Indian Linguistics 80(3-4) 2019: 149-168 ISSN: 0378-0759 AN EYE TRACKING STUDY OF THE MALAYALAM SCRIPT EQUATING PERCEPTUAL SPAN WITH PHONEMIC DENSITY PRINCY PAPPACHAN, KIRAN KISHORE and GAUTAM SENGUPTA University of Hyderabad princypappachan.p@gmail.com ABSTRACT Perceptual span is the size of the region from which useful information is extracted while reading a text. One major factor that affects perceptual span is the properties of the script. This paper aims at investigating the perceptual span in Malayalam, a script with characteristics different from other Brahmi-derived scripts. Four eye tracking experiments were conducted to examine the size and shape of the effective visual field by using the moving window technique of the gaze contingent paradigm. The result showed that the perceptual span of Malayalam was asymmetric with 1-2 akaras to the left and 13-14 akaras to the right of the fixated akara. The finding also posits the estimation of the perceptual span in terms of the information density in scripts. Keywords: Eye tracking study, perceptual span, information density, Malayalam script, Brahmi/ Indic scripts 1. Introduction Research on tracking eye movements while reading dates back to the 19 th century. In reading science, the processing of written information is pivotal in addressing a plethora of questions, namely, how one reads different scripts and the manner in which one reads. The Eye-Mind Hypothesis (EMH) formulated by Just and Carpenter in the 1980’s states that the duration of fixation at a particular point is equivalent to the time taken to process it. The hypothesis highlighted eye movements as indicators of cognitive processing during reading. Prior research has shown that the question of how different scripts are mentally represented and processed can be answered by identifying the perceptual span. The perceptual span is the area of effective vision or the area required to extract necessary information during a fixation in reading. It is the number of characters that must be present for reading such that the reading performance is not hindered or in no ways affected. An increase in the observed span would not enhance reading performance while a decrease in span would affect reading. From earlier research, the perceptual span is observed to differ depending on the properties of the scripts. McConkie and Rayner (1975) introduced the moving window technique of the gaze contingent paradigm to demonstrate the perceptual span of English. In this technique, the information around the fixation point is shown while the information outside this region or window (as per predetermined condition) is masked by X’s. In English, alphabetic script, the perceptual span was found to be around 14-15 characters to the right and 3-4 characters to the left of a fixation (McConkie & Rayner, 1975; Den Buurman, Roersema & Gerrissen, 1981). In Chinese, which has a logographic script, the perceptual span extended 1 character to the left of the fixated character and 3 characters to its right (Chen & Tang, 1998). The shape of the perceptual span of Chinese can thus be seen to be smaller when compared to that of English. This is due to the lack of visual spatial cues like punctuation and the use of spaces which is present in alphabetic scripts. This is also the case in Hebrew and Japanese. In Hebrew, the functional words are attached