters into the operation of both senses. The eyes normally fixate in succession just as the fingers explore in succession. Both senses are active". This statement needs qualifying because the manner of succession in the two methods is different. It is commonly thought that visual reading takes place during a smooth, even focus of the eyes along the lines of print, and that touch reading is a similar process except that information is gathered via the nerve endings in the finger pads instead of using the eyes. A better understanding of the mechanisms involved is necessary before the advantages and disadvantages of reading in the medium of braille can be recognised.
According to Fry (1963) and Watts and Buzan (1973), movement of the eyes during visual reading takes place in a series of 'jerks', that is, short movements from left to right interspersed with short pauses, and the amount taken in at a fixation is more variable than the fixation time depending on the ability of the reader. Therefore, the more that can be taken in at a fixation, the fewer will be the number of fixations necessary and the quicker the reading rate. Foulke (1982, 168-169) referred to "the relatively large field of view of the visual system, which makes possible the observation at one time, of a relatively large number of symbols" because of "the high compatibility between display and the perceptual system".
The tactile system is not as well adjusted for reading as the visual sense in that the field of 'vision' is much reduced because reading is a character by character succession without the fixations of visual reading. In addition, the braille reader cannot glimpse ahead. For example, "In his cap___" could become "In his capable hands" but could alternatively become "In his capacity as ...". Visual readers may be aware of misprints, though often these are not picked up because content meaning is dominant. In braille reading a missed dot can be crucial to the meeting. For example, in a watery setting "The rush..." (first letter ) may anticipate ideas such as tumbling waters, whereas if the lower dot is omitted () it becomes "The hush" anticipating possibilities of a silent pool or the evening calm. It could also be most disconcerting to read "He is dead" instead of "He is deaf", caused by the sign for F being reversed in error. These and other perceptual problems to be discussed later in this chapter contribute to some serious limitations for touch readers.
Comparisons of rates of reading between samples are only reliable when variables such as age, intelligence and the effect of rate on accuracy and comprehension are taken into account. However, average silent visual reading rates for adults are quoted by the de Leeuws (1965, p.28) as 200-250 w.p.m. and by Foulke (1982, p.172) as 250-300 w.p.m. Williams (1971, p.116) concluded from a nation-wide survey of braille reading pupils aged 10 to 16 years that "with narrative requiring to be read and not just skimmed, a braille reading rate of 100 words per minute can be considered average, and extending this on either side, a rate of 80-120 words per minute can be considered as fairly normal".
Whatever the skill, playing the violin or building a brick wall for example, two aspects need to be considered before successful results can be expected. These are the quality of materials used and the techniques needed for execution. The same is true for braille reading.
A badly produced text on poor paper, with unhelpful spacing and poor quality of printing, both in depth of inking and size of symbols, can all cause difficulties and irritation in visual reading. Publishing embossed print has its own problems. One abortive effort was the use of solid dot braille. The intention was to use a stronger yet lighter weight paper with durable dots applied to the surface. It was unpopular with adult readers because the hard surface of the dots was tiring to read (personal comment by braille users) and the dots could be accidentally removed; indeed, in the classroom the paper sometimes became torn and "fingers" enjoyed picking off the dots in idle moments, with disastrous results for the braille content. Two weights of paper are currently provided by the RNIB for personal use, the lightweight one being more typically used for short term work and the heavyweight one for material of more lasting use. "White rag" paper is used for production of "The Braille Radio Times" whose useful life can be little more than a week at most. Brailon consists of plastic sheets used for the production of multiple copies from a braille master copy embossed on manilla paper. It has lasting qualities, but is easily creased and hands tend to become sweaty in use. In spite of a large amount of research on many aspects of the use of braille, nothing had been published on the detailed characteristics of these materials until 1986, yet to have the most suitable of materials is crucial to the braille reader and writer.
In 1986 Cooper, Davies, Lawson-Williams, and Tobin compared these three types of paper together with three synthetic ones for physical characteristics such as weight, thickness, bulk, porosity, smoothness, strain, and "burst". Porosity is important for unless the air can pass through the paper the surface can feel sweaty in a warm atmosphere or occasionally if the reading is carried out in stressful circumstances. According to the 16 subjects who took part in the investigation, there was no overall preference for all the circumstances tested, but in descending order heavyweight manilla, lightweight manilla, and white rag were preferred to the synthetic materials. Brailon was considered to have the smoothest surface.
Though the study was not conclusive, the investigators felt that it would merit further work with a larger sample, longer reading passages, and some more accurate means of assessing the "feel" of braille. The experimenters pointed out (ibid., p.327) that this more detailed knowledge is important "especially if new electronics and other devices are to be offered to blind people with braille as one of the major forms of output".
In 1890 (p.11) Fowler wrote "It is the experience of many that the sharp conical dots, though very distinct at first, soon irritate and confuse the touch; the dots should therefore be made dome-shaped, so as to present a smooth surface to the finger". The symbols of old braille books printed about the turn of the century by the British and Foreign Blind Association are sometimes colloquially referred to as "knitting needle braille" and have received very favourable comments by those blind people who have been able to see copies (personal comments). The master plates were punched on copper sheets and the results probably approached the quality advocated by Fowler. [The writer owns plates used for the libretto of one of the arias in Handel's "Messiah".] Later machine printing shows dots that seem less wide in diameter though still adequate. With the coming of computerised braille which can be easily and speedily replaced there has been a tendency to produce braille that is sometimes of inferior quality and less durable. It is important that dot shape shall be of maximum comfort to the reader.
It has also been found that some beginners and also some whose touch is less perceptive prefer to use the expanded cell, often referred to as "jumbo dot". The height of the dots remains the same but the diameter of the base of the dots is increased as well as the spacing between dots. It would seem that individual differences are important when deciding on the size of cell to be used. For example, a person with less acute touch might prefer the expanded cell at the beginning stage. Tobin (1982), in his self-instructional reading scheme for newly blinded adults, used the expanded cell type of braille for the first book, books 2 and 3 contain identical material so the learner can choose when to progress from jumbo braille to standard size braille, and book 4 is in standard size braille. Young children starting braille have smaller finger pads to cover the symbols, and using the larger cell might encourage "scrubbing" instead of a smooth left to right progression.
