Theories of Reading Development [15, 1 ed.] 2017007488, 2017030260, 9789027218117, 9789027265647

Citation preview

S TUDIES IN WRITTEN LAN GUAGE AN D LITERACY 15

Theories of Reading Development edited by

Kate Cain, Donald L. Compton and Rauno K. Parrila

John Benjamins Publishing Company

Theories of Reading Development

Studies in Written Language and Literacy issn 0929-7324

A multi-disciplinary series presenting studies on written language, with special emphasis on its uses in different social and cultural settings. The series combines sociolinguistic and psycholinguistic accounts of the acquisition and transmission of literacy and brings together insights from linguistics, psychology, sociology, education, anthropology and philosophy. For an overview of all books published in this series, please see http://benjamins.com/catalog/swll

Editors Ludo Verhoeven Radboud University Nijmegen

Paul van den Broek

Leiden University

Volume 15 Theories of Reading Development Edited by Kate Cain, Donald L. Compton and Rauno K. Parrila

Theories of Reading Development Edited by

Kate Cain Lancaster University

Donald L. Compton Florida State University

Rauno K. Parrila University of Alberta

John Benjamins Publishing Company Amsterdam / Philadelphia

8

TM

The paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.

doi 10.1075/swll.15 Cataloging-in-Publication Data available from Library of Congress: lccn 2017007488 (print) / 2017030260 (e-book) isbn 978 90 272 1811 7 isbn 978 90 272 6564 7

(Hb) (e-book)

© 2017 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Company · https://benjamins.com

Table of contents

Introduction Kate Cain, Donald L. Compton and Rauno K. Parrila

1

Part I.  Big questions Introduction to big questions Kate Cain, Donald L. Compton and Rauno K. Parrila Integrating word processing with text comprehension: Theoretical frameworks and empirical examples Joseph Z. Stafura and Charles A. Perfetti Genetic and environmental influences on the development of reading and related skills Richard K. Olson, Janice M. Keenan, Brian Byrne and Stefan Samuelsson

7

9

33

Early literacy across languages Catherine McBride

55

Learning to read alphasyllabaries Sonali Nag

75

Opening the “black box” of learning to read: Inductive learning mechanisms supporting word acquisition development with a focus on children who struggle to read Laura M. Steacy, Amy M. Elleman and Donald L. Compton

99

Part II.  Typical development of word reading and underlying processes Introduction to word reading Donald L. Compton, Kate Cain and Rauno K. Parrila

125

Orthographic mapping and literacy development revisited Linnea C. Ehri

127

Putting the learning into orthographic learning Kate Nation and Anne Castles

147

vi

Theories of Reading Development

Orthographic learning is verbal learning: The role of spelling pronunciations 169 Carsten Elbro and Peter F. de Jong Learning to read morphologically complex words Joanne F. Carlisle and Devin M. Kearns

191

Learning to read in a second language Ludo Verhoeven

215

Part III.  Typical development of comprehension and underlying processes Introduction to reading comprehension Kate Cain, Donald L. Compton and Rauno K. Parrila

237

Vocabulary, morphology, and reading comprehension Mercedes Spencer, Jamie M. Quinn and Richard K. Wagner

239

Reading comprehension: What develops and when? Kate Cain and Marcia A. Barnes

257

Development of reading comprehension: Change and continuity in the ability to construct coherent representations Paul van den Broek and Panayiota Kendeou

283

Part IV.  Atypical reading development Introduction to atypical reading development Rauno K. Parrila, Donald L. Compton and Kate Cain

309

Early identification of reading disabilities Hugh W. Catts

311

Dyslexia and word reading problems Rauno K. Parrila and Athanassios Protopapas

333

Children with specific text comprehension problems Jane Oakhill and Kate Cain

359

Part V.  Instruction and intervention Introduction to instruction and intervention Donald L. Compton, Rauno K. Parrila and Kate Cain

379

Starting from home: Home literacy practices that make a difference Monique Sénéchal, Josée Whissell and Ashley Bildfell

383



Table of contents vii

Early reading interventions: The state of the practice, and some new directions in building causal theoretical models Robert Savage and Emilie Cloutier

409

Morphological instruction and literacy: Binding phonological, orthographic, and semantic features of words John R. Kirby and Peter N. Bowers

437

Reading comprehension instruction and intervention: Promoting inference making Kristen L. McMaster and Christine A. Espin

463

Theoretically guided interventions for adolescents who are poor readers Sharon Vaughn and Colby Hall Child characteristics by instruction interactions, literacy, and implications for theory and practice Carol McDonald Connor and Frederick J. Morrison

489

507

Index525

Introduction Kate Cain, Donald L. Compton and Rauno K. Parrila

Lancaster University / Florida State University / University of Alberta

Printed words are one of the most remarkable inventions of humankind, and learning to read them is one of the most remarkable achievements of human individuals. In recent decades, how people learn to read and reason with printed text has been studied intensively in genetics, education, psychology, and cognitive science, and both the volume of research papers and breadth of the topics examined have increased exponentially. At the same time, we feel that the rapid expansion of empirical research into new topics and questions has not been accompanied by matching access to theoretical advances. Partly the blame rests on current and former journal editors (including the three of us) who focus on publishing new empirical research, and partly with researchers who publish their theories in separate books, if at all. As a result, there is a mass of high quality empirical studies on reading development and disabilities, but integrative theoretical reviews of the literature are much more difficult to locate. Inspired by a Special Issue of Scientific Studies of Reading in 2014 (vol. 18, issue 1), which included four papers focused on theories, the purpose of Theories of Reading Development is to collect, in one place, the latest descriptions of various important and complementary theories of reading development and disabilities. Theoretical reviews provide critically important lenses through which to interpret the accumulating, and sometimes conflicting, empirical findings. They provide necessary routes into the increasingly complex field of reading research, providing the reader with frameworks through which to evaluate and synthesize new findings and arguments. The 22 chapters in this volume are grouped into five sections. In Part 1, five chapters tackle big questions and current debates about literacy acquisition and its development. The other four sections focus on the development of word reading and underlying processes (Part 2), the development of comprehension and component skills (Part 3), atypical development of word reading and reading comprehension (Part 4), and reading instruction and interventions (Part 5). Separate introductions to each section below provide summaries of their content; here we highlight some of the main themes that emerge across the different chapters and sections in this volume. A core theme running through this volume is the critical role of knowledge. In the first chapter of Part 1, Stafura and Perfetti examine word knowledge, and how the quality of information encoded about a word’s phonology, orthography, and semantics influences both word recognition and text comprehension. In the same section, Steacy, doi 10.1075/swll.15.int © 2017 John Benjamins Publishing Company

2

Kate Cain, Donald L. Compton and Rauno K. Parrila

Elleman, and Compton also consider the importance of knowledge for these two aspects of reading, focusing on how knowledge about a word’s form is acquired, and how the vocabulary and general knowledge that support comprehension of content develop. These two chapters provide excellent foundation to the three chapters in Part 3, which focus on reading comprehension development. Each has knowledge at its core: the role of vocabulary, including morphological, knowledge in reading comprehension (Spencer, Quinn, & Wagner); how and when information that underpins knowledge-based inferences about time, space, and causality is acquired (Cain & Barnes); and the critical role of knowledge for ensuring coherence (van den Broek & Kendeou). Knowledge is also considered in the discussion of reading comprehension difficulties in Part 4, where impoverished knowledge or knowledge structures are examined as contributors to the inference making difficulties of poor comprehenders (Oakhill & Cain). The knowledge bases we draw on for reading are acquired; they are not innate. As a consequence, theories of reading development often resemble as much learning theories as they resemble developmental theories. Some of this learning takes place through literacy instruction, and some outside of formal instruction, and even before formal schooling, through our exposure to language (e.g., Catts in Part 4) and different home literacy practices, as explained by Sénéchal, Whissell, and Bildfell in their chapter (Part 5). The theme of learning, in both formal and informal contexts, is at the core of this volume. Indeed, as Nation and Castles (Part 2) conclude “… to understand more about reading development, we need to understand more about how learning happens.” How we learn and acquire new word knowledge is the focus of Steacy et al.’s chapter in Part 1 and how we learn to read words is the central theme of Part 2; indeed, each chapter features ‘learning’ in its title. Ehri’s chapter on the development of sight word reading considers the changing nature of representations between the spoken and written language, and how this is influenced by instruction; Nation and Castles consider how we can advance our understanding of orthographic learning by situating our theory and methods in a more general framework of learning and memory; and Elbro and de Jong discuss learning to read words as an example of verbal learning where the critical content to be learned are the connections between spelling and standard pronunciations of words. Finally, Carlisle and Kearns focus on how little we know about learning to read morphologically complex words, and Verhoeven examines reading acquisition in L2. In turn, Parrila and Protopapas (Part 4) are concerned with children who fail to learn to read words at an expected rate, and Catts (Part 4) extends this concern to reading comprehension disabilities. Oakhill and Cain (Part 4) then provide a closer look at poor comprehenders, children who do not develop comprehension skills commensurate with their word reading skills. How we learn to comprehend is the underlying theme in the chapters in Part 3, and how we can design and deliver better comprehension instruction is the explicit focus of both McMaster and Espin (Part 5) who examine the role of learning mechanisms in comprehension instruction and intervention, and Vaughn and Hall (Part 5) who examine why reading development often stagnates in adolescence and what can

Introduction 3

be done about it. All of these chapters argue for stronger connections between theory and practice, and for better understanding of fundamental learning mechanisms needed to develop skilled reading comprehension. Similar arguments about the central role of theory are made by Savage and Cloutier (Part 5) who examine early reading interventions and advocate strongly for more and better randomized-controlled trials, and Kirby and Bowers (Part 5) who examine both theoretical and empirical arguments for explicit instruction in morphology. The context in which development and learning take place is a central theme for several chapters in this volume. Different languages exist across cultures and environments, and McBride (Part 1) considers the similarities and differences in learning to read words across languages, scripts, and cultures and how this shapes both the process and the outcomes. In the same section, Nag reviews reading acquisition research in Indic languages and highlights the specific and general learning challenges that children face when learning to read in orthographies that use an extended symbol set of mixed granularity. In his chapter on learning to read in a second language, Verhoeven discusses how specific features of the L1 and L2 interact with learner characteristics to influence transfer and learning. The environment in which learning takes place is considered in depth in Part 5, in which the influence of variation in both the classroom (Morrison & Connor) and the home environment (Sénéchal, Whissell, & Bildfell) are considered. Perhaps the final underlying theme that runs across the chapters is complexity. Stafura and Perfetti (Part 1) note that reading itself is too complex to be captured by a single theory, and the complexities of word reading and comprehension development, processes that contribute to them, and instructional interventions that we hope to improve them are evident throughout the chapters. In Part 1, Olson, Keenan, Byrne, and Samuelsson add heritability to the consideration and provide a highly accessible review of the state of knowledge in behaviour genetics of reading. The topic of heritability is picked up in Part 3 on reading disabilities, both in Catts’ examination of using family risk to guide early identification, and in Parrila and Protopapas’ exposition of a developmental systems approach to dyslexia. A better understanding of the complex interactions between our predispositions and environmental factors will enhance our understanding of protective and risk factors alike. This point is well made by Catts and by Parrila and Protopapas who consider reading disabilities within the risk-resilience framework, and how good instruction or intervention may moderate the effects of risk. Fittingly, the last chapter in the book by Morrison and Connor discusses a theoretical framework for examining child-instruction interactions and explicitly calls for adaptive systems thinking to guide instructional studies. Morrison and Connor suggest that we think of learning environments as complex adaptive systems where various agents, each with the specific individual characteristics, act and interact simultaneously to implement developmental change (better reading). In this formulation, the next generation of theories of reading development move from static descriptive theories of individual differences in reading skills to theories describing and modelling the developmental processes and the agents driving it.

4

Kate Cain, Donald L. Compton and Rauno K. Parrila

Our aim was to bring together theoretical reviews of the literature on reading development and difficulties to provide frameworks and inspiration to guide current and future researchers in the field. We hope that this volume meets that aim. All that remains is to thank each of the authors who contributed, the reviewers who provided constructive comments on each chapter, and Nicola Currie for her careful checking of each chapter.

Part I

Big questions

Introduction to big questions Kate Cain, Donald L. Compton and Rauno K. Parrila

Lancaster University / Florida State University / University of Alberta

The first five chapters tackle some of the big questions in reading research: the role of words in text comprehension, the influence of environment and genetics on reading development, reading across different orthographies, and the mechanisms by which word knowledge is acquired. Three of these chapters are extended from papers published in a Special Issue of Scientific Studies of Reading in 2014 (vol. 18, issue 1), which inspired this volume. Stafura and Perfetti’s chapter is an extension of one of the papers that appeared in the special issue. The authors argue for a more central role for word knowledge in a theory of comprehension. They first suggest that reading comprehension is too broad a construct to enable researchers to specify the types of testable models that guide research on word reading. As a result, they argue, reading comprehension research has been guided mainly by flexible general frameworks, such as the Reading Systems Framework they go on to present. The main focus of their chapter is an examination of the influence of word-level processes on comprehension, specifically word-to-text integration, which they suggest has been neglected by most reading comprehension frameworks due to their focus on higher level processes. Stafura and Perfetti present evidence that skilled readers are better able than less skilled readers to integrate new words into their mental models and that the skill differences in integration processes in general may depend on knowledge of word meanings. Olson, Keenan, Byrne and Samuelsson’s paper is an extended version of an earlier paper that appeared in the special issue. In this chapter, Olson et al. review the main results from their two large-scale behaviour genetic studies and examine the consistency of the results across other similar projects in developed countries. Perhaps the most critical points arise in their discussion of what these results (frequently showing a high genetic influence on individual differences) show and don’t show, and how they are currently interpreted and frequently misinterpreted. Irrespective of the reader’s expertise in behaviour genetic research methods and findings, these discussions should be compulsory pre-readings for any discussion on the role of genetic studies in education. While the first two papers focus mostly (but not exclusively) on studies completed with English-speaking subjects reading or learning to read English, the next two chapters expand the coverage to other orthographies. These two chapters are new doi 10.1075/swll.15.01cai © 2017 John Benjamins Publishing Company

8

Kate Cain, Donald L. Compton and Rauno K. Parrila

and were written specifically for this volume. McBride provides first an overview of the larger context of the early reading and writing development in different orthographies by discussing possible interactions between various dimensions of orthographies, cognitive dimensions of children, and the contexts within which learning takes place. She then focuses specifically on how critical phonological (both segmental and suprasegmental), semantic, and orthographic information is conveyed in different scripts, and how these differences may affect the tasks of learning to read and write. Throughout the chapter, her discussion is anchored in her expertise in both Chinese and English orthographies. She concludes that the comparison of multiple scripts can not only greatly enhance our understanding of the universals and specifics in literacy learning, but also lead to discovering new aspect of literacy development that may be relevant to literacy learning across scripts. In the fourth chapter, Nag turns the focus on learning to read Indic alphasyllabaries using aksharas, which are visually complex symbols of mixed granularity, as the orthographic units. A brief introduction to the akshara writing systems succinctly illustrates both how different they are from alphabetic and character based writing systems, but also the heterogeneity among various akshara systems. When compared to alphabetic writing systems, akshara orthographies and their learning contexts present new challenges, which are rare or nonexistent in alphabetic orthographies and in contexts where they are learned. Nag presents evidence that learning to read in akshara orthographies requires both the same and some additional cognitive-linguistic processes than what are needed to learn to read an alphabetic orthography. A further intriguing possibility remains that the relationships between cognitive-linguistic predictors and reading outcomes that statistically appear similar are based on different effective mechanisms. Herein may lie the biggest promise of cross-linguistic studies helping us to both develop new and to test existing theoretical models of literacy acquisition. Perhaps the main take-home message from the chapters by McBride and Nag is that the bar for claiming universality needs to be set much higher than when the focus is solely on alphabetic orthographies. In the final chapter of Part 1, Steacy, Elleman and Compton develop further the ideas outlined in their contribution to the SSR special issue. Similar to the first chapter of this section, they focus on word knowledge and how it is acquired, specifically orthographic and semantic knowledge. As a result, this chapter speaks to the influence of words on comprehension, as well as word recognition processes. They marshal a range of evidence to support the role of inductive learning mechanisms in word acquisition and how these interact with other knowledge structures to support general development and individual differences in reading. They argue that the limited success of many interventions designed to ameliorate reading difficulties is likely because the teaching methods are not aligned to how words are learned. Thus, a critical message from this chapter, and one that is central to the work represented in this volume, is that theory of reading development and difficulties should play a central role in the development of instruction and intervention.

Integrating word processing with text comprehension Theoretical frameworks and empirical examples Joseph Z. Stafura and Charles A. Perfetti University of Pittsburgh

We previously observed that “there is no theory of reading, because reading has too many components for a single theory” (Perfetti & Stafura, 2014, p. 1). 1 In the pursuit of studying reading and reading development, then, research has largely been driven by flexible frameworks and specific problems rather than by precise theoretical testing. How readers comprehend and how skill differences arise are intertwined problems in reading theory. We suggest that the reintroduction of a broad-scope, general framework for reading can aid the formulation of specific hypotheses about reading expertise and reading problems. Following the presentation of the framework we take a closer look at on-line text-based comprehension processes. In this section, we discuss current work in our lab addressing the functionality of word knowledge and word processing in text and discourse comprehension. First, we consider the broader context for comprehension theories.

Theories of reading comprehension Two complementary ideas shaped the modern study of reading comprehension, one that described an enriched level of comprehension beyond the literal meaning of a text – the reader’s situation model (Van Dijk & Kintsch, 1983) – and one that described the cognitive dynamics of text comprehension, the construction-integration (C-I) model (Kintsch, 1988). The C-I model made general assumptions about the reader’s cognitive architecture (e.g., limited memory) and cognitive procedures (e.g., retrieval and carry-over operations) as well as text devices (e.g., argument overlap) that support comprehension. The C-I theory was critical in explaining text comprehension by an interactive combination of top-down (knowledge-driven) and bottom-up (wordbased) processes. 1. The article by Perfetti and Stafura (2014) provides the original material from which this chapter is based. doi 10.1075/swll.15.02sta © 2017 John Benjamins Publishing Company

10

Joseph Z. Stafura and Charles A. Perfetti

Prior to this, ideas about text comprehension had been dominated by demonstrations of knowledge-driven, top-down procedures guided by scripts (Shank & Abelson, 1977) and other forms of schemata (Anderson, 1978; Bartlett, 1932). Following van Dijk and Kintsch (1983) and Kintsch (1988), text comprehension research headed in new directions, building on these models of text-knowledge interaction and developing enriched variations (Goldman & Varma, 1995) and updates on the basic idea (Kintsch & Rawson, 2005). Theories tackled one or more aspects of comprehension (for a review, see McNamara & Magliano, 2009). The landscape model (Van den Broek et al., 1996) targeted the “landscape” of activation patterns that wax and wane during reading and how the reader’s goal of maintaining coherence guided these patterns. The structure building theory (Gernsbacher, 1990) also assumed a central role for coherence, which was viewed as the result of the structures built and connected by the reader, and provided hypotheses about individual differences in comprehension. The event-indexing model (Zwaan, Langston, & Graesser, 1995) elaborated the idea of the situation model toward a more comprehensive multidimensional tracking of various aspects of narrativity. Among the many issues targeted in these theories, an especially important one concerned inferences: that they were necessary for comprehension and how and when they were made (Graesser, Singer, & Trabasso, 1994). A specific contribution of this research was to show that readers make inferences that maintain coherence, including causal inferences that connect actions in narratives (Graesser & Kruez, 1993). The current empirical status of these and other theories (e.g., embodied comprehension; Zwaan, 2003) remains under study and is beyond the scope of the present article. The point here is that broadly-defined contrasts – for example, verbatim versus gist memory, literal versus inferential text processing, coherent versus incoherent texts – can be addressed in models that include interactions among knowledge sources that are initiated by written word reading, rather than solely by top-down knowledge-generating inferences. Still, it is fair to say that attention to how word processes actually contribute to comprehension has been minimal, with a few notable exceptions that include the role of word meaning selection in the structure building framework (Gernsbacher, 1990) and word activation processes in the construction phase of the C-I model (e.g., Kintsch & Mross, 1985). The theories just discussed are global frameworks rather than specific theoretical models. The value of a framework for something as complex as comprehension is that it provides a set of interconnected claims that, with the addition of specific assumptions, can lead to theoretical models with testable propositions and implications. Nevertheless, in contrast to well-defined models of word reading that make precise predictions (e.g., Coltheart, Rastle, Perry, Langdon, & Ziegler, 2001), reading comprehension is too broad a target for precise models. As we illustrate in the next section, there is value in capturing this breadth in a general framework that provides a view of the component subsystems of reading comprehension.



Integrating word processing with text comprehension

Linguistic and Writing System Knowledge Linguistic System

Orthographic System Mapping to Phonology/Morphology

Parser

Text Representation Situation Model

Inferences

Meaning Morphology Syntax - argument structure - thematic roles

Meaning and From Identification

Linguistic Units

Comprehension Processes

Lexicon Word Identification

Visual Input

Orthographic Units

Phonology, Syntax, Morphology

General Knowledge (including text structure)

Figure 1.  The Reading Systems Framework. Note. The components of reading within a language-cognitive architecture from visual processing through higher level comprehension. The key elements are knowledge sources, basic cognitive and language processes, and interactions among them. The framework allows the development of specific models (e.g., word identification models, models of inferences) and allows hypotheses about both the development of reading expertise and reading weaknesses. A particular point of focus is the lexicon, which is a central connection point between the word identification system and the comprehension system. Based on Perfetti (1999).

The Reading Systems Framework A general framework of reading systems must reflect reading more fully by adding word level processes to the higher level processes that are the focus of comprehension research. Figure 1 presents a variation of such a framework, derived from a “blueprint” of the reader (Perfetti, 1999) and used to frame problems in comprehension (Perfetti, Landi, & Oakhill, 2005). This Reading Systems Framework makes the following claims about reading:

1. Three classes of knowledge sources are used in reading: linguistic knowledge, orthographic knowledge, and general knowledge (knowledge about the world, including knowledge of text forms, e.g., text genres).

11

12

Joseph Z. Stafura and Charles A. Perfetti

2. The processes of reading – decoding, word identification, meaning retrieval, constituent building (sentence parsing), inferencing, and comprehension monitoring – use these knowledge sources in both constrained ways (e.g., decoding uses orthographic and phonological knowledge but not general knowledge) and in interactive ways (e.g., inferences use general knowledge and propositional meaning extracted from sentences). 3. These processes take place within a cognitive system that has pathways between perceptual and long-term memory systems and limited processing resources. This framework finds support in neurobiological models of language. Hagoort’s (2005; 2013) MUC model posits a functional core of memory, unification, and control operations in language processing. In terms of reading comprehension, these are functional during encounters with words, as input from the visual orthographic system drives memory operations in the temporal lobes that retrieve associated linguistic and general knowledge from long-term memory. Unification operations in the left inferior frontal gyrus act to integrate the word-level syntactic and semantic knowledge into the ongoing context (e.g., into a sentence). Finally, control operations seated in the dorsolateral prefrontal cortex and anterior cingulate cortex guide efficient processing with limited resources. This neurobiological framework is broadly consistent with C-I models (Kintsch, 1988) of comprehension in its focus on bottom-up activity and incremental integration into ongoing higher level, coherent text representations. The Reading Systems Framework can also guide the formation of novel theories and hypotheses of reading problems. Readers can show weaknesses in specific knowledge sources, which then affect processes that use these knowledge sources in reading. An alternative view, the dominant one, is that it is weaknesses in the processes themselves that lead to comprehension breakdown. It is typically difficult to choose between these two views. Is a measured weakness in decoding due to a processing problem involving the conversion of orthography to phonology? Or is it due to a knowledge weakness about phonology or the rules that link orthography to phonology? A weakness in these links could result in difficulty accessing what are actually high quality phonological representations (Boets et al., 2013). Is an observed problem in inference making due to a weak inference process or to a lack of knowledge that is needed to make the inference? To a limited extent these are empirical questions. For example, in the case of the inference/knowledge debate, one can try to control for knowledge (Cain, Oakhill, Barnes, & Bryant, 2001). However, even with the best of efforts it is difficult to persuasively assess processes in isolation of other processes and, especially, in isolation of the knowledge sources on which they rely. The Reading Systems Framework is useful in the generation of hypotheses about the sources of comprehension problems in several ways. One is to identify reading problems by measureable weaknesses within one or more of the components (knowledge and processes) of the framework. This works especially well with lower level processes, where failure in decoding defines basic reading disability or dyslexia. More specific hypotheses then focus on even more fine-grained components in the visual



Integrating word processing with text comprehension 13

or phonological subsystems as sources of reading disability, with the bulk of the evidence showing phonological processing problems (Vellutino, Fletcher, Snowling, & Scanlon, 2004). This strategy does not work nearly as well with higher level processes, because these depend on receiving high-quality input from word-level and sentence-level sources. Thus, careful testing of higher level sources of comprehension problems must attempt to control for some of the lower level components (e.g., Cain, Bryant, & Oakhill, 2004; Cain, Oakhill, & Bryant, 2000; Oakhill, Cain, & Bryant, 2003). The result is the identification of specific reading comprehension difficulties, of which several have been proposed, with no single difficulty emerging as definitive (Cain, 2010; Cain & Oakhill, 2006; Nation, 2005). Another strategy is to hypothesize pressure points in the reading system. For example, the Lexical Quality Hypothesis (Perfetti, 2007; Perfetti & Hart, 2002) assumes that the lexicon is a critical pressure point in reading comprehension. The lexicon sits astride two reading systems: one, the word identification system, requires high-quality linguistic and orthographic information to enable rapid word identification; the second, the comprehension system, takes its input from the word identification system to build meaning units (propositions). Knowledge of written word forms and meanings, then, is central to reading and thus a pressure point for reading comprehension – a prime candidate for a cause of reading comprehension difficulty, and a critical level of analysis for extending theory. As a scaffold for theory development and hypotheses testing, the Reading Systems Framework allows other hypotheses about pressure points and about the interactions between reading subsystems. In what follows, however, we focus on the lexical component and its interaction with text and discourse representations.

Comprehension skill within the lexical system of the Reading Systems Framework The focus on a lexical subsystem in reading arises from the centrality of word meanings (represented in a long-term memory) as (a) the output of word identification and (b) the input to comprehension processes. This leads to research on the relations between lexical processes and comprehension processes and the following two complementary hypotheses:

1. Text comprehension depends on understanding words and integrating their meaning into a mental model of the text, and more skilled comprehenders do this better than less skilled comprehenders (Perfetti, Yang, & Schmalhofer, 2008; Yang, Perfetti, & Schmalhofer, 2005, 2007; Stafura & Perfetti, 2014). 2. Learning words depends on acquiring information about both word forms and meanings from word-learning events, and more skilled comprehenders do this better than less skilled comprehenders (Bolger, Balass, Landen, & Perfetti, 2008; Perfetti, Wlotko, & Hart, 2005; Van Daalen-Kapteijns & Elshout-Mohr, 1981).

14

Joseph Z. Stafura and Charles A. Perfetti

Both hypotheses are now empirical generalizations insofar as they are consistent with available evidence: correlational evidence for the first and correlational and intervention-based experimental evidence for the second. In the next section, we examine more closely the nature of this evidence and its implications for hypotheses about the sources of reading comprehension problems.

Comprehending texts includes comprehending words In the Reading Systems Framework, a key set of processes links lexical outcomes with comprehension (Figure 1; “meaning and form selection”). Early sentence comprehension processes that build sentence constituents (e.g., noun phrases, verb phrases) and propositions (elementary meaning units) make use of this link. (Notice also the bidirectional link, which allows word learning to result from comprehension.) These links can be studied only by on-line measures that expose word level reading comprehension while it happens and not by observations made after a text has been read. There are three important methods for obtaining such measures: (a) word-byword reading controlled by the reader, (b) eye-tracking, and (c) event-related potentials (ERPs) during text reading. The first has ease of instrumentation, but it allows reader strategies some influence. Eye-tracking and ERPs, which are measures taken without overt decisions required by the reader, provide the clearest evidence of word-to-comprehension links. Each has its advantages: Eye-tracking allows natural movements of the eyes. ERPs, which generally require that the eyes not move, allow multiple word processing components (e.g., visual attention, orthographic recognition, meaning processes, syntactic processes) to be observed on a single word. Next we focus on ERP studies. Word-to-text integration We assume that, for a motivated reader, understanding entails a mental representation of the “situation” described by a text (Van Dijk & Kinstch, 1983). In terms of structure, for our initial purpose we assume only that an unfolding narrative text asserts situations and events and that the reader builds and updates a situation model accordingly (Zwaan & Madden, 2004). For now, we also postpone our discussion of the potential effects of older non-updated information remaining in memory (O’Brien, Rizzela, Albrecht, & Halleran, 1998). A key additional assumption is that comprehension proceeds along multiple input units. For example, a noun is understood through lexical meaning retrieval, a noun phrase is understood through additional referential processes, and a clause that includes the noun phrase is understood through additional lexical and parsing processes. So, for example, in reading the sentence, “The rain ruined her beautiful sweater,” the following comprehension processes are centered on the understanding the word “rain”: (a) retrieval of meaning of “rain,” (b) establishing definite situational reference of “the rain” (cf. “a rain would ruin the picnic”), and (c) extracting a proposition in which “the rain” is the subject of “ruin her beautiful sweater” and thus the cause of the ruining event. Of course, only the first two operate on the reading of “the rain” with the predication about ruining the sweater requiring additional reading. It



Integrating word processing with text comprehension

is these two processes that are in focus in the following analysis. Of specific interest is what happens across a sentence boundary, which is a paradigm case of text integration processes. We begin with a single sentence, from which a situation model can be constructed: (1) While Cathy was riding her bike in the park, dark clouds began to gather, and it started to storm. 2 In Figure 2, a simple scheme represents a possible situation model a reader might have from the reading of (1). The situation includes four referential entities – Cathy, the park, the bike, and dark clouds – and an event – the storm. Referents are essential in the model, because referents are eligible (unequally) for elaboration. Events are normally but not necessarily established through verbs, and these events also become eligible for elaboration. With this situation established, the text adds a new sentence:

- IN THE PARK - CATHY ON BIKE

- STORM

- STORM - CATHY ON BIKE

- DARK CLOUDS < SITUATION 1>

< SITUATION UPDATE>

Figure 2.  A situation model. Note. The model represents what a reader might understand after reading the sentence While Cathy was riding her bike in the park, dark clouds began to gather, and it started to storm. The general form of the model is SITUATION + EVENT. Adapted from Perfetti and Stafura (2014).

(1) While Cathy was riding her bike in the park, dark clouds began to gather, and it started to storm. The rain ruined her beautiful sweater. The noun phrase that begins the new sentence – the rain – is understood immediately in relation to the situation model. It refers to the storm event, to which it can be integrated as part of the model. Figure 2 would now incorporate the new event – the ruination of the sweater. Experiments summarized in Perfetti et al. (2008), using sentences similar to (1), measured the ERP responses initiated by a target word – “rain” in the current example. When the target word appeared, the N400 component, an indicator of the degree of fit between the word and its context experienced by the reader (Kutas & Federmeier, 2000), was reduced in amplitude relative to a baseline condition (2):

2. Our discussion relies on relatively simple narrative texts, which contain clear event structures to expose examples of integration processes. However, our theoretical framework and the wordto-text integration processes we examine apply to texts of all types.

15

16

Joseph Z. Stafura and Charles A. Perfetti

(2) When Cathy saw there were no dark clouds in the sky, she took her bike for a ride in the park. The rain that was predicted never occurred. Here, the N400 on the word “rain” has a more pronounced negative deflection, because initially it requires more processing for meaning activation, meaning integration (with the text), or both. There is no antecedent for “rain” in the preceding sentence. Equivalently, the situation model contains no referent to which the new event, “the rain,” can be attached. Unlike in (1), there is no “storm” event to support the integration of “the rain” into the model. Instead the reader must build a structure around this new event. It is important to emphasize that the difference between texts (1) and (2) is not about their sensibility or coherence. Text (2) is fully sensible. Thus, the N400 comparison of texts (1) and (2) on “rain” is quite subtle compared with the more typical studies of the N400, which use anomalous sentences (e.g., “related anomaly” paradigm; Amsel, DeLong, & Kutas, 2015). For example, in a classic N400 study, an ERP recorded on “eat” in “The pizza was too hot to eat” was compared with the ERP recorded on the anomalous “drink” in “The pizza was too hot to drink” (Kutas & Hillyard, 1980). In these conditions, the N400 differences are dramatic and may be explained by expectancy violations (Kutas & Federmeier, 2000; Lau, Almeida, Hines, & Poeppel, 2009; Van Petten & Luka, 2012). However, in our case, a comparison is made across sensible texts that differ only in the degree to which they invite an immediate word-to-text integration process. (See Brown & Hagoort, 1993, for an N400 interpretation based on post-lexical integration processes.) Expecting a certain word across a sentence boundary seems unhelpful to comprehension; nearly any grammatical sentence beginning can continue with coherent ties to the preceding sentence. Thus the reduction of the N400 in our case is not likely about expectancy violations, but about integration.

The paraphrase effect and comprehension skill The word “rain” is better integrated with text (1) than with text (2). We refer to this as the paraphrase effect, with the understanding that this is not exactly the everyday sense of “paraphrase,” that is, expressing an idea in words different from its original expression. In our usage, paraphrase is an implicit co-referential relation between a word or phrase in one sentence and a word or phrase in a following sentence. The co-referential relation is defined by the contents of the mental representation of the enriched semantic content of the text – the situation model. The paraphrase can update the situation model modestly (or merely reinforce the salience of a referent) while maintaining coherence. Thus, in text (1), “rain” is not another way of saying “storm.” Rather, “rain” fine-tunes – or elaborates – the mental model by identifying a correlate or consequence of the storm, which was established in the first sentence. The paraphrase effect reflects online comprehension, an updating of the situation model that integrates a word with a text representation. In addition to evidence for this integration process in ERP records, we discovered that skilled comprehenders showed the paraphrase effect more robustly than less skilled comprehenders, who



Integrating word processing with text comprehension 17

were described as showing sluggish word-to-text integration (Perfetti et al., 2008). This sluggishness can have consequences for maintaining coherence across sentences, as memory resources, which are required for comprehension repairs, become less available. Recent work has supported this view of individual differences, finding that the paraphrase effect occurred only among highly skilled readers (Stafura & Perfetti, 2014). Thus, these anaphoric elaboration processes – centered on word level processing – are sensitive to reading ability. Word-to-text integration can involve inferences. Indeed, one might argue that the paraphrase effect is a kind of bridging inference (Haviland and Clark. 1974; Graesser, Singer, & Trabasso, 1994; Singer & Halldorson, 1996). In our previous example, a bridging inference could link “the rain” back to the storm, preserving coherence. On this description, we could say that skilled comprehenders make this bridging inference more readily than less skilled comprehenders. However, such a description seems incomplete without a focus on its lexical basis, and, further, it would beg the question of what makes this bridging inference more likely for the skilled comprehenders. Instead, we think describing the rain-to-storm link-up as a lexically based integration process (word-to-text) better captures the cognitive operations involved and frames a hypothesis for why there is a skill difference. Thus, instead of focusing on “broken” bridging processes one focuses on word knowledge and context-sensitive meaning selection that are required for the integration process. There is an important role for bridging inferences in this kind of word-to-text integration, however. If the text of the first sentence has only an implication of rain rather than establishing a rain-related event (storm), the integration process requires bridging, as in text (3): (3) While Cathy was riding her bike in the park, dark clouds began to gather. The rain ruined her beautiful sweater. When the reader encounters “rain” in (3), there is no storm event in the mental model to which “rain” can be attached. Instead, the reader makes a bridging inference, constructing a new event: Rain. This bridging inference (e.g., Graesser et al., 1994; Singer & Halldorson, 1996) is readily made, although with some cost to processing efficiency. Yang et al. (2007) observed that for texts of this type, the N400 amplitude was not significantly different from baseline. Thus, reading “The rain …” in sentence (3) was similar to reading “The rain …” in sentence (2) as far as the ERP record was concerned. The costly bridging inference is unnecessary if, in the first sentence, the reader makes a forward or predictive inference. Such an inference would occur while reading the first sentence of (3), specifically the segment “… dark clouds began to gather.” This inference is a prediction (it will rain) the reader might make (Graesser et al., 1994). The inference has little warrant, however, so adding rain to the mental model is a risky move. Certainly the comprehension of “dark clouds” in the first sentence allows “the rain” to be easily understood when it does appear in the next sentence. (Hence, the N400 to (3) is not more negative than in (2).) However, it does not compel a forward inference (McKoon & Ratcliff, 1992). The N400 results of Yang et al. (2007) strongly

18

Joseph Z. Stafura and Charles A. Perfetti

suggest that skilled readers do not make the forward inference consistently, and thus had to make a bridging inference when they came to the word “rain.”. (For further discussion that connects inferences and word-to-text integration processes see Perfetti & Stafura, 2015) To summarize: Word-to-text integration processes are central to comprehension because they recur with each phrase. They reflect a close coupling of word identification with representations of the meaning of the text, mediated by the retrieval and selection of word meanings. Word-to-text integration processes are pervasive, and the processes that produce the paraphrase effect are only part of the integration picture. Other anaphoric processes, ranging from simple pronoun binding through more complex co-referential expressions are also relevant, as are bridging inferences. All these processes maintain coherence at variable costs to comprehension efficiency. Comprehension skill depends in part on these word-to-text integration processes. Those processes that depend on word meanings are especially likely to show individual differences, because knowledge and use of word meanings is highly variable across individuals. Explaining further the association between reading skill and the paraphrase effect requires more research. Candidate explanations within the Reading Systems Framework include (a) individual differences in the lexicon, either vocabulary size (in a familiarity or passive knowledge sense) or more finely tuned word knowledge that supports the use of words in specific contexts; (b) cognitive architecture factors, including working memory limitations (Just & Carpenter, 1992); and (c) problems in executive functioning (e.g., Cutting, Materek, Cole, Levine, & Mahone, 2009) that can cause less effective inhibition of irrelevant word level semantic information (Gernsbacher, 1990).

Functional mechanisms of word-to-text integration More research is needed to understand the mechanisms of word-to-text integration, aside from skill differences. The cross-sentence paraphrase effect is a general language process, found in listening comprehension as well as reading (Adlof & Perfetti, 2011). Stafura and Perfetti (2014) explored the question of whether the message level of comprehension (what the text means) or the lexical level (the association between prior words and the word being read), or both, are responsible for the paraphrase effect. Although the message level must be involved if the effect is about comprehension, lexical processes initiated by word identification, including associations that a word has with other words in memory and with other words in the text are part of the process. In the C-I model, associations are activated through rapid, automatic processes in the construction stage and may have no consequences for the later integrative stages of comprehension. However, if the text contains words that take advantage of the associations that are evoked unconsciously, then associations may provide a head start on message-level comprehension. Stafura and Perfetti (2014) compared ERP responses to critical words (e.g., rain) that were either strong (4) or weak (5) associates of the referentially-related antecedent words (italicized) in the first sentence.



Integrating word processing with text comprehension 19

(4) While Cathy was riding her bike in the park, dark clouds began to gather, and it started to storm. The rain… (5) While Cathy was riding her bike in the park, dark clouds began to gather, and it started to shower. The rain… Similar to the original reports of the paraphrase effect (Perfetti et al., 2008; Yang et al., 2005; 2007), critical words in both strong association and weak association conditions elicited reduced N400 responses relative to the same critical words in baseline sentences such as (2). Crucially, there were no differences between the strong association and weak association conditions. This finding suggests that, within the parameters of these materials, the message level is the dominant locus for the paraphrase effect. An additional question is whether the lexical component of the text integration process takes advantage of forward association processes or uses memory based backward processes (Stafura, Rickles, & Perfetti, 2015). In our example, does “storm” in the first sentence evoke “rain” as an associate, which is then available to support integration when “rain” is encountered in the next sentence? Or is the more important process that when “rain” is read in the second sentence it resonates with the memory of “storm” from the first sentence. The critical comparison is between forward association conditions like (4) with backward association conditions like (6); (6) While Cathy was riding her bike in the park, dark clouds began to gather, and it started to rain. The storm ruined her beautiful sweater. Stafura and colleagues (Stafura et al., 2015) analyzed both mean evoked amplitudes and principal components extracted from the ERP data, finding both similarities and differences in critical word processing depending on the dominant direction of association. Consistent with the paraphrase effect findings in the past, both association conditions elicited reduced mean N400 amplitudes over central electrodes relative to baseline conditions. Additionally, forward and backward association conditions elicited ERPs that differed in both earlier (~200ms) and later (≥425ms) time-windows. We interpreted the late component in terms of memory processes (Rugg et al., 1998; Rugg & Curran, 2007), specifically, those involved in discourse updating (Burkhardt, 2007). These may be passive resonance processes (O’Brien et al., 1998) leading to activation of the co-referential information in the first sentence. This activity was greatest in the backward association condition wherein the critical word could act as a retrieval cue for the preceding sentences’ meaning structure (e.g., propositional; Ericsson & Kintsch, 1995; O’Brien, Plewes, & Albrecht, 1990; Ratcliff & McKoon, 1988) leading to long-lasting positivity in the ERP wave. Together, this evidence supports a message level locus for the paraphrase effect, at least within the parameters of the short, sensible materials used in these studies. At the same time, there are some tentative indications that lexical level processes play distinct functional roles in word-to-text integration, with words potentially acting as retrieval cues for the meaning of the preceding text.

20 Joseph Z. Stafura and Charles A. Perfetti

Structural aspects of text and word-to-text integration Identifying the structure and situational dimensions of mental representations of text (e.g., Zwaan & Radvansky, 1998) and how they interact as the reader builds an understanding of the text (Rapp & Taylor, 2004) are important topics of comprehension research. One area in which we have begun exploring the effects of text and discourse level factors on word level processing relates to the semantic – or structural – centrality of a text (van den Broek, Helder, & Van Leijenhorst, 2013). A given word’s structural centrality is defined as the degree to which that word is central to the semantic structure of the text. Readers are typically better at recalling information that is central to the semantic structure of texts, compared to non-central information. However, studies on centrality to this point have measured comprehension off-line, after reading (for a review see van den Broek et al., 2013). We recently carried out an experiment to test the effects of structural centrality on on-line, word level processing using ERPs (Helder, Stafura, Calloway, van den Broek, and Perfetti, 2015). Table 1 provides an example narrative in 3 conditions used in this study. The Theme Conditions have semantic structures that differ in terms of their central themes. In Table 1, Theme 1 has “Weather” as a central theme, and Theme 2 has “Clothes” as a central theme. Critical words from which ERPs were measured are bolded if they are related to the first theme (e.g., Weather), and bolded and underlined if they are related to the second theme (e.g., Clothes). Baseline passages that are neutral regarding the experimental themes were used in the first critical word analysis only (due to constraints in making the texts sensible and coherent). Table 1.  Example passages. From Helder et al., 2015. Condition

Example passage

Theme 1 – Weather

Cathy likes to check the weather all the time on her iPhone. She is always very excited when stormy weather is predicted. While Cathy was riding her bike in the park, dark clouds began to gather, and it started to storm. The rain ruined her beautiful sweater. Cathy loves clothes and bought herself a new wardrobe. She is getting ready to go outside and decides to wear her new outfit today. She noticed that a lot of people were looking at her clothes while it started to storm. The rain ruined her beautiful sweater. Cathy lives close to a park. She likes to be there as much as she can during the summer. When Cathy saw there were no dark clouds in the sky, she took her bike or a ride in the park. The rain that was predicted never occurred.

Theme 2 – Clothes

Baseline



Integrating word processing with text comprehension 21

Helder et al (2015) found no effect of centrality at the first critical word (e.g., rain); baseline passages elicited greater average N400 responses than the experimental conditions, which did not differ. However, text-final words (e.g., sweater) revealed effects of centrality with not-thematically-related words eliciting greater average P600 responses relative to thematically-related words. Thus, structural centrality influences on-line word-to-text integration processes at the text-final position. This evidence suggests that the message level influences on-line comprehension processes at the lexical level. At earlier positions in a sentence, local operations such as co-referential binding seem to have a dominant role. At the end of the sentence(/text) message-level processes appear dominant. Here, the P600 effect may reflect the ease of mental model updating when words are related to the central theme of the passage, compared to when they are not related to the central theme (but sensible).

Memory updating and word-to-text integration Text and discourse research has demonstrated that “outdated” discourse information (information no longer available in working memory) can influence later comprehension processes (Albrecht & O’Brien, 1993; O’Brien & Albrecht, 1992; O’Brien et al., 1998). However, it is unclear whether comprehenders are sensitive to this outdated information at the point of the first potentially conflicting word, or only at a later point in reading, as previous methods have almost completely relied on self-paced reading of entire sentences. Currently, our lab is carrying out ERP studies to test the influence of outdated discourse inconsistencies on a word-by-word basis. Findings of immediate sensitivity at the word-level would provide insight into on-line processing connections between message- and word-levels across large stretches of text. Specific ERP components can then allow us to zero-in on mechanisms that may be functional in this process. For example, inconsistent words may attract attention or updating processes related to information in working memory, eliciting larger P300 responses than consistent words (Donchin & Coles, 1998). Additionally, inconsistent words may lead to extended memory retrieval and analysis related to the earlier conflicting “episode”, perhaps through passive resonance mechanisms (Myers & O’Brien, 1998) that have been shown to selectively improve memory of information surrounding inconsistent material (O’Brien & Myers, 1985). These extended processes may lead to a prominent late positivity (P600) at the critical word in inconsistent relative to consistent conditions, such as was seen in the directional association study by Stafura et al. (2015). On the other hand, the findings of Helder et al. (2015) suggest that, in some conditions, effects of message level factors may be delayed, eliciting effects on sentence- or text-final words. We are currently pursuing these and other connections between word level processing and message level processes engaged during on-line reading of longer, relatively natural texts.

22

Joseph Z. Stafura and Charles A. Perfetti

Knowledge of word meanings is instrumental in reading comprehension The paraphrase effect, and word-to-text integration processes in general, demonstrate subtle roles of word meanings in comprehension (in addition to their obvious role in allowing comprehension) and leads to the question of what kinds of word knowledge are responsible for integration and comprehension success. The Lexical Quality Hypothesis (Perfetti, 2007; Perfetti & Hart, 2002) assumes that word knowledge (both form and meaning) is central to reading skill. High-quality form knowledge includes phonological specificity, the lack of which has been linked to problems in reading and word learning (Elbro, 1998; Elbro & Jensen, 2005). It also includes orthographic precision, which has been shown by Andrews and colleagues to have specific consequences beyond the effects of reading skill. Spelling-based lexical expertise effects are seen in lexical access (Andrews & Hersch, 2010; Andrews & Lo, 2012) and spelling/ vocabulary lexical expertise effects can show subtle effects in the balance of top-down and bottom-up processes in comprehension (Hersch & Andrews, 2012). The semantic constituent of lexical quality has a close connection to comprehension, as is well established by correlations between vocabulary and reading comprehension (e.g., Adlof, Catts, & Little, 2006; Cromley & Azevedo, 2007). The impact of vocabulary on reading is usually assumed to be indirect through its role in general language comprehension. However, it is possible that vocabulary also has a direct effect on reading itself. An observation by Braze, Tabor, Shankweiler, and Mencl (2007) is interesting on this point. Across a range of adolescent and adult readers, Braze et al. (2007) found that vocabulary accounted for reading comprehension to a greater degree than it did listening comprehension. They argued that this reflects the fact that written words are more likely to fail to activate lexical representations than are spoken words. In effect, a stronger semantic connection (a more integrated set of word constituents) can compensate for lower orthographically initiated activation. Consistent with this possibility are the results of other cross-sectional and longitudinal studies. Protopapas, Sideridis, Mouzaki, and Simos (2007) in a study of 534 children in Grades 2, 3, and 4 in Greek schools in Crete found a strong relationship among reading comprehension, vocabulary, and decoding. However, the unique contribution of decoding, relative to its shared variance with vocabulary, was negligible with vocabulary taken into account, especially beyond Grade 2. In a longitudinal study that followed 2,143 Dutch children through Grade 6, Verhoeven and Van Leeuwe (2008) found that at Grade 1 reading comprehension was accounted for by a structural model that combined word decoding and listening comprehension. Examining later grades with time-lagged correlations, they found that earlier vocabulary predicted later reading comprehension, whereas earlier listening comprehension did not. Accounting for word meaning knowledge as part of reading provides a challenge for the assumption that decoding a word unlocks all the knowledge associated with the spoken word. The Simple View of Reading (Hoover & Gough, 1990), an expression of this assumption, would need to accommodate the direct effects of vocabulary on reading comprehension by allowing vocabulary knowledge to influence decoding



Integrating word processing with text comprehension 23

(Tunmer & Chapman, 2012). Word meaning would thus contribute to reading both as a component of language comprehension and through word reading. Indeed, recent structural equation models of language ability among children in grades 1–3 support indirect effects of vocabulary on both word reading and listening comprehension (Language and Reading Research Consortium, 2015). Although the spoken language may be the main carrier of word meanings, it is the retrieval of word meanings through orthographic representations (and their integration with text meaning) that is critical in reading. A second aspect of the word knowledge – comprehension connection concerns learning new words. (Figure 1 shows this connection by arrows from comprehension back to the lexicon.) During reading, readers implicitly infer meanings from imperfectly understood text, allowing the establishment of a new lexical entry or the refinement of an existing one. Readers of greater skill, word knowledge, and experience are more effective at this learning. Experimental evidence for this conclusion spans studies of children and adults and reveals skill differences in learning new words implicitly from text, as well as from direct instruction (e.g., Cain, Oakhill, & Lemmon, 2004; Perfetti, Wlotko, & Hart, 2005). In the present context, the relevance of these twin aspects of the word knowledge – comprehension relationship is the centrality of word knowledge in a theory of comprehension. The word-to-text integration evidence is that skilled readers are better able to integrate words into their mental models of the text, and that word level processing reveals text level factors on-line. The correlational evidence on the word knowledge – comprehension relationship and the experimental evidence on learning new words together suggest that skill differences in the integration processes may depend on knowledge of word meanings or the use of this knowledge during text reading. In the final section, we return to the theoretical concerns we raised at the beginning, showing how the Reading Systems Framework can guide more specific hypotheses about comprehension and differences in comprehension skill.

Word comprehension within the Reading Systems Framework Word reading in context is about word comprehension, which is at the center of the Reading Systems Framework. Word comprehension is the output of the word identification system and the input to the comprehension systems (sentence, text, and situation). Figure 3 is a wide-angle lens view of this part of the framework, showing (in an altered spatial orientation) the word identification system on top and the word comprehension system on the bottom. Word comprehension is word-to-text integration in this view. Word meanings stored in memory (the lexicon) are only part of word comprehension, as they (and other memory-driven associations) are activated during reading and then tuned to what the context (the representation of the situation) demands.

24 Joseph Z. Stafura and Charles A. Perfetti

WRITTEN TEXT

Orthographic units

Phonological units

Word identification

Semantic units

Word- to-text integration

Sentence representation

Situation model

Text model

Prior Knowledge

Figure 3.  The connection between two systems that support comprehension, “The word identification system and the word comprehension system,” is illustrated. As words are identified, they are comprehended in relation to the representation of the text. The comprehension process links the word to an existing referent (or event) in a mental model or extends the mental model to include a new or updated referent (or event). Semantic units activated with word identification are part of this process (the activation phase of the CI model), but the selection of meaning is influenced by the reader’s immediate representation of the text. From Perfetti and Stafura (2014).

To return to the review of comprehension theories, we see that the word comprehension model corresponds roughly to the construction (upper part of Figure 3) and integration phases (lower part of Figure 3) of the C-I model (Kintsch, 1988). However, we do not assume that this integration is necessarily an active process. It is at least partly (if not mainly) a memory-driven process, in which words from the recently read text and the propositions they encode (the text model) are highly accessible in memory. A word, as it is read, “resonates” with these memories, and connections are made without an active construction process, which can later tune and correct the representation.



Integrating word processing with text comprehension 25

This process is adaptive for comprehension insofar as what is activated in memory is relevant and consistent with the current state of the situation model. However, even information that has been updated and is no longer relevant can continue to exert an influence on comprehension (O’Brien, Cook, & Guéraud, 2010; O’Brien et al., 1998). Active construction may become necessary when coherence breaks down and requires new structures to be built (Gernsbacher, 1990). However, the value of a more passive memory process is that text integration can occur at low cost to processing resources, and this may be the default integration processing mode within and across sentences. It is not completely clear whether immediate updating of the situation model is sufficient to protect comprehension from the intrusion of no-longer relevant information (O’Brien et al., 2010; Zwaan & Madden, 2004). This issue may not matter much for word-to-text integration in short texts (e.g., ~2 sentences). This is because the relevant memory traces have been established just prior to the word to be read (although across a sentence boundary) and there is not much contradictory discarded information to produce interference. However, with longer stretches of text to contribute more information in memory, interference may be more likely. This is an active area of research in our lab. There is an advantage of localizing a small part of the comprehension process for theoretical focus within the Reading Systems Framework: It allows a tractable number of comprehension processes to be considered. Here is the minimum set of the overlapping processes required for fluent word-to-text integration.

1. Rapid, automatic lexical access based on word form; 2. Rapid, automatic activation of associated knowledge from memory; 3. Access to memory for recently read text at the level of text model, situation model, or both; 4. Knowledge of context-relevant meaning associated with the lexical entry and its rapid retrieval; and 5. Word-to-text integration resulting from the overlapping of these processes. For an expert reader with knowledge of word meanings and sufficient experience, these are not effortful processes. Indeed, each overlapping phase of integration can be executed with minimal resource demands, approaching automaticity. These processes can be modeled through word activation networks with feedback from semantics and memories for recently read text segments. To perform robustly over text variations, models would need to include syntactic processes, which are usually ignored in text comprehension models. The point is that the basics of a testable theory that assumes individual differences in word-to-text integration processes arise from lexical knowledge. Alternative hypotheses can be tested, for example, that such comprehension skill differences arise from memory limitations or word identification processes that are resource demanding. Furthermore, one can examine the deeper question of whether sluggish word-to-text integration (Perfetti et al., 2008) propagates upward to the higher levels, and thus helps to explain retention and higher level as well as lower level comprehension problems.

26 Joseph Z. Stafura and Charles A. Perfetti

Conclusion Theories (or more accurately, frameworks) of reading comprehension have moved from ideas of broad scope to more specifically targeted aspects of the overall problem of comprehension. This has allowed progress in the study of components of comprehension, from the role of memory, the use of inferences, and the updating of mental models. We reintroduced a wide-angle framework that makes central the role of the lexicon, a somewhat neglected component in text comprehension research. The Reading Systems Framework represents the broad set of knowledge sources, and processes that act on these knowledge sources, allowing researchers to examine specific systems and subsystems and the interactions among them. Within this framework, we target a seemingly small yet central and recurring comprehension process, the integration of the currently read word into a mental structure that represents the current understanding of the text (the situation model). These word-to-text integration processes allow readers to continuously tune and update their current understanding. The paraphrase effect reflects a text integration process that is initiated by reading a word whose activated meanings include one that is congruent with the current model of the text, and thus can be integrated into that model. The lexical nature of this process distinguishes it from other integrating processes, such as bridging inferences, which also allow updating and keep the text coherent but at some cost to processing effort. Individual differences in reading comprehension are seen during these word-totext processes, specifically in the lexically driven paraphrase effect. This fact invites closer examination of whether subtle differences in knowledge of word meanings or the conditions of word use might affect word-to-text integration, as well as more global measures of comprehension. The general relationships between global comprehension skill and vocabulary and between comprehension skill and new word learning also suggest this possibility. Additionally, less skilled readers may require more bridging inferences compared to skilled readers more fluid, incremental word-to-text integration processes. These questions can be addressed through additional assessment and the development of new text and discourse stimuli. Finally, research should continue to probe the on-line nature of text and discourse processes, using methods that allow for word level measurement and analysis. Theoretically, our argument entails a closer view, within the Reading Systems Framework, of the interaction between the word identification system and the comprehension system that is mediated by lexical knowledge and manifest in word meaning processing.



Integrating word processing with text comprehension 27

Acknowledgments Research discussed in this article was supported by NICHD award R01HD058566–02 to the University of Pittsburgh, Charles Perfetti, PI. We graciously acknowledge the helpful reviewers of Perfetti and Stafura (2014), who greatly improved that manuscript through their critiques and comments, and, therefore, this chapter.

References Adlof, S. M., Catts, H. W., & Little, T. D. (2006). Should the simple view of reading include a fluency component? Reading and Writing, 19, 933–958.  doi: 10.1007/s11145-006-9024-z Adlof, S. M., & Perfetti, C. A. (2011, July). Are integration processes the same in reading and listening? Paper presented at the 18th annual meeting of the Society for the Scientific Study of Reading, St. Pete Beach, FL. Albrecht, J. E., & O’Brien, E. J. (1993). Updating a mental model: Maintaining both local and global coherence. Journal of Experimental Psychology: Learning, Memory, & Cognition, 19, 1061–1070.  doi: 10.1037/0278-7393.19.5.1061 Amsel, B. D., DeLong, K. A., & Kutas, M. (2015). Close, but no garlic: Perceptuomotor and event knowledge activation during language comprehension. Journal of Memory and Language, 82, 118–132.  doi: 10.1016/j.jml.2015.03.009 Anderson, R. C. (1978). Schema-directed processes in language comprehension. In A. M. Lesgold, J. W. Pellegrino, S. D. Fokkema, & R. Glaser (Eds.), Cognitive psychology and instruction (pp. 67–82). New York, NY: Plenum.  doi: 10.1007/978-1-4684-2535-2_8 Andrews, S., & Hersch, J. (2010). Lexical precision in skilled readers: Individual differences in masked neighbor priming. Journal of Experimental Psychology: General, 139, 299–318 doi: 10.1037/a0018366 Andrews, S., & Lo, S. (2012). Not all skilled readers have cracked the code: Individual differences in masked form priming. Journal of Experimental Psychology: Learning, Memory, and Cognition, 38, 152–163. Bartlett, F. C. (1932). Remembering: A study in experimental and social psychology. Cambridge, UK: Cambridge University Press. Boets, B., de Beeck, H. P. O., Vandermosten, M., Scott, S. K., Gillebert, C. R., Mantini, D., … & Ghesquière, P. (2013). Intact but less accessible phonetic representations in adults with dyslexia. Science, 342, 1251–1254.  doi: 10.1126/science.1244333 Bolger, D. J., Balass, M., Landen, E., & Perfetti, C. A. (2008). Context variation and definitions in learning the meanings of words: An instance-based learning approach. Discourse Processes, 45, 122–159.  doi: 10.1080/01638530701792826 Braze, D., Tabor, W., Shankweiler, D. P., & Mencl, W. E. (2007). Speaking up for vocabulary reading skill differences in young adults. Journal of Learning Disabilities, 40, 226–243. doi: 10.1177/00222194070400030401 Brown, C., & Hagoort, P. (1993). The processing nature of the N400: Evidence from masked priming. Journal of Cognitive Neuroscience, 5, 34–44.  doi: 10.1162/jocn.1993.5.1.34 Burkhardt, P. (2007). The P600 reflects cost of new information in discourse memory. Neuroreport, 18, 1851–1854.  doi: 10.1097/WNR.0b013e3282f1a999 Cain, K. (2010). Reading development and difficulties: An introduction. Oxford, UK: Wiley-Blackwell.

28

Joseph Z. Stafura and Charles A. Perfetti

Cain, K. E., Bryant, P. E., & Oakhill, J. (2004). Children’s reading comprehension ability: Concurrent prediction by working memory, verbal ability, and component skills. Journal of Educational Psychology, 96, 31–42.  doi: 10.1037/0022-0663.96.1.31 Cain, K., & Oakhill, J. (2006). Profiles of children with specific reading comprehension difficulties. British Journal of Educational Psychology, 76, 683–696.  doi: 10.1348/000709905X67610 Cain, K., Oakhill, J. V., Barnes, M. A., & Bryant, P. E. (2001). Comprehension skill, inference-making ability, and their relation to knowledge. Memory & Cognition, 29, 850–859. doi: 10.3758/BF03196414 Cain, K., Oakhill, J. V., & Bryant, P. E. (2000). Investigating the causes of reading comprehension failure: The comprehension-age match design. Reading and Writing, 12, 31–40.  doi: 10.1023/A:1008058319399 Cain, K., Oakhill, J., & Lemmon, K. (2004). Individual differences in the inference of word meanings from context: The influence of reading comprehension, vocabulary knowledge, and memory capacity. Journal of Educational Psychology, 96, 671–681.  doi: 10.1037/0022-0663.96.4.671 Coltheart, M., Rastle, K., Perry, C., Langdon, R., & Ziegler, J. (2001). DRC: a dual route cascaded model of visual word recognition and reading aloud. Psychological Review, 108, 204–256. doi: 10.1037/0033-295X.108.1.204 Cromley, J. G., & Azevedo, R. (2007). Testing and refining the direct and inferential mediation model of reading comprehension. Journal of Educational Psychology, 99, 311–325. doi: 10.1037/0022-0663.99.2.311 Cutting, L. E., Materek, A., Cole, C. A. S., Levine, T., & Mahone, E. M. (2009). Effects of fluency, oral language, and executive function on reading comprehension performance. Annals of Dyslexia, 59, 34–54.  doi: 10.1007/s11881-009-0022-0 Donchin, E., & Coles, M. G. (1998). Context updating and the P300. Behavioral and Brain Sciences, 21, 152–154.  doi: 10.1017/S0140525X98230950 Elbro, C. (1998). When reading is “readn” or somthn. Distinctiveness of phonological representations of lexical items in normal and disabled readers. Scandinavian Journal of Psychology, 39, 149–153.  doi: 10.1111/1467-9450.393070 Elbro, C., & Jensen, M. N. (2005). Quality of phonological representations, verbal learning, and phoneme awareness in dyslexic and normal readers. Scandinavian Journal of Psychology, 46, 375–384.  doi: 10.1111/j.1467-9450.2005.00468.x Ericsson, K. A., & Kintsch, W. (1995). Long-term working memory. Psychological Review, 102, 211–245.  doi: 10.1037/0033-295X.102.2.211 Gernsbacher, M. A. (1990). Language comprehension as structure building. Hillsdale, NJ: Erlbaum. Goldman, S. R., & Varma, S. (1995). CAPing the construction-integration model of discourse comprehension. In C. A. Weaver, S. Mannes, & C. R. Fletcher (Eds.), Discourse comprehension: Essays in honor of Walter Kintsch (pp. 337–358). Hillsdale, NJ: Erlbaum. Graesser, A. C., & Kreuz, R. J. (1993). A theory of inference generation during text comprehension. Discourse Processes, 16, 145–160.  doi: 10.1080/01638539309544833 Graesser, A. C., Singer, M., & Trabasso, T. (1994). Constructing inferences during narrative text comprehension. Psychological Review, 101, 371–395.  doi: 10.1037/0033-295X.101.3.371 Hagoort, P. (2005). On Broca, brain, and binding: A new framework. Trends in Cognitive Sciences, 9, 416–423.  doi: 10.1016/j.tics.2005.07.004 Hagoort, P. (2013). MUC (memory, unification, control) and beyond. Frontiers in Psychology, 4. Haviland, S. E., & Clark, H. H. (1974). What’s new? Acquiring new information as a process in comprehension. Journal of Verbal Learning and Verbal Behavior, 13, 512–521. doi: 10.1016/S0022-5371(74)80003-4



Integrating word processing with text comprehension 29

Helder, A., Stafura, J., Calloway, R., van den Broek, P., & Perfetti, C. (July 2015). Centrality effects on word to text integration during reading: an ERP study. Presented at the 25th Annual Meeting of the Society for Text & Discourse, Minneapolis, MN. Hersch, J., & Andrews, S. (2012). Lexical quality and reading skill: Bottom-up and top- down contributions to sentence processing. Scientific Studies of Reading, 16, 240–262. doi: 10.1080/10888438.2011.564244 Hoover, W. A., & Gough, P. B. (1990). The simple view of reading. Reading and Writing, 2, 127–160. doi: 10.1007/BF00401799 Just, M. A., & Carpenter, P. A. (1992). A capacity theory of comprehension: Individual differences in working memory. Psychological Review, 99, 122–149.  doi: 10.1037/0033-295X.99.1.122 Kintsch, W. (1988). The role of knowledge in discourse comprehension: a construction- integration model. Psychological Review, 95, 163–182.  doi: 10.1037/0033-295X.95.2.163 Kintsch, W., & Mross, E. F. (1985). Context effects in word identification. Journal of Memory and Language, 24, 336–349.  doi: 10.1016/0749-596X(85)90032-4 Kintsch, W., & Rawson, K. A. (2005). Comprehension. In M. J. Snowling & C. Hulme (Eds.), The science of reading: A handbook (pp. 209–226). Malden, MA: Blackwell. doi: 10.1002/9780470757642.ch12 Kutas, M., & Federmeier, K. D. (2000). Electrophysiology reveals semantic memory use in language comprehension. Trends in Cognitive Sciences, 4, 463–470.  doi: 10.1016/S1364-6613(00)01560-6 Kutas, M., & Hillyard, S. A. (1980). Reading senseless sentences: Brain potentials reflect semantic incongruity. Science, 207, 203–205.  doi: 10.1126/science.7350657 Language and Reading Research Consortium (2015). Learning to read: Should we keep things simple? Reading Research Quarterly, 50, 151–169. Lau, E., Almeida, D., Hines, P. C., & Poeppel, D. (2009). A lexical basis for N400 context effects: Evidence from MEG. Brain and Language, 111, 161–172.  doi: 10.1016/j.bandl.2009.08.007 McKoon, G., & Ratcliff, R. (1992). Inference during reading. Psychological Review, 99, 440–466. doi: 10.1037/0033-295X.99.3.440 McNamara, D. S., & Magliano, J. (2009). Toward a comprehensive model of comprehension. In B. Ross (Ed.), The psychology of learning and motivation (Vol. 51, pp. 297–384). New York, NY: Elsevier Science.  doi: 10.1016/S0079-7421(09)51009-2 Myers, J. L., & O’Brien, E. J. (1998). Accessing the discourse representation during reading. Discourse processes, 26, 131–157.  doi: 10.1080/01638539809545042 Nation, K. (2005). Connections between language and reading in children with poor reading comprehension. In H. W. Catts & A. G. Kamhi (Eds.), The connection between language and reading disabilities (pp. 41–54). Mahwah, NJ: Erlbaum. Oakhill, J. V., Cain, K., & Bryant, P. E. (2003). The dissociation of word reading and text comprehension: Evidence from component skills. Language and Cognitive Processes, 18, 443–468. doi: 10.1080/01690960344000008 O’Brien, E. J., & Albrecht, J. E. (1992). Comprehension strategies in the development of a mental model. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 777–784.  doi: 10.1037/0278-7393.18.4.777 O’Brien, E. J., Cook, A. E., & Guéraud, S. (2010). Accessibility of outdated information. Journal of Experimental Psychology: Learning, Memory, and Cognition, 36, 979–991. O’Brien, E. J., & Myers, J. L. (1985). When comprehension difficulty improves memory for text. Journal of Experimental Psychology: Learning, Memory, and Cognition, 11, 12–21. O’Brien, E. J., Plewes, P. S., & Albrecht, J. E. (1990). Antecedent retrieval processes. Journal of Experimental Psychology: Learning, Memory, and Cognition, 16, 241–249. doi: 10.1037/0278-7393.16.2.241

30

Joseph Z. Stafura and Charles A. Perfetti

O’Brien, E. J., Rizzella, M. L., Albrecht, J. E., & Halleran, J. G. (1998). Updating a situation model: A memory-based text processing view. Journal of Experimental Psychology: Learning, Memory, and Cognition, 24, 1200–1210. Perfetti, C. A. (1999). Comprehending written language: A blueprint of the reader. In C. Brown & P. Hagoort (Eds.), The neurocognition of language (pp. 167–208). New York, NY: Oxford University Press. Perfetti, C. A. (2007). Reading ability: Lexical quality to comprehension. Scientific Studies of Reading, 11, 357–383.  doi: 10.1080/10888430701530730 Perfetti, C. A., & Hart, L. (2002). The lexical quality hypothesis. In L. Verhoeven, C. Elbro, & P. Reitsma (Eds.), Precursors of functional literacy (pp. 67–86). Amsterdam, the Netherlands: John Benjamins.  doi: 10.1075/swll.11.14per Perfetti, C. A., Landi, N., & Oakhill, J. (2005). The acquisition of reading comprehension skill. In M. J. Snowling & C. Hulme (Eds.), The science of reading: A handbook (pp. 227–247). Oxford, UK: Blackwell.  doi: 10.1002/9780470757642.ch13 Perfetti, C., & Stafura, J. (2014). Word knowledge in a theory of reading comprehension. Scientific Studies of Reading, 18, 22–37.  doi: 10.1080/10888438.2013.827687 Perfetti, C., & Stafura, J. (2015). Comprehending implicit meanings in text without making inferences. In E. O’Brien, A. Cook, & R. Lorch. (Eds.), Inferences during reading (pp. 1–18). Cambridge, UK: Cambridge University Press.  doi: 10.1017/CBO9781107279186.002 Perfetti, C. A., Wlotko, E. W., & Hart, L. A. (2005). Word learning and individual differences in word learning reflected in event-related potentials. Journal of Experimental Psychology: Learning, Memory, and Cognition, 31, 1281–1292. Perfetti, C., Yang, C. L., & Schmalhofer, F. (2008). Comprehension skill and word-to- text integration processes. Applied Cognitive Psychology, 22, 303–318.  doi: 10.1002/acp.1419 Protopapas, A., Sideridis, G. D., Mouzaki, A., & Simos, P. G. (2007). Development of lexical mediation in the relation between reading comprehension and word reading skills in Greek. Scientific Studies of Reading, 11, 165–197.  doi: 10.1080/10888430701344322 Rapp, D. N., & Taylor, H. A. (2004). Interactive dimensions in the construction of mental representations for text. Journal of Experimental Psychology: Learning, Memory, and Cognition, 30, 988–1001. Shank, R. C., & Abelson, R. (1977). Scripts, plans, goals, and understanding. Hillsdale, NJ: Erlbaum. Ratcliff, R., & McKoon, G. (1988). A retrieval theory of priming in memory. Psychological Review, 95, 385–408.  doi: 10.1037/0033-295X.95.3.385 Rugg, M. D., & Curran, T. (2007). Event-related potentials and recognition memory. Trends in Cognitive Sciences, 11, 251–257.  doi: 10.1016/j.tics.2007.04.004 Rugg, M. D., Mark, R. E., Walla, P., Schloerscheidt, A. M., Birch, C. S., & Allan, K. (1998). Dissociation of the neural correlates of implicit and explicit memory. Nature, 392, 595–598. doi: 10.1038/33396 Singer, M., & Halldorson, M. (1996). Constructing and validating motive bridging inferences. Cognitive Psychology, 30, 1–38.  doi: 10.1006/cogp.1996.0001 Stafura, J. Z., & Perfetti, C. A. (2014). Word-to-text integration: Message level and lexical level influences in ERPs. Neuropsychologia, 64, 41–53.  doi: 10.1016/j.neuropsychologia.2014.09.012 Stafura, J. Z., Rickles, B., & Perfetti, C. A. (2015). ERP evidence for memory and predictive mechanisms in word-to-text integration.. Language, Cognition, and Neuroscience, 30, 1273–1290. doi: 10.1080/23273798.2015.1062119



Integrating word processing with text comprehension 31

Tunmer, W. E., & Chapman, J. W. (2012). The simple view of reading redux vocabulary knowledge and the independent components hypothesis. Journal of Learning Disabilities, 45, 453–466. doi: 10.1177/0022219411432685 Van Daalen-Kapteijns, M. M., & Elshout-Mohr, M. (1981). The acquisition of word meanings as a cognitive learning process. Journal of Verbal Learning and Verbal Behavior, 20, 386–399. doi: 10.1016/S0022-5371(81)90515-6 Van den Broek, P. W., Helder, A., & Van Leijenhorst, L. (2013). Sensitivity to Structural Centrality: Developmental and individual differences in reading comprehension skills. In: M. A. Britt, S. R. Goldman & J-F Rouet (Eds.), Reading: From words to multiple texts (pp. 132–146). New York: Routledge, Taylor & Francis Group. Van den Broek, P., Risden, K., Fletcher, C. R., Thurlow, R., Britton, B., & Graesser, A. (1996). A “landscape” view of reading: Fluctuating patterns of activation and the construction of a stable memory representation. In B. K. Britton & A. C. Graesser (Eds.), Models of understanding text (pp. 165–187). Mahwah, NJ: Erlbaum. Van Dijk, T. A., & Kintsch, W. (1983). Strategies of discourse comprehension. New York, NY: Academic Press. Van Petten, C., & Luka, B. J. (2012). Prediction during language comprehension: Benefits, costs, and ERP components. International Journal of Psychophysiology, 83, 176–190. doi: 10.1016/j.ijpsycho.2011.09.015 Vellutino, F. R., Fletcher, J. M., Snowling, M. J., & Scanlon, D. M. (2004). Special reading disability (dyslexia): What have we learned in the past four decades. Journal of Child Psychology and Psychiatry, 45, 2–40.  doi: 10.1046/j.0021-9630.2003.00305.x Verhoeven, L., & Van Leeuwe, J. (2008). Prediction of the development of reading comprehension: A longitudinal study. Applied Cognitive Psychology, 22, 407–423.  doi: 10.1002/acp.1414 Yang, C. L., Perfetti, C. A., & Schmalhofer, F. (2005). Less skilled comprehenders ERPs show sluggish word-to-text integration processes. Written Language & Literacy, 8, 157–181. doi: 10.1075/wll.8.2.10yan Yang, C. L., Perfetti, C. A., & Schmalhofer, F. (2007). Event-related potential indicators of text integration across sentence boundaries. Journal of Experimental Psychology: Learning, Memory, and Cognition, 33, 55–89. Zwaan, R. A. (2003). The immersed experiencer: Toward an embodied theory of language comprehension. Psychology of Learning and Motivation, 44, 35–62. doi: 10.1016/S0079-7421(03)44002-4 Zwaan, R. A., Langston, M. C., & Graesser, A. C. (1995). The construction of situation models in narrative comprehension: An event-indexing model. Psychological Science, 6, 292–297. doi: 10.1111/j.1467-9280.1995.tb00513.x Zwaan, R. A., & Madden, C. J. (2004). Updating situation models. Journal of Experimental Psychology: Learning, Memory, and Cognition, 30, 283–288. Zwaan, R. A., & Radvansky, G. A. (1998). Situation models in language comprehension and memory. Psychological Bulletin, 123, 162–185.  doi: 10.1037/0033-2909.123.2.162

Genetic and environmental influences on the development of reading and related skills Richard K. Olson1, 2, Janice M. Keenan3, Brian Byrne2, 4 and Stefan Samuelsson2

1University

3University

of Colorado Boulder / 2Linköping University / of Denver / 4University of New England

Modern behavior-genetic studies of twins in the U.S., Australia, Scandinavia, and the U.K. show that genes account for most of the variance in children’s reading ability by the end of the first year of formal reading instruction. Strong genetic influences continue across the grades; however, the relevant genes vary for reading words and for comprehending text. Strong genetic influences do not diminish the importance of the environment for reading development; in fact, some of the genetic influence comes through a gene – environment correlation. They do, however, question setting the same minimal performance criterion for all children. Why do children differ in their development of reading and related skills? A casual perusal of reading research over the past 20 years reveals answers that are predominantly environmental, including preschool language and print exposure, quality and quantity of reading instruction in school, peer and family influences, socioeconomic level, and learning to read in a second language. This environmental focus is understandable from the obvious fact that reading is a learned skill that initially depends on formal instruction. The environmental focus is also supported by many experiments showing significant effects from manipulating the reading environment, and by studies reporting correlations between reading and related skills, such as phonological awareness and phonological decoding. Indeed, these studies are often interpreted as implying an environmentally-based causal pathway from phonological skills to the individual differences and deficits observed in reading. A different perspective on the etiology of individual differences in reading and related skills has been provided by behavior-genetic studies that compare similarities between large samples of identical (monozygotic or MZ) and fraternal (dizygotic or DZ) twin pairs who share their home and school environment, yet differ in their additive genetic similarity (100% for MZ pairs, 50% of segregating genes on average for DZ pairs). This natural experiment is unique in its ability to estimate the average influence on individual differences in reading and related skills that arise from genes, from shared environments that make twins in a pair similar (e.g., books in the home,

doi 10.1075/swll.15.03ols © 2017 John Benjamins Publishing Company

34

Richard K. Olson et al.

general family support for reading, shared teachers, classrooms, peers), and from non-shared environments that make twins different (e.g., different peers, teachers, classrooms, illnesses, birth problems, accidents, and measurement error) (Plomin, DeFries, McClearn, & McGuffin, 2008). In addition to assessing genetic and environmental influences on individual variation in specific reading and related skills, behavior-genetic research can also address genetic and environmental influences on the correlations between different skills, such as between word recognition, listening comprehension, and reading comprehension (cf., Keenan, Betjemann, Wadsworth, DeFries, & Olson, 2006). This knowledge can provide a deeper understanding of why children differ, not only for genetic reasons, but also for reasons related to environmental influences. Before we explore the results from recent behavior-genetic studies of reading, we need to clarify some important general qualifications and limitations of behavior-genetic research using data from identical and fraternal twins reared together. First, we sometimes hear concerns that our behavior-genetic results might imply a genetic basis for mean differences between racial, ethnic, or regional groups’ reading ability. It is important to emphasize that behavior-genetic studies cannot address the etiology of group mean differences. They are assessments of the sources of variance between individuals in a sample, not mean performance of that sample. Thus, environmental differences may completely account for any mean group differences, even when genetic influences explain individual differences within those groups. Second, it is important to understand that estimates of genetic and environmental influence from twin studies are specific to the behaviorally relevant environmental range in the sampled populations. Twin samples with low environmental variation are likely to yield higher genetic and lower environmental estimates when compared to samples with high environmental variation. Third, behavior-genetic studies only describe the average current balance of genetic and environmental influences on individual differences within a sample. They do not specify the genetic and environmental etiology for any individual in that sample. Fourth, estimates of genetic and environmental influences from behavior-genetic studies do not speak to the possibility for changing the average reading level in a population by manipulating the quantity and quality of reading instruction or practice. Nor do they speak to the potential benefits of extraordinary environmental interventions for reading disabilities. Fifth, genes express themselves through the environment. For example, we will present evidence that genes related to reading ability influence the amount of reading practice, resulting in a positive gene – environment correlation. This correlation likely accounts for at least part of our estimates of genetic influence from twin studies. It has important implications for environmental intervention that we will consider in the concluding section. Regardless of these limitations and qualifications, we will argue that considered together, the results of modern behavior-genetic twin studies of individual differences and deficits in reading provide the best evidence for why, on average, children differ in



Why children differ in reading ability 35

their reading and related skills. Following a brief overview of the major behavior-genetic studies of reading conducted over the past 20 years, we will turn to what we think are some of the most important results from those studies. In the final section of our review, we will consider the implications of the behavior-genetic results for education and directions for future research.

Overview of modern behavior-genetic studies of reading In 1992, the National Institutes of Health (NIH) began providing substantial funding for twin research focused on the genetic and environmental etiology of learning difficulties in reading (e.g., “dyslexia”), and more recently on ADHD and math deficits, through the Colorado Learning Disabilities Research Center (CLDRC) (DeFries et al., 1997; Olson, 2004; 2006). Beginning around 2000, NIH also began funding longitudinal twin studies of individual differences in pre-reading and early-reading development that are being conducted in Colorado, Florida, and Ohio. These longitudinal studies in the U.S. are complemented by others being conducted in Australia, Scandinavia, and the U.K. Together, these longitudinal studies provide important cross-language and cross-cultural perspectives on the etiology of individual differences in reading development. Our goal in this chapter is to highlight the most important themes from modern behavior-genetic research on reading and discuss their broader implications. We will concentrate on longitudinal studies of individual differences in unselected population samples, but we will first discuss a few key findings from research on twin pairs wherein at least one member was selected for reading disability.

The genetic and environmental etiology of reading disability Prior to 1985, the limited behavior-genetic research on reading disability used a categorical definition. Evidence for genetic influence was based simply on a comparison of diagnostic concordance rates for MZ and DZ twin pairs (both “dyslexic” or just one member “dyslexic”). Subsequently, DeFries and Fulker (1985) recognized that the continuous normal distribution of reading ability in the population could be used to support a continuous regression method for assessing the average genetic and environmental etiology of twins’ reading disability, based on the similarity of the MZ and DZ cotwins’ regression to the population mean. This “DF” model yielded more statistically powerful and accurate estimates of genetic and environmental influence on group deficits in reading and related skills. When the DF model has been applied to study the etiology of reading disability group membership in the CLDRC and an independent study in the U.K., we find broadly similar results. In describing the results, we will follow behavior-genetic convention and label additive genetic influence as A, shared environment influence as C,

36

Richard K. Olson et al.

and non-shared environment influence as E. Together these three influences account for 100% of the total phenotypic variance, so we can express the A, C, and E influences as accounting for percentages of the phenotypic variance. Here and in the rest of the paper, we only report A (genetic) and C (shared environment) percentages, because E (non-shared environment including measurement error) simply equals 100% – (A% + C%). So for example, if A has been estimated at 40% and C has been estimated at 50%, then it can be assumed that E has been estimated at 10%. Friend, DeFries, and Olson (2008) applied the DF model to a composite of reading and spelling data from 545 same-sex twin pairs at mean age 11.4 years in which at least one member of each pair scored below the 10th percentile. The average influence on reading-disability group membership in this CLDRC sample was A = 61% for genetic and C = 30% for shared environment. Harlaar, Spinath, Dale, and Plomin (2005) found very similar results in their large and representative population sample of seven-year-old twins tested on a composite of word and nonword reading near the end of first grade in the U.K. (A = 59%, C = 30%). In both the Friend et al. and Harlaar et al. studies, genetic, shared environment, and non-shared environment influences were all statistically significant for reading disability, but the average influence of genes was about twice as strong as the shared environment influence. Of course we would like to know the specific genetic and environmental etiology for individual children, but behavior-genetic data cannot provide this answer; it only provides information about the average etiology of reading-disability group membership. It is therefore possible that for some individual children within the group with reading disability, environmental factors may have been the major or only influence, while for others, genes may have been the major or only influence. However, it is possible to expand the DF model to ask if the degree of genetic and environmental influences on reading disability is significantly related to individual differences on other variables, thus bringing us closer to an understanding of differential genetic and environmental etiology within the reading disabled group. Friend et al. (2008) did this using parents’ years of education to explore the possibility of a “genetic-influence by environment” interaction. They found that genetic influence was significantly higher on average for children with reading disability who had parents with higher education, compared to children with lower parent education. For children with lower parent education, shared family environment and genes were, on average, about equally influential. One interpretation of these results is that children who fail in reading in spite of having highly educated parents (and likely a better environment for learning to read) are more likely to have genetic than environmental constraints on their reading development. The interaction between parent education and genetic influence on reading has also been explored for high-reading group performance using the same DF model (Friend et al., 2009). Twin pairs with at least one member performing at least one standard deviation above the population mean on the TOWRE word and nonword reading efficiency composite (Torgesen, Wagner, & Rashotte, 1999) were selected from representative population twin samples in Colorado and the U.K. Genetic influence



Why children differ in reading ability 37

on high reading group membership was substantial (50%–72%), depending on grade level and country, and the level of genetic influence interacted with parents’ education. But interestingly, the pattern for high-level readers was quite different than for low-level readers. For low-level readers, genetic influence was higher with higher parent education; whereas for high-level readers, genetic influence was greater for children with low parent education. In other words, children who read well, in spite of environmental disadvantages that are often associated with low parent education, are more likely to have higher genetic influence on their high performance. Thus, the Friend et al. (2008; 2009) studies show that the balance of genetic and environmental influences on extreme group membership varies depending on the environmental circumstances for children within those groups.

The genetic and environmental etiology of individual differences in reading While it is important to understand the etiology of reading disabilities because adequate reading is so important for broader educational and professional development, it is also important to understand the etiology of the full normally-distributed variation in reading and related skills in the population. In this section we first provide an overview of some recent results from the CLDRC that have focused on the etiology of individual differences in reading comprehension, language comprehension, word recognition, and their correlation among children between age 8 to 18 years. Then we will turn to results from recent longitudinal twin studies of individual differences in pre-reading and early reading development.

Do the same genetic and environmental factors influence word recognition, listening comprehension, and reading comprehension? Twin research in the CLDRC initially focused primarily on the etiology of deficits and individual differences in word reading, spelling, and related skills such as phonological awareness, phonological decoding, and orthographic coding (DeFries, Fulker & Labuda, 1987; Gayán & Olson, 2001; 2003; Olson, Wise, Conners, Rack, & Fulker, 1989). Beginning in 2000, a component of the CLDRC directed by Jan Keenan at the University of Denver introduced new measures of reading and listening comprehension so we could better understand how genetic and environmental influences on basic word reading skills and language skills influenced comprehension of text, the ultimate goal of reading. Keenan et al. (2006) explored the genetic and environmental basis for the “simple view” of reading comprehension proposed by Hoover and Gough (1990), wherein individual differences in reading comprehension can be accounted for by listening comprehension and word recognition or decoding. Keenan et al. found that the genetic correlation (rg) between word recognition and listening comprehension was

38

Richard K. Olson et al.

modest (rg = .37), and that in accord with the simple view, these two skills accounted for all of the significant genetic influences on reading comprehension. Subsequent analyses using an expanded sample from the CLDRC (Betjemann, Keenan, Olson, & DeFries, 2011), and a study by Harlaar et al. (2010) using twin data from Ohio have confirmed these findings. In contrast, shared environmental influences were highly correlated across word recognition, listening comprehension, and reading comprehension in these studies. Thus, it is the largely independent genetic influences on listening comprehension and word recognition that account for their unique contributions to individual differences in reading comprehension. The research on reading comprehension in the CLDRC has raised questions about how different methods of assessing reading comprehension vary in their relations to word recognition and listening comprehension. In essence, the genetic correlations of reading comprehension with word reading and with listening comprehension differ dramatically across reading comprehension tests (Betjemann et al., 2011; Keenan, Betjemann & Olson, 2008). Thus, different tests used to measure the same construct may manifest very different patterns of genetic covariation. Converging evidence on the differential etiology of printed word recognition and oral language has come from our International Longitudinal Twin Study (ILTS). Olson et al. (2011) found that, oral language comprehension (a vocabulary latent trait) and word decoding latent traits had significant independent genetic influences, and together they accounted for all of the high genetic influence (A = 86%, C = 9%) on a reading comprehension latent trait at the end of 4th grade. Byrne et al. (2013) conducted a genetic factor analysis across a wide range of preschool and grade 2 measures in the ILTS that included five on-line learning tasks and measures of vocabulary, letter identification, and word and nonword reading. Three correlated genetic factors emerged, the first factor for vocabulary, the second factor for grade 2 word and nonword reading, grade 2 orthographic learning, and preschool letter knowledge, and the third factor for tests of verbal short-term memory. The second print-related factor showed the most genetic specificity. The results support the importance and distinctive genetic etiology of learning print-speech integration that was partly independent from genetic factors affecting spoken language and verbal short-term memory. Taken together, results from the CLDRC and the ILTS studies have shown low to moderate genetic correlations between word decoding and oral language. They highlight the importance of partly independent genetic influences on paired associate learning between print and speech for the development of word recognition. But the partly independent genetic influences on oral language so critical for reading comprehension also deserve attention in both research and education. They suggest that interventions need to focus not just on print-speech associations but also on higher language skills, perhaps differentially depending on children’s skill profiles.



Why children differ in reading ability 39

Genetic and environmental etiology of longitudinal stability in CLDRC twins The CLDRC twins have been recruited across a broad age range between 8 and 18 years. Most previous examinations of developmental differences in this sample have been cross-sectional (cf., Keenan et al., 2008). However, a sub-sample of CLDRC twins initially tested at a mean age of about 10 years have been retested at a mean age of about 16 years (Betjemann et al., 2008; Hulslander, Olson, Willcutt, & Wadsworth, 2010). Betjemann et al. reported high stability of reading performance and high genetic correlations over this interval for individual word-reading and comprehension measures. Hulslander et al. subsequently found nearly perfect phenotypic stability for individual differences when reading and related skills were modeled as latent traits to remove contamination from measurement error. The latent-trait longitudinal correlations were r = .98 for word recognition, r = 1.0 for phonological awareness, r = .93 for phonological decoding, and r = .95 for spelling. It seems that the vast majority of children establish a very stable developmental trajectory for growth in reading and related skills by age 10 years. To better understand the origin of individual differences in developmental trajectories, we now turn to the development of reading and related skills from preschool at mean age 4 years 10 months through the end of fourth grade at mean age 10 years.

The International Longitudinal Twin Study (ILTS) Much of the twin research that we have reported has come from unselected population samples of twins tested on the same measures in Australia, Colorado, and Scandinavia (Norway and Sweden combined). The twins were initially tested on pre-reading skills in their homes or preschools during the year prior to kindergarten entry (Byrne et al., 2002; Samuelsson et al., 2005). They were subsequently tested on reading and related skills at the end of kindergarten, first grade, and second grade in all three countries, and also at the end of fourth grade in Colorado (Olson et al., 2011; Christopher et al., 2013). At preschool in all three ILTS samples, most individual differences on a print knowledge latent trait, primarily based on letter name and sound knowledge, were due to differences in shared family environment (A = 20%–26%; C = 62%–74%) (Samuelsson et al., 2007). The vast majority of preschool children could not read, so we could not estimate genetic and environmental influences on their reading ability. By the end of kindergarten, most children could read enough words and nonwords on the TOWRE so we could estimate genetic and environmental influences on their individual differences (Samuelsson et al., 2008). Those individual differences were mostly due to genes in Australia (A = 84%; C = 9%) and Colorado (A = 68%; C = 25%). In contrast, individual differences for the Scandinavian twins’ reading at the end of kindergarten were mostly due to shared environment (A = 33%; C = 52%). Samuelsson et al. noted that reading is not formally taught in Scandinavia until the first grade, so the

40 Richard K. Olson et al.

Scandinavian twins’ reading scores were significantly lower than for the Australian and Colorado twins. Thus, it was variation in the twins’ shared home, preschool, and kindergarten environment that was the major influence on individual differences at the end of kindergarten in Scandinavia. However, after all children had received at least a year of formal reading instruction at the end of first grade, genetic influence was about as strong in Scandinavia (A = 79%; C = 7%) as it was at the end of first grade in Australia (A = 80%; C = 2%) and Colorado (A = 83%; C = 7%). Similarly high genetic and low shared environment estimates have been found for spelling and reading comprehension at the end of first grade, and the genetic correlations between word recognition, spelling, and reading comprehension were all above rg = .9 (Byrne et al., 2007). The pattern of high genetic and low shared environment estimates continued to the end of second grade in all three samples (Byrne et al., 2009), and to the end of fourth grade in Colorado for word and nonword reading, reading comprehension, and spelling (Christopher et al., 2013; Olson et al., 2011). The bottom line is that after a year of formal reading instruction, individual differences in word reading, spelling, and reading comprehension are highly influenced by genes in the independent twin samples from the Sydney area of Australia, from the Denver area of Colorado, and from southern Norway and Sweden. Of course environmental influences do have a big effect – they affect the level of reading in the population because we learn to read in classrooms and homes. However, on average, the variation across the twins’ shared home and classroom environment has relatively little influence on individual differences in reading once children have completed a year of formal reading instruction. When compared to the strong shared environment influences on preschool print knowledge, it seems that what formal reading instruction in schools does is to considerably reduce the environmental variance for reading development in all of the ILTS sampling areas. This is a very good result. It would be unfortunate if strong family-based environmental influences persisted beyond the early stages of school. Indeed, this is partly what schools are about, overcoming factors that produce big differences among kids before they go to school, particularly when those environmental influences are negative. A major goal of the ILTS has been to better understand the relations between pre-reading skills and subsequent reading and spelling skills through the early grades. To date we have addressed two basic questions about these relations: (1) the phenotypic relations between poor-reader subtypes and pre-reading skills, and (2) the relative influences of genes and environment on the phenotypic correlations between pre-reading and subsequent reading skills. Elwér, Keenan, Olson, Byrne, and Samuelsson (2013) examined the longitudinal phenotypic relations from pre-reading to reading in two subgroups of poor readers selected at the end of grade 4. One group was selected for poor word decoding with adequate oral language comprehension. The other group had poor oral language comprehension with adequate word decoding. When the groups were compared on reading-related skills at preschool and the end of kindergarten, grade 1, and grade 2, the poor decoder group was consistently lower in rapid naming, while the poor oral



Why children differ in reading ability 41

comprehender group was consistently lower in vocabulary. A second study compared good and poor reading comprehension groups that were matched on word decoding at the end of grade 4 (Elwér, Gustafson, Byrne, Olson, Keenan, & Samuelsson, 2015). Again poor reading comprehenders were shown to have had significantly lower preschool vocabulary, grammar, and verbal memory. Thus, the seeds for these poor-reader subtypes at the end of grade 4 were already sown in their profile of pre-reading skills. Christopher et al. (2015) examined the genetic and environmental etiology of these phenotypic correlations between pre-reading and subsequent reading. They found that for word reading, spelling and early-grade reading comprehension, the etiology was largely due to genetic influences. This was true even when individual differences at preschool were primarily due to shared family environment (print knowledge and vocabulary). However, for 4th grade reading comprehension, the phenotypic correlations with preschool print knowledge (r = .50), phonological awareness (r = .61), and vocabulary (r = .57) were primarily mediated by the twins shared environment. Why are shared family environment influences on three of the preschool skills more closely associated with reading comprehension than with word recognition or spelling, and why is their shared environment influence more strongly related to reading comprehension at the end of grade 4 than at the end of the first grade? The answer may be that individual differences in reading comprehension are more influenced by individual differences in word decoding skills at first grade when many children are still learning to read (Elwér et al., 2013). When children are primarily reading to learn at the end of grade 4, most have sufficient word decoding skills to accurately read the text. Therefore, individual differences in reading comprehension become relatively more dependent on vocabulary and world knowledge. Of course this begs the question of why word recognition and spelling are less related to these shared environmental influences. We suggest that individual differences in word decoding and spelling are more closely tied to basic genetically influenced differences in learning rates for printto-sound (decoding) and sound-to-print (spelling) relations (Byrne et al., 2013). It will be interesting to see if the greater shared environment influences linking pre-reading skills to reading comprehension at the end of grade 4 are maintained when the twins complete their follow-up assessments at the end of grade 9.

Results from other longitudinal twin studies It is important to ask if other twin studies find similar results to the ILTS. Although these studies have not addressed relations between pre-reading skills and later reading, they have tracked twins’ reading development across the early grades. The Twins Early Development Study (TEDS) is being conducted with a very large population sample in England and Wales. Harlaar et al. (2005) administered the TOWRE word and nonword tests over the phone to 3909 twin pairs at mean age 7.07 years near the end of first grade. Based on a composite measure including both TOWRE word and nonword reading, they reported estimates of genetic (A = 66%) and shared environment (C = 18%)

42 Richard K. Olson et al.

averaged across gender, not much different from the ILTS estimates at the end of first grade near the same mean age, and well within our 95% confidence intervals. However, it is possible that the slightly lower genetic and higher shared-environment estimates in the TEDS sample compared to the three ILTS samples reflects a greater environmental range across its national sample in the U.K. It is also possible that their administration of the TOWRE by telephone was more influenced by the home environment than the TOWRE administered directly by testers in the ILTS samples. A recent large twin study in Florida analyzed scores for 2570 twin pairs on the school-administered one-minute Oral Reading Fluency (ORF) test (Taylor & Schatschneider, 2010). They reported genetic (A = 62%) and shared environment (C = 22%) estimates from near the end of first grade at mean age 6.6 years. Their sample was more socioeconomically diverse than the ILTS samples, and they offered that as a potential explanation for their lower genetic and higher shared environment estimates compared to the ILTS results. Also, when they separated their sample into the lower 25%, the middle 26%-74%, and the highest 25% for median family income based on geographic area, they found that the lowest income group had the lowest genetic (A = 45%) and highest shared environment influences (C = 37%) of the three income groups. This result may be similar to what Friend et al. (2008) found for the relations between parents’ years of education and levels of genetic and environmental influences on reading disability. A third twin study is being conducted in Ohio by Stephen Petrill and colleagues. They initially measured several reading and related skills in their twin sample at mean age 6 years (similar to the Colorado ILTS mean age at the end of kindergarten), and they retested the twins each year after that out to mean age 12 years (Logan et al., 2013). In one of their earlier papers, Petrill et al. (2007) examined the etiology and longitudinal stability of early reading, assessed at their first two waves of data. Looking at their first wave (when twins were in kindergarten or first grade), estimates for letter knowledge showed significant genetic and shared environmental influences (A = 35%, C = 38%), word recognition showed genetic and shared environmental influences (A = 55%, C = 34%), pseudoword decoding demonstrated significant genetic and shared environmental influences (A = 56%, C = 26%) and reading comprehension showed genetic and shared environmental effects (A = 50%, C = 21%). Results in the second wave were highly similar, with the exception of passage comprehension, which showed high genetic and nonsignificant shared environmental influences (A = 76%, C = 11%). Subsequent papers reporting estimates from later assessment waves suggest that genetic influences on reading outcomes remain statistically significant whereas shared environmental influences become small or nonsignificant (see Harlaar et al., 2010). In summary, the TEDS and Florida twin studies report results that are basically consistent with those from the three independent ILTS twin samples: genetic influences are substantially greater than shared environment influences when word reading is tested at or near the end of the first year of formal reading instruction, and this pattern continues across the early grades. The attenuated genetic and stronger shared environmental influences found in the early assessment waves in the Ohio twin study



Why children differ in reading ability 43

may have been due to its very broad within-wave range in months of education (beginning of kindergarten through the end of first grade) at their first assessment wave.

Biometric growth models of early reading development Biometric growth models of early reading development were first introduced by Petrill et al. (2010). These models are able to distinguish the average genetic and environmental etiologies for where children start in reading development, defined as the intercept, and for their subsequent growth patterns across the grades. To date, these models have been applied to longitudinal twin data from Australia, Colorado, and Scandinavia (Christopher et al., 2013), Florida (Hart et al., 2013), the U.K. (Harlaar, Dale, HayiouThomas, & Plomin, 2012), and Ohio (Logan et al., 2013; Petrill et al., 2010). The studies vary in exactly how the initial intercept is defined (end of kindergarten, beginning of first grade, or end of first grade), measures employed (word recognition, oral reading fluency, reading comprehension, spelling), and modeling assumptions (linear vs. non-linear, correlation of errors). The details of these recent biometric growth model studies are beyond the scope of this paper. In summary, most of the biometric growth model studies to date showed high A and low C at the first wave intercept as well as high A and low C for univariate estimates across each of the waves. Results from the Ohio twin sample showing high C and low A were an exception to this pattern. We surmise that this was due to the very wide range of months of education that are shared by twins in a pair (beginning kindergarten through end of first grade) at their first wave intercept. The other studies with the opposite high A and low C results at intercept had a much more narrow range for months of schooling. The one exception was for Scandinavia, with approximately equal influences from A and C at the end of kindergarten (Christopher et al., 2013), a result that we previously noted was due to the lack of reading instruction in Scandinavian kindergartens (Samuelsson et al., 2008). The results for linear and quadratic growth were more mixed. The A and C influences on growth seem to depend partly on measures, samples, fitting linear versus nonlinear models, and models allowing or not allowing unique variances (twin similarities not related to growth) to correlate (see Christopher et al., 2013, for a comparison of models with correlated or uncorrelated errors). We concur with Christopher et al. that allowing unique variances to correlate is the most appropriate model, though this is still a debated question. Either way, variance in growth is generally much lower than variance in the intercept, and growth variance has relatively little relation to univariate estimates of genetic and environmental influence across the grades for twin samples from Australia, Colorado, Florida, Scandinavia, and the UK.

44 Richard K. Olson et al.

Summary In summary, the evidence suggests that the answer to the question, “Why do children differ in their development of reading and related skills?” is on average, after the first year of formal literacy instruction, mostly the genetic differences between children. The mostly genetic conclusion is consistently supported from samples of identical and fraternal twins tested in Australia, Colorado, Florida, Scandinavia, and the UK near the end of first grade, and in Ohio by mean age 9 years. It is also supported for older twins tested in the CLDRC and in the U.K. TEDS study. The genetic and shared environment influences on growth are less consistent across studies and measures, but regardless of whether shared environment influences on growth are relatively high or low, variance in growth is low, and it has little influence on univariate A and C estimates at any given age or grade. We also found that the relevant genes depend at least partly on the specific reading and related skills being assessed (e.g., word decoding, listening comprehension, reading comprehension). But there are important qualifications to the “mostly genetic” conclusion. It is to these that we now turn.

