Reading and Language Processing
What goes on during reading? Is reading necessarily an auditory act with a silent inner voice pronouncing the words or can meaning be derived without accessing a phonological code? What can eye movements and pupil dilation tell us about language processing? What aspects of language processing are specific to reading as opposed to listening? If any of these questions attract your interest, you'll find this volume to be a rich source of cutting-edge research.
This collection of 13 papers, most of which were first published as a special issue of the Canadian Journal of Experimental Psychology, sets out to present the state of the art in "some of the most important and interesting areas of contemporary cognitive research" (p. 3). These articles represent a strong if not exhaustive view of major researchers in cognitive psychology. I agree with the editors' claim that this volume will interest not only ". . .experimental psychologists but also those interested in linguistic theories of language structure,. . . studies of natural language processing in computer science,. . . and reading practitioners" (p. 3). However, this book is not a review of current theory clearly explained for the non-specialist; the reader who is not comfortable perusing articles in cognitive psychology may find it hard going. On the other hand, a better opportunity to familiarize oneself with recent experimental techniques being used in reading research would be hard to find.
An appropriate subtitle for this book might be Processing of Visually-Presented Language, for the majority of the articles deal with processing of visual input on the word or sentence level. The book could be divided into two parts: 1) use of technology for tracking eye-movement in parsing, and 2) low-tech approaches to higher-level processing of connected sentences in reading passages.
The book begins with a study by Buchanan and Besner using a word-naming task with Japanese speakers to examine how, for certain languages with a strong, consistent correspondence between spelling and sound (shallow orthography, e.g. Spanish, Italian, Serbo-Croatian), word recognition can bypass the semantic system which is more necessary for "deep orthography" languages (e.g. English, Hebrew).
There follows a series of articles illustrating how recent technological advances have helped to clarify several contested issues in language processing. The advantage of using technology for [-1-] measuring eye fixation in language processing studies is that it allows for tasks that more closely approximate natural reading, in addition to providing very precise measures. Rather than requiring the subject to monitor for some specific aspect or to perform a specific task, readers are instructed to read a passage for comprehension. The technology measures precisely patterns of eye movement and duration of fixation for each part of the text, in milliseconds. Granted, holding your head on a chinrest, close to a computer monitor, teeth clamped on a bite bar to ensure that only your eyes move, and not your head, is quite distant from curling up with a good book and a cup of tea, but the progress in the experimental design is the absence of additional tasks which may alter the reader's normal processing procedures. Several articles using this technique provide definitive responses to debates addressed in earlier research.
In an important study, Daneman and Reingold revise earlier research that claimed that word identification automatically activates the sounds of the word. With the eye fixation technique, they show that ". . . phonological activation appears to be a by-product of lexical access rather than a route to it" (p. 31).
Pollatsek, Raney, Lagasse and Rayner have applied similar technology in investigating the "perceptual span" to show that readers use mainly the information from only the line they are currently reading. The line directly below the current line is rarely accessed for information. Likewise, Henderson and Ferreira reexamine the debate over "parallel vs. sequential attention." In previous studies, the perceptual span for reading left-to-right languages such as English has been shown to be 4 characters to the left of the character fixated and 14-15 characters to the right. Information from characters further to the right is taken in at the same time as the word focused on, in the view of the parallel attention hypothesis. Data presented here support the "sequential attention hypothesis." Processing of the fixated word begins first, before accessing the next word to be fixated.
A study by Clifton introduces a series of articles using eye-fixation techniques in processing "garden-path" sentences. The reduced relative clause has been widely used in cognitive psychology to investigate language processing. The most famous formulation belongs to Bever (1970): "The horse raced past the barn fell". Sentences are constructed in such a way as to trick the reader into expecting a certain construction. In this way the reader is "led down the garden path". In the above sentence, we first interpret raced as the preterit verb of the subject horse. When we arrive at fell, we must reinterpret our initial construction. This results in longer processing time than for a simpler sentence such as "The horse raced past the barn and fell". (See Spivey-Knowlton, Trueswell, and Tanenhaus, this volume, for a thorough review.) [-2-]
Much research has been conducted using variations on garden-path sentences. Here, eye-fixation technology allows a more fine-tuned approach. For example, Ferreira and Henderson caution against overinterpreting longer reading times to mean more complex processing. Subjects tend to speed up as they read longer sentences. Thus, comparing data among variations in garden-path constructions requires strict control for similar length.
Finally, Just and Carpenter, early supporters of eye-fixation research, present a groundbreaking study of pupil dilation as an indicator of "intensity of processing." They also propose suggestions for integrating the concept of intensity into their computer model, CCReader, as a ratio of expenditures to resources. The editors call for more work in this area with the potential goal of pupillometric measures of reading compared to listening.
The last five articles in this collection go beyond the sentence level to examine higher-level processing.
Subjects in Singer's study were presented with sentences such as "Dorothy poured the water on the bonfire, but the bonfire grew hotter." This inconsistent sequence produced faster reaction times to the question "Does water extinguish fire?" than did the temporal sequence "Dorothy PLACED the water BY the bonfire; the bonfire grew hotter." Even though the first sequence is illogical, processing infers the logical association between fire and water.
Other studies measured the accuracy of recall and/or speed of repetition across variations of texts. Moravcsik and Kintsch showed that writing quality, general reading skills, and world knowledge are all important factors in text comprehension.
A study by Levy, Barnes and Martin and a series of six experiments by Masson examined how the visual mode of presentation and small alterations in content affect rereading fluency. Levy, Barnes and Martin found that paraphrasing the original text did not reduce rereading fluency, whereas changing the words of main concepts did. Masson showed that rereading fluency is highest when visual information is presented identically in both trials. Even text presented upside down in the original trial was reread faster in the identical presentation rather than in a normal position.
This volume presents a wealth of cutting-edge "pure" research, much of which will be of definite interest to the second language acquisition researcher. Bringing this research into the applied arena of the second language classroom is not immediately practical. Nonetheless, the insights this research affords will also interest the second language teacher. It is particularly useful as an introduction to current experimental designs using recent technology to track eye movement and/or as an introduction to the extensive research in cognitive psychology using "garden-path" tasks. [-3-]
The book includes both an author index and a subject index.
Bever, T. G. (1970). The cognitive basis for linguistic structure. In J. R. Hayes (Ed.). Cognitive development of language. New York: Wiley.
Universite de Technologie de Compiegne, France
© Copyright rests with authors. Please cite TESL-EJ appropriately.
Editor's Note: Dashed numbers in square brackets indicate the end of each page in the paginated ASCII version of this article, which is the definitive edition. Please use these page numbers when citing this work.