Comprehension Processes in Reading

David A. Balota, G.B. Flores d'Arcais, Keith Rayner, David A. Balota, G.B. Flores d'Arcais, Keith Rayner

• Éditeur: Routledge
• Année: 2012
• ISBN:
  9781136469121
• Édition: 1
• Sujet: Education
• Sous-sujet: Education General

Comprehension Processes in Reading addresses the interrelationship among several areas relevant to understanding how people comprehend text. The contributors focus on the on-line processes associated with text understanding rather than simply with the product of that comprehension -- what people remember from reading. Presenting the latest theories and research findings from a distinguished group of contributors, Comprehension Processes in Reading is divided into four major sections. Each section, concluding with a commentary chapter, discusses a different aspect of reader understanding or dysfunction such as individual word comprehension, sentence parsing, text comprehension, and comprehension failures and dyslexia.

1 COMPREHENSION
PROCESSES:
INTRODUCTION

Keith Rayner

University of Massachusetts, Amherst

INTRODUCTION

During the past 15 years, there has been considerable interest among experimental psychologists, educational psychologists, child psychologists, linguists, and others in the process of reading. In the past couple of years, two textbooks (Just & Carpenter, 1987; Rayner & Pollatsek, 1989) have surveyed research on the psychology of reading. In addition, numerous other edited volumes (e.g., Coltheart, 1987) have appeared that deal with the reading process. Although a considerable amount of information has been learned about reading since 1975, there are many gaps in our knowledge and understanding of the process of skilled reading. In particular, our understanding of moment-to-moment comprehension processes still seems quite limited.

The chapters appearing in the present volume represent an attempt to provide reviews of current research on topics relevant to on-line comprehension processes. It should be noted that the primary focus is on the process of comprehension and not on the product of comprehension (or what people remember from what they have read). What people remember from what they have read is certainly an important topic, but it is not unique to reading. Although some of the chapters in this volume deal with the product of reading, the primary concern is with how readers comprehend written text on a moment-to-moment basis. Accordingly, the four sections of the book are: (a) comprehension of words, (b) syntactic processes in comprehension, (c) comprehension of discourse, and (d) comprehension failures manifested in dyslexia and poor reading. Each section consists of a number of presentations followed by one or two commentary chapters. Whereas some of the commentary chapters deal with specific points made by the authors, others deal more globally with general topics relevant to the section.

COMPREHENSION OF WORDS

When we read, it is necessary to recognize the individual words that are printed on the page. There is a very large literature within experimental psychology dealing with the recognition of printed words (Rayner & Pollatsek, 1989). In many cases, researchers studying word recognition are not really interested in reading per se. Rather, printed letters and words are merely a convenient way to study perceptual processes because they represent a stimulus set in which variables such as frequency, visual angle, legibility, and so on, can readily be controlled. As a result of this fact, the relevance of studies of word recognition to understanding the reading process is often questioned. In addition, the particular tasks used to study word recognition are sometimes considered unrelated to what happens during fluent silent reading. For example, lexical decision, categorization, and naming tasks are typically used to study word-recognition processes, and it can be argued that such tasks are somewhat removed from what happens in normal reading. Likewise, threshold-identification tasks, in which words are degraded by brief tachistoscopic presentations, seem somewhat unlike normal reading. Despite these criticisms, it is clearly the case that in order to read fluently, individual words must be comprehended. In fact, a considerable amount of the variance in reading rate and how long readers look at individual words is accounted for by variables that also influence simple word recognition tasks (such as word length, word frequency, etc., Just & Carpenter, 1980). Hence, it does seem to be the case that understanding how individual words are comprehended is directly related to processes that occur in reading. Lately, there has been considerable interest in this issue (as opposed to the earlier and more frequent interest in perceptual issues related to word recognition).

The chapters in Section 1 address a number of important topics related to lexical access and the processes whereby word recognition occurs. The initial chapter by Balota deals with a somewhat controversial topic, namely is there a magical moment wherein lexical access occurs? More precisely, his issue is: Are there any tasks that psychologists typically use to investigate lexical access that can tap into this magical moment (if it exists)? The chapters by Seidenberg and by Besner are interesting because they highlight current debates in the field over the utility of connectionist models in understanding word recognition processes. The chapters by de Groot and by Schreuder, Grendel, Poulisse, Roelofs, and van de Voort deal with the structure of the mental lexicon. Finally, the chapter by Pollatsek and Rayner deals with the relationship between lexical access and eye movements in reading.

