I have argued that the processes which underlie shape perception by touch differ with the size and composition of patterns. Even the perception of single braille patterns was found to depend on the nature and type of scanning movements that were made possible by the available reference information and/or prior knowledge. The present chapter looks in more detail at hand and finger movements in fluent reading, and how scanning movements become organized for the spatial and verbal aspects of reading connected prose texts.
The crucial point is that, in touch, the intake of information occurs during scanning (Burklen, 1932; Davidson et al., 1980; Foulke, 1982; Kusajima, 1974). The reverse is the case in visual reading. Visual reading involves left-to-right eye-movements to travel over the text. But the intake of information occurs during the time that the eyes fixate a word, and not during the saccadic eye movements which occur between fixations (e.g. Rayner, 1983; Rayner and Pollatsek, 1987, 1989). The times spent in fixations, rather than movement latencies, are thus the important data in eye-movement studies. For braille reading, by contrast, the important data on the intake and processing of verbal information come from the deployment of the hands in reading and the precise timing of finger-movements over small details as well as over larger portions of text.
An on-line recording and timing device was developed for that reason. It is described first, because it provided data on the precise timing of types of finger movements which are essential for understanding the intake of information in reading by touch. The device is analogous to filming eye-movements in visual reading (e.g. Rayner, 1983). But it video-records hand and finger movements from below transparent surfaces, together with (1/100 s) timing and voice output. The device has the added advantage that people can read texts normally without any physical restraints while the recording provides synchronous timing for each millimetre detail that is being touched during reading. Filming finger movements from below, rather than from above or laterally, improves the precision of the data considerably. It is possible to see directly on the monitor which part of the fingerpad touches any dot of a braille cell of the text and at what