Literacy Beyond Text Comprehension

A Theory of Purposeful Reading

M. Anne Britt, Jean-François Rouet, Amanda Durik

• Publisher: Routledge
• Year: 2017
• Print ISBN:
  9781138927018
• eBook ISBN:
  9781317404446
• Topic: Education
• Subtopic: Education General

Literacy Beyond Text Comprehension aims to systematically investigate how readers interpret reading tasks within a situation, and how that interpretation influences reading behavior and comprehension. Presenting a new model of REading as problem SOLVing (RESOLV), the authors describe reading comprehension in terms of how a reader adopts goals within a particular situation that then guide what is read, when, and how. By applying the RESOLV model to a range of reading situations, this book provides evidence to suggest that there is no unitary understanding of a task, because individuals bring their own goals and characteristics to the situation; as such, it demonstrates the importance of understanding how a reader (e.g., student, test-taker, employee completing a work task) represents the context and the specific assignment. Written by internationally recognized learning sciences scholars, Literacy Beyond Text Comprehension advances the state of the art in reading research, but also seeks to inform a broader range of audiences, including those interested in the teaching and the assessment of reading.

1
What Is Reading?

A Review of Current Constructs and Theories

As part of a class on the psychology of reading offered to students with diverse academic backgrounds, one of us has been asking the same question each year at the onset of the first class meeting: “To you, what does it mean that a person is able to read?” Students are invited to think for a few minutes and then to write down a few lines in response to the question. The responses collected across many years of teaching almost invariably include two dimensions, under various wordings: “(It means that) people are able to decode written words/sentence/texts” and “(It means that) people can comprehend/interpret the meaning of written texts.” Most students perceive that these are two important constructs underlying reading literacy. The two elements of that definition, however, are not symmetrical. Decoding is often construed as the means whereby readers achieve comprehension, which is seen as the essential purpose or function of reading in our society.

This anecdotal convergence in lay students’ responses reflects a consensual understanding of reading in developed societies. Interestingly, reading as a product of decoding and comprehension processes also turns out to form the conceptual backbone of most contemporary research into the acquisition and processes of reading (i.e., the so-called “simple view of reading,” see p. 10). In this chapter, we briefly review some of the defining constructs of current reading theories. We also discuss how these constructs have evolved in past decades toward more comprehensive definitions encompassing the contexts and purposes for which people read.

Our review of the ample literature devoted to reading structures and processes is by no means exhaustive. Instead, we focus on some key aspects that will help us highlight the novel aspects of our own model. For more comprehensive reviews see Israel and Duffy (2009); Britt, Goldman, and Rouet (2013); Zwaan and Singer (2003).

1. Reading Processes and Structures

Not surprisingly, much of the text processing research published before 2000 focused on what readers do when reading “to comprehend”—that is, reading with the goal of building a memory representation of the meaning of the text. Thus, most studies of reading in school or laboratory settings implement a reading situation in which the participant’s reading goal is to learn the author’s main point (i.e., create a so-called “Situation Model”; Kintsch, 1998) for a subsequent memory test (Thiede, Griffin, Wiley, & Anderson, 2010). This restriction in the set of reading tasks and purposes reflected the need to clearly delineate the construct in order to ensure replication and comparison across studies, and to make clear and firm progress (Graesser, Singer, & Trabasso, 1994; Kintsch & van Dijk, 1978). Indeed, as recently as 2005, Perfetti, Landi, and Oakhill were still taking a strong stance on limiting the concept of reading comprehension:

We can expect the comprehension of written language to approximate the comprehension of spoken language. When that happens, then reading comprehension has developed, for practical purposes, to its limiting or asymptotic level. (…) All other limitations are imposed by linguistic abilities, relevant knowledge, and general intelligence. If we make things more complex than this, we push onto the concept of reading comprehension all these other important aspects of cognition, with the muddle that results from conceptual conflation.

(p. 227).