No information exists to show how the three variables of spacing of dots within a cell, between cells along the line, and between lines, were determined by the BFBA when braille was first introduced into Britain. A study of braille readability was carried out by the American Commission on Uniform Type in 1920, but no experimentation on spacing variables was carried out until that done by Meyers, Ethington and Ashcroft (1958). Rate of reading in words per minute was the measure for testing 275 blind children on the three variables of dot spacing within cells, between letters and the space between lines. Three values for each variable were selected and material was read in all possible combinations. The middle values for each were found to be the most readable and corresponded closely to the values already being used. Comprehension was controlled. It was suggested that the experiment should be replicated but with a considerably longer period of reading time and with values not included in the current investigation.
The layout of braille material is similar to that of inkprint in that the material is read from left to right and set out in paragraphs. When books were expensive to produce, blank lines were never left but paragraphs were indented, each starting in the third space. Whereas it is an easy matter for the visual reader to move quickly to another part of the text, the blind reader can at least find the next paragraph by running the reading finger down the left side of the page to find the indentations. An indented paragraph preceded by an italics sign indicates a side heading. With the coming of computerized braille the leaving of lines where appropriate may become more prevalent.
An investigation was carried out by Hartley, Tobin, and Trueman in 1987 to determine whether headings were helpful and in what form. The findings were not conclusive. The writer suggests that the content of headings merits attention. Their purpose is to indicate the following subject matter, and therefore, depending on context, a single word can be meaningless and too long a heading wastes valuable reading time. Braille's own view seems pertinent here: "Since our methods of writing and printing take up a lot of space on paper, we must compress the thought into the fewest possible words" (Coltat, 1853, p.16).
The most obvious external differences between braille readers are the ways in which their hands and fingers are used. These variations include which hands are used, how they are held, characteristics of movement involving type of progression from left to right, regressions, scrubbing movements when difficulties are encountered, and even erratic movements employed when knowledge of the code and/or reading efficiency are insufficient. It is not surprising therefore that studies of these aspects have received much attention (Holland and Eatman, 1933; Fertsch, 1946; Kusajima, 1961, 1974; Hermelin and O'Connor, 1971A, 1971B) in the continuing hope that analyses might lead to improved techniques and ultimately an increase in the rate of reading.
Hand movement is a very individual matter depending on such factors as the effect of asymmetry in the brain (which determines which hand shall be dominant), the relative sensitivity of each finger, and possibly the training, if any, received at an early stage of learning. Some read with the left or right hand alone, with one hand merely marking the place, and some use both hands at the same time. It is usual for one or both the forefingers to be the reading finger(s). In 1982 (p.202) Foulke quoted his experiment of (1964) in which subjects read passages with each of the fingers alone on either hand while reading ability was measured against accuracy and time. The results showed rapidly diminishing ability in the progression from the forefinger to the little finger. All the fingers showed some sensory capacity. Contrary to expectation the dominant hand had no connection with whether the reader is left or right handed in everyday activities.
There is also variety when both hands are used at the same time. Beginners sometimes hold both forefingers side by side sometimes touching or having a short space between them. It has been hypothesized that for left handed readers the right hand may pick up some information that is confirmed by the following and dominant hand. For right handed readers the role of the left hand may be to check and reinforce what has been sensed. In both cases the less active hand can be used to mark the beginnings or ends of lines. The most efficient method seems to be the use of both hands but working independently. The left starts reading, the right takes over somewhere along the line, and while it completes the line the left finds and then starts the beginning of the next line. Most readers have one hand slightly more dominant and this will determine how far along the line the right hand takes over. The most obvious advantage of this method is the time saved in the return sweep to find the next line and obviates the consequent interruption in the sense of what is being read. The loss of 6-7% (Fertsch, 1946) of reading time taken up by return sweeps is a serious matter and no doubt is one of the factors that leads to braille being a slower reading medium than print.
Moving picture records of fingers reading braille (Holland and Eatman, 1933; Fertsch, 1947) demonstrated characteristics of the less able readers which, though intended to help, may in themselves cause further problems. When the reader is not sure if a word or words have been interpreted correctly it is natural to regress for one or more words before continuing. It is usually found more convenient for this to be carried out by the left hand while the right hand keeps the place, but both hands being used together for the purpose can be regarded as more typical of performance by a poor reader. Sometimes regression becomes a habit, particularly by the more hesitant reader. If the letter or word presents difficulties the reader may resort to "scrubbing" the symbol or symbols. More pressure is used, some loss of direction may occur as the usual left to right progression is interrupted and often there is a break in concentration. There is also a tendency for beginners to lose the line. This is sometimes due to the fact that it is a more natural movement for hands to move in a curve equidistant from the body and learners therefore need to become used to working along straight lines. Good readers show an even, steady progression.
In 1978 hand movements of school children aged 10 to 12 were filmed for a demonstration by the writer to show different methods of hand use at a braille workshop for teachers of blind children. It was not conducted under experimental conditions, but was used as a talking point to show that there are considerable individual variations in techniques of reading, in addition to knowing which hand or hands are employed. It is the writer's belief that teachers are sometimes so involved in checking the accuracy of oral reading that they sometimes leave little opportunity to observe individual differences of technique.
From a total of approximately 24 children aged between 10 and 12, 7 were selected showing a variety of reading behaviour and some of their performances are briefly described here. It was noticeable how techniques were affected by the size of hand and more particularly the length of fingers.
Lee had very large hands which he found difficult to adapt to the reading of small symbols. The screen seemed filled with fingers and thumbs and the thumbs were sometimes used under the fingers as props to propel the hand along, resulting in rather jerky reading. He was not alone in using the thumbs in this way. At other times the left thumb was held higher than the other fingers. Sometimes readers hold the fingers not being used above the reading finger so that the whole hand is slightly tilted towards the thumb. It seems a tiring position but could be caused by the portion of the finger pad nearest the thumb being the most sensitive part. Lee's reading fingers kept in contact with each other even when moving to a new line.
