Short Reports

An Alternative to Braille

W.R. Thurlow

Abstract: The standard braille code symbols are very difficult to discriminate tactually, and the braille alphabet code is difficult to learn. It is possible to design alphabet code symbols which are significantly easier to discriminate and learn. Tests of an improved code called C5 are described, which show that its symbols can be perceived accurately when formed by a braille printer.

A large percentage of blind people cannot read braille. One reason for this is that braille is a difficult code. The Moon System of Embossed Reading (1968) states, "The majority of people who go blind in later life are unable to master the small dots of the braille system;" the Moon alphabetic code provides "an easier alternative embossed type," and has been in use in England for 140 years.

Our previous research (Thurlow, 1986) showed that standard braille symbols were difficult to discriminate tactually, and that the braille code was difficult to learn. We found that Moon tactual symbols were significantly easier to discriminate and learn than same-size braille symbols. Later experiments showed that alphabetic code symbols could be improved still further by careful analysis of confusions between the tactual symbols, when tested with many subjects. These experimental symbols were constructed from wire and small crystals to represent dots. In two of these experiments, a new symbol set (MN) produced significantly better discrimination than Moon symbols, both when the symbols were of a larger size (11 mm high), and when a smaller size (6 mm high). Analysis of confusion errors in these and other experiments led us to the construction of an even better symbol set (C5), shown in Figure 1. Some of these symbols are similar to Moon symbols, and to the symbols Bliss et al., used in their experiments with air-jet stimulation of the fingers (Bliss, Crane, Link, & Townsend, 1966).

Challenging problems remain, however, in producing these symbols for practical use. Moon code has been printed in type, set by hand-a slow process. These codes could be more advantageously produced by computer-controlled braille printers. Experiments were therefore carried out with the C5 alphabet code as printed bv a braille printer (Ohtsuki, model BT-5000).

The Ohtsuki printer has graphics capability, and was programmed from an Apple II computer to produce C5 symbols. Adjacent dots in the symbols, horizontally or vertically, were spaced 2.5mm from each other, center to center. Two sizes were used: 1) a set in which each symbol was 7.5mm high and 7.5mm wide-with the exception of symbols for A and V, which were made 10mm wide, for symmetry; and 2) a set in which each symbol was 10mm high and 10mm wide.

Subjects - volunteers from an introductory psychology course - were initially made familiar with the C5 symbols visually. Similarity between the C5 symbol and the letter was pointed out. Subjects then attempted to learn the letter corresponding to each symbol. They were tested after two learning trials of one minute each; subjects were then blindfolded. Their left index finger was drawn across each raised symbol in approxirnately one second; they went through all the symbols in this way to become more familiar with them tactually.

Figure 1. The C5 code. (Note: chart will not reproduce)

Then they were asked to draw each symbol, after it had been presented tactually, with their right hand. The 26 symbols were presented in random order three times. Feedback was given, after trials 1 and 2, on any errors.

Results for 10 subjects tested with the 10mm-high symbols showed an average of 21.8 symbols drawn correctly on the third trial (with standard deviation = 1.7). Results for 8 subjects tested with the 7.5mm-high symbols showed an average, of 22.5 correct (SD = 1.8). Another group of 8 subjects was tested with 10mm-high braille tactual symbols, twice the standard size. They obtained an average of 11 out of 26 correct (SD= 6.1). Finally, subjects in the group tested with the 7.5mm C5 symbols were given a test in which they used their right index finger to explore each symbol tactually and then draw what they had perceived. Six of the eight subjects drew all 26 symbols correctly, while the other two subjects made only one error each.

Subjects were able to learn which English letters corresponded to C5 symbols quite readily (an average of 22.4 and 24.2 correct on the second visual learning trial, for the two groups of subjects). Subjects tested with the braille code found it more difficult to learn; they obtained an average of 9.1 letters correctt on the second visual learning trial. Similarity between C5 symbols and English letters obviously aids learning for subjects already familiar with English letters.

We hope that this report will stimulate others to try out a code like C5 for blind persons who find braille too difficult. Further research, however, is essential to test punctuation symbols as well as optimal separation of symbols within and between words.

References

Bliss, J.C., Crane, H.D., Link, S.W., & Townsend, J.T. (1966). Tactile perception of sequentially presented spatial patterns. Perception and Psychophysics, 1, 125-130.

Royal National Institute for the Blind (1968). The Moon system of embossed reading. Reigate, Surrey, England.

Thurlow, W.R. (1986). Some comparisons of characteristics of alphabetic codes for the deaf-blind. Human Factors, 28, 175-18

Willard R. Thurlow, Ph.D., professor emeritus, Department of Psychology, University of Wisconsin, 1202 Johnson Street, Madison, WI 53706.

The autbor is indebted to Alida Moe, of Obtsuki Communications Products, for making copies of the C5 symbols.