2.5 In Retrospect
Despite all the attention paid to physical attributes in type by designers and researchers, it is usually not desirable that the reader of a text likewise pays much attention to the shape of the letters and their constituent parts. In fact, we are most successful at reading when nothing distracts us from the message content. The reader has to be able to identify each character easily, but should not remember anything about the type itself afterwards. If someone would remember specific features of the type used, it would mean that not all attention was attracted by the text content simply because something else in the reader’s environment was too prominent to ignore. The quality that is desired of type described here is transparency.

Note that transparency of type is a determinant of readability and is not desirable in all applications of text. Text on billboards, for example, often is not transparent; in congenial designs like this, it is done intentionally (chapter 4). The discussion in this paragraph does not imply that research on the physical attributes of type is useless since readers will not be aware of their results anyway. Even though the shape of type may not be consciously analyzed by readers, different shapes can still result in differences in reading performance.

The focus in legibility research has shifted from character legibility to legibility of text. It is expected that this approach will point to reliable differences (statistically significant) between typefaces, provided that the measures of legibility are made as sensitive as possible (Poulton, 1965). This does not mean that early research never produced reliable differences. Paterson and Tinker (1932) already found a significant decrease in reading speed when subjects were asked to read a text in Block type after having read the text set in Roman type. The typeface Cloister Black, for instance, was read at 83.5 % of the pace with which Scotch Roman was read. This means that reading a piece of text that would take 10 minutes in Cloister would be read in about 8.5 minutes in Scotch Roman. Of course, speed is not the only concern in choosing a typeface.

Another issue is the reader’s comfort with reading the typeface. It is suggested that comfort and reading speed are correlated, but no conclusive calculations of this are known (Ovink, 1938; Tinker and Paterson, 1942). Block lettering in long texts may annoy a reader, because its elaborate extensions and doubled strokes appear tiring. Still, such styles are used because of their historic ‘feel’. These considerations will be discussed in chapter 4.

One would think that reading habits and cultural biases might have an effect on legibility and readability. Coghill (1980) tested the hypothesis that reading habits may have such an effect. As subjects, she used children who had just started reading and were unfamiliar with all the typefaces used in the experiment except for Gill Sans. Coghill asked the children to read a text with familiar content, printed in various typefaces (originally in Gill Sans), and concluded that the reading of alternative typefaces provided little difficulty. The experiment was conducted to prove that serif and glyphic type are just as suitable for use in educational material as sans serif type, which is regularly used for such purposes. However, one would think that an additional conclusion can be drawn. One might conclude that the outcome of the experiment also means that differences in legibility or readability between typefaces can be attributed to preference rather than to reading habits or the functional properties of these typefaces.

Variations in typeface indeed do not affect legibility heavily (Tinker, 1963; Poulton, 1967; Foster, 1973). Bartram (1982) concludes from this that "the designer has a considerable degree of freedom in the choice of functional type properties where the main concern is one of legibility". Thus, he concludes that the challenge lies not in finding the most legible typeface, but in finding the most appropriate one among the legible typefaces.

In this review, appropriateness will be related to two aspects of typography; semantic appropriateness and compatibility with other typographic factors. We will first look at these other typographic factors, such as word spacing and leading, which need to be in harmony with the chosen typeface. The next chapter is not about type itself , but rather about the way type is organized.