SYNTACTIC PROCESSES IN COMPREHENSION

When we read text, there is an underlying grammatical structure that influences the ease (or difficulty) of understanding. Whereas some approaches to understanding comprehension processes are based on the assumption that readers go from individual words to a more global semantic internal representation of the text (without any involvement of syntactic parsing processes), recent research clearly documents the important role that syntax plays in reading. It is the case that most work on syntactic processing merely uses reading as the context for the research. That is, the questions asked are very much the same whether the subjects in the research are reading or listening to discourse. However, I suspect that understanding the way that individual words are parsed into their appropriate syntactic constituents is vitally relevant to understanding how moment-to-moment comprehension processes occur.

Currently, there are a number of interesting debates concerning parsing strategies. These debates revolve around the extent to which parsing is a modular process that is unaffected by contextual information and real-world knowledge. All theorists realize that context and plausibility factors influence sentence processing. The critical question is when these types of information have their influence. Research by Rayner, Carlson, and Frazier (1983) and Ferreira and Clifton (1986) has suggested that initially the parser does not rely on such information in making syntactic parsing decisions but relies on procedures related to underlying grammatical principles. Thus, the impact of these types of information (context and world knowledge) is on reanalysis procedures. However, other researchers (Altmann, 1988; Altmann & Steedman, 1988; Taraban & McClelland, 1988) have argued that these types of information have an effect on early parsing decisions. The chapters in Section 2 largely deal with this issue and with a second general issue: To what extent are there general syntactic principles that are generalizable across different language systems? In one way or another, the chapters in Section 2 by Perfetti, Taraban and McClelland, Clifton and DeVincenzi, Mitchell, Cuetos, and Zagar, and Frazier all deal with these two issues.

COMPREHENSION OF DISCOURSE

There are many studies in the literature dealing with what people remember from what they have read. Although there are often debates about the most appropriate measure of retention (recall vs. recognition, for example), the general conclusions are quite similar across many studies. It is only in the past few years that there has been a serious attempt to understand moment-to-moment comprehension processes. As a result, there is now much interest in topics like anaphoric reference, causal inferences, elaborative inferences, instrumental inferences, and backward inferences. Much of the current research seems to focus on when anaphoric processes and inferences take place. It isn’t enough to know that such processes have or have not taken place, researchers now want to know when they take place (if, indeed they do take place). The chapters in Section 3 focus on the general issue of on-line comprehension processes.

In addition to dealing with on-line comprehension processes, there are some interesting themes running through the chapters in Section 3. One thing that is very central to the section is the extent to which different types of models can account for comprehension processes. Although the Kintsch and van Dijk (1978; van Dijk & Kintsch, 1983) model received a great deal of notoriety, one could argue that, although it’s interesting at a global level, the specifics are not well worked out. To be sure, the Kintsch and van Dijk model may still serve as a useful framework for researchers interested in comprehension processes, and many of the ideas that they initially proposed still have some currency. However, as the chapters in Section 3 make clear, researchers are now providing models that are much more specific about certain aspects of the comprehension process during reading. Thus, the chapters by Myers, van den Broek, Garrod and Sanford, and Vonk and Noordman deal with rather specific models and aspects of reading comprehension. The chapter by Sharkey, on the other hand, presents more of a global model of comprehension based on connectionist principles.

A second general theme running through the chapters in Section 3 has to do with the appropriate methods to use in assessing on-line comprehension processes. This topic is addressed very clearly by Keenan, Potts, Golding, and Jennings and is also dealt with by McKoon and Ratcliff (who also address the issue of appropriate models of the comprehension process). The chapter by Keenan et al. ties in very nicely with issues touched upon in Section 1 by Balota and others.

COMPREHENSION FAILURES AND DYSLEXIA

Dyslexia is a topic of considerable practical interest. As a result, there has been much recent research dealing with reading disability. Like the work on word recognition, much of the early research focused on perceptual issues related to dyslexia. More recently, the view has emerged that dyslexia reflects various types of language processing deficits (as opposed to a perceptual deficit). However, most of the work has not dealt with language processing per se, and a great deal of recent research has focused on the debate over whether developmental dyslexia represents an extreme point along a continuum of reading ability and on the extent to which there are distinct subtypes of developmental dyslexia.