Qualifications and clarifications Assumptions of additive genetic influence and no assortative mating Here we add a sixth qualification to the five mentioned in the introduction. The models used in all of the reviewed twin studies assume additive genetic influence (no dominance or epistasis) and no assortative mating indicated by correlations between parents. Violations of these two assumptions tend to bias genetic and environmental estimates in opposite directions. Genetic effects would tend to be underestimated if there is assortative mating, because assortative mating results in DZ twins sharing more than 50% of their segregating genes. Genetic effects would be overestimated if there is dominance or epistasis, because this reduces the DZ twins’ genetic similarity below the additive model’s assumption of 50% (Carey, 2003; Keller & Coventry, 2005; Keller, Medland, & Duncan, 2010). However, converging evidence from the Colorado Adoption Project comparing genetically unrelated siblings shows very low shared environment influences on reading (Wadsworth, Corley, Hewitt, Plomin, & DeFries, 2006), consistent with the low shared environment estimates from the reviewed twin studies. This suggests that any potential bias in the genetic and environmental estimates from the reviewed twin studies’ assumptions of additivity and no assortative mating would not be sufficient to challenge the basic finding that individual differences and deficits in reading are primarily due to genetic factors in the sampled populations.



Why children differ in reading ability 45

Dependence of genetic and environmental estimates on the environmental range In the introduction we emphasized that the average balance of genetic and environmental influences across a twin sample depends on the relevant environmental range in that sample. Unfortunately, it is difficult to quantify the relevant environmental range for reading in different published twin studies beyond a comparison of their estimates of environmental influences from the behavior-genetic analyses. But in principle, twin samples with greater reading-relevant environmental variance are likely to show lower average estimates for genetic influences and higher average estimates for environmental influences on individual differences or deficits in reading. Also, it is important to keep in mind that even when behavior-genetic estimates of the average environmental influence within a sample are low, there can be extreme cases of poor reading within the sample that are entirely due to environmental influences, such as a particularly poor home, peer, or classroom environment for reading.

Importance of the environment is not inconsistent with high genetic influence Our acknowledgement of the importance of environmental range and the possibility of strong environmental influences on individual cases within twin samples does not diminish our conclusion that genes are the main average influence on individual differences in children’s reading ability. Yet the reading environment is obviously important for reading development. For example, if the emphasis on reading instruction in the early grades were to double in time and intensity (cf., Sadoski & Wilson, 2006), the resulting increase in average reading ability would be entirely due to this environmental change. However, the etiology of individual differences after this environmental change, i.e. what accounts for the differences between individuals all exposed to increased reading instruction, would continue to be mostly genetic influences.

The evidence for classroom effects Those who believe that the environment is the main influence on individual differences in reading sometimes assert that they are due to differences in teacher quality. This view is constantly reinforced by the U.S. media and politicians. They blame teachers for children’s reading difficulties, although interestingly, they do not readily give teachers credit for the high achievers in their classes; nor is there recognition of environmental and genetic influences on students that are outside teachers’ control. So let’s look at the evidence for this predominant environmental assumption. If there is direct evidence for very strong environmental influences from classrooms (including teachers) on individual-differences variance in early reading, that would contradict the low environmental estimates from behavior-genetic twin studies.

46 Richard K. Olson et al.

Byrne et al. (2010) estimated classroom effects on Australian and Colorado ILTS twins’ reading and spelling performance in kindergarten, grade 1, and grade 2. Based on the difference in correlations for the same- vs. different-classroom twin pairs, classroom effects were estimated at 8% across the same Australian and Colorado ILTS samples that were used in the behavior-genetic analyses reported in the individual differences section of the present paper. When the twins’ performance was controlled by their performance in the previous year, the “value added” classroom effects averaged only 4%. Kovas, Haworth, Dale, and Plomin (2007) estimated classroom effects by comparing their shared environment estimates on TOWRE word and noword reading at the end of first grade between twins in the same class room (C = .17) versus twins in different classrooms (C = .07). While this difference was not statistically significant in their large sample, it suggests a small shared versus non-shared classroom environment effect on individual differences of around 10%. Hart, Taylor, and Schatschneider (2013) tested the difference in oral reading fluency (ORF) at the end of second and third grade for twins that were in first-grade classrooms with growth below the mean versus classrooms with growth above the mean in ORF across the year. The effect sizes for twin pairs discordant for first grade classroom ORF growth were statistically significant but small, and they concluded that the effect of teacher quality on student reading outcomes was small. It could be argued that since twins share their home environment, the effects of having a different classroom might be diminished by their collaboration in reading activities at home. Fortunately, there is one large experimental study with random assignment of teachers to classrooms and students to teachers conducted in Tennessee by Nye, Konstantopoulos, and Hedges (2004). The Nye et al. study found that classroom effects (these include teachers as well as other aspects of the classroom such as paraprofessional resources and peer influences) on average classroom performance accounted for only 7% of the individual-differences variance in children’s reading ability in grades 1–3. In sum, the average classroom effect on individual differences in early grade reading has consistently been found to be small across the three cited studies with twins and the only large and well-designed experimental study on non-twins. Of course, extremely effective or ineffective teachers can have very positive or negative influences that are not obvious from the very modest average influence of classroom differences on early reading development. We do not mean to deny the importance of strong teacher training and monitoring of continued professional development. However, we wish to emphasize that it is important not to characterize classroom performance differences as an index of “teacher quality,” as was done in the Hart et al. (2013) study and the earlier studies by Nye et al. (2004), Taylor and Schatschneider (2010), and Taylor et al. (2010). Using mean classroom performance or classroom growth across the school year for the evaluation of teacher quality is complicated by many other influences on average classroom performance, such as the factors collectively referred to as “classroom climate,” exemplified by the students’ perceptions of their class’s attitude



Why children differ in reading ability 47

to learning. These influences are independent of particular teachers (Marsh, Martin, & Cheng, 2008), and can reflect variation across classrooms in children’s self-regulation skills (Skibbe, Phillips, Day, Brophy-Herb, & Connor, 2012). Average classroom performance will also be influenced by the genetic and classroom-independent environmental influences on individual students’ reading skills. Therefore, the use of classroom performance differences to rate “teacher quality” can be quite misleading, and often unfair, in this era of blaming teachers for children’s learning difficulties in reading and related skills (cf., Alter, 2007).

Implications of genetic influences for educational policy The evidence for strong genetic influences on individual differences and deficits in reading may seem discouraging to many educators, but results from behavior-genetic studies also suggest how genes influence reading development in ways that offer avenues for intervention. The question of how genes influence reading has two educationally relevant answers that we will consider here. One is that genes influence learning rates for reading and related skills. The other is that genes influence the environment through a gene-environment correlation. The Byrne et al. (2013) genetic factor analysis we reviewed earlier and Byrne et al. (2008) included on-line learning measures that demonstrated substantial genetic influences on learning rates for reading and related skills. The implication for education is that, depending on the severity of reading difficulties, much more reading practice, possibly including computer or tutor support for decoding difficulties (Wise, Ring, & Olson, 2000), may be required for a child with genetically constrained learning rates for reading accuracy, fluency, and comprehension to reach or more closely approach the “grade-level” criterion (average performance) that was originally required of “all children” by 2014 (107th Congress, 2002). Unfortunately, the second thing we have learned is that genetic constraints on learning rates for reading development are likely to result in less than normal reading practice, and thus in a gene-environment correlation that works against the need for greater reading practice. Olson and Byrne (2005) noted that there is significant genetic influence on a title recognition measure of print exposure. Harlaar, Dale, and Plomin (2007) found that genetic influence on word recognition at age 7 in the U.K. was highly correlated with genetic influence on an author recognition test at age 10, and with prior reading ability. Moreover, after controlling for genetic and environmental influences shared by word recognition at age 7 and author recognition at age 10, there was evidence for a separate shared environmental link between author recognition at age 10 and word recognition at age 12. In the Ohio twin sample, Harlaar, Deater-Deckard, Thompson, DeThorne, and Petrill (2011) recently confirmed a gene-environment correlation between independent reading rated by the twins and their caregivers at age 11 and their reading achievement at age 10. The educational implications of compromised learning rates and the gene-environment correlation are daunting for children with reading disabilities, their parents,

48 Richard K. Olson et al.

their teachers, and their schools. The extra instruction and reading practice needed to at least partly compensate for children’s reading problems may be difficult to accommodate and motivate in a typical school day filled with other academic demands. Of course, greater reading practice and instruction could be supported for all children if the school day and school year were longer. But with the short school days and years common in most western societies, compensatory reading practice often has to be supported in after-school classes and in the home. This raises a range of motivational issues because many children with reading disabilities would often rather be doing anything but reading after school and in their home. That is why organizations such as the International Dyslexia Association emphasize the importance of developing adequate reading skills, the payoff for working extra hard to improve those skills, and the transfer of that work ethic to other areas of life. Teachers can certainly help to provide this motivation, but they often need the support of the family and broader school environment do so.

Conclusion The foregoing discussion of how genes influence individual differences and deficits in reading raises the question of what are the reasonable expectations for children with reading disabilities, and for their parents, their teachers, and their schools. Here, in conclusion, we will grab the electrified third rail of educational discourse and say that our expectations for a child’s reading achievement may often be too high. We are aware that diminished expectations can result in children failing to reach their “potential” level of reading achievement, and we should guard against that. But the definition of “potential” at the individual level is complicated by genes and sometimes hidden environmental constraints, or at least ones that are beyond control, as well as the values of the society, the family and the child. Certainly the most optimistic and well-meaning, but absurd, criterion is for all children to be at least at “grade level” (typically defined as average performance for grade on standardized tests) as specified by the No Child Left Behind law (107th US Congress, 2001). Similar demands for high literacy in all children are included in the Common Core Standards (2010) adopted by most States in the U.S., which seek to “….ensure that all students are college and career ready in literacy no later than the end of high school.” While some of the requirements of the No Child Left Behind law have been relaxed, the sentiment is still expressed in many State education laws that all children must meet some minimal criterion, that teachers should be evaluated on their students’ reaching that criterion, and that children failing to meet criterion by the end of third grade should be retained in grade until they do (Rose, 2012). The relations of values to criteria for reading achievement are not often considered, but we think they should be, and the values should include those of the child and the family. Just how much additional reading practice and remedial instruction should be expected or required for children who are slow in their reading development? Some



Why children differ in reading ability 49

children with slow learning rates for reading and related skills may choose to devote much more than normal practice in reading at the expense of other activities to reach or more closely approach “grade level.” Other children may place less value on their reading proficiency, and more on other academic and non-academic activities. We believe that all children should have strong support for their reading development, including the opportunity for additional intensive instruction for those with learning difficulties in reading. But the evidence for strong genetic influences on many reading difficulties, including reading fluency that seems most resistant to intervention (Torgesen et al., 2001), requires a much more nuanced approach to reading ability expectations for children than those reflected in U.S. Federal and State laws. Currently these laws and public expectations for all children’s reading achievement can be quite unfair to many children with reading difficulties, their parents, their teachers, and their schools. In this review we have emphasized the role of genes in influencing the course of reading development. Our motivation has been in part to counter the view, common enough in the social sciences, that the environment is where most of the action is (Pinker, 2002). But in conclusion we want to remind readers that environmental influence is important too, as shown by the estimates of around 30% shared environment for reading disability and around 20% shared environment in the TEDS and Florida population studies of individual differences near the end of first grade. In addition, there is evidence that some of the high genetic influence on reading ability is due to a gene-environment correlation for reading practice, further emphasizing the importance of the reading environment in reading development. Even if estimates of shared environmental variance are very low in a twin sample, that may only suggest a very narrow effective environmental range in that sample. It does not preclude changes in the environment to improve reading at the low end of the distribution in that sample, as well as across the whole sample. Thus, regardless of the levels of genetic and environmental influence in a population, there is always room for well-designed interventions, including extended reading practice, and research should continue into the most effective interventions for reading difficulty and for improving literacy in the population as a whole.

References 107th Congress. (2002). The no child left behind act of 2001. Washington, DC: United States Congress. Alter, J. (2007, February 10). Stop pandering on education; It’s time to move from identifying failing schools to identifying failing teachers. Sounds obvious, but it hasn’t happened in American education. Newsweek, p. 55. Betjemann, R. S., Keenan, J. M., Olson, R. K., & DeFries, J. C. (2011). Choice of reading comprehension test influences the outcomes of genetic analyses. Scientific Studies of Reading, 15, 363–382.  doi: 10.1080/10888438.2010.493965

50

Richard K. Olson et al.

Betjemann, R. S., Willcutt, E. G., Olson, R. K., Keenan, J. M., DeFries, J. C., & Wadsworth, S. J. (2008). Word reading and reading comprehension: Stability, overlap, and independence. Reading and Writing: An Interdisciplinary Journal, 21, 539–558.  doi: 10.1007/s11145-007-9076-8 Byrne, B., Coventry, W. L., Olson, R. K., Samuelsson, S., Corley, R., Willcutt, E. G., Wadsworth, S., & DeFries, J. C. (2009). Genetic and environmental influences on aspects of literacy and language in early childhood: Continuity and change from preschool to grade 2. Journal of Neurolinguistics, 22, 219–236.  doi: 10.1016/j.jneuroling.2008.09.003 Byrne, B., Coventry, W. L., Olson, R. K, Hulslander, J., Wadsworth, S., DeFries, J. C., et al. (2008). A behavior-genetic analysis of orthographic learning, spelling, and decoding. Journal of Research in Reading, 31, 8–21.  doi: 10.1111/j.1467-9817.2007.00358.x Byrne, B., Coventry, W. L., Olson, R. K., Wadsworth, S. J., Samuelsson, S., Petrill, S. A., Willcutt, E. G., & Corley, R. (2010). “Teacher effects” in early literacy development: Evidence from a study of twins. Journal of Educational Psychology, 102, 32–42.  doi: 10.1037/a0017288 Byrne, B., Delaland, C., Fielding-Barnsley, R., Quain, P., Samuelsson, S., Hoien, T., Corley, R., DeFries, J. C., Wadsworth, S., Willcutt, E., & Olson, R. K. (2002). Longitudinal twin study of early reading development in three countries: Preliminary results. Annals of Dyslexia, 52, 49–74.  doi: 10.1007/s11881-002-0006-9 Byrne, B., Samuelsson, S., Wadsworth, S., Hulslander, J., Corley, R., DeFries, J. C., Quain, P., Willcutt, E., & Olson, R. K. (2007). Longitudinal twin study of early literacy development: Preschool through Grade 1. Reading and Writing: An Interdisciplinary Journal, 20, 77–102. doi: 10.1007/s11145-006-9019-9 Byrne, B., Wadsworth, S., Boehme, K., Talk, A. C., Coventry, W. L., Olson, R. K., Samuelsson, S., & Corley, R. (2013). Multivariate genetic analysis of learning and early reading development. Scientific Studies of Reading, 17, 224–242.  doi: 10.1080/10888438.2011.654298 Carey, G. (2003). Human genetics for the social sciences. London: Sage Publications. Christopher, M. E., Hulslander, J., Byrne, B., Samuelsson, S., Keenan, J. M., Pennington, B., … & Olson, R. K. (2013). Modeling the etiology of individual differences in early reading development: Evidence for strong genetic influences. Scientific Studies of Reading, 17, 350–368. doi: 10.1080/10888438.2012.729119 Christopher, M. E., Hulslander, J., Byrne, B., Samuelsson, S., Keenan, J. M., Pennington, B., DeFries, J. C, Wadsworth, S. J, Willcutt, E., & Olson, R. K. (2013). The genetic and environmental etiologies of individual differences in early reading growth in Australia, the United States, and Scandinavia. Journal of Experimental Child Psychology, 115, 453–467. doi: 10.1016/j.jecp.2013.03.008 Christopher, M. E., Hulslander, J., Byrne, B., Samuelsson, S., Keenan, J. M., Pennington, B., … Olson, R. K. (2015). Genetic and environmental etiologies of the longitudinal relations between prereading skills and reading. Child Development, 86, 342–361  doi: 10.1111/cdev.12295 Common Core State Standards for English Language Arts & Literacy in History/Social Science, and Technical Science (2010). Accessed October 16, 2012 at http://www.corestandards.org/ assets/CCSSI_ELA%20Standards.pdf DeFries, J. C., Filipek, P. A., Fulker, D. W., Olson, R. K., Pennington, B. F., Smith, S. D., & Wise, B. W. (1997). Colorado Learning Disabilities Research Center. Learning Disabilities: A Multidisciplinary Journal, 8, 7–19. DeFries, J. C., & Fulker, D. W. (1985). Multiple regression analysis of twin data. Behavior Genetics, 15, 467–478.  doi: 10.1007/BF01066239 DeFries, J. C., Fulker, D. W., & LaBuda, M. C. (1987). Evidence for a genetic aetiology in reading disability of twins. Nature, 329, 537–539.  doi: 10.1038/329537a0



Why children differ in reading ability 51

Elwér, Å., Gustafson, S., Byrne, B., Olson, R. K., Keenan, J. M., & Samuelsson, S. (2015). Oral language deficits in poor reading comprehension with adequate word decoding. Scandinavian Journal of Psychology, 56, 157–166.  doi: 10.1111/sjop.12188 Elwér, Å., Keenan, J. M., Olson, R. K., Byrne, B., & Samuelsson, S. (2013). Longitudinal stability and predictors of poor oral comprehenders and poor decoders. Journal of Experimental Child Psychology, 115, 497–516.  doi: 10.1016/j.jecp.2012.12.001 Friend, A , DeFries, J. C., & Olson, R. K. (2008). Parental education moderates genetic influences on reading disability. Psychological Science, 19, 1124–1130.  doi: 10.1111/j.1467-9280.2008.02213.x Friend, A., DeFries, J. C., Olson, R. K., Pennington, B., Harlaar, N., Byrne, B., … Keenan, J. M. (2009). Heritability of high reading ability and its interaction with parental education. Behavior Genetics, 39, 427–436.  doi: 10.1007/s10519-009-9263-2 Gayán, J., & Olson, R. K. (2001). Genetic and environmental influences on orthographic and phonological skills in children with reading disabilities. Developmental Neuropsychology, 20, 487–511.  doi: 10.1207/S15326942DN2002_3 Gayán, J., & Olson, R. K. (2003). Genetic and environmental influences on individual differences in printed word recognition. Journal of Experimental Child Psychology, 84, 97–123. doi: 10.1016/S0022-0965(02)00181-9 Harlaar, N., Cutting, L., Deater-Deckard, K., DeThorne, L. S., Justice, L. M., Schatscheeider, C., … Petrill, S. A. (2010). Predicting individual differences in reading comprehension: A twin study. Annals of Dyslexia, 60, 265–288.  doi: 10.1007/s11881-010-0044-7 Harlaar, N., Dale, P., Hayiou-Thomas, & Plomin, R. (2012). Individual variation in reading achievement trajectories: New evidence from a UK twin study. Paper presented at the meeting of the Society for the Scientific Study of Reading, Montreal, July 12, 2012. Harlaar, N., Deater-Deckard, K., Thompson, L. A., DeThorne, L. S., & Petrill, S. A. (2011). Associations between reading achievement and independent reading in early elememtary school: a genetically informative cross-lagged study. Child Development, 82, 2123–2137.  doi: 10.1111/j.1467-8624.2011.01658.x Harlaar, N., Dale, P. S., & Plomin, R. (2007). Reading exposure: A (largely) environmental risk factor with environmentally-mediated effects on reading performance in the primary school years. Journal of Child Psychology and Psychiatry, 48, 1192–1199.  doi: 10.1111/j.1469-7610.2007.01798.x Harlaar, N., Spinath, F. M., Dale, P. S., & Plomin, R. (2005). Genetic influences on early word recognition abilities and disabilities: A study of 7-year-old twins. Journal of Child Psychology and Psychiatry, 46, 373–384.  doi: 10.1111/j.1469-7610.2004.00358.x Hart, S. A., Logan, J. A. R., Soden-Hensler, B., Kershaw, S., Taylor, J., & Schatschneider, C. (2013). Exploring how nature and nurture affect the development of reading: An analysis of the Florida Twin Project on Reading. Developmental Pscyhology. 49, 1971–1981. doi: 10.1037/a0031348 Hart, S. A., Taylor, J., & Schatschneider, C. (2013). There is a world outside of experimental designs: using twins to investigate causation. Assessment for Effective Intervention, 38, 117–126 doi: 10.1177/1534508412451490 Hoover, W. A. & Gough, P. B. (1990). The simple view of reading. Reading and Writing, 2, 127–160. doi: 10.1007/BF00401799 Hulslander, J., Olson, R. K., Willcutt, E. G., & Wadsworth, S. J. (2010). Longitudinal stability of reading-related skills and their prediction of reading development. Scientific Studies of Reading, 14, 111–136.  doi: 10.1080/10888431003604058 Kovas, Y., Haworth, C. M. A., Dale, P. S., & Plomin, R. (2007). The genetic and environmental origins of learning abilities and disabilities in the early school years. Monographs of the Society for Research in Child Development, 72, 1–144.

52

Richard K. Olson et al.

Keenan, J. M., Betjemann, R. S., & Olson, R. K. (2008). Reading comprehension tests vary in the skills they assess: Differential dependence on decoding and oral comprehension. Scientific Studies of Reading, 12, 281–300.  doi: 10.1080/10888430802132279 Keenan, J. M., Betjemann, R., Wadsworth, S. J., DeFries, J. C., & Olson, R. K., (2006). Genetic and environmental influences on reading and listening comprehension. Journal of Research in Reading, 29, 75–91.  doi: 10.1111/j.1467-9817.2006.00293.x Keller M. C., & Coventry, W. L. (2005). Quantifying and addressing parameter indeterminacy in the classical twin design. Twin Research and Human Genetics, 8, 201–213. doi: 10.1375/twin.8.3.201 Keller, M. C., Medland, S. E., & Duncan, L. E. (2010). Are extended twin family designs worth the trouble? A comparison of the bias, precision, and accuracy of parameters estimated in four twin family models. Behavior Genetics, 40, 377–393.  doi: 10.1007/s10519-009-9320-x Logan, J. A. R., Hart, S. A., Cutting, L., Deater-Deckard, K., Schatschneider, C., & Petrill, S. A. (2013). Reading development in children ages 6 to 12: Genetic and environmental influences. Child Development, 84, 2131–2144.  doi: 10.1111/cdev.12104 Marsh, H. W., Martin, A. J., & Cheng, J. H. S. (2008). A multi-level perspective on gender in classroom motivation and climate: Potential benefits of male teachers for boys? Journal of Educational Psychology, 100, 78–95.  doi: 10.1037/0022-0663.100.1.78 Nye, B., Konstantopoulos, S., & Hedges, L. V. (2004). How large are teacher effects? Educational Evaluation and Policy Analysis, 26, 237–257.  doi: 10.3102/01623737026003237 Olson, R. K. (2004). SSSR, environment, and genes. Scientific Studies of Reading, 8, 111–124. doi: 10.1207/s1532799xssr0802_1 Olson, R. K. (2006). Genes, Environment, and Dyslexia: The 2005 Norman Geschwind Memorial Lecture. Annals of Dyslexia, 56, 205–238.  doi: 10.1007/s11881-006-0010-6 Olson, R. K., & Byrne, B. (2005). Genetic and environmental influences on reading and language ability and disability. In H. Catts, & A. Kamhi (Eds.), The connections between language and reading disabilities (pp. 173–200). Mahwah, NJ: Laurence Erlbaum Associates. Olson, R. K., Keenan, J. M., Byrne, B., Samuelsson, S., Coventry, W. L., Corley, R.,…Hulslander, J. (2011). Genetic and environmental influences on vocabulary and reading development. Scientific Studies of Reading, 15, 26–46.  doi: 10.1080/10888438.2011.536128 Olson, R. K., Wise, B., Conners, F., Rack, J. & Fulker, D. (1989). Specific deficits in component reading and language skills: Genetic and environmental influences. Journal of Learning Disabilities, 22, 339–348.  doi: 10.1177/002221948902200604 Petrill, S. A., Deater-Deckard, K., Thompson, L. A., Schatschneider, C., DeThorne, L. S., & Vandenbergh, D. J. (2007). Longitudinal genetic analysis of early reading: The Western Reserve Reading Project. Reading and Writing, 20, 127–146.  doi: 10.1007/s11145-006-9021-2 Petrill, S. A., Hart, S. A., Harlaar, N., Logan, J., Justice, L. M., Schatschneider, C., … Cutting, L. (2010). Genetic and environmental influences on the growth of early reading skills. Journal of Child Psychology and Psychiatry, 51, 660–667.  doi: 10.1111/j.1469-7610.2009.02204.x Pinker, S. (2002). The blank slate: The modern denial of human nature. New York: Penguin Books. Plomin, R., DeFries, J. C., McClearn, G. E., & McGuffin, P. (2008). Behavioral genetics (5th ed.). New York, NY: Worth. Rose, S. (2012). Third grade reading policies. Education Commission of the States, www.ecs.org, accessed 9/27/2012. Sadoski, M., & Wilson, V. L. (2006). Effects of a theoretically based large-scale reading intervention in a multicultural urban school district. American Educational Research Journal, 43, 137–154.  doi: 10.3102/00028312043001137



Why children differ in reading ability 53

Samuelsson, S., Byrne, B., Olson, R. K., Hulslander, J., Wadsworth, S., Corley, R., DeFries, J. C. (2008). Response to early literacy instruction in the United States, Australia, and Scandinavia: A behavior-genetic analysis. Learning and Individual Differences, 18, 289–295. doi: 10.1016/j.lindif.2008.03.004 Samuelsson, S., Byrne, B., Quain, P., Corley, R., DeFries, J. C., Wadsworth, S., Willcutt, E., & Olson, R. K. (2005). Environmental and genetic influences on pre-reading skills in Australia, Scandinavia, and the U.S. Journal of Educational Psychology, 97, 705–722. doi: 10.1037/0022-0663.97.4.705 Samuelsson, S., Olson, R. K., Wadsworth, S., Corley, R., DeFries, J. C., Willcutt, E., Hulslander, J., & Byrne, B. (2007). Genetic and environmental influences on pre-reading skills and early reading and spelling development: A comparison among United States, Australia, and Scandinavia. Reading and Writing: An Interdisciplinary Journal, 20, 51–75. doi: 10.1007/s11145-006-9018-x Skibbe, L. E., Phillips, B. M., Day, S. L., Brophy-Herb, H. E., & Connor, C. M. (2012). Children’s early literacy growth in relation to classmates’ self-regulation. Journal of Educational Psychology, 104, 541–553.  doi: 10.1037/a0029153 Taylor, J., & Schatschneider, C. (2010). Genetic influence on literacy constructs in kindergarten and first grade: Evidence from a diverse twin sample. Behavior Genetics, 40, 591–602. doi: 10.1007/s10519-010-9368-7 Taylor, J., Roehrig, A. D., Soden Hensler, B., Connor, C. M., & Schatschneider, C. (2010). Teacher quality moderates the genetic effects on early reading. Science, 328, 512–514. doi: 10.1126/science.1186149 Torgesen, J. K., Alexander, A. W., Wagner, R. K., Rashotte, C. A., Voeller, K. K. S., & Conway, T. (2001). Intensive remedial instruction for children with severe reading disabilities: Immediate and long-term outcomes from tow instructional approaches. Journal of Learning Disabilities, 34, 33–58.  doi: 10.1177/002221940103400104 Torgesen, J. K., Wagner, R. K., & Rashotte, C. A. (1999). Test of Word Reading Efficiency (TOWRE). Austin, TX: Pro-Ed. Wadsworth, S. J., Corley, R. P., Plomin, R., Hewitt, J. K. & DeFries, J. C. (2006). Genetic and environmental influences on continuity and change in reading achievement in the Colorado Adoption Project. In A. Huston & M. Ripke (Eds.), Developmental contexts of middle childhood: Bridges to adolescence and adulthood (pp. 87–106). New York: Cambridge University Press.  doi: 10.1017/CBO9780511499760.006 Wise, B. W., Ring, J., & Olson, R. K. (2000). Individual differences in gains from computer-assisted remedial reading with more emphasis on phonological analysis or accurate reading in context. Journal of Experimental Child Psychology, 77, 197–235.  doi: 10.1006/jecp.1999.2559

Early literacy across languages Catherine McBride

The Chinese University of Hong Kong

Is the process of learning to read and write pretty much the same across languages and cultures? The answer to this fundamental question will always be both yes and no because it will depend crucially on the lens through which we address this issue. The goal of learning to read is reading to learn, where reading comprehension is the central focus. Ultimately, the basic ingredients of reading comprehension, from the perspective of individual attributes, including the importance of background knowledge, inference-making, metacognition, memory, and word recognition skills (Snow & Ninio, 1986), appear to be similar across languages. However, the early years of word recognition and writing may require somewhat different skills in different orthographies. In this chapter, I focus on word reading and writing in different orthographies. I first briefly explore the concept of a word itself for the purpose of literacy learning. I then highlight the three basic cognitive dimensions on which word reading and writing are compared, namely, phonological, semantic (or morphological), and orthographic aspects of word recognition (e.g., Seidenberg & McClelland, 1989). Next, contextual aspects of word learning are highlighted at a very general level. Finally, I look in more depth at phonological processing, semantic skills, and visual-orthographic processing for word reading and then for word writing. I conclude by highlighting the fact that there are differences in learning to read and to write words across languages and orthographies. Such differences are important for modeling of early literacy acquisition around the world.

What is a word? As we think about word reading across languages and scripts, it is important to note that the idea of what constitutes a word itself is not always clear from one orthography to another (McBride-Chang et al., 2012). In Spanish, French, English, or German, words are clearly demarcated on a page because they constitute a group of letters with space on either side. Readers of these languages recognize a word visually in this way from early on. However, such visual-spatial cues are not present in every orthography. For example, Thai does not separate words by spacing. Instead, readers of this orthography begin to internalize the concept of word with experience. Younger children are taught to read Thai with spatial cues provided, but older ones transition to the doi 10.1075/swll.15.04mcb © 2017 John Benjamins Publishing Company

56

Catherine McBride

adult method of reading without spaces. In Chinese, there is equal spacing from one character to another throughout the text. Sometimes even Chinese expert adults do not agree on how to demarcate words in text. Finnish is another interesting example. Finnish contains extensive compounding and derivational rules such that words are often up to 20 letters or so in length. Thus, as we talk about early literacy skills across cultures, we must bear in mind that not even the concept of a single word is a universal the way it might be understood in English or French. We have argued (Li & McBride-Chang, 2014) that the unit of recognition for Chinese could be either the character or the word. In this conceptualization, words are comprised of two or more characters. There is some preliminary evidence that Chinese characters are recognized more easily by kindergartners and primary school students when embedded within words than when presented alone without such context. Character recognition may require particular attention to orthographic configurations, such as the positions and types of radicals included. In contrast, word recognition in Chinese may make clearer use of the dimension of morphological awareness that is lexical compounding (e.g., Wang, McBride-Chang & Chan, 2013). When children understand how morphemes fit together to form words, they tend to be able to make clearer educated guesses as to the identity of words even when only part of the word can be identified. I have often wondered whether a similar process might be at work in English. Intuitively, shorter words in English should be easier to recognize because processing them requires the analysis of fewer phonemes and phonological units overall. However, sometimes analyzing a longer word may provide clues as to its identity. For example, basketball might be easier to read than basket, simply because recognizing the morpheme of ball embedded in it may narrow the choices possible for the beginning of the word in a way that reading basket on its own might not. Thus, the unit of word recognition may differ from one script to another. With this brief note on word reading, we turn now to factors explaining it across languages.

Basic dimensions of word reading There are a variety of dimensions on which to compare orthographies that are relevant to word recognition itself. Seymour, Aro and Erskine (2003), for example, focused on shallow and deep orthographies. This work was primarily focused on the extent to which alphabetic spellings are phonologically transparent vs. opaque. Importantly, Chinese, and many other languages/scripts, are not even included on this list of orthographies because the original focus was on Indo-European languages. Beyond this distinction, I find it helpful to think about languages and corresponding orthographies as differing on two somewhat distinct aspects of phonological sensitivity, suprasegmental and segmental phonology, as discussed later in the chapter. Both dimensions of phonological sensitivity are crucial for learning to read (for reviews, see, e.g., Zhang & McBride-Chang, 2010; Ziegler & Goswami, 2005).



Early literacy across languages 57

Beyond the phonological level, how semantic information is communicated in a given script is important for understanding early differences in learning to read. Semantic information is most typically talked about in beginning readers as related morphological information contained in words. This level of understanding of words is crucial for learning to read in Chinese but not necessarily very strong for those learning to read in alphabetic orthographies. As an example, at an extreme, whereas training in phonological awareness has been shown to facilitate word reading in virtually every orthography in which it has been studied (Chiappe, Siegel, & Wade-Woolley, 2002; Vellutino & Scanlon, 1987), such training does not promote better word reading skills in Chinese (Zhou, McBride-Chang, Fong, Wong, & Cheung, 2012). In contrast, a focus on morphological awareness in the form of understanding of homographs/ homophones and lexical compounding (Ku & Anderson, 2003; McBride-Chang, Shu, Zhou, Wat & Wagner, 2003; Packard, 2000; Wu et al., 2009; Zhou et al., 2012) tends to promote better reading skills in Chinese. Morphological awareness training has also been helpful in facilitating word reading in alphabetic orthographies in some contexts for early readers (e.g., Bowers, Kirby, & Deacon, 2010; Deacon, Wade-Woolley & Kirby, 2007), but given the prominence of phonological sensitivity for reading in alphabetic orthographies, early morphological awareness training tends to be less of a focus in these learners. Another dimension proposed by Nag (2007) focuses on the visual elements of a script; she describes this dimension as ranging from extensive to contained orthographies. Extensive orthographies have many visual symbols to be assimilated. Chinese is the best example of an extensive orthography because the number of individual characters is in the thousands. English, with its 26 letters in the alphabet, or Hebrew, with 22, are two examples of what Nag (2007) refers to as contained orthographies. In the middle are those from akshara such as Bengali or Kannada, which sometimes have over 400 different individual symbols to be learned (Nag, 2007). Seidenberg and McClelland’s (1989) triangle theory emphasized all three crucial aspects of reading, i.e., phonological, semantic (or morphological), and orthographic aspects of word recognition.

Early literacy in context Contextual variables interact with the aforementioned cognitive dimensions of differences across orthographies. The way in which literacy skills are taught is one such contextual variable. Clearly, word recognition can be taught in different ways. What works best for one child may not work well for another. Moreover, there are different schools of thought on optimal teaching, leading to different teaching methods. One obvious example is the whole language vs. phonics methods for English (Bruck, Treiman, Caravolas, Genesee, & Cassar, 1998). The nature of the orthography does not necessarily dictate that a single method of teaching of reading will work better than another one. For instance, one of the longest traditions of teaching of Chinese

58

Catherine McBride

focuses on copying and rote memorization (Wu, 1999). However, scholars have also emphasized the importance of grouping Chinese characters and words by common, shared semantic radicals or common characters in order to reduce memory loads for literacy learning (e.g., Tse, Marton, Ki, & Loh, 2007). The phonological coding system of Pinyin is another important dimension of teaching of literacy for Mainland Chinese children. Knowledge of this system has been shown to be an independent predictor of word reading and writing in Chinese even five years later (Pan et al., 2011). In addition, despite the fact that Korean Hangul has been described as very easy and efficient to learn because of its clear phonological principles (Yoon, Bolger, Kwon, & Perfetti, 2002), most child learners do not begin by learning each component grapheme in a syllable block (which can contain both vertically and horizontally oriented graphemes) but rather memorize individual syllables as visual wholes (Park, 2000). Beyond orthographies and teaching methods, learning to read in different languages involves a whole host of environmental factors. Although I cannot give an exhaustive list of these factors, a few of the more prominent ones can be considered here. One is the age at which formal literacy instruction takes place. Hong Kong may be at one extreme of this continuum, with children who begin learning formally in the first year of kindergarten, around the age of 3.5 years. At another extreme are some European countries which do not begin teaching literacy skills until children are in first grade, around the age of 6 to 7 years old. Another environmental factor that affects literacy learning across cultures is parental support. Some orthographies, such as Chinese, are more difficult to learn to read by their very nature. Chinese parents often spend long hours in pre-literacy activities (e.g., Li & Rao, 2000; McBride-Chang, 2004). Formal schooling also differs widely from one place to another worldwide. The school the child attends represents a multitude of factors that can affect literacy learning. In one study across over forty countries, Chiu and McBride-Chang (2006) showed that school is important because of the composition of members. For example, more female classmates tended to be associated with better reading scores in 14-year-old students around the world according to PISA data. Moreover, girls seemed to enjoy the reading process more, and this enjoyment of reading was contagious across peers in that study. Schools also represent commitment and resources from the collective group of parents supporting the schools (e.g., Chiu, McBride-Chang, & Lin, 2012). Big and small factors related to school achievement and school focus all fall under a larger header of school-related environmental factors (Chiu & McBride-Chang, 2006; Chiu et al., 2012). Both early in development and as formal schooling takes place, children’s literacy growth depends strongly on parental scaffolding, or supported interactions as children learn to read and write. Parents’ support of learning to read probably takes many forms. One is early reading and sometimes scaffolding of writing with children. Different cultures may have different approaches to supporting children in the literacy-learning process. For example, Asian parents, particularly Japanese, Korean, and Chinese, are well known for being relatively academically demanding and strict with their children (Lin & Fu, 1990). In contrast, there is the perception that Hispanic



Early literacy across languages 59

families tend to be relatively relaxed about early academic achievement, though this depends on families’ education levels (e.g., Ortiz, 2004). What about print exposure? Parents’ storybook reading has some clear benefits for children’s subsequent literacy outcomes, particularly in building up solid vocabulary knowledge and a sense of narrative or story structure (Brown, 1975; LeFevre, 2001; Nezworski, Stein & Trabasso, 1982; Roberts, 2008). However, storybook reading by itself has little to do with teaching children the mechanics of word recognition including the letters of an alphabet or some of the basic radicals used to make up characters (or characters themselves) in Chinese. Such academically oriented learning often involves parents’ early writing interactions with children (e.g., Levin, Aram, Tolchinsky, & McBride, 2013; Levin & Aram, 2012; Lin, McBride-Chang, Aram, Shu, Levin, & Cho, 2012). Around the world, parents’ focused attentions to building blocks of words (e.g., letters, characters) and to writing and its characteristics tend to be important factors in facilitating early word reading and word writing in children across cultures (e.g., Aram & Levin, 2004; McBride-Chang, 2004). In general, home literacy environment, including literacy resources such as different types of books in the home (Chiu & McBride-Chang, 2006), matters strongly for reading development and can differ sharply by culture (Chiu & McBride-Chang, 2006). Note that culture here could refer to geographic area (e.g., Finland vs. Zambia vs. Malaysia), but it could also refer to socio-economic status in a broad sense, with those with better educations and more financial resources often doing more toward early scaffolding of literacy skills than those who are poorer in these aspects (e.g., Aram & Levin, 2004; Chiu et al., 2012; Chiu & McBride-Chang, 2006). Having said something about the importance of environment for promoting early literacy skills, defined here as learning to read and write individual words, we turn back now to cognitive skills that influence them across cultures. Such cognitive factors are part of the building blocks of solid literacy skills. They are the abilities that researchers test in younger children to look at potential for literacy skills, especially in an effort to identify those who might be at risk for dyslexia or other reading difficulties. These are also the skills that can often be targeted early or even later in development to be honed in order to facilitate better reading. Once these are identified, teachers, clinicians, or educational psychologists can sometimes focus on individual areas of strength and weakness to facilitate the reading and writing process later on. I begin with phonological sensitivity, which is a hallmark of reading skill (Byrne, Freebody, & Gates, 1992; Fletcher et al., 1994; Stanovich & Siegel, 1994). I review both suprasegmental and segmental aspects of phonological processing because researchers have established that both are important and unique facets of the reading process (Wood, Wade-Woolley & Holliman, 2009; Zhang & McBride-Chang, 2010).

60 Catherine McBride

Phonological processing information Both suprasegmental processing and segmental processing can be important for learning to read. Suprasegmental processing takes place across the entire word and is generally applied to words of more than one syllable. In English, such suprasegmental processing is particularly exemplified by stress. For example, there are some basic rules regarding where stress typically falls in 2- or 3-syllable words, and these often depend partly on what part of speech the word is in (e.g., noun, adjective, verb). However, most native English-speaking children are not aware of these rules and simply acquire them through typical speech. Stress is not marked in print in English, though it is in some other languages such as Spanish (Harris, 1983). Researchers have found that stress sensitivity in English and other languages (Holliman, Wood, & Sheehy, 2008; Wood, 2006) is independently associated with better reading skills. In Chinese, though most experts agree that stress is fairly consistent across syllables, lexical tone is important in marking contrasts in meaning across syllables. Researchers (Cheung et al., 2009; Ho, Leung, & Cheung, 2011; Wong, Ciocca, & Yung, 2009) have demonstrated that a lack of sensitivity to lexical tone tends to be associated with dyslexia in Chinese children. In some studies, sensitivity to lexical tone seems to explain variance in reading skills across time (McBride-Chang et al., 2008). Wood, Wade-Woolley and Holliman (2009), Calet, Gutierrez-Palma, Simpson, Gonzalez-Trujillo, and Defior (2015), and others have pioneered ways to measure suprasegmental processing in alphabetic orthographies with consistent results demonstrating that suprasegmental processing is additionally predictive of variation in reading skills beyond traditional tasks of phonological awareness; researchers have demonstrated a similar phenomenon for Chinese children (Ho & Bryant, 1997; McBride-Chang, Tong, Shu, Wong, Leung, & Tardif, 2008). Segmental processing refers to what researchers have traditionally termed phonological awareness. This is awareness of and access to the speech sounds of a language, in larger or smaller segments. In English, phoneme awareness is important, given that each letter of the alphabet can represent a single speech sound and that blending such sounds together is an integral aspect of word recognition (e.g., Ziegler & Goswami, 2005). For example, each letter in the letter sequence of c-a-t makes a unique sound according to the English alphabet. Together, they form a word. Larger phonological units are also important for learning to read (Ziegler & Goswami, 2005). In Chinese, for example, syllable awareness is an excellent predictor and correlate of very early reading because the majority of Chinese words are comprised of two or more syllables, and combining syllables together may help to facilitate word recognition (e.g., McBride-Chang, Tong et al., 2008). A cross-cultural, cross-linguistic focus on dimensions of phonological processing should consider both segmental and suprasegmental processing. Segmental processing is clearly associated with reading outcomes longitudinally in English (Torgesen, Wagner & Rashotte, 1994; Vellutino & Scanlon, 1987). It is also important early on for other alphabetic orthographies such as Finnish. For example, Lyytinen (1997)



Early literacy across languages 61

came up with a country-wide screening procedure for Finnish children. Those with poor phonological awareness at the segmental level are now routinely presented with a game called Graphogame, a simple computer game targeting letter sound and early phonics principles to teach word recognition, sponsored by the Finnish government and asked to play this game routinely around age 4. Playing this game facilitates better letter knowledge, segmental awareness and, ultimately, word reading (for a review, see Richardson & Lyytinen, 2014). This is just one example. Across alphabetic orthographies, training in phonological awareness tends to facilitate better word reading skills, especially when done over time and especially when the phonological segments are represented not only orally but in print (e.g., in French – Casalis & Louis-Alexandre, 2000; in Norwegian – Lyster, 2002; in Dutch – van Goch, McQueen, & Verhoeven, 2014). Future work that examines phonological sensitivity across cultures should take into account both dimensions of phonological processing. More and more studies are focused on both dimensions as they explore reading and writing development (see, e.g., Wang & Arciuli, 2015, for an overview). One question of interest across cultures might be whether the marking of suprasegmental processing (e.g., stress; tone) within a script matters for reading development. For example, is suprasegmental processing a more important marker of reading variability in English than in Spanish given that it is marked in Spanish but not in English? Tonal markings are not indicated in Chinese script either. Is this important? A continuous focus on phonological awareness could also be extended to how to combine training in phonemic units with training in morphological skills (e.g., Dallasheh-Khatib, Ibrahim, & Karni, 2014) or fluency. Thus, there is much to be explored, both in the area of suprasegmental processing, particularly in relation to marked and unmarked scripts, and potential for training in such processing, and in the more traditional area of phonological awareness.

Semantic information How meaning is conveyed within a script is another important difference across orthographies. Much of this information has to do with how morphemes are represented. Morphemes are the smallest units of meaning in a word, and the kinds of morphemes children encounter vary widely. In English, for example, the word “dogs” contains two morphemes, dog, which represents the object of interest and s, which makes this word plural. A relatively long word such as lettuce is only one morpheme. Underhanded consists of a prefix, under, a base word, hand, and the ed, which in this case renders it an adjective. Different kinds of morphemes across words include base words, which are sometimes combined in compound words (e.g., daylight), derivational morphemes such as prefixes and suffixes (unclear; spacious), and inflectional morphology which influences grammar (e.g., ed for past tense or s to make words plural in English). The prevalence of these categories across languages varies widely. In Dutch, German, and English, for instance, inflectional morphology is prevalent in early

62 Catherine McBride

reading and writing, and those who know how to apply the rules of grammar in their language tend to be better readers (e.g., Rispens, McBride-Chang, & Reitsma, 2008). For example, children with good knowledge of morphemes will understand that talked is spelled as it is and not as talkt which is how it sounds, based on the principles of inflectional grammar, i.e., that ed is often used to communicate past tense (Nunes, Bryant, & Bindman, 1997). In older children learning to read and write Indo-European languages, knowledge of derivational grammar is also associated with better reading skills (Carlisle, 1995; 2000; Deacon & Kirby, 2004). These categories are relatively rare in Chinese. However, in Chinese and also in Korean, lexical compounding, or formation of compound words, is particularly prevalent and the ability to manipulate such morphemes is also directly associated with reading skill (McBride-Chang et al., 2005; Tong et al., 2011). Compounding is important for word reading, perhaps particularly in Chinese (e.g., McBride-Chang et al., 2005) in order to disambiguate meanings from one another. For example, we need to hear flower in the context of sunflower in order to understand that it is this spelling to which the word refers and not to flour, instead. There is another aspect of conveying meaning in print that occurs in Chinese but not in other scripts. This is the fact that the semantic radical is generally one component of most Chinese characters. The semantic radical is not pronounced when it is within a larger character (some semantic radicals can stand alone and are pronounceable on their own; but in the context of a character containing the semantic radical and also a phonetic, the semantic radical communicates only semantic, but not phonological, information). However, it conveys some aspect of the meaning of each character. For example, Chinese characters that contain the semantic radical that represents the mouth are found in characters having something to do with the mouth, such as eat, kiss, and sing. The water radical is found in characters having something to with water, such as river, ocean, or sea. This unique aspect of Chinese is not necessarily possible to capture on a continuum across orthographies. There is no analogy to semantic radicals in alphabetic orthographies. In one paper (Wang & McBride, 2016), we have demonstrated that in Chinese, lexical compounding is important for one aspect of learning to read Chinese (i.e., recognizing 2- or 3-character words) and orthographic knowledge is important for learning to read in another aspect, i.e., that of individual Chinese characters. Across orthographies, semantic information is a crucial aspect of word recognition. Semantic information is conveyed by parts of words that are consistent across different words (e.g., inflections, derivations), by spellings that disambiguate words that sound the same but have different meanings (e.g., to, too, two), and by semantic radicals in Chinese. Although the forms of semantic information used are different across different orthographies, the purpose of this information is to indicate specific meanings and sometimes to identify one word or part of a word over another with a similar or identical phonological representation.



Early literacy across languages 63

Visual-orthographic information As Nag (2007) noted, scripts tend to vary in visual-orthographic information contained within them too. At one end are alphabets which make use of a core, very small set of visual symbols, letters which might constitute a relatively few (e.g., 26 for English) unique representations. In the middle are some Indian orthographies that might contain up to 400 or more independent symbols (Nag, 2007). At the other end is Chinese, which boasts literally tens of thousands of independent characters. Many scripts contain diacritics as well. These are additional symbols that indicate changes in meaning or pronunciations. For example, Vietnamese contains nine different diacritics, some falling at the top, bottom, or sides of letters to indicate changes (see Figure 1). Many scripts, including French, Hebrew, Arabic, and Thai include these to indicate various changes in pronunciations not indicated by the alphabet itself. There is perhaps the least information on this continuum of visual-orthographic information in all of its forms so far from research, probably because research on reading arose first or at least primarily with English. English is by all accounts an atypical orthography (Share, 2008). Many of its words were borrowed from various different languages and its irregular (opaque) spelling as compared to other orthographies reflects historical happenstance, rather than a consistent system. Sometimes it is difficult to disentangle how phonological, semantic, and orthographic information is measured entirely separately. For example, distinguishing the meaning of no from know or two from to or too is important for reading, but in some ways requires using all three types of information simultaneously. The words sound identical. They also look different and have different meanings. ã

á

high falling-raising tone

non-flat, high raising tone

à

ả

ạ

low, flat tone

low falling-raising tone

non-flat, low falling tone

a (no diacritics) high, flat tone

Figure 1.  5 out of the 9 diacritics are used to indicate different tones in Vietnamese alphabets. Syllable with different tones have different meaning, for example, ca (to sing), cà (eggplant), and cá (fish).

Orthographic information is essentially having a sense of the look of different words in print. This sense is manifested in a notion of the rules of writing, whether implicit or explicit. We English writers might spell a word on paper and then worry that it simply “doesn’t look right.” With experience, literacy learners of a given script come to understand many implicit ideas about it. For example, in English, a word can end with

64 Catherine McBride

a double s (SS) but cannot begin with this same feature. In Chinese, there are certain rules that are understood relatively quickly as well. For example, phonetic radicals, parts of Chinese characters that indicate something about the sound of a character, typically fall on the right or top of the character, whereas semantic radicals, those parts of Chinese characters that indicate something about the character’s meaning, tend to fall on the left or bottom of the character (for an example or both types of radicals, see Figure 2). Different researchers (Cheung & Chen, 2004; Shu, Anderson, & Wu, 2000) have sometimes asserted that knowledge of such rules in Chinese and knowledge of the radicals themselves in Chinese constitute the same process of orthographic knowledge, though others (e.g., Tong & McBride-Chang, 2010) see them as different processes, at least developmentally. 伙 /huǒ/ (partner)

災 / zāi / (disaster)

Figure 2.  The radical 火 (which reads as /hu /) falls on the right of the character 伙 and acts as a phonetic radical, indicating the sound /hu /. The radical 火 (which means fire) falls on the bottom of the character 災 and acts as a semantic radical, indicating that the character disaster is related to fire.

Visual-orthographic information is again an essential ingredient of word recognition across scripts. However, visual-orthographic information varies both in terms of how many units or configurations must be memorized and also in terms of the nature of these units (including graphemes and even grapheme components and diacritics). But the nature of the information differs both in degree (i.e., more or fewer visual skills are particularly important for learning to read for beginners). For example, young English readers can get easily confused in recognizing and producing letters such as b, d, p, or q because they are essentially comprised of the same shapes but in different configurations. With experience, readers across orthographies begin to internalize the rules of their particular script in order to enhance word recognition skills.

Learning to write across cultures Apart from word reading, an additional issue of interest is how writing of words develops across cultures. Writing is an even more involved process than word recognition, particularly for beginners. Just as implicit memory, a clear example of which might be recognizing the answer in a quiz (i.e., you select the correct answer from among limited choices) tends to be less effortful than producing an answer on one’s own, word recognition is less effortful than word writing. A current area of increasing research is on word writing. There have been excellent models of word writing for many years, at least as applied to English (for a review, see, e.g., Frith, 1980), but what particularly



Early literacy across languages 65

interests me currently is the question of how children learn to write words across cultures. Here, again, a basic cognitive focus extends to phonological processing, semantic skills, and visual-orthographic awareness. Visual-motor skills become pivotal in this process as well. Word writing and word reading can reinforce one another in the literacy learning process (e.g., Levin et al., 2013). In relation to word writing, the importance of phonological processing is perhaps clearest in the realm of invented spelling. From a very early age, children are able to write. They may begin with indecipherable scribbling and then move on to creating discernable letter formations, in alphabetic languages. Chinese children may write character-like representations very early. However, while children’s representations of words, or attempts at writing words, may be somewhat recognizable or clear in alphabetic orthographies (e.g., furst for first; techir for teacher), for Chinese, children do not seem to create new ways of writing words unless they really know what the character representations are. The phonological elements and advantages of alphabets in this respect are clear for alphabetic writing but not for Chinese (e.g., McBride-Chang, 2004). On the other hand, the semantic aspects of writing are inherent from the beginning in Chinese but not clear developmentally for a while in many alphabetic orthographies. Because most Chinese characters have a semantic radical as part of their composition, Chinese children must make use of this semantic aspect in writing. For example, if you give kindergartners in Hong Kong semantic and phonetic radicals and ask them to create pseudocharacters, which are not real characters but which have the structure of real characters, with these, Chinese children can do this (e.g., Tong & McBride, 2014). They understand this aspect of writing very early. Those learning an alphabetic orthography may not be aware of the semantic aspects of writing, particularly the fact that something about meaning makes it important to spell certain words certain ways, in the earliest years of writing. For example, it only makes sense to spell psychology as such when one understands that psyche, meaning breath or life, is a Greek root from which the English word derives. Otherwise, one might just as well spell it as sycology or some other more practical, phonological spelling. Children show this perhaps earlier by spelling shoes with an s rather than a z (which is the sound indicated in the word). Critics might argue that the semantic radicals in Chinese characters and the morphological indicators in English or other alphabetic languages are not parallel. I would agree but point out that this idea precisely underscores the importance of considering how or whether early literacy learning is similar across languages and orthographies. Visual-orthographic skills in relation to writing show interesting developmental patterns, and these patterns extend across a relatively long developmental time period. Even in early childhood, children show some patterns of writing that demonstrate sensitivity to certain writing patterns over others. For example, both American and Brazilian 4-year-old children produce patterns of writing that show evidence of statistical learning (Pollo, Kessler & Treiman, 2009). That is, letters that are more frequent in written texts tend to be reproduced more often by children. This is interesting because in English and Portuguese, respectively, different letters are more

66 Catherine McBride

prominent. Correspondingly, in both cultures, the more frequent letters of the culture appeared more often in children’s writing. In addition, certain patterns, including consonant-vowel bigrams (in which the second letter could be predicted by the first) appeared more frequently. Letters that were part of children’s own names also were more common in children’s writing. Chinese writing is complicated because different Chinese characters look so different and comprise many apparent visual patterns. Treiman and Yin (2011) showed that even very young (ages 2 to 6 years) Chinese children tended to produce different writing output when asked to write as compared to when they are asked to draw. Thus, children distinguish drawing from writing very early. However, how to demonstrate this depends upon the orthography one uses. The contrast between alphabetic orthographies and Chinese is just as striking for writing as it is for word recognition. While the unit of reading and writing for word recognition in alphabetic orthographies initially appears to be fairly uniformly at the letter level (rimes, like the “ight” for night, light, or fight in English are likely learned a bit later), the unit of representation for Chinese may not be so uniform. It could be at the radical or stroke level, or perhaps a mixture of the two (e.g., Lui, Leung, Law, & Fung, 2010). At older ages, there is also evidence of evolving visual-orthographic skills. Even adults occasionally forget how to write a particular word in English or Chinese. Often we rely on pondering, accompanied by some experimentation until the target “looks right.” English and Chinese may be among the best orthographies in which to demonstrate this phenomenon. English is among the most opaque of the alphabetic orthographies (Seymour et al., 2003), and Chinese by its very nature, notably its relative lack of reliance on phonological cues, is much more opaque still. Thus, in both scripts, writers must at least occasionally rely on non-phonological cues, letter or radical configurations that are re-imagined or re-understood as we struggle to reproduce them. Beyond the three key elements of phonological, semantic, and visual-orthographic cues that apply perhaps similarly to learning to read and to write a word, visual-motor skills emerge as important and perhaps unique for word writing. My own feeling about visual-motor applications is that writing in an alphabetic orthography involves letter sequencing, so the rote memorization of letter configurations is primary. In contrast, writing of Chinese requires, at least initially, attention to strokes and stroke orders for each character separately. Automatizing the writing process is fundamental for early literacy development. Children who are slow or who have difficulties in handwriting – variously categorized as aspects of dysgraphia sometimes, such as gripping the pencil too tightly, writing with too much force, gripping the pencil incorrectly, or difficulties in sequencing, among others (e.g., Overvelde & Hulstijn, 2011) – often though not always have difficulties with word recognition as well (e.g., Nicolson & Fawcett, 2011). This phenomenon extends across cultures and writing systems (e.g., Cheng-Lai, Li-Tsang, Chan, & Lo, 2012; Poon, Li-Tsang, Weiss, & Rosenblum, 2010; Rosenblum, Margieh, & Engel-Yeger, 2013).



Early literacy across languages 67

Conclusion Overall, the contrast between learning to read and write words from one orthography to another highlights some fundamental differences in the early development of literacy. Chinese children tend to master Chinese word recognition about 2 years later than do children learning to read English. In turn, English tends to be mastered about 1.5 to 2 years later than does reading in German, for example (Aro & Wimmer, 2003; Seymour et al., 2003). Such differences may at least partly explain why there are real age differences in when children begin formal reading instruction in school. For example, some European (non-English-speaking) countries tend to begin such instruction around the age of 6, whereas Hong Kong, at the other extreme, initiates such formal instruction around age 3.5 years. In most Chinese- and English-speaking societies, formal instruction begins around the age of 5. In addition, although every society now appears to accept the idea that specific reading difficulty is a problem, the method of testing for dyslexia differs depending upon orthography. For example, in English and Chinese, both timed and untimed word reading are tested, but in German, only timed reading is tested because the orthography is regular such that few if any errors in untimed word reading occur after age 7 or so (Landerl, Wimmer & Frith, 1997). One indicator of dyslexia in some languages such as English or Chinese is spelling. Spelling difficulties tend to continue into adulthood for dyslexics, even when their word recognition improves to the point of appearing fairly normal (Bruck, 1990; Bruck, 1993; Tops, Callens, Van Cauwenberghe, Adriaens, & Brysbaert, 2013). A consideration of how early word recognition and production is accomplished across orthographies is particularly important for building comprehensive models of literacy development and impairment, with practical implications. For example, there are dimensions of words that are potentially very important for learning to read in some orthographies and nonexistent in others. The consonant clusters that are so common in English, for example, make good phonemic awareness a strong facilitator of English word reading but not so much of Chinese, where there are no consonant clusters. Lexical tone is strongly related to reading of Chinese but nonexistent in English. Word boundaries must be taught to some extent in Thai and Chinese but not in English or German where they are demarcated on the paper clearly with spaces. A study of only a few orthographies limits our understanding of the dimensions relevant for reading that exist. We are still exploring and expanding what the relevant dimensions are in different scripts. Understanding these dimensions is important because once we understand them, we can test them in order to identify those who are at-risk for reading impairments. We can also explicitly teach some of these, and this explicit teaching typically facilitates children’s literacy learning. Indeed, one “universal” of word recognition is that making the implicit explicit in terms of rules of reading, even if the rules are imperfect, facilitates better reading. For example, teaching that different letters represent different individual sounds helps children to decode words and to write them reasonably phonologically accurately. In English, the consistency of the rules for which letters

68 Catherine McBride

sound which ways is not as strong as for Spanish or German but reasonably so. For example, the letter p usually makes the /p/ sound, but this is not the case in words such as psychology or elephant, so this “rule” of p and /p/ is good but not perfect. Similarly, there are Chinese characters in which radicals appear that can either cue something about the sound or meaning of the character, and these have various accuracy rates (e.g., Shu, Chen, Anderson, Wu, & Xuan, 2003). Parents sometimes worry that there is so much variability in Chinese that it is perhaps not worth it or fair to teach rules. However, rules help readers and writers to scaffold their learning. When a rule works, even if it only does so perhaps 30% of the time, it still cuts down strongly on memory load. A dyslexic child, for example, who has no “rule” for a character or a letter, has no choice but to memorize how to read or to write each individual word separately. All readers, but particularly those with reading difficulties, who sometimes have particular problems with phonological memory (e.g., Ho & Lai, 1999), learn better with rules. Whenever a teacher can teach children a rule that sometimes works and then teach them the instances in which it works first before teaching the exceptions, children benefit because they learn more quickly. In this chapter, I have argued that although the basic principles of word recognition are the same across scripts, the process of learning to read and to write words is different, to a certain extent. Differences have to do with the scripts used, languages mapped onto the scripts, and environments in which literacy is taught. Although the “triangle” of elements related to word reading, i.e., phonological, semantic, and orthographic information, is similar across all scripts, their manifestations differ. I emphasize the difference, rather than the similarity, mostly because I think such a focus might lead us to think in ways we would not otherwise. Such a focus might highlight the potential importance of suprasegmental processing for phonological processing, lexical compounding for semantics, or diacritics or other visual elements for orthographic processing, in ways we have not yet fully explored. We can then compare across scripts to discover new and potentially important aspects of literacy development relevant to teachers, parents, educators, and literacy learners ourselves, across scripts.

References Aram, D., & Levin, I. (2004). The role of maternal mediation of writing to kindergartners in promoting literacy in school: A longitudinal perspective. Reading and Writing, 17, 387– 409.  doi: 10.1023/B:READ.0000032665.14437.e0 Aro, M., & Wimmer, H. (2003). Learning to read: English in comparison to six more regular orthographies. Applied Psycholinguistics, 24, 621–635.  doi: 10.1017/S0142716403000316 Bowers, P. N., Kirby, J. R., & Deacon, S. H. (2010). The effects of morphological instruction on literacy skills: A systematic review of the literature. Review of Educational Research, 80, 144–179.  doi: 10.3102/0034654309359353



Early literacy across languages 69

Brown, A. L. (1975). Recognition, reconstruction, and recall of narrative sequences by preoperational children. Child Development, 46, 156–166.  doi: 10.2307/1128844 Bruck, M. (1990). Word-recognition skills of adults with childhood diagnoses of dyslexia. Developmental Psychology, 26, 439–454  doi: 10.1037/0012-1649.26.3.439 Bruck, M. (1993). Component spelling skills of college students with childhood diagnoses of dyslexia. Learning Disability Quarterly, 16, 171–184.  doi: 10.2307/1511325 Bruck, M., Treiman, R., Caravolas, M., Genesee, F., & Cassar, M. (1998). Spelling skills of children in whole language and phonics classrooms. Applied Psycholinguistics, 19, 669–684. doi: 10.1017/S0142716400010419 Byrne, B., Freebody, P., & Gates, A. (1992). Longitudinal data on the relations of word-reading strategies to comprehension, reading time, and phonemic awareness. Reading Research Quarterly, 27, 141–151.  doi: 10.2307/747683 Calet, N., Gutiérrez-Palma, N., Simpson, I. C., González-Trujillo, M. C., & Defior, S. (2015). Suprasegmental phonology development and reading acquisition: A longitudinal study. Scientific Studies of Reading, 19, 51–71.  doi: 10.1080/10888438.2014.976342 Carlisle, J. F. (1995). Morphological awareness and early reading achievement. In L. B. Feldman (Ed.), Morphological aspects of language processing (pp. 189–209). Hillsdale, NJ: Erlbaum. Carlisle, J. F. (2000). Awareness of the structure and meaning or morphologically complex words: Impact on reading. Reading and Writing, 12, 169–190.  doi: 10.1023/A:1008131926604 Casalis, S., & Louis-Alexandre, M. F. (2000). Morphological analysis, phonological analysis and learning to read French: A longitudinal study. Reading and Writing, 12, 303–335. doi: 10.1023/A:1008177205648 Cheng-Lai, A., Li-Tsang, C. W., Chan, A. H., & Lo, A. G. (2013). Writing to dictation and handwriting performance among Chinese children with dyslexia: Relationships with orthographic knowledge and perceptual-motor skills. Research in Developmental Disabilities, 34, 3372–3383.  doi: 10.1016/j.ridd.2013.06.039 Cheung, H., & Chen, H. C. (2004). Early orthographic experience modifies both phonological awareness and online speech processing. Language and Cognitive Processes, 19, 1–28. doi: 10.1080/01690960344000071 Cheung, H., Chung, K. K., Wong, S. W., McBride‐Chang, C., Penney, T. B., & Ho, C. S. (2009). Perception of tone and aspiration contrasts in Chinese children with dyslexia. Journal of Child Psychology and Psychiatry, 50, 726–733.  doi: 10.1111/j.1469-7610.2008.02001.x Chiappe, P., Siegel, L. S., & Wade-Woolley, L. (2002). Linguistic diversity and the development of reading skills: A longitudinal study. Scientific Studies of Reading, 6, 369–400. doi: 10.1207/S1532799XSSR0604_04 Chiu, M. M., & McBride-Chang, C. (2006). Gender, context, and reading: A comparison of students in 43 countries. Scientific Studies of Reading, 10, 331–362.  doi: 10.1207/s1532799xssr1004_1 Chiu, M. M., McBride-Chang, C., & Lin, D. (2012). Ecological, psychological, and cognitive components of reading difficulties: Testing the component model of reading in fourth graders across 38 countries. Journal of Learning Disabilities, 45, 391–405.  doi: 10.1177/0022219411431241 Dallasheh-Khatib, R., Ibrahim, R., & Karni, A. (2014). Longitudinal data on the relations of morphological and phonological training to reading acquisition in first grade: The case of Arabic language. Psychology, 5, 918–940.  doi: 10.4236/psych.2014.58103 Deacon, S. H., & Kirby, J. R. (2004). Morphological awareness: Just “more phonological”? The roles of morphological and phonological awareness in reading development. Applied Psycholinguistics, 25, 223–238.  doi: 10.1017/S0142716404001110

70 Catherine McBride

Deacon, S. H., Wade-Woolley, L., & Kirby, J. (2007). Crossover: The role of morphological awareness in French immersion children’s reading. Developmental Psychology, 43, 732–46. doi: 10.1037/0012-1649.43.3.732 Fletcher, J. M., Shaywitz, S. E., Shankweiler, D. P., Katz, L., Liberman, I. Y., Stuebing, K. K., … & Shaywitz, B. A. (1994). Cognitive profiles of reading disability: Comparisons of discrepancy and low achievement definitions. Journal of Educational Psychology, 86, 6–23. doi: 10.1037/0022-0663.86.1.6 Frith, U. (Ed.) (1980). Cognitive processes in spelling. Sydney: Academic Press. Harris, J. W. (1983). Syllable Structure and Stress in Spanish: A Nonlinear Analysis. Studies in Second Language Acquisition, 7, 123–124. Ho, C. S. H., & Bryant, P. (1997). Phonological skills are important in learning to read Chinese. Developmental Psychology, 33, 946–951.  doi: 10.1037/0012-1649.33.6.946 Ho, C. S. H., & Lai, D. N. C. (1999). Naming-speed deficits and phonological memory deficits in Chinese developmental dyslexia. Learning and Individual Differences, 11, 173–186. doi: 10.1016/S1041-6080(00)80004-7 Ho, C., Leung, M. T., & Cheung, H. (2011). Early difficulties of Chinese preschoolers at familial risk for dyslexia: deficits in oral language, phonological processing skills, and print‐related skills. Dyslexia, 17, 143–164.  doi: 10.1002/dys.429 Holliman, A. J., Wood, C., & Sheehy, K. (2008). Sensitivity to speech rhythm explains individual differences in reading ability independently of phonological awareness. British Journal of Developmental Psychology, 26, 357–367.  doi: 10.1348/026151007X241623 Ku, Y.-M., & Anderson, R. C. (2003). Development of morphological awareness in Chinese and English. Reading and Writing: An Interdisciplinary Journal, 16, 399–422. doi: 10.1023/A:1024227231216 Landerl, K., Wimmer, H., & Frith, U. (1997). The impact of orthographic consistency on dyslexia: A German-English comparison. Cognition, 63, 315–334.  doi: 10.1016/S0010-0277(97)00005-X LeFevre, J. A. (2001). Storybook reading and parent teaching: Links to language and literacy development. New Directions for Child and Adolescent Development, 2001, 39–52. doi: 10.1002/cd.14 Levin, I., & Aram, D. (2012). Mother – child joint writing and storybook reading and their effects on kindergartners’ literacy: An intervention study. Reading and Writing, 25, 217–249. doi: 10.1007/s11145-010-9254-y Levin, I., Aram, D., Tolchinsky, L., & McBride, C. (2013). Maternal mediation of writing and children’s early spelling and reading: The Semitic abjad versus the European alphabet. Writing Systems Research, 5, 134–155.  doi: 10.1080/17586801.2013.797335 Li, T., & McBride-Chang, C. (2014). How character reading can be different from word reading in Chinese and why it matters for Chinese reading development. In X. Chen, Q. Wang & Y. C. Luo (Eds.), Reading Development and Difficulties in Monolingual and Bilingual Chinese Children (pp. 49–65). Dordrecht, Netherlands: Springer.  doi: 10.1007/978-94-007-7380-6_3 Li, H., & Rao, N. (2000). Parental influences on Chinese literacy development: A comparison of preschoolers in Beijing, Hong Kong, and Singapore. International Journal of Behavioral Development, 24, 82–90.  doi: 10.1080/016502500383502 Lin, C. Y. C., & Fu, V. R. (1990). A comparison of child‐rearing practices among Chinese, immigrant Chinese, and Caucasian‐American parents. Child Development, 61, 429–433. doi: 10.2307/1131104



Early literacy across languages 71

Lin, D., McBride-Chang, C., Aram, D., Shu, H., Levin, I., & Cho, J. R. (2012). Maternal mediation of word writing in Chinese across Hong Kong and Beijing. Journal of Educational Psychology, 104, 121–137.  doi: 10.1037/a0025383 Lui, H. M., Leung, M. T., Law, S. P., & Fung, R. S. Y. (2010). A database for investigating the logographeme as a basic unit of writing Chinese. International Journal of Speech-Language Pathology, 12, 8–18.  doi: 10.3109/17549500903203082 Lyster, S. A. H. (2002). The effects of morphological versus phonological awareness training in kindergarten on reading development. Reading and Writing, 15, 261–294. doi: 10.1023/A:1015272516220 Lyytinen, H. (1997). In search of the precursors of dyslexia: A prospective study of children at risk for reading problems. In C. Hulme & M. Snowling (Eds.), Dyslexia: Biology, cognition and intervention (pp. 97–107). London: Whurr. McBride-Chang, C. (2004). Children’s Literacy Development. London: Edward Arnold/Oxford Press. McBride-Chang, C., Chen, H. C., Kasisopa, B., Burnham, D., Reilly, R., & Leppänen, P. (2012). What and where is the word? Behavioral and Brain Sciences, 35, 295–296. doi: 10.1017/S0140525X1200009X McBride-Chang, C., Cho, J.-R., Liu, H., Wagner, R. K., Shu, H., Zhou, A., Cheuk, C. S. -M., & Muse, A. (2005). Changing models across cultures: Associations of phonological and morphological awareness to reading in Beijing, Hong Kong, Korea, and America. Journal of Experimental Child Psychology, 92, 140–160.  doi: 10.1016/j.jecp.2005.03.009 McBride-Chang, C., Lam, F., Lam, C., Doo, S., Wong, S. W. L., & Chow, Y. Y. Y. (2008). Word recognition and cognitive profiles of Chinese pre-school children at-risk for dyslexia through language delay or familial history of dyslexia. Journal of Child Psychology and Psychiatry, 49, 211–218. McBride-Chang, C., Shu, H., Zhou, A., Wat, C.-P., & Wagner, R. K. (2003). Morphological awareness uniquely predicts young children’s Chinese character recognition. Journal of Educational Psychology, 95, 743–751.  doi: 10.1037/0022-0663.95.4.743 McBride-Chang, C., Tong, X., Shu, H., Wong, A. M. Y., Leung, K. W., & Tardif, T. (2008). Syllable, phoneme, and tone: Psycholinguistic units in early Chinese and English word recognition. Scientific Studies of Reading, 12, 171–194.  doi: 10.1080/10888430801917290 Nag, S. (2007). Early reading in Kannada: The pace of acquisition of orthographic knowledge and phonemic awareness. Journal of Research in Reading, 30, 7–22. doi: 10.1111/j.1467-9817.2006.00329.x Nezworski, T., Stein, N. L., & Trabasso, T. (1982). Story structure versus content in children’s recall. Journal of Verbal Learning and Verbal Behavior, 21, 196–206. doi: 10.1016/S0022-5371(82)90561-8 Nicolson, R. I., & Fawcett, A. J. (2011). Dyslexia, dysgraphia, procedural learning and the cerebellum. Cortex, 47, 117–127.  doi: 10.1016/j.cortex.2009.08.016 Nunes, T., Bryant, P., & Bindman, M. (1997). Morphological spelling strategies: developmental stages and processes. Developmental psychology, 33, 637–649.  doi: 10.1037/0012-1649.33.4.637 Ortiz, R. W. (2004). Hispanic/Latino fathers and children’s literacy development: Examining involvement practices from a sociocultural context. Journal of Latinos and Education, 3, 165–180.  doi: 10.1207/s1532771xjle0303_3

72

Catherine McBride

Overvelde, A., & Hulstijn, W. (2011). Handwriting development in grade 2 and grade 3 primary school children with normal, at risk, or dysgraphic characteristics. Research in Developmental Disabilities, 32, 540–548.  doi: 10.1016/j.ridd.2010.12.027 Packard, J. L. (2000). The morphology of Chinese: A linguistic and cognitive approach. Cambridge, United Kingdom: Cambridge University Press.  doi: 10.1017/CBO9780511486821 Pan, J., McBride-Chang, C., Shu, H., Liu, H., Zhang, Y., & Li, H. (2011). What is in the naming? A 5-year longitudinal study of early rapid naming and phonological sensitivity in relation to subsequent reading skills in both native Chinese and English as a second language. Journal of Educational Psychology,103, 897–908.  doi: 10.1037/a0024344 Park, H.-A. (2000). The development of phonological awareness in children. Korean Journal of Child Studies, 21, 35–44. Pollo, T. C., Kessler, B., & Treiman, R. (2009). Statistical patterns in children’s early writing. Journal of Experimental Child Psychology, 104, 410–426.  doi: 10.1016/j.jecp.2009.07.003 Poon, K. W., Li-Tsang, C. W. P., Weiss, T. P. L., & Rosenblum, S. (2010). The effect of a computerized visual perception and visual-motor integration training program on improving Chinese handwriting of children with handwriting difficulties. Research in Developmental Disabilities, 31, 1552–1560.  doi: 10.1016/j.ridd.2010.06.001 Richardson, U., & Lyytinen, H. (2014). The GraphoGame method: The theoretical and methodological background of the technology-enhanced learning environment for learning to read. Human Technology: An Interdisciplinary Journal on Humans in ICT Environments, 10, 39–60.  doi: 10.17011/ht/urn.201405281859 Rispens, J., McBride-Chang, C., & Reitsma, P. (2008). Morphological awareness and early and advanced word recognition and spelling in Dutch. Reading and Writing, 21, 587–607. doi: 10.1007/s11145-007-9077-7 Roberts, T. A. (2008). Home storybook reading in primary or second language with preschool children: Evidence of equal effectiveness for second‐language vocabulary acquisition. Reading Research Quarterly, 43, 103–130.  doi: 10.1598/RRQ.43.2.1 Rosenblum, S., Margieh, J. A., & Engel-Yeger, B. (2013). Handwriting features of children with developmental coordination disorder – Results of triangular evaluation. Research in Developmental Disabilities, 34, 4134–4141.  doi: 10.1016/j.ridd.2013.08.009 Seidenberg, M. S., & McClelland, J. L. (1989). A distributed, developmental model of word recognition and naming. Psychological Review, 96, 523–68.  doi: 10.1037/0033-295X.96.4.523 Seymour, P. H., Aro, M., & Erskine, J. M. (2003). Foundation literacy acquisition in European orthographies. British Journal of Psychology, 94, 143–174.  doi: 10.1348/000712603321661859 Share, D. L. (2008). On the Anglocentricities of current reading research and practice: the perils of overreliance on an “outlier” orthography. Psychological Bulletin, 134, 584–615. doi: 10.1037/0033-2909.134.4.584 Shu, H., Anderson, R. C., & Wu, N. (2000). Phonetic awareness: Knowledge of orthography – phonology relationships in the character acquisition of Chinese children. Journal of Educational Psychology, 92, 56–62.  doi: 10.1037/0022-0663.92.1.56 Shu, H., Chen, X., Anderson, R. C., Wu, N., & Xuan, Y. (2003). Properties of school Chinese: Implications for learning to read. Child Development, 74, 27–47.  doi: 10.1111/1467-8624.00519 Snow, C. E., & Ninio, A. (1986). The contracts of literacy: What children learn from learning to read books. In W. H. Teale & E. Sulzby (Eds.), Emergent literacy: Writing and reading (pp. 116–138). Norwood, NJ: Ablex.



Early literacy across languages 73

Stanovich, K. E., & Siegel, L. S. (1994). Phenotypic performance profile of children with reading disabilities: A regression-based test of the phonological-core variable-difference model. Journal of Educational Psychology, 86, 24–53.  doi: 10.1037/0022-0663.86.1.24 Tong, X., & McBride-Chang, C. (2010). Developmental models of learning to read Chinese words. Developmental Psychology, 46, 1662–1676.  doi: 10.1037/a0020611 Tong, X., & McBride, C. (2014). Chinese children’s statistical learning of orthographic regularities: Positional constraints and character structure. Scientific Studies of Reading, 18, 291–308.  doi: 10.1080/10888438.2014.884098 Tong, X., McBride-Chang, C., Wong, A. M. -Y., Shu, H., Reitsma, P., & Rispens, J. (2011). Longitudinal predictors of very early Chinese literacy acquisition. Journal of Research in Reading, 34, 315–332.  doi: 10.1111/j.1467-9817.2009.01426.x Tops, W., Callens, M., Van Cauwenberghe, E., Adriaens, J., & Brysbaert, M. (2013). Beyond spelling: The writing skills of students with dyslexia in higher education. Reading and Writing, 26, 705–720.  doi: 10.1007/s11145-012-9387-2 Torgesen, J. K., Wagner, R. K., & Rashotte, C. A. (1994). Longitudinal studies of phonological processing and reading. Journal of Learning Disabilities, 27, 276–286.  doi: 10.1177/002221949402700503 Treiman, R., & Yin, L. (2011). Early differentiation between drawing and writing in Chinese children. Journal of Experimental Child Psychology, 108, 786–801.  doi: 10.1016/j.jecp.2010.08.013 Tse, S. K., Marton, F., Ki, W. W., & Loh, E. K. Y. (2007). An integrative perceptual approach for teaching Chinese characters. Instructional Science, 35, 375–406.  doi: 10.1007/s11251-006-9011-4 van Goch, M. M., McQueen, J. M., & Verhoeven, L. (2014). Learning phonologically specific new words fosters rhyme awareness in Dutch preliterate children. Scientific Studies of Reading, 18, 155–172.  doi: 10.1080/10888438.2013.827199 Vellutino, F. R., & Scanlon, D. M. (1987). Phonological coding, phonological awareness, and reading ability: Evidence from a longitudinal and experimental study. Merrill-Palmer Quarterly, 33, 321–363. Wang, M. & Arciuli, J. (2015). Phonology beyond phonemes: Contribution of suprasegmental information to reading. Scientific Studies in Reading, 19, 1–4. doi: 10.1080/10888438.2014.976790 Wang, Y., & McBride, C. (2016). Character reading and word reading in Chinese: Unique correlates for Chinese kindergarteners. Applied Psycholinguistics. 37, 371–386. Wang, Y., McBride-Chang, C., & Chan, S. F. (2013). Correlates of Chinese kindergarteners’ word reading and writing: the unique role of copying skills? Reading and Writing, 27, 1281–1302. doi: 10.1007/s11145-013-9486-8 Wong, A. M. -Y., Ciocca, V., & Yung, S. (2009). The perception of lexical tone contrasts in Cantonese children with and without Specific Language Impairment (SLI). Journal of Speech, Language and Hearing Research, 52, 1493–1509.  doi: 10.1044/1092-4388(2009/08-0170) Wood, C. (2006). Metrical stress sensitivity in young children and its relationship to phonological awareness and reading. Journal of Research in Reading, 29, 270–287. doi: 10.1111/j.1467-9817.2006.00308.x Wood, C., Wade-Woolley, L., & Holliman, A. J. (2009). Phonological awareness: Beyond phonemes. In C. Wood & V. Connelly (Eds.), Contemporary perspectives on reading and spelling (pp.7–23). London: Routledge. Wu, J. (1999). Syntax and semantics of quantification in Chinese (Doctoral dissertation). Retrieved from department of Linguistics, University of Maryland.

74

Catherine McBride

Wu, X., Anderson, R. C., Li, W., Wu, X., Li, H., Zhang, J., … & Gaffney, J. S. (2009). Morphological awareness and Chinese children’s literacy development: An intervention study. Scientific Studies of Reading, 13, 26–52.  doi: 10.1080/10888430802631734 Yoon, H.-K., Bolger, D. J., Kwon, O.-S., & Perfetti, C. A. (2002). Subsyllabic units in reading: A difference between Korean and English. In L. Verhoeven, C. Elbro, & P. Reitsma (Eds.), Precursors of functional literacy (pp.139–163). Amsterdam: John Benjamins. doi: 10.1075/swll.11.12yoo Zhang, J., & McBride-Chang, C. (2010). Auditory sensitivity, speech perception, and reading development and impairment. Educational Psychology Review, 22, 323–338. doi: 10.1007/s10648-010-9137-4 Zhou, Y. L., McBride-Chang, C., Fong, C. Y. -C., Wong, T. T. -Y., & Cheung. S. K. (2012). A comparison of phonological awareness, lexical compounding, and homophone training for Chinese word reading in Hong Kong kindergarteners. Early Education and Development, 23, 475–492.  doi: 10.1080/10409289.2010.530478 Ziegler, J. C., & Goswami, U. (2005). Reading acquisition, developmental dyslexia, and skilled reading across languages: a psycholinguistic grain size theory. Psychological Bulletin, 131, 3–29.  doi: 10.1037/0033-2909.131.1.3

Learning to read alphasyllabaries Sonali Nag

The Promise Foundation and University of Oxford

The orthographies of several languages across South and Southeast Asia have been collectively called the Indic alphasyllabaries. These orthographies trace their roots back to the ancestral script of Brahmi and hence are also called Brahmi-derived or Brahmic scripts. Individual orthographies in the Indic family are distinct in appearance whilst sharing core principles of architecture and representation. The Indic alphasyllabaries are in use for languages that vary in phonological and morpho-phonological characteristics; a small set of these languages are tonal languages with an additional register of tone markers in the orthography. The direction of reading is left to right and texts are laid out with spacing between words, though there are exceptions such as the absence of inter-word spacing in Thai. Traditionally, longer texts included a single or double perpendicular line to denote ends of sentences, couplets, and paragraphs, but contemporary writing uses the entire register of punctuation marks from European scripts. The focus of this chapter is on the architecture of the Indic alphasyllabaries, 1 the cognitive-linguistic processes that underpin learning to read in this writing system, and on some implications for model building and theorizing about writing systems that research on the Indic alphasyllabaries raise. The first section introduces the writing system, followed by a section on symbol learning, reading and phonological processing. A number of developmental studies are available for reading accuracy, but there is far less research on the other component skills of reading fluency and reading comprehension. This skew will be evident in the following review. A third section covers models that have emerged from akshara-focused research related to inventory size, orthographic learning and skilled reading. A point to note is that this review is focused on the Indic alphasyllabaries. There is one other instance of a similar writing system – the Ethiopic alphasyllabary – which currently has a very small cognitive-linguistic research base. 1. The term alphasyllabary has been challenged by Peter Daniels and colleagues as inappropriate because it gives the notion of a hybrid script with part solutions from an alphabet system and a syllabary. William Bright who coined the term and Salomon who went on to use it, however, explicitly strove to avoid such a misconception (Bright, 1992; Bright, 2000; Salomon, 2000). If there should be a departure from the popular term of Indic alphasyllabary then a name that presents itself is the akshara writing system. The term has wider appeal than suggestions such as ‘abugida’ taken from the first symbols of the Ethiopic alphasyllabary (Daniels, 1990) and ‘arepiconu’ from symbols of the other ancient Indic script, Kharoshthi (Share & Daniels, 2015). doi 10.1075/swll.15.05nag © 2017 John Benjamins Publishing Company

76

Sonali Nag

Investigating languages that use the Indic writing system has far reaching social significance. Country profiles developed by the UNESCO Institute for Statistics highlights that about 150 million learners are in the primary school age population of South Asia alone. Added to this are primary school learners in Southeast Asia, older learners of South and Southeast Asia, the diaspora from these countries, and other learners whose interests range from adding an Asian language to their professional toolkit to a wish to connect with the cultural heritage of the region. The research collated in this chapter covers native speakers and second language learners, children learning in one, two or three scripts, and adults with varied levels of proficiency. The samples are from eight South and Southeast Asian languages, covering early, primary and middle school years, and mature readers who are typically university graduates, but also older adults. Such variations on important parameters confound attempts at synthesizing the literature with an eye to articulating a causal model of literacy development. At the same time, the variety is excellent for drawing out universals across alphasyllabaries as well as orthography-specific phenomenon.

Principles of the writing system The orthographic unit in the Indic alphasyllabaries is known by the generic term akshara. In conventional reading instruction, the symbol system is presented roughly in the following order: vowels (V), consonants with an inherent vowel (Ca), consonants with other vowels (CV), consonant clusters with a vowel (e.g. CCa, CCCa, CCV, CCCV), and consonants (C), (for an exception in order of instruction consider Tamil). Among these, the inherent vowel and the consonant cluster are defining traits of the writing system. The inherent vowel, as the term implies, is a vowel that is implicit and left unmarked in the orthography. This vowel is realized as a mid-central [ə], open-central [a] or open-mid back rounded [ɔ] vowel in different languages. All other vowels have dedicated vowel markers that are appended to consonants. A consonant is assumed to carry the inherent vowel unless assigned a phonemic status (e.g., the chillaksharam of Malayalam), nullified by another vowel marker (e.g., the mātrā 2 of Hindi, gunitā of Kannada), or ‘killed’ by an explicit suppression marker (e.g. the virāma in Sanskrit, halanth in Hindi, the hal lakuna / hal kirīma in Sinhala). The consonant cluster, also called a conjoint cluster, carries either an inherent or marked vowel. All permutations of consonant clusters and consonant-vowel combinations are orthographically possible. Only a sub-set of this repertoire is, however, phonotactically valid and encountered in print. Even so, a conservative estimate of the more and less frequently used akshara in a language runs into hundreds. The size of the symbol inventory therefore defines the pace of learning the symbol set and symbol knowledge 2. The International Alphabet of Sanskrit Transliteration (I.A.S.T.), and its later version, the National Library at Kolkata Romanization, has been used throughout this chapter. This scheme is useful for a Romanized transliteration of Indic languages.



Learning to read alphasyllabaries 77

is a determinant of literacy outcomes in early and more advanced stages of literacy development. Among expert readers and spellers, for example, visual word recognition and spelling may remain low for words with low frequency akshara. A supplementary point to note is that the component phonemic markers in an akshara allow for learning about combinatorial principles. This facility is clearly more economical than learning the several hundred akshara one by one by one as global, undifferentiated units. A good way to approach the complexities in the akshara writing system is to understand that there are some surface characteristics that are easy-to-see, and there are other principles that operate at a deeper linguistic level, and are not immediately obvious. Learning to read in this writing system therefore requires attention to both these layers of representation of the spoken language in writing. An introduction to the writing system and implications for learning is given in Table 1. The rest of this section gives details and illustrations about the writing system and the following sections cover the implications for learning. Table 1.  A shortlist of some obvious and less-obvious characteristics of the akshara writing system and what they mean for learning to read Characteristics of the akshara writing systema Obvious and visible 1 An akshara can represent consonants, vowels, consonantvowel pairs and consonant clusters with vowels. In languages like Thai and Lao the akshara can also include a tone marker. 2 Most akshara are constructed by joining individual markers. One important exception is the inherent vowel, which has no marker. 3 Individual markers have a designated place in the akshara. The location may be non-linear (e.g., before another marker) but this location is entirely predictable. Less obvious and less visible 1 A consonant that comes after a vowel becomes a new akshara (e.g., two-akshara words: , . ).b 2 The akshara within words map to phonology at different levels (e.g., /taa.kaa/ is a word with two open syllables and there is one akshara for each syllable; /kiip/ is a word with a single closed syllable and there is one akshara each for the body and coda segment of the syllable). 3 If there is more than one way of representing codas in a language, the one to use for a specific word will depend on linguistic rules (e.g., rules about etymology of loan words, legal endings of words, and syllable weight).

Implications for learning to read Many symbols have to be recognized. Since symbols are constructed using systematic combinatorial rules, it is efficient to quickly learn these rules.

Visual word recognition sometimes requires more than straight forward sound-symbol decoding. Since linguistic rules underpin the written word form, it is efficient to widen lexical knowledge.

Note a. There may be exceptions to each of these characteristics and there are most certainly differences in type and token frequencies between akshara-based orthographies. b. These are names of currency in Bangladesh, Thailand and Lao respectively.

78

Sonali Nag

Several architectural features characterize the surface organization of Indic symbol sets. First, many akshara are two or more phonemic markers put together to form orthographic syllables. This is a case of mixed granularity because the unit of representation is available in more than one grain size within the same symbol system (e.g., the syllable /ṭu/ in Bengali: ট্ + ◌ু = টু , Hindi: ट् +◌ु = टु, Gurmukhi: ਟ੍ + ◌ੁ = ਟੁ, Kannada: ಟ್  + ◌ು = ಟು and Sinhala: ට් + ◌ු = ටු). Second, two representations are available for most individual phonemes. The primary forms are of full size and written in-line. The secondary form is either another visually distinct unit or a fraction or miniature of the primary form, and seen in a subordinate and sometimes off-the-line position. The use of the appropriate form depends on location within a word, syllable, or syllable string. Among vowels, for example, the primary form is reserved for word initial vowels and morphemes, and the secondary form for the post-consonantal vowel. The inherent vowel is one obvious exception to this positional rule for vowel representation since it is left unmarked in the orthography. A second exception is a morpheme vowel, for example, the Bengali emphatic particles ই (-i) and ও (-o) (e.g., সেই ‘he himself ’; সেও ‘he also’). There are also language-specific exceptions such as the Hindi आ (ā) in कौआ ‘crow’ and कछु आ ‘tortoise’ where the vowel is neither word-initial nor a morpheme. A third architectural feature is the predictable nature of where the markers join within an akshara, although some non-standard ligaturing solutions also exist (e.g., Tamil: த் + /u/ = து, but ப் + /u/ = பு). The fourth feature is that the arrangement of markers may be non-linear. Off-the-line location of markers is one example of non-linearity, evident in the illustrations above. Sequence mismatch is a second type of non-linearity. For example, the symbol block in Kannada, Sinhala and Telugu orthographies represents the phonological sequence /C1C2V/. When viewed on a linear axis, the arrangement is obviously a mis-sequence of the phonological string (and hence non-linear). The ratio of linear : non-linear block arrangements differs across orthographies. Kannada, for example, has fewer linearised arrangements than Hindi, Tamil, Malayalam and Bengali, with more dramatic nonlinearities seen in the heritage scripts of Philippines and Indonesia. Such variety suggests that the nature and weight for perceptual processing would differ in orthography-specific accounts of akshara learning. We turn next to the less-visible rules in the writing system. Akshara-based writing proceeds on a principle of complete phonemic transcription. Several nuances to this general rule of phonological representation need highlighting. The post-vocalic consonants are always represented by a new akshara. In a closed syllable, for example, the body is the first akshara and the coda the next akshara (e.g., Hindi words /āj/ and /kal/, ‘today’ and ‘tomorrow’, are written with two akshara as and ). Such coda representation as a new akshara implies a re-syllabification at the written level. In other words, while phonological syllables and orthographic syllables (the akshara) are isomorphic for CVCV words (e.g., /yo.ga/ and ), this is not the case for words with closed syllables. When the adjacent unit is an open syllable, then representation is carried whole within one akshara block (e.g., /kār.ya / is written as ‘work’), but if it is a closed syllable, then a third akshara is formed for the last coda (e.g.,



Learning to read alphasyllabaries 79

/īsh.war/ ‘God’ is ). In this way, orthographic re-syllabification moves down the phonological string. Further, there are orthography-specific differences in coda representation. A prominent example is the phenomenon of schwa suppression seen in several languages including Hindi, Bengali and Marathi. In schwa suppression, the coda is represented by a consonant with inherent vowel but the inherent vowel is suppressed in reading (e.g., /CVC/ as as rendered in and above). An example of schwa suppression in bi-syllabic words is when /CVC.CV/ is written as . The various solutions for coda representation introduce ambiguity where mere assembling of sounds/symbols does not suffice as a word decoding strategy. Lexical knowledge related to phonototactics and morphophonology is instead critical, as is an appreciation of etymology. In Thai, for example, certain phonemes are not allowed in word-final positions and there is systematic substitution with legal word-final phonemes. Or take the Bengali lexicon where native words and Sanskrit borrowings do not contain complex codas while borrowings from other languages do (Kar, 2009). Since coda solutions differ across languages, one further corollary is that there would be differences also in the extent to which broader linguistic awareness improves visual word recognition and spelling accuracy. The akshara-phonology mapping principle extends over multiple levels of phonology. While the singleton akshara map on to orthographic syllables and the component markers are recognizable as phonemic representations, the rules of re-syllabification lead to the mapping of the akshara to the body of a syllable, a coda-body combination, or a coda-open syllable concatenation. All of these instances of akshara-phonology Table 2.  Frequency estimations of non-linear arrangements and re-syllabification features in a selection of akshara orthographies Language Bengali Hindi Gujarati Kannada Malayalam

Non-linear arrangements

Re-syllabification of /CVC.CV/a, b

on CVs

on CCVs Other

as as as “villa”, car > “Ford”) would be much more common. Instead, common misreadings almost always resemble the correct sound (e.g. house > “horse”, car > “can”). See Ehri (1992) for a more detailed discussion.

182 Carsten Elbro and Peter F. de Jong

Secondly, whilst beginning readers recode words letter by letter, skilled readers may also be able to recode several letters in parallel. This has been suggested in connectionist models of decoding (e.g., Harm & Seidenberg, 1999, 2004), but sparsely researched in human readers. Recently, de Jong and colleagues (de Jong, 2011; van den Boer & de Jong, 2015; van den Boer, Georgiou, & de Jong, 2016) showed that in skilled readers the reading speed of high frequency words is highly related to the speed of the rapid naming of alphanumeric symbols (e.g., 3, 7) – if the stimuli in both tasks are presented in an isolated (discrete) format. De Jong and colleagues reasoned that the naming of isolated alphanumeric symbols is a pure measure for the retrieval of a pronunciation from memory. If such a measure is highly related to isolated word reading, then the pronunciation of these words is most likely also retrieved from memory. In contrast, the relationship of serial naming with isolated word reading was much weaker in skilled readers. Surprisingly, van den Boer and de Jong (2015) found the same relationships with isolated nonword reading: isolated rapid naming was more strongly related to nonword reading than was serial rapid naming. Evidently, the pronunciation of a nonword is not available in memory. Instead, van den Boer and de Jong (2015) suggested that the ability to generate (or activate) phonology in parallel from print should be separated from the availability of orthographic knowledge. In sum, skilled readers may be able to activate a string of phonemes simultaneously when presented with a written word (a string of letters). In order to recognise a word as a whole, all that is needed is that this string of phonemes activate an existing phonological representation in the mental lexicon (Figure 2). However, this is also the point where we have to turn to speculation rather than point to evidence. There are at least two ways to skip the stepping stones of spelling pronunciations. Either they can be jumped over, or they can become part of the firm ground (the mental lexicon). That is, with growing reading experience, spelling pronunciations may either become superfluous because phonological recoding is tuned to provide the standard pronunciation, or spelling pronunciations become so well integrated into the mental lexicon that activating them is lexical activation (i.e., word recognition). Both options are possible in Figure 2. Let us briefly consider these two options. The first option is the more standard one that orthographic code learning closes the gap between spelling pronunciations and the standard pronunciation of words. With growing reading experience, the outcome of the simultaneous recoding – the spelling pronunciations – may gradually approach existing phonological representations (see Perfetti, 1992, for a similar proposal). This can happen for specific words, but readers will also gradually learn conditional letter pronunciations, that is, pronunciations that depend on the neighbouring letters. For example, frequently occurring letter combinations – such as complex graphemes oo, ng, ea, sh, etc., and very frequent morphemes -ing, -ed, in-, etc.) – are treated as orthographic units (e.g., Elbro, 2005; McGuinness, 1998). Such letter patterns can be considered an extension of the alphabet, and their pronunciation follows the extended code. As the “window” into orthographic patterns is gradually widened from single letters to whole words, the accuracy of recoding will increase and ultimately reach perfection. In other words, the recoded phonological



Orthographic learning is verbal learning 183

output – the spelling pronunciations generated by parallel recoding – will gradually approach standard pronunciations. However, long before the extended code is mastered, frequently seen words will be recognised as wholes. This simply means that their unique letter sequence is associated with the unique sound sequences that are represented in the mental lexicon. This view fits with the learning part of connectionist models of reading (e.g., Harm & Seidenberg, 1999) and with the end state of localist models assuming a connection between an orthographic and a phonological lexicon (e.g., Coltheart, Rastle, Perry, Langdon, & Ziegler, 2001). The second option – which does not exclude the first – is that learning spelling pronunciations closes the gap between spelling pronunciations and standard pronunciations. This option was already explained as a kind of language learning earlier. Learning spelling pronunciations was compared to learning a second variant (e.g., a dialect) of one’s own language. This learning is likely to build on a combination of rote learning and rule learning. Once a dialect is learned, the listener immediately and effortlessly recognises the spoken words without taking notice of the “deviant” sound structure. The knowledge of spelling pronunciations may become so entrenched in the mental lexicon that the activation of a spelling pronunciation is just as efficient as (or even more efficient than) activation of a standard phonological representation. Word recognition of the new variant pronunciation becomes automatic with no need for conscious awareness. At least two predictions follow from this second option – that spelling pronunciations are learned as spoken variants of known words. First, learning spelling pronunciations greatly reduces the need for learning the advanced orthographic code as a set of conditional orthography to phonology connections (or rules). This would explain why even good readers have great difficulties explaining the advanced orthographic rules that they appear to follow when reading. Second, if learning spelling pronunciations is an important aspect of orthographic learning, we would expect that spelling pronunciations may sometimes take over from standard pronunciations in words that are common in print. Indeed, there are many examples in support of this prediction. Several words have changed pronunciation (or resisted changes) over the years under the influence of orthography: humour and hotel were originally pronounced without the initial “h” (as in French) from the 14th century till the beginning of the 20th century when the spelling pronunciation with the “h” took over. Blush was originally pronounced “bloosh” until the spelling pronunciation took over. Template used to be “templit” (see Bloomfield, 1984, § 27.6 for additional examples).

Conclusion and future directions In this chapter, we have detailed how learning to read words can be seen as a special case of verbal learning. Evidently, the standard sounds of the letters have to be learned. It is an important recent finding that the visual-verbal learning of letter sounds is related to reading development because of the verbal learning component (Litt et al.,

184 Carsten Elbro and Peter F. de Jong

2013). What matters is that children are able to learn the letter sounds sufficiently well to allow for active production (recall). Mere recognition is not sufficient. By sounding out the letters, beginning readers may be able to produce a spelling pronunciation which they can then often recognise as a real word. The main part of the chapter has described the nature and the recognition of such spelling pronunciations of words already in lexicon. Spelling pronunciations are mostly overlooked in the literature on reading even though anyone who listens to a beginning reader can easily observe them. For beginning readers, spelling pronunciations are “stepping stones” in the process of word recognition. The recognition of words from their spelling pronunciations is no simple task though. It requires a (sometimes high) degree of phonological flexibility – especially in the case of irregularly spelled words. There is strong evidence that orthographic learning influences phonological representations of the words in the reader’s mental lexicon (e.g., Morais & Kolinsky, 2005; Nation & Hulme, 2011; Ranbom & Connine, 2011). Accordingly, we suggest that spelling pronunciations are learned just like any other variant of a spoken language – both on an item basis and by means of rules, that is, patterns that emerge from the relationships between spelling pronunciations and standard pronunciations. This learning is also a kind of verbal learning because the reader has to be able to both produce and recognise the spelling pronunciation. The gap between spelling pronunciations and standard pronunciations are narrowed as the reader becomes familiar with more advanced orthographic conventions (the extended code) and can take into account several letters at once when their sounds are activated. The gap may completely disappear when specific links are formed between the unique letter sequences of words and their unique pronunciations (in “sight words”). Alternatively, the gap may diminish as spelling pronunciations are learned as variants (a “dialect”) of standard pronunciations. In both cases, once established spelling pronunciations will cease to be consciously activated in fluent reading. However, even proficient readers and spellers may activate (recall) spelling pronunciations in order to spell irregularly spelled words. As a result, spelling pronunciations for some words are rehearsed and thus fully retained in the mental lexicon. Although there is ample evidence that both beginning readers and advanced spellers have access to spelling pronunciations, a number of central issues remain unknown: To what extent are beginning readers supported by prior knowledge of spelling pronunciations? To our knowledge, no relevant teaching experiment has been reported. How well does knowledge of patterns of spelling pronunciations (“orthographic conventions”) generalise across words that share the pattern? Are spelling pronunciations stored and activated separately or as variants of already known pronunciations? Functional brain imagery might shed light over this and related questions.



