Introduction

Abstract

In the monograph «The gaze of Schrodinger's cat: Eye-tracking in psycholinguistics» we de- scribe how the interest to studying eye movements developed in time. It is an effective method to investigate how humans process information. Modern eye-tracking devices give opportunity to collect ecologically valid data. We already know a lot about basic parameters of eye-move- ments in reading and complex scene viewing. However, all the attempts to model the system of eye-movement control do not give a definite answer on how oculomotor and cognitive systems interact when processing visual information (including reading).

Cross-linguistic research is carried out to describe the mechanisms of text processing in dif- ferent languages with different writing systems. However, most studies are still done with ref- erence to English and several other European languages (mostly German and Dutch). In this book we present studies on reading in Russian language which has specific features like Cyrillic script and orthography which is neither opaque nor absolutely transparent, rich morphology and flexible word order. This monograph is the result of an integrated research project on verbal information processing at different linguistic levels (from recognizing single letters to compre- hending coherent texts), on processing of polycode texts, including verbal and nonverbal com- ponents, and on the relationship between complex images (such as paintings) visual processing and verbalization.

The current book is organized in three chapters. In the Chapter 1 the development of eye-track- ing methodology is presented from a historical perspective. Also, a short overview of the differ- ent eye-tracking measures and of the models of eye movement control in reading is provided (SWIFT, E-Z Reader). We discuss differences between the models, and describe how modern eye-tracking devices (by SR-Research, SMI, Tobii) work. The Chapter 2 focuses on different stages and level of linguistic processing during reading. We describe the mechanisms of sep- arate graphemes recognition, activation and competition processes in word recognition, con- textual integration processes, syntactic parsing, and anaphora resolution in our eye-tracking experiments. We present results of two boundary experiments, conducted to check if the font (fixed-width Courier New vs. proportional Georgia) influences legibility of Cyrillic letters. We show that letter identification efficiency depends on font: subjects were more efficient nam- ing Georgia letters (72%) than Courier New (46%). Besides, crowded letters in Courier New were harder to identify than the ones in Georgia, while recognition efficiency of isolated letters was almost the same. Since crowded letters imitate real reading, we claim that Georgia is more readable font than Courier New. One of the issues of our particular interest is the process of ambiguity resolution: the possibility of different interpretations of the message depending on the context is one of the key features of human language, and ambiguity can be both an object and a tool of a linguistic study when investigating the mechanisms of access to mental lexicon, syntactic parsing, discourse integration. The results of eye-tracking experiment on processing of ambiguous words with literal and non-literal meanings show that the type of the meaning

is not marked in the mental lexicon. This finding partially proves the parallel model of percep- tion of polysemantic words. Also, the results show that context does not impact lexical access immediately but affects post-lexical processing. The results of eye-tracking experiment with Russian idiomatic and additive sentences support the configuration model; facilitation in idiom processing was found, though as a delayed effect. A referential ambiguity appears when two or more referents pretend to be chosen as an antecedent for a pronoun. In our study we reveal the effect of ambiguity advantage, which is expressed in shorter processing time for ambiguous sentences as compared to unambiguous ones, when experimental task requires a resolution of the pronoun. Furthermore, our results demonstrate that pronoun interpretation is sensitive to syntactic roles of referents. In the last section of this chapter we describe the mechanisms of coherent text processing and the influence of text structure, text type and functional style, text readability and the level of reading skills of the reader as well as the type of the instruction. In the Chapter 3 we address to the investigation of image and complex scene processing and com- paring processing of verbal and non-verbal information as well as the influence of verbalization on visual processing. New data on the influence of text type, instruction, and the competence of the reader on reading process were obtained on Russian language material and presented in this book. The mechanisms for processing polycode text, including verbal and nonverbal com- ponents, have been described as well. Following the Cognitive Theory of Multimedia Learning our results show that integrated text-pictural presentation format (a polycode text) makes the comprehension of the information easier. An image and a written text presented together can contribute to a better understanding of the information than if they are presented separately. We demonstrate the advantages and disadvantages of the infographics (graphical visual representa- tions of complex information) and verbal text. Infographics is read faster and the percentages of correct answers to the after-the-text questions are higher after processing it than after reading a verbal text. Also, we describe the relationship between the perception of a complex visual image and its verbalization. First, the oculomotor behavior during the complex image free viewing (such as paintings) biased the following verbal description of this image: the more fixations are made and the longer the gaze is directed to the certain area of the image, the more words are dedicated to this area in the following description. Second, verbalization of the previously seen image affects the parameters of eye movements during repeated viewing of the same image. As a result, a specific pattern of ambient processing appears, which is usually associated with the initial stage of complex image scanning, and is characterized by short fixations and long sacca- des. It is noteworthy that the re-viewing without verbalization results with the focal processing pattern (longer fixations and shorter saccades).

The new data obtained in our studies give an opportunity to formulate practical recommenda- tions on text legibility and font design, text readability, visual organization of text. Some results may be useful for teachers of Russian as a foreign language and font specialists, for example, when developing a font for children or for patients with optical dyslexia etc.