Although not much attention has been paid to the relationship between reading disability and comprehension processes, it does seem that examining potential relationships might be important. After all, whereas reading rates of dyslexic readers are often quite slow, the primary characteristic of a dyslexic is that he or she does not understand written language. In the past few years, there have been some very interesting proposals concerning the relationship between word recognition failures and dyslexia, between syntactic parsing failures and dyslexia, and between comprehension monitoring failures and dyslexia. The chapters in Section 4 by Crain and Shankweiler, Conners and Olson, Seymour, and Rispens all deal with these issues.

GLOBAL THEMES

As you read the various chapters in the different sections, certain themes that cross the sections will be apparent to you. These themes are especially interesting because they provide a basis of common ground for discussion for workers interested in slightly different topics.

One obvious theme across the different sections relates to the need to determine when on-line processes are occurring. Thus, there is considerable interest in knowing when a particular process (be it lexical access or word recognition, parsing, or inferencing) occurs in real time during reading. The desire for this type of information leads to a second general theme, which concerns the appropriate measures to use to index on-line processes. The chapters by Balota and by Pollatsek and Rayner in Section 1, by Keenan et al. in Section 3, and many of the chapters in Section 2 all address the issue of how best to measure the processes one is interested in studying.

A third general theme crossing many of the sections deals with appropriate models of the processes one is interested in studying. Many of the chapters deal with this issue and it is one that is addressed in many of the commentary chapters. Although such a concern has been with us for some time, the fourth general theme is something of a new development (though obviously very much related to the third theme I have identified). This fourth theme concerns the relevance of connectionist models in studying processes like word recognition, parsing, and discourse comprehension. Some of the chapters (see Seidenberg, Taraban and McClelland, and Sharkey) represent a very pro-connectionist viewpoint, whereas other chapters (see Besner and Norris) take very anti-connectionist viewpoints. It seems that few issues have attracted the attention of so many scientists in different areas (all perhaps under the general umbrella of Cognitive Science) as have the claims of the connectionists, and many of the chapters in this volume reflect both the zeal and the apprehension associated with the enterprise.

A final general theme which I will touch on here doesn’t really cross many of the sections but rather is reflected primarily in Section 4. Although basic researchers often do not realize it, their work has practical implications. In the study of reading, basic research often has relevance for how reading is taught to children and for the development of remedial programs for people who do not read very well. Hence, because understanding the skilled reading process has significant implications for understanding why some people don’t read well, Section 4 has been included in the hope of stimulating further thought on the relationship between word processing, parsing, and comprehension processes on the one hand, and reading problems on the other.

In the final chapter the various issues and general themes that have been described in this chapter will be addressed again. My primary goal in the present chapter has been to set the stage for the different topics discussed in the four sections of this volume.

REFERENCES

Altmann, G. (1988). Ambiguity, parsing strategies, and computational models. Language and Cognitive Processes, 3, 73 - 98.

Altmann, G., & Steedman, M. (1988). Interaction with context during human sentence parsing. Cognition, 30, 191 - 238.

Coltheart, M. (Ed.). (1987). Attention and performance XII: The psychology of reading. Hillsdale, NJ: Lawrence Erlbaum Associates.

Ferreira, F., & Clifton, C. (1986). The independence of syntactic processing. Journal of Memory and Language, 25, 348 - 368.

Just, M. A., & Carpenter, P. A. (1980). A theory of reading: From eye fixations to comprehension. Psychological Review, 87, 329 - 354.

Just, M. A., & Carpenter, P. A. (1987). The psychology of reading and language comprehension. Boston: Allyn & Bacon.

Kintsch, W., & Van Dijk, T. A. (1978). Toward a model of text comprehension and production. Psychological Review, 85, 363 - 394.

Rayner, K., Carlson, M., & Frazier, L. (1983). The interaction of syntax and semantics during sentence processing: Eye movements in the analysis of semantically biased sentences. Journal of Verbal Learning and Verbal Behavior, 22, 358 - ...