Orthographic learning is verbal learning 185

References Anthony, J. L., Aghara, R. G., Dunkelberger, M. J., Anthony, T. I., Williams, J. M., & Zhang, Z. (2011). What factors place children with speech sound disorders at risk for reading problems? American Journal of Speech-Language Pathology, 20, 146–160.  doi: 10.1044/1058-0360(2011/10-0053) Bloomfield, L. (1984, org. 1933). Language. Chicago: University of Chicago Press. Bosman, A. M. T., van Hell, J. G., & Verhoeven, L. (2006). Learning the spelling of strange words in Dutch benefits from regularized reading. Journal of Educational Psychology, 98, 879–890.  doi: 10.1037/0022-0663.98.4.879 Bürki, A., Spinelli, E., & Gareth Gaskell, M. (2012). A written word is worth a thousand spoken words: The influence of spelling on spoken-word production. Journal of Memory and Language, 67, 449–467.  doi: 10.1016/j.jml.2012.08.001 Byrne, B. (1998). The foundation of literacy: The child’s acquisition of the alphabetic principle. Hove, East Sussex: Psychology Press. Caravolas, M., Lervåg, A., Defior, S., Málková, G. S., & Hulme, C. (2013). Different patterns, but equivalent predictors, in growth in reading in consistent and inconsistent orthographies. Psychological Science, 24, 1398–1407.  doi: 10.1177/0956797612473122 Cardoso-Martins, C., Mesquita, T. C., & Ehri L. (2011). Letter names and phonological awareness help children to learn letter-sound relations. Journal of Experimental Child Psychology, 109, 25–38.  doi: 10.1016/j.jecp.2010.12.006 Carney, E. (1994). A survey of English spelling. London: Routledge.  doi: 10.4324/9780203199916 Coltheart, M., Rastle, K., Perry, C., Langdon, R., & Ziegler, J. (2001). DRC: a dual route cascaded model of visual word recognition and reading aloud. Psychological Review, 108, 204–56. doi: 10.1037/0033-295X.108.1.204 Connine, C. M. (2004). It’s not what you hear but how often you hear it: On the neglected role of phonological variant frequency in auditory word recognition. Psychonomic Bulletin & Review, 11, 1084–1089.  doi: 10.3758/BF03196741 Connine, C. M., Ranbom, L. J., & Patterson, D. J. (2008). Processing variant forms in spoken word recognition: The role of variant frequency. Perception and Psychophysics, 70, 403–411.  doi: 10.3758/PP.70.3.403 de Jong, P. F. (2007). Phonological awareness and the use of phonological similarity in letter-sound learning. Journal of Experimental Child Psychology, 98, 131–152. doi: 10.1016/j.jecp.2007.06.003 de Jong, P. F. (2011). What discrete and serial rapid automatized naming can reveal about reading. Scientific Studies of Reading, 15, 314–337.  doi: 10.1080/10888438.2010.485624 de Jong, P. F. & van der Leij, A. (1999). Specific contributions of phonological abilities to early reading acquisition: Results from a Dutch latent variable longitudinal study. Journal of Educational Psychology, 91, 450–476.  doi: 10.1037/0022-0663.91.3.450 de Jong, P. F. & van der Leij, A. (2003). Developmental changes in the manifestation of a phonological deficit in dyslexic children learning to read a regular orthography. Journal of Educational Psychology, 95, 22–40.  doi: 10.1037/0022-0663.95.1.22 Drake, D. A. & Ehri, L. C. (1984): Spelling acquisition: Effects of pronouncing words on memory for their spellings. Cognition and Instruction, 1, 297–320.  doi: 10.1207/s1532690xci0103_2 Ehri, L. C. (1992). Reconceptualizing the development of sight word reading and its relationship to recoding. In P. B. Gough, L. C. Ehri & R. Treiman (Eds.), Reading Acquisition (pp. 107–143). Hillsdale, NJ: Erlbaum.

186 Carsten Elbro and Peter F. de Jong

Ehri, L. C. (2005). Learning to read words: Theory, findings, and issues. Scientific Studies of Reading, 9, 167–188.  doi: 10.1207/s1532799xssr0902_4 Elbro, C. (1996). Early linguistic abilities and reading development: A review and a hypothesis. Reading and Writing: An Interdisciplinary Journal, 8, 453–485.  doi: 10.1007/BF00577023 Elbro, C. (2005). Literacy acquisition in Danish: A deep orthography in cross-linguistic light. In R. M. Joshi & P. G. Aaron (Eds.), Handbook of orthography and literacy (pp. 31–45). Mahwah, NJ: Erlbaum. Elbro, C., Borstrøm, I., & Petersen, D. K. (1998). Predicting dyslexia from kindergarten. The importance of distinctness of phonological representations of lexical items. Reading Research Quarterly, 33, 36–60.  doi: 10.1598/RRQ.33.1.3 Elbro, C., de Jong, P. F., Houter, D., & Nielsen, A.-M. V. (2012). From spelling pronunciation to lexical access: A second step in word decoding? Scientific Studies of Reading, 16, 341–359. doi: 10.1080/10888438.2011.568556 Elbro, C., & Jensen, M. N. (2005). Quality of phonological representations, word learning, and phoneme awareness in dyslexic and normal readers. Scandinavian Journal of Psychology, 46, 375–384.  doi: 10.1111/j.1467-9450.2005.00468.x Elbro, C., Nielsen, I., & Petersen, D. K. (1994). Dyslexia in adults: Evidence for deficits in nonword reading and in the phonological representation of lexical items. Annals of Dyslexia, 44, 205–226.  doi: 10.1007/BF02648162 Elbro, C., & Petersen, D. K. (2004). Long-term effects of phoneme awareness and letter name training. An intervention study with children at risk of dyslexia. Journal of Educational Psychology, 96, 660–670.  doi: 10.1037/0022-0663.96.4.660 Elbro, C., & Scarborough, H. S. (2004a). Early identification. In T. Nunes & P. Bryant (Eds.), Handbook of Children’s Literacy (pp. 339–359). Dordrecht: Kluwer.  doi: 10.1007/978-94-017-1731-1_19 Elbro, C., & Scarborough, H. S. (2004b). Early intervention. In T. Nunes & P. Bryant (Eds.), Handbook of Children’s Literacy (pp. 361–381). Dordrecht: Kluwer.  doi: 10.1007/978-94-017-1731-1_20 Ellefson, M. R., Treiman, R., & Kessler, B. (2009). Learning to label letters by sounds or names: a comparison of England and the United States. Journal of Experimental Child Psychology, 102, 323–41.  doi: 10.1016/j.jecp.2008.05.008 Fowler, A. E., & Swainson, B. (2004). Relationships of naming skills to reading, memory, and receptive vocabulary: Evidence for imprecise phonological representations of words by poor readers. Annals of Dyslexia, 54, 247–281.  doi: 10.1007/s11881-004-0013-0 Harm, M. W., McCandliss, B. D., & Seidenberg, M. S. (2003). Modeling the successes and failures of interventions for disabled readers. Scientific Studies of Reading, 7, 155–182. doi: 10.1207/S1532799XSSR0702_3 Harm, M. W., & Seidenberg, M. S. (1999). Phonology, reading acquisition, and dyslexia: Insights from connectionist models. Psychological Review, 106, 491–528.  doi: 10.1037/0033-295X.106.3.491 Harm, M. W., & Seidenberg MS. (2004). Computing the meanings of words in reading: Cooperative division of labour between visual and phonological processes. Psychological Review, 111, 662–720.  doi: 10.1037/0033-295X.111.3.662 Heltbech, M. A. & Nemholt, P. S. (2007). Indlæring og brug af skriftnære udtaler – en undersøgelse af voksne ordblindes tilegnelse og anvendelse af skriftnære udtaler til lagring af ordspecifikke stavemåder [’Learning and application of spelling pronunciations of word-specific spellings. A study of adult dyslexics’]. Unpublished MA thesis, Copenhagen: Department of Nordic Studies and Linguistics, University of Copenhagen. Hilte, M., Bos, M., & Reitsma, P. (2005). Effects of spelling pronunciations during spelling practice in Dutch. Written Language and Literacy, 8, 137–153.  doi: 10.1075/wll.8.2.06hil



Orthographic learning is verbal learning 187

Hilte, M. & Reitsma, P. (2006). Spelling pronunciation and visual preview both facilitate learning to spell irregular words. Annals of Dyslexia, 56, 301–318.  doi: 10.1007/s11881-006-0013-3 Huang, F. L., Tortorelli, L. S., & Invernizzi, M. A. (2014). An investigation of factors associated with letter-sound knowledge at kindergarten entry. Early Childhood Research Quarterly, 29(2), 182–192.  doi: 10.1016/j.ecresq.2014.02.001 Hulme, C., Goetz, K., Gooch, D., Adams, J., & Snowling, M. J. (2007). Paired-associate learning, phoneme awareness, and learning to read. Journal of Experimental Child Psychology, 96, 150–66.  doi: 10.1016/j.jecp.2006.09.002 Kim, Y.-S., Petscher, Y., Foorman, B. R. & Zhou, C. (2010). The contributions of phonological awareness and letter-name knowledge to letter-sound acquisition – a cross-classified multilevel model approach. Journal of Educational Psychology, 102, 313–326. doi: 10.1037/a0018449 Lervåg, A., Bråten, I., & Hulme, C. (2009). The cognitive and linguistic foundations of early reading development: a Norwegian latent variable longitudinal study. Developmental Psychology, 45, 764–781.  doi: 10.1037/a0014132 Levin, I., Shatil-Carmon, S., & Asif-Rave, O. (2006). Learning of letter names and sounds and their contribution to word recognition. Journal of Experimental Child Psychology, 93, 139– 165.  doi: 10.1016/j.jecp.2005.08.002 Liberman, A. M. (1992). The relation of speech to reading and writing. In R. Frost & L. Katz (Eds.), Orthography, phonology, morphology, and meaning. Advances in Psychology, 94, 167–178. doi: 10.1016/S0166-4115(08)62794-6 Liberman, I. Y., Shankweiler, D. S., Fisher, F. W., & Carter, B. (1974). Explicit syllable and phoneme segmentation in the young child. Journal of Experimental Child Psychology, 18, 201–212. doi: 10.1016/0022-0965(74)90101-5 Litt, R. A., de Jong, P. F., van Bergen, E., & Nation, K. (2013). Dissociating crossmodal and verbal demands in paired associate learning (PAL): What drives the PAL – reading relationship? Journal of Experimental Child Psychology, 115, 137–149.  doi: 10.1016/j.jecp.2012.11.012 Litt, R. A., & Nation, K. (2014). The nature and specificity of paired associate learning deficits in children with dyslexia. Journal of Memory and Language, 71, 71–88. doi: 10.1016/j.jml.2013.10.005 Lyon, G. R., Shaywitz, S. E., & Shaywitz, B. A. (2003). A definition of dyslexia. Annals of Dyslexia, 53, 1–14.  doi: 10.1007/s11881-003-0001-9 Marinus, E., & de Jong, P. F. (2010). Variability in the word-reading performance of dyslexic readers: Effects of letter length, phoneme length and digraph presence. Cortex, 46, 1259– 1271.  doi: 10.1016/j.cortex.2010.06.005 McBride-Chang, C. (1999). The ABCs of the ABCs: The development of letter-name and letter-sound knowledge. Merrill-Palmer Quarterly, 45, 285–308. McBride-Chang, C., Lam, F., Lam, C., Chan, B., Fong, C. Y. C., Wong, T. T. Y., & Wong, S. W. L. (2011). Early predictors of dyslexia in Chinese children: Familial history of dyslexia, language delay, and cognitive profiles. Journal of Child Psychology and Psychiatry, 52, 204–211. doi: 10.1111/j.1469-7610.2010.02299.x McGuinness, D. (1998). Why children can’t read and what we can do about it. A scientific revolution in reading. London: Penguin. Messbauer, V. C. S., & de Jong, P. F. (2003). Word, nonword and visual paired associate learning in Dutch dyslexic children. Journal of Experimental Child Psychology, 84, 77–96. doi: 10.1016/S0022-0965(02)00179-0

188 Carsten Elbro and Peter F. de Jong

Metsala, J. L., & Walley, A. C. (1998). Spoken vocabulary growth and the segmental restructuring of lexical representations: Precursors to phonemic awareness and early reading ability. In J. L. Metsala & L. C. Ehri (Eds.), Word recognition in beginning literacy (pp. 89–120). Mahwah, NJ: Erlbaum.. Moll, K., Loff, A., & Snowling, M. (2013). Cognitive endophenotypes of dyslexia. Scientific Studies of Reading, 17, 385–397.  doi: 10.1080/10888438.2012.736439 Morais, J. & Kolinsky, R. (2005). Literacy and cognitive change. In M. Snowling & C. Hulme (Eds.), The Science of Reading: A Handbook (pp. 188–203). Oxford: Blackwell. doi: 10.1002/9780470757642.ch11 National Early Reading Panel (2008). Developing Early Literacy. Report of the National Early Reading Panel. Jessup, Maryland: National Institute for Literacy. Oakhill, A. E., (1960). Personal communication. Morden, London. Perfetti, C. A. (1992). The representation problem in reading acquisition. In P. B. Gough, L. C. Ehri & R. Treiman (Eds.), Reading acquisition (pp. 145–174). Hillsdale, NJ: Erlbaum. Piasta, S. B., & Wagner, R. K. (2010). Learning letter names and sounds: Effects of instruction, letter type, and phonological processing skill. Journal of Experimental Child Psychology, 105, 324–344.  doi: 10.1016/j.jecp.2009.12.008 Pinker, S. (1994). The language instinct. The new science of language and mind. London: Allen Lane. Ramus, F., & Szenkovits, G. (2008). What phonological deficit? The Quarterly Journal of Experimental Psychology, 61, 129–141.  doi: 10.1080/17470210701508822 Ranbom, L. J., & Connine, C. M. (2011). Silent letters are activated in spoken word recognition. Language and Cognitive Processes, 26, 236–261.  doi: 10.1080/01690965.2010.486578 Schatschneider, C., Fletcher, J. M., Francis, D. J., Carlson, C. & Foorman, B. R. (2004). Kindergarten prediction of reading skills: A longitudinal comparative analysis. Journal of Educational Psychology, 96, 265–282.  doi: 10.1037/0022-0663.96.2.265 Share, D. L. (2004). Knowing letter names and learning letter sounds: A causal connection. Journal of Experimental Child Psychology, 88, 213–233.  doi: 10.1016/j.jecp.2004.03.005 Snowling, M. J. (2000). Dyslexia (2nd ed.). Oxford: Blackwell. Snowling, M. & Hulme, C. (2012). Annual research review: The nature and classification of reading disorders – commentary on proposals for DSM-5. Journal of Child Psychology and Psychiatry, 53, 593–607.  doi: 10.1111/j.1469-7610.2011.02495.x Thaler, V., Landerl, K., & Reitsma, P. (2008). An evaluation of spelling pronunciations as a means of improving spelling of orthographic markers. European Journal of Psychology of Education, 23, 3–23.  doi: 10.1007/BF03173137 Torppa, M., Lyytinen, P., Erskine, J., Eklund, K., & Lyytinen, H. (2010). Language development, literacy skills, and predictive connections to reading in Finnish children with and without familial risk for dyslexia. Journal of Learning Disabilities, 43, 308–321.  doi: 10.1177/0022219410369096 Treiman, R., Tincoff, R., Rodriguez, K., Mouzaki, A., & Francis, D. J. (1998). The foundations of literacy: Learning the sounds of letters. Child Development, 69, 1524–1540. doi: 10.1111/j.1467-8624.1998.tb06175.x Tunmer, W. E., & Chapman, J. W. (2012). Does set for variability mediate the influence of vocabulary knowledge on the development of word recognition skills? Scientific Studies of Reading, 16, 122–140.  doi: 10.1080/10888438.2010.542527 van Bergen, E., de Jong, P. F., Plakas, A., Maassen, B. & van der Leij, A. (2012). Child and parental literacy levels within families with a history of dyslexia. Journal of Child Psychology and Psychiatry, 53, 28–36.  doi: 10.1111/j.1469-7610.2011.02418.x



Orthographic learning is verbal learning 189

van Bergen, E., de Jong, P. F., Maassen, B., & van der Leij, A. (2014). The effect of parents’ literacy skills and children’s preliteracy skills on the risk of dyslexia. Journal of Abnormal Child Psychology, 42, 1187–1200.  doi: 10.1007/s10802-014-9858-9 van den Boer, M. & de Jong, P. F. (2015). Parallel and serial reading processes in children’s word and nonword reading. Journal of Educational Psychology, 107, 141–151.  doi: 10.1037/a0037101 van den Boer, M., Georgiou, G., & de Jong, P. F. (2016). Naming of short words is (almost) the same as naming of alphanumeric symbols: Evidence from two orthographies. Journal of Experimental Child Psychology, 144, 152–165.  doi: 10.1016/j.jecp.2015.11.016 Vellutino, F. R., Steger, J. A., & Pruzek, R. M. (1973). Inter vs. intrasensory deficit in paired-associates learning in poor and normal readers. Canadian Journal of Behavioral Science, 5, 111–123.  doi: 10.1037/h0082336 Venezky, R. L. (1972). Language and cognition in reading (Tech. Rep. No. 188). Madison, WI: Wisconsin Research and Development Center for Cognitive Learning, University of Wisconsin. Wagner, R. K., Torgesen, J. K., Rashotte, C. A., Hecht, S. A., Barker, T. A., Burgess, S. R., et al. (1997). Changing relations between phonological processing abilities and word-level reading as children develop from beginning to skilled readers: A 5-year longitudinal study. Developmental Psychology, 33, 468–479.  doi: 10.1037/0012-1649.33.3.468 Warmington, M. & Hulme, C. (2012). Phoneme awareness, visual-verbal paired-associate learning, and rapid automatized naming as predictors of individual differences in reading ability. Scientific Studies of Reading, 16, 45–62.  doi: 10.1080/10888438.2010.534832 Windfuhr, K. L., & Snowling, M. J. (2001). The relationship between paired associate learning and phonological skills in normally developing readers. Journal of Experimental Child Psychology, 80, 160–173.  doi: 10.1006/jecp.2000.2625 Ziegler, J. C., Perry, C., Ma-Wyatt, A., Ladner, D. & Schulte-Körne, G. (2003). Developmental dyslexia in different languages: Language-specific or universal? Journal of Experimental Child Psychology, 86, 169–193.  doi: 10.1016/S0022-0965(03)00139-5 Zoccolotti, P., De Luca, M., Di Pace, E., Gasperini, F., Judica, A. & Spinelli, D. (2005). Word length effect in early reading and in developmental dyslexia. Brain and Language, 93, 369–373. doi: 10.1016/j.bandl.2004.10.010

Learning to read morphologically complex words Joanne F. Carlisle and Devin M. Kearns University of Connecticut

In learning language, children come to recognize the ways that morphemes (the smallest units of meaning in words) can be combined and recombined to express different meanings and serve particular grammatical roles. From their first exposure to language, they encounter words formed by combining two or more free morphemes (i.e., compounds; e.g., fireman) as well as words formed by combining free and bound morphemes (prefixes or suffixes; e.g., running, runner). Experience with words leads to links among members of a word family (e.g., run, running, runner). Such knowledge is largely implicit, part of the natural process of learning to speak and understand a language. However, the ability to read and understand unfamiliar morphologically complex words becomes particularly important for the development of students’ vocabulary and comprehension of text come the middle and high school years (e.g., Cunningham, 2010; Kirby & Bowers, this volume; Nagy & Anderson, 1984; Snow, 2010). Perhaps because morphemes are units of meaning, the relation of morphological knowledge and vocabulary development has been a topic of considerable debate (e.g., Sparks & Deacon, 2012; Spencer et al., 2015; see also Spencer, Quinn, & Wagner, this volume). Much less attention has been paid to the question of how and when morphological structure plays a role in the accuracy and fluency of children’s word recognition (e.g., Deacon, 2012a), and yet this is potentially critical for comprehensive models of the process of acquiring reading skill. Understandably, researchers and educators have focused primarily on beginning readers’ learning of the alphabetic system in English – decoding written words using grapheme-phoneme correspondences to recover or determine the pronunciation of words. Still, it is increasingly clear that the morphological composition of words should be incorporated in developmental theories (Deacon, 2012a; Nunes, Bryant, & Barros, 2012). Frost (2012) argues that: “a universal model of reading that is a learning model must include, in one way or another, an architecture which considers the intricate relations of orthography, phonology, and morphology (and therefore of meaning) in the language” (p. 277). Thus, the purpose of this chapter is to provide a review of theory and empirical studies of beginning readers’ use of morphological information in word reading.

doi 10.1075/swll.15.11car © 2017 John Benjamins Publishing Company

192 Joanne F. Carlisle and Devin M. Kearns

Even beginning readers encounter morphologically complex words in books read to them and books they read to themselves. Through an informal examination of books commonly read by first and third graders, our informal observation suggests that there is a noticeable increase in the number of polymorphemic words. For example, Father Bear Comes Home (Minarik, 1959), a book appropriate for first graders, contains a number of inflections, such as coming, walked, fishing, and hugged. Morphologically complex words have an even more noticeable presence in Charlotte’s Web (White, 1952), a book appropriate for third graders. The following excerpt comes from the beginning of the book (p. 4): Fern came slowly down the stairs. Her eyes were red from crying. As she approached her chair, the carton wobbled, and there was a scratching noise. Fern looked at her father. Then she lifted the lid of the carton. There, inside, looking up at her, was the newborn pig. It was a white one. The morning light shone through its ears, turning them pink. “He’s yours,” said Mr. Arable. “Saved from an untimely death. And may the good Lord forgive me for this foolishness.”

This short section includes numerous inflections (stairs, crying, approached, wobbled, scratching, looked, lifted, looking, turning, saved), a few derivations (slowly, untimely, foolishness), and two compounds (inside, newborn). Almost every sentence has at least one morphologically complex word. These texts made us wonder when and how students learn to read morphologically complex words and whether the morphemic composition of words facilitates or hinders word reading accuracy and fluency. We begin with a review of children’s learning of morphology and the relation of their morphological knowledge to early word reading achievement. We consider theoretical accounts of learning to read complex words, identifying in the process salient questions about the role morphemes might play in the development of word reading skill. Then we review the findings of empirical studies of elementary students’ reading of polymorphemic words. We close by suggesting priorities for future research that may advance our understanding of the development of word reading skill.

Setting the stage: Young students’ morphological knowledge Children begin to use words of more than one morpheme when the mean length of utterances exceeds two words. They use inflected forms, such as plurals and the present progressive (crying in “stop crying”) (Brown, 1973), and they create novel compounds in order to express concepts or ideas for which they do not have a readily available word in memory. They also use prefixes and suffixes that have familiar meanings or are highly productive early on. Clark (1982) gives the example of a three-year old who said about the dinosaur he had just drawn: “It looks growl-y, doesn’t it?” The -y suffix is productive because its function (i.e., to turn a noun into an adjective) is so useful to preschool children (e.g., sticky, gooey, scary). Even so, first graders know relatively few derivations;



Reading complex words 193

this is an aspect of very rapid growth over the elementary years (Anglin, 1993). Older students are certainly more able to analyze and manipulate word forms and meanings. A summary of key findings from studies of elementary students’ morphological awareness sets the stage for examining the effects of morphological knowledge on word reading in general and (with regard to the main purpose of this chapter) on polymorphemic word reading.

1. As early as kindergarten and first grade, students show sensitivity to the composition of age-appropriate polymorphemic words (e.g., Apel, Diehm & Apel, 2013; Carlisle & Nomanbhoy, 1993; Nagy, Berninger, & Abbott, 2006; Rubin, Patterson & Kantor, 1991; Treiman & Cassar, 1996). Students might be asked to decompose words into morphemes, combine morphemes, or distinguish derived words from similar monomorphemic words (e.g., “Is there a little word in dollar that means something like dollar?”). The ability to identify a familiar base in a derived word increases between the early and late elementary grades (e.g., Berninger, Abbott, Nagy & Carlisle, 2010; Tyler & Nagy, 1989). 2. Elementary students are able to complete analogical reasoning tasks in which the grammatical relation of one pair of words is matched in a second pair (A:B::C:D). Test items might be words or sentences; the student supplies the final word or sentence. For example, Deacon, Kieffer, and Laroche (2014) found that third and fourth graders were able to complete analogies that involved inflections and derivations, as in the following item:“tall:tallest::strong:strongest.” 3. Compared to early elementary students, late elementary students are more likely to attend to the meanings of affixes and syntactic roles of polymorphemic words (e.g., Carlisle & Fleming, 2003). Until the late elementary or middle school years, students are sensitive to the meaning of the base word, often without taking into account the meaning or grammatical role of an affix (e.g., Nagy, Diakidoy, & Anderson, 1993; Tyler & Nagy, 1990). 4. Students’ knowledge of the meaning of derived words increases dramatically between first and fifth grade. Anglin (1993) found increased knowledge of basic words, inflections, derivations, compounds, and idioms from first to third and fifth grades. Most noteworthy was the increase in knowledge of derived words. He estimated that students added to their vocabulary about 5 derived words per day between first and third grades and 14 derived words per day between third and fifth grades (Anglin, 1993, p. 73). 5. Lack of phonological and orthographic transparency of the base word within a derived word form adversely affects students’ awareness of morphological relationships (e.g., Fowler & Liberman, 1995; Tyler & Nagy, 1989). Less skilled readers perform less well than their peers on oral measures of morphological knowledge, but particularly on derived words that lack phonological transparency (e.g., Carlisle, Stone, & Katz, 2001; Casalis, Colé & Sopo, 2004; Fowler & Liberman, 1995). When the base form undergoes a phonological shift, the phonological representation of the whole word (e.g., natural) may interfere with morphological decomposition of the derived word (e.g., nature + -al).

194 Joanne F. Carlisle and Devin M. Kearns

Contribution of morphological knowledge to general word reading The results of studies of morphological awareness suggest that students bring to the process of learning to read considerable knowledge about morphology, but does this knowledge contribute to beginning readers’ developing word reading skill? Studies that have addressed this question report that morphological knowledge is a significant predictor of word reading from first grade on. For example, first through third graders’ morphological knowledge was found to have an increasingly strong relationship to word reading; by third grade, it significantly predicted word reading speed and accuracy, pseudoword reading accuracy, text reading speed, and reading comprehension, after controlling for verbal and nonverbal ability and phonological awareness (Kirby, Deacon, Bowers, Izenberg, Wade-Woolley & Parrila, 2012). Deacon and Kirby (2004) found that morphological knowledge contributed unique variance to both word and pseudoword reading in grades 3 through 5, after controlling for phonological awareness and verbal and nonverbal intelligence. Other researchers have reported similar results for students in grades 3 through 6 (Singson, Mahoney, & Mann, 2000) and grades 4, 6, and 8 (Roman, Kirby, Parrila, Wade-Woolley, & Deacon, 2009). While early on morphological knowledge is tacit, acquired as part of the process of learning to speak and understand English, learning to read and write makes it necessary for students to treat language as an object of thought (e.g., Mattingly, 1984; Tunmer & Herriman, 1984). Their knowledge of morphology becomes more explicit (Nunes, Bryant, & Bindman, 1997), as is evident from their ability to reflect on, analyze, and manipulate the morphological composition of words. Students become increasingly able to use the morphemic structure and meanings of constituent morphemes to identify and understand unfamiliar words in texts (Nagy & Anderson, 1984; Nagy & Scott, 2000). Processing morphologically complex words involves integration of different kinds of linguistic knowledge about words (e.g., morphological, phonological, grammatical, semantic, and orthographic) (Kuo & Anderson, 2003; Berninger, Abbott, Nagy, & Carlisle, 2010). Some researchers regard morphological awareness as a particular component of linguistic awareness, which encompasses different aspects of analysis of word structure and meaning. On the other hand, in some studies, the construct of morphological awareness appears to involve various kinds of linguistic analysis as it relates to word reading. That is, what appears to be the influence of morphological knowledge may reflect processing of different linguistic characteristics (e.g., phonology, semantics). For example, while Deacon (2012b) reported that phonological, orthographic, and morphological awareness made independent contributions to first through third graders’ real and pseudoword reading, most of the variance was shared (R2 = .82).



Reading complex words 195

Morphemic processing in theories of early word reading Developmental theories of learning to read are most concerned with beginning readers’ foundational reading skills (e.g., Adams, 1990; Share, 2011). In English this entails learning letters, the sounds of letters (i.e., grapheme-phoneme correspondences), and common combinations of letters (e.g., onsets and rimes). In most theories, the effects of children’s language development (including morphological knowledge) and language-related word characteristics (e.g., age of acquisition, concreteness, word length) are not central concerns (e.g., Fitzgerald, Elmore, Koons, Hiebert, Bowen, Sanford-Moore, & Stenner, 2015). Ehri’s phase theory (1995; 2005; this volume) follows this pattern in emphasizing development of orthography-phonology connections. The last phase of her model is called consolidation and involves what she calls “unitization.” In this phase, letter patterns that occur frequently in words come to be processed as units, not separate letters; these include bigrams and trigrams (e.g., pr-, str-), phonograms (e.g., -ip in sip, tip), syllables (e.g., sis and ter in sister), and morphemes (e.g., help and less in helpless). Frequent encounters with such units contribute to mental representations that, in turn, lessen the burden of decoding. She gives the example of the word interesting, which if processed as four syllables, is likely to be recognized faster and more accurately than would be the case if the word were read by “sounding out” the eleven graphemes (2005, p. 175). With regard to children’s reading of polymorphemic words, therefore, phase theory indicates that morphemes would be used – as orthographic units. Other developmental models emphasize the importance of morphology late (but not early) in the process of learning to read and spell (e.g., Adams, 1990; Seymour, 1997). In Seymour’s spelling model, knowledge acquired in the logographic and alphabetic stages are the basis on which the orthographic and (thereafter) the morphographic frameworks are built. Similarly, Templeton (1991) suggested that students first learn to process words by basic sound-spelling elements and patterns and only later by syllable juncture or derivational constancy – perhaps the late elementary or middle school years. We argue that even beginning readers are likely to attend to meaningful units (e.g., base words, familiar suffixes) as they encounter them in written texts, just as they do in spoken language. Consider, for example, a child reading the word tasteless in the sentence, “That cake is tasteless.” While it is possible that tasteless is read as a whole word, it seems likely that identifying the meaning of this relatively uncommon word entails three steps of morphological processing, segmentation (i.e., activating the constituent morphemes), licensing (i.e., checking whether the morphemes can be integrated, given their properties), and computation of the lexical representation of the complex word (Schreuder & Baayen, 1995). In short, even for beginning readers, reading for meaning is likely to involve coordinated analysis of the form and meaning of polymorphemic words. For very familiar polymorphemic words, such as the inflected forms in Father Bear Comes Home (e.g., walked, fishing), the process is likely to be relatively automatic. However, less familiar words are bound to require more deliberate analysis. This might the case in the

196 Joanne F. Carlisle and Devin M. Kearns

sentence, “That joke was tasteless.” Students might have no trouble pronouncing the word tasteless, but the meaning of taste in this context is probably not in young readers’ lexical memory. As Perfetti (2007) explains, readers’ fund of lexical knowledge affects their text comprehension; such knowledge includes grammatical roles, spellings, and meanings of words or word parts. Morphemes most certainly serve as units of word reading, but understanding written texts involves the interplay of various sources of linguistic information, drawing on not just orthographic and phonological features but also morphological forms, grammatical roles, and more. Early elementary students are just beginning to acquire the depth of lexical knowledge that skilled readers use to decode and understand morphologically complex words effortlessly.

Do morphemes serve as units in children’s word reading? We focus on this central question by turning again to Ehri’s suggestion that young readers develop mental representations of orthographic units that then serve as the basis for the development of sight word reading. Although both children and adults use a variety of letter patterns in reading words (e.g., rimes, syllables, morphemes) (e.g., Yap & Balota, 2009), relatively little research has focused on students’ use of different units in reading words – a situation that we believe needs to be addressed. Therefore, we prepared a list of the units beginning readers might use, as shown in Figure 1. We thought that examination of this list might help us understand the possible role of morphemes in early word reading. The list may appear to be hierarchical – perhaps starting with the size of the letter pattern or unit. In some ways, this reflects what we know about beginning readers’ learning of phoneme-grapheme correspondences and letter patterns. For them, more so than skilled readers, word length affects speed of word recognition (e.g., De Luca, Barca, Burani, & Zoccolotti, 2008; Gagl, Hawelka, & Wimmer, 2015). In part this may be because it takes time for inexperienced readers to compose the pronunciation of longer as compared to shorter words. On the other hand, the increase in length is actually not a defining characteristic of many units. For example, morphemes can be large or small (cf. a- in atypical versus micro- in microscope), and this is true of syllables as well (cf. e in ego versus splen in splendor). The major difference between syllables and morphemes is that morphemes are units of meaning, whereas syllables are units of vocalization. An additional level of complexity is that some letter patterns can function as either a syllable or a morpheme; for example, -en is a syllable in garden) but both a syllable and a morpheme in harden. Letter patterns in Figure 1 appear to differ in orthographic complexity – and this complexity stems from a number of characteristics. For example, the frequency of letters or letter combinations matters. Young readers are more competent at reading words with single consonants than those with consonant blends and digraphs (e.g., Gagl et al., 2015). They are also more accurate or faster at reading words with many orthographic “neighbors” (i.e., words that differ in one letter – kit, fit, and pit as neighbors of sit) than



Reading complex words 197

Orthographic Units Graphemes (t, ch)

Phonological Units grapheme-phoneme correspondence (GPC)

Onsets (/st/)

Orth. onsets (st) Syll. bodies (-eam)

Sounds (/t/, /t∫/)

phonogram

Rimes (/im/)

Syllables (sym)

Syllables (/sim/)

Morphemes (re)

Morphemes (/ri/)

Words (steam)

Words (/stim/)

Morphemes Concepts (re = {again}) Syntax (ed = {past tense} Words Concepts (steam = {gaseous water}) Syntax (steam = {noun}) Semantic/Syntactic Units

Figure 1.  Interplay of phonological, orthographic and semantic/syntactic elements in word reading

those with few neighbors (e.g., suit) (e.g., Laxon, Masterson, & Moran, 1994). Positional constraints matter as well. For example, -ly commonly appears at the ending but not beginning of words. Children need sufficient experience with longer words to learn what to expect by way of the position of letters and letter patterns within words. Young readers gradually internalize the statistical properties of the language through reading experience (e.g., Gagl et al., 2015; Treiman & Cassar, 1996). Orthographic knowledge does contribute to reading skill for beginning readers (Pasquarella, Deacon, Chen, Commissaire, & Au-Yeung, 2014). However, implied in this knowledge is links between phonological features of letters, syllables, and words (Goswami & Ziegler, 2006). Consistency and regularity of letter-sound relations is a critical factor. For example, consistency in the pronunciation of rimes (e.g., -ink in pink vs. -ead in head or read) facilitates young readers’ word recognition (e.g., Gagl et al., 2015; Laxon et al., 1994). We might then ask how this affects young readers’ recognition of polysyllabic and polymorphemic words. There is some evidence that complex relationships between

198 Joanne F. Carlisle and Devin M. Kearns

orthographic and phonological features that are characteristics of a given language system may still present challenges to the pronunciation of morphologically complex words. In English, one example is the systematic variation in pronunciation of the past tense, depending on the final consonant of the word (e.g., grabbed, fitted, ended). Verhoeven, Schreuder, and Baayen (2006) report that Dutch third and sixth graders gave slower responses on a lexical decision task to plurals that undergo vowel change than those that are regular or involve consonant doubling. This result suggests that graphotactic aspects of pluralization have a systematic effect on pronunciation. Another example is assignment of the stressed syllable in polysyllabic words, which may affect the reader’s perception of the identity of the word. Students have trouble reading words like locality when they do not think to place stress on the second rather than the first syllable. Jarmulowicz, Taran, and Hay (2007) found significant relations between sensitivity to stress production and decoding skills of typically developing third graders. Prosody, variation in phonetic features due to adjacent letters or syllables, and phonotactics all involve coordination of orthographic, phonological, and morphological characteristics of units. Our review of units that are involved in the development of sight word reading suggests that familiar morphemes might at times serve as units that young readers identify and use in word reading. Some evidence comes from tasks in which very young readers show that they can divide written words into morphemic units or circle affixes on written pseudowords (Apel, Diehm, & Apel, 2013; Wolter, Wood, & D’zatko, 2008). Performance on such tasks suggests that they are applying morphological principles. Nonetheless, there are a number of unanswered questions about morphemes as units of word recognition. First, it might be misleading to consider the morpheme a single type of unit because morphemes do not seem to form a homogeneous category. Free morphemes provide the core concepts in morphologically complex words whereas bound morphemes serve to modulate the meaning and/or grammatical role of these free morphemes (Nagy et al., 1989). The two types of bound morphemes (prefixes and suffixes) also differ from one another in form and function; both adjust the meaning of the base word, but suffixes also indicate the grammatical role of a word. Prefixes may obscure the orthographic presence of the base word, particularly in cases where the spelling of the prefix includes the initial consonant of the base word (e.g., immature). Suffixes sometimes affect the pronunciation of the base word, and in doing so may obscure the identity of the base word or the word’s morphemic structure (e.g., nature vs. natural). In particular, familiar free morphemes have a special status among units because of their frequency (high), their average size (greater, on average, than phonograms), and their link to meaning (unique among units). Some researchers have suggested that young readers are likely to learn (and read) morphologically complex words without decomposing them into morphemic constituents; only with exposure to the base word in other members of the word family are such words then decomposed for reading and spelling (e.g., Laxon, Rickard & Coltheart, 1992; Seymour, 1997). However, theory and empirical evidence converge



Reading complex words 199

in supporting the view that for children and adults morphemes function as units that contribute to the speed and accuracy of word recognition. In word recognition studies, adults’ performance on polymorphemic words is faster than on control words matched in spelling and sound (Frost, Grainger & Carreiras, 2008). For elementary age readers as well, results of lexical decision tasks show that familiar morphemes contribute to word recognition (e.g., Quémart, Casalis, & Duncan, 2012). Performance of French fourth graders on a lexical decision task showed that priming with a morphemic relative facilitated response times to derived words more than orthographic primes (Casalis, Dusautoir, Colé, & Ducrot, 2009; see also Quémart, Casalis, & Colé, 2011). That is, morphological primes like laveur (cleaner) speeded responses for morphologically complex words like lavage (cleaning) more than orthographic primes like lavande (lavender) or unrelated primes like moutarde (mustard). To summarize, there are reasons to expect that even for beginning readers morphology may facilitate word reading. Different types of letter patterns are likely to play a role. The process may not be stepwise (i.e., resorting to more basic units if more complex units are not familiar) but rather dependent on the particular features of words. It is possible that reading words with regular syllables (e.g., forgetful) might involve different kinds of processing than reading words with consonant or vowel sequences (e.g., construction, creation) (Eddington, Treiman, & Elzinga, 2013). Acquiring the depth of lexical knowledge that distinguishes skilled readers takes time and experience. Morphology is likely to play an increasingly important role as comprehension of texts involves increasing numbers of morphologically complex words (Nagy & Anderson, 1984). The issues we have reviewed thus far provided the impetus for investigating the findings of recent research on early elementary students’ reading of polymorphemic words.

Studies of polymorphemic word reading We review studies that focus specifically on elementary-age students’ reading of polymorphemic words. Our search of the literature entailed the use of a wide range of keywords related to reading and word structure or morphology (word reading, decoding, morphology, morphological awareness, polymorphemic, polysyllabic, suffixes – to name a few). The selection criteria were as follows: (a) the study was based on an alphabetic writing system and published in English; (b) it included performance on an experimental measure of polymorphemic word reading that was systematically analyzed; and (c) participants were students in the elementary years (K through grade 6). We did not include pre- and post-test measures used in intervention studies. We also did not include lexical decision tasks because they do not assess word reading directly (e.g., Just & Carpenter, 1987; Venezky, 1995). We found 18 studies that fit these criteria. A table in the Appendix provides basic information about each study: an abbreviated citation, a brief description of the

200 Joanne F. Carlisle and Devin M. Kearns

student participants and the polymorphemic word reading measure, and a brief note about results. All of the studies focused on reading derived words except Nagy et al. (2006), which included derivations, inflections, and compound words, and Laxon, Rickard, and Coltheart (1992), which included derived and inflected words. In all of the studies, students were asked to read words in isolation. Through analysis of the 18 studies we identified questions about early elementary students’ reading of polymorphemic words of primary interest to the researchers. We selected six such questions that should provide an overview of recent advances to our understanding of the role of morphology in early word reading. Is there evidence that morphological structure facilitates word reading for elementary students? Morphemes are not simply functioning as orthographic units. Evidence might show that morphemic composition affects word reading speed and accuracy. Laxon et al. (1992) suggested that students need to segment long words into orthographic chunks that lend themselves to orthographic-phonological processing; for polymorphemic words, these chunks lead to recognition of the base word and/or affix. Five of the 18 studies focused on the effects of morphemic structure on the reading of morphologically complex words. One finding is that words made up of two morphemes were read more rapidly and accurately than words made up of one morpheme, when the two types of words were matched on various lexical characteristics. For example, Carlisle and Stone (2005) found that elementary students read derived words (e.g., winner) faster and more accurately than pseudo-derived words (e.g., dinner), matched on spelling, word length, and surface frequency. A second important finding comes from studies that used pseudo-derived words made up of real base words and affixes (that is, novel combinations of two morphemes). The surface form of pseudo-derived words cannot exist in students’ lexical memory so pronunciation of the word must be based on analysis of units within the word. Using polymorphemic words and pseudowords matched on initial phoneme, syllable structure, mean bigram frequency, and orthographic difficulty, Burani and colleagues (Burani, Marcolini, & Stella, 2002; Burani, Marcolini, De Luca, & Zoccolotti, 2008) found that novel derived words with real base forms and affixes were named more rapidly and/or accurately than simple pseudowords and words. Laxon et al. (1992) reported similar results for both pseudo-derived and pseudo-inflected words. A third finding comes from Nunes, Bryant, and Barros’ (2012) large-scale study in which specific kinds of decoding knowledge was required for correct pronunciation of specific words. They found that that 9- and 10-year old students were able to parse words into morphemic constituents to arrive at correct pronunciations. The results of these studies suggest that base words and affixes function as units that early elementary students use in reading both familiar and novel (unfamiliar) words and that processing the morphological composition of words has a positive effect on word reading.



Reading complex words 201

Are there differences in the types of morphemes (base words, suffixes, prefixes) that facilitate word reading? This question focuses on the extent to which morphemic units of different kinds facilitate word reading. Among the 18 studies, four contributed results relevant to this issue (Burani et al., 2002; Colé, Bouton, Leuwers, Casalis & Sprenger-Charolles, 2012; Kearns et al., 2016; Traficante, Marcolini, Luci, Zoccolotti & Burani, 2011). As noted above, Burani et al. (2002) found that Italian students read pseudowords made up of a real base word and suffix (e.g., mammista*, “motherist”) better than morphologically simple pseudowords (e.g., getosto*); their results suggest that base morphemes and affixes are stored in memory and retrieved relatively rapidly during word reading. However, it might be difficult to separate the effects of the base words and suffixes in this study. This problem was addressed by Colé et al. (2012) who found that French pseudowords that are novel combinations of real base words and suffixes (e.g., sauture*) were read faster and more accurately than matched pseudowords made up of a pseudo-base word and real suffix (e.g., seuteur*; seut not a stem), which were read marginally faster than a pseudo-base word and pseudo-suffix (e.g., seutore*, seut not a base word; ore not a suffix). Thus, a real suffix in a novel derived word appears to have facilitated word reading. We note, however, that this finding does not detract from the importance of a familiar base word in these word-reading studies. Support for the special status of the base word comes from Traficante and colleagues (2011), who found that base words have a stronger effect on naming time than suffixes. In another study, Traficante et al. (2014) found that suffix length affected word recognition. The familiarity of the base word is important. Kearns et al. (2016) asked students to read separate lists of base words and their derived counterparts. They found that accurate reading of the derived word was associated with accurate reading of the base word. While the studies cited above focused on derived words with suffixes, two studies included different types of morphologically complex words. Nagy et al. (2006) assessed reading of polymorphemic words with prefixes or pseudo-prefixes, compounds, and derived words; performance was significantly related to students’ morphological awareness. Laxon and colleagues (1992) found that words and pseudowords ending in -er were read more accurately than those ending in -ed. They speculated that variation in the pronunciations of past tense words contributed to this finding (e.g., -ed is prononuced -d in grabbed, -t in tapped, and -ed in batted). Another finding of interest involves grammatical class: Traficante and associates (2014) also found that nouns and verbs might affect derived word recognition differently (e.g., deverbal and denominal nouns). Do students read transparent polymorphemic words more easily than words that undergo phonological shifts? The relation of orthography and phonology at both word and morpheme levels may affect the ease with which students can determine the pronunciation of a polymorphemic word. As noted earlier, transparency of the structure of derived words contributed to students’ performance on oral measures of morphological awareness, so

202 Joanne F. Carlisle and Devin M. Kearns

the question (addressed in five studies) is: is the effect of transparency similar when students are asked to read words aloud (Carlisle, 2000; Carlisle et al., 2001; Deacon, Whalen, & Kirby, 2011; Kearns, 2015; Kearns et al., 2016)? Carlisle (2000) showed a large effect favoring transparent words for third graders (Hedges’ g = 1.00) and a medium effect for fifth graders (Hedges’ g = 0.64). Carlisle et al. (2001) showed an even larger effect (Hedges’ g = 1.95) for students in fourth to ninth grade. Three of the studies also showed interactions. Deacon et al. (2011) showed an interaction of transparency and frequency for fourth graders, indicating that transparency affected accuracy for low frequency but not high frequency words. Kearns (2015) found that for words that undergo phonological or orthographic shifts in the base word (but not transparent words), morphological awareness was related to reading accuracy at all levels of morphological skill. In summary, while transparency appears to affect derived word reading, presumably by obscuring the phonological identity of the base word, this effect may differ based on word characteristics (e.g., frequency; Deacon et al., 2011) and the child’s morphological skills and knowledge (Kearns, 2015; Kearns et al., 2016). Does the frequency of the derived word, the base word, or the word family (i.e., morphological relatives, such as distasteful and tasty) contribute to students’ performance reading polymorphemic words? This question is of interest because measures of frequency provide a way to estimate the extent to which students have previously encountered a word or word family in printed texts. This, is turn, might suggest the likelihood that the words or morphemes are represented in lexical memory. Six studies focused on the effect of word or morpheme familiarity. Kearns et al. (2016) and Kearns (2015) reported that the frequency of the base word mattered in reading of derived words. In Carlisle (2000), third and fifth graders were more accurate reading high than low frequency derived words when both sets of words had high frequency base words, suggesting that surface frequency contributes to the efficiency of word recognition. A different comparison was reported in Deacon, Tong, and Francis (2017); in this study performance was better on derived words with high frequency than low frequency base words. These groups of words had similar surface frequency, underscoring the importance of familiarity of the base word. In comparison, Marcolini et al. (2011) found that skilled sixth-grade readers named low but not high frequency derived words faster than base words, while poor readers showed a speed advantage for morphologically complex words regardless of frequency. This may be because they had not acquired complete orthographic representations for even the high frequency words. Family frequency has also been found to account for variance in derived word reading (Carlisle & Katz, 2006). In short, base word, surface, and family frequency might all be significantly related to reading derived words. If surface frequency was the only measure of familiarity found to be significantly related to derived word reading, we might infer that students did not have or did not access lexical representations of the base words and had not established links among members of word families. However, this was not the case, and so we infer that familiarity with base words and family size do matter.



Reading complex words 203

Is there improvement in the reading of polymorphemic words by grade level? And are skilled elementary readers better at reading such words than less skilled readers? Results of several studies showed that older students were more accurate and faster at naming polymorphemic words than younger students: grades 3 and 5 (Carlisle, 2000); grade-level groups of 4–5, 6–7, and 8–9 (Nagy et al., 2006). They also show that normal readers or skilled readers outperform less skilled readers on tasks of reading derived words (Burani et al., 2008; Carlisle & Katz, 2006; Carlisle et al., 2001; Carlisle & Stone, 2005; Gilbert, Goodwin, Compton, & Kearns, 2013; Kearns et al., 2016; Laxon et al., 1992). Burani et al. (2008) found that skilled younger readers or children with dyslexia benefited more from morphological structure than skilled older readers. Similarly, Marcolini et al. (2011) found that morphological structure resulted in a more pronounced advantage for poor than good readers. In contrast, Traficante et al. (2011) found that frequency of the base word facilitated word naming for both dyslexic and skilled readers. Characteristics of participating students might contribute to these different findings. For example, Kearns et al. (2016) found that late-emerging reading difficulties were associated with a failure to use morphological information to support reading not observed among students with early emerging reading difficulties. Does reading polymorphemic words contribute significantly to measures of students’ vocabulary or reading comprehension? This question provides a way to consider the possible effects of polymorphemic word reading on comprehension processes. Five studies focused in some way on this question. As noted earlier, Nunes, Bryant, and Barros (2012) studied 9- and 10-year old students’ ability to read derived words that required different kinds of decoding knowledge. They found that both large graphophonic and morphemic units significantly predicted students reading comprehension; of the two unit types, morphemes were the stronger predictors. Carlisle (2000) examined the contribution of third and fifth graders’ derived word reading, derived word definitions, and morphological awareness to reading comprehension and vocabulary. While derived word reading made a significant unique contribution to third graders’ reading comprehension and vocabulary and fifth graders vocabulary, for both literacy areas a large portion of the variance that was accounted for by the three morphology measures was shared. This outcome suggests that students’ development of polymorphemic decoding skill and knowledge of word meanings together contribute to their ability to understand words and texts. Nagy et al. (2006) examined the effects of polymorphemic word reading and morphological awareness on reading comprehension and vocabulary. In a Structural Equation Model, they found no direct effect of polymorphemic word reading on either reading comprehension or vocabulary. Instead, skill in decoding (including multiple types of polymorphemic words) was directly related to morphological awareness and had an indirect effect on reading comprehension mediated through morphological awareness. Somewhat similarly, Deacon and her colleagues (2017) found that for 3rd and 5th graders the combined effect of morphological structure awareness, derived word reading, and morphological analysis accounted for 8% of the variance in reading comprehension.

204 Joanne F. Carlisle and Devin M. Kearns

Gilbert et al. (2013) examined the contribution of morphological awareness and derived word reading to comprehension achievement for fifth graders with and without reading difficulty. When both morphological awareness and derived word reading were entered into the model as main effects, neither had a significant relation with reading comprehension, but the interaction was significant. Further analyses showed that for fifth graders with good polymorphemic word reading skill, morphological awareness predicted reading comprehension, but for fifth graders with poor polymorphemic word reading, this was not the case. The less able polymorphemic word readers might have lacked the kind of morphological knowledge that would facilitate reading such words. The relation of polymorphemic word reading skill and reading comprehension has received less attention than seems warranted. Longitudinal studies are needed to understand the extent to which gains in reading comprehension over time are explained by students’ ability to read polymorphemic words.

What have we learned? And what next? Examination of the 18 studies that included measures of polymorphemic word reading have provided confirmation that the ability to read polymorphemic words is an issue of some importance for the further development of theories of learning to read. The major finding is that morphemes contribute to the speed and accuracy of reading morphologically complex words in the early elementary years. Morphological structure affects students’ reading of not only familiar words but also novel or low frequency words (e.g., Burani et al., 2008). The results also shed light on the nature of units used in reading polymorphemic words. That is, base morphemes appear to be most likely to facilitate rapid and accurate word reading, but suffixes attached to novel base words or pseudo-words also contribute to the processing of the internal structure of polymorphemic words. Importantly, there is evidence that base morphemes (e.g., win in winner) facilitate word reading more than syllables in the same word context (e.g., din in dinner). Units of meaning make a difference in early word reading. Overall, the studies we reviewed provide support for Frost’s (2012) proposal that orthography, phonology, and morphology work in concert in the development of word reading. The results point to some factors that negatively affect the facilitation of morphological structure in word reading. In particular, suffixation that affects the phonological transparency of a base word might lead to difficulties formulating a pronunciation of morphologically complex words, particularly low frequency words. Such difficulties might linger into the middle school years for some readers (e.g., Carlisle & Stone, 2005) and warrant further study. To date, the relation of polymorphemic word reading and text comprehension has received relatively little attention. Another important gap involves reconciling differences in views of the age or stage of reading acquisition at which morphological



Reading complex words 205

knowledge is needed. We might benefit from efforts to map developmental changes in the kinds of challenges young readers face in reading polymorphemic words from the early elementary years to the middle school years. While results showed that students’ reading of polymorphemic words improved by grade level (Carlisle, 2000; Deacon et al., 2011; Nagy et al., 2006), there has been relatively little attention to the nature of changes in processes involved in reading polymorphemic words over time. We close by considering directions research might take to develop and test theories of elementary students’ reading of polymorphemic words, building on what we have learned so far. We acknowledge that research in this area is in its early days, and there are many questions that need to be answered. We have chosen three that seem particularly valuable and promising. First, we see the need for studies that examine the relation of polymorphemic word reading and literacy development, including measures of vocabulary and reading comprehension. Such studies might examine a number of student characteristics and word characteristics that are likely to affect the development of word reading (monomorphemic and polymorphemic). Longitudinal studies would be particularly helpful, as it seems likely that the relation of polymorphemic word reading and reading comprehension change as students move through the elementary years. Second, we suggest exploring ways to adapt or expand the concept of unitization to include morphemic elements in early word reading. We note the potential value of theories that integrate processing of orthographic, phonological, and morphological characteristics of words. Studies of sight word reading might be balanced by qualitative analyses of students’ approaches to reading and understanding morphologically complex words – the long and hard words that struggling readers prefer to skip over. In addition, educators would certainly benefit from research that looks at the decoding demands of both monomorphemic and polymorphemic words in the texts they ask students to read in the early and late elementary years. Third, we suggest that more attention be paid to students’ reading and understanding of morphologically complex words in sentences and passages. Perfetti’s (2007) argument for the importance of lexical knowledge is very convincing, but it cannot be studied using conventional measures of reading words in isolation. Instead, we need studies that examine word reading in natural contexts designed to help us understand how students learn to integrate the form and meaning of polymorphemic words as they read. One promising method, used by Kuperman and Van Dyke (2011) in tracking adult readers’ eye movements and fixations, focused on recognition of suffixed words. The approach could be used with beginning readers as well. Use of eye tracking would provide a way to examine the effects of both passage variables and child characteristics on word reading accuracy and efficiency. For example, we wonder whether the density of polymorphemic words in elementary texts affects word reading and passage comprehension. Clearly, there are other issues that hold the potential of contributing to the development of theories of polymorphemic word reading. It seems apparent to us that

206 Joanne F. Carlisle and Devin M. Kearns

research in this area would not only advance our knowledge of elementary students’ reading of polymorphemic words but also contribute more generally to models that seek to account for changes in reading and understanding words across the schoolage years.

References Adams, M. J. (1990). Beginning to read: Thinking and learning about print. Cambridge, MA: MIT Press. Anglin, J. M. (1993). Vocabulary development: A morphological analysis. Monographs of the society for research in child development, 58, 1–186.  doi: 10.2307/1166112 Apel, K., Diehm, E., & Apel, L. (2013). Using multiple measures of morphological awareness to assess the relation to reading, Topics in Language Disorders, 33, 42–56. doi: 10.1097/TLD.0b013e318280f57b Berninger, V. W., Abbott, R. D., Nagy, W., & Carlisle, J. (2010). Growth in phonological, orthographic, and morphological awareness in grades 1–6. Journal of Psycholinguistic Research, 39, 141–163.  doi: 10.1007/s10936-009-9130-6 Brown, R. (1973). A first language: The early stages. Cambridge: Harvard University Press. doi: 10.4159/harvard.9780674732469 Burani, C., Marcolini, S., & Stella, G. (2002). How early does morpholexical reading develop in readers of a shallow orthography? Brain and Cognition, 81, 568–586. Burani, C., Marcolini, S., De Luca, M., & Zoccolotti, P. (2008). Morpheme-based reading aloud: Evidence from dyslexic and skilled Italian readers. Cognition, 108, 243–262. doi: 10.1016/j.cognition.2007.12.010 Carlisle, J. F. (2000). Awareness of the structure and meaning of morphologically complex words: Impact on reading. Reading and Writing: An Interdisciplinary Journal, 12, 169–190. doi: 10.1023/A:1008131926604 Carlisle, J. F., & Fleming, J. (2003). Lexical processing of morphologically complex words in the elementary years. Scientific Studies of Reading, 7, 239–254.  doi: 10.1207/S1532799XSSR0703_3 Carlisle J. F., & Katz, L. A. (2006). Effects of word and morpheme familiarity on reading of derived words. Reading and Writing, 19, 669–693.  doi: 10.1007/s11145-005-5766-2 Carlisle, J. F., & Nomanbhoy, D. M. (1993). Phonological and morphological awareness in first graders. Applied Psycholinguistics, 14, 177–195.  doi: 10.1017/S0142716400009541 Carlisle, J. F., & Stone, C. A. (2005). Exploring the role of morphemes in word reading. Reading Research Quarterly, 40, 428–449.  doi: 10.1598/RRQ.40.4.3 Carlisle, J. F., Stone, C. A., & Katz, L. A. (2001). The effects of phonological transparency on reading derived words. Annals of Dyslexia, 51, 249–274.  doi: 10.1007/s11881-001-0013-2 Casalis, S., Colé, P., & Sopo, D. (2004). Morphological awareness in developmental dyslexia. Annals of Dyslexia, 54, 114–138.  doi: 10.1007/s11881-004-0006-z Casalis, S., Dusautoir, M., Colé, P., & Ducrot, S. (2009). Morphological effects in children word reading: A priming study in fourth graders. British Journal of Developmental Psychology, 27, 761–766.  doi: 10.1348/026151008X389575 Clark, E. (1982). The young word maker: A case study of innovation in the child’s lexicon. In E. Wanner & L. Gleitman (Eds.), Language acquisition: The state of the art (pp. 391–425). Cambridge: Cambridge University Press.



Reading complex words 207

Colé, P., Bouton, S., Leuwers, C., Casalis, S., & Sprenger-Charolles, L. S. (2012). Stem and derivational-suffix processing during reading by French second and third graders. Applied Psycholinguistics 33, 97–120.  doi: 10.1017/S0142716411000282 Cunningham, A. E. (2010). Vocabulary growth through independent reading and reading aloud to children. In E. H. Hiebert & M. L. Kamil (Eds.), Teaching and learning vocabulary: Bringing research to practice (pp. 45–68). NY: Routledge. Deacon, S. H. (2012a). Sounds, letters, and meanings: The independent influences of phonological, orthographic, and morphological skills on early word reading accuracy. Journal of Research in Reading, 35, 456–475.  doi: 10.1111/j.1467-9817.2010.01496.x Deacon, S. H. (2012b). Bringing development into a universal model of reading. Behavioral and Brain Sciences, 35, 284.  doi: 10.1017/S0140525X12000040 Deacon, S. H., Tong, X., & Francis, K. (2017). The relationship of morphological analysis and morphological decoding to reading comprehension. Journal of Research in Reading 1, 1–16.  doi: 10.1111/1467-9817.12056 Deacon, S. H., Kieffer, M. J., & Laroche, A. (2014). The relation between morphological awareness and reading comprehension: Evidence from mediation and longitudinal models. Scientific Studies of Reading, 18, 432–451.  doi: 10.1080/10888438.2014.926907 Deacon, S. H., & Kirby, J. R. (2004). Morphological awareness: Just “more phonological”? The roles of morphological and phonological awareness in reading development. Applied Psycholinguistics, 25, 223–238.  doi: 10.1017/S0142716404001110 Deacon, S. H., Whalen, R., & Kirby, J. R. (2011). Do children see the danger in dangerous? Grade 4, 6, and 8 children’s reading of morphologically complex words. Applied Psycholinguistics, 32, 467–481.  doi: 10.1017/S0142716411000166 De Luca, M., Barca, L., Burani, C., & Zoccolotti, P. (2008). The effect of word length and other sublexical, lexical, and semantic variables on developmental reading deficits. Cognitive Behavior Neurology, 21, 227–235.  doi: 10.1097/WNN.0b013e318190d162 Eddington, D., Treiman, R., & Elzinga, D. (2013). Syllabification of American English: Evidence from a large-scale Experiment, Part II. Journal of Quantitative Linguistics, 20, 75–93. doi: 10.1080/09296174.2013.773136 Ehri, L. C. (1995). Phases of development in learning to read words by sight. Journal of Research in Reading, 18, 116–125.  doi: 10.1111/j.1467-9817.1995.tb00077.x Ehri, L. C. (2005). Learning to read words: Theory, findings and issues. Scientific Studies of Reading, 9, 167–188.  doi: 10.1207/s1532799xssr0902_4 Fitzgerald, J., Elmore, J, Koons, H., Hiebert, E. H., Bowen, K., Sanford-Moore, E. E., & Stenner, A. J. (2015). Important text characteristics for early grades text complexity. Journal of Educational Psychology, 107, 4–29.  doi: 10.1037/a0037289 Fowler, A., & Liberman, I. (1995). The role of phonology and orthography in morphological awareness. In L. Feldman (Ed.), Morphological aspects of language processing (pp.131–154). Hillsdale, NJ: Lawrence Erlbaum. Frost, R. (2012). Towards a universal model of reading. Behavioral and Brain Sciences, 35, 263–329. doi: 10.1017/S0140525X11001841 Frost, R., Grainger, J., & Carreiras, M. (2008). Advances in morphological processing: An introduction. Language and Cognitive Processes, 23, 933–941.  doi: 10.1080/01690960802321925 Gagl, B., Hawelka, S., & Wimmer, H. (2015). On sources of word length effect in young readers. Scientific Studies of Reading, 19, 289–306.  doi: 10.1080/10888438.2015.1026969 Gilbert, J. K., Goodwin, A. P., Compton, D. L., & Kearns, D. M. (2013). Multisyllabic word reading as a moderator of morphological awareness and reading comprehension. Journal of Learning Disabilities, 37, 34–43.

208 Joanne F. Carlisle and Devin M. Kearns

Goswami, U., & Ziegler, J. C. (2006). Fluency, phonology, and morphology: A response to the commentaries on becoming literate in different languages. Developmental Science, 9, 451–453. doi: 10.1111/j.1467-7687.2006.00511.x Jarmulowicz, L., Taran, V. L., & Hay, S. E. (2007). Third graders’ metalinguistic skill, reading skills, and stress production in derived English words. Journal of Speech, Language, and Hearing Research, 50, 1593–1605.  doi: 10.1044/1092-4388(2007/107) Just, M. A., & Carpenter, P. A. (1987). The psychology of reading and language comprehension. Needham Heights, MA: Allyn & Bacon. Kearns, D. M. (2015). How elementary-age children read polysyllabic polymorphemic words. Journal of Educational Psychology, 107, 364–390.  doi: 10.1037/a0037518 Kearns, D. M., Steacy, L. M., Compton, D. L., Gilbert, J. K., Goodwin, A. P., Cho, E., …Collins, A. A. (2016). Modeling polymorphemic word recognition: Exploring differences among children with early-emerging and late-emerging word reading difficulty. Journal of Learning Disabilities 49, 368–394.  doi: 10.1177/0022219414554229 Kirby, J. R., Deacon, S. H., Bowers, P. N., Izenberg, L., Wade-Woolley, L., & Parrila, R. (2012). Children’s morphological awareness and reading ability. Reading and Writing: An Interdisciplinary Journal, 25, 389–410.  doi: 10.1007/s11145-010-9276-5 Kuo, L. J., & Anderson, R. C. (2003). Morphological awareness and learning to read: A cross-language perspective. Educational Psychologist, 41, 161–180.  doi: 10.1207/s15326985ep4103_3 Kuperman, V. & Van Dyke, J. A. (2011). Individual differences in visual comprehension of morphological complexity. In L. Carlson, C. Hoelscher, & T. Shipley (Eds.), Proceedings of the 33rd Annual Conference of the Cognitive Science Society, (pp. 1643–1648). Austin, TX: Cognitive Science Society. Laxon, V., Masterson, J., & Moran, R. (1994). Are children’s representations of words distributed? Effects of orthographic neighbourhood size, consistency and regularity of naming. Language and Cognitive Processes, 9, 1–27.  doi: 10.1080/01690969408402108 Laxon, V., Rickard, M., & Coltheart, V. (1992). Children read affixed words and nonwords. British Journal of Psychology, 83, 407–423.  doi: 10.1111/j.2044-8295.1992.tb02450.x Marcolini, S., Traficante, D., Zoccolotti, P. & Burani, C. (2011). Word frequency modulates morpheme-based reading in poor and skilled Italian readers. Applied Psycholinguistics, 32, 513–532.  doi: 10.1017/S0142716411000191 Mattingly, I G. (1984). Reading, linguistic awareness, and language acquisition. In J. Downing & R. Valtin (Eds.), Language awareness and learning to read (pp. 9–25). NY: Springer-Verlag. doi: 10.1007/978-1-4613-8248-5_2 Minarik, E. H. (1959). Father bear comes home. NY: Harper and Row. Nagy, W. E., & Anderson, R. C. (1984). How many words are there in printed school English? Reading Research Quarterly, 19, 304–330.  doi: 10.2307/747823 Nagy, W., Berninger, V., & Abbott, R. (2006). Contributions of morphology beyond phonology to literacy outcomes of upper elementary and middle school students. Journal of Educational Psychology, 98, 134–147.  doi: 10.1037/0022-0663.98.1.134 Nagy, W. E., Diakidoy, I. N., & Anderson, R. C. (1993). The acquisition of morphology: Learning the contribution of suffixes to the meanings of derivatives. Journal of Reading Behavior, 25, 155–170. Nagy, W. E., Anderson, R. C., Schommer, M., Scott, J. A., & Stallma, A. C. (1989). Morphological families in the internal lexicon. Reading Research Quarterly, 24, 262–282.  doi: 10.2307/747770



Reading complex words 209

Nagy, W., & Scott, J. A. (2000). Vocabulary processes. In M. L. Kamil, P. Mosenthal, P. D. Pearson, & R. Barr (Eds.), Handbook of reading research, Vol. 3 (pp. 269–284). Mahwah, NJ: Erlbaum. Nunes, T., Bryant, P., & Barros, R. (2012). The development of word recognition and its significance for comprehension and fluency. Journal of Educational Psychology, 104, 959–973. doi: 10.1037/a0027412 Nunes, T., Bryant, P., & Bindman, M. (1997). Morphological spelling strategies: Developmental stages and processes. Developmental Psychology, 33, 637–649.  doi: 10.1037/0012-1649.33.4.637 Pasquarella, A., Deacon, H., Chen, B. X., Commissaire, E., & Au-Yeung, K. (2014). Acquiring orthographic processing through word reading: Evidence from children learning to read French and English. International Journal of Disability, Development and Education, 61, 240–257.  doi: 10.1080/1034912X.2014.932579 Perfetti, C. A. (2007). Reading ability: Lexical quality to comprehension. Scientific Studies of Reading, 11, 357–383.  doi: 10.1080/10888430701530730 Quémart, P., Casalis, S., & Colé, P. (2011). The role of form and meaning in the processing of written morphology: A priming study in French developing readers. Journal of Experimental Child Psychology, 109, 478–496.  doi: 10.1016/j.jecp.2011.02.008 Quémart, P., Casalis, S., & Duncan, L. G. (2012). Exploring the role of bases and suffixes when reading familiar and unfamiliar words: Evidence from French young readers. Scientific Studies of Reading, 16, 424–442.  doi: 10.1080/10888438.2011.584333 Roman, A. A., Kirby, J. R., Parrila, R. K., Wade-Woolley, L., & Deacon, S. H. (2009). Toward a comprehensive view of the skills involved in word reading in grades 4, 6, and 8. Journal of Experimental Child Psychology, 102, 96–113.  doi: 10.1016/j.jecp.2008.01.004 Rubin, H., Patterson, P., & Kantor, M. (1991). Morphological development and writing ability in children and adults. Language, Speech, and Hearing Sciences in Schools, 22, 228–235. doi: 10.1044/0161-1461.2204.228 Schreuder, R., & Baayen, R. H. (1995). Modeling morphological processing. In L. B. Feldman (Ed.), Morphological aspects of language processing (pp. 131–154). Hillsdale, NJ: L. Erlbaum. Seymour, P. H. K. (1997). Foundations of orthographic development. In C. A. Perfetti, L. Rieben, & M. Fayol (Eds.), Learning to spell: research, theory, and practice among languages (pp. 319–337). Hillsdale, NJ: Erlbaum. Singson, M., Mahony, D., & Mann, V. (2000). The relation between reading ability and morphological skills: Evidence from derivational suffixes. Reading and Writing: An Interdisciplinary Journal, 12, 219–252.  doi: 10.1023/A:1008196330239 Snow, C. E. (2010). Academic language and the challenge of reading for learning about science. Science, 328, 450–452.  doi: 10.1126/science.1182597 Sparks, E., & Deacon, S. H. (2012). Morphological awareness and vocabulary acquisition: A longitudinal examination of their relationship in English-speaking children. Applied Psycholinguistics, 36, 299–321.  doi: 10.1017/S0142716413000246 Spencer, M., Muse, A., Wagner, R. K., Foorman, B., Petscher, Y., Schnatschneider, C., …Bishop, D. M. (2015). Examination of the underlying dimensions of morphological awareness and vocabulary knowledge. Reading and Writing: An Interdisciplinary Journal, 28, 959–988. doi: 10.1007/s11145-015-9557-0 Templeton, S. (1991). Teaching and learning the English spelling system: Reconceptualizing method and purpose. Elementary School Journal, 92, 185–201.  doi: 10.1086/461687

210 Joanne F. Carlisle and Devin M. Kearns

Traficante, D. Marcolini, S., Luci, A., Zoccolotti, P., & Burani, C. (2011). How do roots and suffixes influence reading of pseudowords: A study of young Italian readers with and without dyslexia. Language and Cognitive Processes, 26, 777–793.  doi: 10.1080/01690965.2010.496553 Traficante, D., Marelli, M., Luzzatti, C., & Burani, C. (2014). Influence of verb and noun bases on reading aloud derived nouns: Evidence from children with good and poor reading skills. Reading and Writing: An Interdisciplinary Journal, 27, 1303–1326. doi: 10.1007/s11145-013-9488-6 Treiman, R., & Cassar, M. (1996). Effects of morphology on children’s spelling of final consonant clusters. Journal of Experimental Child Psychology, 63, 141–170.  doi: 10.1006/jecp.1996.0045 Tunmer, W. E., & Herriman, M. L. (1984). The development of metalinguistic awareness: A conceptual overview. In W. E. Tunmer, C. Pratt, & M. L. Herriman (Eds.), Metalinguistic awareness in children (pp. 12–35). New York: Springer-Verlag.  doi: 10.1007/978-3-642-69113-3_2 Tyler, A., & Nagy, W. E. (1989). The acquisition of English derivational morphology. Journal of Memory and Language, 28, 649–667.  doi: 10.1016/0749-596X(89)90002-8 Tyler, A., & Nagy, W. E. (1990). Use of derivational morphology during reading. Cognition, 36, 17–34.  doi: 10.1016/0010-0277(90)90052-L Venezky, R. L. (1995). How English is read: Grapheme phoneme regularity and orthographic structure in word recognition. In J. Taylor & D. R. Olson (Eds.), Scripts and literacy: Reading and learning to read alphabets, syllabaries, and characters (pp. 111–129). Netherlands: Kluwer Academic Publishers.  doi: 10.1007/978-94-011-1162-1_8 Verhoeven, L., Schreuder, R., & Baayen, R. H. (2006). Learnability of graphotactic rules in visual word identification. Learning and Instruction, 16, 538–548. doi: 10.1016/j.learninstruc.2006.10.003 White, E. B. (1952). Charlotte’s web. NY: Harper and Brothers. Wolter, J. A., Wood, A., & D’zatko, K. W. (2008). The influence of morphological awareness on the literacy development of first-grade children. Language, Speech and Hearing Services in Schools, 40 286–298.  doi: 10.1044/0161-1461(2009/08-0001) Yap, M. J., & Balota, D. A. (2009). Visual word recognition of multisyllabic words. Journal of Memory and Language, 60, 502–529.  doi: 10.1016/j.jml.2009.02.001



Reading complex words 211

Appendix Table 1.  Assessments of elementary students’ reading of polymorphemic words Author(s) & Date

Grade/skill status Polymorphemic reading task(s)

Major results

Burani, Marcolini, & Stella, 2002

8 and 10 year old students

Word naming task

Pseudowords made up of roots and derivational suffixes named more quickly and accurately than matched pseudowords with no morphological constituents

Burani, Marcolini, DeLuca & Zoccolotti, 2008

3 groups: skilled younger (Gr 2–3), skilled older (Gr 6), and dyslexic); adults

Naming latencies to pseudowords composed of roots and derived suffixes; simple pseudowords and words

All four groups faster and more accurate on pseudowords made up of root and suffix than simple pseudowords. Skilled younger readers and dyslexics (but not skilled older or adults) benefited from morphological structure.

Carlisle, 2000

3rd and 5th graders

Word Reading Test (WRT): 2 sets of derived words: Set A: High Frequency (HF) derived words that varied in transparency Set B: transparent words with HF bases and either HF or low frequency surface. Other measures: DEF (Definition of derived words) and TMS (Test of Morphological Structure)

For WRT, 5th graders more accurate than 3rd; for both grades on Set A: Transparent more accurate than Shift words; for Set B: HF more accurate than LF derived words. For 3rd graders, WRT, DEF and TMS accounted for significant variance in vocabulary and reading comprehension; WRT made unique contribution; for Gr 5, WRT, DEF and TMS accounted for significant variance in vocabulary and reading comprehension. WRT made significant contribution to vocabulary but not reading comprehension.

Carlisle & Katz, 4th and 6th Reading Complex Words 2006 graders, skilled and (RCW): two word lists (89 less skilled words in all). Included transparent low frequency derived, high frequency base words varying in transparency; grouped by family size, average family frequency, and derived frequency

Factor analysis suggested 2 factors: morphemic composition (base frequency, average family frequency) and experience with word and word family (derived word frequency, family size). Among good readers, 6th graders better than 4th on RCW; poor readers 6th grade did not differ from 4th.

Carlisle, Stone & Katz, 2001

The effects of phonological transparency were greater for the poor readers than the average readers on the Word Naming than the Lexical Decision task

Grades 4–9, average and poor readers; adults

Word Naming: Computer presentation of shift and stable words; lexical decision task used for comparison

(continued)

212 Joanne F. Carlisle and Devin M. Kearns

Table 1.  (continued) Author(s) & Date

Grade/skill status Polymorphemic reading task(s)

Major results

Carlisle & Stone, 2005, Study 1

Lower el (Gr 2–3) and upper el (Gr 5–6

Word Reading, Set 1: High frequency derived words (e.g., shady) and matched pseudo-derived words (e.g., lady)

(a) Lower elementary students faster on derived than pseudo-derived words; no difference in upper elementary; b) for lower elementary, number of syllables contributed to derived word accuracy and speed; for upper elementary, number of syllables and base word frequency contributed to derived word reading

Colé, Bouton, et al., 2012

2nd and 3rd graders

Experiment 1: reading ”words” that included (a) “illegally” combined real stem and suffix; (b) pseudowords with pseudo-stem and real suffix; (c) pseudowords with pseudo-stem and pseudo suffix. Experiment 2: effect of display time on reading derived words; displayed by syllable, morpheme, morpheme + 1 grapheme, and unsegmented

Experiment 1: Ss read illegally combined real stem and real suffix words faster and more accurately than matched pseudowords with pseudostem; pseudo-stem – real suffix words were read faster and more accurately than pseudo-stem with pseudo-suffix words. Experiment 2: Morpheme + 1 generated the longest reading times; no difference among other three conditions. Syllables, morphemes, and whole words contribute to reading low-frequency derived words.

Deacon, Tong & Third and fifth Francis, 2017 graders

Derived word decoding: half low and half high frequency base; no difference surface frequency or word length

Together morphological structure awareness, decoding and analysis accounted for a significant 8% of the variance in reading comprehension after accounting for other relevant word and child characteristics

Deacon, Whalen, & Kirby, 2011

Derived word naming: 64 two- morpheme words, 8 in each set, varied in frequency and transparency

For all three grade-level groups, performance was faster on derived words with high than with low frequency base words when words were of low surface frequency. Frequency and transparency of the base forms affected accuracy of naming for Gr 4 and 6 but not Gr 8 students.

Polysyllabic polymorphemic word reading: 30 derived words

Morphological awareness had an effect on reading comprehension for children with poor but not good polymorphemic word reading skills.

Grades 4, 6, 8

Gilbert, 5th graders with Goodwin et al., poor readers 2013 over-sampled



Reading complex words 213

Table 1.  (continued) Author(s) & Date

Grade/skill status Polymorphemic reading task(s)

Major results

Kearns et al., 2016

5th graders with poor readers over-sampled (early-emerging and late-emerging word reading problems)

Polysyllabic polymorphemic word reading: 30 derived words. Morphological awareness task: a composite of three suffix choice tasks (Nagy et al, 2003) and the Morphological Relatedness Task items read from a list

Results showed child-by-word effects of root word reading and self-reported familiarity with spoken word; word effect of word frequency; child effects of disability status, morphological awareness, and orthographic choice. Morphological awareness related to word reading more for the earlyemerging group than the late-emerging group. The child-by-item effect of root word reading was larger when words were morphemically transparent.

Kearns, 2015

4th graders

Polysyllabic polymorphemic word reading: 48 words from a list; Morphological awareness task was Carlisle (2000) derivation task

For a polymorphemic model, main effects of word-specific root word reading, general root word reading, vocabulary, and word frequency. There was a morphological awareness by morphemic transparency interaction in that morphological awareness had a greater effect on shift than transparent words. For a combined polysyllabic and polymorphemic model, additional findings were that students’ GPC knowledge affected reading of low frequency words; a bigram frequency effect was also found. At the child level: GPC knowledge, phonological awareness, morphological awareness, and general root word reading skill were significant. Significant interactions indicated that GPC knowledge only helped reading very low frequency words, and morphological awareness was more helpful when words are transparent.

Test of reading affixed words and non-words, ending in -er or -ed

Better and poorer readers read affixed words more accurately than pseudo-derived words and better than pseudowords with a word stem.

Laxon, Rickard, 7, 8, and 9 year& Coltheart, olds divided into 1992 better and poorer readers Marcolini, Traficante, Zoccolotti & Burani, 2011

Good and poor 6th Word reading test: 4 sets graders of 20 words – high and low frequency, derived and simple

Word frequency interacted with naming ability. Skilled readers named low frequency but not high frequency derived words faster than simple words. Poor readers showed an advantage for derived words, regardless of frequency.

(continued)

214 Joanne F. Carlisle and Devin M. Kearns

Table 1.  (continued) Author(s) & Date

Grade/skill status Polymorphemic reading task(s)

Major results

Nagy, Berninger, & Abbott, 2006

3 groups: 4th & 5th graders; 6th and 7th graders; 8th and 9th graders

Path analysis showed a significant contribution of morphological awareness to reading comprehension, vocabulary, and spelling; it also was related to decoding rate for 8th-9th graders and some decoding accuracy measures for 4th-5th graders and 8th9th graders.

4 reading measures: Decoding inflected words, Decoding prefixed irregular stems, Decoding suffixed irregular stems, Decoding sets of morphologically related words. Morphological Relatedness Test assessed morphological awareness.

Nunes, Bryant, 9 and 10 year olds Decoding words, 3 types & Barros, 2012 of knowledge needed: Type 1 letter-sound, Type 2 graphophonic, Type 3 morphological.

Use of larger graphophonic and morpheme units made independent contributions to prediction of reading comprehension (with morphemes the stronger predictor).

Traficante, Marcolini, Luci, Zoccolliti, & Burani, 2011

6th graders with dyslexia and skilled readers age-matched

Reading 4 types of complex pseudowords: 2 morphemic units (root + suffix), root + no-suffix, no-root and suffix, and noroot + no-suffix.

Examining the effects of root vs. suffix: for both groups, the root but not the suffix affected pronunciation. In particular, roots had strong effect on naming time, whereas suffixes did not.

Traficante, Marelli, Luzzatti, & Burani, 2014

4th and 5th grade good and poor readers

Derived word reading: noun and verb base words

Frequency and word length differentually affected reading of words in different grammatical classes; suffix length had an inhibatory effect.

Learning to read in a second language Ludo Verhoeven

Radboud University Nijmegen

The results of neurocognitive studies show second language (L2) reading processes to highly resemble first language (L1) reading processes (cf. Indefrey, 2006). Representation, unification, and control appear to be the basic components in both L1 and L2 reading processes. Representation refers to the identification of words in the mental lexicon; unification refers to the integration of words into syntactic clauses, sentences, and text; and control refers to the monitoring and updating of lexical and syntactic processing in working memory. Notwithstanding the involvement of the same neural networks in L1 and L2 reading processes, substantial differences exist as well. L2 reading, for example, may often require additional neural resources relative to L1 reading. This need for additional resources may be due to limited L2 lexical abilities, syntactic abilities, or oral language comprehension abilities (Verhoeven, 2010). Adult L2 reading processes often require greater working memory than adult L1 reading processes. This additional processing need, however, has been found to relate to the age of the learner, degree of L2 mastery, and the extent of L2 exposure. Greater working memory is needed in cases of late as opposed to early L2 acquisition, for example, and under conditions of limited exposure, in particular when low-proficiency L2 learners are studied (Sebastian, Laird, & Kiran, 2011). From a developmental perspective, it is clear that learning to read in a second language can be a challenging task – most often because L2 learners are doing so in a language that they have not yet fully mastered. Census data show that less than 30% of English-language learners from schools in the USA meet statewide norms on reading comprehension tasks (NAEP, 2005). The question that then arises is just how the differences in reading outcomes for L1 versus L2 learners emerge and evolve during elementary school years and how the individual variation in children’s reading development during these years can be explained. In this chapter, I will attempt to answer these questions. To start, I will consider the foundations for the development of early language and literacy skills. Special attention will be paid to the topic of metalinguistic awareness among bilingual children since it is assumed that exposure to more than one language may foster such awareness. Furthermore, I will review what is known about the development of the essential skills of word decoding and reading comprehension in a second language. For each of these skills, I will explore how linguistic diversity can influence the course doi 10.1075/swll.15.12ver © 2017 John Benjamins Publishing Company

216 Ludo Verhoeven

of its development. Finally, I will explore the role of L1 in L2 reading development, and close with consideration of the educational implications of linguistic diversity for learning to read.

Foundations of early language and literacy skills Emergent bilingualism can be characterized as the development of children’s first language during the preschool years as the result of linguistic input in L1-speaking homes and later a second language via L2 input from playmates and at school. Several studies show that growing up in a bilingual context affects the development of phonological and lexical abilities of young children in a variety of ways. In a large-scale study, Ford, Cabell, Konold, Invernizzi, and Gartland (2013) also noted considerable heterogeneity among Hispanic ESL students in their cognitive and linguistic abilities upon kindergarten entry. Poulin-Dubois, Bialystok, Blaye, Polonia, and Yott (2013) compared the expressive and receptive vocabularies of monolingual versus bilingual toddlers and found significantly different expressive vocabulary sizes but similar receptive vocabulary sizes. In a study of the emergence of translation equivalents in Spanish-Catalan bilingual 18-month-olds, phonological form proximity was found to facilitate early acquisition (Bosch & Ramon-Casas, 2014). Similarly when Parra, Hoff, and Core (2011) examined the associations of language experience with phonological memory, they found memory for English-like nonwords to correlate highly with memory for Spanish-like nonwords and each to relate to the more general development of the children’s vocabularies and grammars in the two languages. Several studies have stressed the importance of straightforward language exposure for the development of language and literacy. When Cattani and colleagues (2014) examined how much exposure to English was necessary for a bilingual toddler to perform like a monolingual toddler on language tests, they found 60% or more exposure to English – as reported by the parents – to be sufficient for all language measures. Place and Hoff (2011) examined the predictors of English and Spanish language development among bilingual 25-month-olds. They found that not only the amount of exposure but also the number of speakers from whom English was heard and the extent of input provided by native speakers predicted the language outcomes for both English and Spanish. Similarly, Hurtado, Gruter, Marchman, and Fernald (2014) found the relative amount of exposure to Spanish versus English during daily interactions to predict the relative efficiency of real-time language processing and vocabulary development of toddlers in each language. Finally, the effects of early education programs and practices on the foundations for early L2 language and literacy skills were considered in a meta-analysis conducted by Buysse, Peisner-Feinberg, Paez, Hammer, and Knowles (2014). The features of the early education programs and practices, including the language and intensity of instruction, and the features of the research methods, including sampling and study design, were examined in relation to the child outcomes for various intervention



Learning to read in a second language 217

program types (center based, professional development based, curricula based, instructional strategy based). The results indicated that young L2 learners benefitted from such widely available, well-regulated programs as Head Start and public preschool. However, very little could be concluded about the specific language of instruction or type of intervention provided because these features were not taken into account systematically in the available studies. All of these studies highlight the importance of language exposure and language intervention during the preschool years. They also stress the importance of evaluating the foundation of language and literacy skills at the entrance of kindergarten.

Bilingualism and metalinguistic awareness The fact that bilingual children deal with inputs from two different languages has led to the suggestion that bilingual children may have higher levels of metalinguistic awareness relative to monolingual peers. To find evidence for this claim, Marinova-Todd, Zhao, and Bernhardt (2010) compared the performance of Mandarin-English bilingual children on tests of phonological awareness with that of Mandarin and English monolingual children. The Mandarin-English bilinguals performed better than the monolingual groups on tests of both Mandarin and English phonological awareness. Likewise, Kuo and Anderson (2012) examined the effects of early Chinese-English bilingualism on the learning of English L2 phonological regularities and found bilingual children in Taiwan, regardless of whether they actively used the second language at home or simply had exposure to it, showed an advantage over monolingual Englishor Chinese-speaking peers. Janssen, Bosman, and Leseman (2013) similarly found bilingual Turkish-Dutch children to show an advantage over monolingual Dutchspeaking children on Dutch phoneme-awareness tasks. Instruction on L1 phonological awareness has also been shown to produce positive effects for both L1 (Spanish) and L2 (English) phonological awareness (Gorman, 2012). Overall, it can be concluded that bilingualism strengthens the development of phonological awareness and that explicit instruction in either L1 or L2 can boost the development of phonological awareness in both L1 and L2. To consider the further generality of these findings, we need to reconsider how particular patterns of bilingual development relate to the development of various aspects of children’s phonological awareness. When Verhoeven (2007) investigated the associations of early bilingualism with the development of phonological awareness of Turkish-Dutch bilingual children in kindergarten, phonotactic awareness and phonemic awareness were found to be predicted by children’s general proficiency in both L1 and L2. Similarly, Atwill, Blanchard, Gorin, and Burstein (2007) demonstrated the influence of bilingual language proficiency on the development of phonemic awareness in Spanish-speaking kindergarten children learning English: they found positive correlations between proficiency and awareness within and across languages. Similar results were reported by Anthony et al. (2009) when they examined the extent to which English and Spanish

218 Ludo Verhoeven

vocabulary and letter knowledge predicted the growth of English and Spanish phonological awareness among preschool Spanish-speaking learners of English. They found that early phonological awareness and vocabulary in Spanish affected the development of phonological awareness in each language. In a meta-analysis of 38 studies reporting cross-language correlations with English phonological measures among children from preschool through the upper grades of elementary school (Branum-Martin, Tao, Garnaat, Bunta, & Francis, 2012), significant cross-language correlations were again found. However, variation in the associations between skills are heavily influenced by the language used, and to some extent by the linguistic units employed in the tasks. This is evident from a study of cross-language associations in linguistically distant languages by Bialystok, McBrideChang, and Luk (2005). They compared 5–6 year old monolingual English-speaking, bilingual English-Chinese speaking, and monolingual Chinese-speaking children. The bilingual children were all just starting to learn English as a second language when the phonological awareness and word decoding of the children in either English, Chinese, or both languages were examined. Phonological awareness developed in response to language exposure and instruction but, once established, transferred across languages for the bilingual learners. In contrast, decoding ability developed separately for each language as a function of general language proficiency and instruction in that language and did not transfer to the other language. In sum, the above studies clearly indicate that phonological awareness in L1 and L2 are highly related. An intriguing question concerns the extent to which specific L1 and L2 components influence the development of phonological awareness in bilingual children. The lexical restructuring hypothesis posits, for example, that the development of vocabulary can enhance phonological awareness. This hypothesis was explicitly tested in bilingual children by Dixon, Chuang, and Quiroz (2012) who found that growth in L2 English vocabulary in interaction with the mother’s level of education predicted significantly the development of Singaporean children’s phonological awareness. In addition, specific L1 proficiency (i.e., Chinese, Malay, or Tamil) was found to be a statistically significant predictor of phonological awareness, suggesting that specific phonological characteristics of the bilingual’s other language influence the development of L2 phonological awareness. In other research, Janssen, McQueen, Segers and Verhoeven (2015) specifically asked whether lexical training could stimulate bilingual children’s phonological awareness. Dutch monolingual and Turkish-Dutch bilingual children were taught new Dutch words with only minimal acoustic-phonetic differences. Both the monolingual and bilingual children showed improved phoneme blending, which is an early aspect of phonological awareness. During the training, the word-blending capacity of the bilingual children caught up with that of the monolingual children for words showing a phonological overlap between Turkish (L1) and Dutch (L2). Janssen et al. concluded that learning minimal word pairs can foster phonemic awareness in both first- and second-language preliterate children and that the phonological overlap between two languages can positively affect training outcomes. Finally, Chen, Ku, Koyama, Anderson,



Learning to read in a second language 219

and Li (2008) examined the phonological awareness of first, second, and fourth grade Cantonese-speaking children receiving immersion Mandarin instruction starting in first grade. The children performed better on the Cantonese onset awareness task in grade one, but this difference disappeared in later grades when performance on the rime and tone awareness tasks was better in Mandarin than in Cantonese. The associations between phonological awareness and word decoding have also been examined throughout the elementary grades. Cheung et al. (2010) examined the relations between speech perception, phonological and morphological awareness, word reading, and vocabulary in three age groups of Chinese (L1) -English (L2) bilingual children. Speech perception was a better predictor of word reading and vocabulary in L1 than in L2. In contrast, morphological awareness uniquely predicted word reading and vocabulary in both languages, whilst phonological awareness only did this in L2 after morphological awareness was controlled. Furthermore, L1 speech perception, phonological and morphological awareness predicted L2 word reading, but not vocabulary, after controlling the corresponding L2 variables. When Farran, Bingham, and Matthews (2012) examined the associations between language and reading in bilingual Arabic-English children, they found that phonological awareness predicted word and pseudoword reading accuracy while vocabulary predicted reading comprehension for both Arabic and English. Finally, Haigh, Savage, Erdos, and Genesee (2011) investigated the link between phoneme and onset-rime awareness and word reading outcomes for English-dominant children in a French immersion education program. They found English phoneme manipulation to be a significant predictor of both English and French reading outcomes after controlling kindergarten knowledge of letter names and word identification in the two languages. French onset-rime knowledge measured in kindergarten also accounted for a significant amount of the variance in the French reading outcome measures. To conclude, research has shown that learning two languages can enhance children’s phonological awareness. When significant linguistic overlap occurs between the phonologies of the two languages, cross-language influences on the development of phonological awareness can also occur. Further, children’s early phonological and morphological awareness can predict their later word decoding even for bilingual children.

Second language word decoding Although L2 reading processes may be impeded for various reasons (cf. Verhoeven, 2010), reviews by Verhoeven (2000) and Siegel (2003) indicate that the word decoding skills of L1 and L2 learners develop to similar levels across the school years. This can be attributed to many L2 learners mastering the essentials of the target language within the domains of phonology and orthography while learning to read. The positive results for the word decoding of L2 learners over time may also be attributed to cross-language transfer. It has indeed been found that overlap with L1 at the level of

220 Ludo Verhoeven

phonology (Ellis & Schmidt, 1997; Lindsey, Manis, & Bailey, 2003; Lopez & Greenfield, 2004) and orthography (Deacon, Wade-Woolley, & Kirby, 2007; Geva & Siegel, 2000) can help L2 learners build their L2 word decoding skills. In a large-scale study of reading development, Lesaux, Rupp, and Siegel (2007) followed native English-speaking (L1) children and English-language learners (ELLs) representing 33 different native languages from kindergarten to fourth grade. Despite initially slightly lower performance by ELLs on several kindergarten tasks, the differences in fourth grade were negligible. Fourth-grade word reading was predicted by the same kindergarten tasks for all language groups, and prediction of reading comprehension differed by only one task (i.e., kindergarten spelling). Finally, the trajectory of word reading development was nonlinear for all the groups with similar predictors accounting for it. In a similar vein, Mancilla-Martinez and Lesaux (2011) modeled the growth rates for the English and Spanish oral language and word reading skills of Spanish-speaking children from low-income households. The students’ growth rates and overall word reading ability from age 4.5 to 11 were on par with national norms. In contrast, the students’ oral language skills started out below national norms and their rates of growth – although surpassing national rates – were not sufficient to reach age-appropriate levels. It can thus be concluded that in the case of closely related languages, bilingual learners have no difficulty in developing word decoding skills in their second language. Turning to more distant languages, Chiappe, Glaeser, and Ferko (2007) examined the roles of speech perception and phonological processing in the word reading and spelling acquisition of native English and nonnative (Korean) speakers of English in first grade. The Korean-speaking children outperformed native English speakers on each of the literacy measures at the start and end of first grade. Speech-related skills (speech perception, phonological processing) appeared to be important contributors to early literacy – independent of oral language skills – for both groups. In other research, Yeong, Fletcher, and Bayliss (2014) examined the importance of English phonological and orthographic processing skills for English word reading and spelling in younger (8–9 years) versus older (11–12 years) children with different language backgrounds (English monolingual, English L1 – Chinese L2, or Chinese L1-English L2). Both the English monolingual and English L1-Chinese L2 children showed better phonological processing skills than the Chinese L1-English L2 children while the younger bilingual children had better orthographic processing skills than the younger monolingual children. Regression analyses showed orthographic processing to be the only significant predictor of word reading and spelling for the monolingual children. In contrast, phonological processing skills were important for the word reading of the bilingual children and for the spelling of the younger bilingual children. Finally, when Yeung and Chan (2013) investigated the linguistic predictors of English reading among fiveyear-old children with Chinese as L1, they found both oral language proficiency and phonological awareness to significantly predict English word reading. L1 phonological awareness at the level of the tone was found to be the only cross-language predictor of L2 word reading.



Learning to read in a second language 221

Morphological awareness has also been found to be a critical predictor of L2 word reading. In a comparison of English monolinguals and English-Arabic bilinguals, Saiegh-Haddad and Geva (2008) found a significant correlation between phonological awareness in English and Arabic but no such relation for morphological awareness. Phonological awareness predicted reading cross-linguistically, and Arabic morphological awareness predicted English word reading, but not vice versa. While both phonological and morphological awareness in English uniquely predicted English word reading, only phonological awareness in Arabic predicted Arabic word reading. In both languages, morphological awareness predicted word reading fluency. Similarly, in both languages, phonological awareness was the single factor predicting pseudoword decoding accuracy. In other research, Wang, Ko, and Choi (2009) examined the importance of morphological awareness for reading development in L1 Korean and L2 English. Morphological awareness explained a significant amount of the variance for word reading and reading comprehension in both English and Korean Hangul, which suggests that morphological awareness is important not only in an opaque orthography but also in a transparent orthography. Furthermore, morphological awareness in one language uniquely predicted significant variance in the reading of real words in the other language, suggesting that morphological awareness can facilitate word reading across different orthographies. A final question concerns the extent to which reading in L1 and L2 show similar levels of phonological processing. Jared, Cormier, Levy, and Wade-Woolley (2012) investigated if children learning to read simultaneously in English and French activated phonological representations for only the language being read or for both languages. The data suggest that phonological activation in bilinguals is not language selective. Similarly, Leafstedt and Gerber (2005) examined phonological processes, cross-linguistic transfer, language of instruction, and word decoding in SpanishEnglish bilingual children learning to read in varying instructional environments. They found phonological processes to exhibit cross-linguistic transfer and concluded that bilinguals’ phonological awareness might best be conceptualized as consisting of two overlapping components. Language of instruction was also found to influence both English and Spanish word reading and Spanish pseudoword decoding but not English pseudoword decoding. Leikin, Schwartz, and Share (2010) examined cross-linguistic transfer of phonemic awareness and word identification skills across two linguistically distant languages (Russian and Hebrew) and found general benefits for phonemic awareness from Russian to Hebrew and specific benefits of bilingualism for the spelling of vowels and consonant clusters in Hebrew. Finally, Wang, Park, and Lee (2006) investigated cross-language phonological and orthographic relationships for children learning to read in Korean and English; they found phonological skills in L1 and L2 to be strongly correlated, but also that Korean phonological skills explained unique variance in English pseudoword reading beyond that explained by the children’s English phonological and orthographic skills. However, only limited orthographic skill transfer between the two systems manifested itself. In a follow-up study, Wang, Yang, and Cheng (2009) investigated the concurrent contributions of

222 Ludo Verhoeven

phonological, orthographic, and morphological awareness to the reading development of grade one Chinese-English bilingual children. Chinese tone and onset awareness explained a significant amount of variance in English real-word reading after controlling English-related variables. Chinese onset awareness also explained a significant amount of variance in English pseudoword reading after controlling English-related variables. Finally, the bilingual children’s morphological awareness of English explained a unique amount of variance in Chinese character reading as well. Taken together, these findings suggest that over time bilingual children attain similar levels of L2 word decoding as their monolingual peers. They also suggest that similar precursors (phonological awareness, morphological awareness) play a role in L1 and L2 word decoding processes. Finally, the findings summarized above further suggest that L2 reading involves not only phonological and morphological processes, which can be partly shared with L1, but also orthographic processes that may be more, but not entirely, language specific.