By contrast, Amanda had short fingers with the pads held very flat on the page making the most of this surface of sensitivity.
Toni used only the right hand, yet according to her previous teacher she was helped during a whole year to use both hands on different lines very successfully. It is interesting that she had reverted to the use of one hand, probably because the habit was more fixed or she may have found using the addition of the left hand slowed her down. It darted across to help if there was a difficulty.
Jean Pierre's method was unique. He was a left handed reader and the hand was rotated 90· so that the reading finger, the second in this instance, pointed towards the right. Fingers 3 and 4 rested lightly on the page above the line being read and finger 1 was below the line. All the fingers travelled lightly towards the right hand which indicated the end of the line. The pupil had presumably found his own best method, and though unorthodox it seemed successful.
Jane showed two-handed reading, both hands reading to the mid part of the line, the right ending the line while the left moved to the next line. She did not read the first part alone with the left hand so had apparently developed the most economical method for a reader using both hands, the left one being less capable.
Enough has perhaps been included here to show that some movements are helpful and some provide difficulties which hinder not only speed of reading but are also liable to contribute towards inaccuracies.
It has been shown in Chapter 7 that in several experiments carried out concerning hand use, the results showed conflicting evidence over which hand would give the best results in reading braille. In 1984 Millar suggested that a pattern could be seen showing that there appears to be a tendency for beginners to rely on texture and show no hand advantage; while still attending to spatial coding and physical characters of the signs readers tend to use the left hand; and highly proficient readers who use verbal strategies prefer the right hands. Her experiments (1984, p.85) showed that "the notion of a generally 'best hand' for braille is untenable" and that (ibid., p.84) "two-handed reading is superior to reading by each hand alone". This information has been repeated here because it shows the dilemma for those attempting to teach hand use, bearing in mind Fertsch's findings (1946, p.131) that "reading habits become established about the time a pupil has reached the third grade and do not change noticeably with increase in reading experience". This finding puts the onus on teachers of young pupils and the process is more complex than would at first appear.
Hand-use training seems to be a neglected aspect which teachers often leave children to discover for themselves. It needs to be carried out with a proper understanding of what is involved or harm can be done. For example, when the writer selected another group of pupils again for filming of hand-use a few years later, it proved very difficult to select a suitable left-handed reader. This was so unusual that enquiries were made, leading to the knowledge that one conscientious teacher had been training the young children to use their right hands because right hands were used for most activities by children with vision.
Hands are usually held slightly arched and the forearms should bear their weight. Children will generally choose to use their most dominant finger for reading and a few may use the second finger held close to the forefinger. This presumably checks or adds more information but, depending on the relative length of the fingers, may cause a less relaxed way of reading. this is because the extra length of the second finger causes it to be more arched so that more of the tip rather than the pad is in contact with the paper. Each hand could be temporarily tried out on its own to see if two-handed reading might eventually be possible thus eliminating the time used for return sweeps. If the dominance of one hand is very marked the other should still not be neglected, for being able to read even one word at the beginning or end of a line would save valuable time. However, this method is not suitable for all readers. Some children start one way and never try any other method, so a little encouragement while the teacher and child are finding the best method should be very worth while. Allowance has to be made for the effect of hand dominance (see previous chapter) and also the possibility that the sensitivity of the forefingers may vary. There are many varieties of hand use and the individual's best must be sought, for, apart from comfort, it is imperative to find the way that will promote the best rate of reading. Late beginners, including adults should benefit from similar training.
Any analysis of braille reading problems should include a knowledge of the mental and perceptual abilities of the reader and problems presented by the medium itself. Ashcroft (1960) considered that the physical mechanisms are extremely important in reading braille, for in part it is a tactual-kinaesthetic process, "However, the most important aspect of the problem would seem to be the accurate perception of ideas from the printed page. No matter how good the mechanics of reading, if errors plague the reader, the result can be neither efficient nor effective in obtaining ideas from his reading" (ibid., p.22).
Ashcroft provided 12 paragraphs of increasing difficulty, each forming a story to be read orally. Specific interest was centred on pupils in Grades 2, 4, and 6, but data were also obtained from pupils in Grades 3 and 5 so that trends could be monitored. 728 pupils took part in the experiment and each child read until 10 successive errors had been made Oral reading errors were examined in terms of orthographic features in the code with the following results shown in descending order of frequency of errors (ibid., p.53):
Rank Title
1 Shortform words.
2 Multiple cell contractions.
3 Combinations of orthography.
4 Lower contractions.
5 Upper contractions.
6 Full spelling.
7 Simple upper wordsigns.
Some of the shortforms were unfamiliar because of infrequent occurrence, and because some had so many letters excluded that they became a burden on memory causing a slow-down rather than a means of increasing speed of reading. This is reminiscent of the similar problem included in the reading of Lucas type (see Chapter 3).
Multiple cell contractions seem to have caused problems in two ways. The extra dots sometimes caused perceptual problems, and mistakes were made because of the varieties of position of dots in the previous half cell making different meanings which had to be learned, e.g. work, word, world.
Lower contractions proved to be the most difficult of the single-cell contractions. They can be confused with the same signs occurring in the upper part of the cell but which have different meanings; the dots occur in the lower part of the cell and these, as will be discussed later, seem to be the ones most likely to not be sensed; and they represent a multiplicity of meanings according to position in a sentence. For example:
, |
; |
: |
. |
! |
() |
" |
" |
||||
deci-mal point |
? |
||||||||||
ea |
con |
dis |
en |
in |
|||||||
bb |
dd |
ff |
gg |
The sign for COM and the apostrophe sign are made from the lowest dots in the cell. Ashcroft recommended that the double letter signs should no longer be included in the code. However, one aspect seems important to the writer. By their inclusion a word can become less cluttered with dots, and therefore may be more legible.
