1 Attitudes of Typographers

Section 1.1 noted the uncritical acceptance of the view that serif typefaces were easier to read on paper and other hard surfaces than were sans serif typefaces. This view has traditionally been adopted by typographers and typography educators (see, e.g., Craig, 1971, pp. 123–125; Hamai, 1986, p. 7; McLean, 1980, p. 44; Williamson, 1966, p. 109). Such authors have often relied on their attitudes and experience (and, sometimes, their authority and influence), but they show little awareness that the issue might be subjected to formal empirical research. The most extreme position was adopted by Morison (1959), who had developed the serif typeface Times New Roman in 1932 for the London newspaper, The Times (see Sect. 1.2); Morison insisted that “the serif is essential to the reading of alphabetical composition in long passages and consecutive pages” (p. xi), but he provided no evidence for this assertion. In short, as far as 20th-century typography was concerned, it was definitely a matter of “everybody knew” that serif typefaces were easier to read than sans serif typefaces when printed on paper.

These views have often been uncritically incorporated into the guidelines for potential authors that have been produced by journal editors and publishers. Such guidelines tend to include assumptions about the legibility of serif and sans serif typefaces, typically without providing any arguments or evidence in support of those assumptions. One characteristic example can be found in the sixth (2010) edition of the Publication Manual of the American Psychological Association. In a section headed “Preparing the Manuscript for Submission,” readers were instructed as follows:

A serif typeface . . . is preferred for text because it improves readability and reduces eye fatigue. (A sans serif type may be used in figures, however, to provide a clean and simple line that enhances the visual presentation.) (American Psychological Association, 2010, pp. 228–229).

This is contrary to good practice in psychology and the social sciences, where it is nowadays expected that assertions of this kind will be supported by citations of published (and, ideally, peer-reviewed) research, not based on presumed authority or ex cathedra pronouncements. Even so, the Association’s guidelines were adopted by the American Educational Research Association and by many other organisations both in the United States and around the world.

Another example can be found in Merriam-Webster’s Manual for Writers and Editors (1998), although this did at least allude to the existence of empirical evidence on the issue:

Serif faces are somewhat easier to read in blocks or paragraphs of text than sans-serif faces. . . . Studies of typeface legibility have tended to demonstrate that standard serif typefaces can be read somewhat more easily and quickly than standard sans-serif typefaces. (pp. 329–330).

However, since no such studies were explicitly cited, it is impossible for an interested but sceptical reader to determine whether or not the Manual’s account was accurate.

2 Dissenting Voices

During the twentieth century, there were few dissenting voices from this dominant view within typography. Although he had been one of Morison’s colleagues, Dreyfus (1985, p. 19) stated: “The outcome of many experiments indicates there is no statistically significant difference between the legibility of a wide variety of text types, even between seriffed and unseriffed types.” Unfortunately, he failed to specify which “experiments” he had in mind.

In the twenty-first century, dissenting voices have tended to come from the editors of journals in medicine and bioscience. They would, of course, be comfortable with the idea that the legibility of different typefaces could be the subject of empirical research, but their comments suggest that they typically lacked specialised knowledge of this research literature. In 2004, for example, the Journal of Psychopharmacology moved from a serif typeface to a sans serif typeface, which the editor claimed (once again without citing any sources) “should improve visual impact and reading ease of the Journal” (Nutt, 2004, p. 5).

In 2011, the editor of the Spanish journal Revista Española de Anestesiología y Reanimación, notified its readers of a “new look” for the journal (Errando Oyonarte, 2011a). (It specialises mainly in anaesthesiology, resuscitation, and pain management.) He mentioned a number of changes in the appearance and the format of the journal with the aim of making it more pleasant to read. It is interesting that the editor did not claim that these changes would necessarily render its contents more legible, simply that they would make its contents more attractive to its target readership. Among other changes, the journal had employed a different typeface, but in his initial announcement the editor did not specify the typeface in question.

One reader, González-Rodriguez (2011), pointed out that the journal had adopted a palo seco or sans serif typeface for both the headings of articles and their text. (The expression palo seco literally means “dry stick” in Spanish. It is used as a technical phrase by Spanish typographers, but there appears to be no counterpart expression to refer to serif typefaces. Some authors use tipografía con remates and tipografía sin remates—literally, with and without finishing—and others use tipografía con adornos and tipografía sin adornos—literally, with and without ornamentation—while others simply borrow the foreign terms serif and sans serif.) The reader complained that the exclusive use of palo seco violated the general custom in Spanish typography of using sans serif typefaces for headings but serif typefaces for the body text of articles. Specifically, he claimed that adopting a serif typeface facilitated a reader’s eye movements in following a line of text. (This idea was discussed but dismissed in Sect. 1.2.) In contrast, he claimed that the adoption of a sans serif typeface meant that readers’ eye movements were in danger of being lost in a “river” of white spaces.