Second language reading comprehension It is known that limited L2 lexical quality (the degree to which words in the L2 are phonologically, semantically and orthographically specified) may place the reading comprehension of L2 learners at risk (see Perfetti & Stafura, 2014; Genesee, LindholmLeary, Saunders, & Christian, 2006). Estimates of oral vocabulary knowledge have revealed major differences between L1 and L2 learners with the smaller L2 vocabularies of second-language learners seriously impeding their L2 reading comprehension (Verhoeven, 2000). Bialystok, Luk, Peets, and Yang (2010) found consistent differences in the receptive vocabularies of bilingual versus monolingual children between 3 and 10 years of age; follow-up analyses showed these differences to not differ across language pairs and to be largely confined to words relevant to the home as opposed to school context. In addition to less extensive vocabularies, L2 learners have also been shown to have fewer associative links between L2 words than L1 learners of the same language (Vermeer, 2001). A similar conclusion was arrived at by Cremer and Schoonen (2013) when they compared the influences of word decoding, word availability, and accessibility of semantic word knowledge on the reading comprehension of monolingual versus bilingual children. Despite equal word decoding abilities, the monolingual children outperformed the bilingual children on reading comprehension and availability of semantic word knowledge. When Schwartz and Katzir (2012) compared the depth and breadth of lexical knowledge of bilingual second-generation immigrant children and monolingual children (7–8 years), they found a significant gap for most measures at the beginning of second grade. Proctor, Silverman, Harring, and Montecillo (2012) also examined the role of vocabulary depth – defined to include morphological awareness, awareness of semantic relations, and syntactic awareness – in the reading comprehension of monolingual



Learning to read in a second language 223

English and bilingual Spanish-English children in grades two through four. They found significant contributions only for initial reading comprehension. Uchikoshi (2013) examined the role of lexical quality factors in the English reading comprehension of second grade Cantonese-speaking English language learners. The results showed that both English vocabulary and English word-decoding skill – measured using real and nonsense words – played significant roles. These results highlight the crucial role of English vocabulary development for the development of L2 English reading comprehension skills. Proctor, August, Carlo, and Snow (2006) explored the opposite; namely the role of L1 Spanish vocabulary knowledge in the prediction of L2 English reading comprehension among Spanish-English bilingual fourth graders. They found a significant main effect of Spanish vocabulary knowledge along with an interaction between Spanish vocabulary and initial English reading fluency, with faster English readers benefitting more from their Spanish vocabulary knowledge than slower English readers. With respect to reading comprehension, the so-called simple view of reading (SVR) (Hoover & Gough, 1990) claims that – in addition to basic word decoding skills – oral language comprehension can strongly predict reading comprehension. From this perspective, Verhoeven and van Leeuwe (2012) explored the reading comprehension of L2 learners compared to L1 learners and found similar levels of word decoding but delays for the L2 learners on both listening and reading comprehension. Nevertheless, the associations between word decoding, listening comprehension, and reading comprehension turned out to be highly comparable for the two groups and longitudinal analyses of the same data showed the simple view of reading to hold for both L1 and L2 learners. When looking across development by grade, the impact of word decoding on reading comprehension decreased while the impact of listening comprehension on reading comprehension increased to the same extent for both groups of learners. However, the reciprocity of the cross-lagged relation between listening comprehension and reading comprehension was less prominent for the L2 learners compared to the L1 learners. In a similar vein, Nakamoto, Lindsey, and Manis (2008) investigated the associations between language comprehension and reading skills of Spanish-speaking learners of English and found both decoding and language comprehension to significantly predict reading comprehension in both languages. The within-language effects were larger than the cross-language effects for both languages, however, and the cross-language effects disappeared when the within-language effects were controlled for. Fender (2001) reviewed the research on L1 and L2 word-to-text integration and the development of L2 word-to-text integration in particular. L1 and fluent L2 speakers appear to utilize similar processing procedures for the integration of words into larger phrase and sentence structures. The results also showed the development of fluent and accurate L2 word integration to largely depend on the development of sufficient L2 syntactic structure-building skill. To arrive at more general findings on L2 reading comprehension and its components, Jeon and Yamashita (2014) conducted a meta-analysis of studies. Clear evidence was found for four core components: L2

224 Ludo Verhoeven

decoding, L2 vocabulary knowledge, L2 grammar knowledge, and L1 reading comprehension. Less clear evidence was found for six other components: L2 phonological awareness, L2 orthographic knowledge, L2 morphological knowledge, L2 listening comprehension, working memory, and metacognition. After taking into account the role of age, L2 proficiency, L1-L2 script and language distance, and measurement characteristics, the results showed L2 decoding, L2 vocabulary knowledge, and L2 grammar knowledge to remain the three strongest correlates of L2 reading comprehension. In several other studies, the prediction of L2 reading comprehension by children’s L1 and L2 oral language capacities has been examined. Swanson, Rosston, Gerber, and Solari (2008) analyzed the roles of oral language comprehension and phonological awareness on the reading performance of grade three bilingual students and found expressive vocabulary and syntax knowledge to best predict L2 reading comprehension. When Babayigit (2014) examined the role of oral language skills in the listening and reading comprehension of L1 versus L2 learners of English, she found that the L1 learners outperformed L2 learners on measures of L2 oral language and text comprehension but that the two groups performed comparably on measures of L2 word reading speed and accuracy. Oral language emerged as the most powerful predictor of both listening and reading comprehension. Although there was a tendency for oral language skill to more strongly relate to L2 reading comprehension, its relationship with listening comprehension was comparable across the two groups. Individual oral language differences emerged, moreover, as the primary explanatory factor for the group differences in text comprehension. Another study on this topic was conducted by Baker, Park, and Baker (2012) who examined pseudoword reading, oral reading fluency, and reading comprehension among Spanish-speaking English learners in grades one through three. Results of hierarchical linear modeling indicated different patterns of reading growth in Spanish and English across measures and grades. The bilinguals at the beginning of first grade had higher scores on pseudoword reading in Spanish than in English and a higher rate of growth on Spanish pseudoword reading. In second and third grades, initial scores on oral reading fluency were comparable in both languages, but oral reading fluency growth rates were higher in English than in Spanish. Regression and path analysis results further showed the students’ initial scores and the growth of reading fluency to be strong, direct predictors of reading comprehension within the same language but not across languages. Finally, in an examination of the role of SES in the prediction of L2 reading comprehension, Kieffer (2011) compared the growth trajectories for the English reading of L1 learners and L2 learners who had differing levels of initial English proficiency before and after controlling for SES. L2 learners who entered kindergarten fluent in English caught up with native English speakers by first grade and maintained nationally average levels of performance through eighth grade. L2 learners who initially had limited English proficiency showed English reading trajectories substantially below national averages but converged with L1 peers from similar socioeconomic backgrounds during middle school. In a follow-up study, Kieffer and Vukovic (2013)



Learning to read in a second language 225

investigated growth in reading-related skills between grades one and four for L2 learners and their native English-speaking classmates from similarly low socioeconomic backgrounds. Growth trajectories were compared for language background and grade four reading difficulties, with the goal of informing decisions about how early L2 learners can undergo screening for risk of reading difficulties. The subset of L2 learners with later word reading difficulties demonstrated major weaknesses in earlier vocabulary, oral comprehension, phonological awareness, and working memory. In contrast, the subset of L2 learners with later specific reading comprehension difficulties demonstrated major weaknesses in earlier vocabulary and oral comprehension, but not phonological awareness or working memory. These weaknesses were apparent in grade one and persisted through grade four. To conclude, research on L2 reading comprehension shows lexical quality to be an extremely important predictor of this skill. Not only the sheer number of words represented in the children’s L2 vocabularies but also the semantic ties between the words can be seen as highly relevant components of L2 reading comprehension. Research further shows the simple view of reading to equally hold for L1 and L2 reading. In both cases, variation in reading comprehension can be seen to be largely the product of children’s word decoding and listening comprehension – including their vocabulary, morphological awareness, and syntactic abilities. The next question concerns the specific contribution of L1 proficiency to L2 reading outcomes, if any.

Role of first language proficiency in second language reading Literacy skills acquired in one language can, in principle, easily transfer to another language. The development of L1 reading abilities can thus be expected to predict the development of corresponding L2 reading abilities and produce significant interdependencies (cf. Verhoeven, 1994). In addition, language transfer can be expected to occur from not only L1 to L2 but also from L2 to L1. This interdependence hypothesis further predicts that optimal input in one language will lead not only to better skills in that language but also to deeper conceptual and linguistic proficiency, which can, in turn, facilitate the transfer of cognitive and academic language skills across languages (Cummins, 1980, 1984, 1991). Cross-language reading research can provide insight into the conditions under which L2 reading processes can be facilitated. When Melby-Lervåg and Lervåg (2011) conducted a meta-analysis of the evidence for the cross-linguistic transfer of oral language skill, word decoding, phonological awareness, and reading comprehension, they found a small correlation between L1 and L2 oral language skill and a moderate to large correlation between L1 and L2 phonological awareness and decoding. This pattern of findings was taken to reflect the complexity of oral language skills compared to phonological awareness and word decoding with a limited number of letter-sound combinations to learn. However, widespread variation in the L1-L2 correlations was found for all language domains. The variation for word decoding was moderated by the writing system and language of instruction. Furthermore,

226 Ludo Verhoeven

the meta-correlation between L1 word decoding and L2 reading comprehension was only small to moderate and decreased with age, whilst a correlation between L1 oral language skill and L2 reading comprehension was largely lacking. A few studies have examined the transfer of L1 phonological abilities to early second language reading. Chung, McBride-Chang, Cheung, and Wong (2013) focused on the associations of general auditory processing, speech perception, phonological awareness, and word reading in Cantonese-speaking children from Hong Kong learning to read in Chinese (L1) and English (L2). L1 auditory processing explained both L1 and L2 word reading while L1 speech perception explained only a unique amount of the variance in L1 word reading and L2 phonological awareness explained only a unique amount of the variance in L2 word reading. In cross-language comparisons, L1 phonological awareness and L1 speech perception were uniquely associated with L2 word reading, suggesting cross-language transfer from L1 to L2 only. Wang, Perfetti, and Liu (2005) investigated cross-language and writing system relationships in children learning to read the two different writing systems of Chinese and English. They found Chinese onset matching skill to significantly correlate with English onset and rime matching skill. Knowledge of Pinyin, an alphabetic phonetic system used to assist children in learning to read Chinese characters, correlated highly with English pseudoword reading. Furthermore, Chinese tone processing skill explained a moderate, but significant, amount of the variance in English pseudoword reading even when English phonemic-level processing skill was controlled. Orthographic processing skill in the one writing system did not predict word reading in the other writing system, however, which suggests that in this case bilingual reading acquisition is a joint function of shared phonological processes and specific orthographic skills. Other studies have investigated more directly the evidence for transfer within the domain of orthographic learning. Commissaire, Duncan, and Casalis (2011) examined the development of word-specific knowledge (e.g., rain-rane) and sensitivity to sublexical regularities (e.g., schoal-sckoal) in L1 French reading of children in grades six and eight and charted the cross-language transfer of orthographic skills during the first three years of L2 English learning. Word-specific orthographic knowledge in L2 correlated with L1 reading speed, and thus revealed direct cross-language associations. However, no evidence was found for the transfer of sensitivity to sublexical regularities. Deacon, Chen, Luo and Ramirez (2013), in turn, examined the transfer of L1 orthographic knowledge to L1 and L2 reading in Spanish-English bilingual children in grades four and seven, and found significant associations for both English and Spanish reading, which is a novel finding for the relatively transparent script of Spanish. Further, orthographic knowledge in Spanish was related to English reading. In another study, Deacon, Commissaire, Chen, and Pasquarella (2013) examined the nature of orthographic processing in English-French bilingual children in the first grade of a French immersion program and, in particular, the relationship between their orthographic processing and reading. They found greater orthographic knowledge for patterns common to the two languages than for patterns unique to one of the



Learning to read in a second language 227

two languages. Across both lexical and sublexical orthographic knowledge measures, however, significant relations to word reading were found only within each language. In a follow-up study by Pasquarella, Deacon, Chen, Commissaire, and Au-Yeung (2014), the within- and cross-language associations between orthographic knowledge and word reading in French and English were examined across grades one and two. Orthographic processing in grade one did not predict word reading in grade two in either language, but word reading in grade one predicted orthographic processing in grade two within each language; cross-linguistically, French word reading significantly predicted growth in English orthographic processing. Still other studies have focused on the transfer from L1 to L2 within the domain of morphological awareness and L2 reading. Ramirez, Chen, Geva, and Kiefer (2010) examined the within- and cross-language relations for grades four and seven Spanishspeaking children learning English, and found cross-linguistic transfer from Spanish to English but not from English to Spanish. These results not only suggest that morphological awareness is particularly important for word reading in Spanish with its shallow orthography but complex morphological system, but also that the morphological awareness developed in Spanish, the children’s L1, can nevertheless facilitate word reading in English, the children’s L2. When Pasquarella, Chen, Lam, Luo, and Ramirez (2011) examined the cross-language transfer of morphological awareness in Chinese-English bilingual children in grades one through four, they found the proposed cross-language model to fit significantly better than the within-language model, suggesting transfer of morphological awareness from English to Chinese. They observed a bidirectional relationship between English compound awareness and Chinese vocabulary but also English compound awareness to significantly predict Chinese reading comprehension but not vice versa. In a similar vein, Ramirez, Chen and Pasquarella (2013) examined the cross-language effects of Spanish derivational awareness on English vocabulary and reading comprehension in Spanish-English bilingual children in grades four and seven. They found significant associations of Spanish derivational awareness with only English cognates and not English noncognates. In addition, there was an indirect contribution of Spanish derivational awareness to English reading comprehension via English cognate vocabulary and English morphological awareness. Finally, more general explorations of linguistic transfer have also been conducted. Montanari (2014), for example, examined how Italian and English literacy skills emerged among 60 children enrolled in an Italian-English dual language program. She found reading fluency to emerge first in Italian for all students but become comparable across languages in a relatively short period of time and even higher in English for the English-speaking students in particular. The correlations between the children’s Italian and English reading fluency scores were moderately high, suggesting that Italian decoding skills transferred to English to help the children in this study develop their English literacy skills while primarily instructed in Italian. Sparks, Patton, Ganschow, Humbach, and Javorsky (2008), in turn, examined early L1 predictors of later L2 reading (i.e., word decoding and comprehension) and spelling skills. They

228 Ludo Verhoeven

found the best predictor of L2 decoding skill to be L1 decoding and the best predictors of L2 spelling to be L1 spelling and L1 phonological awareness. The best predictor of L2 reading comprehension was a measure of L1 reading comprehension. These findings suggest that even several years after learning to read and spell in their L1, the L1 word decoding, spelling, and reading comprehension skills can still transfer from L1 to L2. To conclude, clear and significant interdependencies manifest themselves in the development of L1 and L2 reading abilities. Most striking are those for cognates and within the domains of phonology and morphology when there is sufficient overlap between the L1 and L2. The exact conditions required for transfer of reading abilities from one language to the other are, however, far from fully understood.

Future perspectives The studies reviewed in this chapter make it clear that the development of second language reading can be called a challenging task. In order to prevent second language readers from falling behind first language readers, it is essential that suitable developmental input and ample learning opportunites are available. The present review of the research on L2 reading makes it clear that bilingual children can profit from bilingual instruction and that exposure to L1 and L2 can lead to higher levels of metalinguistic awareness. When a higher level of phonological and morphological awareness is attained, this can, in turn, boost the process of learning to read in a second language. Further, research has shown convincingly that second language learners are capable when it comes to word decoding and that learning to read in a second language does not negatively influence the development of children’s word decoding. But word decoding is only one predictor of L2 reading comprehension. A robust research finding is that L2 learners significantly lag behind L1 learners in vocabulary and that the differences in both the size and depth of the vocabularies of L1 versus L2 learners remain noticeable throughout the elementary grades. A similar conclusion holds for L1 versus L2 reading comprehension with L2 comprehenion systematically lagging behind L1 comprehension throughout elementary school. It also seems clear that these reading comprehension delays can be attributed, not only to ongoing differences in vocabulary, but also to differences in morphological awareness and syntactic abilities. Linguistic transfer can nevertheless occur from L1 to L2 for both lower-level phonological and orthographic skills, and higher-level morphosyntactic skills. The combination of languages being mastered can, in addition, influence the process of learning to read in a second language. Although the state of the art in L2 reading research can be called promising, several limitations are apparent as well. First of all, much of the evidence comes from smallscale studies in specific contexts and may not generalize to other contexts. Although meta-analyses are available, replications across a variety of learning contexts are urgently needed. Furthermore, many of the studies have been either cross-sectional or involved only a brief longitudinal data collection. There is thus a pressing need for



Learning to read in a second language 229

longitudinal studies that systematically compare the development of phonological awareness, orthographic skills, reading vocabulary, and reading comprehension in both L1 and L2 throughout the elementary grades. It should also be acknowledged that much of the current research evidence comes from correlational studies. Longerterm longitudinal studies should be complemented with experimental training studies in which specific hypotheses about the nature of L2 reading processes are tested. Comparative studies are needed to provide more insight into the features of the languages and orthographies being mastered that affect transfer from one language to the other. Issues such as variation in writing systems and orthographic depth need systematic study. Finally, most of the present evidence of reading development comes from behavioral studies while adult L2 reading research has made considerable progress with regard to understanding the underlying brain mechanisms. For the study of L2 reading in children, however, research on brain mechanisms is only in a nascent state. In order to better understand the language universals and language specifics of L1 and L2 reading development in children, it is important that behavioral studies are complemented with studies of underlying neural mechanisms. The present review has important implications for educational practice. To begin with, there is evidence that achieving native results in L2 reading can be seen as an enduring task that can take several years, depending on L2 aptitude and first language skills. L2 reading success also builds upon environmental opportunities. Optimal conditions for L2 learners immersed in a majority L2 society include strong home literacy practices, opportunities to use the L2 informally, specially designed and carefully implemented L2 educational programs employing highly skilled teachers, and sufficient time for L2 literacy instruction. To further understand the worldwide variation in L2 reading success, it is important to examine the roles of specific child and environmental variables in L2 reading development across varying multilingual and sociocultural contexts.

References Anthony, J. L., Solari, E. J., Williams, J. M., Schoger, K. D., Zhang, Z., Branum-Martin, L., & Francis, D. J. (2009). Development of bilingual phonological awareness in Spanish-Speaking English language learners: The roles of vocabulary, letter knowledge, and prior phonological awareness. Scientific Studies of Reading, 13, 535–564.  doi: 10.1080/10888430903034770 Atwill, K., Blanchard, J., Gorin, J. S., & Burstein, K. (2007). Receptive vocabulary and cross-language transfer of phonemic awareness in kindergarten children. The Journal of Educational Research, 100, 336–345.  doi: 10.3200/JOER.100.6.336-346 Babayigit, S. (2014). The role of oral language skills in reading and listening comprehension of text: a comparison of monolingual (L1) and bilingual (L2) speakers of English language. Journal of Research in Reading, 37, S22–S47.  doi: 10.1111/j.1467-9817.2012.01538.x Baker, D. L., Park, Y., & Baker, S. K. (2012). The reading performance of English learners in grades 1–3: the role of initial status and growth on reading fluency in Spanish and English. Reading and Writing, 25, 251–281.  doi: 10.1007/s11145-010-9261-z

230 Ludo Verhoeven

Bialystok, E., Luk, G., Peets, K., & Yang, S. (2010). Receptive vocabulary differences in monolingual and bilingual children. Bilingualism-Language and Cognition, 13(4), 525–531. Bialystok, E., McBride-Chang, C., & Luk, G. (2005). Bilingualism, language proficiency, and learning to read in two writing systems. Journal of Educational Psychology, 97, 580–590. doi: 10.1037/0022-0663.97.4.580 Bosch, L., & Ramon-Casas, M. (2014). First translation equivalents in bilingual toddlers’ expressive vocabulary: Does form similarity matter? International Journal of Behavioral Development, 38, 317–322.  doi: 10.1177/0165025414532559 Branum-Martin, L., Tao, S., Garnaat, S., Bunta, F., & Francis, D. J. (2012). Meta-Analysis of Bilingual Phonological Awareness: Language, Age, and Psycholinguistic Grain Size. Journal of Educational Psychology, 104, 932–944.  doi: 10.1037/a0027755 Buysse, V., Peisner-Feinberg, E., Paez, M., Hammer, C. S., & Knowles, M. (2014). Effects of early education programs and practices on the development and learning of dual language learners: A review of the literature. Early Childhood Research Quarterly, 29, 765–785. doi: 10.1016/j.ecresq.2013.08.004 Cattani, A., Abbot-Smith, K., Farag, R., Krott, A., Arreckx, F., Dennis, I., & Floccia, C. (2014). How much exposure to english is necessary for a bilingual toddler to perform like a monolingual peer in language tests? International Journal of Language & Communication Disorders, 49, 649–671.  doi: 10.1111/1460-6984.12082 Chen, X., Ku, Y.-M., Koyama, E., Anderson, R. C., & Li, W. (2008). Development of phonological awareness in bilingual Chinese children. Journal of Psycholinguistic Research, 37, 405–418.  doi: 10.1007/s10936-008-9085-z Cheung, H., Chung, K. K. H., Wong, S. W. L., McBride-Chang, C., Penney, T. B., & Ho, C. S.-H. (2010). Speech perception, metalinguistic awareness, reading, and vocabulary in ChineseEnglish bilingual children. Journal of Educational Psychology, 102, 367–380. doi: 10.1037/a0017850 Chiappe, P., Glaeser, B., & Ferko, D. (2007). Speech perception, vocabulary, and the development of reading skills in English among Korean- and English-speaking children. Journal of Educational Psychology, 99, 154–166.  doi: 10.1037/0022-0663.99.1.154 Chung, K. K. H., McBride-Chang, C., Cheung, H., & Wong, S. W. (2013). General auditory processing, speech perception and phonological awareness skills in Chinese-English biliteracy. Journal of Research in Reading, 36, 202–222.  doi: 10.1111/j.1467-9817.2011.01500.x Commissaire, E., Duncan, L. G., & Casalis, S. (2011). Cross-language transfer of orthographic processing skills: a study of French children who learn English at school. Journal of Research in Reading, 34, 59–76.  doi: 10.1111/j.1467-9817.2010.01473.x Cremer, M., & Schoonen, R. (2013). The role of accessibility of semantic word knowledge in monolingual and bilingual fifth-grade reading. Applied Psycholinguistics, 34, 1195–1217. doi: 10.1017/S0142716412000203 Cummins, J. (1980). Cross-lingual dimensions of language proficiency: implications for bilingual education and the age issue. TESOL Quarterly, 14, 175–187. Cummins, J. (1984). Wanted: A theoretical framework for relating language proficiency to academic achievement among bilingual students. In C. Rivera (Ed.), Language proficiency and academic achievement (pp. 2–19). Clevedon: Multilingual Matters. Cummins, J. (1991). Conversational and academic language proficiency in bilingual contexts. AILA Review, 8, 75–89.



Learning to read in a second language 231

Deacon, S. H., Chen, X., Luo, Y., & Ramirez, G. (2013). Beyond language borders: Orthographic processing and word reading in Spanish-English bilinguals. Journal of Research in Reading, 36, 58–74.  doi: 10.1111/j.1467-9817.2011.01490.x Deacon, S. H., Commissaire, E., Chen, X., & Pasquarella, A. (2013). Learning about print: The development of orthographic processing and its relationship to word reading in first grade children in French immersion. Reading and Writing, 26, 1087–1109. doi: 10.1007/s11145-012-9407-2 Deacon, S., Wade-Woolley, L., & Kirby, J. (2007). Crossover: The role of morphological awareness in French immersion children’s reading. Developmental Psychology, 43, 732–746. doi: 10.1037/0012-1649.43.3.732 Dixon, L. Q., Chuang, H.-K., & Quiroz, B. (2012). English phonological awareness in bilinguals: A cross-linguistic study of Tamil, Malay and Chinese English learners. Journal of Research in Reading, 35, 372–392.  doi: 10.1111/j.1467-9817.2010.01471.x Ellis, N. C., Schmidt, R. (1997). Morphology and longer distance dependencies: Laboratory research illuminating the A in SLA. Studies in Second Language Acquisition, 19, 145–171. doi: 10.1017/S0272263197002027 Farran, L. K., Bingham, G. E., & Matthews, M. W. (2012). The relationship between language and reading in bilingual English-Arabic children. Reading and Writing, 25, 2153–2181. doi: 10.1007/s11145-011-9352-5 Fender, M. (2001). A review of L1 and L2/ESL word integration skills and the nature of L2/ESL word integration development involved in lower-level text processing. Language Learning, 51, 319–396.  doi: 10.1111/0023-8333.00157 Ford, K. L., Cabell, S. Q., Konold, T. R., Invernizzi, M., & Gartland, L. B. (2013). Diversity among Spanish-speaking English language learners: Profiles of early literacy skills in kindergarten. Reading and Writing, 26, 889–912.  doi: 10.1007/s11145-012-9397-0 Genesee, F., Lindholm-Leary, K., Saunders, W., & Christian, D. (2006). Educating English language learners: A synthesis of research evidence. New York, NY: Cambridge University Press. doi: 10.1017/CBO9780511499913 Geva, E., & Siegel, L. S. (2000). Orthographic and cognitive factors in the concurrent development of basic reading skills in two languages. Reading and Writing: An Interdisciplinary Journal, 12, 1–30.  doi: 10.1023/A:1008017710115 Gorman, B. K. (2012). Relationships between vocabulary size, working memory, and phonological awareness in Spanish-Speaking English language learners. American Journal of SpeechLanguage Pathology, 21, 109–123.  doi: 10.1044/1058-0360(2011/10-0063) Haigh, C. A., Savage, R., Erdos, C., & Genesee, F. (2011). The role of phoneme and onset-rime awareness in second language reading acquisition. Journal of Research in Reading, 34, 94–113. doi: 10.1111/j.1467-9817.2010.01475.x Hoover, W. and Gough, P. (1990). The simple view of reading. Reading and Writing: An Interdisciplinary Journal, 2, 127–160.  doi: 10.1007/BF00401799 Hurtado, N., Gruter, T., Marchman, V. A., & Fernald, A. (2014). Relative language exposure, processing efficiency and vocabulary in Spanish-English bilingual toddlers. Bilingualism: Language and Cognition, 17, 189–202.  doi: 10.1017/S136672891300014X Indefrey, P. (2006). A meta-analysis of hemodynamic studies on first and second language processing: Which suggested differences can we trust and what do they mean? Language Learning, 56, 279–304.  doi: 10.1111/j.1467-9922.2006.00365.x

232 Ludo Verhoeven

Janssen, M., Bosman, A. M., & Leseman, P. P. (2013). Phoneme awareness, vocabulary and word decoding in monolingual and bilingual Dutch children. Journal of Research in Reading, 36, 1–13.  doi: 10.1111/j.1467-9817.2011.01480.x Janssen, C., Segers, E., McQueen, J. M., & Verhoeven, L. (2015). Lexical specificity training effects in second language learners. Language Learning, 65, 358–389.  doi: 10.1111/lang.12102 Jared, D., Cormier, P., Levy, B. A., & Wade-Woolley, L. (2012). Cross-language activation of phonology in young bilingual readers. Reading and Writing: An Interdisciplinary Journal, 25, 1327–1343.  doi: 10.1007/s11145-011-9320-0 Jeon, E. H., & Yamashita, J. (2014). L2 reading comprehension and its correlates: A meta-analysis. Language Learning, 64, 160–212.  doi: 10.1111/lang.12034 Kieffer, M. J. (2011). Converging trajectories: Reading growth in language minority learners and their classmates, Kindergarten to Grade 8. American Educational Research Journal, 48, 1187–1225.  doi: 10.3102/0002831211419490 Kieffer, M. J., & Vukovic, R. K. (2013). Growth in reading-related skills of language minority learners and their classmates: More evidence for early identification and intervention. Reading and Writing, 26, 1159–1194.  doi: 10.1007/s11145-012-9410-7 Kuo, L.-J., & Anderson, R. C. (2012). Effects of early bilingualism on learning phonological regularities in a new language. Journal of Experimental Child Psychology, 111, 455– 467.  doi: 10.1016/j.jecp.2011.08.013 Leafstedt, J. M., & Gerber, M. M. (2005). Crossover of phonological processing skills – A study of Spanish-speaking students in two instructional settings. Remedial and Special Education, 26, 226–235.  doi: 10.1177/07419325050260040501 Leikin, M., Schwartz, M., & Share, D. L. (2010). General and specific benefits of bi-literate bilingualism: A Russian-Hebrew study of beginning literacy. Reading and Writing, 23, 269–292.  doi: 10.1007/s11145-009-9210-x Lesaux, N. K., Rupp, A. A., & Siegel, L. S. (2007). Growth in reading skills of children from diverse linguistic backgrounds: Findings from a 5-year longitudinal study. Journal of Educational Psychology, 99, 821–834.  doi: 10.1037/0022-0663.99.4.821 Lindsey, K. A., Manis, F. R., & Bailey, C. E. (2003). Prediction of first-grade reading in Spanishspeaking English-language learners. Journal of Educational Psychology, 95, 482–494. Lopez, L. M., & Greenfield, D. B. (2004). The cross-language transfer of phonological skills of Hispanic Head Start children. Bilingual Research Journal, 28, 1–18. doi: 10.1080/15235882.2004.1016260 Mancilla-Martinez, J., & Lesaux, N. K. (2011). The gap between Spanish speakers’ word reading and word knowledge: A longitudinal study. Child Development, 82, 1544–1560. doi: 10.1111/j.1467-8624.2011.01633.x Marinova-Todd, S. H., Zhao, J., & Bernhardt, M. (2010). Phonological awareness skills in the two languages of Mandarin-English bilingual children. Clinical Linguistics & Phonetics, 24, 387–400.  doi: 10.3109/02699200903532508 Melby-Lervåg, M., & Lervåg, A. (2011). Cross-linguistic transfer of oral language, decoding, phonological awareness and reading comprehension: A meta-analysis of the correlational evidence. Journal of Research in Reading, 34, 114–135.  doi: 10.1111/j.1467-9817.2010.01477.x Montanari, S. (2014). A case study of bi-literacy development among children enrolled in an Italian-English dual language program in Southern California. International Journal of Bilingual Education and Bilingualism, 17, 509–525.  doi: 10.1080/13670050.2013.833892



Learning to read in a second language 233

Naep (2005). National Center for Education Statistics, National Assessment of Educational Progress (NAEP): Reading and Mathematics. Washington, D.C.: http://nationsreportcard. gov/reading_ math_2005 Nakamoto, J., Lindsey, K. A., & Manis, F. R. (2008). A Cross-Linguistic Investigation of English Language Learners’ Reading Comprehension in English and Spanish. Scientific Studies of Reading, 12, 351–371.  doi: 10.1080/10888430802378526 Parra, M., Hoff, E., & Core, C. (2011). Relations among language exposure, phonological memory, and language development in Spanish-English bilingually developing 2-year-olds. Journal of Experimental Child Psychology, 108, 113–125.  doi: 10.1016/j.jecp.2010.07.011 Pasquarella, A., Chen, X., Lam, K., Luo, Y. C., & Ramirez, G. (2011). Cross-language transfer of morphological awareness in Chinese-English bilinguals. Journal of Research in Reading, 34, 23–42.  doi: 10.1111/j.1467-9817.2010.01484.x Pasquarella, A., Deacon, H., Chen, B. X., Commissaire, E., & Au-Yeung, K. (2014). Acquiring orthographic processing through word reading: Evidence from children learning to read French and English. International Journal of Disability, Development and Education, 61, 240–257.  doi: 10.1080/1034912X.2014.932579 Perfetti, C., & Stafura, J. (2014). Word knowledge in a theory of reading comprehension. Scientific Studies of Reading, 18, 22–37. Place, S., & Hoff, E. (2011). Properties of dual language exposure that influence 2-year-olds’ bilingual proficiency. Child Development, 82, 1834–1849.  doi: 10.1111/j.1467-8624.2011.01660.x Poulin-Dubois, D., Bialystok, E., Blaye, A., Polonia, A., & Yott, J. (2013). Lexical access and vocabulary development in very young bilinguals. International Journal of Bilingualsim, 17, 57–70.  doi: 10.1177/1367006911431198 Proctor, C. P., August, D., Carlo, M. S., & Snow, C. (2006). The intriguing role of spanish language vocabulary knowledge in predicting English reading comprehension. Journal of Educational Psychology, 98(1), 159–169.  doi: 10.1037/0022-0663.98.1.159 Proctor, C. P., Silverman, R. D., Harring, J. R., & Montecillo, C. (2012). The role of vocabulary depth in predicting reading comprehension among English monolingual and SpanishEnglish bilingual children in elementary school. Reading and Writing, 25, 1635–1664. doi: 10.1007/s11145-011-9336-5 Ramirez, G., Chen, X., & Pasquarella, A. (2013). Cross-linguistic transfer of morphological awareness in Spanish-speaking English language learners: The facilitating effect of cognate knowledge. Topics in Language Disorders, 33, 73–92.  doi: 10.1097/TLD.0b013e318280f55a Ramirez, G., Chen, X., Geva, E., & Kiefer, H. (2010). Morphological awareness in Spanishspeaking English language learners: Within and cross-language effects on word reading. Reading and Writing, 23, 337–358.  doi: 10.1007/s11145-009-9203-9 Saiegh-Haddad, E., & Geva, E. (2008). Morphological awareness, phonological awareness, and reading in English-Arabic bilingual children. Reading and Writing, 21, 481–504. doi: 10.1007/s11145-007-9074-x Schwartz, M., & Katzir, T. (2012). Depth of lexical knowledge among bilingual children: The impact of schooling. Reading and Writing, 25, 1947–1971.  doi: 10.1007/s11145-011-9308-9 Sebastian, R., Laird, A. R., & Kiran, S. (2011). Meta-analysis of the neural representation of first language and second language. Applied Psycholinguistics, 32, 799–819. doi: 10.1017/S0142716411000075 Siegel, L. (2003). Bilingualism and reading. In L. Verhoeven, C. Elbro, & P. Reitsma (Eds.), Precursors of functional literacy (pp. 287–302). Amsterdam, The Netherlands: John Benjamins.

234 Ludo Verhoeven

Sparks, R. L., Patton, J., Ganschow, L., Humbach, N., & Javorsky, J. (2008). Early first-language reading and spelling skills predict later second-language reading and spelling skills. Journal of Educational Psychology, 100, 162–174.  doi: 10.1037/0022-0663.100.1.162 Swanson, H., Rosston, K., Gerber, M., & Solari, E. (2008). Influence of oral language and phonological awareness on children’s bilingual reading. Journal of School Psychology, 46, 413–429. doi: 10.1016/j.jsp.2007.07.002 Uchikoshi, Y. (2013). Predictors of English reading comprehension: Cantonese-speaking English language learners in the U.S. Reading and Writing, 26, 913–939.  doi: 10.1007/s11145-012-9398-z Verhoeven, L. (1994). Transfer in bilingual development: The linguistic interdependency hypothesis revisited. Language Learning, 44, 381–415.  doi: 10.1111/j.1467-1770.1994.tb01112.x Verhoeven, L. (2000). Components in early second language reading and spelling. Scientific Studies of Reading, 4, 313–330.  doi: 10.1207/S1532799XSSR0404_4 Verhoeven, L. (2007). Early bilingualism, language transfer, and phonological awareness. Applied Psycholinguistics, 28, 425–439.  doi: 10.1017/S0142716407070233 Verhoeven, L. (2010). Second language reading acquisition. In M. L. Kamil, P. D. Pearson, E. B. Moje & P. Afflerbach (Eds.) Handbook of reading research (pp. 661–683). New York: Taylor & Francis. Verhoeven, L. & Leeuwe, J. van (2012). The simple view of second language reading throughout the primary grades. Reading and Writing, 25, 1805–1818.  doi: 10.1007/s11145-011-9346-3 Vermeer, A. (2001). Breadth and depth of vocabulary in relation to L1/L2 acquisition and frequency of input. Applied Psycholinguistics, 22, 217–234.  doi: 10.1017/S0142716401002041 Wang, M., Ko, I. Y., & Choi, J. (2009). The importance of morphological awareness in KoreanEnglish biliteracy acquisition. Contemporary Educational Psychology, 34, 132–142. doi: 10.1016/j.cedpsych.2008.12.002 Wang, M., Park, Y., & Lee, K. R. (2006). Korean-English biliteracy acquisition: Crosslanguage phonological and orthographic transfer. Journal of Educational Psychology, 98, 148–158.  doi: 10.1037/0022-0663.98.1.148 Wang, M., Perfetti, C. A., & Liu, Y. (2005). Chinese-English biliteracy acquisition: Cross-language and writing system transfer. Cognition, 97, 67–88.  doi: 10.1016/j.cognition.2004.10.001 Wang, M., Yang, C., & Cheng, C. (2009). The contributions of phonology, orthography, and morphology in Chinese-English biliteracy acquisition. Applied Psycholinguistics, 30, 291–314.  doi: 10.1017/S0142716409090122 Yeong, S. H. M., Fletcher, J., & Bayliss, D. M. (2014). Importance of phonological and orthographic skills for English reading and spelling: A comparison of English monolingual and MandarinEnglish bilingual children. Journal of Educational Psychology, 106, 1107–1121. doi: 10.1037/a0036927 Yeung, S. S., & Chan, C. K. K. (2013). Phonological awareness and oral language proficiency in learning to read English among Chinese kindergarten children in Hong Kong. British Journal of Educational Psychology, 83, 550–568.  doi: 10.1111/j.2044-8279.2012.02082.x

Part III

Typical development of comprehension and underlying processes

Introduction to reading comprehension Kate Cain, Donald L. Compton and Rauno K. Parrila

Lancaster University / Florida State University / University of Alberta

The three chapters in this section examine reading comprehension development. Together, the three chapters provide a coherent analysis of the key knowledge, skills, and processes involved in comprehending text and how they develop. In the first chapter, Spencer, Quinn, and Wagner focus on knowledge of words, both semantic knowledge and morphological knowledge, and how this influences reading comprehension. They first examine what it means to know a word: the breadth and depth of word understanding, and the importance of both conceptual knowledge and knowledge of morphology. They then consider the predictive relationship between vocabulary knowledge and reading comprehension and how the two are related across development, exploring the different mechanisms that might underpin this relationship. They garner evidence in support of the instrumentalist hypothesis: that vocabulary is causally related to reading comprehension. However, one challenge to our understanding of the relation between vocabulary and reading comprehension is that whilst longitudinal studies demonstrate a predictive relationship from vocabulary to reading comprehension, intervention studies to foster vocabulary development often have only modest impacts on reading comprehension. Spencer, Quinn and Wagner’s approach to this challenge is to consider how advances in statistical analysis, such as latent change score modeling, can be used to elucidate the relationships between constructs across time. Their analysis of studies using latent change score modeling provides support for the instrumentalist hypothesis that vocabulary affects reading comprehension, rather than the alternative, that reading comprehension affects vocabulary knowledge. The second chapter by Cain and Barnes focuses on comprehension beyond the word and sentence level. Their starting point is to consider ‘what does it mean to comprehend?’ and they take a broad view of comprehension across different media: written texts, texts read aloud, and also visual ‘texts’ in the form of movies and cartoons. For each medium, the core product of comprehension is a mental representation of meaning, namely a situation (or mental) model. In this model, the meanings of different elements within the text are integrated and knowledge-based inferences are encoded to ensure coherence. Cain and Barnes’ critique of different accounts of reading leads them to conclude that these accounts largely describe differences across development but are inadequate accounts of the process of reading comprehension doi 10.1075/swll.15.13cai © 2017 John Benjamins Publishing Company

238 Kate Cain, Donald L. Compton and Rauno K. Parrila

development because they have focused on the end point: skilled adult comprehension. Cain and Barnes then consider what knowledge must be acquired to support the core skills involved in constructing the situation model. A key point they make is that the foundations of reading comprehension and the ability to construct a situation model develop through experience with language and the world around us from very early in development, well before formal literacy instruction. This leads Cain and Barnes to consider how accounts of cognition, such as embodied cognition, might provide insights into how we develop an understanding of the dimensions we encode in a situation model, such as time, space, and causality. The final chapter in this section by van den Broek and Kendeou picks up the themes of inference making and coherence considered by Cain and Barnes. van den Broek and Kendeou examine how reader, text, and task characteristics influence processing for meaning and, critically, how the interactions between these can influence the quality of comprehension. An important message is that comprehenders are constantly refining and updating their mental representation of the text’s meaning, such that the product of comprehension relies on the quality of processing at each point during the reading of the text. In their review of developmental processes, van den Broek and Kendeou focus on continuity and change over time. In tune with Cain and Barnes, they provide evidence that some aspects of inference making and processing for meaning are apparent very early in development – evidence for continuity – but they also show how the growth of knowledge and strategic skills influence the quality of comprehension across development. This analysis of continuity and change is fundamental to understanding the transition from learning to read to reading to learn. van den Broek and Kendeou conclude with suggestions for future research to deepen our theoretical understanding of comprehension development. In particular, they argue that more work on developmental changes in the recruitment and interconnectedness of neural structures and functions during language processing will inform a better understanding of the development of language comprehension, and also how insights into the dynamic process of reading comprehension should inform classroom interventions.

Vocabulary, morphology, and reading comprehension Mercedes Spencer, Jamie M. Quinn and Richard K. Wagner Florida State University

Well before you began formal schooling, you would have learned the meaning and pronunciation of the word run. A likely context was that you were running somewhere you should not have been, you may have fallen or run into something and hurt yourself or were disturbing others around you, and you were told “stop running!” From this context and others to follow, you eventually would be able to answer the typical item on a vocabulary test: “What does the word run mean?” An acceptable answer would be something to the effect of moving faster than a walk using your feet. At this point, run would be considered a word that is part of your vocabulary. But how much of what there is to know about run do you actually know at this point? Most words have more than one sense or definition. How many would you guess there is for the word run? According to WordNET (Fellbaum, 1998; Miller, 1995), there are 16 senses of meaning or usages of the word ‘run’ as a noun. Examples include: a score made in baseball (“The lead hitter scored a run”); a foot race (“mile run”); an unbroken series of events (“a run of bad luck”); a regular trip (“the boat made its daily run to the Keys”); the act of running (“the dog broke into a run”); to participate in an election (“run for office”); and unrestricted access (“the cat has the run of the house”). If this is not enough, there are an additional 41 senses of meaning of run as a verb, for a total of 57 senses of meaning or definitions! As this example makes clear, what it means to know a word goes well beyond merely being able to produce a simple definition. The purpose of this chapter is to begin to answer two questions. The first is what do you know when you know a word? The second is how does growth in vocabulary knowledge affect growth in reading comprehension, and vice versa? There is a positive correlation between how people perform on measures of vocabulary and on measures of reading comprehension. What underlies this correlation? Vocabulary and reading comprehension might be correlated because of their joint association with other constructs, such as general language ability or verbal aptitude. Alternatively, in addition to the correlation resulting from joint association with a superordinate construct, there may be direct influences between the development of vocabulary and the development of reading comprehension. For example, individual differences in vocabulary knowledge might influence the development of

doi 10.1075/swll.15.14spe © 2017 John Benjamins Publishing Company

240 Mercedes Spencer, Jamie M. Quinn and Richard K. Wagner

reading comprehension. Higher levels of vocabulary knowledge may make it possible to better comprehend the intended meaning of a passage, including more nuanced aspects of what the author wanted to convey. Being able to better comprehend the meaning of what one reads could, over time, result in better reading comprehension. Alternatively, individual differences in reading comprehension might influence the development of vocabulary. Being better at reading comprehension might improve a reader’s ability to infer the meanings of unfamiliar words from context. Over time, this could result in greater vocabulary knowledge. Of course, these last two possibilities are not mutually exclusive; the co-development of vocabulary knowledge and reading comprehension might be characterized by bi-directional influences. In this chapter, we begin by reviewing the literature on what it means to know a word. We then turn to the topic of what we know about the co-development of vocabulary knowledge and reading comprehension. We end by drawing some conclusions about the current state of knowledge about these topics and suggest some promising paths forward.

What does it mean to know a word? Two aspects of what it means to know a word that have been the subject of a considerable amount of research are vocabulary knowledge and morphological awareness.

Vocabulary knowledge Vocabulary knowledge has been identified as one of most robust predictors of reading comprehension performance (Beck & McKeown, 1991; McKeown, Beck, Omanson, & Perfetti, 1983; National Institute for Child and Human Development (NICHD), 2000; Ouellette, 2006). A distinction is commonly made between expressive and receptive vocabulary (NICHD, 2000). It is easier to point to a picture representing a word (receptive vocabulary) than it is to produce a definition (expressive vocabulary; Chafe & Tannen, 1987; Kamil & Hiebert, 2005; Lehr, Osborn, & Hiebert, 2004). However, there may be little more to this distinction than method variance associated with response format (i.e., pointing versus responding orally). A potentially more useful distinction is between breadth and depth of knowledge (Anderson & Freebody, 1981; Coyne, McCoach, Loftus, Zipoli, & Kapp, 2009; Nagy & Herman, 1987; Tannenbaum, Torgesen, & Wagner, 2006). Breadth refers to the number of words that an individual knows whereas depth refers to how much an individual knows about words they know. A growing body of research continues to show that vocabulary breadth and depth of knowledge are differentially predictive of various literacy-based skills (Cain & Oakhill, 2014; Dixon, LeFevre, & Twilley, 1988; Ouellette & Beers, 2010; Tannenbaum et al., 2006).



Vocabulary, morphology, and comprehension 241

By the age of 6, children have already acquired nearly 10,000 words and continue to acquire upwards of 3,000 new vocabulary words per year through a combination of direct instruction and incidental word learning (Anglin, 1993; Nagy & Anderson, 1984; Nagy & Herman, 1987). Direct instruction that includes contexts that are richer and that extend words beyond the classroom have positive effects on how quickly children are able to access word meanings which, in turn, impacts reading comprehension (Beck, Pefetti, & McKeown, 1982; McKeown, Beck, Omanson, & Pople, 1985; Perfetti & Hart, 2002). According to the National Reading Panel (NICHD, 2000) report, major sources of incidental word learning include independent book reading and engaging in oral language opportunities, such as listening to stories, being involved in classroom discussions, and engaging in casual oral language interactions (Farkas & Beron, 2004; Nagy, Herman, & Anderson, 1985; NICHD, 2000; Phythian-Sence & Wagner, 2007). Such interactions can provide children with multiple opportunities to better understand how words can and should be used in various contexts. Once young children begin to learn to read, they transition from hearing words spoken to also seeing these words in print (Long, 2001; NICHD, 2000; Olofsson & Nidersoe, 1999; Wise et al., 2007). Because of how words are usually encountered (i.e., within a context-rich environment), vocabulary knowledge rarely exists in isolation. Vocabulary knowledge is highly complex: word learning occurs incrementally, single words often have multiple meanings, and word knowledge is constrained by context dependency in many instances (see Nagy & Scott, 2000). This complexity has led to multiple discussions about the extent to which different aspects of word knowledge beyond definitional knowledge may constitute a comprehensive construct of vocabulary knowledge (e.g., Kieffer & Lesaux, 2012b; Spencer et al., 2015). These features have also resulted in the development of a variety of different vocabulary measures, which presents a challenge for its assessment.

Morphological awareness Morphological awareness refers to an awareness of and ability to manipulate morphemes (Carlisle, 1995). A morpheme is the smallest meaningful unit in a language, and because the English language has a morphophonemic orthography, it often provides valuable information regarding a word’s meaning, spelling, and pronunciation (Carlisle, 1995). Morphemes can be free or bound and inflectional or derivational. A free morpheme, as its name implies, is one that can stand on its own (i.e., a morpheme that is also a word). For example, the word happy is a free morpheme because it is meaningful and cannot be further broken down into smaller units that retain meaning. Bound morphemes are affixes that precede (i.e., prefixes) or follow (i.e., suffixes) root morphemes that are not meaningful on their own but become so once attached to a word. For example, adding the bound morpheme un- to the free morpheme happy to make unhappy. Additionally, morphemes can be inflectional or derivational.

242 Mercedes Spencer, Jamie M. Quinn and Richard K. Wagner

Inflectional morphemes contain grammatical information, such as a change in tense (-ed), possession, (-’s), or plurality (-s) (Bybee, 1988; Tong et al., 2011). Derivational morphemes change the meaning of a word and create new words from original or parent words (Ford, Davis, & Marslen-Wilson, 2010; Tyler & Nagy, 1989). Going back to our original example, un- is a derivational morpheme that changes the meaning of happy (feeling or showing pleasure) to unhappy (not pleased). Children’s knowledge of inflectional morphological awareness precedes their development of derivational morphological awareness, and this knowledge develops fairly early on (Carlisle, 1995; see also Carlisle and Kearns, this volume, for a discussion of learning to read morphologically complex words). Children as young as 2 or 3 years of age show evidence of being able to combine morphemes to refer to novel things (Adams, 1990; Carlisle, 2003, 2007). For example, knowing the words brush and toothbrush allows them to infer the meaning of the novel word dogbrush. Before school entry, children know most suffixes used to create common inflected forms such as plural forms (-s in bakes) and verb tense markers (-ed in baked) and some productive suffixes used to create derived forms (-er in baker). This is further evidenced by young children’s prolific (and sometimes ungrammatical) use of past tense during the early years of language acquisition, such as saying “goed” instead of went. However, morphological derivation remains beyond the grasp of many young children, with most only being able to properly derive words that are transparent in phonology (e.g., quick/quickly) and semantics (e.g., like/unlike) as well as productive words or verbs (Carlisle, 1995). Kindergarten-aged children often demonstrate difficulty completing morphological awareness tasks, which may be due to the fact that young children’s morphological awareness is relatively implicit (Carlisle, 1995). In a seminal study, Carlisle (1988) investigated derivational morphological awareness in fourth, sixth, and eighth grade students and found evidence for the presence of developmental trends. Across multiple tasks, fourth graders consistently made more errors than either sixth or eighth grade students; sixth graders performed better on the task than fourth graders, but consistently obtained scores that were lower than scores obtained by eighth graders. Fourth graders made the most errors on items involving phonological shifts (e.g., heal/health) and on items requiring a combination of phonological and orthographic shifts (e.g., space/spatial). They were also unable to master more complex suffix addition rules. Moreover, tasks requiring written responses were significantly more difficult than those requiring oral responses. This possibility of developmental trends in morphological awareness can potentially lead to the development and administration of targeted morphological awareness assessments (Carlisle, 1988, 1995; Carlisle & Nomanbhoy, 1993; Tyler & Nagy, 1989; Windsor, 1994). Nevertheless, the multifaceted nature of this skill, coupled with the availability of numerous researcher-created morphological awareness measures, makes it equally challenging to assess.



Vocabulary, morphology, and comprehension 243

Relations between vocabulary knowledge and morphological awareness Relations between morphological awareness and vocabulary knowledge are well established (Anglin, 1993; McBride-Chang, Wagner, Muse, Chow, & Shu, 2005; Wagner, Muse, & Tannenbaum, 2007). A significant proportion of the words that children acquire are learned incidentally (Nagy & Anderson, 1984). Morphological awareness provides a means for acquiring new words, which may explain why morphological awareness and vocabulary knowledge are so highly related (Anglin, 1993; McBrideChang et al., 2005; Nagy & Anderson, 1984; Wagner et al., 2007). Longitudinal examinations of these skills show that even though vocabulary and morphological awareness are significantly correlated throughout development, the degree of their relation may change over time such that the two skills may become almost indistinguishable from one another (Kirby et al., 2012; McBride-Chang et al., 2005; Nagy, Berninger, Abbott, Vaughn, & Vermeulen, 2003). The converse may also be true; morphological awareness and vocabulary may become more distinct as children grow older (Deacon, Kieffer, & Laroche, 2014; Deacon, Benere, & Pasquarella, 2013; Kieffer & Lesaux, 2012b). Although relations between vocabulary and morphological awareness are relatively well established, more recent investigations aim to determine whether vocabulary knowledge and morphological awareness are truly separable skills. Namely, there have been several studies that have examined the factor structure of morphological awareness and vocabulary knowledge (Kieffer & Lesaux, 2012a, 2012b; Spencer et al., 2015; Tannenbaum, Torgesen, & Wagner, 2006). In particular, Kieffer and Lesaux (2012b) and Spencer et al. (2015) provide extensive investigations of these constructs. Kieffer and Lesaux (2012b) examined these skills in 584 sixth grade students who were identified as either monolingual English-speaking or English language learners. Vocabulary and morphological awareness were assessed using tasks that measured semantic and relational knowledge, multiple meanings, and morphological decomposition and derivation. Using multi-group confirmatory factor analysis, the authors concluded that a three-factor solution was the preferred model for both language groups: knowledge was represented as vocabulary breadth, contextual sensitivity, and morphological awareness. In a comparable investigation, Spencer et al. (2015) examined vocabulary and morphological awareness in 99 fourth graders (Study 1) and 90 eighth graders (Study 2). In the first study, students were given nine morphological awareness measures and two standardized measures of vocabulary; in the second study, students were given measures of vocabulary and morphological awareness that contained the same words throughout. The tasks assessed definitional knowledge, usage ability, relational knowledge, and knowledge of morphologically-related variants. In both studies, a single combined vocabulary and morphological awareness factor provided the best fit to the data. The equivocal findings presented by Kieffer and Lesaux (2012b) and Spencer et al. (2015) may have resulted from key differences between the studies. For example, tasks administered by Kieffer and Lesaux primarily had a written administration format while several tasks used by Spencer et al. were orally administered. Differences

244 Mercedes Spencer, Jamie M. Quinn and Richard K. Wagner

on written tasks may have emerged as a result of variability associated with reading ability in addition to individual differences in vocabulary performance. Furthermore, Spencer et al. conducted all analyses using raw data while Kieffer and Lesaux analyzed residual scores in an effort to account for nesting (i.e., students nested within classrooms). This method, however, can inadvertently reduce individual student-level variation if there are corresponding differences in average vocabulary performance directly associated with classrooms and schools. In the future, a study that includes sufficient classroom- and school-level clusters would allow student- and classroom levels to be modeled separately and allow for a further investigation of relations between these skills (see Mehta, Foorman, Branum-Martin, & Taylor, 2005). Moreover, morphological awareness is a unique predictor of performance on measures of reading comprehension (Carlisle, 1995; 2000, 2003; Deacon et al. 2014; Kirby et al., 2012; Kruk & Bergman, 2013; see also Shu, McBride-Chang, Wu, & Liu, 2006, and Kieffer & Lesaux, 2008 for evidence that morphological awareness contributes to reading comprehension in non-English languages). Individuals who obtain lower scores on measures of reading comprehension tend to have correspondingly lower scores on measures of morphological awareness (Tong et al., 2011). Furthermore, morphological awareness has been shown to uniquely predict performance on measures of reading comprehension over and above vocabulary knowledge in several instances (Carlisle, 2007; Deacon & Kirby, 2004; Kieffer & Lesaux, 2008, 2012c; Nagy, Berninger, & Abbott, 2006). Relations between morphological awareness and reading comprehension may arise from a number of sources. One is that morphological awareness aids in inferring the meanings of unfamiliar morphologically complex words that are encountered when reading for meaning and may also be helpful in parsing complex sentences (Nagy, 2007). Nagy (2007, p. 64) provides the example of the contrast in meaning between the phrases Observant investigators proceed carefully and Observe investigator’s procedures carefully. The critical suffix -ant in the word observant provides an important signal to the correct meaning of the phrase, and poor readers are more likely to miss these kinds of signals than good readers (Tyler & Nagy, 1990). Thus, it is very likely that morphological awareness impacts reading comprehension both directly and indirectly because sensitivity to morphological information can support the prediction of unknown words (Nagy & Anderson, 1984) and because of the strong relations between morphological awareness and vocabulary knowledge (Kieffer & Lesaux, 2012a).

How do vocabulary knowledge and reading comprehension co-develop? Vocabulary knowledge is an important predictor of reading comprehension, even when word reading skills are controlled (e.g., Muter, Hulme, Snowling, & Stevenson, 2004; Ouellette, 2006). Further, measures of both depth and breadth of vocabulary knowledge predict reading comprehension independently of decoding and listening



Vocabulary, morphology, and comprehension 245

comprehension (Braze, Tabor, Shankweiler, & Mencl, 2007; Ouellette, 2006; Sénéchal, Ouellette, & Rodney, 2006; Tannenbaum et al. 2006). One of the earliest attempts to sort out relations between vocabulary knowledge and reading comprehension was provided by Anderson and Freebody (1981) who identified three hypotheses that could account for the observed correlations between vocabulary knowledge and reading comprehension. According to the instrumentalist hypothesis, vocabulary is causally related to reading comprehension. Knowing more about the words you come across in a passage improves your ability to comprehend the message. According to the knowledge hypothesis, vocabulary knowledge and reading comprehension are correlated because they both are related to conceptual knowledge. Much of our conceptual knowledge is in the form of vocabulary knowledge and relations between words. Conceptual knowledge is what is communicated by most passages. Finally, according to the aptitude hypothesis, vocabulary knowledge and reading comprehension are correlated because both are correlated with the third variable of verbal aptitude. The three hypotheses identified by Anderson and Freebody (1981) represent a good start at thinking about how vocabulary knowledge and reading comprehension might be related, but they do not exhaust the possibilities (Wagner & Meros, 2010). For example, an alternative instrumental hypothesis, in which reading comprehension causally influences vocabulary rather than the other way around, should be considered. If you are good at reading comprehension, you will be better able to understand the message conveyed by a passage. Understanding the message conveyed by a passage is useful for inferring the meanings of new words you come across, and for sharpening the meanings of words you already know. Alternatively, relations between vocabulary knowledge and reading comprehension might be mediated by a third variable. For example, it is easier to perform a phonological awareness task, such as saying the word pact without the /k/ sound (i.e., pat), for known words compared to nonwords, perhaps because it requires less working memory to do these kinds of operations on known words. Therefore, vocabulary knowledge can influence phonological awareness, which in turn affects the ease with which one learns to decode words on the page. Being better able to decode words on the page then affects reading comprehension.

Instrumental relations between vocabulary knowledge and reading comprehension Knowledge of word meanings is essential to understanding the meaning of a text. Reading comprehension for both children and adults is supported by knowledge of words, including precision of orthographic, phonological, and semantic representations (Cain & Oakhill, 2011; Verhoeven & van Leeuwe, 2008, Verhoeven, van Leeuwe, & Vermeer, 2011). The lexical quality hypothesis (Perfetti & Hart, 2001) states that the quality of lexical representations that exist for known words and the number of known words directly affect reading comprehension. This idea is consistent with the

246 Mercedes Spencer, Jamie M. Quinn and Richard K. Wagner

simple view of reading (Gough & Tunmer, 1986; Hoover & Gough, 1990) in which reading comprehension is viewed as the product of decoding and linguistic comprehension, provided we consider vocabulary knowledge to be an indicator of linguistic comprehension. Evidence for an instrumental role for vocabulary in the development of reading comprehension would be supported by studies showing that training in vocabulary knowledge improves reading comprehension. The results are mixed, however. A ­meta-analysis of the effects of vocabulary training on reading comprehension reported a large and significant average weighted effect size of 0.97 for vocabulary interventions on researcher-constructed comprehension outcomes that were designed to be sensitive to the vocabulary interventions examined, and a modest yet statistically significant average weighted effect size of 0.30 for standardized tests of reading comprehension (Stahl & Fairbanks, 1986). Larger effect sizes were associated with activities requiring more depth of processing, pairing of contextual and definitional information, and the type and number of word exposures. The review of vocabulary interventions by the National Reading Panel (NICHD, 2000) found gains on standardized measures of reading comprehension as a consequence of reading instruction for only two of the studies reviewed. Nevertheless, they supported vocabulary instruction as a means for improving comprehension skills. A more recent meta-analysis by Elleman, Lindo, Morphy, and Compton (2009) addressed the effects of vocabulary instruction on passage-level comprehension. The authors began by examining the effects of vocabulary instruction on vocabulary outcomes. If the effects of vocabulary instruction on vocabulary outcomes are negligible, there would be little reason to expect the effects to generalize to reading comprehension outcomes. For researcher-created vocabulary measures that were sensitive to the kinds of vocabulary interventions that were used, effect sizes ranged from −0.11 to 2.28, with a significant average weighted effect size of 0.79. For standardized vocabulary tests, the effect sizes for vocabulary instruction ranged from −0.24 to 0.46, with a significant average weighted effect size of 0.29. Turning to the effects of vocabulary intervention on passage comprehension, the effect sizes ranged from −0.06 to 1.46 for researcher-created measures of reading comprehension, with a significant average weighted effect size of 0.50. However, for standardized measures of passage-reading comprehension, the effect sizes ranged from −0.26 to 0.43, with a non-significant average weighted effect size of 0.10. In sum, the results suggest an effect of vocabulary intervention on reading comprehension that was substantial for researcher-developed measures but modest to non-existent for standardized measures of reading comprehension that were presumably less sensitive to the specific interventions studied than were the researcher-developed measures.



Vocabulary, morphology, and comprehension 247

Indirect effects of vocabulary knowledge on reading comprehension Vocabulary knowledge may affect reading comprehension indirectly through mediating factors. Consider phonological awareness, which refers to the ability to recognize and manipulate the individual sounds of spoken words (Stanovich, 1992; Wagner & Torgesen, 1987). Children who are aware of the similarities and differences between the spoken words cat, rat, and hat – each of which is composed of three phonemes, with different phonemes in the initial position and identical phonemes in the medial and final positions – have an easier time learning their printed forms (Lonigan, 2007; Lonigan, Burgess, & Anthony, 2000; Wagner, Torgesen, & Rashotte, 1994; Wagner, Torgesen, Laughon, Simmons, & Rashotte, 1993). This makes sense because of the rough correspondence between letters of the alphabet and phonemes in English. How vocabulary knowledge figures in is that children demonstrate greater phonological awareness for known vocabulary words. For example, the Comprehensive Test of Phonological Processing – Second Edition (CTOPP-2, Wagner, Torgesen, Rashotte, & Pearson, 2013) has comparable phonological awareness subtests for real word and for nonword items. For words and nonwords that are equivalent in the number of phonemes, performance on the phonological awareness tasks is better for words than for nonwords. Given the fact that (a) vocabulary knowledge affects phonological processing, (b) phonological processing affects decoding, and (c) decoding affects reading comprehension, a plausible indirect influence exists for vocabulary on reading comprehension (Wagner & Meros, 2010).

Instrumental effects of reading comprehension on vocabulary knowledge Reading comprehension could directly influence vocabulary because a substantial amount of vocabulary knowledge is acquired from reading print as opposed to listening to speech (Wagner & Meros, 2010). There is evidence that reading comprehension and crucial reading comprehension-related skills, such as inference making, may facilitate the development of vocabulary knowledge (Cain, 2007; Daneman, 1988; Nagy & Scott, 2000; Nation, Snowling, & Clarke, 2007). Consequently, good reading comprehension ability could facilitate vocabulary knowledge (Wagner & Meros, 2010).

Third variable relations between vocabulary knowledge and reading comprehension The obvious example of a third variable relation between vocabulary and reading comprehension is their joint correlation with verbal aptitude. However, more nuanced possibilities also exist. Nagy (2007) proposed that vocabulary and reading comprehension are correlated because of their joint association with metalinguistic awareness. Metalinguistic awareness refers to an awareness of and an ability to manipulate the structure of oral language (Tunmer, Herriman, & Nesdale, 1988). Specifically,

248 Mercedes Spencer, Jamie M. Quinn and Richard K. Wagner

acquisition of vocabulary knowledge is enhanced by morphological awareness, which is conceptualized as one form of metalinguistic awareness. Reading comprehension is enhanced by two forms of metalinguistic awareness. The first is comprehension monitoring, which refers to being aware of the success one is having in comprehending text. Being aware of comprehension failures can trigger rereading or changing comprehension strategies. The second is syntactic awareness, which can be helpful in parsing syntax for long or convoluted sentences.

Longitudinal studies of relations between vocabulary knowledge and reading comprehension The results from longitudinal studies can have implications for better understanding relations between vocabulary knowledge and reading comprehension. For example, Muter et al. (2004) followed children over a two-year-period and showed that end of second grade reading comprehension was predicted by previous performance on measures of word identification, vocabulary knowledge, and linguistic skills. Similarly, Cain, Oakhill, and Bryant (2004) reported that vocabulary knowledge, verbal IQ, inference skills, and comprehension monitoring predicted reading comprehension in 8- to 11-year-old children. Longitudinal studies provide stronger evidence of relations between constructs, such as vocabulary knowledge and reading comprehension, when the autoregressive effect of prior performance on the target construct is included as a predictor. Two longitudinal studies of relations between vocabulary knowledge and reading comprehension that included the autoregressor of prior performance on the dependent variable of interest have been conducted. In the first study, decoding, vocabulary, and listening comprehension skills at grade three were predictive of reading comprehension at grade five after controlling for reading comprehension at grade three (de Jong & van der Leij, 2002). In the second study, vocabulary knowledge was a strong predictor of subsequent reading comprehension when prior reading comprehension was included as a predictor, but conversely, reading comprehension was only a weak predictor of subsequent vocabulary knowledge when prior vocabulary knowledge was controlled (Verhoeven & van Leeuwe, 2008). Recently, advances in statistical modeling techniques have enabled researchers to conduct more sophisticated analyses of relations between vocabulary knowledge and reading comprehension using longitudinal data. The best example of such an advance is latent change score modeling (for a review, see McArdle, 2009). The best way to understand the advanced approach represented by latent change score modeling is to place it in the context of previous approaches for modeling longitudinal data. There have been two widely used approaches for studying developmental relations over time between a pair of constructs. Cross-lagged panel models or structural equation causal models rely on time precedence to determine whether individual differences in construct A at time 1 account for subsequent individual differences in construct B at time 2, independently of time 1 prior performance on construct B (i.e., the autoregressor



Vocabulary, morphology, and comprehension 249

effect). An advantage of this approach is that it uses time precedence to test models that are consistent with, although not in and of themselves sufficient proof of, causal relations between the developing constructs. A disadvantage of such models is that they are fit to only some of the data. Depending on the model, covariances or correlations are modeled. Exactly the same results are found regardless of whether performance on either construct increases, decreases, or stays the same over time. The second approach is latent growth curve modeling. When growth is modeled in two constructs simultaneously that have been measured repeatedly in the same sample, it is called parallel process latent growth curve modeling. An advantage of latent growth curve models is that they explicitly model growth over time and take advantage of all the data because they are fit to both covariances and means. A disadvantage is that time precedence is lost because the growth curve parameters apply to the entire developmental period. Although it is possible to estimate the correlation between growth rates in two constructs, it is not possible to determine whether status or growth in construct A explains subsequent status or growth in construct B. Latent change score models combine the advantages of cross-lagged panel models or structural equation causal models and parallel process latent growth curve models (Ferrer & McArdle, 2010; McArdle, 2009). One can think of them as parallel process growth curve models after dividing the overall developmental period into discrete segments. The change in latent change score models refers to change in performance from one segment to the next. It does this by estimating growth over multiple discrete time intervals within the overall developmental period. Four examples of latent change score modeling will be reviewed that are relevant to the study of the co-development of reading comprehension and vocabulary. The first two modeled the co-development of reading comprehension and IQ or verbal aptitude as opposed to vocabulary per se. However, the second two studies modeled the co-development of reading comprehension and vocabulary specifically. Ferrer et al. (2007) used latent change score models to study the co-development of IQ and a reading composite composed of the letter-word identification, decoding, and passage comprehension. Participants in the study were from the Connecticut Longitudinal Study, a sample of 445 students who were followed from first through twelfth grade. The results supported bidirectional coupling influences between reading and IQ. In other words, earlier performance in reading accounted for subsequent growth in IQ, and conversely, earlier IQ accounted for subsequent growth in reading. This result also held when looking specifically at coupling between verbal IQ and passage comprehension, which are the two variables in the study that are most relevant to the study of the co-development of vocabulary and reading comprehension. Ferrer et al. (2010) carried out a follow-up analysis using a subset of the same Connecticut Longitudinal Study sample. For this study, participants were categorized as typical readers, compensated readers, and persistently poor readers. Interestingly, the finding of bidirectional coupling of IQ and reading reported from the Ferrer et al. (2010) study applied only to the typical readers. No coupling between IQ and reading was found for both the compensated readers and the persistently poor readers.

250 Mercedes Spencer, Jamie M. Quinn and Richard K. Wagner

Reynolds and Turek (2012) reported results from latent change score modeling of the co-development of crystallized verbal-comprehension knowledge (Gc) and reading comprehension at 3rd, 5th, and 9th grades. However, their measure of crystallized verbal-comprehension knowledge was picture vocabulary. They did not find bi-directional coupling, but rather that vocabulary accounted for subsequent change in reading comprehension; there was no effect of reading comprehension on subsequent vocabulary. Quinn, Wagner, Petscher, and Lopez (2015) also investigated the co-development of vocabulary knowledge and reading comprehension in a sample of 316 students followed from first through fourth grade. Using latent change score modeling, four competing models were fit to the data. The correlated but uncoupled development model specified that the development of vocabulary is potentially correlated with but not coupled with that of reading comprehension. What this means is that children who grow fast in vocabulary grow fast in reading comprehension, but there is no temporal coupling (i.e., performance in one construct accounts for subsequent change in the other construct). The unidirectional coupling from vocabulary knowledge to reading comprehension model specified that vocabulary knowledge accounted, in part, for subsequent changes in reading comprehension. The rationale is that more knowledge about the words in a passage gives greater access to the meaning of the passage, and this improves reading comprehension over time. The unidirectional coupling from reading comprehension to vocabulary model specified that reading comprehension accounted for subsequent changes in vocabulary. The rationale is that text provides an opportunity to learn new vocabulary, a process that is moderated by reading comprehension skill. Finally, a bidirectional coupling model represented a combination of coupling from vocabulary to reading comprehension and from reading comprehension to vocabulary. The results supported unidirectional coupling from vocabulary knowledge to reading comprehension. This outcome is consistent with Anderson and Freebody’s (1981) instrumentalist hypothesis and replicates the results of Reynolds and Turek (2012).

Summary In conclusion, our understanding of what it means to know a word has grown considerably with the availability of innovative methods for studying relations at the level of underlying constructs. By studying relations among latent variables represented by multiple indicators, we are moving to a more robust understanding that is less affected by characteristics of particular variables. At the moment, we have two views (i.e., unidimensional versus multidimensional) that represent polar opposite views of what it means to know a word. Resolving the question of whether vocabulary knowledge is characterized by facets of a single underlying fund of knowledge or ability (Spencer et al., 2015) or by multiple correlated dimensions (Kieffer & Lesaux, 2012b) remains an important and achievable near-term objective. What will be required is a



Vocabulary, morphology, and comprehension 251

study that includes sufficient participants and classrooms so that individual-level and classroom-level variance can be modeled jointly. Turning to the issue of relations between vocabulary and reading comprehension, we are moving from the study of development to the study of co-development. Recently developed statistical models allow testing alternative hypotheses about whether growth in vocabulary is explained in part by skill in reading comprehension, and conversely, whether growth in reading comprehension is explained in part by the extensiveness of vocabulary knowledge. The early results suggest that vocabulary affects the development of reading comprehension more than reading comprehension affects the development of vocabulary. However, additional studies, particularly studies that include older students, are needed.

References Adams, M. J. (1990). Beginning to read: Thinking and learning about print. Cambridge, MA: MIT. Anderson, R., & Freebody, P. (1981). Vocabulary knowledge. In J. T. Guthrie (Ed.), Comprehension and teaching: Research reviews (pp. 77–117). Newark, DE: International Reading Association. Anglin, J. M. (1993). Knowing versus learning words. Monographs of the Society for Research in Child Development, 58, 176–186.  doi: 10.1111/j.1540-5834.1993.tb00377.x Beck, I., & McKeown, M. (1991). Conditions of vocabulary acquisition. In R. Barr, M. L. Kamil, P. B. Mosenthal, & P. D. Pearson (Eds.), Handbook of reading research (pp. 789–814). New York: Longman. Beck, I. L., Perfetti, C. A., & McKeown, M. G. (1982). Effects of long-term vocabulary instruction on lexical access and reading comprehension. Journal of Educational Psychology, 74, 506–521.  doi: 10.1037/0022-0663.74.4.506 Braze, D., Tabor, W., Shankweiler, D. P., & Mencl, W. E. (2007). Speaking up for vocabulary in primary grades. Journal of Educational Psychology, 98, 44–62. Bybee, J. L. (1988). Morphology as lexical organization. In M. Hammond & M. Noonan (Eds.), Theoretical morphology (pp. 119–141). San Diego: Academic Press. Cain, K. (2007). Deriving word meanings from context: Does explanation facilitate contextual analysis? Journal of Research in Reading, 30, 347–359.  doi: 10.1111/j.1467-9817.2007.00336.x Cain, K., & Oakhill, J. (2011) Matthew effects in young readers: Reading comprehension and reading experience aid vocabulary development. Journal of Learning Disabilities, 44, 431–443. doi: 10.1177/0022219411410042 Cain, K. & Oakhill, J. (2014). Reading comprehension and vocabulary: Is vocabulary more important for some aspects of comprehension? L’Année Psychologique, 114, 647–662. doi: 10.4074/S0003503314004035 Cain, K., Oakhill, J., & Bryant, P. (2004). Children’s reading comprehension ability: Concurrent prediction by working memory, verbal ability, and component skills. Journal of Educational Psychology, 96, 31–42.  doi: 10.1037/0022-0663.96.1.31 Carlisle, J. F. (1988). Knowledge of derivational morphology and spelling ability in fourth, sixth, and eighth graders. Applied Psycholinguistics, 9, 247–266.  doi: 10.1017/S0142716400007839 Carlisle, J. F. (1995). Morphological awareness and early reading achievement. In L. Feldman (Ed.), Morphological apsects of language processing (pp. 189–209). Hillsdale, NJ: Lawrence Erlbaum.

252 Mercedes Spencer, Jamie M. Quinn and Richard K. Wagner

Carlisle, J. F. (2000). Awareness of the structure and meaning of morphologically complex words: Impact on reading. Journal of Reading and Writing, 12, 169–190.  doi: 10.1023/A:1008131926604 Carlisle, J. F. (2003). Morphology matters in learning to read: A commentary. Reading Psychology, 24, 373–404.  doi: 10.1080/02702710390227369 Carlisle, J. F. (2007). Fostering morphological processing, vocabulary development, and reading comprehension. In R. K. Wagner, A. E. Muse, & K. R. Tannenbaum (Eds.), Vocabulary acquisition: Implications for reading comprehension (pp. 78–103). New York: Guilford Press. Carlisle, J. F. & Nomanbhoy, D. M. (1993). Phonological and morphological awareness in first graders. Applied Psycholinguistics, 14, 177–195.  doi: 10.1017/S0142716400009541 Chafe, W., & Tannen, D. (1987). The relation between written and spoken language. Annual Review of Anthropology, 16, 383–407.  doi: 10.1146/annurev.an.16.100187.002123 Coyne, M. D., McCoach, D. B., Loftus, S., Zipoli, R., Jr., & Kapp, S. (2009). Direct vocabulary instruction in kindergarten: Teaching for breadth versus depth. The Elementary School Journal, 110, 1–18.  doi: 10.1086/598840 Daneman, M. (1988). Word knowledge and reading skills. In M. Daneman, G. MacKinnon, & T. G. Waller (Eds.), Reading research: Advances in theory and practice (Vol. 6, pp. 145–175). San Diego, CA: Academic Press. de Jong, P., & van der Leij, A. (2002). Effects of phonological abilities and linguistic comprehension on the development of reading. Scientific Studies of Reading, 6, 51–77. doi: 10.1207/S1532799XSSR0601_03 Deacon, S. H., Benere, J., & Pasquarella, A. (2013). Reciprocal relationship: Children’s morphological awareness and their reading accuracy across grades 2 to 3. Developmental Psychology, 49, 1113–1126.  doi: 10.1037/a0029474 Deacon, S. H., Kieffer, M. J., & Laroche, A. (2014). The relation between morphological awareness and reading comprehension: Evidence from mediation and longitudinal models. Scientific Studies of Reading, 18, 432–451.  doi: 10.1080/10888438.2014.926907 Deacon, S. H. & Kirby, J. (2004). Morphological awareness: Just more phonological? The roles of morphological and phonological awareness in reading development, Applied Psycholinguistics, 25, 223–238.  doi: 10.1017/S0142716404001110 Dixon, P., LeFevre, J. A., & Twilley, L. C. (1988). Word knowledge and working memory as predictors of reading skill. Journal of Educational Psychology, 80, 465–472.  doi: 10.1037/0022-0663.80.4.465 Elleman, A. M., Lindo, E. J., Morphy, P., & Compton, D. L. (2009). The impact of vocabulary instruction on passage level comprehension of school age children: A meta-analysis. Journal of Research on Educational Effectiveness, 2, 1–44.  doi: 10.1080/19345740802539200 Farkas, G. & Beron, K. (2004). The detailed age trajectory of oral vocabulary knowledge: Differences by class and race. Social Science Research, 33, 464–97.  doi: 10.1016/j.ssresearch.2003.08.001 Fellbaum, C. (1998, ed.) WordNet: An Electronic Lexical Database. Cambridge, MA: MIT Press. Ferrer, E. & McArdle, J. (2010). Longitudinal modeling of developmental changes in psychological research. Current Directions in Psychological Science, 19, 149–154. doi: 10.1177/0963721410370300 Ferrer, E., McArdle, J. J., Shaywitz, B. A., Holahan, J. M., Marchione, K., & Shaywitz, S. E. (2007). Longitudinal models of developmental dynamics between reading and cognition from childhood to adolescence. Developmental Psychology, 43, 1460–1473. doi: 10.1037/0012-1649.43.6.1460 Ferrer, E., Shaywitz, B. A., Holahan, J. M., Marchione, K., & Shaywitz, S. E. (2010). Uncoupling of reading and IQ over time: Empirical evidence for a definition of dyslexia. Psychological Science, 21, 93–101.  doi: 10.1177/0956797609354084



Vocabulary, morphology, and comprehension 253

Ford, M. A., Davis, M. H., & Marslen-Wilson, W. (2010). Derivational morphology and base morpheme frequency. Journal of Memory and Language, 63, 117–130. doi: 10.1016/j.jml.2009.01.003 Gough, P. B., & Tunmer, W. E. (1986). Decoding, reading, and reading disability. Remedial and Special Education, 7, 6–10.  doi: 10.1177/074193258600700104 Hoover, W. A., & Gough, P. B. (1990). The simple view of reading. Reading and Writing, 2, 127–160. doi: 10.1007/BF00401799 Kamil, M. L., & Hiebert, E. H. (2005). Teaching and learning vocabulary: Perspectives and persistent issues. In E. Hiebert & M. Kamil (Eds.), Teaching and learning vocabulary: Bringing research to practice (pp. 1–23). Mahwah, NJ: Erlbaum. Kieffer, M. J., & Lesaux, N. K. (2008). The role of derivational morphological awareness in the reading comprehension of Spanish-speaking English language learners. Reading and Writing, 21, 783–804.  doi: 10.1007/s11145-007-9092-8 Kieffer, M. J. & Lesaux, N. K. (2012a). Development of morphological awareness and vocabulary knowledge in Spanish-speaking language minority learners: A parallel process latent growth curve model. Applied Psycholinguistics, 33, 23–54.  doi: 10.1017/S0142716411000099 Kieffer, M. J., & Lesaux, N. K. (2012b). Knowledge of words, knowledge about words: Dimensions of vocabulary in first and second language learners in sixth grade. Reading and Writing, 25, 347–373.  doi: 10.1007/s11145-010-9272-9 Kieffer, M. J., & Lesaux, N. K. (2012c). Direct and indirect roles of morphological awareness in the English reading comprehension of native English, Spanish, Filipino, and Vietnamese speakers. Language Learning, 62, 1170–1204.  doi: 10.1111/j.1467-9922.2012.00722.x Kirby, J. R., Deacon, S. H., Bowers, P. N., Izenberg, L., Wade-Woolley, L., & Parrila, R. (2012). Children’s morphological awareness and reading ability. Reading and Writing, 25, 389–410. doi: 10.1007/s11145-010-9276-5 Kruk, R. S., & Bergman, K. (2013). The reciprocal relations between morphological processes and reading. Journal of Experimental Child Psychology, 114, 10–34.  doi: 10.1016/j.jecp.2012.09.014 Lehr, F., Osborn, J., & Hiebert, E. H. (2004). Research-based practices in early reading series: Focus on vocabulary. Pacific Resources for Education and Learning PREL. Long, M. (2001). Psychology of education. United Kingdom: Routledge. Lonigan, C. J. (2007). Vocabulary development and the development of phonological awareness skills in preschool children. In R. Wagner, A. Muse & K. Tannenbaum (Eds.), Vocabulary acquisition: Implications for reading comprehension (pp. 15–31). New York: Guilford Press Lonigan, C. J., Burgess, S. R., & Anthony, J. L. (2000). Development of emergent literacy and early reading skills in preschool children: Evidence from a latent variable longitudinal study. Developmental Psychology, 36, 596–613.  doi: 10.1037/0012-1649.36.5.596 McArdle, J. J. (2009). Latent variable modeling of longitudinal data. Annual Review of Psychology, 60, 577–605.  doi: 10.1146/annurev.psych.60.110707.163612 McBride-Chang, C., Wagner, R. K., Muse, A., Chow, B. W.-Y., & Shu, H. (2005). The role of morphological awareness in children’s vocabulary acquisition in English. Applied Psycholinguistics, 26, 415–435.  doi: 10.1017/S014271640505023X McKeown, M. G., Beck, I. L., Omanson, R. C., & Perfetti, C. A. (1983). The effects of long-term vocabulary instruction on reading comprehension: A replication. Journal of Literacy Research, 15, 3–18.  doi: 10.1080/10862968309547474 McKeown, M. G., Beck, I. L., Omanson, R. C., & Pople, M. T. (1985). Some effects of the nature and frequency of vocabulary instruction on the knowledge and use of words. Reading Research Quarterly, 20, 522–535.  doi: 10.2307/747940

254 Mercedes Spencer, Jamie M. Quinn and Richard K. Wagner

Mehta, P. D., Foorman, B. R., Branum-Martin, L., & Taylor, W. P. (2005). Literacy as a unidimensional multilevel construct: Validation, sources of influence, and implications in a longitudinal study in grades 1 to 4. Scientific Studies of Reading, 9, 85–116. doi: 10.1207/s1532799xssr0902_1 Miller, G. A. (1995). WordNet: A Lexical Database for English. Communications of the ACM Vol. 38, No. 11: 39–41.  doi: 10.1145/219717.219748 Muter, V., Hulme, C., Snowling, M. J., & Stevenson, J. (2004). Phonemes, rimes, and language skills as foundations of early reading development: Evidence from a longitudinal study. Developmental Psychology, 40, 663–681.  doi: 10.1037/0012-1649.40.5.665 Nagy, W. (2007). Metalinguistic awareness and the vocabulary-comprehension connection. In R. K. Wager, A. E. Muse, & K. R. Tannenbaum (Eds.), Vocabulary acquisition: Implications for reading comprehension (pp. 52–77). New York: Guilford. Nagy, W. E., & Anderson, R. C. (1984). How many words are there in printed school English? Reading Research Quarterly, 19, 304–330.  doi: 10.2307/747823 Nagy, W. E., Berninger, V., & Abbott, R. (2006). Contributions of morphology beyond phonology to literacy outcomes of upper elementary and middle school students. Journal of Educational Psychology, 98,134–147.  doi: 10.1037/0022-0663.98.1.134 Nagy, W. E., Berninger, V., Abbott, R., Vaughn, K., & Vermeulen, K. (2003). Relationship of morphology and other language skills to literacy skills in at-risk second grade readers and at-risk fourth grade writers. Journal of Educational Psychology, 95, 730–742.  doi: 10.1037/0022-0663.95.4.730 Nagy, W. E. & Herman, P. A. (1987). Breadth and depth of vocabulary knowledge: Implications for acquisition and instruction. In M. McKeown and M. Curtis (Eds.), The Nature of Vocabulary Acquisition, (pp. 19–35). Hillsdale, NJ: Erlbaum Associates. Nagy, W. E., Herman, P. A., & Anderson, R. (1985). Learning words from context. Reading Research Quarterly, 20, 304–330.  doi: 10.2307/747758 Nagy, W. E., & Scott, J. A. (2000). Vocabulary Processes. In M. L. Kamil, P. B. Mosenthal, P. D. Pearson, & R. Barr (Eds.), Handbook of reading research, Vol III (pp. 269–284). Mahwah, NJ, US: Lawrence Erlbaum Associates Publishers. Nation, K., Snowling, M. J., & Clarke, P. (2007). Dissecting the relationship between language skills and learning to read: Semantic and phonological contributions to new vocabulary learning in children with poor reading comprehension. Advances in Speech Language Pathology, 9, 131–139.  doi: 10.1080/14417040601145166 National Institute of Child Health and Human Development. (2000). Report of the National Reading Panel. Teaching children to read: An evidence-based assessment of the scientific research literature on reading and its implications for reading instruction. (NIH Publication No. 00–4769). Washington, D.C.: U.S. Government Printing Office. Olofsson, A. & Niedersoe, J. (1999). Early language development and kindergarten phonological awareness as predictors of reading problems. Journal of Learning Disabilities, 32, 464–472. doi: 10.1177/002221949903200512 Ouellette, G. (2006). What’s meaning got to do with it: the role of vocabulary in word reading and reading comprehension. Journal of Educational Psychology, 98, 554–566. doi: 10.1037/0022-0663.98.3.554 Ouellette, G., & Beers, A. (2010). A not-so-simple view of reading: How oral vocabulary and visual-word recognition complicate the story. Reading and Writing, 23, 189–208. doi: 10.1007/s11145-008-9159-1



Vocabulary, morphology, and comprehension 255

Perfetti, C. A., & Hart, L. (2001). The lexical bases of comprehension skill. In D. Gorfien (Ed.), On the consequences of meaning selection (pp. 67–86). Washington, DC: American Psychological Association. Perfetti, C., & Hart, L. (2002). The lexical quality hypothesis. In L. Verhoeven, C. Elbro, & P. Reitsma (Eds.), Precursors of functional literacy (pp. 189–213). Amsterdam, Netherlands: John Benjamins.  doi: 10.1075/swll.11.14per Phythian-Sence, C., & Wagner, R. K. (2007). Vocabulary acquisition: A primer. In R. Wagner, A. Muse, & K. Tannenbaum (Eds.), Vocabulary acquisition: Implications for reading comprehension (pp. 1–14). New York, NY: Guilford. Quinn, J. M., Wagner, R. K., Petscher, Y., & Lopez, D. (2015). Developmental relations between vocabulary knowledge and reading comprehension: A latent change score modeling study. Child Development, 86, 159–175.  doi: 10.1111/cdev.12292 Reynolds, M. R., & Turek, J. J. (2012). A dynamic developmental link between verbal comprehension-knowledge (Gc) and reading comprehension: Verbal comprehension knowledge drives positive change in reading comprehension. Journal of School Psychology, 50, 841–863. doi: 10.1016/j.jsp.2012.07.002 Sénéchal, M., Ouellette, G., & Rodney, D. (2006). The misunderstood giant: On the predictive role of early vocabulary to future reading. In S. B. Neuman & D. Dickinson (Eds.) Handbook of early literacy research. New York, NY: Guilford Press. Shu, H., McBride-Chang, C., Wu, S., & Liu, H. (2006). Understanding Chinese developmental dyslexia: Morphological awareness as a core cognitive construct. Journal of Educational Psychology, 98, 122–133.  doi: 10.1037/0022-0663.98.1.122 Spencer, M., Muse, A., Wagner, R. K., Foorman, B., Petscher, Y., Schatschneider, C., Tighe, E. L, & Bishop, D. (2015). Examining the underlying dimensions of morphological awareness and vocabulary knowledge. Reading and Writing, 28, 959–988.  doi: 10.1007/s11145-015-9557-0 Stahl, S. A., & Fairbanks, M. M. (1986). The effects of vocabulary instruction: A model-based meta-analysis. Review of Educational Research, 56, 72–110.  doi: 10.3102/00346543056001072 Stanovich, K. E. (1992). Speculations on the causes and consequence of individual differences in early reading acquisition. In P. Gough, L. Ehri, & R. Treiman (Eds.), Reading acquisition (pp. 307–342). Mahwah, NJ: Erlbaum Tannenbaum, K. R., Torgesen, J. K., & Wagner, R. K. (2006). Relationships between word knowledge and reading comprehension in third-grade children. Scientific Studies of Reading, 10, 381–398.  doi: 10.1207/s1532799xssr1004_3 Tong, X., Deacon, S. H., Kirby, J. R., Cain, K., & Parrila, R. (2011). Morphological awareness: A key to understanding poor reading comprehension in English. Journal of Educational Psychology, 103, 523–534.  doi: 10.1037/a0023495 Tunmer, W., Herriman, M., & Nesdale, A. (1988). Metalinguistic abilities and beginning reading. Reading Research Quarterly, 23, 134–158.  doi: 10.2307/747799 Tyler A. & Nagy, W. E. (1989). The acquisition of English derivational morphology. Journal of Memory and Language, 28, 649–667.  doi: 10.1016/0749-596X(89)90002-8 Tyler, A. & W. Nagy. (1990). Use of English derivational morphology during reading. Cognition, 36, 17–34.  doi: 10.1016/0010-0277(90)90052-L Verhoeven, L., & van Leeuwe, J. (2008). Prediction of the development of reading comprehension: A longitudinal study. Applied Cognitive Psychology, 22, 407–423.  doi: 10.1002/acp.1414 Verhoeven, L., van Leeuwe, J., & Vermeer, A. (2011). Vocabulary growth and reading development across the elementary school years. Scientific Studies of Reading, 15, 8–25. doi: 10.1080/10888438.2011.536125

256 Mercedes Spencer, Jamie M. Quinn and Richard K. Wagner

Wagner, R. K., & Meros, D. (2010). Vocabulary and reading comprehension: direct, indirect, and reciprocal influences. Focus on Exceptional Children, 43, 1–10. Wagner, R. K., Muse, A. E., & Tannenbaum, K. R. (Eds.). (2007). Vocabulary acquisition: Implications for reading comprehension. New York, NY: Guilford Press. Wagner, R. K., & Torgesen, J. K. (1987). The nature of phonological processing and its causal role in the acquisition of reading skills. Psychological Bulletin, 101, 192–212. doi: 10.1037/0033-2909.101.2.192 Wagner, R. K., Torgesen, J. K., Laughon, P., Simmons, K., & Rashotte, C. A., (1993). Development of young readers’ phonological processing abilities. Journal of Educational Psychology, 85, 1–20.  doi: 10.1037/0022-0663.85.1.83 Wagner, R. K., Torgesen, J. K., & Rashotte, C. A. (1994). The development of reading-related phonological processing abilities: New evidence of bi-directional causality from a latent variable longitudinal study. Developmental Psychology, 30, 73–87.  doi: 10.1037/0012-1649.30.1.73 Wagner, R. K., Torgesen, J. K., Rashotte, C. A., & Pearson, N. A. (2013). Comprehensive Test of Phonological Processing – Second Edition. Austin, TX: PRO-ED. Windsor, J. (1994). Children’s comprehension and production of derivational suffixes. Journal of Speech, Language, and Hearing Research, 37, 408–417.  doi: 10.1044/jshr.3702.408 Wise, J. C., Sevcil, R. A., Morris, R. D., Lovett, M. W., & Wolf, M. (2007). The relationship among receptive and expressive vocabulary, listening comprehension, pre-reading skills, word identification skills, and reading comprehension by children with reading disabilities. Journal of Speech, Language, and Hearing Research, 50, 1093–1109.  doi: 10.1044/1092-4388(2007/076)

Reading comprehension What develops and when? Kate Cain and Marcia A. Barnes University of Texas at Austin

‘What does it mean to comprehend?’ There are numerous theories of (reading) comprehension. At their core, they all seek to explain the nature of the mental representation that readers and listeners construct as they process text, and the knowledge and mechanisms that underpin this (McNamara & Magliano, 2009). The consensus is that successful comprehenders construct a coherent and integrated representation of the meaning of that text, rather than a verbatim record of its specific words, syntax, or structure. This representation is typically referred to as a mental model or situation model (Zwaan & Radvansky, 1998). It is generally assumed that the principles underpinning theories of comprehension generalise across media: written texts as well as those read aloud, and static and dynamic cartoons and film. Studies of children and adults demonstrate significant associations between the ability to comprehend narratives presented in these different media (Bishop & Adams, 1992; Gallagher et al., 2000; Gernsbacher et al., 1990; Kendeou et al., 2008) and there is also overlap with the brain regions and networks engaged when processing text by ear or by eye, indicating supramodal processing (Braze et al., 2011; Gernsbacher & Kaschak, 2003; Lindenberg & Scheef, 2007). A separate line of research, which falls under the broad umbrella of embodied cognition, shows that the processing of language about perception, action, and emotion engages regions of the brain involved in perception, action, and emotion. For example, when words related to sound, vision, taste and smell are read, those modality specific perceptual areas are activated (Glenberg et al., 2013; Pulvermüller, 2013). These different strands of research suggest that to understand reading comprehension, we need to look beyond written text and consider the language and cognitive processes that influence our ability to construct situation models. A wealth of cross-sectional and longitudinal research details the factors predictive of reading comprehension development and those that are weak in children with reading comprehension difficulties. What we lack are comprehensive accounts of how the ability to understand text develops. Critical questions that we address in our review are: What develops? and How does the development of our ability to make sense of the world around us relate to the development of reading comprehension? There are currently no adequate models of how reading comprehension develops and improves doi 10.1075/swll.15.15cai © 2017 John Benjamins Publishing Company

258 Kate Cain and Marcia A. Barnes

with age, as we discuss below. To identify the critical precursors of reading comprehension and what develops over time, we have examined the early developmental literature. Critically, because we view comprehension as a proxy for making meaning from stimuli in general, not just written stimuli, we draw on research into memory and cognitive development during early childhood, and individual differences in adults’ reading comprehension. In the next section, we consider what it means to comprehend. We then review models of reading that speak to reading comprehension development, as well as skilled comprehension, and a critique of their explanatory power within a developmental framework. In the fourth section, we explore a developmental explanation of the skills identified as central to comprehension. Finally, we conclude with suggestions for future research directions that we believe are necessary to test theories of comprehension development.

1. Comprehension by ear and by eye: What is involved? 1.1

Text comprehension: The product and the process

Comprehension can be considered “… the processing of information to extract meaning.” (McNamara & Magliano, 2009, p. 298). According to this viewpoint, the product of comprehending a text is the same regardless of modality: whether we read the text ourselves, listen to someone else read a text aloud, or view that ‘text’ as a set of pictures or an animated sequence, we construct a representation of the text’s meaning across different dimensions, such as time, space, and causality (Gernsbacher et al., 1990; Kendeou et al., 2008; Zwaan & Radvansky, 1998). Adopting this perspective, we use the term comprehension throughout this chapter when discussing general comprehension processes, only using terms such as ‘listening comprehension’ and ‘reading comprehension’ when relevant for specific research studies. The representation that skilled comprehenders produce is not a literal or verbatim record. Three strands of evidence illustrate this point. First, adults encode the meaning of sentences, rather than their form, and will therefore not remember if a sentence in a passage was in its active or passive form (Sachs, 1967). Second, adults are guided by context and background knowledge, such that ‘fried’ is as good a recall cue as ‘cooked’ for the sentence ‘She cooked the chips’ (Garnham, 1979; see also Garnham, 1981, for instantiation of nouns). Third, adults and children construct a representation that reflects the state of affairs described by the text. When presented with sentences such as “Three turtles rested on a floating log, and some fish swam beneath them.” they cannot later reliably distinguish between the original sentence and one that describes the same situation with different wording, e.g., “…. and some fish swam beneath it.” (Barnes, Huber, Johnston, & Dennis, 2007; Bransford et al., 1972). This and other evidence from cognitive psychology (see Zwaan & Radvansky, 1998, for a review) supports the view that the product of comprehension is a situation



Reading comprehension 259

model: a representation of the state of affairs described by the ‘text’ rather than a representation of the text itself. Thus, we agree with Zwaan and Radvansky (1998, p. 163) that rather than asking the question “How do readers comprehend text?” we should be asking “How do readers construct a coherent situation model?”. In Section 3, we evaluate how well current models of reading can explain when and how children develop the skills to construct accurate and coherent situation models. In terms of development, we believe that a focus either on readers or written text will not result in an adequate developmental account of comprehension, because the language skills, knowledge and processes that contribute to skilled reading comprehension also underpin the construction of situation models in the aural modality, and from non-linguistic input such as pictures and videos as well. Thus the question that we focus on is: ‘How does the ability to construct a coherent situation model develop?’. To construct a coherent situation model, comprehenders draw on a range of language skills, knowledge, and cognitive processes. Consider this brief text: “Ruby carried the glass of juice. She tripped on the step. Mum fetched the mop”. You likely found this text easy to process: you decoded the individual words and retrieved their meanings, and used your knowledge of syntax to assemble these word meanings into sensible sentences. Beyond individual word and sentence meanings, readers and listeners engage in integrative and constructive processing to extract the sense of the text (Bransford et al., 1972). They combine the meanings of different elements in the text guided, for example, by cohesive devices such as the pronoun ‘she’ in our example, which signals a reference back to a person mentioned earlier (Garnham, 2001; Graesser et al., 1994; Singer, 1994). As the text unfolds, good comprehenders monitor the consistency of the message: they are aware of the adequacy of their comprehension (Baker, 1984; Garner, 1980). In order to integrate the final sentence in our example “Mum fetched the mop.” into the situation model constructed so far, skilled comprehenders recognise the need to generate an inference to make sense of Mum’s actions and to ensure that the situation model is coherent. The degree to which monitoring for meaning and inference making is relatively automatic and effort-free may depend on the information to be integrated, the nature of the task, and reader characteristics, such as skill level and age. When text is considerate (i.e. cohesive and coherent), word-by-word and sentence-by-sentence integration may be considered as ‘fine tuning’ of the situation model (Perfetti & Stafura, 2015; see also Stafura and Perfetti this volume), something that can happen with relatively little cognitive effort in contrast to situations where information is only weakly associated and strategic inferences must be made to ensure coherence (see also Cook & O’Brien, 2015). Likewise, evaluating incoming information in a text is a routine and non-strategic aspect of skilled comprehension that is less effortful when the new information to be validated shares a close relationship with both the earlier text and reader knowledge (Cook & O’Brien, 2014; Isberner & Richter, 2014). Inference making may require the comprehender to go outside of the text, drawing on background knowledge to make sense of unspecified details (Graesser et al., 1994). In our example, knowledge about the likely consequence of tripping while carrying a

260 Kate Cain and Marcia A. Barnes

glass of juice – spilling the drink – enables us to make sense of the information and to generate a motive for Mum’s action. Knowledge differences impact on comprehension: domain-specific knowledge can compensate for low verbal ability, such that experts in a subject area demonstrate superior memory and comprehension of texts than novices, even when novices have superior verbal abilities (Ericsson & Kintsch, 1995; Schneider et al., 1989). Modelling of the component skills of reading comprehension shows that knowledge affects reading comprehension directly, but also indirectly through inference making (Cromley & Azevedo, 2009). Interpretation of a text can also be guided by knowledge of text structure (Zwaan, 1994). Comprehenders know what to expect from narratives because they contain conventional features and elements (Stein & Glenn, 1982), and this knowledge can help them to link together information in the situation model. In sum, to adequately comprehend beyond the word and sentence level when reading or listening, comprehenders integrate information between the elements within a text and they generate knowledge-based inferences as and when necessary to ensure coherence (see also van den Broek and Kendeou, this volume for a discussion of coherence). They are also conscious of the adequacy of their understanding because they monitor for consistency, and they can use their knowledge of text structure as a framework to guide their understanding. These processes are essential for adequate text comprehension. The process of constructing the situation model happens in real time. Comprehenders constantly revise or update the content of the situation model, integrating each new word and phrase into the situation model. When additional information disambiguates a character or supports a particular interpretation, the situation model is updated to reflect that (Zwaan & Madden, 2004). Thus, an incorrect inference will be discarded and one that is more appropriate to the context will be encoded (Lorsbach et al., 1998; Perez et al., 2015). The situation model also serves as the context for interpreting specific words and phrases. This occurs when we instantiate general verbs and nouns when the text supports a more specific interpretation (Garnham, 1979, 1981). Context also supports the selection of the most appropriate meanings of words with multiple meanings, such as ‘date’, ‘bank’, or ‘spade’, and the interpretation of non-literal, that is figurative, expressions such as ‘they were in the same boat’ (Barnes et al., 2004; Cain et al., 2009; Gernsbacher & Faust, 1991). Thus, when we consider text comprehension and what develops, we need to consider the construction of the situation model in real time.

1.2

The longitudinal prediction of comprehension development: The story so far

Cross-sectional studies and longitudinal studies demonstrate that a range of language skills concerned with meaning predict reading and listening comprehension. For 7- to 9-year-olds, vocabulary, grammar, and the higher-level skills of inference making and comprehension monitoring each explain unique variance in concurrent



Reading comprehension 261

reading comprehension outcomes after controlling for word reading ability (Kim, 2015; LARRC & Logan, in press). Longitudinally, between 7 to 11 years, vocabulary knowledge, inference making, comprehension monitoring, and knowledge of text structure are each unique predictors of reading comprehension over and above word reading ability (Oakhill & Cain, 2012). When we consider development in pre-readers or children in the first year or two of reading instruction, we find that a similar range of language and cognitive skills are associated with comprehension level (Kendeou et al., 2009; Lepola et al., 2012; Lynch et al., 2008; Muter et al., 2004; Silva & Cain, 2015). These studies are limited in their account of the development of reading comprehension because they identify the component skills that are associated with comprehension at a given point in time, which makes them essentially static in nature; they have not examined how those component skills themselves develop and how the changing nature of these skills affects reading comprehension and vice versa. This is a critical issue in detailing a developmental theory, where we need to specify what develops and whether there are qualitative or quantitative shifts across development. For example, there is evidence that language skills and knowledge bases interact across the course of development: early vocabulary predicts subsequent reading comprehension which, in turn, predicts subsequent vocabulary knowledge (Verhoeven & van Leeuwe, 2008). In a similar way, reading comprehension at 7 years predicts subsequent inference making which then predicts later reading comprehension at 11 (Oakhill & Cain, 2012). Thus, when we seek to identify what develops and the specific drivers of change, we need to consider also the dynamics across time – the interactions between skills.

2. Theoretical models of reading comprehension: A critical review 2.1

Language-based accounts of reading comprehension: How well do they explain development?

Both the simple view of reading (Gough et al., 1996; Gough & Tunmer, 1986) and the verbal efficiency hypothesis (Perfetti, 1985, 1988) stipulate that as readers develop faster and more efficient word reading, they have greater cognitive resources to devote to deeper comprehension processing. Thus, an individual’s listening comprehension skills become an increasingly important predictor of reading comprehension relative to their word reading ability during the first few years of reading instruction. There is a wealth of support for this sketch of reading development: the strength of the relation between word reading and reading comprehension lessens and, conversely, the strength of the relation between listening comprehension and reading comprehension increases (Catts et al., 2005; de Jong & van der Leij, 2002; Garcia & Cain, 2014; LARRC, 2015). Perfetti’s lexical quality hypothesis (LQH) (Perfetti, 2007) is another wordbased account that speaks to sentence- and passage-level comprehension. The LQH builds on verbal efficiency by emphasising the importance of knowledge. The basic

262 Kate Cain and Marcia A. Barnes

premise is that the quality of word-level representations of form (at the phonological and orthographic levels) and meaning can explain individual differences in reading comprehension (Perfetti & Hart, 2001). An individual with precise and well-specified representations of words will have better comprehension because efficient access to stable representations frees up cognitive resources for comprehension, enables lowcost integrative processing, and results in less interference from form or meaning competitors. In support of this viewpoint, concurrent performance on indicators of lexical quality (speed and accuracy of access to semantic, phonological and orthographic representations of words) is related to reading comprehension skill in children and adults (Perfetti & Hart, 2001; Richter et al., 2013). Further, differences in reading comprehension skill between children in consecutive grades are related to increasing efficiency of access to these representations with age (Richter et al., 2013). The LQH is not a developmental theory, but it could be integrated into an explanation of comprehension development. First, as Richter and colleagues have shown in their cross-sectional study of young readers, there is a stronger association between reading comprehension and performance on measures of lexical quality with increasing grade (Richter et al., 2013). Clearly, knowledge about words and, therefore, the quality of their representations will be enhanced by meaningful exposure to language. Thus, we would expect that, over time, the strengthening phonological and orthographic representations, as well as the additional links between semantic associates, will lead to more efficient integration and inference making, core skills involved in the construction of a situation model. For skilled adult readers, there is good evidence that readily accessible vocabulary and world knowledge results in relatively cost-free inference making and integration (Cook & O’Brien, 2015; Keenan et al., 1984; Perfetti & Stafura, 2015). In contrast, less skilled comprehenders do not obtain the same processing advantage for more vs. less causally related sentences (Barnes et al., 2015). From an individual differences perspective, a lexical quality explanation for differences in text comprehension might be that the knowledge base is set up differently – more deeply or better connected in some children than others, to support precise but flexible retrieval of word meanings. The outcome would be better text comprehension skills. The LQH provides an explanation for improvements in both the process and the product of reading comprehension over time. However, its focus on the written word limits its application to more general processes of comprehension that have been demonstrated by the strong associations between comprehension of text and other media (Gernsbacher et al., 1990; Magliano et al., 2001; Kendeou et al., 2008; Kendeou et al., 2005; Paris & Paris, 2003). Further, children with specific reading comprehension difficulties have parallel problems with listening comprehension: their understanding of passages read aloud is weak relative to same-age peers (Cain et al., 2000). 1 These 1. Perfetti notes that lexical quality can be applied to spoken language, with a focus on phonological form and meaning (Perfetti, 2007, p. 361), but this account has not been developed to date.