The upper contractions include whole and part-word signs. AND, FOR, OF, THE, and WITH are frequently occurring words but Ashcroft found that they caused more errors when used in words. Perhaps this is because they contribute to the clutter of dots in a word, or they are so frequent as words that they are not always recognised at once as partwords. Partwords, nevertheless, did not cause much difficulty.
The words in full spelling caused little difficulty and words represented by a single alphabet word were the easiest of all, though some were not recognised because they occur so infrequently, e.g. for KNOWLEDGE. The fact that individual letters are recognised more easily emphasises the fact that though contractions can save a lot of time when familiar to the reader, it is the multiplicity of meanings which seems to cause problems.
In addition, Ashcroft categorized the error types and the following table (ibid., p.43) should be useful to teachers, particularly those engaged in remedial work.
Table 4. Table to Show Distribution of Eight Braille Error-type Groups. (Ashcroft (1960), p.43, Table 6).
Error Group |
Number of Errors |
Percentage of Errors |
Missed Dots Ending Problems Reversals Added Dots Association Gross Substitutions Up and Down Alignment Left and Right Alignment |
1702 1599 1434 1392 1358 1335 1140 973 |
15.6 14.6 13.1 12.7 12.4 12.2 10.4 8.9 |
Totals |
10,933 |
99.9 |
To get a clearer understanding of the problems involved these errors may be classified as follows:
Perception:missed dots, added dots, ending errors;
Orientation: reversals, vertical alignment, horizontal alignment;
Meaning: association errors, gross substitutions.
The results showed that space-saving devices contributed substantially to the difficulties encountered, also the failure to suspend judgement until the whole of the symbol had been sensed, and perceptual errors, particularly of missed dots. By using continuous prose in the form of short stories Ashcroft's subjects were likely to have been well-motivated, and the better readers would have been able to make use of context cues which is not possible when single letters or words form the test material. His analysis of errors provided a wealth of insights into problems encountered by young learners when reading braille. In addition, he suggested ways of improving the code, gave ideas to help teachers, and made suggestions for future research. These included (ibid., p.89) a consideration of the effectiveness of different approaches to teaching; evaluation of space-saving devices in terms of reading and comprehension; development of means of increasing the rate of reading by scientific evaluation of the code itself; and recommendation of the use of controlled testing of progress in reading.
The word method of reading (a recognition of whole words rather than a synthesis of individual sign meanings to make a word) was the acceptable practice of teaching braille reading in America on the advice of Maxfield (1928). Nolan and Kederis (1969, preface) set out "to study factors in braille word recognition in order to delineate more clearly the cues that make braille reading by the whole word method possible" for blind pupils in a series of 9 related studies.
A tachistotactometer was used which was capable of exposing characters for controlled periods of time. The braille characters were punched on plasticized paper and they could be pushed up through a line of holes corresponding to 36 braille cells. The use of the tachistotactometer could be criticized in that it does not conform closely enough to the braille reading situation. Instead of the reading finger getting stimulation from the progress along the line of characters the machine raised the dots up to the fingers, so reduced finger movement occurred. This view is supported by Foulke (1982, p.184). The dot locator (the 6-dot cell) was included for testing single cell signs to help distinguish lower signs from the same shapes in the upper position of the cell, but for some readers the resulting conglomeration of dots may have been confusing until they became used to this method of presentation.
The investigation carried out by Ashcroft in 1960 had provided considerable information concerning the difficulties in recognition of individual characters within words. Nolan and Kederis (1969) also measured the time taken to recognise single cell signs. Only the total times were recorded which represented times for recognition of the character plus its naming. 36 subjects in Grades 4 through 12 took part. The exposure times were gradually increased by steps of .01 sec. until all the characters were recognised by all the participants. The results were tabulated in ascending order of mean recognition times (ibid., p.61) and the range was from .02 to .19 sec. Though taking longer time, the slow readers (study 8) showed a similar difficulty order. Foulke (1982, p.178) pointed out that it was not possible to adjust the tachistotactometer used by Nolan and Kederis to measure the threshold values for a few of the characters which required the shortest recognition times. The minimum time possible by the apparatus of .02 sec. was therefore given in these instances. The order of difficulty was found to be similar to that found by Ashcroft both for fast and slow readers.
The recognition times for characters increased according to the number of dots in a cell. This is demonstrated in the following short table:
Table 5. Table to Show Recognition Times of Characters Grouped According to Number of Dots (Nolan and Kederis, 1969, p.62).
Number of Dots |
1 |
2 |
3 |
4 |
5 |
6 |
Time (secs.) |
.030 |
.033 |
.058 |
.091 |
.128 |
.190 |
Excluding the 2 dot configurations, it was found that "characters having a greater space at the bottom and/or the right required 22% more time to be recognised", the lower signs requiring 55%, and dot 6 was missed more than those in other positions in the cell (ibid., p.63). No suggestions were made as to why this should be. Braille is read sequentially from left to right so the finger reaches the left hand dots first and information gained from the left side of the cell must be held over until the whole of the cell has been covered.
Another possibility is tentatively suggested by the writer: braille is set out in straight horizontal lines but it is natural for the hands to move in a curve equidistant from the body; the reading finger is shorter than the middle finger, and these factors result in the fingers being held in a slightly arched position at approximately 30· from the page. There must therefore be a tendency for the finger pad to be held sloping slightly downwards which may result in a more definite recognition of the upper position of the cells and the lower parts of the pads hardly touching the page or not making contact.
The signs seemed to be recognised as dot patterns and when dots were omitted the sign was confused with another one with some similarity. Table 8 (ibid., 65-66) is very revealing. For example, X was confused with M , ING with U , and Q is shown to be confused with its parts and .
It is interesting to note that these tactual aspects were taken into account by Louis Braille approximately 170 years ago when he was first considering the make-up of his code. It has been shown (Chapter 2) that when he selected 10 characters from the 15 that were possible using the top four dots in the cell, he omitted , , and . In addition, when he added dots in the lower positions to make further symbols the only alphabet letter to be made with the addition of dot 6 was W . The remaining signs of the first line with dot 6 added were used for accents. These signs were used for one-space upper wordsigns incorporating dot 6 when the code was adapted for the English language in 1870.