This statement reflects the tensions inherent to any attempt to design scientific theories of complex phenomena that involve a large number of factors, some of them very difficult to manipulate or control. Indeed, reading is a complex human activity. Unlike spoken language, reading needs to be learned through explicit instruction. It takes children several years to learn how to read well (Catts, Hogan, & Adlof, 2005; Perfetti, 1985; Vellutino, Tunmer, Jaccard, & Chen, 2007), and many people struggle with it throughout their lives (Perfetti, Marron, & Foltz, 1996; Sabatini, Shore, Holtzman, & Scarborough, 2011; Unesco, 2010). During reading, the eyes move, information gets to the visual field and is then distributed across various parts of the brain (Rayner, Pollatsek, Ashby, & Clifton, 2012). This is only possible thanks to a sophisticated cognitive architecture. Here we very briefly summarize the current understanding of reading and general aspects of the architecture that enables reading.

Like the students who we asked to define reading, theorists usually describe reading as the product of two distinct processing components: the decoding of written words, and the comprehension of the text’s meaning. This forms the basis of the so-called “simple view of reading” (Gough & Tunmer, 1986; Hoover & Gough, 1990; see also Florit & Cain, 2011; Kendeou, Savage, & van den Broek, 2010, for more recent discussions). Put in simple words, the simple view of reading posits that reading requires both some ability to recognize written words (“decoding”) and the ability to comprehend the meaning of the message. The simple view further states that whereas decoding is specific to the written modality of language (casting aside the important role of phonology in the process), comprehension processes are relatively independent from modality, and perhaps even shared across different representations of information (e.g., verbal vs. pictorial; see Gernsbacher, Varmer, & Faust, 1990; Ganis, Kutas, & Sereno, 1996).

Decoding and comprehension vary in their neurocognitive underpinnings, their acquisition, and their sensitivity to cultural and contextual variations. The ability to decode words is closely related to the characteristics of the visual systems in humans (Dehaene, 2009; Dehaene & Cohen, 2007) and perhaps in other primates (Grainger, Dufau, Montant, Ziegler, & Fagot, 2012). Word decoding is usually learned through direct instruction during the early school years. Skilled decoding means effortless and largely automated recognition of nearly any printed word (Perfetti, 1985). Decoding is achieved through similar mechanisms in nearly every culture, regardless of the characteristics of the writing system (though see Ziegler & Goswami, 2005; Ziegler et al., 2011, on the influence of orthographic characteristics of different languages). Most researchers agree that competent readers use two main routes to decoding a printed word: a direct route, whereby the visual input is directly matched to a lexical entry in the readers’ long-term memory; and an indirect or assembly route, whereby letter strings are progressively assembled into a unit thanks to the activation of morphological and phonological traces (Coltheart, Rastle, Perry, Langdon, & Ziegler, 2001; Jobard, Crivello, & Tzourio-Mazoyer, 2003). The direct route is the most usual one for experienced readers reading familiar words; the indirect route is used by developing readers and by all readers when faced with new, unfamiliar, or complex words.

Humans’ ability to comprehend language is arguably more complex and multifaceted than their ability to decode words. Language comprehension is usually defined as the construction of a meaningful mental representation from a set of words, sentences, or larger discourse units (Zwaan & Singer, 2003). Comprehension relies on the person’s knowledge of the situations, concepts, and relations conveyed in the message. It requires an activation of word meanings and their integration into a coherent whole (Kintsch, 1998). Integration means that the reader draws connections between the various sentences and passages of a text (McNamara & Magliano, 2009). Whereas some connections are explicitly cued by linguistic devices (e.g., repetition, connectives), others remain more implicit. For instance, most languages use pronouns to refer to previously mentioned persons or other entities. It is up to the reader to find out which entity the pronoun refers to, a task that is often less trivial than it may seem, especially for developing readers (Yuill & Oakhill, 1988). Comprehension also requires the integration of information at the deeper level of temporal, causal, and other types of relations among discourse statements (van den Broek, Rapp, & Kendeou, 2005; Zwaan & Radvansky, 1998). Finally, comprehension involves judgments of information validity and relevance (McCrudden, Magliano, & Schraw, 2011; Richter & Rapp, 2014). Validation can be fast and automatic, such as when a reader recognizes that a statement violates their world knowledge (e.g., “soap is edible,” Richter, Schroeder, & Wöhrmann, 2009), or it may require conscious reflection and reasoning about the plausibility of the statements, the reasons or arguments provided by the author, or other, more peripheral cues such as whether the author is trustworthy or whether the text comes from a reliable source (Bråten, Strømsø, & Britt, 2009; Britt, Richter, & Rouet, 2014).