Reading comprehension 263

considerations limit the potential of the LQH with its current focus on the written word to explain comprehension development in general. To understand the development of comprehension more generally, across different modalities, we need an account that considers the quality of representations of general knowledge, as well as knowledge about individual word meanings, syntactic structures, etc., and one that does not start with the written word. We focus on the development of knowledge and the factors that must be considered to specify a framework in which to understand the development of language comprehension in Section 4.

2.2

Cognitive resources: The role of executive function and working memory

Comprehension is a dynamic process: the situation model is constructed in real time, moment-by-moment, as discourse or text unfolds. For that reason, the cognitive resources of working memory and executive function have featured prominently in research on comprehension, most notably in accounts of individual differences in children’s and adults’ reading and listening comprehension (Daneman & Merikle, 1996; Radvansky & Copeland, 2001). Whilst there is evidence that good language skills and coherent supportive texts make integrative and constructive processing less resource demanding (Cook & O’Brien, 2014, 2015; Perfetti & Stafura, 2015), there is also evidence that working memory and executive function are predictive of reading comprehension concurrently in both children and adults (Cain, Oakhill, & Bryant, 2004; Daneman & Merikle, 1996; Sesma et al., 2009). Significantly, early measures of working memory and executive function are predictive of later text comprehension and related skills, such as inference making, that are critical to the construction of the situation model, an influence that is apparent from preschool through to 9 years (Pike et al., 2013). The influence of executive function skills is specific to comprehension, not reading in general: executive function at 4 years predicts oral language comprehension, but not letter-word identification, two years later (Fuhs et al., 2014). These studies suggest that cognitive resources both support and drive the foundational language skills that enable good listening and reading comprehension. Thus, we need to understand the role of cognitive resources in relation to comprehension outcomes, and how these influence comprehension development. The dominant view has been that working memory capacity constrains language comprehension, because fixed capacity processing resources will limit the integrative and constructive processes necessary to construct a situation model (Just & Carpenter, 1992). This view underpins verbal efficiency and the LQH and is implicit in the simple view of reading as well. In support of it, we find that independent measures of working memory are related to adults’ sentence processing (Just & Carpenter, 1992) and children’s reading comprehension and inference making (Cain, Oakhill, & Bryant, 2004; Pike et al., 2010; Seigneuric et al., 2000). Critically, however, several studies have demonstrated that, when pitted against language processing skills that draw heavily on cognitive processing resources, it is these broader language skills that predict unique

264 Kate Cain and Marcia A. Barnes

variance in reading comprehension over and above working memory (Cain, Oakhill, & Bryant, 2004; Currie & Cain, 2015; LARRC, Jiang, & Farquharson, under review). A different set of explanations for the relation between cognitive resources and reading comprehension focus on memory mechanisms and the quality of information activated and processed in working memory, rather than the capacity of working memory per se (Carretti et al., 2005; Van Dyke & Johns, 2012). For example, the specificity or distinctness of retrieval cues in the text, rather than the distance between individual text elements, can account for why adults find some sentences more difficult to process than others (Van Dyke & Johns, 2012). Such findings have been related to lexical quality: robust and precise lexical representations will enable efficient and accurate retrieval of meaning during language processing and limit interference from competitors (Van Dyke & Shankweiler, 2013). A developmental account might posit that interference would be reduced with meaningful exposure to language, as knowledge representations become more precise and robust over time. In support of this, younger readers experience greater interference than older readers when pictures are inconsistent with the inferences supported by text (Pike et al., 2010). These studies address a fundamental question that has not been answered to date: is a particular level of working memory skill necessary to enable the processing of language or do good language skills free up the necessary working memory resources for processing language? Longitudinal studies are required to address this point in relation to comprehension development and recent work supports the notion that language supports growth of these cognitive resources: Early communicative gestures at 15 months of age predict expressive language skills at 2 and 3 years (including vocalisation to request toys and understanding of language concepts), and these early language abilities, in turn, predict the development of later executive abilities, such as working memory, inhibitory control, and attention shifting (Kuhn et al., 2014). Thus, language may be a precursor for the development of executive abilities such as working memory. Findings such as these suggest that the influence of early language on later comprehension may be both direct and also indirect; that is, through language’s influence on the development of executive abilities such as working memory. When we consider developmental models of comprehension and how best to foster comprehension skills, we need to not only consider the nature of the relation between cognitive resources and language comprehension over time, but also whether some relations are developmentally limited. For example, working memory resources might be less predictive of performance in the early stages of comprehension when processing of written words, vocabulary, and syntax is less fluent; its influence may increase with age, as language experience develops knowledge structures. Recent empirical work supports this point: in a study of children in grades 1 through 3, working memory explained a greater proportion of variance in comprehension in the older sample (LARRC et al., under review). Other work points to the need for a more detailed analysis of what skills our measures of working memory and executive skills are tapping and how they relate to comprehension processing. For example, the memory task used by Pike et al. (2013) found to be predictive of reading comprehension longitudinally



Reading comprehension 265

from preschool through to 9 years also tapped executive processes such as planning and inhibitory control, which have been implicated in the breakdown of reading comprehension (Borella et al., 2010; Locascio et al., 2010). Indeed, future studies need to pinpoint which aspects of executive function, of which working memory is one component, are fundamental to comprehension processing.

2.3

Beyond language and cognitive skills: The role of motivation

A different approach that extends beyond the language and cognitive bases of reading comprehension also merits mention; a component account of reading development (Joshi & Aaron, 2000). This framework includes ecological (such as home language) and motivational factors. Motivation is a unique predictor of children’s text comprehension: it explains variance in reading comprehension over and above initial reading comprehension level and exposure to print (Guthrie et al., 1999; Wang & Guthrie, 2004). Motivation may influence comprehension because it taps into an individual’s perseverance on task, similar to the personality trait grit (Duckworth et al., 2007). In older children with long histories of reading difficulties, lower feelings of self-efficacy and control predict reading comprehension even when reading is valued (Wolters et al., 2014). This is important because, within the domain of motivational constructs, it is lower feelings of perceived competence that have been most strongly related to reduced task engagement and persistence (Zimmerman & Schunk, 2006). Although the component account is not a model of development, we believe that future research on development must consider motivational factors to better understand comprehension development, and that some of these factors may come into play at different points in reading development. In sum, different accounts of reading provide a description of reading comprehension differences across development, but they are largely silent about development. What changes across time and what is constant? What are the mechanisms by which children develop the skills to construct more coherent and more accurate situation models? We should note that many of these models and frameworks were not set up to describe the development of either reading comprehension or the situation model more specifically. But we need to ask why all of these are inadequate as descriptions of comprehension development. One reason may be that they start with the end point – skilled comprehension of adults – and work backwards. Here we might learn a useful lesson from the neuropsychology literature, where it has been eloquently argued that that the adult brain model for acquired skills is not a good model to apply to skill development or difficulties (Bishop, 1997; Karmiloff-Smith, 2009).

266 Kate Cain and Marcia A. Barnes

3. Returning to the situation model: What’s in a situation model and how does that develop? The comprehension models discussed above accommodate aspects of the acquisition of comprehension skills, but they are inadequate accounts of the development of what is needed to construct the situation model. Zwaan and Radvansky (1998) propose that comprehenders construct situation models that contain the following dimensions: spatial, temporal, causal, protagonist/object, and goals or intentionality. Whether all of these dimensions are necessary or separate is questionable (e.g., Therriault, Rinck, & Zwaan, 2006); however, situation models at least capture causal relations between events, objects, and characters (van den Broek et al., 2005). What is known about the development of children’s understanding of these causal relations and the application of this knowledge during comprehension?

3.1

Individual dimensions in situation models: Causal, temporal and spatial relations

The ability to make simple causal inferences about object-based events develops in the first 6 months of life and is linked to early sensory and motor experiences (Cohen, Chaput, & Cashon, 2002). However, when cognitive load is increased by using complex real world stimuli this event-based causal reasoning is not observed for several more months (Cohen et al., 2002). Although simple causal situation models develop early in life, we are most interested in whether and how features of situation models are constructed by young children when events are described (or pictured) rather than physically experienced. In studies with 3- to 5-year-olds, Fecica and O’Neill (2010) investigated the processing of physical motion events during narrative comprehension. Children pressed a button to listen to succeeding sentences. In one study, children were told that a character was either driving or walking to his/her aunt’s house. For the portions of text describing what the character saw while driving or walking, sentence processing times were faster for the “driving” versus the “walking” condition, with no effects of age. Young children with causal knowledge of movement events appear to use that information by adopting the perspective of the protagonist. Another study showed age invariance from 8 to 14 years in the ability to construct temporal-spatial aspects of situation models during reading (Barnes et al., 2014). After learning about the spatial layout of a market and objects outside each stall, children read a story about a protagonist who moved from one explicitly mentioned market stall (source) to another (goal), passing outside an unmentioned stall (path). Periodically, children had to judge whether two objects (words) were from the same stall or a different stall. Children’s responses were consistent with them tracking the perspective of the protagonist in space and time. They were more accurate and faster on the object judgement task when the objects were from the goal stall or the unmentioned path stall



Reading comprehension 267

than when the objects were from a stall that the character had not traversed. Accuracy in judging that two objects were from the unmentioned pathway uniquely predicted reading comprehension. Although these studies suggest that children use spatial-temporal information to construct situation models during comprehension, there are developmental constraints on these processes that arise from how events are represented linguistically. These processing constraints are revealed by studies that investigated developmental changes in the construction of spatial (Uttal, Fisher, & Taylor, 2006) and temporal (Pyykkönen & Järvikivi, 2012) aspects of situation models. Spatial relations are typically experienced in a holistic fashion (e.g., scenes). In contrast, spatial relation information in language is coded sequentially, requiring that serial verbal descriptions be integrated into a coherent spatial representation. Uttal et al. (2006) found that in comparison to 10-year-olds and adults, the mental models of 8-year-old children more often preserved the sequential nature of the linguistic descriptions resulting in less-well integrated spatial representations that led to difficulties in making accurate spatial inferences. These younger children were aided by graphic representations suggesting that it is the sequential nature of language that affected their construction of the spatial situation model. A similar study comparing skilled and less skilled comprehenders showed that the latter had difficulty making inferences from texts in which spatial relation information was presented sequentially (Barnes, Huber, Johnston, & Dennis, 2007). In sum, younger children and children with comprehension difficulties are less likely than older children and more skilled comprehenders to construct spatial situation models from linguistic input, which requires the integration of sequentially presented information. Spatial-temporal events in the stories used by Fecica and O’Neill (2010) and Barnes et al. (2014) occurred in chronological order, but language does not always express events in chronological order. For example, compare the sentence, “Lauren went to the store before she played with her cat” to one in which language codes events in reverse order, “Lauren went to the store after she played with her cat”. Adults do not differ in their comprehension of these two sentences, but children between 8–12 years of age are more accurate in their interpretation of sentences in which the language used matches event-based chronological order (Pyykkönen & Järvikivi, 2012). Similar findings were obtained in a study with 3- to 7-year-olds who were shown an animation of each clause in the sentence and asked to touch the thing on the screen ‘that happened first’ (Blything, Davies, & Cain, 2015). In that study, a memory measure predicted variance in performance, which is compatible with the idea that cognitive load increases when language codes events out of order. Further support for this hypothesis comes from Pyykkönen and Järvikivi (2012) who reported that the advantage for sentences containing before vs. after is not found when these prepositions occur at the beginning of the sentence (e.g., After Lauren went to the store, she played with her kitten), suggesting that the processing cost associated with coding events out of order is related to having to revise the situation model.

268 Kate Cain and Marcia A. Barnes

These findings from studies on spatial and temporal situation model construction reveal two inter-related points about the development of children’s situation models. First, when there is a conflict between how events are typically experienced in the real world and the language that is used to code those events, children’s situation model construction will be affected. Second, these negative effects on children’s situation model construction may be related to the need for cognitively demanding revision and integration processes engendered by reverse temporal coding and sequential descriptions of complex spatial relations between objects and characters. Resource limitations are not specific to children and also affect comprehension in adults (Munte et al., 1998). However, because executive functions, like working memory, show developmental change from early childhood through adolescence, they could account for some of these age-related findings in situation model construction.

3.2

Individual dimensions in situation models: Intentions and motivations

In another of the studies conducted by Fecica and O’Neill (2010) 4- and 5-year-olds listened to stories in which the child was primed about how “eager” a character was to go to a particular destination. For the part of the story where the character was described as getting ready to go, sentence processing times were faster when children knew the character was eager (vs. not eager) to go to the destination, suggesting that young children adopted the motivational perspective of the protagonist during story comprehension. These findings are compatible with theories of social inference (e.g., Gilbert, Pelham, & Krull, 1988) proposing fast acting, relatively “automatic” processing of social information. In adults, this type of processing is followed by slower acting, more effortful processing when situation-based information requires one to revise the interpretation that was derived from initial “default” processing (Gilbert et al., 1988). To the extent that more effortful processing requires cognitive resources, these findings for adults support the idea that during development, processing resources may constrain the construction of accurate situation models in cognitively demanding situations. In a cross-sectional study, Haga, Garcia-Marquez, and Olson (2014) found age effects when situation-based correction procedures needed to be applied to accurately interpret a character’s actions. In contrast to older children and adults, kindergarten children thought that a child who was behaving in a sad way was sad by disposition regardless of whether they received additional information that the situation was one that typically produced feelings of sadness or happiness. In another study (Haga et al., 2014), children from kindergarten to the 6th grade had more difficulty than 9th graders and college students in a cognitively demanding task requiring calibration of the emotional valence of a situation while at the same time taking dispositional information into account. Here they had to rate the sadness of a question being asked of a child given prior information about the child’s disposition as someone who often cries or who is a cheerful person.



Reading comprehension 269

In sum, these studies suggest that by 4–5 years of age children can quickly form and integrate perspective taking inferences in spoken language, such as identifying what the character’s motivational perspective is, to make inferences about their behavior. However, similar to the findings for spatial and temporal aspects of situation model construction, these character-based inferences are less likely to be made in situations with high cognitive demands such as when conflicting situation-based information is presented that requires the child to revise their initial interpretation of a character’s actions (also see review in San Juan, Khu, & Graham, 2015).

3.3

Pulling it together: Encoding several dimensions in a situation model

Thus far, we have considered developmental aspects of the construction of individual features of situation models. Studies of adults suggest that they construct situation models that contain all or most of the dimensions identified by the event-indexing theory (Zwaan & Radvansky, 1998). To our knowledge, only two studies address children’s simultaneous construction of multiple features of the situation model. In their third study with preschool children, Fecica and O’Neill (2010) combined the motivational manipulation “eager vs. non-eager” with the physical movement manipulation “driving vs. walking”. The protagonist’s motivational set affected children’s sentence processing times for the walking scenario where speed of walking is under the character’s control, but not for the driving scenario where the protagonist has no control over how fast the car goes. These findings suggest that during narrative comprehension, pre-schoolers simultaneously include in their situation models causal information about character goals and motivations and physical events. However, there is also evidence that school-age children are less likely than adults to include and monitor all aspects of the situation model simultaneously. A study of 9- to 12-year-olds found that causal information, and particularly emotionally charged character-based information, was more likely than spatial and temporal information to be coded in the situation model for narratives (Wassenburg, Becker, van den Broek, & van der Schoot, 2015). A study comparing 12-year-olds and adults also suggests that children routinely monitor causality, but not other dimensions. In this study, children and adults read stories sentence by sentence: both groups showed the predicted reading time increase when there was a shift in the dimension of causality, but only adults also showed the same behaviour when there was a temporal or spatial shift (Bohn-Gettler et al., 2011). Off-line measures indicated that the children had a relatively sophisticated appreciation of these different dimensions, but they were less likely to process shifts in these dimensions during on-line reading. Of note, these studies (like many others in the field) have focused on narrative texts. Whether spatial and temporal features are routinely included in children’s situation models for social studies and science texts, in which temporal and spatial dimensions may be more important than they are in narratives, is unknown.

270 Kate Cain and Marcia A. Barnes

In sum, the research on developmental aspects of situation model construction suggests that children use their knowledge about objects, people, and events to construct situation models during comprehending. Children can fail to construct situation models that are as accurate and comprehensive as those constructed by adults: (1) when there is a lack of knowledge, and (2) when there is knowledge, but it is not applied during comprehension. One hypothesis about the lack of application of knowledge comes from the findings above that causal knowledge about events, objects, and character motivation and intention is less likely to be used in situation model construction under conditions of high cognitive load. High cognitive load could be related to stimulus complexity and the need to ignore task-irrelevant information, from the need to revise previously constructed representations, and from the need to integrate multiple sources of information. In terms of reading comprehension, there may also be another source of developmental differences. The ability to automatically associate printed words with their real world referents appears to have a fairly protracted developmental course. In a cross-sectional fMRI study comparing neural activation to pictures and words, Dekker, Mareschal, Johnson and Sereno (2014) showed that 7- to 10-year-olds, like adults, activated regions of brain functionally associated with pictures of animals and tools. In contrast, the written words for these objects activated similar regions of brain as did the pictures, but only in adults. Consistent with the LQH, these findings suggest that it may take considerable exposure to print to develop high quality lexical representations in which orthography, phonology and semantics are tightly integrated in ways that simulate real world events through reading. The findings are also relevant for embodied cognition approaches to comprehension. For example, work by Glenberg and colleagues shows that comprehension in beginning readers is enhanced by manipulating toy objects while reading sentences. Improvements in comprehension are thought to arise because the activity encourages children to map words to the objects that they represent, thereby strengthening the links between phonology, orthography and the real world situation described by the text (Glenberg, Gutierrez, Levin, Japuntich, & Kaschak, 2004). Consistent with dual process models of cognition (reviewed in Evans & Stanovich, 2013), situation models appear to be constructed fairly routinely and at young ages when children have the domain knowledge. However, for the reasons discussed above, accurate situation model construction can also require more effortful processing that relies on limited cognitive resources, such as working memory. Furthermore, it may take considerable experience with orthographic-phonological representations for readers to efficiently and consistently activate situation-based representations during reading. We suggest that these processing constraints have implications for the development of situation model construction during comprehension.



Reading comprehension 271

4. Conclusions The questions posed in the title of this chapter on reading comprehension development was “what develops and when?” In this section, we summarise our findings in relation to this question and make suggestions for future research directions to inform theories of comprehension development.

4.1

Models of reading comprehension: What do they tell us about what we need in a model of development?

Our review shows that there are many adequate models describing key skills and processes involved in reading comprehension, but no current models of reading comprehension development. We believe that this is (partly) because we have been looking in the wrong place, misled by our focus on adult models of skilled comprehension and the end-point of development. Theories of reading comprehension or reading ability more broadly have not focused on the development of understanding; instead they have focused on the printed word (Gough et al., 1996; Perfetti, 2007), the general processes involved in constructing the product of skilled comprehension – the situation model (Gernsbacher, 1990; Kintsch, 1998; van den Broek et al., 2005), or have provided detailed accounts of how we retrieve and use general knowledge and generate inferences in the process of constructing the situation model (Graesser et al., 1994; McKoon & Ratcliff, 1992; Myers & O’Brien, 1998; Zwaan & Radvansky, 1998). We are not suggesting that these accounts are at fault; they were developed as frameworks to describe and understand reading ability or comprehension, not their development. In addition, studies of children’s comprehension to date cannot provide an adequate answer to our questions because they have largely focused on the source of difficulties experienced by select groups of poor readers (see Oakhill & Cain, this volume) or have only considered development with ‘static’ longitudinal models (Kendeou et al., 2009; Muter et al., 2004; Oakhill & Cain, 2012), rather than examined what develops and when. These varied strands of research have provided a wealth of information that we must now use to build developmental theories. For adults, we have a clear understanding of the nature of comprehension: what is encoded and what is not encoded in the situation model (Zwaan & Radvansky, 1998); and for children (and adults) we have a wealth of knowledge about the critical skills that aid the construction of the situation model and which differentiate reading comprehension success and failure, such as inference making and working memory (Cain & Oakhill, 1999; Daneman & Merikle, 1996). Further, when we examine competence in infancy and preschool, there is clear evidence that children are attuned to the dimensions encoded in a situation model and draw on their prior knowledge and experience to make sense of events from an early age. What is not clear is how the ability to monitor and encode accurately these different dimensions develops or, alternatively, what constrains comprehension in the early years. Future research needs to address these issues to build a model of comprehension development.

272 Kate Cain and Marcia A. Barnes

4.2 Building a model of reading comprehension development: The role of knowledge and cognitive resources Clearly knowledge will be central to any developmental account of comprehension. We acquire knowledge across the lifespan, through interactions with the world around us, schooling, and less formal means of acquisition, such as access to various media. This viewpoint accords with the strong relationship between vocabulary knowledge and reading comprehension (Carroll, 1993). However, our review has also demonstrated that basic knowledge about the dimensions encoded in a situation model, critical to understanding both narrative and also expository texts, develops at a young age (see also Case et al., 1996, for evidence of early conceptual knowledge) and we know that stored knowledge is not always used in the service of comprehension (Barnes et al., 1996; Bohn-Gettler et al., 2011). Thus, although knowledge must play a key role, change in how much we know about our language and the world around us provides an inadequate vehicle for developmental gains in reading comprehension. One alternative that has gained ground when considering the relationship between reading comprehension, word reading, and vocabulary is that we need to consider depth of knowledge (Ouellette, 2006; Tannenbaum et al., 2006). In relation to the situation model, richer and better connected semantic networks may facilitate inference making and the coherence of situation models because critical concepts will be more accessible when encountered or more easily inferred and integrated (Cook & O’Brien, 2015). Further, accessibility of knowledge plays a role in explaining developmental gains in inference making (Barnes et al., 1996). When we consider development, there is direct evidence that experience with the world and exposure to information may change the nature, as well as the strength, of links between lexical items: for example, functional associates, many of which will be learned through observation and experience, are more strongly associated than category associates in young children (Blewitt & Toppino, 1991). Thus, our knowledge structures will change across the lifespan and this may, in part, explain improvements in reading (and other forms of) comprehension. Future research needs to explore better how our semantic representations and knowledge structures develop and how their organisation impacts reading comprehension. As noted, knowledge may be present but not routinely applied in the process of comprehending a text, particularly when the cognitive load is high. When we consider developmental gains in comprehension, it is clear how cognitive resources may play a role in limiting younger children’s performance. For example, when compared with older individuals on the same texts, those texts will be more complex to younger children, access to relevant information may be more effortful, and the ability to ignore task-irrelevant information and to revise and update the situation model will draw on executive functions that are still developing. Thus, we need to understand the interplay between knowledge, language, and cognitive resources to understand what constrains comprehension at different points in development. Some recent correlational studies suggest that the relation between text comprehension and a cognitive resource, such as working memory, is indirect, mediated by vocabulary knowledge (Currie & Cain,



Reading comprehension 273

2015; Van Dyke et al., 2014). But other work with younger children suggests a primary influence of memory in constructing situation models (Blything et al., 2015). Longitudinal studies that map the development and interaction between the skills over time is required to understand how cognitive resources both support and drive the foundational language skills that enable good listening and reading comprehension. Critical to understanding the interplay between text, cognitive load, and the reader’s own knowledge and processing resources, are studies that examine the moment-by-moment processing of text to understand better the dynamics of text processing in real time. We know that adults (and children) take longer to process information that requires a shift in focus (or dimension) or an inference to be made (Bohn-Gettler et al., 2011; Zwaan et al., 1995) and that processing signals in text can facilitate the integrative processes involved in the construction of a situation model (Cain & Nash, 2011). Further, the study of the real time processing of language by very young children can provide insights into their comprehension and simulations of a narrative (Fecica & O’Neill, 2010). Future studies that measure moment-by-moment processing of text across development are needed to understand the continuities and discontinuities across development.

4.3

Building a model of reading comprehension development: Where does knowledge come from?

When thinking about the development of knowledge, and how the ability to construct meaning-based representations of discourse and text develops, we need to return to the definition of a situation model: the actual situation described by the text. If we think about development and we do not focus only on reading comprehension, we can gain from understanding cognitive development more broadly; specifically, children’s development of their understanding of the world and how constructing the situation model means to map text onto a real world situation. This brings us to an account of cognition that we believe can usefully inform a future account of comprehension development: embodied cognition (Glenberg et al., 2013; Zwaan, 2014). Embodied cognition is a current ‘hot topic’ in the fields of psychology and neuroscience. A basic premise is that mental representations of a concept, object, or event are often grounded in perception and action, resulting in activation of perceptual, somatosensory, and motor areas of the brain during cognitive processing (Barsalou, 2010; Glenberg et al., 2013; Pulvermüller, 2013). In this view, cognition, and therefore language comprehension, involves the mental simulation of events (Barsalou, 1999). Critically, in terms of development, there is a key role for action experience in learning, which goes back to classical theory of psychological development (Piaget, 1952). Certainly for adults, there is growing evidence that cognition is in some form embodied 2 and we speculate that this account might be a useful bridge between infants’ experiences with their environment, through direct action and observation, and an 2. Distinctions between the different theoretical accounts are not relevant to our discussion here.

274 Kate Cain and Marcia A. Barnes

understanding of the dimensions we encode in a situation model, such as time, space, and causality. A limitation in this approach is that accounts of embodied cognition and language comprehension to date have focused on words and sentences at the neglect of discourse comprehension (Fischer & Zwaan, 2008; Zwaan, 2014). This clearly limits the extent to which we can apply this account to a developmental account of how children learn to construct situation models beyond the single word or sentence. Further, we must not neglect the importance of our experience with text itself in reading comprehension: The vocabulary, syntax, and text structures used in written text are less frequent and more complex than those used in spoken conversation (Hayes & Ahrens, 1988; Scott, 2009) and text is an important means of knowledge acquisition across the lifespan (Stanovich, 1993). Despite these limitations, we believe that this approach could be usefully applied to understand better the early development of the proposed dimensions encoded in situation models of what we see, hear, and read, and how they provide a foundation for later reading comprehension.

4.4 Comprehension development: A summary We have not outlined a model of reading comprehension development that enables us to specify what should be taught and at what age. Rather, we have shown that, in many ways, children are similar to adults in their appreciation of the critical aspects of text comprehension and that, even in infancy, children are constructing situation models and making inferences about the world around them. In our view, it seems that the foundations of reading comprehension and the ability to construct a situation model develop through experience with language and the world around us from very early in development, well before formal literacy instruction.

Acknowledgements Marcia Barnes’ contribution was supported by the Institute of Education Sciences, U.S. Department of Education, through Grant R305F100013 to the University of Texas-Austin as part of the Reading for Understanding Research Initiative and by Award Number P50 HD052117, Texas Center for Learning Disabilities, from the Eunice Kennedy Shriver National Institute of Child Health and Human Development to the University of Houston. The opinions expressed are those of the authors and do not necessarily represent views of the Institute or the U.S. Department of Education or official views of the Eunice Kennedy Shriver National Institute of Child Health and Human Development or the National Institutes of Health.



Reading comprehension 275

References Baker, L. (1984). Children’s effective use of multiple standards for evaluating their comprehension. Journal of Educational Psychology, 76, 588–597.  doi: 10.1037/0022-0663.76.4.588 Barnes, M. A., Ahmed, Y., Barth, A., & Francis, D. J. (2015). The Relation of Knowledge-Text Integration Processes and Reading Comprehension in 7th-to 12th-Grade Students. Scientific Studies of Reading, 19, 253–272.  doi: 10.1080/10888438.2015.1022650 Barnes, M. A., Dennis, M., & Haefele-Kalvaitis, J. (1996). The effects of knowledge availability and knowledge accessibility on coherence and elaborative inferencing in children from six to fifteen years of age. Journal of Experimental Child Psychology, 61, 216–241. doi: 10.1006/jecp.1996.0015 Barnes, M. A., Faulkner, H., Wilkinson, M., & Dennis, M. (2004). Meaning construction and integration in children with hydrocephalus. Brain and Language, 89, 47–56. doi: 10.1016/S0093-934X(03)00295-5 Barnes, M. A., Huber, J., Johnston, A. M., & Dennis, M. (2007). A model of comprehension in spina bifida meningomyelocele: Meaning activation, integration, and revision. Journal of the International Neuropsychological Society, 13, 854–864.  doi: 10.1017/S1355617707071172 Barnes, M. A., Raghubar, K. P., Faulkner, H., & Denton, C. A. (2014). The construction of visual – spatial situation models in children’s reading and their relation to reading comprehension. Journal of experimental child psychology, 119, 101–111.  doi: 10.1016/j.jecp.2013.10.011 Barsalou, L. W. (1999). Perceptual symbol systems. Behavoiral and Brain Sciences, 22, 577–660. Barsalou, L. W. (2010). Grounded cognition: past, present, and future. Topics in Cognitive Science, 2, 716–724.  doi: 10.1111/j.1756-8765.2010.01115.x Bishop, D. V. M. (1997). Cognitive neuropsychology and developmental disorders: Uncomfortable bedfellows. The Quarterly Journal of Experimental Psychology: Section A, 50, 899–923. doi: 10.1080/713755740 Bishop, D. V. M., & Adams, C. (1992). Comprehension problems in children with specific language impairment: Literal and inferential meaning. Journal of Speech and Hearing Research, 35, 119–129.  doi: 10.1044/jshr.3501.119 Blewitt, P., & Toppino, T. C. (1991). The development of taxonomic structure in lexical memory. Journal of Experimental Child Psychology, 51, 296–319.  doi: 10.1016/0022-0965(91)90038-T Blything, L., Davies, R., & Cain, K. (2015). Young children’s comprehension of temporal relations in complex sentences: the influence of memory on performance. Child Development, 86, 1922–1934.  doi: 10.1111/cdev.12412 Bohn-Gettler, C. M., Rapp, D. N., van den Broek, P., Kendeou, P., & White, M. J. (2011). Adults’ and children’s monitoring of story events in the service of comprehension. Memory and Cognition, 39, 992–1011.  doi: 10.3758/s13421-011-0085-0 Borella, E., Carretti, B., & Pelegrina, S. (2010). The specific role of inhibition in reading comprehension in good and poor comprehenders. Journal of Learning Disabilities, 43, 541–552 doi: 10.1177/0022219410371676 Bransford, J. D., Barclay, J. R., & Franks, J. J. (1972). Sentence memory; a constructive versus interpretive approach. Cognitive Psychology, 3, 193–209.  doi: 10.1016/0010-0285(72)90003-5 Braze, D., Mencl, W. E., Tabor, W., Pugh, K. R., Constable, R. T., Fulbright, R. K., … Shankweiler, D. P. (2011). Unification of sentence processing via ear and eye: an fMRI study. Cortex, 47, 416–431.  doi: 10.1016/j.cortex.2009.11.005

276 Kate Cain and Marcia A. Barnes

Cain, K., & Nash, H. (2011). The influence of connectives on young readers’ processing and comprehension of text. Journal of Educational Psychology, 103, 429–441.  doi: 10.1037/a0022824 Cain, K., Oakhill, J., & Bryant, P. (2004). Children’s reading comprehension ability: Concurrent prediction by working memory, verbal ability, and component skills. Journal of Educational Psychology, 96, 31–42.  doi: 10.1037/0022-0663.96.1.31 Cain, K., Oakhill, J., & Bryant, P. E. (2000). Investigating the causes of reading comprehension failure: The comprehension-age match design. Reading and Writing, 12, 31–40. doi: 10.1023/A:1008058319399 Cain, K., & Oakhill, J. V. (1999). Inference making and its relation to comprehension failure. Reading and Writing. An Interdisciplinary Journal, 11, 489–503.  doi: 10.1023/A:1008084120205 Cain, K., Towse, A. S., & Knight, R. (2009). The development of idiom comprehension: An investigation of semantic and contextual processing skills Journal of Experimental Child Psychology, 102, 98–108.  doi: 10.1016/j.jecp.2008.08.001 Carretti, B., Cornoldi, C., De Beni, R., & Romano, M. (2005). Updating in working memory: A comparison of good and poor comprehenders. Journal of Experimental Child Psychology, 91, 45–66.  doi: 10.1016/j.jecp.2005.01.005 Carroll, J. B. (1993). Human cognitive abilities: A survey of factor-analytic studies. New York: Cambridge University Press.  doi: 10.1017/CBO9780511571312 Case, R., Okamoto, Y., Griffin, S., McKeough, A., Bleiker, C., Henderson, B., … Keating, D. P. (1996). The role of central conceptual structures in the development of children’s thought. Monographs of the Society for Research in Child Development, 61, 1–295. doi: 10.1111/j.1540-5834.1996.tb00535.x Catts, H. W., Hogan, T. P., & Adlof, S. M. (2005). Developmental changes in reading and reading disabilities. In H. W. Catts & A. G. Kamhi (Eds.), The connections between language and reading disabilities (pp. 25–40). Mahwah, NJ: Lawrence Erlbaum Associates. Cohen, L. B., Chaput, H. H., & Cashon, C. H. (2002). A constructivist model of infant cognition. Cognitive Development, 17, 1323–1343.  doi: 10.1016/S0885-2014(02)00124-7 Cook, A. E., & O’Brien, E. J. (2014). Knowledge activation, integration, and validation during narrative text comprehension. Discourse Processes, 51, 26–49.  doi: 10.1080/0163853X.2013.855107 Cook, A. E., & O’Brien, E. J. (2015). Passive activation and instantiation during reading. In E. J. O’Brien, A. E. Cook & R. F. Lorch (Eds.), Inference during reading (pp. 19–41). Cambridge: Cambridge University Press.  doi: 10.1017/CBO9781107279186.003 Currie, N. K., & Cain, K. (2015). Children’s inference generation: the role of vocabulary and working memory. Journal of Experimental Child Psychology, 137, 57–75. doi: 10.1016/j.jecp.2015.03.005 Daneman, M., & Merikle, P. M. (1996). Working memory and language comprehension: A meta-analysis. Psychonomic Bulletin and Review, 3, 422–433.  doi: 10.3758/BF03214546 de Jong, P. F., & van der Leij, A. (2002). Effects of phonological abilities and linguistic comprehension on the development of reading. Scientific Studies of Reading, 6, 51–77. doi: 10.1207/S1532799XSSR0601_03 Dekker, T. M., Mareschal, D., Johnson, M. H., & Sereno, M. I. (2014). Picturing words? Sensorimotor cortex activation for printed words in child and adult readers. Brain and Language, 139, 58–67.  doi: 10.1016/j.bandl.2014.09.009 Duckworth, A. L., Peterson, C., Matthews, M. D., & Kelly, D. R. (2007). Grit: perseverance and passion for long-term goals. Journal of Personality and Social Psychology, 92, 1087–1101. doi: 10.1037/0022-3514.92.6.1087



Reading comprehension 277

Ericsson, K. A., & Kintsch, W. (1995). Long-term working memory. Psychological Review, 102, 211–245.  doi: 10.1037/0033-295X.102.2.211 Evans, J. S. B., & Stanovich, K. E. (2013). Dual-process theories of higher cognition: Advancing the debate. Perspectives on Psychological Science, 8, 223–241. Fecica, A. M., & O’Neill, D. K. (2010). A step at a time: Preliterate children’s simulation of narrative movement during story comprehension. Cognition, 116, 368–381. doi: 10.1016/j.cognition.2010.05.014 Fischer, M. H., & Zwaan, R. A. (2008). Embodied language: A review of the role of the motor system in language comprehension. The Quarterly Journal of Experimental Psychology, 61, 825–850.  doi: 10.1080/17470210701623605 Fuhs, M. W., Nesbitt, K. T., Farran, D. C., & Dong, N. (2014). Longitudinal associations between executive functioning and academic skills across content areas. Developmental Psychology, 50, 1698–1709.  doi: 10.1037/a0036633 Gallagher, H. L., Happé, F., Brunswick, N., Fletcher, P. C., Frith, U., & Frith, C. D. (2000). Reading the mind in cartoons and stories: an fMRI study of ‘theory of mind’in verbal and nonverbal tasks. Neuropsychologia, 38, 11–21.  doi: 10.1016/S0028-3932(99)00053-6 Garcia, J. R., & Cain, K. (2014). Decoding and reading comprehension: a meta-analysis to identify which reader and assessment characteristics influence the strength of the relationship in English Review of Educational Research, 84, 74–111.  doi: 10.3102/0034654313499616 Garner, R. (1980). Monitoring of understanding: An investigation of good and poor readers’ awareness of induced miscomprehension of text. Journal of Reading Behaviour, XII, 55–63. Garnham, A. (1979). Instantiation of verbs. Quarterly Journal of Experimental Psychology, 31, 207–214.  doi: 10.1080/14640747908400720 Garnham, A. (1981). Anaphoric reference to instances, instantiated and non‐instantiated categories: A reading time study. British Journal of Psychology, 72, 377–384. doi: 10.1111/j.2044-8295.1981.tb02197.x Garnham, A. (2001). Mental models and the interpretation of anaphora. Philadelphia, PA: Psychology Press. Gernsbacher, M. A. (1990). Language comprehension as structure building. Hillsdale, N.J.: Lawrence Erlbaum. Gernsbacher, M. A., & Faust, M. (1991). The mechanism of suppression: A component of general comprehension skill. Journal of Experimental Psychology: Learning, Memory and Cognition, 17, 245–262. Gilbert, D. T., Krull, D. S., & Pelham, B. W. (1988). Of thoughts unspoken: Social inference and the self-regulation of behavior. Journal of Personality and Social Psychology, 55, 685–694. Gernsbacher, M. A., & Kaschak, M. P. (2003). Neuroimaging studies of language production and comprehension. Annual Review of Psychology, 54, 91–114. doi: 10.1146/annurev.psych.54.101601.145128 Gernsbacher, M. A., Varner, K. R., & Faust, M. (1990). Investigating differences in general comprehension skill. Journal of Experimental Psychology: Learning, Memory and Cognition, 16, 430–445. Glenberg, A. M., Witt, J. K., & Metcalfe, J. (2013). From the revolution to embodiment 25 years of cognitive psychology. Perspectives on Psychological Science, 8, 573–585. doi: 10.1177/1745691613498098 Glenberg, A. M., Gutierrez, T., Levin, J. R., Japuntich, S., & Kaschak, M. P. (2004). Activity and imagined activity can enhance young children’s reading comprehension. Journal of Educational Psychology, 96, 424–436.  doi: 10.1037/0022-0663.96.3.424

278 Kate Cain and Marcia A. Barnes

Gough, P. B., Hoover, W. A., & Peterson, C. L. (1996). Some observations on a simple view of reading. In C. Cornoldi & J. Oakhill (Eds.), Reading comprehension difficulties: Processes and interventions (pp. 1–13). Mahwah, NJ: Erlbaum. Gough, P. B., & Tunmer, W. E. (1986). Decoding, reading and reading disability. Remedial and Special Education, 7, 6–10.  doi: 10.1177/074193258600700104 Graesser, A. C., Singer, M., & Trabasso, T. (1994). Constructing inferences during narrative text comprehension. Psychological Review, 101, 371–395.  doi: 10.1037/0033-295X.101.3.371 Guthrie, J. T., Wigfield, A., Metsala, J. L., & Cox, K. E. (1999). Motivational and cognitive predictors of text comprehension and reading amount. Scientific Studies of Reading, 3, 231–256.  doi: 10.1207/s1532799xssr0303_3 Hagá, S., Garcia-Marques, L., & Olson, K. R. (2014). Too young to correct: A developmental test of the three-stage model of social inference. Journal of Personality and Social Psychology, 107(6), 994 1012. Hayes, D. P., & Ahrens, M. (1988). Vocabulary simplification for children: A special case of ‘motherese’. Journal of Child Language, 15, 395–410.  doi: 10.1017/S0305000900012411 Isberner, M. B., & Richter, T. (2014). Does validation during language comprehension depend on an evaluative mindset? Discourse Processes, 51, 7–25.  doi: 10.1080/0163853X.2013.855867 Joshi, R. M., & Aaron, P. G. (2000). The component model of reading: Simple view of reading made a little more complex. Reading Psychology, 21, 85–97.  doi: 10.1080/02702710050084428 Just, M. A., & Carpenter, P. A. (1992). A capacity theory of comprehension: individual differences in working memory. Psychological Review, 99, 122–149.  doi: 10.1037/0033-295X.99.1.122 Karmiloff-Smith, A. (2009). Nativism versus neuroconstructivism: rethinking the study of developmental disorders. Developmental Psychology, 45, 56–63.  doi: 10.1037/a0014506 Keenan, J. M., Baillet, S. D., & Brown, P. (1984). The effects of causal cohesion on comprehension and memory. Journal of Verbal Learning and Verbal Behavior, 23, 115–126. doi: 10.1016/S0022-5371(84)90082-3 Kendeou, P., Bohn-Gettler, C., White, M., & van den Broek, P. (2008). Children’s inference generation across different media. Journal of Research in Reading, 31, 259–272. doi: 10.1111/j.1467-9817.2008.00370.x Kendeou, P., Lynch, J. S., van den Broek, P., Espin, C., White, M., & Kremer, K. E. (2005). Developing successful readers: Building early narrative comprehension skills through television viewing and listening. Early Childhood Education Journal, 33, 91–98. doi: 10.1007/s10643-005-0030-6 Kendeou, P., van den Broek, P., White, M., & Lynch, J. S. (2009). Predicting reading comprehension in early elementary school: the independent contributions of oral language and decoding skills. Journal of Educational Psychology, 101, 765–778.  doi: 10.1037/a0015956 Kim, Y. S. (2015). Language and cognitive predictors of text comprehension: Evidence from multivariate analysis. Child Development, 86, 128–144.  doi: 10.1111/cdev.12293 Kintsch, W. (1998). Comprehension: A Paradigm for Cognition. New York, NY: Cambridge University Press. Kuhn, L. J., Willoughby, M. T., Wilbourn, M. P., Vernon‐Feagans, L., & Blair, C. B. (2014). Early communicative gestures prospectively predict language development and executive function in early childhood. Child Development, 85, 1898–1914. LARRC. (2015). Learning to read: should we keep things simple? Reading Research Quarterly, 50, 151–169. LARRC & Logan, J. (in press). Pressure points in reading comprehension: A quantile multiple regression analysis. Journal of Educational Psychology.



Reading comprehension 279

LARRC, Jiang, H., & Farquharson, K. (under review). Are memory and attention equally important for reading and listening comprehension? A developmental comparison. Lepola, J., Lynch, J. S., Laakkonen, E., Silven, M., & Niemi, P. (2012). The role of inference making and other language skills in the development of narrative listening comprehension in 4-6-year-old children. Reading Research Quarterly, 47, 259–282. Lindenberg, R., & Scheef, L. (2007). Supramodal language comprehension: role of the left temporal lobe for listening and reading. Neuropsychologia, 45, 2407–2415. doi: 10.1016/j.neuropsychologia.2007.02.008 Locascio, G., Mahone, E. M., Eason, S., & Cutting, L. E. (2010). Executive dysfunction among children with reading comprehension deficits. Journal of Learning Disabilities, 43, 441–454. doi: 10.1177/0022219409355476 Lorsbach, T. C., Katz, G. A., & Cupak, A. J. (1998). Developmental differences in the ability to inhibit the initial misinterpretation of garden path passages. Journal of Experimental Child Psychology, 71, 275–296.  doi: 10.1006/jecp.1998.2462 Lynch, J. S., van den Broek, P., Kremer, K., Kendeou, P., White, M. J., & Lorch, E. P. (2008). The development of narrative comprehension and its relation to other early reading skills. Reading Psychology, 29, 327–365.  doi: 10.1080/02702710802165416 Magliano, J. P., Miller, J., & Zwaan, R. A. (2001). Indexing space and time in film understanding. Applied Cognitive Psychology, 15, 533–545.  doi: 10.1002/acp.724 McKoon, G., & Ratcliff, R. (1992). Inference during reading. Psychological Review, 99, 440–466. doi: 10.1037/0033-295X.99.3.440 McNamara, D. S., & Magliano, J. (2009). Toward a comprehensive model of comprehension. In B. Ross (Ed.), Psychology of learning and motivation (Vol. 51, pp. 297–384). Burlington: Academic Press.  doi: 10.1016/S0079-7421(09)51009-2 Munte, T. F., Schiltz, K., & Kutas, M. (1998). When temporal terms belie conceptual order. Nature, 395, 71–73.  doi: 10.1038/25731 Muter, V., Hulme, C., Snowling, M., & Stevenson, J. (2004). Phonemes, rimes, vocabulary and grammatical skills as foundations of early reading development: evidence from a longitudinal study. Developmental Psychology, 40, 665–681.  doi: 10.1037/0012-1649.40.5.665 Myers, J. L., & O’Brien, E. J. (1998). Accessing the discourse representation during reading. Discourse Processes, 26, 131–157.  doi: 10.1080/01638539809545042 Oakhill, J., & Cain, K. (2012). The precursors of reading comprehension and word reading in young readers: Evidence from a four-year longitudinal study. Scientific Studies of Reading, 16, 91–121.  doi: 10.1080/10888438.2010.529219 Ouellette, G. P. (2006). What’s meaning got to do with it: The role of vocabulary in word reading and reading comprehension. Journal of Educational Psychology, 98, 554–566. doi: 10.1037/0022-0663.98.3.554 Paris, A. H., & Paris, S. G. (2003). Assessing narrative comprehension in young children. Reading Research Quarterly, 38, 36–76.  doi: 10.1598/RRQ.38.1.3 Perez, A., Cain, K., Castellanos, M. C., & Bajo, T. (2015). Inferential updating in narrative texts: an ERP study. Memory and Cognition, 43, 1105–1135.  doi: 10.3758/s13421-015-0528-0 Perfetti, C. A. (1985). Reading Ability. New York: Oxford University Press. Perfetti, C. A. (1988). Verbal efficiency in reading ability. In M. Daneman, G., MacKinnon & T. G. Waller (Eds.), Reading Research: Advances in Theory and Practice (Vol. 6). San Diego: Academic Press. Perfetti, C. A. (2007). Reading ability: Lexical quality to comprehension. Scientific Studies of Reading, 11, 357–383.  doi: 10.1080/10888430701530730

280 Kate Cain and Marcia A. Barnes

Perfetti, C. A., & Hart, L. (2001). The lexical basis of comprehension skill. In D. S. Gorfein (Ed.), On the consequences of meaning selection: Perspectives on resolving lexical ambiguity (pp. 67–86). Washington, DC, US: APA.  doi: 10.1037/10459-004 Perfetti, C. A., & Stafura, J. Z. (2015). Comprehending implicit meanings in text without making inferences. In E. J. O’Brien, A. E. Cook & R. F. Lorch (Eds.), Inferences during reading (pp. 1–18). Cambridge: Cambridge University Press.  doi: 10.1017/CBO9781107279186.002 Piaget, J. (1952). The origins of intelligence in children. New York, New York: International Univeristy Press.  doi: 10.1037/11494-000 Pike, M., Swank, P., Taylor, H., Landry, S., & Barnes, M. A. (2013). Effect of preschool working memory, language, and narrative abilities on inferential comprehension at school-age in children with Spina Bifida Myelomeningocele and typically developing children. Journal of the International Neuropsychological Society, 19, 390–399.  doi: 10.1017/S1355617712001579 Pike, M. M., Barnes, M. A., & Barron, R. W. (2010). The role of illustrations in children’s inferential comprehension. Journal of Experimental Child Psychology, 105, 243–255. doi: 10.1016/j.jecp.2009.10.006 Pulvermüller, F. (2013). Semantic embodiment, disembodiment or misembodiment? In search of meaning in modules and neuron circuits. Brain and language, 127, 86–103. doi: 10.1016/j.bandl.2013.05.015 Pyykkönen, P., & Järvikivi, J. (2012). Children and situation models of multiple events. Developmental Psychology, 48, 521–529. Radvansky, G. A., & Copeland, D. E. (2001). Working memory and situation model updating. Memory & Cognition, 29, 1073–1080.  doi: 10.3758/BF03206375 Richter, T., Isberner, M.-B., Naumann, J., & Neeb, Y. (2013). Lexical quality and reading comprehension in primary school children. Scientific Studies of Reading, 17, 415–434. doi: 10.1080/10888438.2013.764879 Sachs, J. S. (1967). Recognition of semantic, syntactic, and lexical changes in sentences Psychonomic Bulletin, 1, 17–18. San Juan, V., Khu, M., & Graham, S. A. (2015). A New Perspective on Children’s Communicative Perspective Taking: When and How Do Children Use Perspective Inferences to Inform Their Comprehension of Spoken Language?. Child Development Perspectives, 9, 245–249. doi: 10.1111/cdep.12141 Schneider, W., Körkel, J., & Weinert, F. E. (1989). Domain-specific knowledge and memory performance: A comparison of high-and low-aptitude children. Journal of Educational Psychology, 81, 306–312.  doi: 10.1037/0022-0663.81.3.306 Scott, C. M. (2009). A case for the sentence in reading comprehension. Language, Speech, and Hearing Services in Schools, 40, 184–191.  doi: 10.1044/0161-1461(2008/08-0042) Seigneuric, A., Ehrlich, M.-F., Oakhill, J. V., & Yuill, N. M. (2000). Working memory resources and children’s reading comprehension. Reading and Writing. An Interdisciplinary Journal, 13, 81–103.  doi: 10.1023/A:1008088230941 Sesma, H. W., Mahone, E. M., Levine, T., Eason, S. H., & Cutting, L. E. (2009). The contribution of executive skills to reading comprehension. Child Neuropsychology, 15, 232–246. doi: 10.1080/09297040802220029 Silva, M. T., & Cain, K. (2015). The relations between lower- and higher-level oral language skills and their role in prediction of early reading comprehension Journal of Educational Psychology, 107, 321–331.  doi: 10.1037/a0037769 Singer, M. (1994). Discourse inference processes. In M. A. Gernsbacher (Ed.), Handbook of psycholinguistics (pp. 479–515). San Diego, CA US: Academic Press.



Reading comprehension 281

Stanovich, K. E. (1993). Does reading make you smarter? Literacy and the development of verbal intelligence. In H. Reese (Ed.), Advances in Child Development and Behavior (Vol. 24, pp. 133–180): Academic Press. Stein, N. L., & Glenn, C. G. (1982). Children’s concept of time: the development of story schema. In W. Friedman, J. (Ed.), The developmental psychology of time (pp. 255–282). New York: Academic Press. Tannenbaum, K. R., Torgesen, J. K., & Wagner, R. K. (2006). Relationships between word knowledge and reading comprehension in third-grade children. Scientific Studies of Reading, 10, 381–398.  doi: 10.1207/s1532799xssr1004_3 Therriault, D. J., Rinck, M., & Zwaan, R. A. (2006). Assessing the influence of dimensional focus during situation model construction. Memory & Cognition, 34, 78–89. Uttal, D. H., Fisher, J. A., & Taylor, H. A. (2006). Words and maps: developmental changes in mental models of spatial information acquired from descriptions and depictions. Developmental Science, 9, 221–235. van den Broek, P., Rapp, D. N., & Kendeou, P. (2005). Integrating memory-based and constructionist approaches in accounts of reading comprehension. Discourse Processes, 39, 299–316.  doi: 10.1080/0163853X.2005.9651685 Van Dyke, J. A., & Johns, C. L. (2012). Memory interference as a determinant of language comprehension. Language and Linguistics Compass, 6, 193–211.  doi: 10.1002/lnc3.330 Van Dyke, J. A., Johns, C. L., & Kukona, A. (2014). Low working memory capacity is only spuriously related to poor reading comprehension. Cognition, 131, 373–403. doi: 10.1016/j.cognition.2014.01.007 Van Dyke, J. A., & Shankweiler, D. P. (2013). From verbal efficiency to lexical quality. In M. A. Britt, S. R. Goldman & J. F. Rouet (Eds.), Reading – from words to multiple texts (pp. 115–131). New York: Routledge. Verhoeven, L., & van Leeuwe, J. (2008). Prediction of the development of reading comprehension: a longitudinal study. Applied Cognitive Psychology, 22, 407–423.  doi: 10.1002/acp.1414 Wang, J. H. Y., & Guthrie, J. T. (2004). Modeling the effects of intrinsic motivation, extrinsic motivation, amount of reading, and past reading achievement on text comprehension between US and Chinese students. Reading Research Quarterly, 39, 162–186.  doi: 10.1598/RRQ.39.2.2 Wolters, C. A., Denton, C. A., York, M. J., & Francis, D. J. (2014). Adolescents’ motivation for reading: group differences and relation to standardized achievement. Reading and Writing, 27, 503–533. Zimmerman, B., & Schunk, D. H. (2006). Competence and control beliefs: Distinguishing the means and ends. In P. A. Alexander & P. H. Winne (Eds) Handbook of educational psychology (2nd Ed, pp 349–367). New York: Routledge. Zwaan, R. A. (1994). Effect of genre expectations on text comprehension. Journal of Experimental Psychology: Learning, Memory, and Cognition, 20, 920–933. Zwaan, R. A. (2014). Embodiment and language comprehension: reframing the discussion. Trends in cognitive sciences, 18, 229–234.  doi: 10.1016/j.tics.2014.02.008 Zwaan, R. A., Langston, M. C., & Graesser, A. C. (1995). The construction of situation models in narrative comprehension: An event-indexing model. Psychological Science, 6, 292– 297.  doi: 10.1111/j.1467-9280.1995.tb00513.x Zwaan, R. A., & Madden, C. J. (2004). Updating Situation Models. Journal of Experimental Psychology: Learning, Memory, and Cognition, 30, 283–288. Zwaan, R. A., & Radvansky, G. A. (1998). Situation models in language comprehension and memory. Psychological Bulletin, 123, 162–185.  doi: 10.1037/0033-2909.123.2.162

Development of reading comprehension Change and continuity in the ability to construct coherent representations Paul van den Broek and Panayiota Kendeou Leiden University / University of Minnesota

The ultimate goal of reading is to extract meaning from written text, and access and use that meaning at a later point in time. For example, the Organization for Economic Co-operation and Development defines reading as the capacity “to understand, use, reflect on and engage with written texts, in order to achieve one’s goals, to develop one’s knowledge and potential, and to participate in society” (OECD, 2009, p. 4). A reader’s success accomplishing these aims depends to a great extent on his or her language comprehension skills, the ability to extract meaning from linguistic messages. Language comprehension is broad: in the context of oral language it involves listening comprehension, in the context of written language it involves reading comprehension, and in the context of internet searches and other multimedia it involves comprehension and integration of textual and other symbolic and non-symbolic information (such as graphs, pictures, video). At the core of language comprehension in any of these contexts is the construction of a coherent mental representation of the presented information (Gernsbacher, 1990; Kintsch & van Dijk, 1978; Trabasso, van den Broek, & Suh, 1989; van den Broek, Young, Tzeng, & Linderholm, 1999). In a coherent representation, semantic relations connect the various elements of the message to each other and, importantly, to the individual’s background knowledge. This representation is the outcome of the comprehension process, and it is the foundation from which the individual can perform specific tasks such as evaluating the message, retrieving the information at a later point in time, answering questions about the message, and applying the knowledge gained from the message. The processes by which an individual constructs a mental representation of a message are influenced by the characteristics of the individual, by properties of the message itself, by the demands of the specific task (Snow, 2002; National Reading Panel, 2000), and by the interactions between these factors. Because the individual engages in these processes in an attempt to achieve coherence, they are moderated by the individual’s standards of coherence for a particular reading occasion. The standards may be explicit or implicit to the individual (Ferreira & Patson, 2007; van den Broek, doi 10.1075/swll.15.16van © 2017 John Benjamins Publishing Company

284 Paul van den Broek and Panayiota Kendeou

Risden & Husebye-Hartmann, 1995; van den Broek, Bohn-Gettler, Kendeou, Carlson, & White, 2011), and they determine the type and depth of processing in which the individual will engage and, therefore, determine the completeness and coherence of the resulting representation. As described below, developmental and individual differences in reading comprehension are closely aligned with an individual’s standards of coherence and his or her effective facility with the metacognitive, inferential strategies necessary to attain these standards. Our aim in this chapter is to provide an overview of the skills that are central in building coherence during language comprehension, and of the application of these skills in the context of reading. First, we discuss reader, text, and task factors that influence comprehension. Next, we discuss the developmental continuity and change of these factors, their respective interactions and interdependencies, as well as their reciprocal relation to background knowledge. We conclude with a discussion of implications for current theory and practice, as well as for future research in this area.

Building a coherent representation Successful reading comprehension depends on the execution and integration of many cognitive processes. These processes take place at different grain sizes within the text. For instance, to understand a sentence, readers visually process the individual words, identify and access the word’s phonological, orthographic, and semantic representations, and connect these representations in accordance with the rules of syntax to form an understanding of the underlying meaning of the sentence (Perfetti, 2007; Verhoeven & Perfetti, 2011). Similarly, to comprehend a set of sentences that form a text, readers process and connect individual idea units, resulting (if all goes well) in the construction of a complete and coherent mental representation of the text as a whole (Graesser & Clark, 1985; Kintsch, 1998; Kintsch & van Dijk, 1978; Trabasso, Secco & van den Broek, 1984). In this chapter we focus on the development of the ability to comprehend texts as a whole (see also Cain & Barnes, this volume). The construction of a mental representation is a gradual process, involving various component processes that occur moment-by-moment as the reader moves through the text. As the reader proceeds from one text element (e.g., sentence) to the next, each new element is understood in the context of the representation of prior text elements and of the reader’s background knowledge. This requires that the reader identifies semantic connections between these various sources of information, often via inferences. Some of these connections are readily identified. For example, information in the current element usually is easily connected to information that is still in the reader’s working memory as a result of processing of the immediately preceding element. Likewise, connections to background knowledge that has strong prior associations to the concepts in the current element tend to be made with little effort. Other connections are more difficult to identify. For example, connections to information that is distant in the text or that is less accessible in the reader’s text memory are less easily identified.



Development of reading comprehension 285

Likewise, connections to background knowledge that is only weakly associated to concepts in the current element require more effort. The latter connections are more at risk of being unrecognized by the reader or may require special efforts on the part of the reader. If these special efforts require strategic processes, then whether or not they are executed depends on the reader’s standards of coherence at the time of reading the text (van den Broek et al., 1995, 1999, 2011). We return to inference generation and the role of standards of coherence later in this chapter. As the reader proceeds through a text, each newly read text element may trigger a somewhat different combination of processes, resulting in fluctuations in processing activities. The connections that are identified at a particular element result in the updating of the text representation that has been constructed so far; the updated representation, in turn, forms the backdrop for interpreting the next element, which results in a new cycle of updating, and so on (Goldman & Varma, 1995; Kintsch & van Dijk, 1978; van den Broek et al., 1999). The recursive processes of connecting and updating continue until the end of the text is reached and reading stops. Details of these fluctuating processes and the gradual emergence of a mental representation of the text can be found elsewhere (van den Broek, Risden, Fletcher, & Thurlow, 1996; van den Broek et al., 1999; van den Broek, Rapp, & Kendeou, 2005). The important point here is that the quality and type of processes that occur at each point in the reading process determine the quality, completeness, and coherence of the representation that the reader has at the end of reading the text.

Reader, text, and task factors that affect comprehension Many factors determine the success or failure of the processes that occur during reading. These factors can be grouped into three categories: reader characteristics, text characteristics, and the demands of the reading task (van den Broek, Fletcher, & Risden, 1993).

General reader characteristics Several general reader characteristics set boundary conditions for a reader’s ability to effectively execute comprehension processes. One important set of characteristics concerns the reader’s ability to translate the written code into words that refer to concepts and have meaning. This ability depends heavily on linguistic skills, such as phonological and orthographic processing (e.g., Perfetti & Hart, 2001), fluency (e.g., Fuchs, Fuchs, Hosp, & Jenkins, 2001; LaBerge & Samuels, 1974), and vocabulary (e.g., Anderson & Freebody, 1981; Beck, Perfetti, & McKeown, 1982; Nagy, Herman, & Anderson, 1985; Snow, 2002; Stanovich, 1986; Wagner et al., 1997). A second set of characteristics concern non-linguistic cognitive skills that play a critical role, such as executive functions (McVay & Kane, 2012; Sesma, Mahone, Levine, Eason, &

286 Paul van den Broek and Panayiota Kendeou

Cutting, 2009), working-memory capacity (Cain, Oakhill, & Bryant, 2004; Daneman & Carpenter, 1980; Linderholm & van den Broek, 2002; Swanson & O’Connor, 2009), and attention-allocation skills (Glenberg, Wilkinson, & Epstein, 1982; Liu, Reichle, & Gao, 2013) (see also Oakhill & Cain, this volume for the role of cognitive skills in poor comprehension). A third important characteristic that determines success and failure in comprehension consists of the reader’s prior knowledge. In the context of reading comprehension several forms of prior knowledge are particularly relevant: domain knowledge about the specific field or area of the discourse (e.g., science, medieval history); topic knowledge relevant to the content of a particular discourse; general world knowledge that captures information assumed to be known by the general population; vocabulary knowledge about the meaning of words, particularly as this knowledge concerns the depth and breadth of the reader’s vocabulary (Perfetti & Adlof, 2012; Tannenbaum, Torgesen, & Wagner, 2006); and genre knowledge about the specific purposes, structures, and audiences of different texts (Cain & Oakhill, 2012). These and related characteristics comprise the cognitive infrastructure on which the comprehension processes take place.

Text characteristics Properties of the text also influence the comprehension processes. On the one hand, different genres and types of text invite different kinds of comprehension because they tend to serve different purposes. Genre refers to a classification based on the intent of the communicator; for example, to inform (expository), to tell a story (narrative), to convince (persuasive), to give instructions or directions (procedural), and so on. To be proficient, readers must recognize these various purposes and adjust their processing accordingly. On the other hand, characteristics of the particular text at hand also influence the processing by the reader. Besides the content of the text, its structure and cohesion influence the type and success of the reader’s processing. Structure refers to the way the material is organized in the text by the communicator; for example problem-solution, cause-effect, compare-contrast, description, and others (Meyer & Poon, 2001). Cohesion refers to the grammatical and lexical linking within a text that holds it together and gives it unity (Halliday & Hasan, 1976); for example, high cohesion texts include connectives, transitions, and other cues that hold the text together, whereas low cohesion texts do not (Graesser, McNamara, Louwerse, & Cai, 2004; Lemarié, Lorch, Eyrolle, & Virbel, 2008). The structural and cohesive properties of a text provide processing cues to the reader, both in a facilitative and in a directive sense (McNamara & Kintsch, 1996; Williams et al., 2002). They may facilitate by supporting the reader in identifying connections and deriving meaning, they may direct by steering comprehension towards a particular interpretation (and away from another interpretation), and so on.



Development of reading comprehension 287

Reading goals and task characteristics Comprehension processes are influenced significantly by the general reading goals that are assigned to the reader, as well by specific task or relevance instructions (Cerdán, & Vidal-Abarca, 2008; Linderholm & van den Broek, 2002; Lorch, Lorch, & Klusewitz, 1993; van den Broek, Lorch, Linderholm, & Gustafson, 2001). Common reading goals include reading to study, reading for entertainment, reading to search for information (scanning or skimming), reading to learn, reading to integrate information from multiple sources (e.g., multiple texts or text and graphics), reading to evaluate, critique, and use information, and reading for general comprehension (Kendeou, Bohn-Gettler, & Fulton, 2011). Task or relevance instructions are those that prompt individuals to read for a specific reason in relation to specific text information (Kaakinen & Hyönä, 2005; McCrudden, Magliano, & Schraw, 2011; Schraw, Wade & Kardash, 1993), such as to highlight important information, to identify arguments that support a position, to answer specific questions, to write a summary. Goals and task characteristics affect the allocation of attention as well as the specific processes that the reader executes. For example, proficient readers with the goal to study for a test engage in more reprocessing (looking back, rereading) of the text (Yeari, van den Broek, & Oudega, 2015) and produce more knowledge-based elaborations and connecting inferences (Linderholm & van den Broek, 2002) than when they read for entertainment. Such differences in processing as a result of reading goal or task translate into differences in the actual representation for the text as evidenced, for example, by differences in the content and amount of textual information that is remembered by the reader (Anderson & Pichert, 1978; Ross, Green, Salisbury-Glennon & Tollefson, 2006). It appears though, that the ability to adjust one’s reading to the reading goal or task at hand is stronger for proficient than for struggling readers (Rouet, Vidal-Abarca, Bert-Erboul & Millogo, 2001). We return to this point later.

When the reader, the text, and the task meet: The spark of comprehension When a particular reader encounters a particular text, the characteristics of the reader, of the text, and of the specific goal or task for reading together determine the comprehension processes that will take place. These interactions are perhaps most evident in the context of two essential components of the comprehension process, inference generation and the application of standards of coherence. In the context of reading comprehension, an inference involves the recruitment of information to fill in conceptual gaps in the text (Elbro & Buch-Iversen, 2013; Kintsch, 1998; McNamara & Magliano, 2009; Oakhill, 1984; van den Broek, 1994). This information may come from the reader’s background knowledge, from earlier portions of the text at hand, or from other texts. Inferences are essential to comprehension because they create the network of semantic connections between elements from the text as a whole and between elements from the text and elements from background knowledge

288 Paul van den Broek and Panayiota Kendeou

or other texts. Generating an inference depends on two interacting sets of processes in which (a) information in the current text element activates (i.e., brings into working memory) information from prior text and/or long-term memory and (b) the current information becomes integrated with the activated information to form a new semantic connection (McKoon & Ratcliff, 1992; Singer, 2013). Activation and integration of information are parallel but asychronous processes – integration cannot begin unless activation has started (Cook & O’Brien, 2014). They are in part passive and effortless, as a result of spread-of-activation or resonance through long-term memory or through the representation of prior portions of the text (Kintsch, 1988; Myers & O’Brien, 1998; O’Brien & Myers, 1999; van den Broek, 1994). They may also include strategic processes, specifically aimed at comprehension and requiring cognitive effort. For example, readers may go back in the text to reread information or may search for specific information in their memory for the text or in their background knowledge. Whereas passive, associative processes take place continually as the reader proceeds through the text, strategic comprehension-oriented processes tend to occur only when needed for adequate understanding of the text. The balance between passive and strategic inferential processes is influenced by the readers’ implicit or explicit standards of coherence (Kendeou, 2014; van den Broek, Beker & Oudega, 2015). If, when reading a particular text element, the concepts activated through the passive processes yield enough coherence to meet a reader’s standards then no strategic processes are initiated and the reader continues through the text. However, if the standards are not met (e.g., if referential or causal/logical coherence is lacking or if retrieved information is inconsistent with the current text input) then strategic processes are initiated. Note that even the strategic processes may take place without the reader being aware, for instance because they have become routinized through repeated experience (Thurlow & van den Broek, 1997). A reader’s standards of coherence consist of the criteria for comprehension that the reader employs during reading (van den Broek et al., 1995, 1999, 2011). These criteria may be explicit or implicit, and concern both the types of connections (e.g., causal, referential, spatial, etc.) and the strength of the connections that are pursued. Standards vary between individuals as well as within an individual across various reading situations. For example, proficient readers adapt their standards to the goal for reading the text more than do weaker readers (Linderholm & van den Broek, 2002; Cain, 1999). Compared to weak readers, proficient readers have better knowledge of standards and of strategies to attain and have stronger comprehension-monitoring skills. There also are systematic developmental differences in children’s knowledge of standards and their ability to apply them. We discuss those next, in the section on developmental differences in reading comprehension. The standards of coherence that a reader applies are situation specific, in that they are assembled as a function of the unique confluence of text, reader, and reading context characteristics (Jenkins, 1979; van den Broek et al., 1993). The ‘context of reading’ includes the reader’s goals and interpretations of task and instructions but also external features of the situation (e.g., the presence/absence of distractors or secondary tasks).



Development of reading comprehension 289

The conceptualization of inference generation and standards of coherence as the joint products of reader-text-task characteristics has important implications. One implication pertains to the assessment of these constructs. Because the constructs are the result of complex interactions, they are, by definition, situation- and context-specific. Even though some transfer is expected across situations and contexts (Kendeou, Bohn-Gettler, White, & van den Broek, 2008) there are also situation- and context-specific manifestations of these constructs that are expected to be unique to those situations and contexts. Consequently, the assessment of these constructs must be conducted and interpreted with these situations and contexts in mind (see for example, scenario-based assessments by Sabatini, O’Reilly, Halderman, & Bruce, 2014). A second implication pertains to the relation between these constructs. Because they operate in synchrony during reading comprehension and exert a joint, dynamic influence on building coherence, they are interdependent. This interdependency is strengthened further because both draw in part on the same knowledge structures in long-term memory. Thus, standards of coherence influence inference generation, these inferences influence standards of coherence, which in turn influence further inference generation, and so on.

Development: Continuity and change in building coherence The reader, text and task factors and their interactions described above influence (reading) comprehension at all stages of development. However, reader characteristics do change with age and experience. As a result, the relative contribution of these characteristics to determining success and failure in comprehension of texts changes across developmental stages. For example, in the early elementary school grades decoding skills are a major contributor to individual differences in reading comprehension, but from the later elementary school years onward, comprehension skills take over as the strongest predictors (Catts, Hogan, & Fey, 2003; Ehri, Nunes, Stahl, & Willows, 2001; Tilstra, McMaster, van den Broek, Kendeou & Rapp, 2009). This shift in major determinants coincides with a transition in the function of texts and reading, namely that from learning to read to reading to learn (Chall, Jacobs, & Baldwin, 1990; Snow & Sweet, 2003), and fits with current conceptualizations of development as a dynamic system that is driven by interactions between processes and contexts, and characterized by the waxing and waning patterns of different processes and skills (e.g., Dynamic Systems theory, van Geert, 1991; Thelen & Bates, 2003; Overlapping Waves Model, Siegler, 1996). The development of reading comprehension skills is characterized by continuity and change. The continuity across development lies, first, in the fact that already at a very young age children are involved in the construction of coherence. Indeed, there is evidence that 2-year old children can infer causal relations between sequences of events (Bauer, 2007; Bauer & Lukowski, 2010). Also, 4-year old children can identify semantic connections in discourse and show sensitivity to structural centrality (van den Broek, Helder, & van Leijenhorst, 2013) as evidenced, for example, by their

290 Paul van den Broek and Panayiota Kendeou

ability to identify themes and the central goals of characters (Bauer & San Souci, 2010; Kendeou, van den Broek, White, & Lynch, 2007; Lynch, van den Broek, Kremer, Kendeou, White, & E. Lorch, 2008; van den Broek, Lorch, & Thurlow, 1996). Second, comprehension skills at an early age predict comprehension skills at later ages. Already at an early age comprehension and decoding skills form relatively distinct clusters of skills. The results of longitudinal investigations show that each cluster follows a relatively independent developmental trajectory, with comprehension predicting comprehension and decoding skills predicting decoding skills. The two clusters jointly, and each uniquely, predict reading comprehension by the middle elementary school years (Cain, Oakhill, Barnes, & Bryant, 2001; Kendeou et al., 2008; Kendeou, van den Broek, White, & Lynch, 2009; Oakhill, & Cain, 2012; Storch & Whitehurst, 2002; Whitehurst & Lonigan, 1998). Thus, there is a considerable degree of continuity, both in terms of the processes and skills used during language comprehension by children at different ages and in terms of the predictive power of individual differences across development. In addition to continuity, with development there also are systematic changes in comprehension abilities, as children develop and hone their strategies and skills through experience and instruction. The developmental changes involve a deepening of the ability to create coherence as well as broadening of the range of comprehension contexts in which these skills can be applied (in the reading context, for example, from narrative to other text types, from single texts to multiple texts). Specifically, there is considerable evidence that inference skills develop in non-reading contexts as early as preschool and transfer to reading contexts with the beginning of formal reading instruction (e.g., Kendeou et al., 2008; Kendeou, 2015; Lynch et al., 2008). Inference skills continue to develop in elementary school and are highly predictive of overall reading comprehension performance (Cain et al., 2004; Kendeou et al., 2008; Pike, Barnes, & Barron, 2010). The efficiency of inference generation in establishing coherence improves considerably in secondary school, where it is a strong predictor of overall reading comprehension performance (Barth, Barnes, Francis, Vaughn, & York, 2015; Cromley & Azevedo, 2007; Karasinski & Ellis-Weismer, 2010). With age, children become increasingly capable of identifying semantic relations between the elements that they encounter in a text and between the text elements and their background knowledge. This development runs along several dimensions. In general, even very young children are able to identify simple semantic relations, those that (a) involve concrete events and objects, (b) involve a small number of elements, (c) are physical in nature (e.g., X physically causes Y) as opposed to logical (e.g., if X, then necessarily Y) or psychological (e.g., internal state X brings about action Y), and (d) involve elements that are presented close together in the surface structure of the text (Bourg, Bauer & van den Broek, 1997; Thompson & Myers, 1985; van den Broek et al., 2013). With development, children increasingly are able to recognize and identify relations that involve multiple elements, that extend over greater distance in the text (s), and that are more abstract and/or involve more implied relations. For example, text themes, human motives, and so on become more often extracted routinely (Lynch & van den Broek, 2007; van den Broek, Lynch, Naslund, Ievers-Landis & Verduin, 2003;



Development of reading comprehension 291

Williams, 1993). As a result of these developments, children’s representations of texts expand both quantitatively (i.e., in number of connections) and qualitatively (i.e., in types of connections). These developments are aided by their expanding knowledge of text structure and text genres (Compton, Appleton, & Hosp, 2004; Eason, Goldberg, Young, Geist, & Cutting, 2012; Williams, Hall, Lauer, Stafford, De Sisto, & deCani, 2005). The result of these developments is that children become increasingly sensitive to structural properties of the texts they encounter and to the fact that some text elements are more central to that structure than others. The results of reading time and eye-tracking studies show that proficient readers tend to devote more attention and deeper processing during reading to text elements that are central to text structure, as indicated by having a large number of semantic relations to other text elements, than to text elements that are not central (McMaster et al., 2012; Yeari, van den Broek et al., 2015) and also tend to include those central elements prominently in their mental representations. For example, in proficient adult readers, the frequency with which text elements are recalled, included in summaries, and rated as important is strongly predicted by the number of semantic connections that each element has (Trabasso & van den Broek, 1985; van den Broek, 1989, 1990, 1997). In addition, when the number of connections of a text element is modified, its frequency of inclusion in recall changes accordingly (van den Broek & Trabasso, 1986; van den Broek, 1988). Even young children display such sensitivity to structural centrality in process and product of reading (van den Broek, Lorch & Thurlow, 1996), as do children with (mild) learning disabilities (Wolman, van den Broek, & Lorch, 1997), children with Down Syndrome (Kim, Kendeou, van den Broek, White, & Kremer, 2008) and children with attention-deficit disorders (Lorch, Berthiaume, Milich & van den Broek, 2007; Lorch, et al., 2004; Lorch et al., 1999). But children’s sensitivity gradually develops as they gain experience with reading and reading strategies, expand their knowledge of the structure of texts of different genres and of types of semantic relations, and refine their reading skills (Oakhill & Cain, 2012; van den Broek et al., 2013). Thus, one of the main overall patterns of development in reading comprehension that emerges from the development of the various component skills and factors described above is that, with age and experience, children become increasingly sensitive to the structure of the text and, hence, to the structural centrality of text elements. Along with the inference generation skills and strategies for creating coherence, children’s standards of coherence also change. In early elementary school, the consistency and effectiveness of children’s use of different standards during reading is relatively limited. Young and less experienced readers not only use fewer criteria for comprehension but they also have difficulty applying different standards simultaneously (e.g., checking external consistency, internal consistency, word understanding, coherence of the text representation constructed so far) due to limited attentional resources and general world knowledge (Baker, 1984). In later elementary school, readers become more adapt in task-oriented reading and in adjusting their processes to meet specific reading goals (Tilstra & McMaster, 2013). In secondary school, the efficiency and effectiveness of the use of different standards during reading increases

292 Paul van den Broek and Panayiota Kendeou

significantly as executive functions (Luna, Garver, Urban, Lazar & Sweeney, 2004) and attention-allocation skills (Liu et al., 2013) further mature. Stronger and more efficient use and monitoring of standards of coherence relates also to the development of meta-comprehension skills, that is a reader’s ability to accurately judge his or her level of understanding during learning from text (Rawson, Dunlosky, & Thiede, 2000; Redford, Thiede, Wiley, & Griffin, 2012). In addition to a deepening knowledge of the appropriate standards for different reading situations, a child’s strategies for attaining his or her standards expand, become more efficient and, ideally, automated.

The role of knowledge in building coherence One factor that has clear continuity and change, and that has a strong influence on the reader’s ability to construct coherence is prior knowledge. Prior knowledge has the potential to both facilitate and interfere with comprehension of information from texts (Kendeou & O’Brien, 2015). The facilitative influence of prior knowledge on reading comprehension is well established – readers have the tendency to understand and interpret new information based on pre-existing structures in memory (Kintsch & van Dijk, 1978; Kintsch 1988). In this context, the effects of prior knowledge have been shown to influence the time necessary to comprehend a text, the generation of inferences, and the final mental representation of the text (Afflerbach, 1990; Alexander, Murphy, Woods, Duhon, & Parker, 1997; Alexander & Murphy, 1998; Alexander & Jetton, 2000). The interference of prior knowledge on text comprehension is also well established – readers’ pre-existing misconceptions have a pervasive effect on comprehension and learning (Vosniadou & Brewer, 1994; van den Broek, 2010). In this context, incorrect prior knowledge or misconceptions influence the actual cognitive processes during reading, as well as the content of those processes (Kendeou, & van den Broek, 2005; Kendeou, & van den Broek, 2007; Kendeou, Muis, & Fulton, 2011). Specifically, in an effort to build coherence, readers activate and integrate their prior knowledge with the textual information. If readers activate and integrate pre-existing incorrect knowledge or misconceptions, the content of their knowledge-based inferences (e.g., explanatory, predictive, and other inferences) will reflect these misconceptions and inaccuracies, resulting in impoverished mental representations. It is worth noting that the relation between a reader’s background knowledge and his or her comprehension of texts is reciprocal. On the one hand, as described above, background knowledge provides the backdrop for comprehension of texts. On the other hand, texts are an important source of learning, when reading a text leads to changes in the reader’s background knowledge. These changes in a reader’s background knowledge can involve the acquisition or addition of new knowledge and the revision or updating of pre-existing knowledge. The acquisition of new knowledge from texts takes place in both formal and informal learning environments. This can be accomplished with the reading of single or multiple texts (Goldman, 2004; Rouet, 2006; van den Broek & Kendeou, 2015), for



Development of reading comprehension 293

example in the context of Google Searches, Wikipedia, and so on. There is little doubt that learning from texts occurs and there are a considerable number of methods for manipulating texts and for teaching children strategies that have been found effective in fostering learning from texts (e.g., Pearson & Hamm, 2005; Pressley & McCormick, 1995). However, the precise processes by which learning occurs are far less understood. Models of learning from text posit that the processes involved in constructing knowledge are parallel to, yet extend those involved in constructing a mental representation of a text (Kintsch, 1998; Goldman, 1997). The representations involved in learning are long term and decontextualized, whereas the memory representations that result from comprehension of texts are contextualized and (often) shorter term (Kintsch, 1988; Beker, Jolles & van den Broek, 2016). Thus, comprehension is necessary but not sufficient for learning from texts. The processes involved in the revision or updating of pre-existing knowledge have become considerably better specified in recent years. Empirical work suggests that the revision of prior knowledge is most likely to occur when newly acquired information makes contact with, and activates, the pre-existing knowledge so that both pieces of information co-activate in working memory (Kendeou et al., 2007; van den Broek & Kendeou, 2008). The actual processes of knowledge revision during reading are exemplified in the KReC framework (Kendeou & O’Brien, 2014). KReC outlines basic comprehension processes and text factors that can be enhanced to increase the potential for successful knowledge revision during reading, by systematically mitigating the interference from pre-existing incorrect knowledge. The main mechanism of knowledge revision is the competing activation between pre-existing incorrect and new correct knowledge structures. As new information is encoded, it is integrated with, and becomes a part of, the pre-existing knowledge base. In the early stage of the knowledge revision process, the knowledge base will be dominated by the incorrect pre-existing knowledge. As the amount and quality of new correct information integrated into the knowledge base increases, it gradually will win over the competition for activation during reading (Kendeou, Smith, & O’Brien, 2013); only after that has occurred will knowledge revision become evident (Kendeou, Walsh, Smith, & O’Brien, 2014).

Theoretical and applied implications The development of comprehension skills is marked by continuity and change. Constructing a representation in which text elements and background knowledge are interconnected is central to comprehension across development, but the component skills and strategies change with development and, as a consequence, so does the nature and quality of the representation. The reviewed research has provided profound insights in the continuities and changes across developmental stages, and presents a solid foundation for further investigations aimed at deepening our theoretical understanding of reading comprehension and comprehension development.

294 Paul van den Broek and Panayiota Kendeou

The theoretical perspectives discussed above also present a foundation for investigations in new research directions. One of those directions involves the neuropsychological investigation of processes and structures involved in reading comprehension. There is a considerable body of research on neurological aspects of letter, word, and sentence level processing in children (Frishkoff, Perfetti, & Collins-Thompson, 2010; Prat & Just, 2011; van Berkum, 2009), but the investigation of the comprehension processes in text is sparse – and the research available is mostly restricted to adult readers (Hagoort & Indefrey, 2014). The adult models suggest that comprehension involves a diversity of neurological structures and functions and that the exact combination of structures and functions varies from point-to-point in the text, as well as across text types (for a review on ‘the extended language network’, see Ferstl, Neumann, Bogler, & von Cramon, 2008). In one of the few developmental studies, children and young adults used remarkably similar networks of brain regions when detecting inconsistencies in texts, although there appeared to be differences in execution of the detection task, with children drawing less on regions typically associated with cognitive control (van Leijenhorst, Karlsson, Helder, & van den Broek, 2014). As the theoretical model of development of comprehension skills becomes more extensive, explorations into developmental changes in the recruitment and interconnectedness of neural structures and functions become possible. Such extensions lead to a better understanding of neural functioning on one hand, and a deeper understanding of comprehension skill development on the other. A second direction fostered by a thorough understanding of reading comprehension skills and their development concerns the design of interventions and other educational applications (see also McMaster and Espin, this volume). Given the centrality of the construction of a coherent representation to comprehension, methods for fostering inference generation have the potential of improving comprehension (e.g., Pressley & McCormick, 1995; Trabasso, van den Broek, & Liu, 1988). Indeed, questioning techniques aimed at eliciting the generation of inferences have proven to be effective in improving comprehension in school-age children (Fricke, Bowyer-Crane, Haley, Hulme, & Snowling, 2013; McMaster et al., 2012; Pressley, Graham, & Harris, 2006; van den Broek, Tzeng, Risden, Trabasso, & Basche, 2001; Yuill & Oakhill, 1988), and even in preschool children (van den Broek, Kendeou, Lousberg, & Visser, 2011; van Kleeck, 2008). In designing such instructional approaches it is important to consider individual and developmental differences between the children that are to receive the intervention. For example, the type of questioning that is most effective depends on the developmental stage of the targeted children. Questioning during reading appears to be very effective in older children (7th grade and up) but less effective in young children (4th grade) and even lead to weaker comprehension in this age group; apparently, the task of having to answer questions burdens the attentional capacities of these young children to the point that they generate fewer inferences than when they are left to their own devices (van den Broek et al., 2001.). As a second example, close inspection of the cognitive processes described in this chapter has shown that struggling readers fall into distinct subgroups, each with its own unique profile of difficulties (McMaster, Espin,



Development of reading comprehension 295

& van den Broek, 2014). Short and efficient interventions can be targeted specifically at each of the subgroups. A complete review of educational applications is beyond the scope of this chapter, but would include the design of instructional materials, the construction of diagnostic tools, and so on (e.g., Vaughn, Roberts, Wexler, Vaughn, Fall, & Schnakenberg, 2014; Vaughn et al., 2013; Williams, 2008). Theory and practical application have a reciprocal relation (Compton, Miller, Elleman, & Steacy, 2014): theoretical insights can inform diagnosis and intervention; conversely, diagnosis and intervention provide an important testing ground for and pose challenges to theoretical models of the development of language comprehension skills. A third new direction for research concerns the extension of the study of reading and comprehension processes to situations beyond single texts and beyond building a meaningful representation. With regard to the former, partly as the result of the availability of near-endless resources on the internet, children (and adults) increasingly are asked to integrate information across multiple texts and multiple formats (Bråten, Britt, Strømsø, & Rouet, 2011). Doing so requires unique skills, but these skills no doubt build to a considerable degree on the development of the skills described above (Alexander, 2012; van den Broek & Kendeou, 2015). With regard to the latter, comprehension does not end with the construction of a complete and coherent representation. As the definition of literacy by OECD quoted at the outset of this chapter indicates, successful reading may include other abilities such as those involved in reflection on the text, application to other situations, and so on. These forms of successful reading build on the mental representation that the reader has created of the text’s meaning. Indeed some aspects may be closely tied to the representation. For example, the theme of a text emerges from the causal/semantic structure with elements with many connections likely forming much of the theme. But other aspects may require considerable cognitive steps beyond creating a representation, for example when a critical reflection or evaluation needs to be given. In conclusion, there is growing understanding of the development of language comprehension skills and their application in the context of reading. Direct investigations of the role of these skills in reading comprehension have shown that this role is considerable, over and above the role of other, basic language skills. These insights are important for an understanding of the development of the uniquely human cognitive ability to utter and comprehend language, but they also provide a foundation for gaining a deeper understanding of neurological functioning and development and for the improvement of educational practice.

References Afflerbach, P. P. (1990). The influence of prior knowledge on expert readers’ main idea construction strategies. Reading Research Quarterly, 25, 31–46.  doi: 10.2307/747986 Alexander, P. A., & Jetton, T. L. (2000). Learning from text: A multidimensional and developmental perspective. Handbook of Reading Research, 3, 285–310.

296 Paul van den Broek and Panayiota Kendeou

Alexander, P. A., & Murphy, P. K. (1998). Profiling the differences in students’ knowledge, interest, and strategic processing. Journal of Educational Psychology, 90, 435–447. doi: 10.1037/0022-0663.90.3.435 Alexander, P. A., Murphy, P. K., Woods, B. S., Duhon, K. E., & Parker, D. (1997). College instruction and concomitant changes in students’ knowledge, interest, and strategy use: A study of domain learning. Contemporary Educational Psychology, 22, 125–146. doi: 10.1006/ceps.1997.0927 Alexander, P. A., & the Disciplined Reading and Learning Research Laboratory (2012). Reading into the future: Competence for the 21st century. Educational Psychologist, 47, 1–22. doi: 10.1080/00461520.2012.722511 Anderson, R. C., & Freebody, P. (1981). Vocabulary knowledge. In J. Guthrie (Ed.), Comprehension and Teaching: Research Reviews (pp. 77–117). Newark, DE: International Reading Association. Anderson, R. C., & Pichert, J. W. (1978). Recall of previously unrecallable information following a shift in perspective. Journal of Verbal Learning and Verbal Behavior, 17, 1–12. doi: 10.1016/S0022-5371(78)90485-1 Baker, L. (1984). Spontaneous versus instructed use of multiple standards for evaluating comprehension: Effects of age, reading proficiency and type of standard. Journal of Experimental Child Psychology, 55, 289–311.  doi: 10.1016/0022-0965(84)90127-9 Barth, A. E., Barnes, M., Francis, D., Vaughn, S., & York, M. (2015). Inferential processing among adequate and struggling adolescent comprehenders and relations to reading comprehension. Reading and Writing, 28, 587–609.  doi: 10.1007/s11145-014-9540-1 Bauer, P. J. (2007). Recall in infancy: A neurodevelopmental account. Current Directions in Psychological Science, 16, 142–146.  doi: 10.1111/j.1467-8721.2007.00492.x Bauer, P. J., & Lukowski, A. F. (2010). The memory is in the details: Relations between memory for the specific features of events and long-term recall during infancy. Journal of Experimental Child Psychology, 107, 1–14.  doi: 10.1016/j.jecp.2010.04.004 Bauer, P. J., & San Souci, P. (2010). Going beyond the facts: Young children extend knowledge by integrating episodes. Journal of Experimental Child Psychology, 107, 452–465. doi: 10.1016/j.jecp.2010.05.012 Beck, I. L., Perfetti, C. A., & McKeown, M. G. (1982). Effects of long-term vocabulary instruction on lexical access and reading comprehension. Journal of Educational Psychology, 74, 506–521.  doi: 10.1037/0022-0663.74.4.506 Beker, K., Jolles, D., Lorch, R. F. Jr., & van den Broek, P. (2016). Learning from texts: Activation of information from previous texts during reading. Reading & Writing, 29, 1161–1178. Bourg, T., Bauer, P. J., & van den Broek, P. (1997). Building the bridges: The development of event comprehension and representation. In P. van den Broek, P. Bauer & T. Bourg (Eds.), Developmental spans in event comprehension and representation: Bridging fictional and actual events (pp. 385–407). Hillsdale, NJ: Erlbaum. Bråten, I., Britt, M. A., Strømsø, H. I., & Rouet, J.-F. (2011). The role of epistemic beliefs in the comprehension of multiple expository texts: Towards an integrated model. Educational Psychologist, 46, 48–70.  doi: 10.1080/00461520.2011.538647 Cain, K. (1999). Ways of reading: How knowledge and use of strategies are related to reading comprehension. British Journal of Developmental Psychology, 17, 293–309. doi: 10.1348/026151099165285 Cain, K., & Oakhill, J. (2012). Reading comprehension development from seven to fourteen years: Implications for assessment. In J. P. Sabatini, E. E. Albro & T. O’Reilly (Eds.), Measuring up: Advances in how to assess reading ability (pp. 59–76). Lanham: Rowman & Littlefield Education.



Development of reading comprehension 297

Cain, K., Oakhill, J. V., Barnes, M. A., & Bryant, P. E. (2001). Comprehension skill, inference-making ability, and their relation to knowledge. Memory & Cognition, 29, 850–859. doi: 10.3758/BF03196414 Cain, K., Oakhill, J., & Bryant, P. (2004). Children’s reading comprehension ability: Concurrent prediction by working memory, verbal ability, and component skills. Journal of Educational Psychology, 96, 31–42.  doi: 10.1037/0022-0663.96.1.31 Catts, H. W., Hogan, T. P., & Fey, M. E. (2003). Subgrouping poor readers on the basis of individual differences in reading-related abilities. Journal of Learning Disabilities, 36, 151–164 doi: 10.1177/002221940303600208 Cerdán, R., & Vidal-Abarca, E. (2008). The effects of tasks on integrating information from multiple documents. Journal of Educational Psychology, 100, 209–222.  doi: 10.1037/0022-0663.100.1.209 Chall, J. S., Jacobs, V. A., & Baldwin, L. E. (1990). The reading crisis: Why poor children fall behind. Cambridge, MA: Harvard. Cook, A. E., & O’Brien, E. L. (2014). Knowledge activation, integration, and validation during narrative text comprehension. Discourse Processes, 51, 26–49.  doi: 10.1080/0163853X.2013.855107 Compton, D. L., Appleton, A. C., & Hosp, M. K. (2004). Exploring the Relationship Between Text‐Leveling Systems and Reading Accuracy and Fluency in Second‐Grade Students Who Are Average and Poor Decoders. Learning Disabilities Research & Practice, 19, 176–184. doi: 10.1111/j.1540-5826.2004.00102.x Compton, D. L., Miller, A. C., Elleman, A. M., & Steacy, L. M. (2014). Have we forsaken reading theory in the name of “quick fix” interventions for children with reading disability? Scientific Studies of Reading, 18, 55–73.  doi: 10.1080/10888438.2013.836200 Cromley, J. G., & Azevedo, R. (2007). Testing and refining the direct and inferential mediation model of reading comprehension. Journal of Educational Psychology, 99, 311–325. doi: 10.1037/0022-0663.99.2.311 Daneman, M., & Carpenter, P. A. (1980). Individual differences in working memory and reading. Journal of Verbal Learning and Verbal Behavior, 19, 450–466. doi: 10.1016/S0022-5371(80)90312-6 Eason, S. H., Goldberg, L. F., Young, K. M., Geist, M. C., & Cutting, L. E. (2012). Reader – text interactions: How differential text and question types influence cognitive skills needed for reading comprehension. Journal of Educational Psychology, 104, 515–528. doi: 10.1037/a0027182 Ehri, L. C., Nunes, S. R., Stahl, S. A., & Willows, D. M. (2001). Systematic phonics instruction helps students learn to read: Evidence from the National Reading Panel’s meta-analysis. Review of Educational Research, 71, 393–447.  doi: 10.3102/00346543071003393 Elbro, C., & Buch-Iversen, I. (2013). Activation of background knowledge for inference making: effects on reading comprehension. Scientific Studies of Reading, 17, 435–452. doi: 10.1080/10888438.2013.774005 Ferreira, F., & Patson, N. (2007). The ‘good enough’ approach to language comprehension. Language and Linguistics Compass, 1, 71–83.  doi: 10.1111/j.1749-818X.2007.00007.x Ferstl, E. C., Neumann, J., Bogler, C., & von Cramon, D. Y. (2008). The extended language network: a meta‐analysis of neuroimaging studies on text comprehension. Human Brain Mapping, 29, 581–593.  doi: 10.1002/hbm.20422 Fricke, S., Bowyer‐Crane, C., Haley, A. J., Hulme, C , & Snowling, M. J. (2013). Efficacy of language intervention in the early years. Journal of Child Psychology and Psychiatry, 54, 280–290. doi: 10.1111/jcpp.12010

298 Paul van den Broek and Panayiota Kendeou

Frishkoff, G. A., Perfetti, C. A., & Collins-Thompson, K. (2010). Lexical quality in the brain: ERP evidence for robust word learning from context. Developmental Neuropsychology, 35, 376–403.  doi: 10.1080/87565641.2010.480915 Fuchs, L. S., Fuchs, D., Hosp, M. K.& Jenkins, J. R. (2001). Oral reading fluency as an indicator of reading competence: A theoretical, empirical, and historical analysis. Scientific Studies of Reading, 5, 239–256.  doi: 10.1207/S1532799XSSR0503_3 Gernsbacher, M. A. (1990). Language comprehension as structure building. Hillsdale, NJ: Erlbaum. Glenberg, A. M., Wilkinson, A. C., & Epstein, W. (1982). The illusion of knowing: Failure in the self-assessment of comprehension. Memory & Cognition, 10, 597–602. doi: 10.3758/BF03202442 Goldman, S. R. (1997). Learning from text: Reflections on the past and suggestions for the future. Discourse Processes, 23, 357–398.  doi: 10.1080/01638539709544997 Goldman, S. R. (2004). Cognitive aspects of constructing meaning through and across multiple texts. In N. Shuart-Faris & D. Bloome (Eds.), Uses of intertextuality in classroom and educational research (pp. 317–351). Greenwich, CT: Information Age. Goldman, S. R., & Varma, S. (1995). CAPping the construction-integration model of discourse comprehension. In C. A. Weaver, S. Mannes, & C. R. Fletcher, (Eds.) Discourse comprehension: Essays in honor of Walter Kintsch, (pp. 337–358). Mahwah NJ: Lawrence Erlbaum Associates. Graesser, A. C., & Clark, L. F. (1985). Structures and procedures of implicit knowledge. Norwood, NJ: Ablex. Graesser, A. C., McNamara, D. S., Louwerse, M. M., & Cai, Z. (2004). Coh-Metrix: Analysis of text on cohesion and language. Behavior Research Methods, Instruments, & Computers, 36, 193–202.  doi: 10.3758/BF03195564 Hagoort, P., & Indefrey, P. (2014). The neurobiology of language beyond single words. Annual Review of Neuroscience, 37, 347–362.  doi: 10.1146/annurev-neuro-071013-013847 Halliday, M. A. K. & Hasan, R. (1976). Cohesion in English. London, UK: Longman. Jenkins, J. J. (1979). Four points to remember: A tetrahedral model of memory experiments. In L. S. Cermak & F. I. Craick (Eds.) Levels of Processing in Human Memory (pp. 429–446). New York, NY: Psychology Press. Kaakinen, J. K., & Hyönä, J. (2005). Perspective effects on expository text comprehension: Evidence from think-aloud protocols, eyetracking and recall. Discourse Processes, 40, 239–257. doi: 10.1207/s15326950dp4003_4 Karasinski, C., & Weismer, S. E. (2010). Comprehension of inferences in discourse processing by adolescents with and without language impairment. Journal of Speech, Language, and Hearing Research, 53, 1268–1279.  doi: 10.1044/1092-4388(2009/09-0006) Kendeou, P. (2015). A general inference skill. In E. J. O’Brien, A. E. Cook, & R. F. Lorch, Jr. (Eds.), Inferences during reading (pp. 160–181) Cambridge: Cambridge University Press. doi: 10.1017/CBO9781107279186.009 Kendeou, P. (2014). Validation and comprehension: An integrated overview. Discourse Processes, 51, 189–200.  doi: 10.1080/0163853X.2013.855874 Kendeou, P., Bohn-Gettler, C., & Fulton, S. (2011). What we have been missing: The role of goals in reading comprehension. In M. T. McCrudden, J. P. Magliano, & G. Schraw (Eds.), Text relevance and learning from text (pp. 375–394). Charlotte, NC: Information Age Publishing, Inc Kendeou, P., Bohn‐Gettler, C., White, M. J., & van den Broek, P. (2008). Children’s inference generation across different media. Journal of Research in Reading, 31, 259–272. doi: 10.1111/j.1467-9817.2008.00370.x



Development of reading comprehension 299

Kendeou, P., Muis, K. R., & Fulton, S. (2011). Reader and text factors in reading comprehension processes. Journal of Research in Reading, 34, 365–383.  doi: 10.1111/j.1467-9817.2010.01436.x Kendeou, P., & O’ Brien, E. J. (2016). Prior knowledge: Acquisition and revision. In P. Afflerbach (Ed.), Handbook of individual differences in reading: Text and context (pp. 151–163) New York: Routledge Publishing. Kendeou, P. & O’Brien, E. J. (2014). The knowledge revision components (KReC) framework: Processes and mechanism. In D. N., Rapp & J. Braasch (Eds.), Processing inaccurate information: theoretical and applied perspectives from cognitive science and the educational sciences. (pp. 353–377). Cambridge, MA: MIT Press. Kendeou, P., Smith, E. R., & O’Brien, E. J. (2013). Updating during reading comprehension: why causality matters. Journal of Experimental Psychology: Learning, Memory, & Cognition, 39, 854–865. Kendeou, P., Walsh, E., Smith, E. R., & O’Brien, E. J. (2014). Knowledge revision processes in refutation texts. Discourse Processes, 51, 374–397.  doi: 10.1080/0163853X.2014.913961 Kendeou, P., & van den Broek, P. (2005). The effects of readers’ misconceptions on comprehension of scientific text. Journal of Educational Psychology, 97, 235–245.  doi: 10.1037/0022-0663.97.2.235 Kendeou, P., & van den Broek, P. (2007). The effects of prior knowledge and text structure on comprehension processes during reading of scientific texts. Memory & Cognition, 35, 1567–1577.  doi: 10.3758/BF03193491 Kendeou, P., van den Broek, P., White, M. J., & Lynch, J. S. (2007). Comprehension in preschool and early elementary children: Skill development and strategy interventions. In D. S. McNamara (Ed.), Reading comprehension strategies: Theories, interventions, and technologies (pp. 27–45). Mahwah, NJ: Erlbaum. Kendeou, P., van den Broek, P., White, M. J., & Lynch, J. S. (2009). Predicting reading comprehension in early elementary school: The independent contributions of oral language and decoding skills. Journal of Educational Psychology, 101, 765.  doi: 10.1037/a0015956 Kim, O., Kendeou, P., van den Broek, P., White, M. J., & Kremer, K. (2008). Cat, Rat, and Rugrats: Narrative comprehension in young children with Down Syndrome. Journal of Developmental and Physical Disabilities, 20, 337–351.  doi: 10.1007/s10882-008-9101-0 Kintsch, W. (1988). The role of knowledge in discourse comprehension: a construction-integration model. Psychological Review, 95, 163–182.  doi: 10.1037/0033-295X.95.2.163 Kintsch, W. (1998). Comprehension: A paradigm for cognition. New York: Cambridge University Press. Kintsch, W., & van Dijk, T. A. (1978). Toward a model of text comprehension and production. Psychological Review, 85, 363–394.  doi: 10.1037/0033-295X.85.5.363 LaBerge, D., & Samuels, S. J. (1974). Toward a theory of automatic information processing in reading. Cognitive Psychology, 6, 293–323.  doi: 10.1016/0010-0285(74)90015-2 Lemarié, J., Lorch, R. F., Jr Eyrolle, H., & Virbel, J. (2008). SARA: A text-based and reader- based theory of signaling. Educational Psychologist, 43, 27–48.  doi: 10.1080/00461520701756321 Linderholm, T., & van den Broek, P. (2002). The effects of reading purpose and working memory capacity on the processing of expository text. Journal of Educational Psychology, 94, 778. doi: 10.1037/0022-0663.94.4.778 Liu, Y., Reichle, E. D., & Gao, D. G. (2013). Using reinforcement learning to examine dynamic attention allocation during reading. Cognitive Science, 37, 1507–1540.  doi: 10.1111/cogs.12027 Lorch, E. P., Berthiaume, K. S., Milich, R., & van den Broek, P. (2007). Story comprehension impairments in children with ADHD. In K. Cain & J. Oakhill (Eds.), Cognitive bases of children’s language comprehension difficulties (pp. 128–156). New York: Guilford.