Contractions were invented to save space and therefore reading time, and also the cost of production. If the learning load is not too great and if they are sufficiently known, their use is of great benefit, even though the wisest allocation of meanings was not always made (see Chapter 6). A main difficulty is the insufficiency of meanings that are possible using the 2 x 3 matrix. Solutions have been the use of the same sign in the initial, medial and final positions of a word, each with a different meaning; if a sign is 2 dots high it can be used in the upper or lower position; and a third device is the placing of a dot or dots in the half cell immediately preceding the sign. These strategies all present extra learning compared with visual reading where only 26 letters have to be remembered in their upper and lower case forms. Two examples will demonstrate the extra learning involved:
do |
day |
dis initial |
dd medial |
fullstop final |
4 |
every |
ever |
ance |
ence |
5 |
It will be noticed that the 33 composite wordsigns (beginning with dot 5, dots 4-5, or dots 4-5-6) include the first "sound" in the string of letters. However, the final signs, now known as composite group signs (beginning with dots 4-6, dots 5-6, or dot 6) use the last "letter" of the string. This divergence from the norm may cause confusion to some readers, perhaps because they may be anticipating the composite group signs to also begin with the first sound of the string: for example:
Composite Wordsigns |
Composite Groupsigns |
PART WORD MANY |
ANCE NESS ALLY |
These examples are included to show the precise knowledge that needs to be acquired before contractions can provide the intended help. Contractions concentrate more information per character and thus in principle speed up reading.
In study 4 Nolan and Kederis (1969) studied the influence of contractions upon recognition thresholds for words. Lists of words were provided containing one or two contractions in familiar and unfamiliar words. Upper and lower contractions were included. Although lower contractions can cause problems, Nolan and Kederis found that "familiar words having lower contractions are recognised more easily than familiar words having dots in all the rows of the braille cells" (ibid., p.99). It is suggested that this is probably due to the fact that a more open space occurring between signs containing more dots makes for easier recognition. Unfamiliar words with lower contractions were found less easy to recognise.
The position of the contractions is also important, for Nolan and Kederis (ibid., p.99) found that in familiar words contractions in the medial position are the easiest to recognise and those at the ends of words the hardest. It is suggested that this may sometimes be due to Ashcroft's finding (1960, p.66) that meanings are sometimes guessed before the whole word has been sensed. In unfamiliar words the difficulty diminishes as the position of the contraction moves from the beginning to the end of the word. A likely explanation may be that the meaning is being synthesized as the word is being covered but no guessing occurs. It would seem that recognition of words with the help of contractions is more complex than many have recognised.
The Effect of Word Length, Familiarity, and Orthography on Recognition Thresholds for Braille Words
High school students from three residential schools who read braille with above average comprehension were selected by Nolan and Kederis (ibid., p.72) and then ranked for fast to slow performance. From these, 15 students were selected from each of the upper and lower thirds of the distribution. A multivariate design was necessary. Students read separate characters and also familiar and unfamiliar words of 3, 5, and 7 letters in length, the presentation times being systematically increased until each subject could recognise all the words. In addition, the cover time (time taken until at least one finger had encountered all the letters), and the synthetic time (the sums of recognition times for all the letters in the word) were recorded.
The following (ibid., p.25) were among the findings: the effects on word recognition times of increase in word length and decrease in familiarity augmented one another; the single and combined effects of the variables were proportionately greater for slow readers; and the order of legibility of braille characters was the same for the fast and slow readers. It was also found (ibid., p.81) that of the 36 words read by 30 students (total 1,080 words) 111 words or 10.27% were recognised before the reading fingers had covered all the letters in the word. Other cues must therefore play a part in recognition.
The experiment was replicated with elementary-aged pupils (Study 7) and with pupils with lower intelligence (Study 8), so comparisons could be made demonstrating the development of reading skills at this level. It was found (ibid., p.43) that "the development of reading skills is retarded and reading, even at the upper elementary level, may proceed in a fairly mechanical manner. Once this basic maturational process has been concluded, the child is free to make rapid growth". This overall picture was later to be refined in experiments carried out by Millar (1984; 1985; 1988) which will be referred to later in this chapter. It needs to be remembered that the Nolan and Kederis experiments described above were carried out with the use of the tachistotactometer which does not provide the natural reading situation so the results need to be interpreted with caution.
The most striking result was the difference between the synthetic times and the times required for recognition of these words. To take one example (ibid., p.79, from Table 15): the mean recognition time for an uncontracted familiar five-letter word by a fast reader was .63 sec. and .42 sec. for synthetic recognition time. The corresponding times for slow readers were 1.11 sec. and .79 sec.
The last result had a profound effect on teaching method for it implied that "the process of word recognition appears to be a sequential one in which word recognition is the result of the accumulation of information over a temporal period (ibid., p.39) and that "whole word reading" is not characteristic of the braille readers studied and that the perceptual unit in word recognition is the braille cell" and this seemed to be the answer that Nolan and Kederis had set out to discover. It was to be remembered that all the experiments carried out by Nolan and Kederis (1969), except for Study 9 on the effect of character training, involved testing by means of the tachistotactometer involving oral reading under rigid timing conditions. It would seem that more natural conditions of reading might have given more flexible conditions for strategy choice.
Nolan and Kederis (ibid., 47-48) added the suggestion "that below a certain level of mental ability, braille ceases to be an effective medium for education" so that "for students whose IQ is below 85% it is an extremely inefficient medium of communication". In the opinion of the writer the suggested level of IQ 85 can be taken liberally, for there are a small proportion of readers who may have a comparatively low IQ, who show poor evidence of recall, as evaluated in answers to comprehension questions, and yet who seem to gain much satisfaction from reading. Their rates of reading may be faster than their intelligence would suggest. Such children will progress slowly in education, but their sense of achievement and enjoyment of reading is remarkable in the circumstances. Braille readers show many individual differences in coping with a complex medium.
