On Cyrillic Letters Recognition Mechanisms in Reading: The Role of Font Type

This article investigates processing of Cyrillic letters with regard to the font type: monospaced (Courier New) or proportional (Georgia). The aim of the experiment was to describe letter recognition mechanisms during reading by testing whether the font type influences the accuracy of letter identification. Given that visual object recognition during reading begins at the step of parafoveal processing (that is, before the eyes actually move to it), the experiment focused on identifying a letter in the parafovea. Thirty-three lowercase letters of the Russian alphabet were used in the study. We had two versions of the experiment. In the first case, these letters were crowded, that is, each of them was surrounded by asterisks "*" (for example, "*ф*"), and in the second case they were isolated (for example, "ф"). Usage of the asterisks "*" imitates a letter within a word, and therefore allows us to bring the experimental conditions closer to real reading. Forty-eight people took part in the experiment. They were asked to name the letter on the screen. The experimental design included the invisible boundary paradigm: that is, hidden lines (boundaries) are placed between a screen-centered fixation cross and a target letter, presented to the left or to the right. As soon as the subject's gaze crosses the boundary, the letter disappears. Thus, the subject does not have time to focus on the stimulus, and it is always processed in the parafovea. This parafoveal region is considered to be 2-5° of visual angle to the right and to the left of gaze fixation point. Based on previous studies, the authors have decided to show the stimuli at 5° of visual angle to the right and to the left of the fixation cross. They found that both the presentation type and the font have a significant effect on recognition accuracy (p = 0.002 and p = 0.001 respectively). It turned out that identification of crowded letters is more difficult than that of isolated ones (a well-known crowding effect), and Courier New is a much less legible font than Georgia. This can be explained by the fact that proportional letters differ from each other greater than the ones of a monospaced font. Another explanation might refer to Georgia (but not Courier New) being originally developed for reading from screen. The result shows that probably proportional fonts have more grounds to be used in experimental studies of reading, for they are more readable. More importantly, significant interaction between factors "font" and "type of presentation" was found (p = 0.002): recognition accuracy depends on font when the letter is a part of a sequence, but not when isolated. As font influences letter recognition only when this letter is masked and we know that different fonts highlight different letter features, then it can be concluded that feature-based letter recognition mechanism prevails when letter is a part of a word during reading. In contrast, isolated letters are recognized as a whole.

Keywords

Authors

References