300 Paul van den Broek and Panayiota Kendeou

Lorch, E. P., Eastham, D., Milich, R., Lemberger, C. C., Sanchez, R. P., Welsh, R., & van den Broek, P. (2004). Difficulties in comprehending causal relations among children with ADHD: The role of cognitive engagement. Journal of Abnormal Psychology, 113, 56–63. doi: 10.1037/0021-843X.113.1.56 Lorch, E. P., Sanchez, R. P., van den Broek, P., Milich, R., Murphy, E. L., Lorch, R. F., Jr. & Welsh, R. (1999). The relation of story structure properties to recall of television stories in young children with attention-deficit hyperactivity disorder and non-referred peers. Journal of Abnormal Child Psychology, 27, 293–309.  doi: 10.1023/A:1022658625678 Lorch, R. F., Jr., Lorch, E. P., & Klusewitz, M. A. (1993). College students’ conditional knowledge about reading. Journal of Educational Psychology, 85, 239–252.  doi: 10.1037/0022-0663.85.2.239 Luna, B., Garver, K. E., Urban, T. E., Lazar, N. A., & Sweeney (2004). Maturation of cognitive processes from late childhood to adulthood. Child Development, 75, 1357–1372. doi: 10.1111/j.1467-8624.2004.00745.x Lynch, J. S., & van den Broek, P. (2007). Understanding the glue of narrative structure: Children’s on-and off-line inferences about characters’ goals. Cognitive Development, 22, 323–340. doi: 10.1016/j.cogdev.2007.02.002 Lynch, J. S., van den Broek, P., Kremer, K. E., Kendeou, P., White, M. J., & Lorch, E. P. (2008). The development of narrative comprehension and its relation to other early reading skills. Reading Psychology, 29, 327–365.  doi: 10.1080/02702710802165416 McCrudden, M. T., Magliano, J. P., & Schraw, G. (2011). Relevance in text comprehension. In M. T. McCrudden, J. P. Magliano & G. Schraw (Eds.), Text relevance and learning from text (pp. 1–17). Charlotte, NC: Information Age Publishing. McKoon, G., & Ratcliff, R. (1992). Inference during reading. Psychological Review, 99, 440–466. doi: 10.1037/0033-295X.99.3.440 McMaster, K. L., Espin, C. A., & van den Broek, P. (2014). Making connections: Linking cognitive psychology and intervention research to improve comprehension of struggling readers. Learning Disabilities Research & Practice, 29, 17–24.  doi: 10.1111/ldrp.12026 McMaster, K. L., van den Broek, P., Espin, C. A., White, M. J., Rapp, D. N., Kendeou, P., Carlson, S. (2012). Making the right connections: Differential effects of reading intervention for subgroups of comprehenders. Learning and Individual Differences, 22, 100–111. doi: 10.1016/j.lindif.2011.11.017 McNamara, D. S., & Kintsch, W. (1996). Learning from texts: Effects of prior knowledge and text coherence. Discourse Processes, 22, 247–288.  doi: 10.1080/01638539609544975 McNamara, D. S., & Magliano, J. (2009). Toward a comprehensive model of comprehension. In B. Ross (Ed.), The psychology of learning and motivation (pp. 297–384). New York: Elsevier. doi: 10.1016/S0079-7421(09)51009-2 McVay, J. C., & Kane, M. J. (2012). Why does working memory capacity predict variation in reading comprehension? On the influence of mind wandering and executive attention. Journal of Experimental Psychology: General, 141, 302–320.  doi: 10.1037/a0025250 Meyer, B. J., & Poon, L. W. (2001). Effects of structure strategy training and signaling on recall of text. Journal of Educational Psychology, 93, 141–159.  doi: 10.1037/0022-0663.93.1.141 Myers, J. L., & O’Brien, E. J. (1998). Accessing the discourse representation during reading. Discourse Processes, 26, 131–157.  doi: 10.1080/01638539809545042 Nagy, W. E., Herman, P. A., & Anderson, R. C. (1985). Learning words from context. Reading Research Quarterly, 20, 233–253.  doi: 10.2307/747758 National Reading Panel (2000). Teaching children to read: An evidence-based assessment of the scientific research literature on reading and its implications for reading instruction. Available: http://www.nichd.nih.gov/publications/nrp/report.cfm



Development of reading comprehension 301

Oakhill, J. (1984). Inferential and memory skills in children’s comprehension of stories. British Journal of Educational Psychology, 54, 31–39.  doi: 10.1111/j.2044-8279.1984.tb00842.x Oakhill, J. V., & Cain, K. (2012). The precursors of reading ability in young readers: Evidence from a four-year longitudinal study. Scientific Studies of Reading, 16, 91–121. doi: 10.1080/10888438.2010.529219 O’Brien, E. J., & Myers, J. L. (1999). Text comprehension: A view from the bottom up. In S. R. Goldman, A. C. Graesser & P. van den Broek (Eds.), Narrative comprehension, causality, and coherence: Essays in honor of Tom Trabasso (pp. 35–53). Mahwah, NJ: Erlbaum. Organisation for Economic Co-operation and Development (OECD) (2009). PISA 2009 assessment framework: Key competencies in reading, mathematics and science. Paris, France: OECD. Retrieved from http://www.oecd.org/pisa/pisaproducts/pisa2009assessmentframeworkkeycompetenciesinreadingmathematicsandscience.htm Pearson, P. D., & Hamm, D. N. (2005). The assessment of reading comprehension: A review of practices – Past, present, and future. In S. G. Paris & S. A. Stahl (Eds.), Children’s reading comprehension and assessment (pp. 13–69). Mahwah, NJ: Erlbaum. Perfetti, C. A., (2007). Reading ability: Lexical quality to comprehension. Scientific Studies of Reading, 11, 357–383.  doi: 10.1080/10888430701530730 Perfetti, C. A., & Adlof, S. M. (2012). Reading comprehension: A conceptual framework from word meaning to text meaning. In J. P. Sabatini, E. R. Albro & T. O’Reilly (Eds.), Measuring up: Advances in how to assess reading ability (pp. 3–20). Lanham: Rowman & Littlefield Education. Perfetti, C. A., & Hart, L. (2001). The lexical bases of comprehension skill. In D. Gorfien (Ed.), On the consequences of meaning selection (pp. 67–86). Washington, DC: American Psychological Association. Pike, M., Barnes, M., & Barron, R. W. (2010). The role of illustrations in children’s inferential comprehension. Journal of Experimental Child Psychology, 105, 243–255. doi: 10.1016/j.jecp.2009.10.006 Prat, C. S., & Just, M. A. (2011). Exploring the neural dynamics underpinning individual differences in sentence comprehension. Cerebral Cortex, 21, 1747–1760. doi: 10.1093/cercor/bhq241 Pressley, M., Graham, S., & Harris, K. (2006). The state of educational intervention research as viewed through the lens of literacy intervention. British Journal of Educational Psychology, 76, 1–19.  doi: 10.1348/000709905X66035 Pressley, M., & McCormick, C. (1995). Advanced educational psychology for educators, researchers, and policymakers. New York, NY: Harper Collins. Rawson, K. A., Dunlosky, J., & Thiede, K. W. (2000). The rereading effect: Metacomprehension accuracy improves across reading trials. Memory & Cognition, 28, 1004–1010. doi: 10.3758/BF03209348 Redford, J. S., Thiede, K. W., Wiley, J., & Griffin, T. D. (2012). Concept mapping improves metacomprehension accuracy among 7th graders. Learning and Instruction, 22, 262–270. doi: 10.1016/j.learninstruc.2011.10.007 Ross, M. E., Green, S. B., Salisbury-Glennon, J. D., & Tollefson, N. (2006). College students’ study strategies as a function of testing: An investigation into metacognitive self-regulation. Innovative Higher Education, 30, 361–375.  doi: 10.1007/s10755-005-9004-2 Rouet, J. F. (2006). The skills of document use: From text comprehension to web-based learning. New York: Psychology Press.

302 Paul van den Broek and Panayiota Kendeou

Rouet, J. F., Vidal-Abarca, E., Bert-Erboul A. & Millogo, V. (2001). Effects of information search tasks on comprehension of instructional texts. Discourse Processes, 21, 163–186. doi: 10.1207/S15326950DP3102_03 Sabatini, J., O’Reilly, T., Halderman, L. & Bruce, K. (2014). Integrating scenario-based and component reading skill measures to understand the reading behavior of struggling readers. Learning Disabilities Research & Practice, 29, 36–43.  doi: 10.1111/ldrp.12028 Schraw, G., Wade, S. E., & Kardash, C. A. (1993). Interactive effects of text-based and task-based importance on learning from text. Journal of Educational Psychology, 85, 652–661. doi: 10.1037/0022-0663.85.4.652 Sesma, H. W., Mahone, E. M., Levine, T., Eason, S. H., & Cutting, L. E. (2009). The contribution of executive skills to reading comprehension. Child Neuropsychology, 15, 232–246. doi: 10.1080/09297040802220029 Siegler, R. S. (1996). Emerging Minds: The process of change in children’s thinking. Oxford, UK: Oxford University Press. Singer, M. (2013). Validation in reading comprehension. Current Directions in Psychological Science, 22, 361–366.  doi: 10.1177/0963721413495236 Snow, C. E. (2002). Reading for understanding:Towards an R & D program in reading comprehension. Santa Monica, CA: RAND Corporation. Snow, C. E., & Sweet, A. P. (2003). Reading for comprehension. In A. P. Sweet & C. E. Snow (Eds.), Rethinking reading comprehension, (pp. 1–11). New York: Guilford Press. Stanovich, K. E. (1986). Matthew effects in reading: Some consequences of individual differences in the acquisition of literacy. Reading Research Quarterly, 21, 360–407.  doi: 10.1598/RRQ.21.4.1 Storch, S. A., & Whitehurst, G. J. (2002). Oral language and code-related precursors to reading: evidence from a longitudinal structural model. Developmental Psychology, 38, 934–947. doi: 10.1037/0012-1649.38.6.934 Swanson, H. L., & O’Connor, R. (2009). The role of working memory and fluency practice on the reading comprehension of students who are dysfluent readers. Journal of Learning Disabilities, 42, 548–575.  doi: 10.1177/0022219409338742 Tannenbaum, K. R., Torgesen, J. K., & Wagner, R. K. (2006). Relationships between word knowledge and reading comprehension in third-grade children. Scientific Studies of Reading, 10, 381–398.  doi: 10.1207/s1532799xssr1004_3 Thelen, E., & Bates, E. (2003). Connectionism and dynamic systems: Are they really different? Developmental Science, 6, 378–391.  doi: 10.1111/1467-7687.00294 Thompson, J. G., & Myers, N. A. (1985). Inferences and recall at ages four and seven. Child Development, 56, 1134–1144.  doi: 10.2307/1130228 Thurlow, R., & van den Broek, P. (1997). Automaticity and inference generation during reading comprehension. Reading & Writing Quarterly, 13, 165–181.  doi: 10.1080/1057356970130205 Tilstra, J., & McMaster, K. L. (2013). Cognitive processes of middle grade readers when reading expository text with an assigned goal. Learning and Individual Differences, 28, 66–74. doi: 10.1016/j.lindif.2013.09.011 Tilstra, J., McMaster, K., van den Broek, P., Kendeou, P., & Rapp, D. N. (2009). Simple but complex: Components of the simple view of reading across grade levels. Journal of Research in Reading, 32, 383–401.  doi: 10.1111/j.1467-9817.2009.01401.x Trabasso, T., Secco, T., & van den Broek, P. (1984). Causal cohesion and story coherence. In H. Mandl, N. L. Stein & T. Trabasso (Eds.), Learning and comprehension of text (pp. 83–111). Hillsdale, NJ: Erlbaum. Trabasso, T., & van den Broek, P. (1985). Causal thinking and the representation of narrative events. Journal of Memory and Language, 24, 612–630.  doi: 10.1016/0749-596X(85)90049-X



Development of reading comprehension 303

Trabasso, T., van den Broek, P., & Liu, L. (1988). A model for generating questions that assess and promote comprehension. Questioning Exchange, 2, 25–38 Trabasso, T., van den Broek, P., & Suh, S. Y. (1989). Logical necessity and transitivity of causal relations in stories. Discourse Processes, 12, 1–25.  doi: 10.1080/01638538909544717 van Berkum, J. J. A. (2009). The neuropragmatics of ‘simple’ utterance comprehension: An ERP review. In U. Sauerland, & K. Yatsushiro (Eds.), Semantics and pragmatics: From experiment to theory (pp. 276–316). Basingstoke: Palgrave Macmillan. van den Broek, P. (1988). The effects of causal relations and hierarchical position on the importance of story statements. Journal of Memory and Language, 27, 1–22. doi: 10.1016/0749-596X(88)90045-9 van den Broek, P. (1989). Causal reasoning and inference making in judging the importance of story statements. Child Development, 60, 286–297.  doi: 10.2307/1130976 van den Broek, P. (1990). Causal inferences maker: Towards a process model of inference generation in text comprehension. In D. A. Balota, G. B. F. d’Arcais & K. Rayner (Eds.), Comprehension processes in reading (pp. 423–446). Hillsdale, NJ: Lawrence Erlbaum Associates. van den Broek, P. (1994). Comprehension and memory of narrative texts: Inferences and coherence. In M. A. Gernsbacher (Ed.), Handbook of psycholinguistics (pp. 539–588). New York: Academic Press. van den Broek, P. (1997). Discovering the cement of the universe: The development of event comprehension from childhood to adulthood. In P. van den Broek, P. J. Bauer & T. Bourg (Eds.), Developmental spans in event comprehension and representation: Bridging fictional and actual events (pp. 321–342). Hillsdale, NJ: Erlbaum. van den Broek, P. (2010). Using texts in science education: Cognitive processes and knowledge representation. Science, 328, 453–456.  doi: 10.1126/science.1182594 van den Broek, P., Beker, K., & Oudega, M. (2015). Inference generation in text comprehension: Automatic and strategic processes in the construction of a mental representation. In E. J. O’Brien, A. Cook & R. F. Lorch (Eds.), Inferences during reading (pp. 99–121). Cambridge, UK: Cambridge University Press.  doi: 10.1017/CBO9781107279186.006 van den Broek, P., Bohn-Gettler, C., Kendeou, P., Carlson, S., & White, M. J. (2011). When a reader meets a text: The role of standards of coherence in reading comprehension. In M. T. McCrudden, J. P. Magliano & G. Schraw (Eds.), Text relevance and learning from text (pp. 123–139). Greenwich, CT: Information Age Publishing. van den Broek, P., Fletcher, C. R., & Risden, K. (1993). Investigations of inferential processes in reading: A theoretical and methodological integration. Discourse Processes, 16, 169–180. doi: 10.1080/01638539309544835 van den Broek, P., Helder, A., & van Leijenhorst, L. (2013). Sensitivity to Structural Centrality: Developmental and individual differences in reading comprehension skills. In M. A. Britt, S. R. Goldman & J.-F. Rouet (Eds.), Reading: From words to multiple texts (pp. 132–146). New York: Routledge. van den Broek, P., & Kendeou, P. (2008). Cognitive processes in comprehension of science texts: The role of co-activation in confronting misconceptions. Applied Cognitive Psychology, 22, 335–351.  doi: 10.1002/acp.1418 van den Broek, P., & Kendeou, P. (2015). Building coherence in web-based and other non-traditional reading environments; Cognitive opportunities and challenges. In J. J. Spiro, M. Deschryver, M. S. Hagerman, P. Morsink & P. Thompson (Eds.), Reading at a crossroads? Disjunctures and continuities in current conceptions and practices (pp. 104–114) New York: Routledge.

304 Paul van den Broek and Panayiota Kendeou

van den Broek, P., Kendeou, P., Lousberg, S., & Visser, G. (2011). Preparing for reading comprehension: Fostering text comprehension skills in preschool and early elementary school children. International Electronic Journal of Elementary Education, 4, 259–268. van den Broek, P., Lorch, R. F., Linderholm, T., & Gustafson, M. (2001). The effects of readers’ goals on inference generation and memory for texts. Memory & Cognition, 29, 1081–1087. doi: 10.3758/BF03206376 van den Broek, P., Lorch, E. P., & Thurlow, R. (1996). Children’s and adults’ memory for television stories: The role of causal factors, story‐grammar categories, and hierarchical level. Child Development, 67, 3010–3028.  doi: 10.2307/1131764 van den Broek, P., Lynch, J. S., Naslund, J., Ievers-Landis, C. E., & Verduin, K. (2003). The Development of comprehension of main ideas in narratives: Evidence from the selection of titles. Journal of Educational Psychology, 95, 707–718.  doi: 10.1037/0022-0663.95.4.707 van den Broek, P., Rapp, D. N., & Kendeou, P. (2005). Integrating memory-based and constructionist processes in accounts of reading comprehension. Discourse Processes, 39, 299–316. doi: 10.1080/0163853X.2005.9651685 van den Broek, P., Risden, K., Fletcher, C. R., & Thurlow, R. (1996). A “landscape” view of reading: Fluctuating patterns of activation and the construction of a stable memory representation. In B. K. Britton & A. C. Graesser (Eds.), Models of understanding text (pp. 165–187). Hillsdale, NJ: Erlbaum. van den Broek, P., Risden, K., & Husebye-Hartmann, E. (1995). The role of readers’ standards for coherence in the generation of inferences during reading. In J. R. F. Lorch & E. J. O’Brien (Eds.), Sources of coherence in reading (pp. 353–373). Hillsdale, NJ: Erlbaum. van den Broek, P., & Trabasso, T. (1986). Causal networks versus goal hierarchies in summarizing text. Discourse Processes, 9, 1–15.  doi: 10.1080/01638538609544628 van den Broek, P., Tzeng, Y., Risden, K., Trabasso, T., & Basche, P. (2001). Inferential questioning: Effects on comprehension of narrative texts as a function of grade and timing. Journal of Educational Psychology, 93, 521–529.  doi: 10.1037/0022-0663.93.3.521 van den Broek, P., Young, M., Tzeng, Y., & Linderholm, T. (1999). The landscape model of reading: Inferences and the online construction of a memory representation. In H. van Oostendorp & S. R. Goldman (Eds.), The construction of mental representations during reading (pp. 71–98). Mahwah, NJ: Erlbaum. (Reprinted in R. B. Ruddell & N. J. Unrau (Eds.), Theoretical models and processes of reading (1244–1269). Newark, NJ: International Reading Association). van Geert, P. (1991). A dynamic systems model of cognitive and language growth. Psychological Review, 98, 3–53.  doi: 10.1037/0033-295X.98.1.3 van Kleeck, A. V. (2008). Providing preschool foundations for later reading comprehension: The importance of and ideas for targeting inferencing in storybook‐sharing interventions. Psychology in the Schools, 45, 627–643.  doi: 10.1002/pits.20314 van Leijenhorst, L., Karlsson, J., Helder, A., & van den Broek, P. (2014, July). Neural correlates of revising a situation model for sentences with multiple events in children and young adults. Presentation at the biannual meeting of International Association for the Study of Child Language, Amsterdam, The Netherlands. Vaughn, S., Roberts, G., Wexler, J., Vaughn, M., Fall, A.-M., & Schnakenberg, J. B. (2014). High school students with reading comprehension difficulties: Results of a randomized control trial of a two-year reading intervention. Journal of Learning Disabilities, 48, 546–558. doi: 10.1177/0022219413515511



Development of reading comprehension 305

Vaughn, S., Swanson, E. A., Roberts, G., Wanzek, J., Stillman-Spisak, S. J., Solis, M., & Simmons, D. (2013) Improving reading comprehension and social studies knowledge in middle school. Reading Research Quarterly, 48, 77–93. Verhoeven, L.,& Perfetti, C. A. (Eds.) (2011). Vocabulary growth and reading skill. Scientific Studies of Reading, 15, 1–108.  doi: 10.1080/10888438.2011.536124 Vosniadou, S., & Brewer, W. F. (1994). Mental models of the day/night cycle. Cognitive Science, 18, 123–183.  doi: 10.1207/s15516709cog1801_4 Wagner, R. K., Torgesen, J. K., Rashotte, C. A., Hecht, S. A., Barker, Th. A., Burgess, S. R., Donahue, J., & Garon, T. (1997). Changing relations between phonological processing abilities and word-level reading as children develop from beginning to skilled readers: A 5-year longitudinal study. Developmental Psychology, 33, 468–479.  doi: 10.1037/0012-1649.33.3.468 Whitehurst, G. J., & Lonigan, C. J. (1998). Child development and emergent literacy. Child Development, 69, 848–872.  doi: 10.1111/j.1467-8624.1998.tb06247.x Williams, J. P. (1993). Comprehension of students with and without learning disabilities: Identification of themes and idiosyncratic text representation. Journal of Educational Psychology, 85, 631–641.  doi: 10.1037/0022-0663.85.4.631 Williams, J. P. (2008). Explicit instruction can help primary students learn to comprehend expository text. In C. C. Block & S. Parris (Eds.), Comprehension instruction: Research-based best practices 2nd edition (pp. 171–182). New York: Guilford Press. Williams, J. P., Hall, K. M., Lauer, K. D., Stafford, K. B., De Sisto, L. A., & deCani, J. S. (2005). Expository text comprehension in the primary grade classroom. Journal of Educational Psychology, 97, 538–550.  doi: 10.1037/0022-0663.97.4.538 Williams, J. P., Lauer, K. D., Hall, K. M., Lord, K. M., Gugga, S. S., Bak, S. J., Jacobs, P. R., & deCani, J. S. (2002). Teaching elementary school students to identify story themes. Journal of Educational Psychology, 94, 235–248.  doi: 10.1037/0022-0663.94.2.235 Wolman, C., van den Broek, P., & Lorch, R. F. (1997). Effects of causal structure on immediate and delayed story recall by children with mild mental retardation, children with learning disabilities, and children without disabilities. The Journal of Special Education, 30, 439–455.  doi: 10.1177/002246699703000405 Yeari, M., van den Broek, P., & Oudega, M. (2015). Processing and memory of central versus peripheral information as a function of reading goals: Evidence from eye-movements. Reading and Writing, 28, 1071–1097. Yuill, N. & Oakhill, J. (1988). Effects of inference awareness training on poor reading comprehension. Applied Cognitive Psychology, 2, 33–45.  doi: 10.1002/acp.2350020105

Part IV

Atypical reading development

Introduction to atypical reading development Rauno K. Parrila, Donald L. Compton and Kate Cain

University of Alberta / Florida State University / Lancaster University

Although the majority of children acquire adequate word reading and reading comprehension skills across various educational contexts and instructional methods, a minority of children fail to do so. For these children, poor reading can impact their educational attainment, employment prospects, health outcomes, and social lives. The three chapters in this section deal with the atypical development of both word reading and reading comprehension, each examining the underpinning skills and impacts for theory and practice. Catts’s chapter covers both word reading and reading comprehension disabilities. He argues persuasively that we must consider the early factors related to reading disability to enable early identification and intervention, rather than wait until children have experienced years of reading instruction and failed to make adequate progress. Catts first examines the research on family history and deficits in early language development that could be used to identify children at risk of later difficulties in learning to read. The language factors considered (late talkers, specific language impairment, phonological processing deficits) are each risk factors, which he argues should be considered in relation to other variables for early diagnosis; however, he finds little support for non-linguistic factors as primary causes of reading disability. Catts then considers this work in the context of the risk-resilience framework, and how good instruction or intervention may moderate the effects of risk, in addition to individual characteristics such as attention, engagement, and grit. He concludes with advocating an approach that combines information from screening for risk factors with dynamic assessment approaches and with examination of how children respond to multi-tiered interventions. The second chapter by Parrila and Protopapas focuses on developmental dyslexia and word reading problems in alphabetic languages. They first discuss difficulties in defining dyslexia as a behavioural disorder and settle on a definition with minimal exclusion and no inclusion criteria. They then review various single-deficit and subtyping theories and conclude that none of these is likely to provide satisfactory explanations of dyslexia as a behaviorally defined developmental disorder. They then explore recent multiple deficit theories and how an interaction between a range of risk and protective factors (genetic, neural, cognitive, behavioral, and environmental) may provide multiple developmental pathways to high or low reading performance. doi 10.1075/swll.15.17par © 2017 John Benjamins Publishing Company

310 Rauno K. Parrila, Donald L. Compton and Kate Cain

In this way, the chapter complements Catts’ analysis of multiple risk factors. A critical innovation in their analysis is the interpretation of multiple deficit theories as developmental systems theories; Parrila and Protopapas conclude that as developmental systems theories, theories of dyslexia will need to move beyond describing individual differences in performance to explanations of various developmental pathways to inaccurate or inefficient word reading. In the final chapter, Oakhill and Cain focus on children who experience reading comprehension difficulties in the presence of age-appropriate word reading skills. They first examine models of (skilled) text comprehension and conclude that no single theory or model to date provides an adequate account of this ability. Instead, they argue, an analysis of component skills that support text comprehension might provide a more productive means to understand why some children fail to develop adequate text comprehension. Their analysis focuses on two core components – vocabulary knowledge and inference making. Vocabulary is shown to be essential for going beyond words and sentences to make connections between elements in a text and to support inferences that are essential for coherence; inference making is shown to support vocabulary acquisition.

Early identification of reading disabilities Hugh W. Catts

Florida State University

Given appropriate opportunity and experience most children learn to read in the early school grades and go on to use their reading skills for educational and recreational purposes. A small proportion of children, however, experience significant difficulties learning to read. These difficulties often in turn lead to a host of negative consequences including academic failure, poor self-concept, substance abuse, truancy, delinquency, and limited employment opportunity (Beitchman, Wilson, Douglas, Young, & Adlaf, 2001; Kirk & Reid, 2001; Spear-Swerling & Sternberg, 1996). Fortunately, research indicates that the severity of reading problems and associated negative consequences can be reduced with early intervention (National Reading Panel, 2000). However, for early intervention to take place, children must be identified in a timely fashion. Because the primary symptom of RD is difficulty learning to read, practitioners and educators have typically had to wait until adequate reading instruction has been provided before a diagnosis could be made. This practice often has delayed identification until Grade 2 or later. Fortunately, recent research has begun to uncover early factors related to RD as well as educational practices that allow practitioners and educators to identify children at risk for RD prior to, or at the very least, the beginning of formal reading instruction. This chapter will consider what we know about the early identification of RD. Most of the research has focused on the early identification of word reading problems and less attention has been given to difficulties in reading comprehension. Nevertheless, what evidence that is relevant to the latter problems will be discussed.

Family history Perhaps the earliest indicator of risk for RD is a family history of reading difficulties. Research shows that approximately 40–60% of children with a parent or sibling with RD will have reading problems themselves (Gilger, Hanebuth, Smith, & Pennington, 1996; Scarborough, 1990; Snowling, Gallagher, & Frith, 2003). This is consistent with what is known about the genetic etiology of reading and reading difficulties. There is now strong evidence from twin studies and association studies for the role of genes in the development of reading (Elliott & Grigorenko, 2014; Olson, Keenan, Byrne, & doi 10.1075/swll.15.18cat © 2017 John Benjamins Publishing Company

312 Hugh W. Catts

Sanuelsson, 2014, see also Olson et al., this volume). Currently, over 20 genes or gene locations have been linked with RD and many more are expected to be identified in the future. Each of these genes may have a small-to-moderate effect on the development of reading and it is the combination of multiple genes and various environmental conditions that leads to RD (Elliot & Grigorenko, 2014). Research further demonstrates that family risk for RD is continuous as opposed to discrete. That is, not only do affected family members have RD but many of those without RD nevertheless, have mild difficulties in reading (Boets, De Smelt, Cleuren, Vandewalle, Wouters, & Ghesquiere, 2010; Dandache, Wouters, & Ghesquire, 2014; Pennington & Lefty, 2001; Snowling, Muter, Carrol, 2007; van Bergen, de Jong, Plakas, Maassen, & van der Leij, 2012). Studies also show that family risk for RD is associated with numerous other reading-related problems. For example, children with family risk for RD often have problems in the development of oral language (Lyytinen, Eklund, & Lyytinen, 2005; Scaborough, 1990, Snowling et al., 2003). In a recent study, Nash, Hulme, Gooch, & Snowling (2013) reported that nearly 30% of preschool children with family risk for dyslexia met the criteria for having a specific language impairment (SLI). In addition to SLI, difficulties in phonological awareness, verbal memory, speech production, and rapid naming have been reported (Carroll, Mundy, & Cunningham, 2014; Elbro, Borstrom, & Petersen, 1998; Nash et al., 2013; Pennington & Lefty, 2001; van Bergen et al., 2012). As with literacy skills, some of these secondary symptoms also appear to have a more continuous distribution. For example, some have reported phonological awareness to be impaired in both affected and unaffected siblings with a family risk of RD (Boets et al., 2010; Snowling et al., 2003; van Bergen et al., 2012). Interestingly, these studies further show that those affected with RD may have oral language or rapid naming problems not seen in their unaffected siblings (Moll, Loff, & Snowling, 2013; Pennington & Lefty, 2001; van Bergen et al., 2011). Such results are consistent with the view that multiple genes interact with the environment to cause RD. Finally, research suggests that the impact of family history on literacy may go well beyond the association with early problems in language development. Puolakanaho et.al (2007) and Carroll et al. (2014) found that family history accounted for future reading abilities over and above early phonological and/or language abilities. Thus, the familial mechanism that underlies RD likely involves other genetic and environmental components that act on cognitive abilities than those typically examined in family studies (also see van Bergen, Bishop, Zuijen, & de Jong, 2015). Finally, family history may also carry with it information about the environment. Siblings often share the same environment and this environment interacts with genetic predispositions to influence reading development. In summary, these various results indicate that if a family member has RD, or a history of RD, it should be considered an important factor in evaluating a child’s risk for RD.



Early identification of reading disabilities 313

Early deficits in language development As noted above, early deficits in the development of oral language are often associated with subsequent difficulties in learning to read. Theoretically, there should be a close link between oral language development and reading, especially reading comprehension. The Simple View of Reading, a theoretical view that has guided much research in reading in recent years, proposes that reading comprehension is the product of word recognition and language (or listening) comprehension (Gough & Tunmer, 1986; Hoover & Gough, 1990). Numerous studies have provided support for the Simple View of Reading and the role of listening comprehension in reading comprehension (Aaron, Joshi, & Williams, 1999; Catts, Hogan, & Adlof, 2005; de Jong & van der Liej, 2002; Hoover & Gough, 1990; Language and Reading Research Consortium, 2015). Some studies have taken a developmental approach and have shown a strong link between early oral language abilities and reading comprehension (Kendeou, van den Broek, White, & Lynch, 2009; Storch & Whitehurst, 2002). Most have proposed a direct link between emerging language abilities and subsequent skills in reading comprehension, whereas others have provided support for an indirect route through listening comprehension (Kim, 2015). From a theoretical perspective, early language abilities should also be associated with the development of word reading abilities. Specifically, the lexical quality hypothesis (Perfetti & Hart, 2002) argues that reading is dependent on high quality representations that include knowledge of semantic-syntactic information. The latter information should be particularly important to word reading when the orthography or phonology is of lower quality. For example, this would be the case when a word has a less transparent spelling or when the same phonological form may have multiple orthographic representations (i.e., homophones).

Late talkers One of the earliest indicators of problems in oral language, and in some cases, subsequent reading problems is failure to begin talking at the appropriate developmental stage. Most children acquire their first spoken words by one year of age and go on to produce numerous words and two-word combinations by 18 months (Owens, 2000). A small percentage of children, on the other hand, demonstrate delayed expressive language. These children, sometimes referred to as late talkers, produce few words and/or no two combinations by two years of age but show adequate development in other domains (Paul, Murray, Clancy, & Andrews, 1997; Rescorla, 2002; Preston et al., 2010). Follow-up investigations have shown that children who were late talkers are generally less skilled than typical children in reading and spelling throughout the school years (Lyytinen et al., 2005; Preston et al., 2010; Rescorla, 2002). Preston and colleagues (2010) also reported that these children were approximately 4 times more likely to be diagnosed with RD than were children who were not late talkers. Despite

314 Hugh W. Catts

the increased risk as a group, most late talkers do not have severe enough reading/ spelling problems to be identified as having RD unless additional risk factors are present. Specifically, Lyytinen et al. (2005) has shown that when early expressive language delays are combined with problems in receptive language and/or family history of RD, children frequently have the most significant deficits in reading and spelling. Thus, late talking can be considered a significant early indicator of RD, when combined with these other risk factors.

Specific language impairment Another developmental language problem associated with later RD is a specific language impairment (SLI). Children with SLI have deficits in vocabulary and/or grammar, but normal nonverbal cognitive abilities (Leonard, 2014; Tomblin & Nippold, 2014). SLI is typically identified in children between three and five years of age and generally persists throughout the school years (Tomblin, Nippold, Fey, & Zhang, 2014). Most children with SLI also have a history of late talking. Numerous studies have documented that children with SLI not only continue to have language problems but frequently experience reading disabilities during the school years (Bishop & Adams, 1990; Catts, 1993; Catts, Fey, Tomblin, & Zhang, 2002; Catts, Adlof, Hogan, & Weismer, 2005). Catts et al. (2005) specifically examined the association between SLI and dyslexia in an epidemiologic sample and showed that 18–36% of children with SLI in kindergarten had dyslexia in the school years. The percentage, which was 2–3 times that of children with typical language development, varied depending on the definition of dyslexia and the grade at which it was identified. Further analyses, showed that children with SLI were also more likely to be dyslexic if they had deficits in nonword repetition (Catts & Adlof, 2011; also see Kelso, Fletcher, & Lee, 2007; McArthur & Castles, 2013). Others have reported that children with SLI have a higher incidence of dyslexia if they also perform poorly in rapid serial naming (Bishop McDonald, Bird, & Hayiou-Thomas, 2009; McArthur & Castles, 2013). Given the significant deficits that children with SLI often have in spoken language comprehension, it is not surprising that many of these children also have problems in reading comprehension (Bishop et al., 2009; Catts et al., 2002). Catts et al. (2002), again using an epidemiologic sample, showed that about 40% of children with SLI in kindergarten had deficits in reading comprehension in second and fourth grades. Of the remaining children with SLI, most had at least mild problems in reading comprehension. Further analyses examined children with language impairments and non-verbal cognitive deficits. These children with non-specific language impairments (NLI) had IQ scores above 70 and are sometimes included within the definition of SLI (TagerFlusberg & Cooper, 1999). Results showed that the majority of these children (approximately 65%) had deficits in reading comprehension in second and fourth grades. In addition to examining reading problems in children with language impairments, researchers have taken the opposite perspective and investigated language



Early identification of reading disabilities 315

problems in children identified with specific deficits in reading comprehension. The latter children, sometimes referred to as poor comprehenders, are typically reported to have a history of problems in language development, with approximately 30% having severe enough difficulties to be diagnosed with SLI (Catts, Adlof, & Weismer, 2006; Nation Cocksey, Taylor, & Bishop, 2010). Taken together, this work indicates that the presence of SLI (and especially NLI) should be considered an early risk factor for RD. Early language intervention, therefore, could not only be critical for addressing the spoken language problems of these children but could assist in reducing the reading problems that many of these children may develop.

Phonological processing deficit The above sections concerned the risk for RD by children with early language deficits primarily in vocabulary and grammar. However, much of the attention directed toward language problems in children with RD has focused on the phonological deficits hypothesis (Brady & Shankweiler, 1991; Stanovich, 1988). The basic tenant of this hypothesis is that many children with RD have deficits in the storage and/or retrieval of phonological information, and this makes it difficult for them to develop links between spoken words and orthographic spellings. There is now considerable support for this hypothesis and it is a mainstay of the literature in dyslexia (Elliott & Grigorenko, 2014; Kamhi & Catts, 2012). Research has focused on a variety deficits in the phonological domain including problems in phonological awareness, phonological memory, and rapid automatized naming (Melby-Lervåg, Lyster, & Hulme, 2012; Wagner & Torgesen, 1987; Wolf & Bowers, 1999). Difficulties in phonological production have also sometimes been observed (Catts, 1989; Snowling, 1981). Some have argued that these deficits stem from limitations in the quality of underlying phonological representations (Elbro et al., 1998), whereas others have proposed that these problems are due to retrieval or output deficits (Hulme & Snowling, 1992; Ramus & Ahissar, 2012). Most of the attention has been focused on phonological awareness, one’s ability to reflect on and/or manipulate the speech sounds of language. The relationship between phonological awareness and word reading is complex in that one’s phonological awareness both influences the development of word reading and is impacted by this development (Castles & Coltheart, 2004). This reciprocal relationship has likely inflated the correlation between phonological awareness and reading in the early school grades. However, there is still strong evidence that preschool deficits in phonological awareness are associated with and may be a causal factor in subsequent word reading problems (Boets et al., 2010; Dandache, et al., 2014; Lyytinen, Erskine, Tolvanen, Torppa, Poikkeus, & Lyytinen, 2006). Furthermore, as noted above, studies of children with a family history of RD, find problems in phonological awareness to be among the most common deficits in affected siblings (Boets et al., 2006; Lyytinen et al. 2006). However, phonological awareness deficits are also often found among unaffected siblings, leading some to argue that these deficits are an endophenotype of RD (Boets et

316 Hugh W. Catts

al., 2010; Moll, Loff, & Snowling, 2013; van Bergen et al., 2012). An endophenotype is a heritable trait that is associated with a disorder or illness but is state-independent, that is, it may be present in both affected and unaffected family members (Bidwell, Willcutt, deFries, & Pennington, 2007). Others have also reported that some children with SLI may have early deficits in phonological awareness but not subsequent word reading difficulties (Catts & Adlof, 2011; Bishop et al., 2009). Thus, these various results suggest that whereas phonological awareness deficits are often associated with RD, by themselves they may not be the most reliable early indicator of RD (see also Catts, McIlraith, Nielsen, & Bridges, 2017). Nevertheless, because of their link to RD, these deficits still could provide useful information for early identification especially when found in combination with other language problems.

Nonlinguistic precursors of RD Research has also provided growing evidence for potential risk factors that go beyond oral language and phonological processing. For example, both behavioral and neurophysiological evidence has shown a connection between RD and problems in a major visual pathway in the central nervous system (Stein, 2001). This pathway, the magnocellular pathway, is involved in visual sequencing, visual motor control, and other aspects of vision. It is proposed that deficits in this pathway disrupt the temporal processing of printed words and lead to difficulties in the accuracy and/or fluency of word reading. Other visual processing problems have also been associated with RD. Specifically, researchers have argued that some individuals may have a sluggish visual attentional system that makes it difficult for them to disengage when presented with a sequence of letters (Franceschini, Gori, Ruffino, Pedrolli, & Facoetti, 2012; Lallier et al., 2010). Others have proposed that children with RD have limitations in visual attention span that reduce the number of letters/words that can be processed at a given time (Prado, Dubois & Valdois, 2007). Whereas there is a large and converging body of evidence indicating a connection between various visual problems and RD, it remains unclear the extent to which these deficits play a causal role in RD. It may be that for a small percentage of those with RD, visual deficits are a primary causal factor in the disorder or are a moderator for other primary risk factors. For others, however, visual deficits may simply co-occur with more primary risk factors and contribute little to the RD (Olulade, Napoliello, Eden, 2013). As work in this area proceeds, we may find stronger evidence for the causal basis for specific visual deficits and clearer implications for the role of these in early identification and intervention. Other research suggests that deficits in motor control and balance are associated with RD and could serve as early indicators of risk. The most advanced theory concerning the association between RD and motor/balance problems has been put forth by Nicolson and Fawcett (2006). They argue that motor/balance disturbances are not causal agents in RD, but rather are indicators of an underlying cerebellar dysfunction. This cerebellar impairment is further proposed to disrupt procedural learning and



Early identification of reading disabilities 317

slow the automaticity of a variety of motor and cognitive skills including those related to phonological processing and reading. Although the cerebellar dysfunction theory is an intriguing one, direct tests of its predictions, including those related to early identification, have been met with mixed results (Barth et al., 2010; Loras, Sigmundsson, Stensdotter, & Talcott, 2014). In a review of research related to this theory, Stoodley and Stein (2013) conclude that while the existence of a cerebellar dysfunction can account for numerous behavioral deficits associated with RD, it is unlikely to be the primary cause of RD.

Risk and resilience It should be clear from the previous sections that there are a number of early risk factors that are associated with RD. These factors likely interact with each other as well as combine with other positive or negative influences to increase or decrease the probability of RD. A framework that may help account for the interaction of multiple factors in the manifestation of RD is the risk-resilience framework (Kirby & Fraser, 1997; Rutter, 1985). This framework has been applied for some time to disorders of psychopathology, such as schizophrenia, depression, and bulimia. In such disorders, it has been observed that individuals with very similar risk factors can have very different outcomes. Some individuals seem to show resilience against even the strongest risk factors, while others do not. Researchers have sometimes explained these differences in resilience in terms of protective and vulnerability mechanisms (Rutter, 1985). Protective mechanisms are processes that reduce or buffer the impact of risk factors, whereas vulnerability mechanisms intensify the impact. Most importantly, protective/vulnerability mechanisms have their greatest impact in combination with risk factors. In other words, they serve as moderators of risk and have their impact primarily when risk factors are present. In the absence of risk, they have no impact or their impact is reduced considerably. A protective mechanism that could moderate the effects of risks for RD is instruction/intervention. That is, good instruction and/or well-tailored intervention can potentially reduce the impact of risk for RD. For example, Connor and colleagues (Connor, Morrison, & Katch, 2004) found that teacher-managed explicit code focused instruction had a significant impact on first graders’ reading skills, and this impact was greater for poor readers than for good readers. In another study, Foorman, Breier, and Fletcher (2003) found that a prescriptive kindergarten curriculum that included phonological awareness instruction differentially raised the letter-naming and phonological awareness skills of the lowest performing students. Child-level variables can also serve as protective mechanisms for risk for RD. For instance, it has been suggested that good oral language skills and/or rapid naming can be protective factors (Snowling, 2008; Pennington & Lefty, 2001; van Bergen et al., 2011). These factors could moderate the effects of deficits in other areas such as phonological awareness. However, in some cases oral language or rapid naming may better be

318 Hugh W. Catts

considered as causal ingredients than as moderators. In this case, they do not moderate the effects of other factors as much as they add to or subtract from the overall risk. For example, in the double deficit model, a deficit in rapid naming can combine with a deficit in phonological awareness but can also be a sufficient cause of word reading problems (Wolf & Bowers, 1999). The relationship of problems in oral language and RD may be more complex. In English, an oral language problem such as SLI may be better conceived of as a moderator of word reading difficulties. Problems in vocabulary and syntax associated with SLI likely influence the acquisition of word reading abilities, but given that many children with SLI have normal word reading skills (Catts & Adlof, 2011), they may not be sufficient to cause significant word reading difficulties. However, in a morphologically-based language such as Chinese, language problems in SLI might be a sufficient risk factor for word reading difficulties. Furthermore, oral language problems are clearly a causal factor in problems in reading comprehension, especially cases of specific comprehension deficits (Catts et al., 2006; Nation et al., 2010). Another potential moderator of RD is attention. Research has demonstrated that 25–40% of children with RD also have ADHD (Pennington, 2006). The relationship between these disorders is quite complex and far from fully understood. However one possible relationship is that the symptoms of ADHD may interact with a phonological deficit to increase the likelihood of RD. By itself ADHD does not appear to be a primary causal factor of RD (Snowling, 2008). But when combined with risk factors such as a phonological deficit, it could increase the probability of RD. While this is a plausible theoretical account, research in support of it is still quite limited. In addition, the relationship is further complicated by evidence of an interaction of processing speed (and its correlate rapid naming) with both ADHD and RD (Willcutt, Pennington, Olson, Chhabildas, & Hulslander, 2005). Eklund, Torppa, and Lyyntinen (2013) have suggested that task-focused behavior may also be a moderator of dyslexia. They defined such behavior as the tendency to remain highly engaged in tasks and/or to be persistent in the face of failure. Eklund et al. (2013) found that high levels of task-focused behavior in children with a family history of dyslexia was associated with the absence of reading problems irrespective of the presence of a phonological deficit. While task-focused behavior has just begun to be examined in relationship to RD, others have documented its association with reading outcomes more generally (Hirvonen, Georgiou, Lerkkanen, Aunola, & Nurmi, 2010; Stephenson, Parrila, Geogiou, & Kirby, 2008). Still further research has examined similar behavior in relationship to success/failure in other domains. For example, Duckworth and colleagues have investigated personality characteristics referred to as grit and self-control in relationship to positive life outcomes (Duckworth, Peterson, Matthews, & Kelly, 2007; Duckworth & Seligman, 2005). Grit is defined as the tendency to sustain interest in and effort toward long-term goals, while self-control is the regulation of behavior and attention in the face of momentary temptations or diversions. Duckworth and colleagues have found that grit and self-control account for success/failure across numerous outcomes, including academics, independent of talent or cognitive abilities.



Early identification of reading disabilities 319

Other work suggests that whereas individuals vary inherently in these personality characteristics, there is considerable within individual variability across domains in how these traits are expressed (Tsukayama, Duckworth, & Kim, 2012). Thus, it may not just be the level of grit or self-control that a person has but may also be dependent on one’s interest and/or motivation in a given domain. There is a large body of research that documents a relationship between motivation/interest in reading and reading achievement (Morgan & Fuchs, 2007). This work shows that good readers are more motivated to read and have more interest in reading than do poor readers. Although this relationship is likely to be bidirectional, some investigations do show that early differences in motivation/interest can impact later reading achievement (Lepola, Salnen, & Vaurus, 2000; Salonen, Lepola, Niemi, 1998).

Universal screening Not all children with RD will show a positive family history or will clearly present with language or other reading-related difficulties during the preschool years. As a result, the identification of risk for RD may not be feasible until these children begin school. As noted above, identification of RD has sometimes not taken place until children have entered school and actually demonstrated reading problems. To avoid this situation, it has become increasingly common to screen all children for risk for RD as they begin their formal education. This type of screening, referred to as universal screening, generally involves directly measuring reading and/or literacy-related skills (e.g., letter knowledge, non-word or sight-word reading). Early literacy screening batteries have also been supplemented with measures of phonological awareness, rapid naming, and verbal short-term memory (Foorman et al, 1998; Good & Kaminski, 2002; Wood, Hill, Meyer, & Flowers, 2005). Also, in the attempt to identify risk for problems in reading comprehension, oral language measures such as vocabulary, listening comprehension, or sentence repetition have been included (Catts, Fey, Zhang, & Tomblin, 2001; Foorman, Torgesen, Crawford, & Petscher, 2009; Wood et al., 2005). Research has indicated that these multivariate universal screening batteries can be accurate in the early identification of RD and that problems in both word reading and reading comprehension can be detected with acceptable levels of accuracy (Catts, Nielsen, Bridges, Bontempo, & Liu, 2015; Compton, Fuchs, Fuchs, & Bryant, 2006; O’Connor & Jenkins, 1999; Schatshneider, Fletcher, Francis, Carlson, & Foorman, 2004; Vellutino et al., 2006; Wood et al., 2005). Puolakanaho et al. (2007) and Carroll et al. (2014) have further shown that prediction can be improved by considering family risk for RD. The accuracy of universal screening has been an issue of concern for both educators and researchers alike. Predicting the future is always difficulty and no less so for predicting a complex outcome like reading achievement. Screening accuracy is often quantified in terms of sensitivity and specificity. Sensitivity is an index of the percentage of children with RD who fail a screening instrument, whereas specificity is the percentage of children without RD who pass the screening. As such, sensitivity

320 Hugh W. Catts

and specificity rates represent correct screening decisions. Corresponding screening errors are referred to as false positives or false negatives. A false positive occurs when an instrument identifies a child without RD as at risk, whereas a false negative occurs when a child with RD is missed by the screening. The percentage of false positive and negative errors will vary with the accuracy of the instrument and the cut-points selected. Too often cut-points are chosen without a consideration of the implications they may have for screening errors and educational practice. Jenkins (2003) has argued that particular attention should be given to false negatives in that it may be far worse to miss a child who is at risk than to falsely identify one who is not. He recommends that screening protocols strive to reach no greater than a 10% false negative rate. Of course, maintaining a low false negative rate can lead to an undesirably high false positive rate. The latter may cause undue concern for parents and children and result in the expense of unnecessary intervention. It can also be difficult for school personnel to use false positive rates in planning intervention practices. Petscher, Kim, and Foorman (2011) have suggested that it may be better to consider positive prediction power rather than false positive rate in making intervention decisions. Positive prediction power is an estimate of the percentage of children deemed to be at risk on the basis of a screening tool who actually turn out to have reading problems. This estimate may provide a better indication of the cost-benefit ratio of intervention than the false positive rate. Basically, it tells the school or district approximately how many more children need to be provided with intervention than actually may need it. Thus, this allows educators to consider the trade-off of identification vs. non-identification. Also, positive prediction power is a sample-based index, whereas false positive rate is a population-based estimate. As a result, positive prediction power is impacted by the number of children within a school or district who have RD while false positive rate is not. Thus, because schools or districts vary in demographic characteristics, curriculum, and performance levels and may have different rates of RD, positive prediction power can provide more relevant information on which to base a decision concerning a screening instrument. No matter which error rates or indices are considered, they can be improved by increasing the overall accuracy of an instrument. Several approaches have been considered. One concerns the timing of the screening. Catts et al. (2009) showed that if screening measures are given too early in development, floor effects can occur and lead to a considerable reduction in screening accuracy. They found that floor effects were not tied to a particular grade but were a function of the timing of the assessment of a specific skill. Children’s literacy and literacy-related skills are rapidly developing in the early school grades and matching the assessment with this development may be analogous to hitting a moving target (Speece, 2005). However, if measures are matched with developing skills, screening can be quite accurate. For example, Catts et.al. (2009) showed that a screening measure like letter name fluency may have a large floor effect and poor accuracy at the beginning of kindergarten but a more normal distribution and increased precision by midyear. Furthermore, the appropriateness of a screening measure can change as the curriculum changes. For instance, the increased attention on literacy in the United States has led to reading instruction being pushed down into



Early identification of reading disabilities 321

kindergarten and even preschool. As a result, a measure of letter name fluency at the beginning of kindergarten now can have less of a floor effect and better accuracy than it did just 5–10 years ago (Catts et al., 2015). Compton et al. (2010) has suggested that screening accuracy may also be increased by using a two-step gated procedure. In the first step, a brief screening instrument with a low false negative cut-score is used. A second, multivariate assessment follows and attempts to differentiate the resulting false positives from the true positives. Compton et al. (2010) showed that such an approach could effectively increase the accuracy of screening in first grade children (also see Bridges & Catts, 2011). Another approach that incorporates multiple steps is response to intervention. As described in the next section, this approach uses universal screening and progress monitoring to identify children at risk for RD.

Response to intervention Response to Intervention (RTI) is an educational model that has been put forth to improve the early identification and prevention of learning disabilities, including reading disabilities (Fuchs, Fuchs, & Speece, 2002). In this multi-tiered model, all children are provided with high-quality research-based reading instruction in the general education classroom. Children’s response to this instruction is assessed by universal screening instruments that are administered periodically throughout the school year. Children identified as at risk on the basis of this screening are provided with short-term supplemental intervention. This intervention may progress from small group to individual instruction based on children’s needs. Progress monitoring is used to measure children’s response to intervention. Those who fail to respond to supplemental intervention are considered to be “truly” at risk for RD and may qualify for more specialized instruction provided within a special education setting. Thus, within this model, screening instruments are combined with response to intervention to identify children with RD. There is some emerging evidence in support of RTI for the early identification of RD (Compton, Fuchs, Fuchs, & Bryant, 2006; VanDerHeyden, Witt, & Gilbertson., 2007; Vellutino et al., 2006). Compton et al. (2006) identified 248 at-risk children at the beginning of first grade and provided 9 weeks of small-group intervention to a random subsample of these children. Children were followed through the end of third grade at which time they were designated as RD or non RD on the basis of a composite of word reading and reading comprehension. Results suggested that response to intervention (measured in terms of growth in word reading) added to the prediction of reading outcomes over and above the initial screening assessment. Vellutino et al. (2008) reported that a screening battery administered at the beginning of kindergarten provided only limited accuracy in the prediction of first grade reading outcomes. However, when response to intervention during kindergarten was added, the accuracy of identification for children with RD in first grade was significantly increased.

322 Hugh W. Catts

In a more recent study, Catts, Nielsen, Bridges, and Liu (2016) examined if response to language intervention specifically would add uniquely to the prediction of deficits in reading comprehension. We administered a screening battery at the beginning of kindergarten that included measures of vocabulary and other reading-related measures. A portion of at-risk children were given intervention directed at improving vocabulary and narration. Analyses showed that response to vocabulary intervention (but not narration) added significantly to the prediction of outcomes in reading comprehension at the end of third grade. On the surface, the notion that response to intervention might be useful in the identification of RD seems quite plausible. A child who does not respond to instruction under the best of circumstances would appear to be a good candidate for a child with RD. However, the approach has been criticized on a number of accounts. Some have criticized the approach for not taking into consideration the cognitive and neuropsychological profiles of children when making diagnostic decisions based on response to intervention (Hale, Kaufman, Naglieri, & Kavale, 2006; Reynolds & Shaywitz, 2009). Traditionally, learning disabilities, which encompass RD, have been considered to arise from factors intrinsic to the child. Without documentation of cognitive or neuropsychological deficits, some claim that RD loses its distinction from low achievement or general learning problems (McKenzie, 2009). It is also argued that a comprehensive assessment is necessary to plan appropriate intervention (Reynolds & Shaywitz, 2009). Others, however, have pointed out that there is very little evidence that such an assessment leads to more effective intervention (Compton et al. 2012; Fletcher & Vaughn, 2009). The RTI approach has also been challenged because it provides no mechanism for identifying especially bright children who do not read at a level consistent with their other cognitive skills. Opponents of RTI, argue that these bright and talented students may not qualify for the intervention they need within a RTI approach (Reynolds & Shaywitz, 2009). On the other hand, research indicates that high IQ or special talent may have little to do with learning to read, at least learning to read words (Stuebing et al., 2002). As such, bright underachieving children may be no more deserving or in need of intervention than are low achieving children with less cognitive intellect. Finally, some have argued that the psychometric obstacles involved in quantifying and comparing level of instruction to level of reading achievement are too substantial for the RTI approach to be used in a reliable fashion (Gerber, 2005). The issue of instructional variability alone is enough to seriously undermine RTI. Teachers and schools vary considerably in terms of the amount and quality of instruction, and this variability presents a considerable challenge in a RTI approach. Children could potentially be identified as RD in one school or district and not in another as a direct result of the quality of instruction alone. Also, as instructional models change, the nature of those with RD would change as well.



Early identification of reading disabilities 323

Dynamic assessment Another approach that may have potential to improve the accuracy of early identification of RD is dynamic assessment (O’Connor & Jenkins, 1999; Elbro, Daugaard, & Gellert, 2010). Like progress monitoring in an RTI model, dynamic assessment measures response to instruction. However, in the case of dynamic assessment, response is measured over a shorter time period and in relationship to more targeted instruction. Dynamic assessment is linked theoretically to Vygotsky’s (1987) notion of the “zone of proximal development.” This zone represents the potential gain between what a child can do independently versus what he/she can do with assistance. Applied to early identification, a dynamic assessment provides an index of how well a child might be expected to respond to instruction in the classroom. Rather than a measure of a learned product (i.e., static assessment), it is a measure of the potential to learn (Grigorenko & Sternberg, 1998). Therefore, dynamic assessment might be a particularly sensitive index of risk for difficulties in learning to read. By measuring how well children respond to targeted instruction and feedback concerning reading or reading-related skills, the examiner can gauge how difficult it will be for children to learn to read over a more extended period of instruction. A few studies have examined the usefulness of dynamic assessment in the identification of RD. For example, Elbro et al. (2010) used a dynamic test to identify dyslexia in adult second language learners. This test involved the learning of three novel letter shapes and their sounds and subsequent reading of non-words using the three letters. Instructions were non-verbal in nature. Their results provided some preliminary evidence that the dynamic test could differentiate dyslexic and non-dyslexic second language learners. Fuchs, Compton, Fuchs, Bouton,and Caffrey (2011) taught first grade children to read a small set of novel words and used their response to instruction to predict their future reading skills. More recently, Cho, Compton, Fuchs, Fuchs, and Bouton (2012) found that a similar dynamic assessment of decoding explained unique variance in children’s responsiveness to Tier 2 intervention. Dynamic assessments have also been developed to assess reading-related cognitive skills. For example, Bridges and Catts (2011) used a dynamic assessment of phonological awareness in kindergarteners to predict first grade word reading. We found that this assessment explained variability in reading achievement over and above a static measure of phonological awareness. However, when used in a multivariate screening battery it did not provide unique information in the prediction of word reading difficulties (Catts et al., 2015). To the contrary, O’Conner and Jenkins (1999) found a dynamic assessment of phonological awareness to be useful within a multivariate screening battery. Finally, dynamic assessment has also been used to assess developing language abilities that support later reading comprehension. For instance, Camilleri and Law (2007) have developed a dynamic assessment of receptive vocabulary to identify children with language-learning problems. In other work, Pena, Gillam, and Bedore (2014) have used a dynamic assessment of narration to uncover language impairments in young English language learners.

324 Hugh W. Catts

Summary Important advancements have been made in the early identification of RD. No longer do practitioners and educators have to wait for reading problems to develop before they can be identified and addressed. Research has uncovered factors related to RD that can identify risk for RD prior to beginning formal education. This work shows that a family history of RD and/or the presence of developmental language problems often foretell later reading difficulties. Universal screening, response to intervention, and/or dynamic assessment can also be used as children begin school to identify those children who are at risk for RD. For each of these approaches to be maximally beneficial they need to be tied to an appropriate intervention program. In other chapters in this book, intervention strategies that can be used to reduce the risk of RD will be discussed.

References Aaron, P. G., Joshi, M., & Williams, K. A. (1999). Not all reading disabilities are alike. Journal of Learning Disabilities, 32, 120–137.  doi: 10.1177/002221949903200203 Barth, A. E., Denton, C. A., Stuebing, K. K., Cirino, P. T., Francis, D. J., & Vaughn (2010). A test of the cerebellar hypothesis of dyslexia in adequate and inadequate responders to reading intervention. Journal of International Neuropsychological Society, 16, 526–536. doi: 10.1017/S1355617710000135 Beitchman, J., Wilson, B., Douglas, L., Young, A., & Adlaf, E. (2001). Substance use disorders in young adults with and without LD: predictive and concurrent relationships. Journal of Learning Disabilities, 34, 317–332.  doi: 10.1177/002221940103400407 Bidwell, L., Willcutt, E. G., DeFries, J. C., & Pennington, B. F. (2007). Testing for neuropsychological endophenotypes in siblings discordant for attention-deficit/hyperactivity disorder. Biological Psychiatry, 62, 991–998.  doi: 10.1016/j.biopsych.2007.04.003 Bishop, D. V. M., & Adams, C. (1990). A prospective study of the relationship between specific language impairment, phonological disorders and reading retardation. Journal of Child Psychology and Psychiatry, 31, 1027–1050.  doi: 10.1111/j.1469-7610.1990.tb00844.x Bishop, D. V., McDonald, D., Bird, S., & Hayiou-Thomas, M. E. (2009) Children who read words accurately despite language impairment: Who are they and how do they do it? Child Development, 80, 593–605.  doi: 10.1111/j.1467-8624.2009.01281.x Boets, B., De Smedt, B., Cleuren, L., Vandewalle, E., Wouters, J., & Ghesquiere, P. (2010). Towards a further characterization of phonological and literacy problems in Dutch‐speaking children with dyslexia. British Journal of Developmental Psychology, 28, 5–31. doi: 10.1348/026151010X485223 Boets, B., Wouters, J., Van Wieringen, A., & Ghesquière, P. (2006). Auditory temporal information processing in preschool children at family risk for dyslexia: Relations with phonological abilities and developing literacy skills. Brain and Language, 97, 64–79. doi: 10.1016/j.bandl.2005.07.026 Bridges, M. S., & Catts, H. W. (2011). The use of a dynamic screening of phonological awareness to predict risk for reading disabilities in kindergarten children. Journal of Learning Disabilities, 44, 330–338.  doi: 10.1177/0022219411407863



Early identification of reading disabilities 325

Brady, S. A., & Shankweiler, D. P. (Eds.) (1991). Phonological processes in literacy. Hillsdale, NJ: Erlbaum. Camilleri, B., & Law, J. (2007). Assessing children referred to speech and language therapy: Static and dynamic assessment of receptive vocabulary. International Journal of Speech-Language Pathology, 9, 312–322.  doi: 10.1080/14417040701624474 Carroll, J. M., Mundy, I. R., & Cunningham, A. J. (2014). The roles of family history of dyslexia, language, speech production and phonological processing in predicting literacy progress. Developmental Science, 17, 727–742.  doi: 10.1111/desc.12153 Castles, A., & Coltheart, M. (2004). Is there a causal link from phonological awareness to success in learning to read? Cognition, 91, 77–111.  doi: 10.1016/S0010-0277(03)00164-1 Catts, H. W. (1989). Speech production deficits in developmental dyslexia. Journal of Speech and Hearing Disorders, 54, 422–428.  doi: 10.1044/jshd.5403.422 Catts, H. W. (1993). The relationship between speech-language impairments and reading disabilities. Journal of Speech and Hearing Research, 36, 948–958. Catts, H. W., & Adlof, S. (2011). Phonological and other language deficits associated with dyslexia. In S. Brady, D. Braze, & C. A. Fowler (Eds.), Explaining individual differences in reading: Theory and evidence, (pp. 137–151). New York: Psychology Press. Catts, H., Adlof, S., Hogan, T., Ellis Weismer, S. (2005). Are specific language impairment and dyslexia distinct disorders? Journal of Speech-Language-Hearing Research, 48, 1378–1396. doi: 10.1044/1092-4388(2005/096) Catts, H. W., Adlof, S., & Weismer, S. E. (2006). Language deficits in poor comprehenders: A case for the simple view of reading. Journal of Speech, Language, and Hearing Research, 49, 278–293.  doi: 10.1044/1092-4388(2006/023) Catts, H. W., Fey, M. E., Tomblin, J. B., & Zhang, X. (2002). A longitudinal investigation of reading outcomes in children with language impairments. Journal of Speech, Language, and Hearing Research, 45, 1142–1157.  doi: 10.1044/1092-4388(2002/093) Catts, H. W., Fey, M., Zhang, X., & Tomblin, J. B. (2001). Estimating risk for future reading difficulties in kindergarten children: A research-based model and its clinical implications. Language, Speech, and Hearing Services in Schools, 32, 38–50.  doi: 10.1044/0161-1461(2001/004) Catts, H. W., Hogan, T. P., & Adlof, S. M. (2005). Developmental changes in reading and reading disabilities. In H. W. Catts, & A. G. Kamhi (Eds.) Connections between language and reading disabilities (pp. 25–40), Mahwah, NJ: Erlbaum. Catts, H. W., McIlraith, A., Bridges, M., & Nielsen, D. (2017). Viewing a phonological deficit within a multifactorial model of dyslexia. Reading and Writing: An Interdisciplinary Journal, 30, 613–629. Catts, H. W., Nielsen, D., Bridges, M., Bontempo, D., & Liu, Y. (2015). Early identification of reading disabilities within an RTI framework. Journal of Learning Disabilities,48, 281–297. doi: 10.1177/0022219413498115 Catts, H. W., Nielsen, D., Bridges, M., & Liu, Y. (2016). Early identification of reading comprehension difficulties. Journal of Learning Disabilities 49, 451–465.  doi: 10.1177/0022219414556121 Catts, H. W., Petscher, Y., Schatschneider, C., Bridges, M. S., & Mendoza, K. (2009). Floor effects associated with universal screening and their impact on the early identification of reading disabilities. Journal of Learning Disabilities, 42, 163–176.  doi: 10.1177/0022219408326219 Cho, E., Compton, D. L., Fuchs, D., Fuchs, L. S., & Bouton, B. (2012). Examining the predictive validity of a dynamic assessment of decoding to forecast response to Tier 2 intervention. Journal of Learning Disabilities, 47, 409–423.  doi: 10.1177/0022219412466703

326 Hugh W. Catts

Compton, D. L., Fuchs, D., Fuchs, L. S., Bouton, B., Gilbert, J. K, Barquero, L. A, … & Crouch, R. C. (2010). Selecting at-risk first-grade readers for early intervention: Eliminating false positives and exploring the promise of a two-stage gated screening process. Journal of Educational Psychology, 102, 327–340.  doi: 10.1037/a0018448 Compton, D. L., Fuchs, D., Fuchs, L. S., & Bryant, J. D. (2006). Selecting at-risk readers in first grade for early intervention: A two-year longitudinal study of decision rules and procedures. Journal of Educational Psychology, 98, 394–409.  doi: 10.1037/0022-0663.98.2.394 Compton, D. L., Gilbert, J. K., Jenkins, J. R., Fuchs, D., Fuchs, L. S., Cho, E., … & Bouton, B. (2012). Accelerating chronically unresponsive children to tier 3 instruction: What level of data is necessary to ensure selection accuracy? Journal of Learning Disabilities, 45, 204–216. doi: 10.1177/0022219412442151 Connor, C. M., Morrison, F. J., & Katch, L. E. (2004). Beyond the reading wars: Exploring the effect of child-instruction interactions on growth in early reading. Scientific Studies of Reading, 8, 305–336.  doi: 10.1207/s1532799xssr0804_1 Dandache, S., Wouters, J., & Ghesquière, P. (2014). Development of reading and phonological skills of children at family risk for dyslexia: A longitudinal analysis from kindergarten to sixth grade. Dyslexia, 20, 305–329.  doi: 10.1002/dys.1482 de Jong, P. F., & van der Leij, A. (2002). Effects of phonological abilities and linguistic comprehension on the development of reading. Scientific Studies of Reading, 6, 51–77. doi: 10.1207/S1532799XSSR0601_03 Duckworth, A. L., Peterson, C., Matthews, M. D., & Kelly, D. R. (2007). Grit: Perseverance and passion for long-term goals. Journal of Personality and Social Psychology, 92, 1087–1101. doi: 10.1037/0022-3514.92.6.1087 Duckworth, A. L., & Seligman, M. E. (2005). Self-discipline outdoes IQ in predicting academic performance of adolescents. Psychological Science, 16, 939–944.  doi: 10.1111/j.1467-9280.2005.01641.x Eklund, K. M., Torppa, M., & Lyytinen, H. (2013). Predicting reading disability: Early cognitive risk and protective factors. Dyslexia, 19, 1–10.  doi: 10.1002/dys.1447 Elbro, C., Borstrøm, I., & Petersen, D. K. (1998). Predicting dyslexia from kindergarten: The importance of distinctness of phonological representations of lexical items. Reading Research Quarterly, 33, 36–60.  doi: 10.1598/RRQ.33.1.3 Elbro, C., Daugaard, H. T., & Gellert, A. S. (July, 2010). Specific reading difficulties may be hard to diagnose in a second language. Paper presented at the Seventeenth Annual Meeting of the Society for the Scientific Study of Reading, Berlin. Elliott, J. G., & Grigorenko, E. L. (2014). The dyslexia debate. New York: Cambridge University Press.  doi: 10.1017/CBO9781139017824 Fletcher, J. M., & Vaughn, S. (2009). Response to intervention: Preventing and remediating academic difficulties. Child Development Perspectives, 3, 30–37.  doi: 10.1111/j.1750-8606.2008.00072.x Foorman, B. R., Breier, J. I., & Fletcher, J. M. (2003). Interventions aimed at improving reading success: An evidence-based approach. Developmental Neuropsychology, 24, 613–639. doi: 10.1080/87565641.2003.9651913 Foorman, B. R., Francis, D. J., Fletcher, J. M., Schatschneider, C., & Mehta, P. (1998). The role of instruction in learning to read: Preventing reading failure in at-risk children. Journal of Educational Psychology, 90, 37–55.  doi: 10.1037/0022-0663.90.1.37 Foorman, B. R., Torgesen, J. K., Crawford, E., & Petscher, Y. (2009). Assessments to guide reading instruction: Decisions supported by the new Florida system. Perspectives on Language and Literacy, 4, 13–19.



Early identification of reading disabilities 327

Franceschini, S., Gori, S., Ruffino, M., Pedrolli, K., & Facoetti, A. (2012). A causal link between visual spatial attention and reading acquisition. Current Biology, 22, 814–819. doi: 10.1016/j.cub.2012.03.013 Fuchs, D., Compton, D. L., Fuchs, L. S., Bouton, B., & Caffrey, E. (2011). The construct and predictive validity of a dynamic assessment of young children learning to read: Implications for RTI frameworks. Journal of Learning Disabilities, 44, 339–347.  doi: 10.1177/0022219411407864 Fuchs, L. S., Fuchs, D., & Speece, D. L. (2002). Treatment validity as a unifying construct for identifying learning disabilities. Learning Disability Quarterly, 25, 33–45.  doi: 10.2307/1511189 Gerber, M. M. (2005). Teachers are still the test limitations of response to instruction strategies for identifying children with learning disabilities. Journal of Learning Disabilities, 38, 516–524. doi: 10.1177/00222194050380060701 Gilger, J. W., Hanebuth, E., Smith, S. D., & Pennington, B. F. (1996). Differential risk for developmental reading disorders in the offspring of compensated versus non-compensated parents. Reading and Writing, 8, 407–417. Good, R. H., & Kaminski, R. A. (Eds.) (2002). Dynamic Indicators of Basic Early Literacy Skills (6th ed.). Eugene, OR: Institute for the Development of Educational Achievement. Gough, P. B., & Tunmer, W. E. (1986). Decoding, reading, and reading disability. Remedial and Special Education, 7, 6–10.  doi: 10.1177/074193258600700104 Grigorenko, E. L., & Sternberg, R. J. (1998). Dynamic testing. Psychological Bulletin, 124, 75–111. doi: 10.1037/0033-2909.124.1.75 Hale, J. B., Kaufman, A., Naglieri, J. A., & Kavale, K. A. (2006). Implementation of IDEA: Integrating response to intervention and cognitive assessment methods. Psychology in the Schools, 43, 753–770.  doi: 10.1002/pits.20186 Hirvonen, R., Georgiou, G. K., Lerkkanen, M. K., Aunola, K., & Nurmi, J. E. (2010). Task focused behaviour and literacy development: A reciprocal relationship. Journal of Research in Reading, 33, 302–319.  doi: 10.1111/j.1467-9817.2009.01415.x Hoover, W. A., & Gough, P. B. (1990). The simple view of reading. Reading and Writing, 2, 127–160. doi: 10.1007/BF00401799 Hulme, C., & Snowling, M. (1992). Deficits in output phonology: An explanation of reading failure? Cognitive Neuropsychology, 9, 47–72.  doi: 10.1080/02643299208252052 Jenkins, J. R. (2003). Candidate measures for screening at-risk students. Paper presented at the NRCLD Responsiveness-to-Intervention Symposium, Kansas City, MO. Kamhi, A. G., & Catts, H. W. (Eds.) (2012). Language and reading disabilities. Boston, MA: Pearson. Kelso, K., Fletcher, J., & Lee, P. (2007). Reading comprehension in children with specific language impairment: An examination of two subgroups. International Journal of Language & Communication Disorders, 42, 39–57.  doi: 10.1080/13682820600693013 Kendeou, P., van den Broek, P., White, M. J., & Lynch, J. S. (2009). Predicting reading comprehension in early elementary school: The independent contributions of oral language and decoding skills. Journal of Educational Psychology, 101, 765–778.  doi: 10.1037/a0015956 Kim, Y. S. (2015). Language and cognitive predictors of text comprehension: Evidence from multivariate analysis. Child Development, 86, 128–144.  doi: 10.1111/cdev.12293 Kirby, L. D., & Fraser, M. W. (1997). Risk and resilience in childhood. In M. Fraser (Eds.), Risk and resilience in childhood: An ecological perspective (pp.10–33), Washington DC: NASW Press. Kirk, J., & Reid, G. (2001). An examination of the relationship between dyslexia and offending in young people and the implications for the training system. Dyslexia, 7, 77–84. doi: 10.1002/dys.184

328 Hugh W. Catts

Language and Reading Research Consortium. (2015). Learning to read: should we keep things simple? Reading Research Quarterly, 50, 151–169. Lallier, M., Tainturier, M. J., Dering, B., Donnadieu, S., Valdois, S., & Thierry, G. (2010). Behavioral and ERP evidence for amodal sluggish attentional shifting in developmental dyslexia. Neuropsychologia, 48, 4125–4135.  doi: 10.1016/j.neuropsychologia.2010.09.027 Leonard, L. (2014). Children with specific language impairment, 2nd edition. Cambridge, MA: MIT Press. Lepola, J., Salonen, P., & Vauras, M. (2000). The development of motivational orientations as a function of divergent reading careers from pre-school to the second grade. Learning and Instruction, 10, 153–177.  doi: 10.1016/S0959-4752(99)00024-9 Loras, H., Sigmundsson, H., Stensdotter, A., & Talcott, J. B. (2014). Postural control is not systematically related to reading skills: Implications for the assessment of balance as a risk factor for developmental dyslexia. PLoS One, 9, e98224.  doi: 10.1371/journal.pone.0098224 Lyytinen, P., Eklund, K., & Lyytinen, H. (2005). Language development and literacy skills in latetalking toddlers with and without familial risk for dyslexia. Annals of Dyslexia, 55, 166–192. doi: 10.1007/s11881-005-0010-y Lyytinen, H., Erskine, J., Tolvanen, A., Torppa, M., Poikkeus, A. M., & Lyytinen, P. (2006). Trajectories of reading development: A follow-up from birth to school age of children with and without risk for dyslexia. Merrill-Palmer Quarterly: Journal of Developmental Psychology, 52, 514–546.  doi: 10.1353/mpq.2006.0031 McArthur, G., & Castles, A. (2013). Phonological processing deficits in specific reading disability and specific language impairment: Same or different? Journal of Research in Reading, 36, 280–302.  doi: 10.1111/j.1467-9817.2011.01503.x McKenzie, R. G. (2009). Obscuring vital distinctions: The oversimplification of learning disabilities within RTI. Learning Disability Quarterly, 32, 203–215.  doi: 10.2307/27740373 Melby-Lervåg, M., Lyster, S. A. H., & Hulme, C. (2012). Phonological skills and their role in learning to read: A meta-analytic review. Psychological Bulletin, 138, 322–352. doi: 10.1037/a0026744 Moll, K., Loff, A., & Snowling, M. J. (2013). Cognitive endophenotypes of dyslexia. Scientific Studies of Reading, 17, 385–397.  doi: 10.1080/10888438.2012.736439 Morgan, P. L., & Fuchs, D. (2007). Is there a bidirectional relationship between children’s reading skills and reading motivation? Exceptional Children, 73, 165–183.  doi: 10.1177/001440290707300203 Nash, H. M., Hulme, C., Gooch, D., & Snowling, M. J. (2013). Preschool language profiles of children at family risk of dyslexia: Continuities with specific language impairment. Journal of Child Psychology and Psychiatry, 54, 958–968.  doi: 10.1111/jcpp.12091 National Reading Panel (2000). Report of the national reading panel. Teaching children to read: An evidence based assessment of the scientific research literature on reading and its implications for reading instruction. Washington, DC: National Institute of Child Health and Human Development, National Institutes of Health. Nation, K., Cocksey, J., Taylor, J. S., & Bishop, D. V. (2010). A longitudinal investigation of early reading and language skills in children with poor reading comprehension. Journal of Child Psychology and Psychiatry, 51, 1031–1039.  doi: 10.1111/j.1469-7610.2010.02254.x Nicolson, R. I., & Fawcett, A. J. (2006). Do cerebellar deficits underlie phonological problems in dyslexia? Developmental Science, 9, 259–262.  doi: 10.1111/j.1467-7687.2006.00486.x O’Connor, R. E., & Jenkins, J. R. (1999). The prediction of reading disabilities in kindergarten and first grade. Scientific Studies of Reading, 3, 159–197.  doi: 10.1207/s1532799xssr0302_4



Early identification of reading disabilities 329

Olson, R. K., Keenan, J. M., Byrne, B., & Samuelsson, S. (2014). Why do children differ in their development of reading and related skills? Scientific Studies of Reading, 18, 38–54. doi: 10.1080/10888438.2013.800521 Olulade, O. A., Napoliello, E. M., & Eden, G. F. (2013). Abnormal visual motion processing is not a cause of dyslexia. Neuron, 79, 180–190.  doi: 10.1016/j.neuron.2013.05.002 Owens, R. (2000). Language development: An introduction (5th ed.). Boston, MA: Allyn & Bacon. Paul, R., Murray, C., Clancy, K., & Andrews, D. (1997). Reading and metaphonological outcomes in late talkers. Journal of Speech, Language, and Hearing Research, 40, 1037–1047. doi: 10.1044/jslhr.4005.1037 Peña, E. D., Gillam, R. B., & Bedore, L. M. (2014). Dynamic assessment of narrative ability in English accurately identifies language impairment in English language learners. Journal of Speech, Language, and Hearing Research, 57, 2208–2220.  doi: 10.1044/2014_JSLHR-L-13-0151 Pennington, B. F. (2006). From single to multiple deficit models of developmental disorders. Cognition, 101, 385–413.  doi: 10.1016/j.cognition.2006.04.008 Pennington, B. F., & Lefly, D. L. (2001). Early reading development in children at family risk for dyslexia. Child Development, 7, 816–833.  doi: 10.1111/1467-8624.00317 Perfetti, C. A., & Hart, L. (2002). The lexical quality hypothesis. In L. Verhoeven, C. Elbro, & P. Reitsma (Eds.) Precursors of functional literacy (pp. 67–86). Amsterdam: John Benjamins Publishing Company.  doi: 10.1075/swll.11.14per Petscher, Y., Kim, Y. S., & Foorman, B. R. (2011). The importance of predictive power in early screening assessments: Implications for placement in the response to intervention framework. Assessment for Effective Intervention, 36, 158–166.  doi: 10.1177/1534508410396698 Prado, C., Dubois, M., & Valdois, S. (2007). The eye movements of dyslexic children during reading and visual search: Impact of the visual attention span. Vision Research, 47, 2521–2530. doi: 10.1016/j.visres.2007.06.001 Preston, J. L., Frost, S. J., Mencl, W. E., Fulbright, R. K., Landi, N., Grigorenko, E., … & Pugh, K. R. (2010). Early and late talkers: School-age language, literacy and neurolinguistic differences. Brain, 133, 2185–2195.  doi: 10.1093/brain/awq163 Puolakanaho, A., Ahonen, T., Aro, M., Eklund, K., Leppanen, P. H., Poikkeus, A. M.,…& Lyytinen, H. (2007). Very early phonological and language skills: Estimating individual risk of reading disability. Journal of Child Psychology and Psychiatry, 48, 923–931. doi: 10.1111/j.1469-7610.2007.01763.x Ramus, F., & Ahissar, M. (2012). Developmental dyslexia: The difficulties of interpreting poor performance, and the importance of normal performance. Cognitive Neuropsychology, 29, 104–122.  doi: 10.1080/02643294.2012.677420 Rescorla, L. (2002). Language and reading outcomes to age 9 in late-talking toddlers. Journal of Speech, Language, and Hearing Research, 45, 360–371.  doi: 10.1044/1092-4388(2002/028) Reynolds, C. R., & Shaywitz, S. E. (2009). Response to Intervention: Ready or not? Or, from waitto-fail to watch-them-fail. School Psychology Quarterly, 24, 130–145.  doi: 10.1037/a0016158 Rutter, M. (1985). Resilience in the face of adversity. British Journal of Psychiatry, 147, 598–611. doi: 10.1192/bjp.147.6.598 Salonen, P., Lepola, J., & Niemi, P. (1998). The development of first graders’ reading skill as a function of pre-school motivational orientation and phonemic awareness. European Journal of Psychology of Education, 13, 155–174.  doi: 10.1007/BF03173087 Scarborough, H. S. (1990). Very early language deficits in dyslexic children. Child Development, 61, 1728–1743.  doi: 10.2307/1130834

330 Hugh W. Catts

Schatchneider, C., Fletcher, J. M., Francis, D. J., Carlson, C. D., & Foorman, B. R. (2004). Kindergarten prediction of reading skills: A longitudinal comparative analysis. Journal of Educational Psychology, 96, 265–282. Snowling, M. J. (1981). Phonemic deficits in developmental dyslexia. Psychological Research, 43, 219–234.  doi: 10.1007/BF00309831 Snowling, M. J. (2008). Specific disorders and broader phenotypes: The case of dyslexia. The Quarterly Journal of Experimental Psychology, 61, 142–156.  doi: 10.1080/17470210701508830 Snowling, M. J., Gallagher, A., & Frith, U. (2003). Family risk of dyslexia is continuous: Individual differences in the precursors of reading skill. Child Development, 74, 358–373. doi: 10.1111/1467-8624.7402003 Snowling, M. J., Muter, V., & Carroll, J. (2007). Children at family risk of dyslexia: A follow‐up in early adolescence. Journal of Child Psychology and Psychiatry, 48, 609–618. doi: 10.1111/j.1469-7610.2006.01725.x Spear-Swerling, L., & Sternberg, R. J. (Eds.) (1996). Off track: When poor readers become “learning disabled.” Boulder, CO: Westview Press. Speece, D. L. (2005). Hitting the moving target known as reading development. Some thoughts on screening children for secondary interventions. Journal of Learning Disabilities, 38, 487–493.  doi: 10.1177/00222194050380060301 Stanovich, K. E. (1988). The right and wrong places to look for the cognitive locus of reading disability. Annals of Dyslexia, 38, 154–177.  doi: 10.1007/BF02648254 Stein, J. (2001). The magnocellular theory of developmental dyslexia. Dyslexia, 7, 12–36. doi: 10.1002/dys.186 Stephenson, K. A., Parrila, R. K., Georgiou, G. K., & Kirby, J. R. (2008). Effects of home literacy, parents’ beliefs, and children’s task-focused behavior on emergent literacy and word reading skills. Scientific Studies of Reading, 12, 24–50.  doi: 10.1080/10888430701746864 Stoodley, C. J., & Stein, J. F. (2013). Cerebellar function in developmental dyslexia. The Cerebellum, 12, 267–276.  doi: 10.1007/s12311-012-0407-1 Storch, S. A., & Whitehurst, G. J. (2002). Oral language and code-related precursors to reading: Evidence from a longitudinal structural model. Developmental Psychology, 38, 934–947. doi: 10.1037/0012-1649.38.6.934 Stuebing, K. K., Fletcher, J. M., LeDoux, J. M., Lyon, G. R., Shaywitz, S. E., & Shaywitz, B. A. (2002). Validity of IQ-discrepancy classifications of reading disabilities: A meta-analysis. American Educational Research Journal, 39, 469–518.  doi: 10.3102/00028312039002469 Tager-Flusberg, H., & Cooper, J. (1999). Present and future possibilities for defining a phenotype for specific language impairment. Journal of Speech, Language, and Hearing Research, 42, 1275–1278.  doi: 10.1044/jslhr.4205.1275 Tomblin, J. B., & Nippold, M. A. (Eds.). (2014). Understanding individual differences in language development across the school years. New York, NY: Psychology Press. Tomblin, J. B., Nippold, M. A., Fey, M. E., & Zhang, X. (2014). The character and course of individual differences in spoken language. In J. B. Tomblin & M. A. Nippold (Eds.) Understanding individual differences in language development across the school years (pp. 47–78). New York, NY: Psychology Press. Tsukayama, E., Duckworth, A. L., & Kim, B. (2012). Resisting everything except temptation: Evidence and an explanation for domain‐specific impulsivity. European Journal of Personality, 26, 318–334.  doi: 10.1002/per.841



Early identification of reading disabilities 331

van Bergen, E., Bishop, D. V. M., van Zuijen, T. L., & de Jong, P. F. (2015). How does parental reading influence children’s reading? A study of cognitive mediation. Scientific Studies of Reading, 19, 325–339.  doi: 10.1080/10888438.2015.1050103 van Bergen, E., de Jong, P. F., Plakas, A., Maassen, B., & van der Leij, A. (2012). Child and parental literacy levels within families with a history of dyslexia. Journal of Child Psychology and Psychiatry, 53, 28–36.  doi: 10.1111/j.1469-7610.2011.02418.x van Bergen, E., de Jong, P. F., Regtvoort, A., Oort, F., van Otterloo, S., & van der Leij, A. (2011). Dutch children at family risk of dyslexia: Precursors, reading development, and parental effects. Dyslexia, 17, 2–18.  doi: 10.1002/dys.423 VanDerHeyden, A. M., Witt, J. C., & Gilbertson, D. (2007). A multi-year evaluation of the effects of a response to intervention (RTI) model on identification of children for special education. Journal of School Psychology, 45, 225–256.  doi: 10.1016/j.jsp.2006.11.004 Vellutino F. R., Scanlon D. M., Small S., & Fanuele D. P. (2006). Response to intervention as a vehicle for distinguishing between children with and without reading disabilities: Evidence for the role of kindergarten and first-grade interventions. Journal of Learning Disability, 39, 157–169.  doi: 10.1177/00222194060390020401 Vellutino, F. R., Scanlon, D. M., Zhang, H., & Schatschneider, C. (2008). Using response to kindergarten and first grade intervention to identify children at-risk for long-term reading difficulties. Reading and Writing, 21, 437–480.  doi: 10.1007/s11145-007-9098-2 Vygotsky, L. S. (1987). The development of scientific concepts in childhood. In R. W. Rieber & A. S. Carton (Eds.) & N. Minick (Trans.), The collected works of L. S. Vygotsky (Vol. 1, pp. 167–241). New York: Plenum Press. Wagner, R. K., & Torgesen, J. K. (1987). The nature of phonological processing and its causal role in the acquisition of reading skills. Psychological Bulletin, 101, 192–212. doi: 10.1037/0033-2909.101.2.192 Willcutt, E. G., Pennington, B. F., Olson, R. K., Chhabildas, N., & Hulslander, J. (2005). Neuropsychological analyses of comorbidity between reading disability and attention deficit hyperactivity disorder: In search of the common deficit. Developmental Neuropsychology, 27, 35–78.  doi: 10.1207/s15326942dn2701_3 Wolf, M., & Bowers, P. G. (1999). The double-deficit hypothesis for the developmental dyslexias. Journal of Educational Psychology, 91, 415–438.  doi: 10.1037/0022-0663.91.3.415 Wood, F. B., Hill, D. F., Meyer, M. S., & Flowers, D. L. (2005). Predictive assessment of reading. Annals of Dyslexia, 55, 193–216.  doi: 10.1007/s11881-005-0011-x

Dyslexia and word reading problems Rauno K. Parrila and Athanassios Protopapas University of Alberta / University of Oslo

Developmental dyslexia is the most common learning disability in children, with prevalence estimates varying between 3% and 20% of all school age children (see e.g., Shaywitz, 1996; Snowling, 2013). It is acknowledged to affect children across languages, writing systems, and educational approaches to reading instruction. Developmental dyslexia is also the most widely studied behaviourally defined developmental disorder, with a rapidly expanding evidence base on associated genetics, neural functioning, cognitive skills, and environmental influences. In this chapter, we provide an overview of widely available cognitive theories of developmental dyslexia. Our review is by no means exhaustive in terms of theories included or the evidence for and against each of them – not even a book-length treatment (see, e.g., Elliot & Grigorenko, 2014) could achieve that. We hope, however, that we cover the main theories and references driving the cognitive research on dyslexia at the moment. Further, we limit our discussion of developmental dyslexia to alphabetic orthographies, and mainly to European alphabetic orthographies that have been studied most extensively. We make no claims about the universality of ideas presented (see McBride-Chang, this volume; McCardle, Miller, Lee, & Tzeng, 2011; Nag, this volume; Share, 2008, for cross-linguistic issues in dyslexia and reading research) but acknowledge that as the theoretical models of developmental dyslexia develop and move from single-deficit models to multiple-deficit and hybrid models, their potential for accommodating specific features of different writing systems likely improves. Below, we will first define developmental dyslexia and then review what we call the single-cause theories of dyslexia that have traditionally dominated the field. We then advance to more recent double- and multiple-deficit models, and conclude with a discussion considering the place of development and of the individual with dyslexia in developmental dyslexia research.

Developmental dyslexia defined We define developmental dyslexia as a persistent and unexpected difficulty in developing age- and experience-appropriate word reading skills. For us, word reading skills include both accuracy and efficiency, defined as correctly read words per unit of time. doi 10.1075/swll.15.19par © 2017 John Benjamins Publishing Company

334 Rauno K. Parrila and Athanassios Protopapas

Difficulty includes both performance that falls at the low end of the distribution for any given age and educational level as well as performance that may not be clinically low but can be maintained only with extraordinary effort. We take it for granted that word reading abilities are continuously distributed in a population and our definition potentially identifies as dyslexics those individuals who consistently fall at the lower end of the distribution on some word reading test(s). The cut-off between dyslexics and non-dyslexics is arbitrary and signifies no qualitative differences between those on different sides of the border. By focusing on word reading problems, we distinguish between dyslexia and reading disability and argue that dyslexia is a subset of the latter. According to ICD-10, for example, specific reading disability requires impairment in reading comprehension, word recognition, oral reading, or in tasks that require reading. While dyslexia frequently leads to oral reading and reading comprehension problems, we suggest that dyslexia is present when the primary reading problem is at the level of words and the additional problems are either comorbid or secondary to the word reading difficulty. By persistent we mean that the difficulty has to be present over some period of time and not easily remedied by an alternative instructional method. For example, if some children fail to learn to read words in grade 1 with one instructional method but then make clear progress with a different instructional method, they would not qualify as dyslexics (see e.g., Vellutino et al., 1996). Instead, we would call them “teaching disabled” (following Tunmer & Greaney, 2010). Note, however, that persistence does not necessarily require early onset (see e.g., Catts, Compton, Tomblin & Bridges, 2012; Torppa, Eklund, van Bergen & Lyytinen, 2015). Finally, our definition includes the element of unexpectedness. Despite potential problems in operationalization, this is necessary for distinguishing dyslexia from word reading difficulties in general. Unexpectedness requires that we can establish reasonable expectations not simply based on age. One such basis could be oral language comprehension (cf. Tunmer & Greaney, 2010) but this is not typically included in definitions of dyslexia and may be problematic in that language skills and word reading are intertwined (e.g., Nation & Snowling, 1998; Ricketts, Nation & Bishop, 2007). Instead, many widely adopted definitions of dyslexia, such as those in DSM-5 and in ICD-10, that include the idea of unexpectedness state that poor general cognitive ability, sensory perception problems, or inadequate educational opportunities must be ruled out as possible causes of poor reading before a diagnosis of specific learning or reading disorder can be ascertained (see also International Dyslexia Association, 2002). Contrary to many definitions of dyslexia, we have no inclusion criteria. The most common inclusion criterion is an associated phonological processing deficit. We acknowledge that most individuals with dyslexia will exhibit a phonological processing deficit. However, a phonological deficit does not seem to be a necessary condition for dyslexia and many individuals with considerable word reading problems do not exhibit depressed phonological awareness scores (e.g., Georgiou, Parrila, Cui, & Papadopoulos, 2013; Pennington et al., 2012; Torppa et al., 2013; van Bergen, Bishop,



Dyslexia and word reading problems 335

van Zuijen & de Jong, 2015). Thus, it seems problematic to confine the use of the term – and the access to additional resources and accommodations that may come with it – to only those with phonological processing deficits, no matter how large a proportion they form of the total developmental dyslexia population.

Single-cause theories of developmental dyslexia A traditional and most common theoretical approach to developmental dyslexia has been to posit a specific deficit in some cognitive or perceptual process to account for word reading difficulties. The deficit is typically observed on a signature nonreading task (or a narrow set of tasks) that is meant to expose some crucial underlying weakness. In most cases, the posited deficit is meant to account for impairments in learning to read rather than for dysfunction in the cognitive mechanism of mature reading; therefore, such proposals are best viewed as developmental, rather than neuropsychological, accounts of reading difficulties. In the terminology of Castles and Coltheart (2004), these theories concern distal causes rather than proximal causes, necessitating additional theoretical steps (and empirical demonstrations) to link them with observed reading performance. The following discussion considers some important aspects of a subset of influential approaches. Further information can be found in recent reviews by Ramus and Ahissar (2012) and Elliot and Grigorenko (2014).

Phonological deficits The currently dominating theory of developmental dyslexia posits a “phonological deficit” at the core of the problem for all or a large majority of children with difficulties learning to read words (Bishop & Snowling, 2004; Ramus et al., 2003; Vellutino, Fletcher, Snowling, & Scanlon, 2004). The phonological family of approaches to reading difficulties is empirically based on a set of tasks known as phonological awareness tasks in which children are asked to segment, blend, delete, or otherwise manipulate phonemes in oral tasks not directly associated with reading. A specific causal link between deficits in phonological awareness and word reading difficulties has been difficult to demonstrate conclusively due to interactions between reading and phonological awareness skills, such as the implication of orthographic processing in phonological awareness tasks (Castles & Coltheart, 2004). However, this should not detract from the fact that poor performance in phonological awareness tasks is concurrently and longitudinally associated with dyslexia across languages (Ziegler & Goswami, 2005). In most phonological deficit theories, phonological awareness is assumed to be causally related to word reading because being able to deliberately individuate and identify phonemes in a spoken word is a prerequisite to consciously linking graphemes to those phonemes. However, additional theoretical steps are required to explain why reading difficulties persist past the initial stages or why phonological awareness

336 Rauno K. Parrila and Athanassios Protopapas

predicts advanced orthographic knowledge as measured by word recognition and spelling tests. Moreover, difficulty in phoneme individuation and identification, impeding grapho-phonemic decoding, would be expected to cause major difficulties in reading unfamiliar letter strings, such as novel words and pseudowords. Indeed, poor pseudoword decoding has long been considered a hallmark of dyslexia. It is, however, increasingly acknowledged that word recognition is the most severely affected domain, with larger effect sizes between typical and poor readers than pseudoword decoding (Van den Broeck & Geudens, 2012). Alternative approaches to a causal link between phonological awareness tasks and reading development focus on phonological representations, assuming that the representations of phonemes are impaired, somehow improperly or insufficiently specified (Perfetti, 1992; Ramus, 2003; Snowling, 2000). Poor phonemic representations account for poor phonological awareness and grapho-phonemic decoding because phonemic representations are needed both for conscious manipulation and for efficient mapping to graphemes. However, de-emphasizing the role of awareness overlooks the fact that tasks in which poor readers exhibit poor performance are consistently those in which phonological representations must be explicitly manipulated. There is little evidence that speech perception or production are affected, as would be expected if phonemic representations were impaired (Ramus & Ahissar, 2012). Perception studies have reported inconsistent findings, including somewhat poorer (e.g., Rosen & Manganari, 2001), no different (e.g., Hazan, Messaoud-Galusi, Rosen, Nouwens, & Shakespeare, 2009), or enhanced (e.g., Serniclaes, Van Heghe, Mousty, Carré, & Sprenger-Charolles, 2004) discrimination of speech sounds. The online uptake of acoustic information in matching lexical candidates also appears normal (Magnuson et al., 2011). Given these challenges to the representation account, a deficit in phonological access, rather than representation, has been proposed (Boets et al., 2013; Ramus & Szenkovits, 2008, 2009). According to this account, phonemes are properly specified but they are not efficiently accessible for operations such as those required for explicit phonological awareness tasks and for mapping between visual and phonological codes. It remains to be seen how this idea might account for the observed deficits once molded into specific theoretical hypotheses with associated empirical implications. The dominance of the “phonological deficit” in theorizing about reading difficulties remains undisputed and most individuals with developmental dyslexia perform poorly in the signature tasks; at the same time, many children with phonological deficits develop into adequate readers. Other single-cause approaches tend to acknowledge these facts and fall into two categories: In the first, complementary or alternative domains of impairment are posited to explain word reading problems of different types. In the second category, the phonological deficit itself is attributed to a more general or lower-level dysfunction.



Dyslexia and word reading problems 337

Rapid naming deficit A second major branch of theorizing is based on “rapid automatized naming” (RAN) tasks (Denckla & Rudel, 1976; Norton & Wolf, 2012; Wolf & Bowers, 1999; Wolf, Bowers & Biddle, 2000). In these tasks, participants are shown an array of symbols (letters, digits, color patches, or objects) and asked to name them aloud sequentially as fast as possible. RAN tasks have been described as “an early, simpler approximation of the reading process,” including “rapid, serial processing and integration of attentional, perceptual, conceptual, lexical, and motoric subprocesses” (Wolf et al., 2000, p. 393). The time to go through the entire array differentiates poor readers from good readers and is moderately to strongly correlated with word reading fluency, both concurrently and longitudinally, across languages and over a large age range (see review in Kirby, Georgiou, Martinussen, & Parrila, 2010). Low performance in RAN tasks is termed a “naming deficit” and is considered causal to reading performance as an additional critical factor not subsumed under phonological processing. RAN tasks are commonly thought to expose “rate problems” because of their multiple “processing speed requirements” (Wolf & Bowers, 1999), thereby constituting a preferred predictor for reading fluency. Stated this way, it sounds like a method variance issue, with a timed predictor accounting for time-limited measures, but RAN also predicts reading accuracy (Kirby et al., 2010; Parrila, Kirby & McQuarrie, 2004). Despite pronouncements regarding task complexity, in practice the theoretical emphasis has been placed on a general “processing speed” construct that affects cognitive components required for symbol processing. However, naming single symbols displayed individually is not a strong predictor of reading performance or reading difficulty (Jones, Branigan, & Kelly, 2009; Zoccolotti et al., 2013; Zoccolotti, De Luca & Spinelli, 2015). The array presentation format, with multiple stimuli displayed simultaneously, is critical to the predictive power of the task (Georgiou et al., 2013). Therefore processes involved in naming individual symbols cannot account for the RAN-reading relationship, regardless of their speed requirements. Instead, it must be the efficiency in sequentially naming an array of symbols that brings out variance uniquely related to reading (Gordon & Hoedemaker, 2016; Protopapas, Altani, & Georgiou, 2013). At the moment, the “naming deficit” theory concerns primarily an assessment issue rather than a coherent explanatory approach to reading development. Performance on RAN tasks helps identify children with reading difficulty that cannot be attributed to phonological deficits. Moreover, children with low RAN performance in addition to poor phonological awareness tend to be the poorest readers (Kirby, Parrila & Pfeiffer, 2003; Torppa et al., 2013) who likely benefit least from traditional intervention (Kirby et al., 2010). The well-documented relationships between RAN and reading remain largely unexplained (see Georgiou & Parrila, 2012, for a review). Although current research efforts focus on the crucial aspect of the task format (de Jong, 2011; van den Boer, Georgiou, & de Jong, 2016), it is not theoretically necessary that the difficulties

338 Rauno K. Parrila and Athanassios Protopapas

arise directly from the cognitive components responsible for processing sequences (Zoccolotti et al., 2015). Perhaps a relatively minor difficulty in naming individual symbols becomes exacerbated due to the relentless requirements for rapid integration among cognitive processes when going through a sequence of symbols. Access to phonological representations or visual symbol identification cannot provide the explanation because articulation of the symbol names is necessary for the crucial individual differences to emerge (Georgiou et al., 2013). Yet, articulation rates are not the answer either because silent intervals between symbols (“pause times”) within the RAN task are also correlated with reading (Georgiou, Aro, Liao, & Parrila, 2015). While RAN performance has been actively studied for years across multiple laboratories and languages, at present the “naming deficit” approach remains a placeholder for the future identification of putative cognitive and neural processes underlying efficient word and text reading.

