Reading hybrid texts : remarks on text/image transitions

The reading of printed materials implies the visual processing of information originated in two distinct semiotic systems. The rapid identification of redundancy, complementation or contradiction rhetoric strategies between the two information types may be crucial for an adequate interpretation of bimodal materials. Hybrid texts (verbal and visual) are particular instances of bimodal materials, where the redundant information is often neglected while the complementary and the contradictory ones are essential.Studies using the 504 ASL eye-tracking system while reading either additive or exhibiting captions (Baptista, 2009) revealed fixations on the verbal material and transitions between the written and the pictorial in a much higher number and duration than the initially foreseen as necessary to read the verbal text. We therefore hypothesized that confirmation strategies of the written information are taking place, by using information available in the other semiotic system.Such eye-gaze patterns obtained from denotative texts and pictures seem to contradict some of the scarce existing data on visual processing of texts and images, namely cartoons (Carroll, Young and Guertain, 1992), descriptive captions (Hegarty, 1992 a and b), and advertising images with descriptive and explanatory texts (cf. Rayner and Rotello, 2001, who refer to a previous reading of the whole text before looking at the image, or even Rayner, Miller and Rotello, 2008 who refer to an earlier and longer look at the picture) and seem to consolidate findings of Radach et al. (2003) on systematic transitions between text and image.By framing interest areas in the printed pictorial material of non redundant hybrid texts, we have identified the specific areas where transitions take place after fixations in the verbal text. The way those transitions are processed brings a new interest to further research.

Using eye-tracking to study anaphoric relations processing in European Portuguese

In this study we evaluate processing costs of different types of anaphoric expressions during reading. We consider three types of anaphoric expressions in Subject sentential position: a null pronoun (pro), and two gaps produced by syntactic movement: a WHvariable and a NP copy. Given that coreferential pro exhibits more referential weight than wh- and NP-gaps, and grounded on theories of referential processing based on relations of hierarchy and accessibility of the antecedent, we raise the hypothesis that the more dependent on its antecedent the anaphoric null constituent is, and the more minimal is the distance in terms of hierarchical structure between the anaphoric null element and its antecedent, the lower are the cognitive costs in processing. To test our hypothesis, we registered the eye movements with R6-HS ASL system of 20 Portuguese adult native speakers. Text regions including the selected anaphoric expressions were delimited and tagged. We analyzed the reading time of each region taking into account the number and duration of eye fixations per region; we used the reading time by character in milliseconds in order to compare values between regions of different length. We found a significant advantage in the reading time of the gaps arising from movement over the reading time of pro.