To illustrate the various processes and knowledge that support representations created from reading, consider the two documents explaining coral bleaching presented in Table 1.1.

To begin with, readers have to be able to decode the written words to a correct mapping and fluently recognize words in context. Both word recognition and decoding are more of a challenge for beginning readers (García & Cain, 2013), but depending on the text and person, they can still be a challenge for some adults (e.g., zooxanthellae and symbiotic). Readers must have syntactic knowledge and skill as well as vocabulary knowledge (Oakhill, Cain, & McCarthy, 2015; Perfetti, 2010). Students trying to understand the italicized sentences of Text 2 have to use their knowledge of words (e.g., photosynthesis, eject) as well as knowledge of the concepts and situations, such as knowing that “symbiotic” is an interdependent or mutually beneficial relationship. Because the amount of information a person can keep active at any one time in working memory is limited (Baddeley, 1992; Miller,

TABLE 1.1. Two documents explaining coral bleaching. The first document shows how one initiating factor (human pollution) explains coral bleaching. The second document, the biologist’s journal article, shows how a different initiating factor (an increase in trade winds) leads to that outcome by way of a chain of events.

Text 1. High school student's essay posted on school website. You may wonder how coral bleaching occurs. It all begins when humans pollute the ocean. For example, many people dump trash in the Pacific Ocean where corals live. Some of these objects hit the coral. If the objects are big enough, they damage the coral and this process leads to coral bleaching.

Text 2. Excerpt from a biologist's journal article. Coral are animals that live in groups mostly in the eastern part of the Pacific Ocean. Coral are usually colorful, but their color actually comes from the zooxanthellae algae that live among them: the algae and the coral have a symbiotic relationship. The algae make food while the coral protect the algae. Sometimes there are changes in the environment that affect this relationship. Occasionally, the trade winds that travel westward across the ocean begin to slow down or reverse direction completely. This causes the warm surface water to be pushed to the east making the eastern ocean water warmer than normal. When the water warms up, it takes in less carbon dioxide, and this disrupts photosynthesis. Without photosynthesis, the algae cannot make enough sugars to give to the coral. Stressed corals are more likely to eject the algae that live among them. Without the algae present, the coral looks white, and this is called coral bleaching.

1956), the reader constantly needs to activate and organize information in order to understand a text (McNamara & Magliano, 2009). Some of this information is retrieved from long-term memory (e.g., definition of symbiotic); other parts are reactivated from representations of prior text in the current reading situation (e.g., Text 2 proposition: algae and the coral have a symbiotic relationship). In multiple-document reading situations, this can mean reactivating and integrating information from the current text (e.g., Text 2 proposition: algae and the coral have a symbiotic relationship) in addition to information from other texts just read (e.g., Text 1 proposition: pollution is the cause). The reader also can benefit from their knowledge of text structures (e.g., Oakhill & Cain, 2012; Trabasso & Nickels, 1992). To adequately represent texts 1 or 2, readers need to have a schema for an explanation that includes initiating causes connected to intervening states and events that then lead to an outcome.

According to current models of text comprehension, the specific knowledge retrieved or reactivated is controlled by automatic and strategic processes (Long & Lea, 2005). In terms of automatic processes, memory-based models assume that processes such as resonance control knowledge activation based on factors like semantic overlap, degree of elaboration, and distance or amount of text between the relevant prior text and the current sentence (Albrecht & Myers, 1995, 1998; Albrecht & O’Brien, 1993; McKoon, Gerrig, & Greene, 1996; Myers & O’Brien, 1998). For instance, the terms “algae” and “sugars” in the first italicized sentence, “Without photosynthesis, the algae cannot make enough sugars to give to the coral,” can reactivate prior information that the “algae make food.” Also the term photosynthesis may reactivate the concept that photosynthesis was disrupted in the prior sentence. Once activated, this prior information can be used to interpret the second italicized sentence, “Stressed corals are more likely to eject the algae that live among them,” by, for instance, encouraging the inference that “coral get stressed when they don’t have enough food.”