Auditory processing The lack of direct evidence for impaired phonological representations notwithstanding, a number of research programs have sought to account for the deficient phonological representations by addressing either general auditory processing or speech-specific processing underlying phonetic perception. Each of these approaches has been based on a signature measure, or a narrow set of measures, in which significant differences are often found between groups with developmental dyslexia and groups of typically developing readers. For example, “rapid auditory processing” (Gaab, Gabrieli, Deutsch, Tallal & Temple, 2007; Tallal, 1980) is assessed with a “repetition test,” in which two brief stimuli are presented in rapid succession and the participant must report them in the correct sequence; “temporal sampling” (Goswami, 2011, 2015) is assessed with a “risetime perception test,” in which stimuli differing in onset abruptness must be distinguished; an “allophonic mode of speech perception” (Noordenbos, Segers, Serniclaes, & Verhoeven, 2013; Serniclaes et al., 2004) is assessed with categorical perception tasks including identification and discrimination of synthetic speech syllables; and so on. At the moment, robustness and interpretation of the initial results remains controversial (see Protopapas, 2014, and Ramus & Ahissar, 2012, for discussion and references). First, the purported deficits do not reliably emerge in every study; failures to replicate and partial replications abound. Second, despite the significant group differences, when individual performance is examined it is invariably found that a majority of participants in the reading impaired group perform within the range of performance of the typically developing group. This is unlike phonological awareness, low performance in which is typically observed for the majority of reading impaired individuals. Moreover, besides group differences in the target tasks, differences are systematically observed in tasks that, according to the theory being tested, should not be affected. Because the tasks pose complex cognitive requirements for successful



Dyslexia and word reading problems 339

performance, it remains plausible that performance differences may be attributable to perceptual or cognitive aspects of carrying out the task other than the hypothesized auditory processing requirements. Moreover, for a causal interpretation of differences in auditory processing, the hypothesized deficits must demonstrably precede and predict mediating deficits, which in turn must precede and predict reading difficulties. Precedence can only be established in longitudinal studies beginning at pre-reading ages (Boets et al., 2011), a goal not easy to achieve in practice. The existing comparisons to age-matched control groups confound performance with all kinds of experience and expertise associated with reading. Far from solving this problem, the alternative reading-level match designs confound group with age and can only reveal developmental and distributional aspects of the measures rather than theoretically important differences among individual children (Van den Broeck & Geudens, 2012). As it is becoming clear that phonological representations may not be impaired in the sense originally thought, it remains to be determined whether and how low performance in various psychophysical tasks may be involved in the formation of and access to phonological representations or otherwise in learning to read.

Visual attention An entirely different set of alternative approaches to explaining developmental dyslexia have focused on visual-spatial attention (Vidyasagar & Pammer, 2010). One proposal is based on a “letter span” task, in which a set of five letters is flashed briefly on the screen and the participant is asked to report either all of the letters or a single letter in a position cued after their disappearance. It is hypothesized that this task assesses the number of visual elements that can be processed simultaneously, as required for efficient reading. A “visual attention span” deficit is posited as complementary to the phonological deficit approach or it can exist independently and account for reading problems in the absence of phonological deficits (Bosse, Tainturier, & Valdois, 2007; Zoubrinetzky, Bielle, & Valdois, 2014). The use of letters as stimuli in the critical task admits alternative interpretations besides visual attention. For example, uptake of visual letter information may be limited due to insufficient reading experience, or inefficient graphophonemic connections may slow down activation of phonological codes for the letters through feedback loops. Diminished effects have been reported with stimuli other than letters or digits (Ziegler, Pech‐Georgel, Dufau, & Grainger, 2010), though subsequent visual categorization data (Lobier, Zoubrinetzky, & Valdois, 2012) and neuroimaging data (Lobier, Peyrin, Pichat, Le Bas, & Valdois, 2014) were unaffected by stimulus type. Still, the causal directionality of the visual attention span remains to be independently verified because reading practice may conceivably affect visual attentional efficiency and multi-element processing of non-alphanumeric stimuli as well (Dehaene et al., 2010).

340 Rauno K. Parrila and Athanassios Protopapas

A more recent proposal is based on a spatial cueing task, in which the location of a symbol on the screen is briefly cued prior to its appearance. Italian children with dyslexia were found to benefit less from correct cueing than typically developing readers (Facoetti et al., 2006, 2010). In a prospective longitudinal study, the performance of preschoolers on this task was found to predict their reading performance in grades 1 and 2. This has been interpreted as evidence for a fundamental deficit in orienting visual attention, termed “sluggish visual attention,” theorized to underlie letter and word recognition (Franceschini, Gori, Ruffino, Pedrolli, & Facoetti, 2012). However, this proposal is not specifically related to letter and word identification performance. Rather, the link to word recognition is through phonological decoding, via a multimodal attention mechanism that mediates efficient orthographic-phonological binding (Gori & Facoetti, 2015).

Discussion of single-deficit theories Most of the single-deficit theories of dyslexia offer little more than observation of an association between poor reading performance and low performance in a signature nonreading task (or in a narrow set of tasks). The most successful among them, the phonological deficit and the naming speed deficit theories, include evidence for longitudinal associations and findings that are more robust across studies. However, much more will be needed before a “theory of dyslexia” can be proclaimed. We highlight here two obstacles evident in current theorizing about dyslexia, namely, (a) understanding task performance and (b) constructing causal theories of reading development that involve the necessary theoretical constructs to connect reading with the signature tasks. Signature task performance is often taken to index a specific target construct, ignoring other cognitive and perceptual requirements for successful performance. For example, performance on phonological awareness tasks may be taken to index the quality of phonological representations even though successful task performance also requires accurate perception, retention in short-term and working memory, manipulation, and formation and execution of an articulatory response. Weaknesses in any of these steps or in their integration will impair task performance without necessarily involving poor quality of the phonological representations. Similarly, performance in rapid naming tasks, risetime perception tasks, letter report tasks, etc., will necessarily involve a multitude of perceptual and cognitive processes and representations, any of which might be implicated in poor performance. Importantly, the “weak link” need not be a single step or process: perhaps two or more elements might need to be compromised before performance decrements can be observed. This possibility cannot be addressed in the absence of in-depth task analyses. The validity of a task indexing a construct cannot be determined a priori. Convergent and divergent validity must be demonstrated by reference to additional tasks that do or do not share the purported critical representation or process. For a theoretical proposal to stand on solid ground, the crucial theoretical constructs must be properly operationalized by their empirical indices. This must include a wide range of



Dyslexia and word reading problems 341

tasks hypothesized to involve the construct in question. Crucially, it must also exclude tasks of similar form and comparable difficulty that do not involve the construct. In the absence of well-defined constructs, theoretical connections are posited in a vacuum. Second, the theoretical constructs must be unambiguously implicated in a causal theory that clearly shows how they underlie poor reading performance. Admittedly, it is not clear how this can be achieved in the absence of a well-defined theory of reading development and reading performance. Still, it behooves the proponents of specific theories of dyslexia to explain how learning to read depends on the hypothesized theoretical constructs and to demonstrate the dependence in properly controlled longitudinal and experimental studies that simultaneously assess convergent and divergent construct validity and also include tasks assessing alternative hypotheses for the same individual differences. Intervention studies in particular can be effective if training in a distal domain can be shown to affect reading skill. However, it must be clearly demonstrated that the effect arises specifically due to the theoretically hypothesized aspect of training. This is only achievable in the context of not simply active control groups but of a well-matched control training regime that differs only in the theoretically critical feature from the proposed intervention.

Acknowledging heterogeneity: Subtypes of dyslexia One principled approach to the heterogeneity of word reading difficulties has been through subtyping. That is, if a single-deficit theory cannot account for all cases of reading failure, perhaps a combination of theories will. There may be two (or more) kinds of problems potentially impeding word reading development, therefore two (or more) kinds of developmental dyslexia. As a consequence, children can still be “dyslexic” if their word reading is below par, but assigned to “type A,” “type B,” or a “combined A+B” dyslexic group depending on some relevant profiling. The categorization may be based on measures of reading performance or on other cognitive, linguistic, or perceptual indices, including the signature tasks discussed above. We review here the most prominent examples of such subtyping.

Phonological vs. surface dyslexia Perhaps the best-established approach to subtyping is based on a theoretical distinction originally drawn in neuropsychological English-speaking patients (Coltheart, 2012). Specifically, it was observed that, following brain damage, certain patients have reading difficulties that manifest differentially with different types of letter strings: Some patients have more difficulty with unfamiliar or made-up words (pseudowords) whereas other patients have more difficulty with familiar but inconsistent words, that is, words that are pronounced differently from other words with similar spelling patterns (Woollams, 2014).

342 Rauno K. Parrila and Athanassios Protopapas

The distinction between consistent and inconsistent words is especially relevant for the English orthography, in which there is a great range of graphophonemic consistency, including words with pronunciation hardly licensed by their spelling, such as yacht and rough. An absolute division between “regular” and “irregular” words has been imposed on a continuum of consistency, based on the theoretical assumption of content-independent “rules” for graphophonemic conversion. Any word not fully pronounceable by the rules is termed “irregular” regardless of its relations to other words. For example, the word pint is deemed irregular for failing to adhere to the same pattern as mint and hint, even though the usual mappings hold for three out of its four letters and pronunciation of the letter i in this context is consistent with the word pine (Plaut, McClelland, Seidenberg, & Patterson, 1996). The developmental plausibility and cross-linguistic relevance of this rule-based distinction remain controversial. The differential patterns of impairment seen in patients with acquired dyslexia, in conjunction with the theoretical hypothesis of absolute graphophonemic rules, have led to a two-pronged approach to word reading, including a “nonlexical” route assigned to applying rules and a “lexical” route to recognize familiar words (Coltheart, Rastle, Perry, Langdon, & Ziegler, 2001). Accordingly, two alternative routes to failure can be posited: Damage to the nonlexical route impedes graphophonemic conversion whereas damage to the lexical route impedes recognition of familiar words. The effects of damage would be most obvious on pseudowords and irregular words, respectively, because these can only be read by the corresponding route. In contrast, regular words can be read correctly by either route, and therefore they are not diagnostic. Children with difficulties in pseudoword reading are termed “phonological” dyslexics whereas those with difficulties in irregular word reading are termed “surface” dyslexics, consistent with the classification of neuropsychological patients (Castles & Coltheart, 1993). Although theoretically attractive in its simplicity, this proposal has met with empirical difficulty in that the vast majority of children with reading difficulty exhibit low performance with all kinds of words and pseudowords (Manis, Seidenberg, Doi, McBride-Chang, & Petersen, 1996), raising concerns about the parsimony of positing simultaneous impairment in both routes. The surface subtype has also been elusive in investigations using reading-level match designs (Manis et al., 1999; Stanovich, Siegel, & Gottardo, 1997). As for the phonological subtype, it has long been known that the pseudoword reading deficit is contingent on item difficulty, emerging mainly with complex pseudowords that do not resemble words (Rack, Snowling, & Olson, 1992). Recent advances in understanding the psychometric issues in comparing performance decrements across domains, such as between words and pseudowords (Van den Broeck & Geudens, 2012), have further undermined the potential for defining subtypes on the basis of relative performance in such tasks.



Dyslexia and word reading problems 343

Double-Deficit theories Proponents of visual and naming deficit theories reviewed above have taken a slightly different approach to subtyping. Low performance in the signature task (letter span or rapid naming) is attributed to an underlying deficit that impedes reading development independently of phonological deficits. Therefore, the visual or naming deficits are seen as alternative or additional to the phonological deficit, supporting classification into four quadrants: Children without difficulty, children with a phonological deficit, children with a visual/naming deficit, and children with a double deficit (Bosse et al., 2007; Wolf & Bowers, 1999). This permits explanation of poor reading in the absence of phonological problems, which remains a thorny issue for proponents of the phonological core approach. The double deficit approaches also provide an additional dimension of severity, possibly associated with poorer response to intervention, insofar as children with a double deficit would suffer from more pervasive and severe difficulties, which may also be more difficult to ameliorate (Wolf et al., 2000). How deficits are associated with reading performance varies across the double deficit approaches. In the original double deficit theory of Wolf and Bowers (1999), naming speed deficits impede the development of efficient word recognition, whereas phonological deficits affect the development of accurate decoding; combined deficits impede reading development on both fronts, resulting in more severe difficulties. Bosse et al. (2007), in turn, suggested that visual attention deficits impair visual word processing more generally, manifesting themselves in decoding as well, even in the absence of frank phonological deficits. Thus, the visual attention span is proposed to account for reading problems potentially including those typically attributed to impaired phonological processing. Nevertheless, visual attention span problems are expected to affect sight word reading more severely than phonological decoding. Finally, proponents of visual attention deficits based on spatial cueing tasks have claimed that deficits in spatial attention specifically affect processes related to phonological decoding (Ruffino, Gori, Boccardi, Molteni, & Facoetti, 2014). In this approach, subtypes do not distinguish among patterns of reading performance but, rather, among patterns of cognitive skills that underlie similar difficulties in word reading.

Discussion of subtyping theories The move from a single phonological-core approach towards encompassing alternative cognitive substrates of word reading difficulties seems welcome in the context of the established cognitive heterogeneity in the low end of the word reading performance spectrum. Freedom from the tyranny of a single cause may pave a path toward a more pervasive acceptance of a multitude of potential routes to word reading difficulties. However, theories defining subtypes on the basis of low performance in specific signature tasks are more strongly related to single-cause theories than to multiple deficit alternatives discussed below in that they seek to identify circumscribed, distinct

344 Rauno K. Parrila and Athanassios Protopapas

causes for specific patterns of word reading difficulties. Thus, they also inherit problems associated with single-cause approaches, including those stemming from the heterogeneity of reading problems, which may not fit entirely within any particular classification. For example, the alternatives reviewed above would already lead one to expect a future triple- or quadruple-deficit theory of dyslexia. Moreover, subtyping proposals raise additional concerns regarding the reliability and stability of the classification, in addition to the validity of the constructs underlying the classification. Reliability is difficult to assess conclusively due to the arbitrariness of cutoff points placed on continuous performance distributions. As task performance is inherently noisy, the reliability of classification depends on the reliability of the signature tasks. Here, issues of theoretical importance (e.g., intraindividual variability) may be dismissed as trivial measurement noise. Stability of subtyping has been found to be moderate for surface vs. phonological dyslexia (Manis et al., 1999) and for naming speed vs. phonological deficit (Steacy, Kirby, Parrila, & Compton, 2014). Although studies addressing a specific distinction may not generalize to other subtyping approaches, it behooves the proponents of subtypes to demonstrate the reliability and stability of classification over large and representative populations across languages.

Multiple-Deficit models of dyslexia The evidence reviewed above suggests that single-, double- or triple-cause theories of developmental dyslexia are unlikely to provide satisfactory explanations of dyslexia as a behaviorally defined developmental disorder. There is now widespread consensus that the term “dyslexia” refers to the low end of a word reading distribution rather than to a discrete condition. If there is no discrete condition that accounts for the great majority of children with word reading difficulties, the search for specific causes is greatly challenged. Rather, the focus is turning to multifactorial developmental pathways that can give rise to brains that differ in their propensity for learning to read (or learning math, or acquiring any other cultural artifact our evolution has not specifically equipped us for). Individual differences in language and reading development are increasingly attributed to a multitude of interacting genetic, neural, cognitive, behavioral, and environmental factors, potentially leading to high or low reading performance via multiple developmental pathways. In some theoretical models, cognitive multiplicity is assumed to reduce to a single genetic or neural cause for developmental dyslexia, but these calls for simplicity face sizeable empirical challenges. The genetic studies of dyslexia started with the expectation of dominant inheritance controlled by a single gene (Hallgren, 1950). Specific susceptibility genes were sought with single-gene strategies (such as genetic linkage, targeted association, and chromosome translocation or deletion). Instead, multiple loci with multiple susceptibility genes have been identified (Kere, 2014). Molecular and behavior genetic studies of dyslexia now agree that the genetic architecture associated with dyslexia is complex, polygenic (two or more genes contribute to the phenotype),



Dyslexia and word reading problems 345

and heterogenic (the same behavioral outcome can be associated with multiple different causes; Carrion-Castillo, Franke, & Fisher, 2013; Elliot & Grigorenko, 2014). There are examples of families where the inheritance pattern is consistent with a rare mutation of a single gene (e.g., de Kovel et al., 2004; Nopola-Hemmi et al., 2000). However, all genes identified in rare familial forms of dyslexia jointly explain a tiny fraction of the variance in reading ability when tested with larger samples. Instead, most cases of dyslexia are probably affected by a very large number of genes, each with only weak effects (Carrion-Castillo et al., 2013), further complicated by a multitude of gene × gene and gene × environment interactions (Bishop, 2015; Jablonka & Lamb, 2014). Neural level examinations have not fared much better in simplifying dyslexia theories. Recent meta-analyses of functional neuroimaging studies (Maisog, Einbinder, Flowers, Turkeltaub, & Eden, 2008; Richlan, Kronbichler, & Wimmer, 2009, 2011) have identified more than a hundred foci of differences between dyslexic and normally developing readers. The results from the meta-analyses were mostly consistent with the typical neurophysiological account of developmental dyslexia for adults (e.g., Pugh et al., 2000) but highlighted the need for refinement in the developmental account. In particular, Richlan et al. (2011) found no brain areas typically associated with phonological coding to be reliably underactivated across studies. Their results did not support the assumption that the primary and early emerging dysfunction resides in the left ­temporo-parietal cortex housing the dorsal reading subsystem. Instead, they suggested that an early and limited left occipito-temporal dysfunction becomes extended over time and is accompanied by a left temporo-parietal dysfunction by adulthood. As individual functional imaging studies continue to produce widely varying results, it may be necessary to examine how much of this variability is related to heterogeneity (beyond age) in the dyslexia samples and how much is related to variability in imaging and analysis methods and the signature tasks used in different laboratories. Notably, substantial variability is not limited to functional imaging but is also present in structural imaging, as demonstrated in meta-analyses of voxel-based morphometry studies of gray matter (Linkersdörfer, Lonnemann, Lindberg, Hasselhorn, & Fiebach, 2012; Richlan, Kronbichler, & Wimmer, 2013; see review in Jednorog et al., 2015) and white matter (Vandermosten, Boets, Wouters & Ghesquière, 2012) comparing disabled and typically developing readers. As structure and function of the brain are both altered by experience (Gabrieli, 2009; Krafnick, Flowers, Luetje, Napoliello, & Eden, 2014; Simos et al., 2002), the relevant heterogeneity in the developmental dyslexia samples is not limited to the reading and cognitive measures usually used to identify the dyslexics.

Probabilistic multiple deficit models The assumption that behaviorally defined developmental disorders can have a single cause at any level of analysis has been further challenged by multifactorial etiological models (e.g., Gottlieb & Halpern, 2002; Lyytinen et al., 1998; Pennington, 2006; van Bergen, van der Leij, & de Jong, 2014). In an influential paper, Pennington (2006)

346 Rauno K. Parrila and Athanassios Protopapas

reviewed genetic, neural, cognitive, and comorbidity studies of developmental dyslexia and concluded that converging evidence precipitates a major reconceptualization of the existing theoretical models. He argued that probabilistic multiple deficit models (PMDM) are needed to provide realistic accounts of developmental disorders, their comorbidity, and the nondeterministic relationships between disorders and their presumed causes. He suggested further that such PMDMs must include protective and risk factors, multiple levels of analysis, bidirectional connections between constructs within each level (horizontal or intralevel interactions), and bidirectional connections between levels (vertical or interlevel interactions) to account for interactions between protective and risk factors functioning at different levels of analysis (see also, Ford & Lerner, 1992; Gottlieb, 1983, 1997; Gottlieb, Wahlsten, & Lickliter, 2006; Lyytinen et al., 1998). Figure 1 shows a simplified version of Pennington’s (2006) PMDM with interlevel connections omitted. The left side of the model shows the levels of analysis – etiological, neural, cognitive, and behavioral – and the right side displays the mechanisms that underlie horizontal interactions in each level. According to Pennington, the etiological level of any behaviorally defined developmental disorder – including dyslexia – is multifactorial and involves interaction of multiple risk and protective factors that can be either genetic or environmental; these jointly and probabilistically influence the development of neural systems and, further, the cognitive processes they support. At the behavioral level, the disorder is jointly and probabilistically produced by multiple Level of Analysis

Etiologic Risk and Protective Factors

Neural Systems

Cognitive Processes

Complex Behavioral Disorders

Non-independence at each level G1

E1

G2

E2

G3

G×E interaction & G-E correlation

N1

N2

N3

Pleiotropy

C1

C2

C3

Interactive development

D1

D2

D3

Comorbidity

KEY: G = genetic risk or protective factor; E= environmental risk or protective factor; N = neural system; C = cognitive process; D = disorder

Figure 1.  Pennington’s (2006) Probabilistic Multiple Deficit Model. Note that causal connections (and feedback loops) between levels not shown in the figure but acknowledged in the text.



Dyslexia and word reading problems 347

Assortative mating GTm

PTm

PTp

GTp

Parents

Child

Evocative rGE and other env. influences

Passive rGE and cultural & genetic transmission

Active rGE

Extra-parental environment

Shared environment confound

Evocative rGE

G1

G2

G3

N1

N2

N3

Neural Systems

C1

C2

C3

Cognitive Processes

D1

D2

D3

Genetic factors

Behavioural disorders

Figure 2.  van Bergen, van der Leij and de Jong’s (2014) intergenerational multiple deficit model (see van Bergen et al., 2014, for detailed descriptions of components).

cognitive risk and protective factors, each influenced by multiple etiological factors. Some of the etiological risk and protective factors influence several disorders (causing comorbidity) whereas others are specific to one disorder. No single etiological, neural, or cognitive factor is sufficient; a combination of several may be necessary to produce the behavioral symptoms that define the disorder. Finally, the liability distribution for any given disorder is continuous. An individual’s position on the distribution is affected by risk and protective factors at any level. Recently, van Bergen et al. (2014) extended Pennington’s PMDM to allow for intergenerational transfer of risk and protective factors. To explicitly account for parental effects, they proposed the intergenerational multiple deficit model (iMDM; see Figure 2), which includes not only genetic transmission from parents to children but also passive and evocative gene-environment correlations and cultural transmission from parents. In Figure 2, environment as shaped and selected by the parents is separate from extra-parental environment and from genetic effects in the etiological level of Pennington’s formulation. In the iMDM, parental skills (as expressed in their phenotype, PT) are transmitted both genetically and via the home environment. This extended home environment can exert a direct effect (cultural transmission in Figure 2) on children in that parents’ cognitive phenotypes impact the environments they can

348 Rauno K. Parrila and Athanassios Protopapas

offer for their children, including the protective and risk factors in these environments. If home environment correlates with parents’ and child’s genotype, we observe passive gene-environment correlation (rGE in Figure 2); these influences could, for example, include highly literate parents having a higher income and thus access to better schools. Evocative rGE arises when some of the child’s highly heritable characteristics (e.g., good phonological awareness) elicit a response from the environment (e.g., more rhyming and alliteration games and playing with letters) that further strengthens the child’s reading development. Shared environmental confound contributes to parent-child resemblance by affecting reading ability in both generations. For example, access to education, particularly for females, is a significant shared environmental confound in many parts of the world. Another such confound could be a shared home language that is different from the language of education. Separating environment from genes as etiological factors allowed van Bergen et al. to also differentiate between environment as shaped by parents and extra-parental environment that parents have less influence over; this could include reading instruction method, access to print and digital media, peer influences, legislation of special provisions and resources in the schools for students with dyslexia, and the value of literacy in the society at large. Van Bergen et al. also distinguish between active rGE (e.g., a child who learns to read easily is motivated to read more and seeks out opportunities to do so in the environment) and evocative rGE, where children’s genetically influenced ability elicits differential reactions from the environment, such as good readers being given more demanding materials to read. In the educational literature, a similar difference is made between the active and evocative impact a child can have on teachers’ behavior (e.g., Nurmi, 2012).

Discussion of multiple deficit models Multiple deficit models provide an interesting meta-theoretical framework to advance developmental dyslexia theorizing and research. These models are examples of dynamic or developmental systems models that have a long history in developmental embryology and biology (see e.g., Gottlieb, 2002) and have permeated developmental sciences for some time (see e.g., Ford & Lerner, 1992, and Thelen & Smith, 1994, for introductions to earlier approaches, and Molenaar, Lerner & Newell, 2014, for newer formulations). While several prominent authors have recently acknowledged the limitations of traditional models (e.g., Catts, this volume; Snowling & Melby-Lervåg, 2016), systems approaches in general have had little traction in dyslexia research (however, see Morrison & O’Connor, this volume, for an example of a systems approach to reading development), perhaps because systems approaches pose formidable empirical challenges and theoretical questions the answers to which poorly match our dominant research traditions and presuppositions. While there are multiple families of developmental or dynamic systems theories, the PMDM models as outlined by Pennington (2006) and van Bergen et al. (2014)



Dyslexia and word reading problems 349

bear a close theoretical resemblance to developmental psychobiological systems theories (e.g., Gottlieb et al., 2006). In these theories, development is conceptualized as a sequential emergence of new structural and functional properties and competencies at all levels of analysis as a consequence of horizontal and vertical interactions among the parts (Gottlieb et al., 2006). This implies that any causal explanation of a developmental outcome, such as dyslexia, must describe the developmental system that, over time, led to the observed outcome. While we can study components of the system, such as rapid naming tasks or individual genes, in relative isolation, individual components neither explain nor cause (normal or abnormal) development in any meaningful sense without an account of the rest of the system; such an account must include the organism and the physical, biological, and social factors (“developmental niche”) that interact with and shape it over time. As the components, or interactants, in these developmental networks are themselves largely products of earlier development, developmental explanations require that we study their interactions over a period of time. As a result, we need to rethink both what constitutes an explanation and what kinds of observations are required to understand the developmental pathways to the observed outcomes. As developmental systems models, PMDMs inherit the idea that an explanation requires understanding the developmental system with all of its risk and protective factors. Thus, the unit of explanation is not an individual but a relational causal network. Relational causal networks include the idea that no single element or level in the developing system has causal primacy, and the functional significance of any element on development can only be understood in the context of the developmental system of which they are part. At each level of the developmental system, the effect of any element is dependent on the rest of the system, making all factors potentially interdependent and mutually constraining (Gottlieb, 1991). The PMDMs reviewed above imply this kind of multidimensionality where assigning causal priority to any level is problematic. However, these ideas seem difficult to reconcile with current theorizing that assigns causal priority at the genetic level and assumes that genetic and environmental influences are additive. In contrast, the developmental systems view leads to conceptualizing genetic and environmental effects as interdependent. At least in principle, genetic (and other inherited) effects leading to individual differences cannot be understood apart from development occurring in a specific environmental context. Complex human behaviors, such as reading, “are influenced by hundreds or thousands of proteins encoded in hundreds or thousands of genes of small effect that interact with one another, the environment, and the epigenome in complex ways” (Charney & English, 2012, p. 30). What then constitutes an explanation of development and developmental outcome is much more complex than most developmental models in developmental dyslexia research currently acknowledge. Methodologically, we need to supplement current nomothetic variable-centered studies with idiographic studies and person-centered analyses (see examples in Molenaar et al., 2014). Nomothetic variable-centered studies that have driven most

350 Rauno K. Parrila and Athanassios Protopapas

of the theory development described above are informative of general tendencies and the components that explanatory models need to include, but they cannot predict how the process of development unfolds over time. We suggest that we need to develop and test “dynamic mechanistic explanations” (Bechtel & Abrahamsen, 2005, 2010) of how different interactants work together in producing observable outcomes. While we currently have few such models to build on (see Giraud & Ramus, 2013, for an exception), these kinds of models are not uncommon in other fields of inquiry (see, e.g., Becher et al., 2014, for a complex model of honeybee colony dynamics). Dynamic mechanistic explanations require longitudinal and experimental studies, going beyond individual differences to observations of developmental mechanisms, as well as computational modeling (e.g., agent-based models, see Railsback & Grimm, 2011) of these mechanisms grounded in empirical observations and aiming to understand their functioning where observations and experimentation are not possible. For example, if we conceptualize genetic, neural, and cognitive interactants suggested by Giraud and Ramus (2013) as agents and model their functioning and interactions with agentbased models (see e.g., Railsback & Grimm, 2011; Wilensky & Rand, 2015; examples at https://ccl.northwestern.edu/netlogo/) that first simulate what we already know, we can then start to posit mechanisms and conditions under which they operate. As we learn more, the models will get more complex by having to include simulations of new empirical findings and they will produce new hypotheses to examine empirically (see Bechtel & Abrahamsen, 2010, for an example of increasing model complexity). The scope of the studies does not have to be any more expansive than the studies we already conduct, but when we use models as first approximations of the developmental interactions between the components we observe, the explicitness of our theories will increase because we will have to focus on the mechanisms and not only on associations among measures. Finally, the notion of equifinality has become an axiom of developmental systems theory (see e.g., Ford & Lerner, 1992; Gottlieb et al., 2006). In this view, organisms with different early – or “initial” – conditions can reach the same endpoint and organisms with the same initial conditions can take different routes or pathways to reach a common endpoint. Equifinality is an important principle in psychological development, but the concept is seldom discussed in developmental dyslexia research. However, if PMDM models are interpreted as developmental systems theories, it follows that development is influenced by many risk and protective factors that interact to produce the reading behavior we use to diagnose dyslexia. Such complex probabilistic networks are bound to produce similar observable states with different interactants (e.g., different explanations of word reading failure; Snowling & Melby-Lervåg, 2016). Examples are bound to proliferate with increased emphasis on person-centered and idiographic methods.



Dyslexia and word reading problems 351

Conclusion Theories of developmental dyslexia cannot simply be theories of individual differences in word reading development; instead, they need to progress towards dynamic mechanistic explanations of various developmental pathways to inaccurate or inefficient word reading. The former are general and meant to account primarily for associations among general constructs, effectively describing the average situation that may not apply to any individual (see e.g., Velicer, Babbin, & Palumbo, 2014). However, if the goal is to develop a theory of dyslexia as a specific condition rather than a diagnostic label, then we need a much more specific theory, of a more applied nature. This theory should account for each and every child that deservedly receives the diagnostic label “dyslexia” following extensive testing, examination, and possibly failed intervention. There is no room for letting some children slip. However large and heterogeneous, groups of children with dyslexia must be fully accounted for by any theory purporting to be a theory of dyslexia. Otherwise it is not really a theory of dyslexia but maybe a theory of some of the difficulties of some of the children who fail to learn to read words. The requirement for full diagnostic coverage seems extremely unlikely to be satisfied by any approach focusing on single causes or single-factor characterizations, and it seems very likely to include multiple pathways to the same behavioral condition. The above discussion has indicated a number of potential interactants from the genetic through the environmental level that need to be considered in a developmental systems theory of developmental dyslexia. These interactants were identified in variable-centered studies focusing on individual differences because (a) we know to measure them, and (b) they co-vary sufficiently with the dependent variable in the examined samples. We suspect that the interactants we currently know of as “actual difference makers” (Waters, 2007) are but a small subset of those needed for a dynamic mechanistic theory of developmental dyslexia. We may have already included some “potential difference makers” (Waters, 2007; see also Griffiths & Tabery, 2013, and Tabery, 2014) in our empirical studies but failed to recognize their significance for a developmental theory of dyslexia because they either did not vary sufficiently to produce the statistical association, or the heterogeneity of their expression in the samples drowned the signal. However, there are undoubtedly more potential difference makers at all levels of analysis that are yet to be identified. Here is where single-case studies and exploratory computer modeling can capture extreme cases to enhance theory development. The truly developmental science of developmental dyslexia requires that we “deconstruct” the phenomenon at each level into its constituents, but also that we then attempt to reconstruct the developing system to test hypotheses about interactions between levels and mechanisms of effect. The theories we want are the ones that not only explain why a child with dyslexia reads differently from another child with or without dyslexia, but also where in that developmental system we can intervene successfully.

352 Rauno K. Parrila and Athanassios Protopapas

References Becher, M. A., Grimm, V., Thorbek, P., Horn, J., Kennedy, P. J. & Osborne, J. L. (2014). BEEHAVE: A systems model of honeybee colony dynamics and foraging to explore multifactorial causes of colony failure. Journal of Applied Ecology, 51, 470–482.  doi: 10.1111/1365-2664.12222 Bechtel, W. & Abrahamsen, A. (2005). Explanation: A mechanistic alternative. Studies in History and Philosophy of Biological and Biomedical Sciences, 36, 421–441. doi: 10.1016/j.shpsc.2005.03.010 Bechtel, W. & Abrahamsen, A. (2010). Dynamic mechanistic explanation: Computational modeling of circadian rhythms as an exemplar for cognitive science. Studies in History and Philosophy of Science, 41, 321–333.  doi: 10.1016/j.shpsa.2010.07.003 Bishop, D. V. M. (2015). The interface between genetics and psychology: Lessons from developmental dyslexia. Proceedings of the Royal Society B, 282: 20143139.  doi: 10.1098/rspb.2014.3139 Bishop, D. V. M., & Snowling, M. J. (2004). Developmental dyslexia and specific language impairment: Same or different? Psychological Bulletin, 130, 858–886.  doi: 10.1037/0033-2909.130.6.858 Boets, B., de Beeck, H. P. O., Vandermosten, M., Scott, S. K., Gillebert, C. R., Mantini, D., … & Ghesquière, P. (2013). Intact but less accessible phonetic representations in adults with dyslexia. Science, 342, 1251–1254.  doi: 10.1126/science.1244333 Boets, B., Vandermosten, M., Poelmans, H., Luts, H., Wouters, J., & Ghesquière, P. (2011). Preschool impairments in auditory processing and speech perception uniquely predict future reading problems. Research in Developmental Disabilities, 32, 560–570. doi: 10.1016/j.ridd.2010.12.020 Bosse, M. L., Tainturier, M. J., & Valdois, S. (2007). Developmental dyslexia: The visual attention span deficit hypothesis. Cognition, 104, 198–230.  doi: 10.1016/j.cognition.2006.05.009 Carrion-Castillo, A., Franke, B. & Fisher, S. E. (2013). Molecular genetics of dyslexia: An overview. Dyslexia, 19, 214–240.  doi: 10.1002/dys.1464 Castles, A., & Coltheart, M. (1993). Varieties of developmental dyslexia. Cognition, 47, 149–180. doi: 10.1016/0010-0277(93)90003-E Castles, A., & Coltheart, M. (2004). Is there a causal link from phonological awareness to success in learning to read? Cognition, 91, 77–111.  doi: 10.1016/S0010-0277(03)00164-1 Catts, H. W., Compton, D., Tomblin, J. B., & Bridges, M. S. (2012). Prevalence and nature of late-emerging poor readers. Journal of Educational Psychology, 105, 166–181. doi: 10.1037/a0025323 Charney, E. & English, W. (2012). Candidate genes and political behavior. American Political Science Review, 106, 1–34.  doi: 10.1017/S0003055411000554 Coltheart, M. (2012). Dual-route theories of reading aloud. In J. S. Adelman (Ed.), Visual word recognition, volume 1: Models and methods, orthography and phonology (pp. 3–27). Hove, UK: Psychology Press. Coltheart, M., Rastle, K., Perry, C., Langdon, R., & Ziegler, J. (2001). DRC: A dual route cascaded model of visual word recognition and reading aloud. Psychological Review, 108, 204–256.  doi: 10.1037/0033-295X.108.1.204 de Kovel, C. G., Hol, F. A., Heister, J. G., Willemen, J. J., Sandkuijl, L. A., Franke, B.,…Padberg, G. W. (2004). Genomewide scan identifies susceptibility locus for dyslexia on Xq27 in an extended Dutch family. Journal of Medical Genetics, 41, 652–657.  doi: 10.1136/jmg.2003.012294 Dehaene, S., Pegado, F., Braga, L., Ventura, P., Nunes Filho, G., Jobert, A., … Cohen, L. (2010). How learning to read changes the cortical networks for vision and language. Science, 330, 1359–1364.  doi: 10.1126/science.1194140



Dyslexia and word reading problems 353

de Jong, P. F. (2011). What discrete and serial rapid automatized naming can reveal about reading. Scientific Studies of Reading, 15, 314–337.  doi: 10.1080/10888438.2010.485624 Denckla, M. B., & Rudel, R. G. (1976). Rapid ‘automatized’ naming (R.A.N.): Dyslexia differentiated from other learning disabilities. Neuropsychologia, 14, 471–479. doi: 10.1016/0028-3932(76)90075-0 Elliott, J. G., & Grigorenko, E. L. (2014). The dyslexia debate. Cambridge University Press. doi: 10.1017/CBO9781139017824 Facoetti, A., Trussardi, A. N., Ruffino, M., Lorusso, M. L., Cattaneo, C., Galli, R., … & Zorzi, M. (2010). Multisensory spatial attention deficits are predictive of phonological decoding skills in developmental dyslexia. Journal of Cognitive Neuroscience, 22, 1011–1025. doi: 10.1162/jocn.2009.21232 Facoetti, A., Zorzi, M., Cestnick, L., Lorusso, M. L., Molteni, M., Paganoni, P., … & Mascetti, G. G. (2006). The relationship between visuo-spatial attention and nonword reading in developmental dyslexia. Cognitive Neuropsychology, 23, 841–855.  doi: 10.1080/02643290500483090 Ford, D. H. & Lerner, R. M. (1992). Developmental systems theory: An integrative approach. Newsbury Park, CA: Sage. Franceschini, S., Gori, S., Ruffino, M., Pedrolli, K., & Facoetti, A. (2012). A causal link between visual spatial attention and reading acquisition. Current Biology, 22, 814–819. doi: 10.1016/j.cub.2012.03.013 Gaab, N., Gabrieli, J. D. E., Deutsch, G. K., Tallal, P., & Temple, E. (2007). Neural correlates of rapid auditory processing are disrupted in children with developmental dyslexia and ameliorated with training: An fMRI study. Restorative Neurology and Neuroscience, 25, 295–310. Gabrieli, J. D. E. (2009). Dyslexia: A new synergy between education and cognitive neuroscience. Science, 325, 280–283.  doi: 10.1126/science.1171999 Georgiou, G. K., Aro, M., Liao, C. H., & Parrila, R. (2015). The contribution of RAN pause time and articulation time to reading across languages: Evidence from a more representative sample of children. Scientific Studies of Reading, 19, 135–144.  doi: 10.1080/10888438.2014.956927 Georgiou, G. & Parrila, R. (2012). Rapid naming and reading. In L. Swanson, K. Harris & S. Graham (eds.), Handbook of learning disabilities (2nd ed., pp. 169–185). New York: Guilford. Georgiou, G. K., Parrila, R., Cui, Y., & Papadopoulos, T. C. (2013). Why is rapid automatized naming related to reading? Journal of Experimental Child Psychology, 115, 218–225. doi: 10.1016/j.jecp.2012.10.015 Giraud, A.-L. & Ramus, F. (2013). Neurogenetics and auditory processing in developmental dyslexia. Current Opinion in Neurobiology, 23, 37–42.  doi: 10.1016/j.conb.2012.09.003 Gordon, P. C., & Hoedemaker, R. S. (2016). Effective scheduling of looking and talking during rapid automatized naming. Journal of Experimental Psychology: Human Perception and Performance, 42, 742–760. Gori, S., & Facoetti, A. (2015). How the visual aspects can be crucial in reading acquisition? The intriguing case of crowding and developmental dyslexia. Journal of Vision, 15, 1–20. doi: 10.1167/15.1.8 Goswami, U. (2011). A temporal sampling framework for developmental dyslexia. Trends in Cognitive Sciences, 15, 3–10.  doi: 10.1016/j.tics.2010.10.001 Goswami, U. (2015). Sensory theories of developmental dyslexia: Three challenges for research. Nature Reviews Neuroscience, 16, 43–54.  doi: 10.1038/nrn3836 Gottlieb, G. (1983). The psychobiological approach to developmental issues. In M. M. Haith & J. J. Campos (Eds.), Handbook of child psychology (4th ed., vol. 2, pp. 1–26). New York: John Wiley & Sons.

354 Rauno K. Parrila and Athanassios Protopapas

Gottlieb, G. (1991). Experiential canalization of behavioral development: Theory. Developmental Psychology, 27, 4–13.  doi: 10.1037/0012-1649.27.1.4 Gottlieb, G. (1997). Synthesizing nature-nurture: Prenatal roots of instinctive behavior. Mahwah, NJ: Lawrence Erlbaum Associates. Gottlieb, G. (2002). Individual development and evolution. Mahwah, NJ: Lawrence Erlbaum Associates. Gottlieb, G. & Halpern, C. T. (2002). A relational view of causality in normal and abnormal development. Development and Psychopathology, 14, 421–435.  doi: 10.1017/S0954579402003024 Gottlieb, G., Wahlsten, D., & Lickliter, R. (2006). The significance of biology for human development: A developmental psychobiological systems view. In R. M. Lerner & W. Damon (Eds.), Handbook of child psychology (6 ed., Vol. 1, pp. 210–257). Hoboken, NJ: John Wiley & Sons. Griffiths, P. E. & Tabery, J. (2013). Developmental systems theory: What does it explain, and how does it explain it? In R. M. Lerner & J. B. Benson (eds.), Advances in child development and behavior (vol. 44, pp. 65–94). Waltham, MA: Academic Press. Hallgren, B. (1950). Specific dyslexia (congenital word-blindness); a clinical and genetic study. Acta Psychiatrica et Neurologica Scandinavica Supplementum, 65, 1–287. Hazan, V., Messaoud-Galusi, S., Rosen, S., Nouwens, S., & Shakespeare, B. (2009). Speech perception abilities of adults with dyslexia: Is there any evidence for a true deficit? Journal of Speech, Language, and Hearing Research, 52, 1510–1529.  doi: 10.1044/1092-4388(2009/08-0220) International Dyslexia Association (2002). Definition of dyslexia. Retrieved from http://eida. org/definition-of-dyslexia/ Jablonka, E. & Lamb, M. J. (2014). Evolution in four dimensions (rev. ed.). Cambridge, MA: The MIT Press. Jednorog, K., Marchewka, A., Altarelli, I., Monzalvo Lopez, A. K., van Ermingen-Marbach, M., Grande, M., Grabowska, A., Heim, S. & Ramus, F. (2015). How reliable are gray matter disruptions in specific reading disability across multiple countries and Languages? Insights from a Large-scale voxel-based morphometry study. Human Brain Mapping, 36, 1741–1754.  doi: 10.1002/hbm.22734 Jones, M. W., Branigan, H. P., & Kelly, M. L. (2009). Dyslexic and nondyslexic reading fluency: Rapid automatized naming and the importance of continuous lists. Psychonomic Bulletin & Review, 16, 567–572.  doi: 10.3758/PBR.16.3.567 Kere, J. (2014). The molecular genetics and neurobiology of developmental dyslexia as model of a complex phenotype. Biochemical and Biophysical Research Communications, 452, 236–243.  doi: 10.1016/j.bbrc.2014.07.102 Kirby, J. R., Georgiou, G. K., Martinussen, R., & Parrila, R. (2010). Naming speed and reading: From prediction to instruction. Reading Research Quarterly, 45, 341–362. doi: 10.1598/RRQ.45.3.4 Kirby, J. R., Parrila, R. K. & Pfeiffer, S. L. (2003). Naming speed and phonological awareness as predictors of reading development. Journal of Educational Psychology, 95, 453–464. doi: 10.1037/0022-0663.95.3.453 Krafnick , A. J., Flowers, D. L., Luetje, M. M., Napoliello, E. M. & Eden, G. F. (2014) An investigation into the origin of anatomical differences in dyslexia. The Journal of Neuroscience, 34, 901–908.  doi: 10.1523/JNEUROSCI.2092-13.2013 Linkersdörfer, J., Lonnemann, J., Lindberg, S., Hasselhorn, M., & Fiebach, C. J. (2012). Grey matter alterations co-localize with functional abnormalities in developmental dyslexia: An ALE meta-analysis. PLoS ONE, 7, e43122.  doi: 10.1371/journal.pone.0043122



Dyslexia and word reading problems 355

Lobier, M. A., Peyrin, C., Pichat, C., Le Bas, J. F., & Valdois, S. (2014). Visual processing of multiple elements in the dyslexic brain: Evidence for a superior parietal dysfunction. Frontiers in Human Neuroscience, 8, 479.  doi: 10.3389/fnhum.2014.00479 Lobier, M., Zoubrinetzky, R., & Valdois, S. (2012). The visual attention span deficit in dyslexia is visual and not verbal. Cortex, 48, 768–773.  doi: 10.1016/j.cortex.2011.09.003 Lyytinen, H., Ahonen, T., Aro, M., Aro, T., Närhi, V., & Räsänen, P. (1998). Learning disabilities: A view of developmental neuropsychology. In R. Licht, A. Bouma, W. Slot, & W. Koops (Eds.), Child neuropsychology: Reading disability and more (pp. 29–54). Delft, NL: Eburon. Magnuson, J. S., Kukona, A., Braze, D., Johns, C. L., Van Dyke, J. A., Tabor, W., Mencl, W. E., Pugh, K. R., & Shankweiler, D. (2011). Phonological instability in young adult poor readers: Time course measures and computational modeling. In P. McCardle, B. Miller, J. R. Lee, & O. J. L. Tseng (Eds.). Dyslexia across languages (pp. 184–201). Baltimore, MD: Paul H. Brookes. Manis, F. R., Seidenberg, M. S., Doi, L. M., McBride-Chang, C., & Petersen, A. (1996). On the bases of two subtypes of development dyslexia. Cognition, 58, 157–195. doi: 10.1016/0010-0277(95)00679-6 Manis, F. R., Seidenberg, M. S., Stallings, L., Joanisse, M., Bailey, C., Freedman, L., … & Keating, P. (1999). Development of dyslexic subgroups. Annals of Dyslexia, 49, 105–134. doi: 10.1007/s11881-999-0021-1 Maisog, J. M., Einbinder, E. R., Flowers, D. L., Turkeltaub, P. E. & Eden, G. F. (2008). A meta-analysis of functional neuroimaging studies of dyslexia. Annals of the New York Academy of Sciences, 1145, 237–259.  doi: 10.1196/annals.1416.024 McCardle, P., Miller, B., Lee, J. R., & Tzeng, O. J. L. (2011). Dyslexia across languages: Orthography and the brain-gene-behavior link. Baltimore, MD: Paul H. Brookes. Molenaar, P. C. M., Lerner, R. M., & Newell, K. M. (eds.)(2014). Handbook of developmental systems theory and methodology. New York: Guilford. Nation, K., & Snowling, M. (1998). Semantic processing and the development of word-recognition skills: Evidence from children with reading comprehension difficulties. Journal of Memory and Language, 39, 85–101.  doi: 10.1006/jmla.1998.2564 Noordenbos, M. W., Segers, E., Serniclaes, W., & Verhoeven, L. (2013). Neural evidence of the allophonic mode of speech perception in adults with dyslexia. Clinical Neurophysiology, 124, 1151–1162.  doi: 10.1016/j.clinph.2012.12.044 Nopola-Hemmi, J., Taipale, M., Haltia, T., Lehesjoki, A-E., Voutilainen, A. & Kere, J. (2000). Two translocations of chromosome 15q associated with dyslexia. Journal of Medical Genetics, 37, 771–775.  doi: 10.1136/jmg.37.10.771 Norton, E. S., & Wolf, M. (2012). Rapid automatized naming (RAN) and reading fluency: Implications for understanding and treatment of reading disabilities. Annual Review of Psychology, 63, 427–452.  doi: 10.1146/annurev-psych-120710-100431 Nurmi, J.-E. (2012). Students’ characteristics and teacher-child relationships in instruction: A meta-analysis. Educational Research Review, 7, 177–197.  doi: 10.1016/j.edurev.2012.03.001 Parrila, R. K., Kirby, J. R., McQuarrie, L. (2004). Articulation rate, naming speed, verbal shortterm memory, and phonological awareness: Longitudinal predictors of early reading development? Scientific Studies of Reading, 8, 3–26.  doi: 10.1207/s1532799xssr0801_2 Perfetti, C. A. (1992). The representation problem in reading acquisition. In P. B. Gough, L. C. Ehri, & R. Treiman (Eds.), Reading acquisition (pp. 145–174). Hillsdale, NJ: Lawrence Erlbaum. Pennington, B. F. (2006). From single to multiple deficit models of developmental disorders. Cognition, 101, 385–413.  doi: 10.1016/j.cognition.2006.04.008

356 Rauno K. Parrila and Athanassios Protopapas

Pennington, B. F., Santerre-Lemon, L., Rosenberg, J., McDonald, B., Boada, R., Friend, A. … Olson, R. K. (2012). Individual prediction of dyslexia by single versus multiple deficit models. Journal of Abnormal Psychology, 121, 212–224.  doi: 10.1037/a0025823 Plaut, D. C., McClelland, J. L., Seidenberg, M. S., & Patterson, K. (1996). Understanding normal and impaired word reading: Computational principles in quasi-regular domains. Psychological Review, 103, 56–115.  doi: 10.1037/0033-295X.103.1.56 Protopapas, A. (2014). From temporal processing to developmental language disorders: mind the gap. Philosophical Transactions of the Royal Society B: Biological Sciences, 369, 1–11. Protopapas, A., Altani, A., & Georgiou, G. K. (2013). Development of serial processing in reading and rapid naming. Journal of Experimental Child Psychology, 116, 914–929. doi: 10.1016/j.jecp.2013.08.004 Pugh, K. R., Mencl, W. E., Jenner, A. R., Katz, L., Frost, S. J., Lee, J. R., et al. (2000). Functional neuroimaging studies of reading and reading disability (developmental dyslexia). Mental Retardation & Developmental Disabilities Research Reviews, 6, 207–213. doi: 10.1002/1098-2779(2000)6:33.0.CO;2-P Rack, J. P., Snowling, M. J., & Olson, R. K. (1992). The nonword reading deficit in developmental dyslexia: A review. Reading Research Quarterly, 27, 29–53.  doi: 10.2307/747832 Railsback, S. F. & Grimm, V. (2011). Agent-based and individual-based modeling: A practical introduction. Princeton, NJ: Princeton University Press. Ramus, F. (2003). Developmental dyslexia: Specific phonological deficit or general sensorimotor dysfunction? Current Opinion in Neurobiology, 13, 212–218. doi: 10.1016/S0959-4388(03)00035-7 Ramus, F., & Ahissar, M. (2012). Developmental dyslexia: The difficulties of interpreting poor performance, and the importance of normal performance. Cognitive Neuropsychology, 29, 104–122.  doi: 10.1080/02643294.2012.677420 Ramus, F., Rosen, S., Dakin, S. C., Day, B. L., Castellote, J. M., White, S., & Frith, U. (2003). Theories of developmental dyslexia: Insights from a multiple case study of dyslexic adults. Brain, 126, 841–865.  doi: 10.1093/brain/awg076 Ramus, F., & Szenkovits, G. (2008). What phonological deficit? The Quarterly Journal of Experimental Psychology, 61, 129–141.  doi: 10.1080/17470210701508822 Ramus, F., & Szenkovits, G. (2009). Understanding the nature of the phonological deficit. In K. Pugh & P. McCardle (Eds.), How children learn to read (pp. 153–169). New York, NY: Psychology Press. Richlan, F., Kronbichler, M., & Wimmer, H. (2009). Functional abnormalities in the dyslexic brain: A quantitative meta-analysis of neuroimaging studies. Human Brain Mapping, 30, 3299–3308.  doi: 10.1002/hbm.20752 Richlan, F., Kronbichler, M., & Wimmer, H. (2011). Meta-analyzing brain dysfunctions in dyslexic children and adults. Neuroimage, 56, 1735–1742.  doi: 10.1016/j.neuroimage.2011.02.040 Richlan, F., Kronbichler, M., & Wimmer, H. (2013). Structural abnormalities in the dyslexic brain: A meta-analysis of voxel-based morphometry studies. Human Brain Mapping, 34, 3055–3065.  doi: 10.1002/hbm.22127 Ricketts, J., Nation, K., & Bishop, D. V. M. (2007). Vocabulary is important for some, but not all reading skills. Scientific Studies of Reading, 11, 235–257.  doi: 10.1080/10888430701344306 Rosen, S., & Manganari, E. (2001). Is there a relationship between speech and nonspeech auditory processing in children with dyslexia? Journal of Speech, Language, and Hearing Research, 44, 720–736.  doi: 10.1044/1092-4388(2001/057)



Dyslexia and word reading problems 357

Ruffino, M., Gori, S., Boccardi, D., Molteni, M., & Facoetti, A. (2014). Spatial and temporal attention in developmental dyslexia. Frontiers in Human Neuroscience, 8, 331. doi: 10.3389/fnhum.2014.00331 Serniclaes, W., Van Heghe, S., Mousty, P., Carré, R., & Sprenger-Charolles, L. (2004). Allophonic mode of speech perception in dyslexia. Journal of Experimental Child Psychology, 87, 336– 361.  doi: 10.1016/j.jecp.2004.02.001 Share, D. L. (2008). On the anglocentricities of current reading research and practice: The perils of overreliance on an “outlier” orthography. Psychological Bulletin, 134, 584–615.  doi: 10.1037/0033-2909.134.4.584 Shaywitz, S. E. (1996). Dyslexia. Scientific American, 275, 98–104. doi: 10.1038/scientificamerican1196-98 Simos, P. G., Fletcher, J. M., Bergman, E., Breier, J. I., Foorman, B. R., Castillo, E. M., Davis, R. N., Fitzgerald, M., & Papanicolaou, A. C. (2002). Dyslexia-specific brain activation profile becomes normal following successful remedial training. Neurology, 58, 1203–1213. doi: 10.1212/WNL.58.8.1203 Snowling, M. J. (2000). Dyslexia (2nd ed.). Oxford, UK: Blackwell. Snowling, M. J. (2013). Early identification and interventions for dyslexia: A contemporary view. Journal of Research in Special Educational Needs, 13, 7–14.  doi: 10.1111/j.1471-3802.2012.01262.x Snowling, M. J. & Melby-Lervåg, M. (2016). Oral language deficits in familial dyslexia: A meta-analysis and review. Psychological Bulletin, 142, 498–545. Stanovich, K. E., Siegel, L. S., & Gottardo, A. (1997). Converging evidence for phonological and surface subtypes of reading disability. Journal of Educational Psychology, 89, 114–127. doi: 10.1037/0022-0663.89.1.114 Steacy, L. M., Kirby, J. R., Parrila, R., & Compton, D. L. (2014). Classification of double deficit groups across time: An analysis of group stability from kindergarten to second grade. Scientific Studies of Reading, 18, 255–273.  doi: 10.1080/10888438.2013.873936 Tabery, J. (2014). Beyond versus: The struggle to understand the interaction between nature and nurture. Cambridge, MA: The MIT Press.  doi: 10.7551/mitpress/9780262027373.001.0001 Tallal, P. (1980). Auditory temporal perception, phonics, and reading disabilities in children. Brain and Language, 9, 182–198.  doi: 10.1016/0093-934X(80)90139-X Thelen, E. & Smith, L. B. (1994). A dynamic systems approach to the development of cognition and action. Cambridge, MA: The MIT Press. Torppa, M., Eklund, K., van Bergen, E., & Lyytinen, H. (2015). Late-emerging and resolving dyslexia: A follow-up study from age 3 to 14. Journal of Abnormal Child Psychology, 43, 1389–1401.  doi: 10.1007/s10802-015-0003-1 Torppa, M., Parrila, R., Niemi, P., Poikkeus, A.-M., Lerkkanen, M.-K., & Nurmi, J.-E. (2013). The double deficit hypothesis in the transparent Finnish orthography: A longitudinal study from kindergarten to grade 2. Reading and Writing: An Interdisciplinary Journal, 26, 1353–1380.  doi: 10.1007/s11145-012-9423-2 Tunmer, W. & Greaney, K. (2010). Defining dyslexia. Journal of Learning Disabilities, 43, 229–243.  doi: 10.1177/0022219409345009 van Bergen, E., Bishop, D. V. M., van Zuijen, T. L., & de Jong, P. F. (2015). How does parental reading influence children’s reading. A study of cognitive mediation. Scientific Studies of Reading, 19, 325–339.  doi: 10.1080/10888438.2015.1050103 van Bergen, E., van der Leij, A. & de Jong, P. F. (2014). The intergenerational multiple deficit model and the case of dyslexia. Frontiers in Human Neuroscience, 8, 346. doi: 10.3389/fnhum.2014.00346

358 Rauno K. Parrila and Athanassios Protopapas

van den Boer, M., Georgiou, G. K., & de Jong, P. F. (2016). Naming of short words is (almost) the same as naming of alphanumeric symbols: Evidence from two orthographies. Journal of Experimental Child Psychology, 144, 152–165.  doi: 10.1016/j.jecp.2015.11.016 Van den Broeck, W., & Geudens, A. (2012). Old and new ways to study characteristics of reading disability: The case of the nonword-reading deficit. Cognitive Psychology, 65, 414– 456.  doi: 10.1016/j.cogpsych.2012.06.003 Vandermosten, M., Boets, B., Wouters, J. & Ghesquière, P. (2012). A qualitative and quantitative review of diffusion tensor imaging studies in reading and dyslexia. Neuroscience and Biobehavioral Reviews, 36, 1532–1552  doi: 10.1016/j.neubiorev.2012.04.002 Velicer, W. F., Babbin, S. F., & Palumbo, R. (2014). Idiographic applications: Issues of ergodicity and generalizability. In P. C. M. Molenaar, R. M. Lerner & K. M. Newell (eds.), Handbook of developmental systems theory and methodology (pp. 425–441). New York: Guilford. Vellutino, F. R., Scanlon, D. M., Sipay, E. R., Small, S. G., Pratt, A., Chen, R. S., & Denckla, M. S. (1996). Cognitive profiles of difficult-to-remediate and readily remediated poor readers: Early intervention as a vehicle for distinguishing between cognitive and experiential deficits as basic causes of specific reading disability. Journal of Educational Psychology, 88, 601–638.  doi: 10.1037/0022-0663.88.4.601 Vellutino, F. R., Fletcher, J. M., Snowling, M. J., & Scanlon, D. M. (2004). Specific reading disability (dyslexia): What have we learned in the past four decades? Journal of Child Psychology and Psychiatry, 45, 2–40.  doi: 10.1046/j.0021-9630.2003.00305.x Vidyasagar, T. R., & Pammer, K. (2010). Dyslexia: A deficit in visuo-spatial attention, not in phonological processing. Trends in Cognitive Sciences, 14, 57–63.  doi: 10.1016/j.tics.2009.12.003 Waters, C. K. (2007). Causes that make a difference. Journal of Philosophy, 104, 551–579. doi: 10.5840/jphil2007104111 Wilensky, U. & Rand, W. (2015). An introduction to agent-based modeling: Modeling natural, social and engineered complex systems with NetLogo. Cambridge, MA: The MIT Press. Wolf, M., & Bowers, P. G. (1999). The double-deficit hypothesis for the developmental dyslexias. Journal of Educational Psychology, 91, 415–438.  doi: 10.1037/0022-0663.91.3.415 Wolf, M., Bowers, P. G., & Biddle, K. (2000). Naming-speed processes, timing, and reading: A conceptual review. Journal of Learning Disabilities, 33, 387–407.  doi: 10.1177/002221940003300409 Woollams, A. M. (2014). Connectionist neuropsychology: Uncovering ultimate causes of acquired dyslexia. Philosophical Transactions of the Royal Society B: Biological Sciences, 369, 20120398. Ziegler, J. C., & Goswami, U. (2005). Reading acquisition, developmental dyslexia, and skilled reading across languages: A psycholinguistic grain size theory. Psychological Bulletin, 131, 3–29.  doi: 10.1037/0033-2909.131.1.3 Ziegler, J. C., Pech‐Georgel, C., Dufau, S., & Grainger, J. (2010). Rapid processing of letters, digits and symbols: what purely visual‐attentional deficit in developmental dyslexia? Developmental Science, 13, F8–F14.  doi: 10.1111/j.1467-7687.2010.00983.x Zoccolotti, P., De Luca, M., Lami, L., Pizzoli, C., Pontillo, M., & Spinelli, D. (2013). Multiple stimulus presentation yields larger deficits in children with developmental dyslexia: A study with reading and RAN-type tasks. Child Neuropsychology, 19, 639–647.  doi: 10.1080/09297049.2012.718325 Zoccolotti, P., De Luca, M., & Spinelli, D. (2015). Discrete versus multiple word displays: A re-analysis of studies comparing dyslexic and typically developing children. Frontiers in Psychology, 6, 01530.  doi: 10.3389/fpsyg.2015.01530 Zoubrinetzky, R., Bielle, F., & Valdois, S. (2014). New insights on developmental dyslexia subtypes: Heterogeneity of mixed reading profiles. PloS one, 9, e99337. doi: 10.1371/journal.pone.0099337

Children with specific text comprehension problems Jane Oakhill and Kate Cain

University of Sussex / Lancaster University

This chapter is about children with poor reading comprehension. We focus on children whose text comprehension difficulties are unexpectedly poor in relation to their age-appropriate word reading skills and whose text comprehension difficulties are apparent in listening, as well as reading, comprehension tasks (Cain & Oakhill, 2007). We first provide an overview of different models of skilled reading comprehension to inform our examination of children’s text comprehension problems. Our subsequent examination of probable causes of poor comprehension in this group focuses on the theoretical and empirical support for two main classes of model: those that link reading comprehension difficulties to inadequate or incomplete word level representations (e.g., Perfetti’s Lexical Quality Hypothesis, Perfetti, 2007) and those that relate these difficulties to insufficient use of text-level processes (e.g., Cain & Oakhill, 2007). We present recent research that explores the relations between different aspects of vocabulary knowledge (a word-level process) and different types of inference skill (a text-level process). Our approach supports a rapprochement between the two different theoretical perspectives. In the final section, we discuss possible avenues of future research that can lead to both theoretical and practical developments in this area.

Background Reading comprehension: what’s involved? Reading comprehension is a complex construct informed by a range of skills and knowledge bases. We first consider what is generally agreed as constituting successful text comprehension. We then use the criteria for success as a framework within which the relevance and importance of different component skills in reading comprehension can be understood. Many skills and aspects of language knowledge are involved in reading comprehension: words are decoded and their meanings retrieved, sentence meanings are computed, the information presented in successive sentences is integrated, and inferences are made to fill in details and to connect up the text into a coherent whole. doi 10.1075/swll.15.20oak © 2017 John Benjamins Publishing Company

360 Jane Oakhill and Kate Cain

Thus, understanding a text can be viewed as a constructive process that results in a coherent and integrated representation of the state of affairs described in the text. This text representation will be very similar whether it is formed via reading or listening comprehension though, of course, reading comprehension requires the additional skill of being able to recognize the words. The construction of a coherent and integrated representation requires the orchestration of all the different processes that, in a skilled comprehender, typically go on in parallel and without conscious strategic control. These include retrieval of word meanings, their integration with the meanings of the previous sentences and paragraphs (see Stafura and Perfetti, this volume), and the identification of key themes and ideas in the text (see van den Broek and Kendeou, this volume). Information that the author has left implicit will need to be filled out with inferences, which might happen relatively automatically, or which might sometimes demand more explicit reasoning. In addition to the integration of information across sentences (resolution of anaphoric links, use of causal and temporal connectives, etc.), the information in the text must be integrated with prior knowledge, and this integration will support inference making. Our own research has demonstrated that, even once word decoding skills and vocabulary are taken into account, higher-level skills, such as integration and inference making, contribute to children’s reading comprehension, both within and across time (Oakhill & Cain, 2012; Oakhill, Cain, & Bryant, 2003).

Models of text comprehension In the areas of word reading and syntactic processing there are distinct, opposing and empirically falsifiable theories of how words are read or sentences are understood. In contrast, distinct theories of reading comprehension do not exist, because the current “theories” are better described as “frameworks” in that they are typically not competitive and falsifiable (as noted earlier in Stafura and Perfetti, this volume). This situation is attributable to the fact that text comprehension is far more complex than other aspects of reading, such as word recognition and sentence parsing, because it requires the successful orchestration of a number of different cognitive skills and abilities. In addition, comprehension is not ‘black or white’ in the sense that it either happens or it does not. For example, although we can state that a person can or cannot accurately read certain words in a text, it does not make sense to say that a certain person makes inferences from text or does not; the issues with text comprehension are much more subtle. Rather than asking whether or not a person makes inferences, we can ask under what circumstances (text variables, person variables) that a reader makes certain types of inference, and whether they make sufficient inferences to ensure adequate comprehension of the text. We can also ask when those inferences are made: immediately as the text is read, or later on when prompted by a question or a post-reading task.



Children with specific text comprehension problems 361

There are many frameworks or models of skilled text comprehension (see McNamara & Magliano, 2009, for a review) but none that focus on comprehension development (see Cain & Barnes, this volume, for further discussion on this point). McNamara and Magliano (2009) reviewed the seven models of text comprehension that have received most attention, and came to the conclusion that there is not one overall “winner”; rather, the different theoretical frameworks fit particular niches. These models of comprehension, despite their differences, have a number of commonalities. All emphasise the need for the reader to establish text coherence through processes such as integration and inference, and the requirement that the end product of comprehension should be an integrated and coherent representation of the text as a whole. This mental representation is variously termed a Mental Model (Johnson-Laird, 1983) or a Situation Model (Kintsch, 1998; Kintsch & van Dijk, 1978) depending on the details of the theoretical perspective (but those details will not concern us here). The lack of a single theory (in the strict sense of being falsifiable, and competitive with other theories in the area) that can adequately account for text comprehension leads us to propose that it is more productive to think in terms of the different component processes in reading comprehension, their relative contributions and inter-independencies, and which of these might be deficient in children who experience comprehension problems. In this chapter we focus on two critical component skills – vocabulary knowledge (a word-level process) and inference skill (a text-level process), their influence on each other, as well as their influence on reading comprehension. But first, we consider what is meant by poor text comprehension.

Who has text comprehension difficulties? There are different types of poor reader and poor comprehender, and the reasons for their comprehension difficulties may differ. For instance, some children may have difficulties understanding text because they are unable to read the words. Thus, children who do poorly on typical comprehension tests may have word reading problems, reading comprehension problems, or both. Our focus is on a group of children who experience specific text comprehension difficulties. This term refers to the specificity of the problem, in that it is specific to comprehension and does not stem from issues with other aspects of reading, such as inability to read the words in the text to an adequate level. Thus, children with dyslexia have specific difficulties with reading words, and poor comprehenders have specific difficulties with comprehension of text that cannot be attributed to poor word reading. These children’s difficulties with reading comprehension are typically related to more general issues with language comprehension; that is, they have trouble understanding text that is read aloud to them (presented in the auditory modality) as well as text that they read themselves (presented visually). Thus, their problems are not usually specific to reading (see e.g., Cain & Oakhill, 2006, 2011; Spencer, Quinn, & Wagner, 2014), although it is in the context of reading that they typically become apparent and most theories have, explicitly or implicitly, addressed reading and not listening comprehension development and difficulties.

362 Jane Oakhill and Kate Cain

The research base supporting word-level and text-level sources of comprehension failure In typical studies comparing good and poor comprehenders (e.g., Cain & Oakhill, 1999; Cain, Oakhill, & Bryant, 2000; Catts, Adlof, & Weismer, 2006; Megherbi, Seigneuric, & Ehrlich, 2006; Nation & Snowling, 2000) the word level reading of children is controlled for and, therefore, discounted. However, accuracy and fluency of word (and non-word) reading is not the only word level aspect that might be of concern. In this chapter we consider broader word-level variables that might have an impact on text comprehension. Such variables, unlike word decoding, are important in both reading and listening to text. Perfetti’s views on the word-level factors that limit reading comprehension have been very influential in this respect, in particular the Lexical Quality Hypothesis (e.g., Perfetti, 2007; Perfetti & Hart, 2001; Perfetti, Wlotko, & Hart, 2005), which has been developed in the Reading Systems Framework (see Perfetti & Stafura, 2014, also this volume). Perfetti’s main hypothesis, put simply, is that reading comprehension depends to a considerable extent on knowledge of words, at least for successful wordto-text integration. We discuss below how vocabulary specifically supports reading comprehension and examine the evidence that knowledge of words is causally implicated in reading comprehension failure. We then turn to the evidence that supports a text-level integration and inference account, before proposing an integrated account that provides a rapprochement of these two, often separate, lines of research.

Knowledge of word meanings: A critical ‘word-level’ skill Vocabulary knowledge is fundamental to comprehension skill. At a superficial level, vocabulary knowledge is crucial for comprehension because, even if words can be decoded, the text cannot be understood unless the meanings of (most of) the words are understood. Vocabulary knowledge is one of the best concurrent predictors of reading comprehension ability (Carroll, 1993; Thorndike, 1973). For instance, Thorndike found correlations of between .66 and .75 between vocabulary knowledge and reading comprehension. There are various ways in which vocabulary and reading comprehension might be related, not all of them causal (see Quinn, Wagner, Petscher, & Lopez, 2015, for a discussion). However, several studies with children have produced evidence for a causal relation in the direction from vocabulary to comprehension. De Jong and van der Leij (2002) showed that vocabulary in grade 3 predicted significant variance in reading comprehension in grade 5, after controlling for grade 3 comprehension (the autoregressor). Similarly, Verhoeven and van Leeuwe (2008) found a strong influence of early vocabulary (grade 1) on later reading comprehension ability (grade 6), over and above the autoregressive effect of reading comprehension; in contrast, early reading comprehension was only a weak predictor of later vocabulary. Quinn et al.’s (2015) data also support the conclusion that previous levels of vocabulary knowledge



Children with specific text comprehension problems 363

are important indicators of growth in reading comprehension (between grades 1 and 4), whereas there was no support in their data for the reverse causal relation. These studies all looked at how simple measures of vocabulary size are related to reading comprehension, with the assumption that word meanings are the building blocks of passage meaning. A more refined view of the reason for the relation between vocabulary and reading comprehension comes from Perfetti’s Lexical Quality hypothesis. The crucial idea is that a low quality lexical code, which is retrieved with effort, can interfere with comprehension processes that are dependent on a high quality code. The concept of Lexical Quality includes a range of knowledge about word forms (phonology, orthography, grammar) and also meaning. The focus here is on that last aspect: the quality of a word’s meaning representation. From this viewpoint, the availability of associative links between words – the consequence of a rich (deep) vocabulary – might aid comprehension by supporting a skill such as inference making. In particular, rich, precise, and well-connected semantic representations will permit the rapid activation not only of word meanings, but also of related concepts, and this activation of semantic networks will provide the underpinning for integration of information into the mental model (see Perfetti, 2007; Perfetti, Yang, & Schmalhofer, 2008). Most studies of the Lexical Quality Hypothesis have used adult participants and there are no studies of children that speak directly to the relation between quality of vocabulary knowledge and inference making. However, research on semantic priming suggests that this is an important area for future research. Poor comprehenders (who have established inference making difficulties, e.g., Cain & Oakhill, 1999) are less sensitive to the abstract semantic relations than good comprehenders (Nation & Snowling, 2000), suggesting that they will be less likely to automatically access semantic information about a word.

Inference making: A critical ‘text-level’ skill The word-level processes detailed above can be contrasted with so-called “text-level” processes, which have been the focus of much of our own work in children (for a review, see Cain & Oakhill, 2007). On this view, it is the child’s ability to use text-level processes, such as inference, which crucially limits reading comprehension skill during development, rather than the word level skills, such as word decoding and semantic access. It goes without saying that effective reading comprehension cannot occur in the absence of good word decoding and access to word meanings. However, our thesis has been that these word level processes, though necessary, are not sufficient to support comprehension, and that other processes are essential too. Indeed, it is quite possible to find children who have good vocabulary skills, in terms of vocabulary size (as measured by a standardized receptive vocabulary assessment, such as the Peabody Picture Vocabulary Test (PPVT; Dunn & Dunn, 2007), or in the UK, the British Picture Vocabulary Scale (BPVS; Dunn, Whetton & Burley, 1997)) but, nevertheless, have problems with text comprehension (Cain, Oakhill, & Lemmon, 2004).

364 Jane Oakhill and Kate Cain

Skilled readers make relevant inferences to establish local cohesion between sentences and global coherence of the text overall with little apparent effort, whereas children who have reading comprehension problems have difficulties making inferences, even relatively simple ones that connect up sentences (e.g., Oakhill, 1982). There is evidence that poor comprehenders have problems with both local cohesion and global coherence inferences. In Cain and Oakhill’s (1999) study, the children read a number of short stories each followed by several questions. Examples of these materials are provided in Table 1. Some questions tapped their literal comprehension, and there were two types of inference question. The poor comprehenders performed poorly on both local cohesion and global coherence inference questions, but did not have any problems answering the literal questions. Table 1.  Sample story and questions from Cain & Oakhill (1999) Debbie was going out for the afternoon with her friend Michael. By the time they got there they were very thirsty. Michael got some drink out of his duffel bag and they shared that. The orange juice was very refreshing. Debbie put on her swimming costume but the water was too cold to paddle in, so they made sandcastles instead. They played all afternoon and didn’t notice how late it was. Then Debbie spotted the clock on the pier. If she was late for dinner her parents would be angry. They quickly packed up their things. Debbie changed and wrapped her swimming costume in her towel. She put the bundle in her rucksack. Then they set off for home, pedaling as fast as they could. Debbie was very tired when she got home, but she was just in time for dinner. Questions Literal information: 1. Who did Debbie spend the afternoon with? 2. Where was the clock? Local Cohesion inference: 3. Where did Michael get the orange juice from? 4. Where did Debbie put her towel when she packed up her things? Global Coherence inference: 5. Where did Debbie and Michael spend the afternoon? 6. How did Debbie and Michael travel home?

The finding that good and poor comprehenders differ in their inference making skills has been replicated many times and cannot be attributed simply to poor comprehenders’ worse memory for the text (Oakhill, 1984) or to lack of background knowledge. Even when good and poor comprehenders are matched for background knowledge, the poor comprehenders still show specific difficulties with inference questions (Cain, Oakhill, Barnes, & Bryant, 2001), which can probably be attributed to failures to activate and use relevant knowledge. Further evidence that it is not simply a matter of having, but knowing how to use, background knowledge comes from a study by Elbro and Buch-Iversen (2013). The study showed that training with



Children with specific text comprehension problems 365

graphic organisers, which focused on the contribution of background knowledge for text comprehension, improved children’s ability to make global coherence inferences. Training studies and research using comprehension-age match and longitudinal designs provide evidence that difficulties with inference are causally implicated in children’s reading comprehension, and particularly in reading comprehension breakdown. Poor comprehenders make fewer inferences than younger children matched for absolute level of comprehension skill on a standardized reading test, which suggests a causal relation between inferences and comprehension skill, because this result rules out a causal link in the opposite direction (Cain & Oakhill, 1999; a detailed discussion the interpretation of the comprehension-age match group design can be found in Cain, Oakhill, & Bryant, 2000). In addition, training poor comprehenders to make inferences results in gains on standardised assessments of reading comprehension, in addition to better inference making skills (Yuill & Oakhill, 1988). Finally, inference skills predict reading comprehension longitudinally even when the autoregressive effect of reading comprehension, verbal IQ, vocabulary size, and other potentially contributory variables have been controlled (Oakhill & Cain, 2012). Thus, there is a wealth of evidence that poor inference making is causally implicated in reading comprehension development and difficulties.

Bringing things together: A more coherent model Vocabulary and inference making: How might they be related? As discussed, both vocabulary knowledge and inference making are shown to support reading comprehension and to be implicated in reading comprehension difficulties. One hypothesis therefore is that poor comprehenders’ difficulties with inference making and reading comprehension in general stem from weak vocabulary skills. However, poor comprehenders do not necessarily have deficient vocabularies (at least in terms of vocabulary breadth), as noted above (Cain, Oakhill, & Lemmon, 2004). Thus, even though vocabulary knowledge and reading comprehension are typically highly correlated, and studies have demonstrated a causal link between vocabulary skills and later comprehension (Quinn et al., 2015, Verhoeven & Leeuwe, 2008), problems with reading comprehension can arise even in the presence of a broad vocabulary and, in general, the difficulties that poor comprehenders experience with reading comprehension cannot be accounted for by their level of vocabulary alone (Oakhill & Cain, 2012). Such findings lead us to the conclusion that other, higher-order, text processes must be important, although it should be noted that the above-mentioned studies have primarily controlled only for breadth of vocabulary (see below). We hypothesise that, in children, a rich (deep) vocabulary knowledge will aid inference making in comprehension because many of the inferences in texts – global coherence inferences in particular – are dependent on semantic links between words in the text (we provide specific examples below). This activation of semantic links can

366 Jane Oakhill and Kate Cain

then provide the basis for many of the inferences that are crucial for the construction of a coherent representation of a text. In addition, we hypothesise that inference making skills help to support the further development of vocabulary and the rich semantic networks that aid successful inference making. We review the evidence for both types of links in the sections below.

What is vocabulary knowledge: A deeper look? The most commonly used tests of vocabulary in the studies outlined above are measures of vocabulary at shallow levels (e.g., assessments of ability to select one from a choice of pictures to go with a word). Such measures are typically described as assessments of breadth of vocabulary, and the knowledge about the words assessed can be fairly superficial. However, recent research suggests that measures of vocabulary knowledge at greater depths (typically referred to depth) might be more important for reading comprehension (Ouellette, 2006; Tannenbaum, Torgesen, & Wagner, 2006). Roughly, breadth corresponds to how many words a person knows (and is what is typically measured in vocabulary assessments, such as the PPVT or BPVS), whereas depth corresponds to what is known about those words (e.g., knowing multiple, or more subtle, meanings; being able to provide synonyms or definitions). Thus, a reader might be able to match up a word with a picture (breadth), but might have rather little idea about the broader meanings and uses of that same word (depth). Ouellette (2006) and Tannenbaum, Torgesen and Wagner (2006) have shown that assessments of depth and breadth of vocabulary are distinguishable and that the two aspects make separate contributions to comprehension skill. Depth of vocabulary knowledge is likely to be more important than breadth in supporting inference making because rich and well-connected semantic representations of words will permit the rapid activation not only of a word’s meaning but also the meanings of related concepts.

Evidence that vocabulary supports inference making Some preliminary work with children provides evidence that depth of vocabulary knowledge is strongly related to making global coherence inferences from text. In a recent study, we showed that depth, but not breadth, of vocabulary knowledge was an important predictor of global coherence inferences, and that this relation held even when word reading skill and literal memory for the text had been taken into account (Cain & Oakhill, 2014). A recent study of children aged 6 to 10 years also indicates that vocabulary is a more important predictor of global coherence inferences than of inferences required to link adjacent sentences in text (Currie & Cain, 2015). Specifically, Currie and Cain (2015) found that vocabulary (a composite of breadth and depth measures) was a unique predictor of both local and global coherence inference making in 6- and 8-year-olds and a unique predictor of global, but not local, coherence inference making for 10-year-olds, with comparable or stronger prediction of global coherence inferences in each age group.



Children with specific text comprehension problems 367

Turning to children with reading comprehension difficulties, we have evidence that good and poor comprehenders differ in their ability to automatically derive themes from word lists. Weekes, Hamilton, Oakhill and Holliday (2008) used the DRM (Deese, 1959; Roediger & McDermott, 1995) false memory paradigm in which the children were required to listen to, and to try to memorise, a short list of words. For example, one word list was: rest, bed, snooze, dream, tired, blanket. The children then completed a recognition test in which they were asked to differentiate between words that had/had not occurred in the lists they had been read. The good comprehenders were more likely than poor comprehenders to falsely claim that sleep had been in the original list in the example given above, even though they did not have poorer word memory more generally. This result can be taken as an indication that the good comprehenders are more likely to automatically derive “themes” from the word lists (even though this is not a requirement of the task) and, thus, falsely remember words that capture the theme of the list. This ability to spontaneously derive themes is very likely to have the same source as having a rich vocabulary, which enables the activation not only of core meanings, but also of categories and semantic associates of words. This propensity to derive themes from word lists might well carry over to inference making in text comprehension because very often the main theme or the setting of a text can be derived from a number of specific words in a text. For example, if you were to read a text that contained the words: trolley, shelves, tins, packets, aisle, scan, bags, pack, till, you might reasonably infer that the text is situated in a shop or supermarket. No one of those words in isolation will support that inference about the setting, but taken together they connect up to provide a coherent overall indication of the setting of the text. The results we have reported suggest that the child’s depth of vocabulary knowledge, which will enable them to activate a rich semantic network and make associations between thematically related words, might underpin ability to make global coherence inferences from a text, in particular. The way in which this might work can be illustrated by considering the following snippet of text: It was a warm Summer’s night in the park. Ann and David walked around the different stalls trying to decide what to buy. Ann bought some candy floss and David bought a toffee apple. There were bright flashing lights everywhere. Ann thought it was really exciting. They had a ride on the Big Wheel and then they went on the Ghost Train. After that they had spent all their money, so they decided to walk home.

In order to answer the question, “What was happening in the park?”, the reader needs to appreciate that several words taken together (stalls, candy floss, toffee apple, flashing light, Big Wheel and Ghost Train) indicate that there is a funfair in the park. If the reader can make associations between the meanings of different words in the text quickly and easily, he or she will very easily be able to infer what is going on in the story, and the question will be trivially easy.

368 Jane Oakhill and Kate Cain

There is also evidence that it is not enough simply to have a rich vocabulary, but that speed and automaticity of access to vocabulary knowledge is also important. The idea that the automaticity with which meanings can be accessed is important in reading comprehension is not new: Early models of reading (e.g., Laberge & Samuels, 1974) emphasized the importance of fluency and automaticity of access to word meanings. In correlational studies, good and poor comprehenders have been shown to differ on measures of semantic fluency, such as the ability to rapidly produce a number of instances of a category (Nation & Snowling, 1998). Recently, Oakhill, Cain, McCarthy and Field (2012) found a strong and specific link between speed of semantic access in vocabulary demanding tasks (synonym and hypernym 1 judgment tasks) and reading comprehension (measured by a standardized task). The link was specific because it was not entirely mediated by word reading skill, or by knowledge about words (assessed by a synonym and hypernym production task), and neither was it related to a simple association between comprehension skill and generally faster response times in a control task that required non-semantic identity judgments (i.e., the ability to say whether two words were orthographically identical or not). These findings suggest that it is not sufficient to know the meanings of words to understand a text, but that knowledge of the interrelations between words and rapid access to the semantic representations of words are both also important for comprehension. Comprehension happens in real time, and if appropriate meanings and associations of words are not accessed very rapidly, the reader will often have moved on in the text and the opportunity for semantic information to support inference and integration of the text will have been missed. The fact that rapid and automatic activation of word meanings is important in children’s inference, integration and broader comprehension skills does not, of course, mean that all children’s inference making is rapid and automatic. Indeed, other work of our own (Cain & Oakhill, 1999) has shown that poor comprehenders are able to make appropriate inferences in response to questions strategically if they are directed back to the text and, in particular, if the parts of the text that support the inference are pointed out to them. Such strategic inferences might also be a product of children’s effective comprehension monitoring, which alerts them to the fact that an inference is required.

Evidence that inference making supports vocabulary development The sections above explored the ways in which vocabulary skills might support inference making. In this section, we review the evidence for a link in the opposite direction and consider the role that inference skill, and extensive and effective reading more generally, can play in supporting vocabulary development. It is generally agreed that written text is an important source of new vocabulary (and the refinement of existing vocabulary) once children become relatively fluent 1. Hypernyms are terms for superordinate categories. For example, rain/weather, dog/mammal, chair/furniture.



Children with specific text comprehension problems 369

readers (Cunningham & Stanovich, 1998; Nagy & Scott, 2000). Reading provides opportunities to acquire, refine, and consolidate vocabulary knowledge via inference from context (we continue to refine our vocabulary in this way throughout our lives, providing we read sufficiently challenging texts). Printed text affords more learning opportunities than does spoken language, mostly because writers tend to use a different register to speakers, and the written register is likely to include more obscure vocabulary items. Again, these studies support the view that substantial reading comprehension practice and exposure to print is crucial in the development of reading-related skills and acquisition of vocabulary. There is experimental evidence to show that poor comprehenders do have more difficulty inferring the meanings of unknown vocabulary items from context, particularly when the memory demands are high. Cain, Oakhill and Lemmon (2004) asked good and poor comprehenders (9- to 10-year-olds) to try to work out the meanings of unknown words (in fact, nonwords) in short texts. The poor comprehenders were worse at coming up with a reasonable meaning for the unknown word, particularly when the clues were more distant from the target word. Thus, it seems that poorer compehenders do have the potential to make such inferences about vocabulary (at least when explicitly required to infer meanings), but if they have to integrate information across several sentences in the text in order to do so they find the task more difficult than do good comprehenders. In another study of vocabulary learning from text we capitalized on some of the data from a longitudinal study (Oakhill & Cain, 2012). The results of those analyses (reported in Cain & Oakhill, 2011) showed that even if good and poor comprehenders have rather similar levels of vocabulary skills at age 8 years, poor reading comprehension at 8 years results in poorer vocabulary 2–3 years later. The pattern of results was identical whether oral (BPVS: Dunn, Dunn, Whetton, & Pintillie, 1982) or written (MacGinitie, MacGinitie, Maria, & Dreyer, 2000) vocabulary was measured. Furthermore, to investigate the relations between reading experience and growth in vocabulary knowledge across time, a series of fixed-order hierarchical multiple regressions was conducted. These analyses were conducted on data from the entire data set, and not only the data from the good and poor comprehenders reported in the previous analyses above. The aim of these analyses was to explore the extent to which reading experience and earlier reading comprehension (both measures taken at age 8) accounted for individual differences in vocabulary growth, once cognitive ability (non-verbal IQ) had been taken into account. Separate analyses were conducted to assess the predictors of written vocabulary skill at a later age (either 11, 14 or 16 years) as the outcome variable. In each analysis, the predictor variables were the measures taken at the first time point, when the children were aged 8. Cognitive ability was entered at the first step, followed by vocabulary. At the third and final step, either the score obtained on a reading questionnaire (which assessed amount of independent leisure reading as a measure of “reading experience”) or (in a separate analysis) the score obtained on the reading

370 Jane Oakhill and Kate Cain

comprehension assessment was entered. The results (see Figure 1) show that reading experience explained later vocabulary competence over and above general cognitive ability and an earlier measure of vocabulary. In addition, reading comprehension also explained later vocabulary even after the same controls. 0.4

Read Habits at 8 Read Comp at 8

0.35

Vocab at 8

ΔR2

0.3

NV IQ at 8

0.25 0.2 0.15 0.1 0.05 0

11 years

11 years

14 years

14 years

16 years

16 years

Prediction of written vocabulary knowledge

Figure 1.  Contributions to vocabulary learning over time: Outcome – Gates-MacGinitie

In summary, poor comprehenders are not good at deriving plausible meanings of new vocabulary items from context. This difficulty is probably related to the poor comprehenders’ inferior inference skills, and seems to be exacerbated when the task requires integration of information across several sentences in the text (and is, therefore, more memory demanding). There is also evidence that comprehension level influences vocabulary development across several years. In particular, poor reading comprehension at age 8 was predictive of poorer vocabulary 2–3 years later and not only reading comprehension ability but also amount of reading experience predicted the children’s later vocabulary competence, at least as assessed by “breadth” measures.

Concluding comments At the outset of this chapter, we considered different models of reading comprehension and concluded that, although they differ in perspective and details, all models are in agreement that inference and integration skills provide the crucial underpinning for the mental model of a text. There is substantial and reliable evidence that (among other problems) poor comprehenders have difficulties in making inferences from text. There is also substantial evidence (in particular from adults) that lexical-level skills



Children with specific text comprehension problems 371

are important for text representation. In this chapter, we considered the way in which inference and word-level skills might interact in reading comprehension. We conclude that a fully comprehensive model of text comprehension needs to consider processes at all levels and how they interact both during development and in skilled reading. Such a model is beyond the scope of the present chapter, but we have chosen some examples, with evidence, of some of these interactions. First, evidence is emerging to show that semantic representations, and in particular depth of vocabulary knowledge, support inference skills in children, and reading comprehension more generally. We found that depth of vocabulary knowledge was an important predictor of one type of inference in particular – global coherence inferences – even when word reading skill and literal memory for the text had been taken into account. There are many studies in the literature that emphasize the importance of vocabulary for reading comprehension, but our studies provide evidence for a more specific link: that is, that a particular aspect of vocabulary (depth) is related to a specific component skill of comprehension (the ability to make global coherence inferences). This link is likely to occur because deep vocabulary supports associative links between the meanings of words in a text, leading to the reader making connections between entities and ideas. This perspective is supported by the finding that better comprehension is associated with a propensity to derive the theme of word lists, even though that processing is not required by the task and is in fact deleterious to the memory task. We suggested above that this propensity to derive the semantic links between words, even when not prompted, might be important in underpinning inference skills and, in particular, the propensity to draw global coherence inferences, which often depend on such semantic links. Second, inference skills also support vocabulary development because they enable children to use context to infer the meanings of words in a text that they do not know the meaning of, and there is growing evidence that reading comprehension level and inference skills support vocabulary learning (Cain, Oakhill, & Lemmon, 2004; Cain, Oakhill, & Elbro, 2003) and predict vocabulary development (Cain & Oakhill, 2011). We have evidence that reading supports the development of vocabulary breadth, but it would be helpful to have data on the extent to which children are (also) able to develop their depth of vocabulary through reading experience. Verhoeven, van Leeuwe and Vermeer (2011) provided evidence of reciprocity between reading comprehension and vocabulary development during the early school years but did not explore depth vs. breadth of vocabulary. More generally, studies are needed to explore the interrelations between the development of both depth and breadth of vocabulary over time, and the relation of these vocabulary components to comprehension skills more generally. Thus, in order to further understand the links between vocabulary knowledge and comprehension, it is important to consider vocabulary and comprehension not as unitary constructs, but to look in more detail at the processes that contribute to effective comprehension, and to consider how they are related to different aspects of vocabulary. Third, we have examined the extent to which processes at different levels might contribute to automated vs. strategic inference making in children. There is little extant data to address this issue, but we do have some preliminary data that suggest

372 Jane Oakhill and Kate Cain

that speed and automaticity of activation of semantic information is a crucial factor in children’s reading comprehension. In a study discussed earlier (Oakhill, Cain, & McCarthy, 2015), we found a strong and specific link between speed of semantic access in meaning judgment tasks (synonym and hypernym judgment tasks) and reading comprehension. Those findings suggest that it is not sufficient to “know” the meanings of words in a text, but that the facility with which the semantic representations of words can be accessed is also important for comprehension (though we did not test inference skills specifically in that study). As discussed above, the speed of access to semantic representations of words is likely to be important for efficient comprehension because comprehension happens in real time and if appropriate meanings and associations of words are not accessed very rapidly the reader will have moved on to the next word, or the next sentence, and the opportunity for semantic information to support inference and integration of the text will have been missed. Thus, there are promising indications that children who are better at text comprehension are better able to make inferences in an effortless (non strategic) manner because they are not only better, but also faster, at activating the relevant links in the text that support the inferences. The data we have presented to support this argument are correlational. In order to inform teaching and interventions, it would be important to ascertain whether there is a causal link developmentally between depth of vocabulary knowledge and reading comprehension and inference skill, or whether it is level of reading comprehension that drives the learning of, and increasing depth of knowledge of, word meanings. Further work is needed to clarify the direction of causality in this area, and also to explore whether the link between speed of semantic access and comprehension skill is mediated by inference making. As we pointed out above, the findings thus far do not preclude the idea that children who differ in text comprehension skill also differ in their ability and propensity to make strategic inferences. Indeed, we have data to show that they do (Cain & Oakhill, 1999) and suggest that the need for such inferences might be guided by the better comprehenders’ superior comprehension monitoring skills and understanding and knowledge of text structure. In this chapter, we have considered only the relation between two major categories of skill in relation to text comprehension: Word level semantic skills and text-level skills, and conclude that these should not be viewed as alternative, but rather as complementary and interactive, in accounting for ability in text comprehension, and that the complementarity of skills should be better represented in models of text comprehension. Furthermore, there are other possible interactions that we have not considered. For example recent work suggests that the relation between working memory and reading comprehension may be mediated by their association with vocabulary (Van Dyke, Johns, & Kukona, 2014). In addition, vocabulary may be related to reading comprehension through its relation with both listening comprehension and word decoding (LARRC, 2015). Thus, further work is required to explore these interactions and to consider how they are best represented in theoretical models. Furthermore, as we point out at the beginning of this chapter, existing models of text comprehension and representation are almost all intended as models of skilled



Children with specific text comprehension problems 373

adult ability. These models require modification and extension in order to be able to capture the range of skills that are known to be crucial to children’s comprehension, and to provide hypotheses about how those skills work together both in successful comprehension in children and during the development of reading ability.

References Cain, K., Lemmon, K., & Oakhill, J. V. (2004). Individual differences in the inference of word meanings from context: The-influence of reading comprehension, vocabulary knowledge, and memory capacity. Journal of Educational Psychology, 96, 671–681.  doi: 10.1037/0022-0663.96.4.671 Cain, K., & Oakhill, J. V. (1999). Inference making ability and its relation to comprehension failure in young children. Reading and Writing, 11, 489–503.  doi: 10.1023/A:1008084120205 Cain, K., & Oakhill, J. V. (2006). Profiles of children with specific reading comprehension difficulties. British Journal of Educational Psychology, 76, 683–696.  doi: 10.1348/000709905X67610 Cain, K., & Oakhill, J. V. (2007). Reading comprehension difficulties: Correlates, causes, and consequences. In K. C. J. Oakhill (Ed.), Children’s comprehension problems in oral and written language: A cognitive perspective (pp. 41–75). New York: Guilford Press. Cain, K., & Oakhill, J. V. (2011). Matthew effects in young readers: Reading comprehension and reading experience aid vocabulary development. Journal of Learning Disabilities, 44, 431–443.  doi: 10.1177/0022219411410042 Cain, K., & Oakhill, J. V. (2014). Reading comprehension and vocabulary: Is vocabulary more important for some aspects of comprehension? L’Année Psychologique, 114, 647–662. doi: 10.4074/S0003503314004035 Cain, K., Oakhill, J. V., Barnes, M. A., & Bryant, P. E. (2001). Comprehension skill, inference-making ability, and their relation to knowledge. Memory & Cognition, 29, 850–859. doi: 10.3758/BF03196414 Cain, K., Oakhill, J. V., & Bryant, P. E. (2000). Phonological skills and comprehension failure: A test of the phonological processing deficit hypothesis. Reading and Writing, 13, 31–56. doi: 10.1023/A:1008051414854 Cain, K., Oakhill, J. V., & Elbro, C. (2003). The ability to learn new word meanings from context by school-age children with and without language comprehension difficulties. Journal of Child Language, 30, 681–694.  doi: 10.1017/S0305000903005713 Cain, K., Oakhill, J., & Lemmon, K. (2004). Individual differences in the inference of word meanings from context: The influence of reading comprehension, vocabulary knowledge, and memory capacity. Journal of Educational Psychology, 96, 671–681.  doi: 10.1037/0022-0663.96.4.671 Carroll, J. B. (1993). Human cognitive abilities: A survey of factor-analytic studies. New York: Cambridge University Press.  doi: 10.1017/CBO9780511571312 Catts, H. W., Adlof, S. M., & Weismer, S. E. (2006). Language deficits in poor comprehenders: A case for the simple view of reading. Journal of Speech Language and Hearing Research, 49, 278–293.  doi: 10.1044/1092-4388(2006/023) Cunningham, A. E., & Stanovich, K. E. (1998). What reading does for the mind. American Educator, 22, 8–15. Currie, N. K., & Cain, K. (2015). Children’s inference generation: The role of vocabulary and working memory. Journal of Experimental Child Psychology, 137, 57–75. doi: 10.1016/j.jecp.2015.03.005

374 Jane Oakhill and Kate Cain

De Jong, P. F., & van der Leij, A. (2002). Effects of phonological abilities and linguistic comprehension on the development of reading. Scientific Studies of Reading, 6, 51–77. doi: 10.1207/S1532799XSSR0601_03 Deese, J. (1959). On the prediction of occurrence of particular verbal intrusions in immediate recall. Journal of Experimental Psychology, 58, 17–22.  doi: 10.1037/h0046671 Dunn, L. M., & Dunn, L. M. (1997). Peabody Picture Vocabulary Test, 3rd edition. Circle Pines, MN: American Guidance Service. Dunn, L. M., Dunn, L. M., Whetton, C. & Burley, J. (1997). British Picture Vocabulary Scale, 2nd edition. Windsor: NFER-Nelson. Dunn, L. M., Dunn, L. M., Whetton, C., & Pintillie, D. (1982). British Picture Vocabulary Scale. Windsor, UK: NFER-Nelson. Elbro, C., & Buch-Iversen, I. (2013). Activation of background knowledge for inference making: Effects on reading comprehension. Scientific Studies of Reading, 17, 435–452. doi: 10.1080/10888438.2013.774005 Johnson-Laird, P. N. (1983). Mental models: Towards a cognitive science of language, inference, and consciousness. Cambridge, MA: Harvard University Press. Kintsch, W. (1998). Comprehension: A paradigm for cognition. New York: Cambridge University Press. Kintsch, W., & van Dijk, T. A. (1978). Toward a model of comprehension and production. Psychological Review, 85, 363–394.  doi: 10.1037/0033-295X.85.5.363 Laberge, D., & Samuels, S. J. (1974). Toward a theory of automatic information processing in reading. Cognitive Psychology, 6, 293–323.  doi: 10.1016/0010-0285(74)90015-2 LARRC. (2015). Learning to read: Should we keep things simple? Reading Research Quarterly, 50, 151–169. MacGinitie, W. H., MacGinitie, R. K., Maria, K., & Dreyer, L. G. (2000). Gates-MacGinitie Reading Tests (4th ed ed.). Itasca, IL: Riverside. McNamara, D. S., & Magliano, J. (2009). Toward a comprehensive model of comprehension. In B. Ross (Ed.), The psychology of learning and motivation (Vol. 51, pp. 297–384). Academic Press.  doi: 10.1016/S0079-7421(09)51009-2 Megherbi, H., Seigneuric, A., & Ehrlich, M.-F. (2006). Reading comprehension in French 1st and 2nd grade children: Contribution of decoding and language comprehension. European Journal of Psychology of Education, 21, 135–147.  doi: 10.1007/BF03173573 Nagy, W. E., & Scott, J. (2000). Vocabulary processes. In P. M. M. Kamil, P. D. Pearson & R. Barr (Ed.), Handbook of reading research (Vol. 3, pp. 269–284). Mahwah, NJ: Erlbaum. Nation, K., & Snowling, M. J. (1998). Semantic processing and the development of word-recognition skills: Evidence from children with reading comprehension difficulties. Journal of Memory and Language, 39, 85–101.  doi: 10.1006/jmla.1998.2564 Nation, K., & Snowling, M. J. (2000). Factors influencing syntactic awareness skills in normal readers and poor comprehenders. Applied Psycholinguistics, 21, 229–241. doi: 10.1017/S0142716400002046 Oakhill, J. V. (1982). Constructive processes in skilled and less skilled comprehenders memory for sentences. British Journal of Psychology, 73, 13–20.  doi: 10.1111/j.2044-8295.1982.tb01785.x Oakhill, J. V. (1984). Inferential and memory skills in children’s comprehension of stories. Educational Psychology, 54, 31–39. Oakhill, J. V., & Cain, K. (2012). The precursors of reading ability in young readers: Evidence from a four-year longitudinal study. Scientific Studies of Reading, 16, 91–121. doi: 10.1080/10888438.2010.529219



Children with specific text comprehension problems 375

Oakhill, J. V., Cain, K., & Bryant, P. E. (2003). The dissociation of word reading and text comprehension: Evidence from component skills. Language and Cognitive Processes, 18, 443–468. doi: 10.1080/01690960344000008 Oakhill, J. V., Cain, K., & McCarthy, D. (2015). Inference processing in children: The contributions of depth and breadth of vocabulary knowledge. In E. O’Brien, A. Cook & R. Lorch (Ed.), Inferences during reading (pp. 140–159). Cambridge: Cambridge University Press. doi: 10.1017/CBO9781107279186.008 Ouellette, G. P. (2006). What’s meaning got to do with it: The role of vocabulary in word reading and reading comprehension. Journal of Educational Psychology, 98, 554–566. doi: 10.1037/0022-0663.98.3.554 Perfetti, C. (2007). Reading ability: Lexical quality to comprehension. Scientific Studies of Reading, 11, 357–383.  doi: 10.1080/10888430701530730 Perfetti, C., & Hart, L. (2001). The lexical basis of comprehension skill. In D. S. Gorfein (Eds.), On the consequences of meaning selection: Perspectives on resolving lexical ambiguity (pp. 67–86). Washington, DC: APA.  doi: 10.1037/10459-004 Perfetti, C., Wlotko, E. W., & Hart, L. A. (2005). Word learning and individual differences in word learning reflected in event-related potentials. Journal of Experimental Psychology-Learning, Memory and Cognition, 31, 1281–1292.  doi: 10.1037/0278-7393.31.6.1281 Perfetti, C., Yang, C. L., & Schmalhofer, F. (2008). Comprehension skill and word-to-text integration processes. Applied Cognitive Psychology, 22, 303–318.  doi: 10.1002/acp.1419 Quinn, J. M., Wagner, R. K., Petscher, Y., & Lopez, D. (2015). Developmental relations between vocabulary knowledge and reading comprehension: A latent change score modeling study. Child Development, 86, 159–175.  doi: 10.1111/cdev.12292 Roediger, H. L., & McDermott, K. B. (1995). Creating false memories – remembering words not presented in lists. Journal of Experimental Psychology: Learning, Memory and Cognition, 21, 803–814.  doi: 10.1037/0278-7393.21.4.803 Spencer, M., Quinn, J. M., & Wagner, R. K. (2014). Specific reading comprehension disability: Major problem, myth, or misnomer? Learning Disability Research & Practice, 29, 3–9. doi: 10.1111/ldrp.12024 Tannenbaum, K. R., Torgesen, J. K., & Wagner, R. K. (2006). Relationships between word knowledge and reading comprehension in third-grade children. Scientific Studies of Reading, 10, 381–398.  doi: 10.1207/s1532799xssr1004_3 Thorndike, R. L. (1973). Reading comprehension education in fifteen countries. New York: Wiley. Van Dyke, J. A., Johns, C. L., & Kukona, A. (2014). Low working memory capacity is only spuriously related to poor reading comprehension. Cognition, 131, 373–403. doi: 10.1016/j.cognition.2014.01.007 Verhoeven, L., & Van Leeuwe, J. (2008). Prediction of the development of reading comprehension: A longitudinal study. Applied Cognitive Psychology, 22, 407–423.  doi: 10.1002/acp.1414 Verhoeven, L., van Leeuwe, J., & Vermeer, A. (2011). Vocabulary growth and reading development across the elementary school years. Scientific Studies of Reading, 15, 8–25. doi: 10.1080/10888438.2011.536125 Weekes, B. S., Hamilton, S., Oakhill, J. V., & Holliday, R. E. (2008). False recollection in children with reading comprehension difficulties. Cognition, 106, 222–233. doi: 10.1016/j.cognition.2007.01.005 Yuill, N., & Oakhill, J. V. (1988). Effects of inference awareness training on poor reading-comprehension. Applied Cognitive Psychology, 2, 33–45.  doi: 10.1002/acp.2350020105

Part V

Instruction and intervention

Introduction to instruction and intervention Donald L. Compton, Rauno K. Parrila and Kate Cain

Florida State University / University of Alberta / Lancaster University

The fifth, and final, section of the book contains six chapters that provide us with some new and some updated insights into theoretical issues associated with instruction. This group of chapters offers a broad and unique look at both the challenges and the potential payoffs of various instructional approaches with regards to improving the skills of developing readers across a wide developmental range. Topics range from preschool to adolescent instruction, early intervention to general classroom instruction, instruction for typically developing to children with reading difficulties, and whole group to more individualized instruction. The chapters also provide insights into the potential efficacy of providing instruction across a wide range of skills associated with skilled reading performance including early oral language and code-oriented skills, basic word reading and decoding, morphological processing, inference-making, comprehension monitoring, vocabulary, background knowledge, and motivation. In the first offering, Sénéchal, Whissel and Bildfell start their chapter with the simple observation that young children start school with already existing individual differences in early literacy skills that are associated with their ease of learning to read and write. They assume that parent-child interactions before school-age are an important source of these individual differences, and review research on two different kinds of interactions derived from the Home Literacy Model: (1) formal, or code-­ oriented, where the main focus of the interaction is on the features of print, that is, on letters, their use, their combinations, as well as attempts to read and print words, and (2) informal, or meaning-oriented, where the main focus is on the meaning carried by the print. Sénéchal et al.’s review finds abundant evidence that code-oriented, but not meaning-oriented, activities are linked to children’s early literacy skills, and that meaning-oriented, but not code-oriented, activities are robustly associated with stronger vocabulary skills. They conclude that while we now understand better how parents can help their children learn early literacy skills, it is debatable whether we should encourage them to take on what traditionally was seen as the task for schools and tutor their children. This is a nice reminder that children have multiple developmental tasks to conquer, not only that of reading development. In a far-reaching chapter, Savage and Cloutier examine cognitively focused early reading interventions. They provide a partly tertiary and partly primary review of what we know about the effectiveness of phonics, reading comprehension, and doi 10.1075/swll.15.21com © 2017 John Benjamins Publishing Company

380 Donald L. Compton, Rauno K. Parrila and Kate Cain

instructional technology interventions in improving reading outcomes; they achieve this by mainly summarizing meta-analyses but also by examining more recent intervention studies. Further, Savage and Cloutier provide us with a thoughtful discussion of the role of intervention studies in general and randomized controlled trials in particular in developing and testing theories of reading development. They end with a call for more (well-designed) RCTs to be completed and published – a call we can all join as former journal editors seeking those studies! In the third chapter, Kirby and Bowers focus on morphological instruction and examine both theoretical and empirical arguments in favour of it. They first note that theoretically, morphology could be viewed as a binding agent that helps learners make sense of the semantics, phonology, and orthography of words, including their interrelationships. Kirby and Bowers then review the intervention studies and conclude that morphological instruction (a) contributes positively to literacy outcomes, (b) performs at least as well as, and often better than, other experimental interventions, (c) is effective with younger and less able learners in particular, and (d) possibly works more effectively when integrated with other aspects of instruction. Finally, they note that while there already is sufficient theoretical and empirical evidence to recommend that teachers include explicit instruction in morphology in their literacy classes, we still know relatively little about how morphological instruction works and how it could work better. Kirby and Bowers conclude their chapter with a set of design principles and hypotheses about how to develop and research the next generation of morphological instruction. In the next chapter, McMaster and Espin identify inference making as a critical component of reading comprehension and ask a simple question: Do instructional approaches and interventions exist that promote inference making and improve reading comprehension? They then review a variety of both content-focused and strategy-focused approaches and conclude that while many have shown promise for improving comprehension, it is unclear whether or how they change the fundamental learning mechanisms needed to develop reading comprehension skills. Further, they speculate that the lack of focus on fundamental learning processes may explain why instruction and interventions seldom have generalizable effects on children’s reading development. McMaster and Espin conclude that a stronger connection between reading comprehension theory and intervention programs is needed to find ways to improve children’s reading comprehension. In the fifth chapter, Vaughn and Hall shift the focus to older readers. They note that reading growth seems to plateau in middle or high school, yet many students do not have the reading skills they need to succeed in post-secondary studies. Vaughn and Hall argue that one possible reason for this stagnation in skill development is that we seldom provide adolescents reading instruction, never mind theoretically sound reading instruction, on the skills they need to master increasingly more complex and specialized content area texts. They go on to describe recently developed educational interventions aimed at providing better reading instruction for poor readers in secondary schools, and review the evidence for their effectiveness. They conclude that the



Introduction to instruction and intervention 381

recent development of theoretically informed reading comprehension interventions for secondary readers gives us a cause for optimism and a promise for all students, including the persistently struggling readers. In the final chapter, Morrison and Connor tackle the difficult problem of differentiating instruction within the framework of whole classroom instruction to better meet the needs of distinct subtypes of learners. Specifically, they review their recent work investigating child characteristic X instruction (CXI) interactions. The basic premise of CXIs is that different children might benefit from different educational experiences (a form of attribute by treatment interactions). Morrison and Connor’s CXI work is heavily influenced by Bronfennbrenner’s (1977, 1979) conceptualization of development in which sources of change go beyond the child to include the proximal (family, school) as well as more distal environment (neighborhood, socio-economic level, race-ethnicity, culture). The driving force behind their research is the realization that child and context are inextricably linked, thus necessitating the need to incorporate proximal and distal environmental factors in research designs examining the development of reading skills in children. In this chapter, the authors first explore the nature of relations between child characteristics and instruction using correlational studies and then move on to explore causal relations through the use of RCTs. Finally, Morrison and Connors conclude the chapter by previewing their current work in which they conceptualize classrooms as “complex adaptive systems” in which interdependent and reciprocal effects between teacher and child over time shape the learning environment. All of this in the hope that one day we can realize the potential of “personalized instruction” to help children maximize their learning.