Gomulicki (1961, p.51) found that "at the age of 5 the blind child was at a distinct disadvantage as compared with the sighted one, taking decidedly longer to produce results that were markedly inferior", but "by their mid-teens or thereabouts, the blind children became ... as good as the sighted children of the same age". He regarded this catching up from a slow start as being "at enormous cost of prolonged effort of the intellect" (ibid., p.52). Gomulicki was writing in more general terms than in the sphere of braille learning, but nevertheless, this background needs to be taken into account in any review of the blind child's reading development.
It is well known that many of the experiences that contribute to reading readiness for sighted children are necessary for the blind child too. However, because the surroundings are less accessible in the absence of sight, many have to be deliberately brought to the notice of the young blind child, and emphasis laid on the remaining senses in order to help concept development. As a result, reading by means of the braille code has a slow start too, intelligence being an important factor.
As a direct result of Gibson's enquiries (1962; 1968) and researches carried out in the '70s and '80s, the type of research into braille reading began to have a different slant. Instead of stressing reading behaviour, researchers were concentrating more on why and how. That is, how does mental development affect braille learning and use, what strategies are used for discrimination and recognition of symbols, and at what stage of development? Because these factors are all interrelated the problems posed by braille reading are seen to be more complex but need to be probed so that a greater understanding can ultimately help the braille reader.
As shown in Chapter 7, short-term memory is a major factor leading to the recognition of braille characters and words, but little experimental work has been carried out in this area. Recall may involve tactual memory or verbal memory, or both occurring concomitantly. Millar (1975, p.194) conducted an enquiry to test the effects of tactual and phonological features of braille consonants on tactual recall by blind children. Lists were compiled containing up to 6 braille letters (ibid., p.195), which were:
(There is a misprint in List 1 for H () is shown as F ().)
The set size, that is number of items, for each child was determined by a test showing the number of serial items which produced a score of 60%-100% correct responses. The results showed that recall of braille letters by blind children is affected by both verbal and tactual features and that this interacts with the number of letters on which they were tested. Further experiments are needed to determine the proportion of these features affecting recall, but the findings, so far, indicate that verbal coding is associated with high levels of total recall (ibid., p.200).
Ashcroft (1960) and Nolan and Kederis (1969) thought the perceptual unit of reading comprised outline shape of the characters, as helped by recognition of dot numerosity. For many years teachers have based their teaching methods on these findings.
Because braille characters are read sequentially the first half of the cell comes under the reading finger pad before the second half can be integrated with it, and during this process the relative positions of dots and spaces have to be registered before recognition can take place. This must be a slow process if not helped by other strategies. In 1984 (p.568), Millar posed the question of whether "shape can actually help with coding, or whether it relates to coding by sound and meaning at different levels of reading efficiency". As a result of observations on a retarded blind child's use of sound in memorising how to write braille characters, Millar suggested that phonological strategies may not only help retarded blind children, but may also be used in the early braille learning of normal children. She therefore carried out investigations (1984) to determine which of phonological, shape, dot numerosity and semantic strategies were used at different levels of reading ability, by retarded and normal children. Two hypotheses were tested (ibid., p.569):
For hypothesis (1), subjects were divided into 3 groups according to reading rates, and for testing hypothesis (2), reading age relative to mental age norms were used (ibid., 569-570).
For the first experiment Millar selected four-letter words from which the subject had to choose the 'odd man out'. The words were carefully chosen so that the choice would reveal which strategy or strategies had been used. For example, for semantic versus phonological choice, the selected words were GIRL, CURL, and LADY, where CURL is odd semantically, and LADY is odd phonologically. In the second experiment the same strategies were tested, but the critical stimulus word was given as well as the coding instruction to be used. Pseudo-words were also included to assess whether the coding instructions could be maintained.
The results showed that for all types of readers shape was the most difficult of the strategies - a radical change from previous thinking. Dot numerosity could be used as additional help. In general the results showed a 'mixed model' of processing, that is more than one type was used. The retarded reader tended to ignore the harder strategies, such as determination by shape and to rely more on phonology. The differences in coding suggested an interaction with mental age differences. It would seem that some readers may need to be helped to adopt additional strategies, thus leading to greater enjoyment of reading with an improved rate of reading.
In 1985, Millar backed up these findings by mounting two experiments on matching braille characters in dot pattern and in outline shape respectively, followed by an analysis of subjects' drawings of braille outline shapes. It was found that even the fastest readers were better at dot patterns than outline shapes and that "subjects were less accurate and slower at matching outline shapes than dot patterns at all levels efficiency". In fact (ibid., 16-17), "facilitation from shape coding is more a matter of individual differences in coding strategies rather than either the cause or effect of faster reading as such". It will be remembered that Ashcroft included errors of orientation in his list of braille difficulties. Here (ibid., p.17), Millar suggested that some readers seem to depend less on confusing mirror image shapes than on confusing small differences in spatial positions of dots within a letter, and also on confusion about the main reference axis in reading. She considered these factors to be more important than practice in detecting global letter shapes (ibid., p.17). Millar's findings on strategy choices by young blind children (1984; 1985) as extended and developed by further research, should have a far reaching effect on their education.
To understand more about the strategies used by faster, more fluent readers, three assumptions can be made (Millar, 1988, p.89):
Millar (1988) demonstrated these variables by measuring their effect on speed as shown by hand movements during oral reading. For this measure of performance a detailed and exactly timed mechanism was needed, much in advance of the more primitive apparatus used by Holland and Eatman (1933) and Fertsch (1946). Hand movements were video recorded from below a transparent reading surface, and synchronised with cumulative timing and voice output. Details of text showed up against the finger pads of the reading finger. Critical words, involving mis-spellings and context changes were included in the narrative, and problems caused by them were indicated in the oral reading and, more particularly, by a change in the tempo of reading.