FIGURE 1.1. Levels of cognitive representation derived from reading written text (see text below for a definition of the labels).

When reading a text, readers construct multiple levels of representation (Figure 1.1). The surface level is a verbatim representation of the text such as distinguishing that Text 2 stated “Without photosynthesis, the algae cannot make enough sugars to give to the coral,” rather than “the zooxanthellae need photosynthesis to be able to make food for the coral.” There are times, such as in argumentation, where the verbatim level representation is very important (Britt, Kurby, Dandotkar, & Wolfe, 2008), but in general, this representation is fleeting (Kintsch, Welsch, Schmalhofer, & Zimny, 1990), giving way to the next level, the textbase. The textbase representation is a network of propositions that represent the separate ideas explicitly mentioned in the text (e.g., “algae make food for the coral”). Though still a representation of the information present in the text, it may no longer include certain verbatim aspects. The textbase next becomes elaborated with inferences from prior knowledge to form the Situation Model (Kintsch, 1998; Schmalhofer & Glavanov, 1986; van Dijk & Kintsch, 1983). For example, readers connect that the algae create sugars as food for the coral by way of photosynthesis. Readers construct the Situation Model as they integrate text information with their prior knowledge and generate inferences. There are a number of different kinds of inferences (e.g., Graesser et al., 1994; McNamara & Magliano, 2009). These include anaphoric mapping (e.g., in the phrase “it takes in less carbon dioxide,” “it” means “the water in the Pacific Ocean”), coherence-building inferences (e.g., when algae don’t make sugars, the coral don’t have food which results in stressed coral), and elaborative inferences (e.g., knowing that coral can get food from other sources).

Most text comprehension theories were developed to account for situations involving a single text. However, reading often involves situations where one must read multiple, related texts. Comprehending multiple texts poses the question as to how readers manage multiple representations of the same situation. To address that issue, we proposed the Documents Model framework (Britt, Perfetti, Sandak, & Rouet, 1999; Perfetti, Rouet, & Britt, 1999; Rouet, 2006). Because the Documents Model is the most general framework for representing information from texts, we will use it to represent what readers represent from reading. The Documents Model includes two main structures: an Integrated Model and the Intertext Model. The Integrated Model is a representation of the content. As just mentioned, the Situation Model is a representation of inference-elaborated content from a single text. It was called a Situation Model because the texts were mostly from narratives describing situations. Thus, the most studied dimensions of the Situation Model included agents and objects, spatiality, temporality, causality, and intentionality (Zwaan, Magliano, & Graesser, 1995; Zwaan & Radvansky, 1998). More recently, text researchers have begun thinking more generally about the concept of a Situation Model to include structures other than a situation. In this book, we will use the term Integrated Model as the more general term to refer to the Mental Model as the semantic content incorporated from one or more texts and organized around the structure of the text(s) or task (Britt & Rouet, 2012; Griffin, Wiley, Britt, & Salas, 2012; Wiley, Britt, Griffin, Steffens, & Project READi, 2012). For narratives, the structure of the Integrated Model would be a situation, but for an explanation it would more likely be something like a branching of factors that connect together to produce an outcome. For example, Figure 1.2 shows the two separate explanations from the two texts in Table 1.1.

As shown in Figure 1.2, there are two separate chains of causes, or explanations, beginning with different initiating cau...

Table of contents

• Cover
• Title
• Copyright
• Contents
• List of Tables
• List of Figures
• Foreword
• Acknowledgements
• Introduction
• 1. What Is Reading?: A Review of Current Constructs and Theories
• 2. Reading as a Purposeful Activity
• 3. A Framework for Reading as Problem Solvine
• 4. Reader Resources: Motivation and Cognition
• 5. Readers' Mental Models of Reading Contexts and Tasks
• 6. RESOLV Processing Decisions
• 7. Family of Hypotheses Derived from RESOLV
• 8. Fostering Effective Reading Strategies
• References
• Glossary
• Appendix
• Author Index
• Subject Index