Starting from home: Home literacy practices that make a difference Monique Sénéchal, Josée Whissell and Ashley Bildfell Carleton University

Young children start school with skills and knowledge that can facilitate their learning to read. Indeed, young children’s knowledge of the alphabet and their awareness of the phonemic structure of spoken language are excellent predictors of word reading growth in elementary school. Moreover, vocabulary knowledge in kindergarten is an excellent predictor of their reading comprehension in later grades. Individual differences in literacy, therefore, are established early and can have long-lasting effects. Notably, children who have difficulty learning to read in grade 1 are more likely to have difficulty in other school domains later on, are less likely to complete high school, and are less likely to pursue their education beyond high school (Alexander, Entwisle, & Horsey, 1997; Entwisle, Alexander, & Olson, 2005). Understanding how young children learn these predictive skills at home is important because the preschool and early-school years is a period of rapid cognitive, linguistic, and academic growth that has been shown to be sensitive to environmental influences (Hart & Petrill, 2009; Landry, Smith, Swank, & Guttentag, 2008). Moreover, this understanding might explain individual differences and eventually, guide intervention to optimize reading skills early in a child’s life. At present, the home environment is viewed as a source of three broad categories of literacy experiences for young children: (1) experiences in which children interact with their parents in reading and writing situations; (2) experiences in which children explore print on their own; and (3) experiences in which children observe their parents model literate behaviors when they read or write themselves (Teale & Sulzby, 1986). This broad multi-faceted description depicts the richness afforded by taking a home literacy perspective. In this chapter, we focus on parent-child interactions that promote 3- to 5-year-old children’s learning of oral and written language, that is, before reading instruction in school. The chapter consists of three main sections. In the first section, we present a theoretical model to guide our examination of the home literacy environment. In the second section, we review cross-cultural and cross-linguistic research to assess the generalizability of the proposed theoretical model. In the third section, we briefly examine research on the quality of parent-child interactions in meaning- and print-focused activities. Finally, we conclude with a critical evaluation of what is currently known, and what remains to be discovered.

doi 10.1075/swll.15.22sen © 2017 John Benjamins Publishing Company

384 Monique Sénéchal, Josée Whissell and Ashley Bildfell

The home literacy model Young children’s experience of literacy at home may take different forms, and children may learn differently as a function of these forms. In this section, we present a theoretical model that allows one to differentiate the types of home literacy activities that promote oral and written language. The purpose of the model is to explain the relations between home literacy activities and child outcomes, and as such it is intended to be descriptive, not prescriptive. The model, proposed by Sénéchal and her colleagues, is based on correlational findings from longitudinal research (Sénéchal & LeFevre, 2002; Sénéchal, LeFevre, Thomas, & Daley, 1998). The distinguishing feature of the Home Literacy Model is the idea that early home literacy experiences should be analyzed as a function of how much focus there is on the print itself. In this view, children’s literacy experiences have been categorized as formal or informal literacy activities, increasingly referred to as code-related and meaning-related interactions, respectively (e.g., Hindman, Connor, Jewkes, & Morrison, 2008). To facilitate comprehension, these latter terms are used in this chapter. Code-related activities are those where the main focus of the interaction is on the features of print, that is, on letters, their use, their combinations, as well as attempts to read and print words. It is important to consider that these interactions about the form of print do not presuppose that the parent-child interactions are structured. To be clear, these interactions can be playful (e.g., asking the child to find the letter b that is hiding in words), informative (e.g., pointing out a word that has two es in it), or didactic (e.g., encouraging the child to print the letter o by showing that it is like making a circle). In contrast, meaning-related activities are those for which the meaning carried by the print is the main focus of the interaction. A typical example is when parents, while reading a book to their child, stop reading to discuss the story. In sum, the difference between code-related and meaning-related interactions is whether the main focus of the interaction is the form of the print itself, or the meaning carried by the print. Table 1 provides definitions of these and other key terms. Before examining the full predictions of the Home Literacy Model, it is necessary to consider children’s learning about and from literacy activities. Many different terms, such as emergent literacy, early literacy, reading readiness skills, and pre-reading skills, have been used to refer to this aspect of children’s learning. These are often viewed as including knowledge pertaining to both oral and written language. However, Sénéchal, LeFevre, Smith-Chant, and Colton (2001) argued that including both oral and written language in the same global construct could lead to confounds when assessing links between the home literacy environment and child outcomes. In their synthesis of the research published between 1988 and 1997, they found that researchers were just as likely to analyze oral and written language together as to analyze them separately. Using archival data, Sénéchal et al. demonstrated how the use of a global construct led one to conclude that shared reading had a much broader influence than it did in reality. They argued that it was best to treat oral vocabulary, phonological awareness, procedural early literacy skills (e.g., alphabet knowledge),



Home literacy practices 385

Table 1.  Definitions of key terms Home literacy environment

Areas of the home environment that expose children to print. It is viewed as a source of three broad categories of literacy experiences in which children can observe, explore and learn about print.

Home literacy experiences

Fall into 3 categories in which children: (1) interact with their parents in reading and writing situations; (2) explore print on their own; and (3) observe their parents model reading and writing behaviours.

Home literacy activities

Expose children to print, such as shared reading or writing. During activities, meaning- and code-related interactions can occur.

Meaning-related interactions Focus on the meaning carried by print (e.g., stop reading to relate the story to the child’s life). Synonym: informal literacy activities Code-related interactions

Focus on the features of print (e.g., on letters, their use, their combinations, as well as attempts to read and print words). Synonym: formal literacy activities

Parent teaching early literacy Parents tutoring children on specific early literacy skills, such as letter names and sounds, to print and read words. Relevant for children before formal literacy instruction in grade school. Related terms: tutoring, training Parent teaching literacy

Parent tutoring of reading and writing to children after formal literacy instruction has begun in grade school.

Child outcomes Oral language

Children’s knowledge of spoken language, and most commonly assessed with measures of receptive and expressive vocabulary, but sometimes includes word definitions and listening comprehension.

Early literacy skills

Children’s procedural literacy knowledge, including knowledge of letter names and sounds, as well as initial reading and spelling attempts. Relevant before children begin formal literacy instruction. Related Terms: emergent literacy, pre-reading, and reading readiness

Phonological awareness

Children’s awareness of the phonological (sound) structure of words and speech.

Reading skills

Word reading, reading fluency, reading comprehension.

and conceptual knowledge about print (e.g., knowing that we read the text and not the illustrations in a picture book) as distinct constructs. Sénéchal et al. also showed that children’s conceptual knowledge about print did not predict eventual literacy skills after controlling for vocabulary, phonological awareness, and procedural knowledge – hence, conceptual knowledge is not discussed further.

386 Monique Sénéchal, Josée Whissell and Ashley Bildfell

The current landscape has changed little, as an examination of our database of 321 empirical reports published between 1998 and 2014 still does not provide a consensus in how best to analyze child outcomes. This examination also reveals that the term early literacy seems to have replaced emergent literacy. In this chapter, early literacy skills refer to children’s procedural literacy knowledge, including knowledge of letter names and sounds, as well as initial reading and spelling attempts. In addition, early literacy skills are treated as a different, but related, construct to children’s oral language (e.g., vocabulary), phonological awareness, and eventual reading and writing skills. Support for this latter view comes from longitudinal twin studies in Scandinavia, Australia, and the United States, demonstrating that phonological awareness measured in preschool seemed to be more influenced by genetics, whereas vocabulary and early literacy skills were more influenced by the shared environment (Samuelsson et al., 2007). The advantage of categorizing home literacy experiences into code-related or meaning-related activities is that it allows one to make specific predictions about their respective role in children’s learning (cf. Burgess, Hecht, & Lonigan, 2002; Burgess, 2011; Yeo, Ong, & Ng, 2014). In the Home Literacy Model, the two types of home literacy experiences are differentially related to early literacy skills, oral vocabulary, phonological awareness, and eventual reading skills. Specifically, four predictions are made: (1) code-related activities promote the acquisition of early literacy skills; (2) meaning-related activities promote the development of oral language; (3) it is through their impact on early literacy skills and vocabulary that home literacy experiences become associated with children’s phonological awareness; and (4) similar to hypothesis 3, it is through their impact on early literacy skills and vocabulary that home literacy experiences in kindergarten become associated with children’s reading skills in grade school. In the Home Literacy Model, the frequency and variety of parent-child book reading, or shared reading, has been used to index meaning-related literacy experiences. During shared reading, parent and child can enjoy the language and content of children’s books as well as the accompanying illustrations. When asked to rate the importance of a variety of reasons to read books to their 4- and 5-year old children, parents strongly endorsed the statements that they read to their children for enjoyment and to share quality time with their child (Audet, Evans, Williamson, & Reynolds, 2008). During shared reading, children can also learn. Sénéchal, LeFevre, Hudson, and Lawson (1996) described three characteristics of shared reading that can foster learning about the world and about language. First, the language used in books is typically more complex than during conversation (Hayes & Ahrens, 1988). Further, the language used by mothers is more complex during shared reading than during free play or when remembering events (Crain-Thoreson, Dahlin, & Powell, 2001). Therefore, children may be exposed to and may learn new syntactic, grammatical, and lexical forms during shared reading episodes (for a demonstration of learning new syntactic forms from book exposure, see Montag & MacDonald, 2015). Second, during shared reading a child has the undivided attention of an adult who can define, explain, and ask questions to facilitate the child’s understanding or impart new knowledge. Third and finally, during shared reading books can be re-read, thus



Home literacy practices 387

providing repeated exposure to new knowledge (Sénéchal, 1997). Because of these features, shared reading is the single most studied aspect of children’s home literacy environment. Although shared reading has the potential to enhance many facets of children’s oral language and world knowledge, the abundant quantitative research on shared reading has almost exclusively focused on demonstrating that it can be a source of vocabulary acquisition for children (Sénéchal, 2017). Shared reading can also present occasions for parents and children to discuss the printed text. Observations of parent-child interactions, however, reveal that parents seldom comment on print during these interactions. For example, in a study by Hindman, Skibbe, and Foster (2014) less than 1% of American mothers (N = 700) pointed out letters or letter sounds during shared reading, and only 10% asked the children to read out words in the text. In sharp contrast, 85% of mothers directed the child to illustrations, 63% expanded on the story, and 46% related the story to their child’s life. Similar findings have been reported in other observational studies (Audet et al., 2008; Deckner, Adamson, & Bakeman, 2006; Korat, Klein, & Segal-Drori, 2007; Stadler & Mcevoy, 2003). Observational research has also shown that 4- and 5-year-olds tend to look at the illustrations, not the written words, during shared reading (Evans & Saint-Aubin, 2005), except when their attention is explicitly drawn to the print (Justice, Pullen, & Pence, 2008). Moreover, intervention research in which parents are asked to focus on letters during shared reading did not yield statistically significant improvements in letter knowledge in samples of young children (Justice & Ezell, 2000; Justice, Skibbe, Mcginty, Piasta, & Petrill, 2011). A study conducted with at-risk children also did not show a statistically significant effect size (Effect Size [ES] = .21) in print knowledge for children whose parents participated in a program to promote shared reading versus children in a control group (Anthony, Williams, Zhang, Landry, & Dunkelberger, 2014). If shared reading is not used frequently to stimulate alphabet knowledge, if young children do not look at print readily, and if young children do not learn more when parents highlight letters during shared reading, then how do parents stimulate their child’s early literacy skills? A recent study by Martini and Sénéchal (2012) might help answer this question. They documented the contexts that parents used to help their 5-year-old child learn about early literacy skills such as the alphabet, reading, and printing. Examples of learning contexts include using familiar household items, street signs, games, the mail, newspapers, as well as children’s books. Martini and Sénéchal found that parents generally reported using a wide variety of learning contexts: Of the 18 contexts presented, parents selected on average 14 different contexts that they used at least some of the time. Moreover, Martini and Sénéchal found that parents who reported teaching early literacy skills more frequently tended to use a greater variety of learning contexts. The authors concluded that parents preferred naturally occurring activities to impart knowledge about the alphabet, reading and printing words. The reported frequency of teaching, along with the numerous learning contexts used, might also suggest that these teaching moments are not very long in duration. Importantly, frequent teaching moments that are varied and short in duration might

388 Monique Sénéchal, Josée Whissell and Ashley Bildfell

indicate parents’ sensitivity to the attention span and interest of their young child, as well as the difficulty of the task at hand. Learning the alphabet, for example, requires learning to discriminate the different forms of letters, learning that letters are symbols that represent individual speech sounds, and learning that letters have names and sounds that may or may not be the same (for an excellent description of the complexity of learning to write words, see Treiman & Kessler, 2014). Additional support for the view that parents stimulate early literacy skills in a variety of contexts comes from analyses of everyday parent-child conversations, demonstrating that parents sometimes talk to their children about letters and ask questions about letter shapes and letter-word associations (Robins, Treiman, & Rosales, 2014). Given the variability of learning contexts that parents use to impart early literacy skills to their children, researchers have relied on more general questions to survey parents on the frequency of code-related literacy experiences at home. In this chapter, the term teaching is used because it tends to be easily understood by parents. Typically, parents are asked to report on the frequency with which they teach their child about letters, to read and print words. The term teaching, however, does not presuppose that parents use structured activities, but rather that they use naturally occurring occasions to stimulate early literacy skills. Importantly, these interactions focusing on the form of print can be playful, informative, or didactic. It is also noteworthy that parents can continue to teach about reading and writing once the children are in grade school. Support for the distinct roles played by code-related and meaning-related home literacy activities in children’s oral language and literacy acquisition comes from two recent meta-analyses. The National Early Literacy Panel (2008) conducted a meta-analysis on shared reading intervention studies, where parents were taught dialogic reading techniques. Dialogic reading encourages parents to initiate interactions with children by asking questions, following child answers with further questions, repeating child utterances with enrichments, and offering praise and encouragement (Arnold, Lonigan, Whitehurst & Epstein, 1994). The meta-analysis found that increasing meaning-related interactions during shared reading enhanced oral vocabulary (ES = .60, 9 studies) but not alphabet knowledge (ES = −.06, 2 studies) or phonological awareness (ES = .11, 2 studies). The powerful role of parent teaching early literacy skills was shown in Sénéchal’s (2014) synthesis of intervention studies designed to promote early literacy (e.g., alphabet knowledge and beginning reading) by showing parents how to tutor their child (see also, Sénéchal & Young, 2008). Here, parent teaching was effective in increasing their preschool- or kindergarten-age child’s early literacy skills (ES = .94, 6 studies, 282 families). A smaller, but still statistically significant effect was found in interventions that combined training parents to tutor their child during specific literacy activities with training them to increase parent-child verbal exchanges during shared reading (ES = .33, 6 studies, 551 families). In contrast, training parents in shared reading alone did not produce statistically significant effects on these early literacy outcomes (ES = .09, 9 studies, 509 families).

Home literacy practices 389

In short, the meta-analyses on shared reading and parent teaching are in accord with the Home Literacy Model. Meaning-related interactions are related to oral vocabulary development, which will eventually be important for reading comprehension. Code-related interactions are specifically helpful for learning about the mechanics of the reading process such as alphabet knowledge, beginning reading and spelling skills. In the next section, we examine support for the Home Literacy Model across cultures and languages. The evidence reviewed is correlational in nature and will provide an overview of recent findings. A comprehensive review is beyond the scope of this chapter.

Evaluating the support for the home literacy model across cultures Research on children’s home literacy environment is being conducted in many countries differing in the nature of the written scripts, as well as in cultural values about the role of parents. In our review, we found 23 studies testing at least one prediction of the Home Literacy Model. These studies were conducted in 11 countries and represented 9 different languages. In all studies, children had attended preschool or kindergarten starting, in most cases, at age 4 or 5. Reading instruction would start at age 6 in all but two countries. Formal reading instruction begins at age 5 in England and at age 7 in Finland. A synopsis of the study characteristics, organized by type of orthography and ease of acquisition, is presented in Table 2. Because the ease with which children learn to read varies according to the nature of the orthographic script, we present first alphabetic scripts and then Chinese. In Chinese, each word, or syllables within words, are represented by a unique character consisting of interweaving strokes (Chen, Zhou, Zhao & Davey, 2010). As a result, learning to read requires the memorization of thousands of characters. Alphabetic scripts vary in the consistency with which letters and letter combinations map onto the same speech sounds in the language. For example, the grapheme ea can represent multiple English speech sounds as in bead, head, and bear. In addition, speech sounds can be spelled inconsistently too, as is the case with the sound [i] in bee, key, and beat. Children’s rate of reading acquisition is affected by this consistency, with less consistent languages requiring more time to learn (Seymour, Aro & Erskine, 2003). In the present chapter, alphabetic scripts are described in the following order. First, Indo-European languages are presented because they show great variability in orthographic consistency (Seymour, 2005). The Indo-European languages, in order of increasing consistency, that are considered in this chapter are English, French, Greek, and Spanish. These are followed by Finnish, the most consistent European orthography. Next are the Semitic languages of Hebrew and Arabic. Semitic languages are read from right to left and are considered consistent, with the exception that long vowels are not marked with letters. It is noteworthy that diacritics can be placed on letters to mark long vowels, and this is used during the initial phases of reading instruction.

390 Monique Sénéchal, Josée Whissell and Ashley Bildfell Table 2. Home literacy studies, which predictions of the home literacy model were tested, and their characteristics presented in order of written language complexity Study and prediction

Study characteristics Language Country Longitudinal

Child characteristics

Parent measures

Child outcome measures

N

Age beg

Age end

SES

Shared reading

Teaching

Oral language

Early literacy

PA Reading

Hood et al. (2008)

1, 2, 3, 4

English

AUS

Y

105

5

7

2

Ch, Fr, NB

LK, R, W

Vr

ER, W

Y

–

Evans et al. (2000)

1, 2, 3

English

CA

Y

66

5

7

3

Ch, O

LK

Vr

LK, ER

Y

R, RC

–

Frijters et al. (2000)

2

English

CA

N

92

5

–

–

Ch, Fr, NB, O

Levy et al. (2006)

1

English

CA

N

474

4–6

–

2

Factor score Factor Score

Martini & Sénéchal (2012)

1

English

CA

N

108

5

–

3

Vr

LK

N*

–

Vr

LK, ER, W

N*

R

LK, ER

N

–

–

LK, R, W

–

Skwarchuk et al. (2014) 1, 2

English

CA

Y

121

5

6

2

Ch

LK, R, W

Vr

LK, ER

N*

–

Sénéchal & LeFevre (2014)

1, 2, 3, 4

English

CA

Y

84

5

7

2

Ch, Fr, NB

LK, R, W

Vr

LK, ER, W

Y

R

Sénéchal et al. (2008)

2

English

CA

N

106

4

–

2

Ch, Fr, O

–

Ve

–

N

–

Stephenson et al. (2008) 1, 3, 4

English

CA

Y

53

5

6

3

Fr, NB

LK, R

Vr

LK

Y

–

Atkinson et al. (2014)

1, 2, 3, 4

English

UK

Y

60

3

6

–

Fr, O

R, W

LC

LK, ER

Y

RC

Kalia & Reese (2009)

2, 3

English

IND

N

50

4

–

2

Ch, Fr, NB, O

R, W

Vr

LK

Y

–

Foy & Mann (2003)

3

English

USA

N

40

4–6

–

2

Ch, Fr, O

R, W,

Ve

LK, ER

Y

–

Sparks & Reese (2012)

1, 2

English

USA

N

60

4

–

1

Fr, O

R, W

Vr, Ve

ER

N*

–

Sénéchal (2006)

1, 2, 3, 4

French

CA

Y

65

6

10

2

Fr, NB

LK, R, W

Vr

LK

Y

G1: R G4: RF, RC

Manolitsis et al. (2011) 1, 2, 3, 4

Greek

GRC

Y

70

5

9

–

Fr, NB

LK, R

Vr

LK

Y

–

Manolitsis et al. (2013) 1, 2, 3, 4

Greek

GRC

Y

82

5

6

2

Fr, NB

LK, R

Vr

LK

Y

RF

Home literacy practices 391 Table 2. (continued) Study and prediction

Study characteristics Language Country Longitudinal

Child characteristics

Parent measures

N

Age beg

Age end

SES

Shared reading

Child outcome measures

Teaching

Oral language

Early literacy

PA Reading

Strasser er & issi si (200

1,, 2 3, 4

Spanish

CHL

Y

126

5

6

3

Fr, NB, O

LK

Vr

LK, W

Y

–

Silinskas et al. (2010)

1, 4

Finnish

FIN

Y

1529

6

6

3

Fr

LK, R

–

LK, ER

N*

–

Silinskas et al. (2012)

R

N

RF

4

Finnish

FIN

Y

1436

6

7

3

Fr

LK, R

–

Aram & Aviram (2009) 2

Hebrew

ISR

N

40

5

–

2

Ch, O

–

Def

LK, ER, W N

–

Kim (2009)

Korean

KOR

Y

192

4

5

1

Fr, O

LK

Vr

LK, ER, W

Y

–

1, 2, 3

Chen et al. (2010)

1, 2

Chinese

CHN

N

97

6

–

3

Ch, Fr, NB

R, W

Vr

ER

N

–

Li et al. (2008)

1, 4

Chinese

CHN

Y

88

5

8

3

HLQ FA

R, W

Ve

ER

N

–

Note. Predictions in bold found support for the Home Literacy Model Prediction (tested) = 1: Parent teaching predicts early literacy; 2: Shared reading predicts oral language; 3: Home literacy is indirectly linked to phonological awareness; 4: Home literacy is indirectly linked to reading skills in grade school SES: 1 = Low income, 2 = Middle/high income; 3 = Mixed Country: AUS = Australia; CA = Canada; CHL = Chile; CHN = China; FIN = Finland; GRC = Greece; IND = India; KOR = Korea; NLD = Netherlands; ISR = Isreal; SG = Singapore; UK = England; USA = United States of America Shared Reading Measure: Ch = Checklist (parent checklist for child book title, author or key sentence recognition); Fr = Frequency of shared reading; HLQ FA = Home Literacy Questionnaire Factor Analysis; NB = Number of Books; O = Other (1+ other self-report questions, e.g.: frequency of visits to the library) Teaching frequency Measure: LK = Letter Knowledge (frequency of teaching the alphabet, letter names and letter sounds); R = Reading (frequency of teaching child to recognize or read words); W = Writing (frequency of teaching child to write or print) Oral Language: Vr = Receptive Vocabulary; Ve = Expressive Vocabulary; Def = Definitions; LC = Listening Comprehension; Nar = Narrative Early Literacy: LK = Letter Knowledge; ER = Early Reading (e.g., word recognition, simple decoding); W = Writing (printing, invented spelling, spelling) Literacy: Int = Interest in reading; R = Reading (word reading); RF = Reading Fluency; RC = Reading Comprehension; Sp = Spelling PA (Phonological Awareness): Y = assessed and tested; N = not assessed; N* = assessed, not tested as outcome measure

392 Monique Sénéchal, Josée Whissell and Ashley Bildfell

It should also be noted that despite the consistency of the Arabic script, learning to read is challenging for many children due to differences in phonology and vocabulary between spoken and written Arabic. Lastly, Korean is presented. As described by Kim (2009), the Hangul alphabetic orthography is highly consistent. This diversity of languages, orthographies, and countries will provide a good test of the generalizability of the Home Literacy Model. In addition to providing a synopsis of study characteristics, Table 2 also indicates for each study which of the four key predictions of the Home Literacy Model was tested. Here, a number in bold signifies that support for a prediction was found. As shown in Table 2, code-related activities were typically measured as parent teaching early literacy skills (e.g., teaching letter names) whereas meaning-related activities were typically measured as the frequency and/or variety of shared reading at home. Sometimes book exposure was used as a proxy measure for shared reading. Because the initial longitudinal study by Sénéchal and colleagues (Sénéchal et al., 1998; Sénéchal & LeFevre, 2002) was instrumental to the Home Literacy Model and has already been described, it is not included in the table. The research testing each of the four predictions of the Home literacy Model is described next.

Prediction 1: Parent teaching predicts early literacy The first prediction of the Home Literacy Model is that the frequency of parent teaching about literacy is directly linked to early literacy skills. In the present review, support for this prediction was found in 15 studies representing six languages, namely, English, French, Greek, Spanish, Finnish and Chinese. For the majority of these studies, teaching of early literacy was assessed through questionnaires, with parents reporting the frequency of teaching the alphabet, letter sounds, to print and/or to read words. As indicated in Table 2, early literacy outcomes included one or more of the following measures: alphabet knowledge (letter names and/or sounds), early reading, and early writing (e.g., printing letters, attempts to write words). Some studies assessed early literacy with a standardized subtest called letter-word identification that combines letter knowledge and word reading. In English-speaking countries, parents report teaching early literacy frequently. For example, 81%, 77%, and 57% of Australian parents reported teaching often to very often the alphabet, to write the child’s name, and to read words, respectively (Hood, Conlon, & Andrews, 2008). Conducting stringent mediation analyses that included child vocabulary, Hood et al. confirmed that parent teaching was directly linked to child early literacy (i.e., letter knowledge, early reading), but that shared reading was not. Similar results were found in England (Atkinson, Powell, Slade & Levy, 2014), in Canada with both English (Evans, Shaw, & Bell, 2000; Levy, Gong, Hessels, Evans, & Jared, 2006; Martini & Sénéchal, 2012; Stephenson, Parrila, Georgiou, & Kirby, 2008), and French-speaking parents (Sénéchal, 2006), in India with families whose children were schooled in English (Kalia & Reese, 2009), and in Greece (Manolitsis, Georgiou, & Parrila, 2011). There was also some evidence of the role of parent teaching in Finnish (Silinskas, Parrila, Lerkkanen, Poikkeus, Niemi, & Nurmi, 2010). Here, parent reports of the



Home literacy practices 393

frequency with which they teach reading, measured in the spring of kindergarten, significantly differentiated children who learned to read words during kindergarten from those who did not. Shared reading did not differentiate kindergarten readers from non-readers, nor did it differentiate those who learned to read some words from children who were already reading many words at the beginning of kindergarten. Chen et al. (2010) corroborated this finding in a sample of 97 families from mainland China. Prior to formal reading instruction at the beginning of grade 1, parent teaching predicted Chinese character recognition after controlling for child age, intelligence, and parent education.

Longitudinal support

In the research so far, parent teaching and child early literacy skills were assessed concurrently, and this makes it difficult to interpret the direction of the correlation despite having included stringent controls in the analyses. In the next four studies, however, the relations are assessed longitudinally. Longitudinal tests provide stronger assessments of the predictive nature of parent teaching to early literacy skills. First, Skwarchuk, Sowinski, and LeFevre (2014) found that parent reports of the frequency of teaching, assessed in the summer before kindergarten, significantly predicted children’s early literacy skills one year later after controlling for a number of variables such as income, phoneme awareness, vocabulary, and shared reading. Second, Strasser and Lissi (2009) found that Chilean mothers’ reports of teaching frequency predicted emergent writing growth during kindergarten, after controlling for mother education, shared reading, and classroom instruction. Third, Manolitsis, Georgiou and Tziraki (2013) assessed Greek children from kindergarten to grade 1. In this sample, parents, on average, reported teaching letter sounds and identification of letters a few times a week. Path analyses that included shared reading and child kindergarten skills showed that it was parent teaching, not shared reading, in kindergarten that significantly predicted child letter knowledge at the beginning of grade 1. A fourth study also confirmed that parent teaching, not shared reading, predicted growth in early literacy skills from kindergarten to the beginning of grade 1 (Sénéchal & LeFevre, 2014). The novelty of this particular study is that the home and school languages were different. Specifically, the families spoke English – the dominant language in the community, but the children were schooled in a minority language, namely French. Because all assessments were in English, the design allowed for more separation between learning about early literacy skills at home and at school than in previous studies.

Lack of support

Four studies, however, did not find support for Prediction 1. First, the frequency of teaching reported by Indian parents did not predict their children’s knowledge of the English alphabet after controlling for age and maternal education (Kalia & Reese, 2009). Second, both shared reading and parent teaching predicted child reading in

394 Monique Sénéchal, Josée Whissell and Ashley Bildfell

Chinese in a sample of families from Hong Kong (Li, Corrie, & Wang, 2008). Of note, this latter study did not include a measure of child vocabulary, and therefore, did not provide a strong test of whether shared reading would still predict reading with vocabulary entered in the analysis. Third, Sparks and Reese (2012) examined the home literacy environment of 60 4-year-old children when they entered the Head Start program – a program aimed to increase the cognitive, social, and emotional development in preschoolers from low-income homes. A reduced sample (size unknown) of parents completed a survey on the frequency of shared reading and teaching to read and print words. Here, it was shared reading, not parent teaching to read that was associated with child early literacy skills. Although unclear, it might be the case that questions on teaching more basic skills such as the alphabet might have yielded a different pattern of results as suggested by Martini and Sénéchal (2012) and shown in Sénéchal (2006). Fourth and final, Kim (2009) found that Korean parents’ reports of reading at home (including the frequency of child independent reading) positively predicted child letter knowledge and word reading growth over the course of the study. In contrast, parent teaching of Hangul (i.e., the orthographic script) and helping with homework was not related to children’s letter knowledge, but was negatively related to word reading and pseudoword reading (i.e., a measure of their ability to recode letters into sounds). These findings, however, are difficult to interpret because the preschool children were learning to read over the course of the study (from reading 20% to 66% of words in a 60-word reading task). Moreover, parents completed the questionnaire 10 months into the study, making interpretation even more difficult. Hence, independent reading might have provided word-reading practice and, therefore, is not a measure of shared reading, whereas help with homework might be indicative of parents helping when their child struggled. These study design issues are unfortunate given an otherwise very interesting study. All things considered, there is abundant cross-cultural and cross-linguistic evidence that parent teaching, but not shared reading, is robustly linked to children’s early literacy skills. The four studies that do not support this prediction had a number of design characteristics that might explain the discrepancy in findings.

Prediction 2: Shared reading predicts oral language The second prediction of the Home Literacy Model is that shared reading, not parent teaching, is directly associated with child oral language. Support for this prediction was found in 13 studies across six languages, namely, English, French, Greek, Hebrew, Korean, and Chinese. In the majority of these studies, the frequency of shared reading was assessed with questionnaires and child oral language was assessed with measures of receptive vocabulary. Hood et al. (2008) conducted stringent analyses and confirmed that shared reading was directly linked to child vocabulary after controlling for age, memory, nonverbal intelligence, and letter knowledge. This pattern was found in England (Atkinson et al., 2014),



Home literacy practices 395

English Canada (Frijters, Barron, & Brunello, 2000; Sénéchal, Pagan, Lever, & Ouellette, 2008; Skwarchuk et al., 2014), India (Kalia & Reese, 2009), French Canada (Sénéchal, 2006), Greece (Manolitsis et al., 2013), Israel with Hebrew-speaking participants (Aram & Aviram, 2009), and China (Chen et al., 2010). As seen in Table 2, this pattern held for children’s receptive vocabulary, expressive vocabulary, and their ability to define words. Further, these studies also clearly showed that shared reading did not predict letter knowledge (Aram & Aviram, 2009; Frijters et al., 2000; Hood et al., 2008; Manolitsis et al., 2013; Sénéchal, 2006; Skwarchuk et al., 2014). To be clear, when an association between shared reading and letter knowledge is found, it is no longer significant once we control for child vocabulary (e.g., Hood et al., 2008; Foy & Mann, 2003) or other closely related variables to letter knowledge such as phoneme awareness (Frijters et al., 2000).

Longitudinal support

In addition to the concurrent associations above, Sénéchal and LeFevre (2014) showed that shared reading in kindergarten predicted growth in child vocabulary at the beginning of grade 1. Similarly, Kim (2009) found, in a Korean sample, that the measure of reading at home predicted growth in vocabulary over 13 months (starting when the children were, on average, 56 months of age). The pattern also holds with very young children: Using structural equation modeling, Torppa et al. (2007) confirmed that parent reports of the frequency of shared reading when children were 2 years predicted child vocabulary at age 3.5 years.

Lack of support

Not all studies, however, found complete support for this prediction. For instance, Kim (2009) found that while home reading positively predicted vocabulary growth, parent teaching was a negative predictor. As mentioned previously, Kim’s results are difficult to interpret due to the study’s design. In other research for which the samples varied in socio-economic status, the association between shared reading and child vocabulary was no longer significant once the researchers controlled for mothers’ education (e.g., Evans et al., 2000; Manolitsis et al., 2011; Strasser & Lissi, 2009). This is important because in the research conducted by Sénéchal, mothers’ education is often not correlated with child outcomes, given their samples of mostly middle-class families. Sénéchal (2006) presented the argument that mothers’ education, however, is a more distal explanatory variable than shared reading (see Marjoribanks, 1979). Hence, researchers who are interested in understanding the effect of shared reading should study it within social strata. To do so would provide a stronger test of whether the variability in the frequency of shared reading explains differences in child vocabulary in samples of parents who have similar levels of education. In sum, there is both cross-cultural and cross-linguistic evidence that the frequency of shared reading, as reported by parents, is associated with stronger vocabulary skills. This relation holds even after entering strict controls. Importantly, shared reading was not found to be linked directly to children’s letter knowledge and parent teaching was typically not found to be directly associated to children’s vocabulary.

396 Monique Sénéchal, Josée Whissell and Ashley Bildfell

Parent teaching and shared reading are distinct The underlying assumption of the Home Literacy Model is that code-related and meaning-related activities are distinct. This distinction has been tested and support for it found (Hood et al., 2008; Levy et al., 2006; Li et al. 2008). In these studies, parents responded to questionnaires and their responses were analyzed for commonalities: parent teaching and shared reading questions loaded on different factors. These results confirm the theoretical view adopted in the Home Literacy Model, and show that, methodologically, omnibus measures (i.e., adding all responses without theory or verification) of parent-child home literacy activities should be avoided because they might lead to faulty conclusions (e.g., Yeo et al., 2014).

Prediction 3: Home literacy indirectly predicts phoneme awareness The third prediction of the Home Literacy Model is that any association found between home literacy activities and phoneme awareness should be mediated through children’s early literacy skills and vocabulary. That is, parent-child activities enhance children’s early literacy and vocabulary, which, in turn, enhance phoneme awareness. In eight studies conducted with English, French, and Greek families, the links between home literacy activities and phoneme awareness were mediated through child vocabulary and/or early literacy (Atkinson et al., 2014; Foy & Mann, 2003; Evans et al., 2000; Hood et al., 2008; Manolitsis et al., 2013; Sénéchal, 2006; Sénéchal & LeFevre 2014; Stephenson et al., 2008). In two other studies, the association between home literacy activities and phoneme awareness was mediated by other factors such as child’s age, mother’s education, or kindergarten classroom activities (Kalia & Reese, 2009) or not related to home literacy at all (Strasser & Lissi, 2009). Further evidence of the role of vocabulary and letter knowledge to phoneme awareness was found in Torppa et al. (2007). They analyzed predictors of growth in phonological awareness in a sample of 186 children followed from birth to age 6. Using structural equation modeling, Torppa et al. confirmed that the frequency of shared reading did not predict phoneme awareness. Over time, it was child vocabulary and letter knowledge that predicted phoneme awareness and vice versa. The study by Kim (2009), however, did not reproduce this general pattern of findings. In this study, home reading positively predicted phoneme awareness growth whereas parent teaching was a negative predictor after controlling for all appropriate child measures. As mentioned previously, Kim’s results are difficult to interpret due to the study’s design. Moreover, Manolitsis et al. (2011) also reported that parent teaching was negatively associated with phoneme awareness once appropriate controls were entered in the analyses, but these findings were not replicated in Manolitsis et al. (2013). On the whole, the findings of 11 out of 13 studies reviewed for Prediction 3 provide support for the notion that home literacy activities are not directly linked to children’s phoneme awareness.



Home literacy practices 397

Prediction 4: Home literacy indirectly predicts eventual reading skills The fourth prediction of the Home Literacy Model is that the association between home literacy and reading skills in grade school is explained by children’s early literacy skills and/or vocabulary. This prediction was clearly supported in eight longitudinal studies. Hood et al. (2008) found that any long-term links between parent teaching in kindergarten and reading in grades 1 and 2 were mediated through children’s early literacy skills. Similar mediated links were found for word reading in grade 1 (Sénéchal, 2006; Stephenson et al., 2008), reading fluency in grade 1 (Manolitsis et al., 2013), for reading in grade 2 (Atkinson et al., 2014; Sénéchal & LeFevre, 2014), and reading comprehension in grade 4 (Manolitsis et al., 2011; Sénéchal, 2006). Not controlling for children’s early skills might lead one to conclude that there is a direct predictive role of home literacy in kindergarten to reading skills in grade school (see Li et al., 2008). Two other studies, however, yielded mixed findings by showing an absence of association (i.e., simple correlations) between teaching in kindergarten and reading in grade 1 (Silinskas, Lerkkanen, Tolvanen, Niemi, Poikkeus, & Nurmi, 2012); or that it was home storybook exposure, not parent teaching, that predicted growth in reading from kindergarten to the end of grade 1 (Strasser & Lissi, 2009). Yet, other studies have nuanced Prediction 4. In most longitudinal studies reviewed, reading in grade school was assessed with measures of word reading or reading comprehension. The study by Sénéchal (2006) stands out here because she also assessed reading fluency and children’s reported frequency of reading for pleasure in grade 4. Sénéchal found statistically significant relations between French-speaking parents’ reports of teaching in kindergarten and children’s grade 4 reading fluency after controlling for early literacy skills and grade 1 reading. This finding might suggest that parents who report teaching frequently might also listen to their child reading more once the children enter grade school; however, it was not replicated in a Greek study in which reading fluency was tested in grade 4 (Manolitsis et al., 2011). Whether the difference across these two studies is due to the nature of the orthography (the more difficult French versus the easier Greek) remains to be ascertained. The second interesting result in Sénéchal (2006) was that parents’ reported frequency of shared reading in kindergarten predicted children’s reported frequency of reading for pleasure in grade 4 after controlling for parents’ education, parent teaching, grade 1 reading, and grade 4 reading comprehension. This is, to our mind, the first prospective demonstration of the relation between shared reading and children’s eventual independent reading. Moreover, Atkinson et al. (2014) also extended the Home Literacy Model by showing that children’s listening comprehension in grade 1 was significantly predicted by their kindergarten vocabulary as well as parent reports of shared reading (but not teaching). Another nuance to Prediction 4 comes from studies in which parent reports of teaching frequency in kindergarten are negatively related to subsequent child outcomes. For instance, Kim (2009) found negative paths between parent teaching in the first year of preschool and their children’s later reading skills. Kim interpreted the findings as

398 Monique Sénéchal, Josée Whissell and Ashley Bildfell

showing that parents of children who were lagging behind may have increased their rate of teaching. Silinskas, Leppänen, Aunola, Parrila, and Nurmi (2010) found that children’s reading performance at the beginning of grade 1 was negatively related to parent reports of teaching the alphabet in the spring of grade 1, r(161) = −.31. Moreover, Silinskas et al. (2012) replicated the negative association between parent reports of teaching reading and word reading at the end of grade 1, r(684 and 752) = −.37 and −.35 for girls and boys, respectively. Taken together, findings from these three studies might show that parents teach their children for different reasons in kindergarten than they do once the children are in grade school (Silinskas, Parrila et al., 2010). A recent study by Sénéchal and LeFevre (2014) supported this possibility by showing that parents seemed to adapt their teaching behaviors according to children’s reading performance. Children whose parents decreased their teaching from grade 1 to grade 2 had significantly higher reading scores at the end of grade 1 than children whose parents increased their teaching. Parents who increased their teaching also reported increasing the frequency with which they listened to their child read. The results of the four studies, therefore, suggest that parents are sensitive to their child’s reading performance and adjust their code-related literacy activities accordingly. Moreover, these results help understand the well-known pattern that parent help with homework is negatively related to achievement later on (for a meta-analysis, see Hill & Tyson, 2009). As a whole, the findings reviewed in this sub-section are mixed: Support for Prediction 4 was found in 6 longitudinal studies, two other studies revealed other types of associations, and yet four other studies nuanced the prediction. Further work is required to ascertain whether Prediction 4 of the Home Literacy Model needs to be modified.

Implications The cross-cultural and cross-linguistic findings reviewed tend to be consistent with the assumption and the main predictions of the proposed Home Literacy Model. This seems to be the case only when researchers avoid omnibus measures and include appropriate controls in their analyses. First, support for the prediction that it is parent teaching that is positively associated with child early literacy skills has been found in English, French, Spanish, Greek, Finnish, and Chinese. Second, support for the prediction that it is shared reading that is positively associated with child oral language has been found in English, French, Greek, Finnish, Hebrew, Korean, and Chinese. Third, support for the prediction that it is child vocabulary and early literacy, not home literacy, that are directly linked to phoneme awareness was also found in English, French, Spanish, and Greek. Fourth, there is some evidence in English, French, and Greek that any long-term links between parent teaching in kindergarten and reading in grade school are mediated through children’s early literacy and vocabulary. To conclude, the Home Literacy Model seems to be a useful and generalizable paradigm that leads to a better understanding of how parents help their child learn oral and written language.



Home literacy practices 399

Assessing the quality of parent-child interactions Most of the research on the Home Literacy Model has focused on the frequency and variety of literacy activities, be it shared reading or parent teaching. In this section, we examine whether the quality of those interactions matters. Our intent is not to conduct a comprehensive review, but rather to illustrate the type of research conducted.

Shared reading During shared reading, parents can engage their children in verbal exchanges, and the quality of those exchanges can impact child oral language learning (Mol, Bus, de Jong, & Smeets, 2008). Parent talk during shared reading has been classified on a continuum ranging from verbal requests anchored in the present to verbalizations that extend beyond the present context such as when parents connect story elements to their child’s personal experiences, or when they asked their child to predict what will happen next in the story (de Jong & Leseman, 1998). The hypothesis is that parent talk that is decontextualized (i.e., beyond the present context) is richer and more conducive to learning. There is some evidence that this is the case. For instance, Hindman and colleagues found positive relations between the quality of parent extra-textual talk and child vocabulary in two large-scale American studies (Hindman et al., 2008, 2014). Moreover, de Jong and Leseman (2001) found, in a sample of 69 Dutch families, that both shared reading frequency and quality were significantly related to grade 1 vocabulary and decoding. However, shared reading frequency and quality were not significant predictors of grade 3 reading comprehension after controlling for grade 1 vocabulary. In a series of studies, Korat and her colleagues examined whether the quality of mothers’ extra-textual talk during shared reading would be linked to children’s early literacy and oral language. The findings showed positive associations between mother-talk quality and children’s early literacy (e.g., alphabet knowledge, word recognition as well as phoneme awareness) in Hebrew-speaking families (Korat, 2009; Korat et al., 2007), and in Arabic-speaking families (Korat, Hassunah-Arafat, Aram, & Klein, 2013). In this latter study, mother-talk quality was also linked to children’s oral language (a combination of vocabulary and listening comprehension) in kindergarten, but it was not associated with reading skills in grade 1. The results showing a predictive relation between the quality of shared reading and literacy skills, however, are difficult to interpret. In light of the Home Literacy Model, it is unclear whether mother-talk quality would still be linked to children’s early literacy were parent teaching and child oral language considered as control variables. The results are also difficult to interpret because Korat et al. (2007) showed that, on average, only 5% of parent talk during shared reading was code-related (i.e., about the written system). If parents talk very little about the writing system during shared reading, then one wonders by what mechanism learning about letter names and sounds occurs. Hence, there is some evidence that the quality of parent-talk during shared reading is linked to child oral language, but the robustness of the link between talk quality and children’s early literacy remains to be ascertained.

400 Monique Sénéchal, Josée Whissell and Ashley Bildfell

Parent teaching The quality of parent-child interactions during code-related activities has received little attention. In this section, we focus on research by Aram, Levin and colleagues – observational research showing that parents adapt their level of scaffolding to their child’s early literacy skills in a writing task. In this research, the degree of parental guidance (or “mediation” in Aram’s terminology) can range from a lower quality level (e.g., mother prints the letter or word for the child and asks the child to copy) to a higher quality level (e.g., the mother encourages the child to identify the sounds in words and to find a corresponding letter for each sound). In their study of 41 Hebrew mother-child dyads from low socio-economic backgrounds, Aram and Levin (2001, 2002) demonstrated that the quality of parents’ guidance during a word printing task predicted the children’s word writing, after controlling for SES, parent literacy, and parent reports of home literacy activities and tools. Although the correlational findings do not allow us to make causal assertions, they do suggest that mothers were sensitive to their kindergarten children’s early literacy skills. In order to examine the issue of maternal sensitivity further, Aram (2007) observed that, in a sample of 28 sets of twins, mothers adapted the quality of their guidance to each twin’s printing level. Fathers too adapt their guidance style to child skills as shown in a study of 51 families (Aram, 2010). The findings so far were obtained with Israeli Hebrew- and Arabic-speaking families. Recent work in Chinese has shown that the pattern of association generalizes to this culture (Mainland China: Lin et al., 2009; Hong Kong: Lin, McBride-Chang, Aram & Levin, 2011). Moreover, Aram, Korat, and Hassunah-Arafat (2013) found that the quality of parent-child interactions during a word printing task in kindergarten was a robust positive predictor of grade 1 reading and writing after controlling for SES, as well as children’s vocabulary and letter naming in kindergarten. Shared reading, however, was not a significant predictor of grade 1 reading. This study showed the importance of going beyond measures of frequency. It will be important to replicate this longitudinal pattern while controlling for more advanced early literacy skills, such as early reading and spelling. The correlational findings presented so far are strengthened by an intensive 7-week training study with low-SES Israeli families (Levin & Aram, 2012). In a print-guidance condition, 30 parents of 5-year-olds were coached on how to scaffold printing during print-focused games that involved letters and spelling. Among the three control conditions, the shared reading condition is of particular interest. In this condition, 30 parents were trained to increase the quality of parent-child language interactions during shared reading. On pre-test, there were no differences across conditions in parent guidance and in children’s alphabet knowledge. On both the immediate and delayed (2.5 months later) post-tests, parents produced the desired behaviors appropriate for their condition. Moreover, children in the print-guidance condition had greater alphabet knowledge on the two post-tests than the children in the shared reading condition. Notably, the researchers also tested whether training parents in



Home literacy practices 401

one condition (e.g., shared reading) would generalize to the activities in the other training condition (e.g., printing). Unfortunately, teaching parents about scaffolding in one task did not transfer to the other task. As with the research on the quality of parent talk during shared reading, there is some evidence that the quality of parents’ interactions during a writing task is linked to child early literacy as well as reading in grade school.

Implications This section highlights the importance of considering the quality of the parent-child interactions during home literacy activities. Doing so might enrich the Home Literacy Model. At the same time, this research also needs to acknowledge the interplay between young children’s oral language and early literacy skills in order to clarify the nature of the obtained findings. The intervention research conducted by Levin and Aram (2012) provides evidence that parents can improve the quality of their teaching behaviors and that those improvements were linked to lasting improvements in children’s alphabet knowledge (for meta-analyses on parent tutoring, see Sénéchal, 2014; van Steensel, McElvany, Kurvers, & Herppich, 2011).

Conclusion Most of the evidence presented in this chapter was in accord with the proposed Home Literacy Model: Informal (meaning-related) literacy activities such as shared reading can help young children learn oral language whereas parent-child formal (code-related) literacy interactions seem to be necessary for gains in early literacy skills. In addition, the relation between home literacy and child phoneme awareness appears to be mediated through child vocabulary and letter knowledge. Over time, it is children’s early skills that were predictive of eventual success in reading in most studies. The research reviewed also allows one to reflect on how to pursue this field of inquiry. First, there is certainly a need for researchers to increase the strength of the evidence about the role of home literacy by conducting more longitudinal research. Doing so would allow one to test predictors of growth in child outcomes rather than concurrent relations. To do so, however, will require larger samples and the use of more sophisticated statistical analyses. Second, there is a need to move beyond SES comparisons. In our opinion, we do not need additional research showing that middle-class parents report providing more literacy resources (books, games) and engaging in more literacy activities with their children (e.g., shared book reading) than parents of lower SES. Nor do we need additional research showing that children from higher-SES families outperform children from low-SES families. The key question, to our mind, is whether variation within socio-economic strata explains variance in child outcomes (e.g., Farver, Xu, Eppe, & Lonigan, 2006; Farver, Xu, Lonigan & Eppe, 2013).

402 Monique Sénéchal, Josée Whissell and Ashley Bildfell

Shared reading Parents read books to their child for enjoyment and the quality time it affords. At the beginning of the chapter, we reviewed reasons why shared reading is an occasion for learning oral language (and world knowledge, for that matter) and showed evidence that this can be the case. In contrast to these positive effects, the review of the home literacy literature showed that shared reading is not a source of early literacy learning for children – at least not their learning of procedural knowledge (e.g., alphabet knowledge, early spelling attempts). Hence, it seems that parents of young children are right to limit the number of print-focused interactions during shared reading. This is food for thought for researchers and practitioners who might be tempted to transform shared reading into a source of early literacy learning. At the same time, numerous questions still need to be addressed about shared reading. For example, rather than improving early literacy, it could be that shared reading increases children’s motivation to read for pleasure. Hints of this were found in Sénéchal (2006) where the frequency of shared reading with 5-year-olds predicted longitudinally the frequency of children reading for pleasure four years later.

Parent teaching about literacy Parents across a variety of cultures and languages report teaching about the mechanics of reading and writing at home. The children of parents who report teaching more frequently tend to have stronger early literacy skills than children whose parents report teaching less. Moreover, parents seem to take advantage of naturally occurring occasions to impart knowledge about letters, reading, and writing. Experimental and quasi-experimental studies show that, when parents are trained to do specific activities that focus on the mechanics of literacy, their children do better than children whose parents were not trained or did alternative activities. Hence, the evidence reviewed shows that parents can help their children learn early literacy skills. The issue to consider is whether parents should. The Home Literacy Model was meant to describe what parents do at home and to show accurately the relation between what parents report doing and different child outcomes. The Model, however, was not intended to be prescriptive. It is therefore debatable whether the findings on parent teaching should be used to encourage parents to tutor their young children about literacy. Perhaps it is time to turn our attention to studying how to build strong partnerships between the home and the school in order to optimize early literacy in young children.



Home literacy practices 403

References Alexander, K. L., Entwisle, D. R., & Horsey, C. (1997). From first grade forward: Early foundation of high school dropout. Sociology of Education, 70, 87–107.  doi: 10.2307/2673158 Anthony, J. L., Williams, J. M., Zhang, Z., Landry, S. H., & Dunkelberger, M. J. (2014). Experimental Evaluation of the Value Added by Raising a Reader and Supplemental Parent Training in Shared Reading. Early Education and Development, 25, 493–514.  doi: 10.1080/10409289.2013.812484 Aram, D. (2007). Sensitivity and consistency of maternal writing mediation to twin kindergartners. Early Education and Development, 18, 71–92.  doi: 10.1080/10409280701274733 Aram, D. (2010). Writing with young children: A comparison of paternal and maternal guidance. Journal of Research in Reading, 33, 4–19.  doi: 10.1111/j.1467-9817.2009.01429.x Aram, D., & Aviram, S. (2009). Mothers’ storybook reading and kindergartners’ socioemotional and literacy development. Reading Psychology, 30, 175–194.  doi: 10.1080/02702710802275348 Aram, D., Korat, O., & Hassunah-Arafat, S. (2013). The contribution of early home literacy activities to first grade reading and writing achievements in Arabic. Reading and Writing, 26, 1517–1536  doi: 10.1007/s11145-013-9430-y Aram, D., & Levin, I. (2001). Mother-child joint writing in low SES. sociocultural factors, maternal mediation and emergent literacy. Cognitive Development, 16, 831–852. doi: 10.1016/S0885-2014(01)00067-3 Aram, D., & Levin, I. (2002). Mother-child joint writing and storybook reading: Relations with literacy among low SES kindergartners. Merrill-Palmer Quarterly, 48, 202–224. doi: 10.1353/mpq.2002.0005 Arnold, D.S., Lonigan, C.J., Whitehurst, G.J., & Epstein, J.N. (1994). Accelerating language development through picture-book reading: Replication and extension to a videotape training format. Journal of Educational Psychology, 86, 235–243. Atkinson, L., Powell, D., Slade L., & Levy J. P. (2014, July). Investigating the influence of home literacy practices on the developmental pathways of early reading comprehension. Symposium presentation at the annual meeting of the Society for the Scientific Study of Reading, Santa Fe, New Mexico. Audet, D., Evans, M. A., Williamson, K., & Reynolds, K. (2008). Shared Book Reading: Parental Goals Across the Primary Grades and Goal – Behavior Relationships in Junior Kindergarten. Early Education & Development, 19, 112–137.  doi: 10.1080/10409280701839189 Burgess, S., Hecht, S. A., & Lonigan, C. J. (2002). Relations of the Home Literacy Environment (HLE) to the Development of Reading-Related Abilities: A One-Year Longitudinal Study. Reading Research Quarterly, 37, 408–426.  doi: 10.1598/RRQ.37.4.4 Burgess, S. R. (2011). Home literacy environments (HLEs) provided to very young children. Early Child Development and Care, 181, 445–462.  doi: 10.1080/03004430903450384 Chen, X., Zhou, H. U. I., Zhao, J., & Davey, G. (2010). Home Literacy Experiences and Literacy Acquisition among Children in Guangzhou, South China1. Psychological Reports, 107, 354– 366.  doi: 10.2466/04.11.17.21.28.PR0.107.5.354-366 Crain-Thoreson, C., Dahlin, M. P., & Powell, T. A. (2001). Parent-child interaction in three conversational contexts: Variations in style and strategy. In P. R. Britto & J. Brooks-Gunn (Eds.), Role of family literacy environments in promoting young children’s emerging literacy skills (pp. 7–23). San Francisco: Jossey-Bass. de Jong, J., & Leseman, P. P. M. (2001). Lasting effects of home literacy on reading achievement in school. Journal of School Psychology, 39, 389–414.  doi: 10.1016/S0022-4405(01)00080-2

404 Monique Sénéchal, Josée Whissell and Ashley Bildfell

Deckner, D. F., Adamson, L. B., & Bakeman, R. (2006). Child and maternal contributions to shared reading: Effects on language and literacy development. Applied Developmental Psychology, 27, 31–41.  doi: 10.1016/j.appdev.2005.12.001 Entwisle, D. R., Alexander, K. L., & Olson, L. S. (2005). First grade and educational attainment by Age 22: A new story. American Journal of Sociology, 110, 1458–1502.  doi: 10.1086/428444 Evans, A. R., & Saint-Aubin, J. (2005). What children are looking at during shared storybook reading. Evidence from eye movement monitoring. Psychological Science, 16, 913–920. doi: 10.1111/j.1467-9280.2005.01636.x Evans, M. A., Shaw, D., & Bell, M. (2000). Home literacy activities and their influence on early literacy skills. Canadian Journal of Experimental Psychology, 54, 65–75.  doi: 10.1037/h0087330 Farver, J. M., Xu, Y., Lonigan, C. J., & Eppe, S. (2013). The home literacy environment and Latino Head Start Children’s emergent literacy skills. Developmental Psychology, 49, 775–791. doi: 10.1037/a0028766 Farver, J. M., Xu, Y., Eppe, S., & Lonigan, C. J. (2006). Home environments and young Latino children’s school readiness. Early Childhood Research Quarterly, 21, 196–212. doi: 10.1016/j.ecresq.2006.04.008 Foy, J. G., & Mann, V. (2003). Home literacy environment and phonological awareness in preschool children: Differential effects for rhyme and phoneme awareness. Applied Psycholinguistics, 24, 59–88.  doi: 10.1017/S0142716403000043 Frijters, J. C., Barron, R. W., & Brunello, M. (2000). Direct and mediated influences of home literacy and literacy interest on prereaders’ oral vocabulary and early written language skill. Journal of Educational Psychology, 92, 466–477.  doi: 10.1037/0022-0663.92.3.466 Hart, S. A., & Petrill, S. A. (2009). The genetics and environments of reading: A behavioral genetic perspective. In Y. Kim (Ed.) Handbook of behavior genetics Part II (pp. 113–123). New York: Springer.  doi: 10.1007/978-0-387-76727-7_8 Hayes, D. P., & Ahrens, M. G. (1988). Vocabulary simplification for children: a special case of ‘motherese’? Journal of Child Language, 15, 395–410.  doi: 10.1017/S0305000900012411 Hill, N. E., & Tyson, D. F. (2009). Parental involvement in middle school: A meta-analytic assessment of the strategies that promote achievement. Developmental Psychology, 45, 740–763. doi: 10.1037/a0015362 Hindman, A. H., Connor, C. M., Jewkes, A. M., & Morrison, F. J. (2008). Untangling the effects of shared book reading: Multiple factors and their associations with preschool literacy outcomes. Early Childhood Research Quarterly, 23, 330–350.  doi: 10.1016/j.ecresq.2008.01.005 Hindman, A. H., Skibbe, L. E., & Foster, T. D. (2014). Exploring the variety of parental talk during shared book reading and its contributions to preschool language and literacy: evidence from the Early Childhood Longitudinal Study-Birth Cohort. Reading and Writing, 27, 287–313. doi: 10.1007/s11145-013-9445-4 Hood, M., Conlon, E., & Andrews, G. (2008). Preschool home literacy practices and children’s literacy development: A longitudinal analysis. Journal of Educational Psychology, 100, 252–271. doi: 10.1037/0022-0663.100.2.252 Justice, L. M., & Ezell, H. K. (2000). Enhancing children’s print and word awareness through homebased parent intervention. American Journal of Speech-Language Pathology, 9, 257–269. doi: 10.1044/1058-0360.0903.257 Justice, L. M., Pullen, P. C., & Pence, K. (2008). Influence of verbal and nonverbal references to print on preschoolers’ visual attention to print during storybook reading. Dev Psychol, 44, 855–866.  doi: 10.1037/0012-1649.44.3.855



Home literacy practices 405

Justice, L. M., Skibbe, L. E., McGinty, A. S., Piasta, S. B., & Petrill, S. A. (2011). Feasibility, Efficacy, and Social Validity of Home-Based Storybook Reading Intervention for Children With Language Impairment. Journal of Speech, Language, and Hearing Research, 54, 523–538. doi: 10.1044/1092-4388(2010/09-0151) Kalia, V., & Reese, E. (2009). Relations between Indian children’s home literacy environment and their english oral language and literacy skills. Scientific Studies of Reading, 13, 122–145. doi: 10.1080/10888430902769517 Kim, Y. S. (2009). The relationship between home literacy practices and developmental trajectories of emergent literacy and conventional literacy skills for Korean children. Reading and Writing: An Interdisciplinary Journal, 22, 57–84.  doi: 10.1007/s11145-007-9103-9 Korat, O. (2009). The effect of maternal teaching talk on children’s emergent literacy as a function of type of activity and maternal education level. Journal of Applied Developmental Psychology, 30, 34–42.  doi: 10.1016/j.appdev.2008.10.001 Korat, O., Hassunah-Arafat, S., Aram, D., & Klein, P. (2013). Book reading mediation, SES, home literacy environment and children’s literacy: Evidence from Arabic-speaking families. First Language, 33, 132–154.  doi: 10.1177/0142723712455283 Korat, O., Klein, P., & Segal-Drori, O. (2007). Maternal mediation in book reading, home literacy environment, and children’s emergent literacy: A comparison between two social groups. Reading and Writing: An Interdisciplinary Journal, 20, 361–398. doi: 10.1007/s11145-006-9034-x Landry, S. H., Smith, K. E., Swank, P. R., & Guttentag, C. (2008). A responsive parenting intervention: The optimal timing across early childhood for impacting maternal behaviors and child outcomes. Developmental Psychology, 44, 1335–1353.  doi: 10.1037/a0013030 Levin, I., & Aram, D. (2012). Mother – child joint writing and storybook reading and their effects on kindergartners’ literacy: An intervention study. Reading and Writing: An Interdisciplinary Journal, 25, 217–249.  doi: 10.1007/s11145-010-9254-y Levy, B. A., Gong, Z., Hessels, S., Evans, M. A., & Jared, D. (2006). Understanding print: early reading development and the contributions of home literacy experiences. Journal of Experimental Child Psychology, 93, 63–93.  doi: 10.1016/j.jecp.2005.07.003 Li, H., Corrie, L. F., & Wong, B. K. M. (2008). Early teaching of Chinese literacy skills and later literacy outcomes. Early Child Development and Care, 178, 441–459.  doi: 10.1080/03004430600789365 Lin, D., McBride-Chang, C., Aram, D., & Levin, I. (2011). Mother – child joint writing in Chinese kindergarteners: Metalinguistic awareness, maternal mediation and literacy acquisition. Journal of Research in Reading, 34, 426–442.  doi: 10.1111/j.1467-9817.2010.01446.x Lin, D., McBride-Chang, C., Aram, D., Levin, I., Cheung, R. Y. M., Chow, Y. Y. Y., & Tolchinsky, L. (2009). Maternal mediation of writing in chinese children. Language and Cognitive Processes, 24, 1286–1311.  doi: 10.1080/01690960801970615 Manolitsis, G., Georgiou, G. K., & Parrila, R. (2011). Revisiting the home literacy model of reading development in an orthographically consistent language. Learning and Instruction, 21, 496–505.  doi: 10.1016/j.learninstruc.2010.06.005 Manolitsis, G., Georgiou, G. K., & Tziraki, N. (2013). Examining the effects of home literacy and numeracy environment on early reading and math acquisition. Early Childhood Research Quarterly, 28, 692–703.  doi: 10.1016/j.ecresq.2013.05.004 Marjoribanks, K. (1979). Families and their learning environments: An empirical analysis. London: Routledge & Kegan Paul. Martini, F., & Sénéchal, M. (2012). Learning literacy skills at home: Parent teaching, expectations, and child interest. Canadian Journal of Behavioural Science, 44, 210–221. doi: 10.1037/a0026758

406 Monique Sénéchal, Josée Whissell and Ashley Bildfell

Mol, S. E., Bus, A. G., De Jong, M. T., & Smeets, D. J. H. (2008). Added value of dialogic parent-child book readings: A meta-analysis. Early Education and Development, 19, 7–26. doi: 10.1080/10409280701838603 Montag, J. L., & MacDonald, M. C. (2015). Text exposure predicts spoken production of complex sentences in eight and twelve year old children and adults. Journal of Experimental Psychology: General, 144, 447–468.  doi: 10.1037/xge0000054 National Early Literacy Panel. (2008). Developing early literacy: Report of the National Early Literacy Panel. Washington, DC: National Institute for Literacy. Robins, S., Treiman, R., & Rosales, N. (2014). Letter knowledge in parent – child conversations. Reading and Writing: An Interdisciplinary Journal, 27, 407–429.  doi: 10.1007/s11145-013-9450-7 Samuelsson, S., Olson, R., Wadsworth, S., Corley, R., Defries, J. C., Willcutt, E., … Byrne, B. (2007). Genetic and environmental influences on prereading skills and early reading and spelling development in the United States, Australia, and Scandinavia. Reading and Writing: An Interdisciplinary Journal, 20, 51–75.  doi: 10.1007/s11145-006-9018-x Sénéchal, M. (2017). Shared reading: an informal literacy activity par excellence. In N. Kucirkova, C. Snow, V. Grover and C. McBride (Eds.). The Routledge International Handbook of Early Literacy Education (pp. 273–283). Oxford: Routledge. Sénéchal, M. (2014 online first). Young Children’s Home Literacy Experiences. In A. Pollatsek & Treiman, R. (Eds.). The Oxford Handbook of Reading (chap. 18). Oxford: Oxford University Press. Sénéchal, M. (2006). Testing the home literacy model: Parent involvement in kindergarten is differentially related to grade 4 reading comprehension, fluency, spelling, and reading for pleasure. Scientific Studies of Reading, 10, 59–87.  doi: 10.1207/s1532799xssr1001_4 Sénéchal, M. (1997). The differential effect of storybook reading on preschoolers’ acquisition of expressive and receptive vocabulary. Journal of Child Language, 24, 123–138. doi: 10.1017/S0305000996003005 Sénéchal, M., & LeFevre, J. A. (2002). Parental involvement in the development of children’s reading skill: A five-year longitudinal study. Child Development, 73, 445–460. doi: 10.1111/1467-8624.00417 Sénéchal, M., & LeFevre, J. A. (2014). Continuity and change in the home literacy environment as predictors of growth in vocabulary and reading. Child Development, 85, 1552–1568. doi: 10.1111/cdev.12222 Sénéchal, M., LeFevre, J. A., Hudson, E., & Lawson, E. P. (1996). Knowledge of storybooks as a predictor of young children’s vocabulary. Journal of Educational Psychology, 88, 520–536. doi: 10.1037/0022-0663.88.3.520 Sénéchal, M., LeFevre, J. A., Smith-Chant, B. L., & Colton, K. V. (2001). On refining theoretical models of emergent literacy the role of empirical evidence. Journal of School Psychology, 39, 439–460.  doi: 10.1016/S0022-4405(01)00081-4 Sénéchal, M., LeFevre, J. A., Thomas, E. M., & Daley, K. E. (1998). Differential effects of home literacy experiences on the development of oral and written language. Reading Research Quarterly, 33, 96–116.  doi: 10.1598/RRQ.33.1.5 Sénéchal, M., & Young, L. (2008). The Effect of Family Literacy Interventions on Children’s Acquisition of Reading From Kindergarten to Grade 3: A Meta-Analytic Review. Review of Educational Research, 78, 880–907.  doi: 10.3102/0034654308320319 Seymour, P. H. K. (2005). Early reading development in European orthographies. In M. J. Snowling & C. Hulme (Eds.). The science of reading: A handbook. (pp. 296–315). Malden: Blackwell Pu blishing.  doi: 10.1002/9780470757642.ch16



Home literacy practices 407

Seymour, P. H., Aro, M., & Erskine, J. M. (2003). Foundation literacy acquisition in European orthographies. British Journal of psychology, 94, 143–174.  doi: 10.1348/000712603321661859 Silinskas, G., Leppänen, U., Aunola, K., Parrila, R., & Nurmi, J.-E. (2010). Predictors of mothers’ and fathers’ teaching of reading and mathematics during kindergarten and grade 1. Learning and Instruction, 20, 61–71.  doi: 10.1016/j.learninstruc.2009.01.002 Silinskas, G., Lerkkanen, M.-K., Tolvanen, A., Niemi, P., Poikkeus, A.-M., & Nurmi, J.-E. (2012). The frequency of parents’ reading-related activities at home and children’s reading skills during kindergarten and grade 1. Journal of Applied Developmental Psychology, 33, 302– 310.  doi: 10.1016/j.appdev.2012.07.004 Silinskas, G., Parrila, R., Lerkkanen, M.-K., Poikkeus, A.-M., Niemi, P., & Nurmi, J.-E. (2010). Mothers’ reading-related activities at home and learning to read during kindergarten. European Journal of Psychology of Education, 25, 243–264.  doi: 10.1007/s10212-010-0014-9 Skwarchuk, S. L., Sowinski, C., & LeFevre, J. (2014). Formal and informal home learning activities in relation to children’s early numeracy and literacy skills: The development of a home numeracy model. Journal of Experimental Child Psychology, 121, 63–84. doi: 10.1016/j.jecp.2013.11.006 Sparks, A., & Reese, E. (2012). From reminiscing to reading: Home contributions to children’s developing language and literacy in low-income families. First Language, 33, 89–109. doi: 10.1177/0142723711433583 Stadler, M. A., & McEvoy, M. A. (2003). The effect of text genre on parent use of joint book reading strategies to promote phonological awareness. Early Childhood Research Quarterly, 18, 502–512.  doi: 10.1016/j.ecresq.2003.09.008 Stephenson, K. A., Parrila, R. K., Georgiou, G. K., & Kirby, J. R. (2008). Effects of Home Literacy, Parents’ Beliefs, and Children’s Task-Focused Behavior on Emergent Literacy and Word Reading Skills. Scientific Studies of Reading, 12, 24–50.  doi: 10.1080/10888430701746864 Strasser, K., & Lissi, M. R. (2009). Home and instruction effects on emergent literacy in a sample of Chilean kindergarten Children. Scientific Studies of Reading, 13, 175–204. doi: 10.1080/10888430902769525 Teale, W. H., & Sulzby, E. (1986). Emergent literacy: Writing and reading. Norwood, NJ: Ablex Publishing Corporation. Torppa, M., Poikkeus, A. M., Laakso, M. L., Tolvanen, A., Leskinen, E., Leppänen, P. H. T., … Lyytinen, H. (2007). Modeling the early paths of phonological awareness and factors supporting its development in children with and without familial risk of dyslexia. Scientific Studies of Reading, 11, 73–103.  doi: 10.1080/10888430709336554 Treiman, R., & Kessler, B. (2014). How children learn to write words. New York, NY: Oxford University Press. Retrieved from http://search.proquest.com/docview/1542013732?accountid= 9894  doi: 10.1093/acprof:oso/9780199907977.001.0001 van Steensel, R., McElvany, N., Kurvers, J., & Herppich, S. (2011). How effective are family literacy programs? Results of a meta-analysis. Review of Educational Research, 81, 69–96. doi: 10.3102/0034654310388819 Yeo, L. S., Ong, W. W., & Ng, C. M. (2014). The Home Literacy Environment and Preschool Children’s Reading Skills and Interest. Early Education and Development, 25, 791–814. doi: 10.1080/10409289.2014.862147

Early reading interventions The state of the practice, and some new directions in building causal theoretical models Robert Savage and Emilie Cloutier McGill University

In this chapter, we examine cognitively focused early reading interventions, their role in furnishing new insight about the process of learning to read, and specifically, their role in developing and testing theories about reading acquisition. We start with a critical analysis of what we know and do not know about effective reading interventions, and an analysis of the extent to which they are linked to theoretical models of learning to read. For this first section we overview recent research before we focus on central domains that have been intensively studied both by us and by others. These are: (i) phonics (ii) reading comprehension, fluency, and related language capabilities, the (iii) the role of educational technologies. The final section of the paper draws together a range of observations and evidence to suggest a more integrated intervention model that might enhance scientific understanding and practice.

The evidence from school-based intervention research: What do we really (confidently) know? The starting point for this section is two previously published tertiary reviews of meta-analyses of interventions by Savage (Savage 2008; Savage & Pompey, 2008) on school-based reading interventions. This work sought to draw attention to the ‘big picture’ of results of reading intervention studies, and the extent to which conclusions about what we know about early reading interventions are epistemologically legitimate claims to knowledge. We argue that the highest levels of epistemological legitimacy are achieved by reviewing evidence available, namely from systematic reviews and meta-analysis of research in a given domain. For this chapter, research on phonics, reading comprehension, fluency and educational technology are reviewed. The major conclusions of the earlier papers will be briefly discussed in each section below before more recent meta-analyses are reviewed.

doi 10.1075/swll.15.23sav © 2017 John Benjamins Publishing Company

410 Robert Savage and Emilie Cloutier

The methodology of our review of reviews In order to identify what is now known regarding early interventions in reading, five searches were undertaken using the PsychInfo database for meta-analytic review articles. The main key words used were ‘reading intervention’ and ‘reading comprehension’ while using meta-analysis and systematic review as a limit term. Several articles were found, but only seven were related to the topic and therefore selected. Another search was made using ‘meta-analysis’ as a keyword, but it did not add any study. Eric and ProQuest database searches using the same keywords identified 30 potential articles encountered of which five were selected using our selection criteria outlined below. Similar searches were undertaken on Medline and PsychArticles, but no new studies were added. A search was also undertaken on Google Scholar using the terms ‘early reading intervention ‘ and ‘meta-analysis’ and two more articles were added. Finally, using a snowball method, reference sections of the more recent reviews were scanned to identify missed articles. By doing so, five more reviews were added to make a grand total of 19 studies summarized in Table 1. Selection criteria for studies to be included were the following: (a) the publication involves early reading intervention, (b) the publication is considered a meta-analysis or a systematic review, (c) the publication has a method that allows comparison of effect sizes, (d) participants were aged from preschool to elementary school, (e) the publication was located in a peer-reviewed journal, (f) the systematic review (but not necessarily the papers within the reviews) was reported in English, and (g) the paper was published in 2008 or later. Selecting reviews for inclusion on these criteria led to important differences in the focus and nature of reviews across domains of reading explored. This follows from the fact that the inclusion criteria for studies were driven by the wider stance undertaken in the original systematic reviews. As a result, this tertiary review led to different emphases across sections below. For example, the Phonics section includes reviews that focused on studies selected on the basis of their intervention content (i.e., phonics instruction). Whereas, the Reading Comprehension and Fluency sections focus mainly on intervention effects on these target constructs as outcomes. The Educational Technology section has different foci again in that the studies are selected on distinct methodological grounds. The fact that methodologies varied in this way was an unavoidable consequence of the tertiary review approach taken here. We discuss the consequences of these variations later.

Phonics One of the main areas addressed by Savage (2008) was early reading interventions using phonics. The National Reading Panel (NRP; 2000) defined phonics as ‘a way of teaching that stresses the acquisition of letter-sound correspondences and their use in reading and spelling’ (pp. 2–89). Six systematic reviews on this topic were identified in Savage (2008): Bus and van IJzendoorn (1999); Camilli, Vargas, and Yurecko (2003);

Early reading interventions 411 Table 1. Summaries of chosen articles according to inclusion criteria Primary Authors field of research

Type of intervention

Methods of the studies Number of included studies and participants

Effective interventions

Early reading skills

Phonics

McArthur Phonics training Random allocation, et al. (2012) minimization and quasi-randomization

Piasta and Alphabet learning, letter Wagner and sound (2010) knowledge, and letter-sound correspondence

Slavin, Lake, Chambers, Cheung, and Davis (2009)

Reading curricula, instructional technology, instructional process program, and combination of curricula and instructional process

Effect sizes

Major conclusions

11 studies with Statistically significant effect sizes of – Phonics training has large effect on many 736 participants phonics training on: reading skills such as non-word reading accuracy, word reading accuracy and letter– Non-word reading accuracy: sound knowledge SDM 0.76 – Preliminary evidence suggest that phonics – Word reading accuracy: training has small effect on reading SDM 0.47 comprehension – Letter-sound knowledge: SDM – More research is needed to explore effect of 0.35 phonics training other reading skills such as reading comprehension

Quasi-experimental or experimental studies

g= 0.65 for letter sound knowledge, – Effects tend to be larger when multicomponent 63 studies with children g = 0.43 for letter name knowledge instruction was provided less than 18 – Larger effects were observed when the of age (most alphabet components taught matched the kindergarten) outcome, and when alphabetic instruction was combined with phonological awareness – No study had a longitudinal design to examine the long term impact of alphabet knowledge

Experimental or quasi-experimental with a control group, minimum of 12 weeks study, at least 15 students and 2 teachers in each treatment

63 beginning – For beginning reading studies: reading The overall sample sizestudies with weighted mean ES of 0 22 and students from 0 27 for decoding and 0 20 for kindergarten comprehension to grade 1 – For upper-elementary studies: and 79 upper Mean ES of .13, curricula ES of elementary .06, technology ES of .06, and studies with instructional process programs students from of 21 grade 2 to 9

– Successful programs at beginning and upper-

elementary levels provided professional development to teachers and included cooperative learning – Effective beginning programs all included phonics and phonemic awareness instruction – Alternative curricula and instructional technologies generate small effect sizes

(continued)

412 Robert Savage and Emilie Cloutier Table 1. (continued) Primary Authors field of research

Writing instruction and reading comprehension

Phonics and reading comprehension

Suggate (2010)

Suggate (2014)

Type of intervention

Methods of the studies Number of included studies and participants

Experimental or Intervention quasi-experimental modalities (nature of intervention and time when it was administered)

Analysis of follow-up effects for early reading interventions in phonemic awareness, phonics, fluency, reading comprehension

Experimental or quasiexperimental studies with either randomly assigned or matched control groups. All studies contained a follow-up design and a follow-up assessment

85 studies with preschool to grade 7 students

16 studies

Impact of writing Quasi-experimental and 95 studies Graham experimental designs and Hebert and writing instruction on (2011) reading

Effect sizes

Major conclusions

Over all mean effect size: d = .49 (SD = .23) Quasi-experimental studies effect size vs experimental: d = .64 vs d = .41 Mean weighted ES for phonics: d = .32 Mean weighted ES for phonological awareness: d = .18 (SD = .22)

– The nature of the intervention as well as when

– – –

Mean effect sizes at posttests vs – follow-up : d = 0.37 vs d = 0.22 ES for grade 1 and grade 2 at posttest – vs follow-up: d = 0.40 vs d = 0.26 – ES for grade 3 at posttest vs follow-up: d = 0.35 vs d = 0.43 Phonics vs phonological awareness at follow-up ES: d = 0.07 vs d = 0.29 For standardized measures of reading comprehension: for quality ES = .37 for quantity ES = .35

it was administered, when considered together significantly improved the proportion of variance explained in the mean effect size. Risk status was not a predictor of response to intervention Effect sizes were larger for quasi-experimental studies than experimental studies Phonological awareness and phonics interventions were most effective when they were administered in kindergarten and grade 1 Two key moderator variables emerged when looking at the effect sizes Intervention type and grade Effect sizes were larger at grade 3 than grade 1 and 2 There was a distinct advantage for phonological awareness vs phonics at follow-up

– There was evidence that writing about a text enhanced comprehension of it

– Higher effect for middle school than for high school

(continued)

Early reading interventions 413 Table 1. (continued)

Morphology

Primary Authors field of research

Effect sizes

Type of intervention

Methods of the studies Number of included studies and participants

Bowers, Kirby and Deacon (2010)

Morphological instruction on literacy skills

Goodwin and Ahn (2010)

Morphological instruction

Experimental and quasi- 22 studies were Morphological instruction: identified experimental studies – Morphological sublexical with control groups outcomes showed highest ES: either untrained or d = 0.65 alternative intervention – compared to alternative training: d = 0.51 – ES with less able children:Morphological sublexical: d = 0.99 – Morphological lexical: d = 0.58 – Morphological supralexical: d = 0.67 Effect of morphological instruction on: Experimental and quasi- 16 studies experimental design. – Literacy achievement: d = 0.33 Single case studies were – Phonological awareness: d = 0.49 excluded – Morphological awareness: d = 0.40 – Vocabulary: d = 0.40 – Reading comprehension: d = 0.24 – Spelling: d = 0.20 Overall effect: d = 0.32 30 studies Experimental, quasiSignificant effect size of morphology experimental and noninstruction on: experimental designs – Morphological knowledge: d = 0.44 – Phonological awareness: d = 0.48 – Vocabulary: d = 0.34 – Decoding: d = 0.59 – Spelling: d = 0.30

Goodwin and Ahn (2013)

Morphological intervention

Major conclusions

– Morphological instruction was effective at

enhancing morphological abilities of young children especially for less able children

– Alternative training was as effective – Stronger effect size for less able children

– Morphology instruction was most effective with children with disabilities

– Morphology instruction had moderate effect on many reading skills

– Morphology instruction was best when

combined with other important components

– Significant and moderate effect on many reading skills

– In contrast with other reviews, no significant effect was found on reading comprehension and fluency

– effect was greater with younger children

(continued)

414 Robert Savage and Emilie Cloutier

Table 1. (continued)

vocabulary

Primary Authors field of research

Type of intervention

Methods of the studies Number of included studies and participants

Elleman, Lindo, Morphy and Compton (2009)

Impact of vocabulary instruction on passage-level comprehension

Marulis and Neuman (2010)

The effect of vocabulary interventions with pre-K and kindergarten children on oral development

Experimental and quasiexperimental designs with either a pretest – posttest control group design, posttest control with randomization, or pretest – posttest within-subject design using counterbalanced conditions. Experimental and quasiexperimental design including a randomized controlled trial, a pre-test, posttest intervention with a control group, or a postintervention comparison between preexisting groups. Experimental and quasiexperimental designs using measures of expressive and receptive vocabulary

Mol, Bus, de Jong, and Smeets (2008)

Dialogic parent-child book reading to promote vocabulary

37 studies

Effect sizes

Major conclusions

Effect of vocabulary instruction on passage level comprehension: – with custom measure: d = 0.50 – With standardized measure: d = 0.10 (not statistically significant)

– Vocabulary instruction had a positive effect on

The overall effect size was g = 0.88, 67 studies were included SE = 0.06, CI95 = 0.76, 1.01, p < .0001 in the study, participants aging from pre-K to grade 3

comprehension, but it was very small and not statistically significant when only standardized measured were used – Because if this difference attributable to test differences, moderators were difficult to identify

– Vocabulary instruction seems to have a significant impact on vocabulary

development – Programs with explicit instruction of vocabulary had a better impact

– Implicit instruction alone was less effective

16 studies with – Moderate correlation between – When the analysis only included studies that kindergarten the intervention and compound assessed expressive vocabulary the relation students of linguistic skills of r = .29 between the intervention and compound of linguistic skills became stronger – When using only studies with expressive vocabulary measures, – There was evidence that dialogic reading did the overall effect size was: d = .59 not have a great impact with older children

(continued)

Early reading interventions 415 Table 1. (continued)

vocabulary

Primary Authors field of research Mol, Bus, and de Jong (2009)

Type of intervention

Methods of the studies Number of included studies and participants

Interactive book reading to stimulate print knowledge and oral language

Experimental and Quasi-experimental studies

Reading fluency

Strickland, Repeated reading Boon, and to increase fluency Spencer (2013) Swanson et Read-aloud al. (2011) interventions

31 studies

19 studies with student aged from kindergarten to fifth grade Treatment-comparison, 29 studies with preschool to multiple treatment, single group, or single- third grade subject research designs students Experimental or quasiexperimental, a pretest/ posttest case design, or a single-subject design

Effect sizes

Major conclusions

Effect of Interactive book reading on: Oral language skills, d = 0.54, Alphabetic knowledge, d = 0.39, phonological sensitivity, d = 0.43, orthographic awareness, d = 0.41

– About 6% of the growth of language

skills can be explain by interactive book reading in school (r = 25) and expressive vocabulary skills benefit the most from the intervention – Studies where the researchers were conducting the intervention seemed to be more effective than those were the teachers were conducting it

Effect sizes for each study was – moderate to large growth in reading fluency compared, but no mean effect sizes for practiced passages, but rather small were calculated effect sizes to unpracticed passages. – No significant differences were found when compared with continuous reading – Language: – The amount of high quality research d = 0.29, t(23) = 3.16, p = .005 regarding read-aloud interventions has increased – Phonological awareness: d = 0.78, t(21) = 5.23, p = .001 – Read-aloud interventions have positive effects on language, phonological – Print concepts: awareness, print concepts, comprehension d = 0.86, t(10) = 3,17, p = .01 and vocabulary. – Comprehension: d = 0.70, t(20) = 4.94, p = .001 – Dialogic reading had the most causal evidence – Vocabulary: d = 1.02, t(50) = 5.83, p = .001 – Only a small amount of variance was accounted for by the intervention type

(continued)

416 Robert Savage and Emilie Cloutier Table 1. (continued)

Multiple fields

Educational technologies

Primary Authors field of research

Type of intervention

Effectiveness of Archer, technologies in Savage, Sanghera- classrooms Sidhu, wood, Gottardo and Chen (2014) Features of Cheung and Slavin educational technologies (2012)

Kim and Quinn (2013)

Methods of the studies Number of included studies and participants This study sought at including systematic reviews on the topic. Therefore, reviews included meat comparable review criteria: control groups, study duration, and valid achievement measures Randomly assigned or matched control groups

Summer reading Experimental designs and non-experimental intervention designs with research based instruction on reading comprehension, fluency and decoding

Sénéchal Family literacy and Young interventions (2008)

Experimental or quasi-experimental

Four systematic reviews were examined and 38 papers were retrieved from these reviews

Effect sizes

Major conclusions

Overall effect size: ES = 0.18 When training and support were entered as moderator: ES = 0.57

– Overall effect size was positive, but small, as in previous studies

– When training and support were entered as moderators, the effect size was moderate

– 55% of the studies did not mention in their

design if they included training and support or failed to give details about it

– No significant effect of implementation fidelity

was found 84 studies and Overall effect size of educational – Educational technologies interventions technologies when compared to over 60 000 produced a positive, but small effect size traditional method: d = 0.16 participants – Computer-assisted instruction, which is very Computer-assisted instruction effect commonly used in school for the past decades, size: d = 0.11 generates a very small effect size – The more methodologically rigorous the study was the lower the effect size was Overall mean effect size : d = 0.10 – Classroom and home interventions had larger Effect of summer intervention on: effect sizes on decoding abilities – There was more positive effect sizes for – reading comprehension only : classroom using research-based instruction d = 0.23 – fluency and decoding combined – Low-income children benefit more from summer reading interventions (.28 standard : d = 0.24 deviation higher for children with low– decoding only: d = 0.43 income background than children with mixed background) 15 studies were Mean weighted effect size of 0.65 – The results of this study suggest that family included with (with a 95% CI from 0.53 to 0.76) literacy interventions can have a moderately student from large effect kindergarten to – Very few studies meet the inclusion criteria grade 3

35 studies and students aged from kindergarten to grade 12

Early reading interventions 417

NRP (2000)/Ehri et al. (2001); Hammill (2004); Torgerson, Brooks, and Hall (2006); and Troia, (1999). As all six reviews found at least some significant positive effects of phonics on literacy, Savage concluded that there was a reasonably strong basis for arguing that phonics teaching improves reading. As an example, Ehri et al. (2001) identified all studies comparing a phonics treatment program to control. A significant but relatively modest effect of d = .41 from 38 treatment-control studies was reported. In addition, several reviews noted that interventions appeared to work best when phonological awareness training was combined with training of letter-sound knowledge and reading (e.g., Bus & van IJzendoorn, 1999). However, Torgerson et al. (2006) criticized existing meta-analytic studies for including both randomized control trials (RCT) and non-RCTs studies and for including few treatment controls (as opposed to unseen ‘business as usual’ classrooms). Both of these approaches, they argued, could lead to over-estimates of intervention effect sizes. Torgerson et al. argued further that the evidence-base from RCT studies alone was still small and could justifiably be described as ‘weak’. More generally, Savage (2008) concluded there was only evidence of significant short-term effects of phonics instruction. Moreover, there were very few well-designed studies comparing different approaches to phonics (synthetic, analytic, analogy phonics etc.), and few studies systematically varying the length, time, content or titration of delivery of the intervention, or other distinctive and practically relevant features of intervention, such as the influence of teaching ‘sight words’. Finally, some systematic reviews have cautiously suggested diminishing returns for interventions of longer durations (e.g., NRP, 2000), and of small effects for field-based teacher-delivered interventions, though others have suggested that there is simply not enough high quality data available to decide the issue (e.g., Torgerson et al., 2006).

The more recent reviews on phonics We identified three recent meta-analytic reviews relevant specifically to phonics. Two additional reviews were relevant but were not centered on phonics: Piasta and Wagner (2010) explored alphabet learning specifically, and reported that it was most effective when it was part of multicomponent instruction and combined with phonics and phonological awareness instruction; and Slavin, Lake, Chambers, Cheung, and Davis (2009) sought to review ‘what worked’ much more generally in elementary school literacy interventions. They report that among other features, the most effective beginning programs all included phonics and phonemic awareness instruction. They concluded that interventions aimed at changing teaching practice more generally are the most effective interventions. Of the three recent reviews that specifically focused on phonics, the first by Suggate (2010) examined 85 experimental or quasi-experimental studies with 116 treatment – control groups (n = 7,522) from preschool to Grade 7. His review suggests that phonics interventions are only effective in the earliest years of elementary school with apparent diminishing returns for phonics programs after grade 1, and that other broader language-related interventions are more effective in the later elementary years. More

418 Robert Savage and Emilie Cloutier

generally, Suggate’s results are consistent with the view that there exist substantial minorities of ‘treatment non-responders’ lacking core meta-cognitive abilities, and who struggle to master phonics even if thoroughly taught (for systematic reviews, see Al Otaiba & Fuchs, 2002; Nelson, Benner, & Gonzalez, 2003) and imply that there is a relatively limited developmental window in which one can affect word-level reading skills. It should, however, be noted that in a later review (discussed below) Suggate (2014) nuances his comments to clarify that interventions for older struggling readers should not be neglected when they appear needed. Most clearly, for instance, if a child is deprived of phonics intervention in first grade for any reason, he would likely benefit from a phonic program in later years. In evaluating Suggate’s review, one perhaps needs to bear in mind that he only included published studies, and that the reported effects were larger for the quasi-experiments than the experimental studies. Furthermore, these effects are themselves measured over the short-term, i.e., at immediate post-test, and do not reflect the input needed to attain (and maintain) longer-term growth in literacy. A strong version of the hypothesis that there is a small developmental window for teaching reading is arguably countered by preliminary evidence that many low-literate adults can learn to read better through principled interventions (e.g., Torgerson, Porterhouse, & Brooks, 2003). Moreover, some phonics programs for older students have been proven effective when combined with other strategies (e.g., metacognition in Lovett & Steinbach, 1997) or content (e.g., orthographic connections in Blachman et al., 2014). As we argue below, these additional elements beyond phonics are probably needed to improve word reading skills of older students. In terms of the optimal duration of phonics programs, there are arguably two related issues that may cloud this issue: The orthography in which the interventions are delivered and the precise content of such programs. On the former issue, for example, there are good empirical reasons for thinking that ‘deep’ orthographies, such as English, can take up to 2.5 years to learn compared to around 1 year for shallow orthographies (Seymour et al., 2003). Arguably, this ‘orthographic depth hypothesis’ (ODH) suggests that more or additional word study or phonic content will need to be taught to English language learners compared to learners of, for example, Spanish or German, and into Grades 2 and 3. Given the complexity of English orthography and theoretical diversity of models of the orthography, how far then should educators go in teaching GPCs to beginner readers in the simplest yet most efficient manner? This issue is also relevant to data on apparent diminishing returns, which might be obtained for longer interventions because they cover the same (perhaps relatively) limited range of GPCs and phonic activities as shorter interventions, just over longer periods of time. As such, they may thus be redundant (and/or may be perceived as such by students themselves, thereby affecting motivation). Ultimately, we would argue that to fully explore Age x Effect Size interactions, researchers need to evaluate the effects of sustained ‘word-study’ interventions rather than traditional phonics interventions alone. Similarly, we will have to study the effects of early versus late wider language comprehension interventions more formally.



Early reading interventions 419

Following this line of reasoning, we have argued that some interpretations of the ODH (Seymour et al., 2003) suggest a role for teaching of at least some more ‘complex’ GPCs to children, and connecting them to their shared book reading. As it stands, children are typically taught only the 20–30 most common GPC rules in grade 1. Additionally, as there remains debate about whether older children benefit from phonic programs compared to younger children (Suggate, 2010), such interventions could be undertaken and evaluated in both Grade 1 and Grade 2 classrooms. To this end, we recently sought to explore whether students benefit from learning the extended list of ‘common complex’ GPC units suggested by a ‘Simplicity Principle’ (Vousden, Ellefson, Solity, & Chater, 2011). We thus taught 30 common GPC units, such as “a_e”,“pp”, “tch”, “igh”, “ed”. These were the maximally useful GPCs for reading real books after the 30 or so most common GPCs typically taught (namely consonant and vowel singletons and the most common vowel digraphs such as ‘oa’ ‘ee’ ‘ea’ etc.). One reading program taught children complex GPCs ordered by their frequency of occurrence in children’s texts (a ‘Simplicity Principle’). The other reading program taught children word usage, focusing on word meanings. Thirty-eight students participated in the 9-week program of 30 supplemental small group sessions. Participants in the Complex GPC group performed significantly better at post-tests with generally large value-added effect sizes (Cohen’s d) at both by-participant and by-item for Spelling, d = 1.85, d = 1.16, Word Recognition with words containing taught GPCs, d = 0.96, d = 0.95, Word Recognition, d = 0.79, d = 0.61, and Reading Motivation, d = 0.34, d = 0.56. These findings provide preliminary evidence that the Simplicity Principle aids in structuring effective supplemental phonic interventions. Children in grade 2 benefited as much as children in grade 1. This work, while a well-designed treatment-control RCT, was small-scale and is now being replicated at scale within a broad RtI model. However, we provide it here as (preliminary) evidence that ‘phonic’ programs can be highly effective beyond grade 1 if the content is appropriate and theoretically-driven to match the task demands of learning ‘deep’ orthographies. For this same reason, we predict that mixed advanced phonics-morphology interventions will be useful well beyond grade 1 given the morpho-phonological nature of English spelling. The results from theoretically-driven experimental intervention studies contrasting or titrating the teaching of complex GPCs and or common morphological structures we think will provide rich insight into how to unpack the deep orthography of English, both in pedagogical and conceptual terms. Turning to the next meta-analysis we identified, Suggate (2014) provided a provocative analysis of interventions that includes, but steps well beyond, the domain of phonics. His review is distinct in exploring ‘longer term’ effects of intervention, that is, the impact of interventions measured at delayed rather than immediate post-tests. Suggate first reports that in 85 experimental or quasi-experimental studies identified, only 21 reported any longer-term follow-up, this ‘follow-up’ was typically 11 months, and that only one study reported a follow-up longer than two years. For phonics interventions specifically, the results showed that intervention effects diminished significantly at post-test, while effects for phonological awareness, spelling, and

420 Robert Savage and Emilie Cloutier

comprehension were generally maintained. Suggate also identifies that both treatment integrity and experimental design (RCT versus non-RCT) and ‘dosage’ (specifically, the use of supplemental programs) positively affected outcomes. In evaluating this review, it should first be noted that a publication bias was suggested by Q- analyses of the complete data set. Another issue that affects interpretation of implementation was the decision to exclude studies with nested designs. As nested designs are the most appropriate way to analyze hierarchically dependent school data (e.g., Savage, Burgos, Wood, & Piquette, 2015), Suggate’s suggestion of very low teacher effects for intervention (d = .10) versus trained intervener, computer, or researcher (d = .30, .25, and .36, respectively) should perhaps be viewed with caution. More fundamentally, as the ‘longer-term’ effects reported were on average based on tests given 11 months after the immediate post-test, comments about the duration of the intervention need to be evaluated in light of the relatively short follow-up time. Our (testable) hunch is that for really long-term impacts it will be necessary to explore sustained interventions, as perhaps suggested by effects for supplemental programs reported in Suggate’s review. To this end, one study considered for Suggate’s (2014) review (Blachman, Schatschneider, Fletcher, Murray, Munger, & Vaughn, 2014) was by some distance the longest follow-up study, exploring genuinely long-term effects of reading intervention. This paper was excluded by Suggate as being untypically long in its follow up period! Blachman et al. present data from a 10-year follow up of their RCT intervention trial for struggling 2nd and 3rd grade readers. Children were originally sampled if they had low reading scores (