González-Rodriguez allowed two exceptions to this generalisation. One involved the preparation of texts for younger readers; the other was the situation of reading written texts on computer screens. In both cases, he argued, sans serif typefaces provided a clearer image. He cited one study with regard to younger readers, who will be considered in Chap. 6 of this book; nevertheless, his claim is clearly not relevant to the task of experienced clinicians reading articles in a specialist journal. He cited no empirical evidence with regard to reading from computer screens, and so this is just another example of “everyone knows” discussed in Sect. 1.1. The issue of reading from screens will be the focus of Part II of this book.

In his response, the editor identified the sans serif typeface that the journal was now using as Akzidenz-Grotesk (Errando Oyonarte, 2011b). As he pointed out, this was devised as long ago as 1898 and had since been used by a wide variety of organisations and agencies; hence, it was by no means a novel and untested intrusion into the publishing world. Even so, he acknowledged that the panel which had recommended the “new look” for the journal had not included any experts on typography. He also argued that the journal should be open to the possibility that readers would increasing tend to access articles online rather than printed on paper. On González-Rodriguez’s (2011) account, therefore, the adoption of a sans serif typeface should actually enhance the journal’s legibility for those readers. The editor agreed to keep the situation under review in the future, although in fact at the time of writing the journal still uses the same sans serif typeface throughout. Indeed, since 2013 the journal has published articles in both Spanish and English using the same appearance, the same format, and the same sans serif typeface.

There is one example of a journal that has flip-flopped on this matter. In 2009, the journal Brain moved from a serif typeface to a sans serif typeface, but in 2015 it moved back to a serif typeface; on the latter occasion, the editor made the comment: “Whether serif or sans-serif typefaces are more readable has been addressed sporadically with psychophysical studies, without a clear conclusion” (Kullmann, 2015, p. 1). Whether this statement provides an accurate account of the relevant research literature is the focus of Part I of this book.

3 Are Serifs Purely Decorative?

One simple idea can be rejected at the outset. This is the assumption that serifs are purely decorative and superfluous to the task of identifying individual letters (e.g., Arditi & Cho, 2005; Burt, 1959, p. 8). There is good experimental evidence that readers identify the specific typefaces that they are reading before they identify the individual letters or words presented in those typefaces. For instance, words take longer to identify if their constituent letters are printed in different typefaces than if they are printed in one single typeface. This is true, in particular, if the letters are printed in both serif and sans serif typefaces (Adams, 1979; Krulee & Novy, 1986; Sanocki, 1987, 1988). Neurophysiological research indicates that typeface-specific information is processed within the right hemisphere of the brain, whereas typeface-independent information is then processed within the left hemisphere of the brain (Schweinberger et al., 2006; Vaidya et al., 1998).

Weaver (2014) described the case of a 52-year-old woman with a history of complex epileptic seizures that in recent years had been triggered specifically by reading. She herself had observed that her seizures were associated with reading material printed in serif typefaces (such as Palatino or Times New Roman) but not with reading material printed in sans serif typefaces (such as Arial or Verdana), although she also had intermittent spontaneous seizures. Her observation was confirmed electrophysiologically by asking her to read the first three pages of Charles Dickens’ novel, A Tale of Two Cities, typed in Times New Roman or Arial typefaces. Weaver suggested that the presence of serifs constituted more complex visual information that led to the activation of the hyperexcitable neuronal network responsible for her seizures.

The neurological condition of synaesthesia, in which perceiving an object in one mode stimulates the perception of a quite different mode, provides another example. A common form is grapheme–colour synaesthesia, in which specific characters, while printed in black, are seen as coloured. Weaver and Hawco (2015) described a patient who tended to perceive the letters ll (as in silly) in a vivid blue colour. The effect was more vivid for words presented in serif typefaces than for words presented in sans serif typefaces. However, if the ll was printed in a sans serif typeface but the rest of the word was printed in a serif typeface, the effect was more vivid than if the ll was printed in a serif typeface but the rest of the word was printed in a sans serif typeface. These findings show that serifs have a functional role that is not superfluous to letter recognition. This objection also applies to the idea that serifs only serve as visual noise (or as “cluttering incoming visual information”, as was suggested by Woods et al., 2005, p. 97).

4 Conclusions

This chapter introduced Part I of this book by summarising the attitudes of 20th-century typographers, who almost without exception considered that serif typefaces were easier to read than sans serif typefaces when printed on paper. During the twenty-first century, any dissenting voices have mainly come from journal editors in medicine and bioscience, who have tended to recommend the use of sans serif typefaces for the contents of their journals but have not provided any supporting evidence. This chapter also considered but dismissed the idea that serifs are purely decorative and superfluous to the task of identifying individual letters.