Theories about the rate of braille reading include fast letter recognition, the fast coding of shapes, the use of context cues, and a knowledge of syntax. The hypotheses to be tested by Millar (1988, p.89) were:
The subjects read braille prose passages aloud. Each of the 6 stories was presented in 6 versions according to the 6 test conditions. These were:
Each subject read 6 different stories bringing in 6 different test conditions, the stories being counterbalanced across the subjects. It was found that stimulus quality, coherence in the text, and mis-spellings affected overall prose reading speeds, but their full effect in reading words and full text on speed is not yet fully understood.
Using the same apparatus, Millar (1987) had shown that when two hands are used in prose reading, they do not process different parts of the prose simultaneously. The left hand, when moving to a new line, does not start reading until the right hand has completed reading the previous line. The evidence suggests (ibid., p.120) "that fluent reading depends to a considerable extent on fast intermittent alterations in function between two hands".
Millar's comments (1984, 74-75) in connection with hand use, and the development in braille reading shown in the above experiments, are summarised when she wrote "highly proficient reading depends mainly on verbal strategies and skill ...; less proficient reading demands attention to spatial coding of the physical characters ...; whilst early learning subjects rely on non-spatial 'texture' (e.g. dot density) features of braille characters ...". Clear cut stages are not apparent, for children develop at differing rates and, as has been seen (Millar, 1984), use a variety of strategic choices, but the main trend can be recognised.
Ever since 1870 attempts have been made to change the code so that it would become easier to read and to use. These attempts have been chronicled in Part 1 as far as the London conference of 1978. Current work on the code will be discussed in the next chapter.
Few alterations have been made to the braille code since the British revision of 1905. Teachers have always known that the medium is less efficient and more difficult than visual reading, so great efforts have been made to improve methods of teaching as demonstrated in books, articles in journals and in presentations during conferences. By observation it was known that the development of blind children was slow compared with that of their sighted peers. Comparisons with the development of sighted children at any age are difficult and this must be particularly so when the complexities of braille reading are being considered. Standardized tests for sighted children have been converted in the absence of anything more appropriate, but braille reading especially needs to be assessed by tests standardized on samples of blind children.
The Tooze Braille Speed Test (1962). The test is intended to assess the child's attainment in "actually reading braille symbols" for children of primary school age. It consists of 120 three-letter words that contain no contractions. Reading is timed and the raw scores obtained in one minute can then be used with chronological age to transform the scores into reading ages and standard scores.
Lorimer Braille Recognition Test (1962). In 1962 (p.5), Lorimer, J., wrote that when reading tests of comprehension were attempted, "there was no way of knowing if or to what extent results were affected by difficulties with braille contractions". It was for this reason that he provided a standardized test based on a population of 332 children of primary school age, which was intended to measure "the braille factor". 174 unrelated but carefully chosen words were provided, each containing a contraction. For example, too much familiarity with a word would mask whether the contraction was recognised, so less familiar words were chosen. The test was terminated after 10 successive failures, and norms are given for each half year from 7.0 to 12.6 years. The test was diagnostic in that it determined the types of errors likely to be made when reading braille, and it could also provide guidance in the construction of teaching material and for remedial help.
Neale Analysis of Reading Ability, adapted for use with blind children (1977). Neither the Tooze test (1962) nor the Lorimer test (1962) was intended to test comprehension, and Lorimer, J., felt that "there was an urgent need for a test which not only provides reliable quantitative measures of accuracy, comprehension and rate in reading ... but also yields diagnostic information which reveals specific difficulties and indicates the type of remediation needed" (1977, 1-2). The blind population in Britain is comparatively small and therefore it would not have been possible to find sufficient numbers needed for trials and final versions of a new test. It was therefore necessary to use a well-tried test for sighted pupils and standardize it for braille-reading pupils. The Neale test for sighted pupils provided 3 parallel forms, which were of comparable standard to make retesting possible. Each form consisted of 6 graded reading passages with questions provided to test comprehension. Based on a sample of 299 blind children, the Lorimer adaptation has proved successful not only in testing accuracy, comprehension, and rate of reading, but also in having diagnostic value. Testing provision is made for recognition of the error types typified by Ashcroft (1960) as well as fundamental reading difficulties specific to the braille code. The test is currently being restandardized and will be referred to again in the last chapter.
Having determined that the braille character is the perceptual unit for braille reading, Nolan and Kederis (1969), investigated the effect of training on the rate of reading. They used 3 types of training, the rate and error scores in oral reading of individual characters and in words, and the rate and comprehension in silent reading. 12 students took part in each of Grades 3 through 6. Pre- and post-tests were administered and after training the results revealed that for the experimental group, an increase in time of 42% for individual characters and 15% for words, and for error scores a decrease of 83% for individual letters and 28% for words. The corresponding scores for the control group were an increase of 15% on time for individual letters, with a decrease of error scores of 19% but the scores for rate and error for words changed hardly at all. Unfortunately, the scores for comprehension were unreliable because some of the students were so highly motivated by monetary rewards that comprehension suffered. The results were encouraging but a replication would have been desirable because the sample was so small. Other training programmes to increase rate of reading included those by Flanigan and Joslin (1969) and Umsted (1972).
At the beginning of this chapter reference was made to the work of Fry (1963) and Watts and Buzan (1973) who attempted to teach sighted adults to read faster. Braille reading is considerably slower for most readers, not only because of perceptual problems and difficulties inherent in the code, but teaching has sometimes inadvertently encouraged this tendency. Stress had been laid on accurate oral reading and silent reading was perhaps not encouraged because of lack of reading material. Now that most blind children are integrated into the main stream, uncorrected silent reading is more common. Material well within the pupil's grasp will encourage a desire to hurry on when enjoying the story.
In America McBride (1974) organised several two-week workshops intended to help increase the reading rates of blind adults. Candidates were encouraged to experiment individually for "each person developed his reading skills in his own way, through suggestions from workshop directors and through comparing his own techniques with others in the class". Each subject provided his own material and comprehension was tested by other participants; meanwhile purposeful reading, active responses, elimination of sub-vocalising, continuous effort to read faster yet with flexibility of speed were all emphasised. A series of exercises were also employed involving rapid scanning of pages of braille using one or both hands and using one or more fingers. The result showed an increase in average reading rate from 138 w.p.m. at the beginning of the course to 710 w.p.m. at the conclusion. It was found that the subjects in the sample were highly motivated professionals. Though not conducted under rigid conditions the workshops stimulated others into action, and within the year (1975) two more studies were mounted.
Crandell and Wallace (1975) divided the participants into two groups, one having training in rapid reading and also code recognition, the other having training in rapid reading only. The results showed gains for the experimental group of approximately 39% and the control group showed only marginal improvement. The experimental group gained speeds of up to 225 w.p.m.
Olson, Harlow, and Williams (1975) divided their participants into three groups, two braille reading and one using large print. One group was taught by an ex-member of one of McBride's workshops and followed his informal methods. The other braille group was taught by more formal methods but also had informal sessions. Both informal post tests revealed substantial gains, but not in the realms of those achieved by McBride. They found (ibid., p.395) that "age had a negative effect on one's chances to increase his reading rate. It is obvious, then, that we should concentrate our training efforts on young children who have not yet established their reading habits". This fits in with the finding made by Fertsch (1946), p.131) that reading habits are established by the third grade, and Olson (1976) published an article entitled "Faster reading: preparation at the reading readiness level" (1976).
In 1977, Lorimer, J., published Outlines of a short course to improve the braille reading efficiency of children in lower senior classes. The study is unusual in that it was carried out under normal classroom conditions. Six classes in a residential school were divided into two groups of three. Half of them contained the 11 pupils regarded as the experimental group and half, the control group, were contained in the parallel three classes. The mean age and IQ of the experimental group was 12.10 years and 96.5 and the control group 13.2 years and 95.6 respectively. All the pupils in the 6 classes took part even if their efforts were not included in the results, and the pupils were unaware of experimental conditions, regarding the proceedings as extra reading training. Some showed little motivation at first, but there was noticeable enthusiasm as the training proceeded. Each class had 19 periods of 40 minutes training. The experimental group showed a gain of 84% in w.p.m. with only a slight drop in comprehension. There was a small gain for the control group. The training involved practice in techniques of hand use and speed reading with and without comprehension, Lorimer, J., stressed that the course was not regarded as comprehensive and that "more research-based information about the process and the limitations of reading by touch is needed before the design of a complete post-primary course in rapid reading can confidently be attempted" (ibid., p.10).
The selection of research findings concerned with reading in the medium of braille included in this chapter spans more than 60 years. In the earlier years much of the work was observational and equipment for measuring results was primitive compared with the more sophisticated devices of present day, but this groundwork formed the basis of much of present day enquiries. Since the late sixties more is being found out about the psychophysical properties of touch perception so a more holistic approach is now possible.
New research throws up further ideas of what needs to be investigated. For example, Millar's findings concerning the strategies used by young children when learning braille show that the processes concerned involve wider choices than had already been realised and were different for stages of development and for those of lesser intelligence. Further research in this area seems of prime importance for the more that is understood, the more will be the benefit to learners and teachers alike. Obviously, teachers need a fuller knowledge of the strategies being used and therefore where to strengthen the learning process.
Because tactile reading takes longer to learn than visual reading there is less time available to learn such skills as the use of context cues, syntax and dictionary skills. There is therefore an urgent need for blind children to be given extra help in this area for them to reach their full potential. More research is needed for a better understanding of when such help is appropriate for each child.
Lorimer, J.'s, research (1977) on improving the braille reading efficiency of children in the lower senior classes could be suitably adapted for use with fluent younger children, based on more research-based information about the stages of development in braille reading at that age.
More research is needed regarding verbal processing and finger and hand movements. This is also important because it has direct bearing on the early stages of reading when the learners are determining their best hand use before their reading habits become more set.
So far little has been said about adult reading because there has been little research carried out in this area. Children were easier to select where whole samples could be found in a special school, and, now that children are more integrated into normal schools in Britain, travelling for researchers is expensive and time consuming. Adult learning of braille is very different for two main reasons; adults are not learning 'to read' at the same time as they are learning braille, and they have the disadvantage of having to learn a new skill without the old word patterns intervening. The teenager who must learn quickly so as to get on with training or for a waiting job, the adult suddenly blinded who has to adjust to new working conditions, the retired professional who wants a purposeful new life, and the elderly blind person who may learn slowly but needs braille for simple organisation in the home, are but a few of the adults who need new literary skills. It is obvious that their needs are very different.
In conclusion it seems to the writer who has been both teacher and researcher, that there is not a sufficient link between the valuable findings of researchers and their application in the learning situation. It is necessary that findings are written up in research journals so that others may build on previous work. Understandably, teachers often do not always understand the statistics involved or even the special vocabulary that is necessary to convey exact meaning. No blame is attached anywhere, for it is the natural result of the working of separate disciplines, and all have problems of lack of time. Yet the matter needs to be addressed for the sake, ultimately of those for whom the research is carried out. The Braille Authorities of the United Kingdom and North America have also supported the desirability of having this knowledge made more available. The International Conference on English Braille, Grade 2, held in Washington, D.C., in 1982, passed resolution 7.5 (p.248) which stated, "Little appears to be known about the structure and functioning of the touch sense or about the psychophysical factors which affect tactile sensitivity in perceiving braille characters. Research on this topic is clearly beyond the competence of any braille authority to carry out. It is therefore proposed that efforts to be made to persuade properly qualified specialists to undertake this task.". No reference was made to this aspect in the following conference held in London in 1988.
The intention of this chapter has been to explore some of the factors in the braille code which cause it to be a comparatively slow medium of reading. Inevitably, therefore, the difficulties became highlighted in order that they might be mitigated to some extent. To end on a positive and kindly note, let Ashcroft (1960, p.52) end the chapter, "The low incidence of errors, about 5 errors per 100 words, is a positive and encouraging finding of the study. Braille is often described as a difficult, cumbersome, and illogical system ... . Nevertheless, this study, which concentrated upon errors, and held a widely inclusive definition of error, revealed relatively little difficulty in a large sample of children of wide range in age, grade, and ability.".