Introduction to This Special Issue: Dyslexia Across Languages and Writing Systems

Research suggests that a phonological deficit primarily underlies developmental dyslexia. However, the existing evidence is mainly based on studies in children learning to read English. In recent years, the research base has broadened, as research around the world has provided new insights into the neural and cognitive foundations of developmental dyslexia in different languages. The present issue aims to bring together observations on reading disabilities across languages and writing systems within cross-linguistic and cross-writing system perspectives while taking into account a broad multidisciplinary scope. An attempt is made to advance theoretical models of developmental dyslexia through systematic analyses of reading disabilities in Arabic, Chinese, Dutch, English, German, Greek, and Spanish in perspective of universal and particular underlying cognitive principles. It is also discussed how withinand cross-language observations may affect the conceptualization of developmental dyslexia. Reading involves decoding written language in order to understand it. In learning to read, children implicitly learn how their writing system encodes their spoken language and thus how they can decode printedwords into spokenwords to getmeaning (seeVerhoeven&Perfetti, 2017). However,many children around the world have problems in learning to read, failing to acquire fluent decoding, and are thus diagnosed as being dyslexic. A large body of research supports the notion that underlying developmental dyslexia is primarily a phonological deficit that takes various forms.However,much of the existing evidence and debate is based on studies in children learning to read English, which has an opaque orthography and is one of the hardest orthographies to master. In recent years, the research base has broadened, as studies around the world have provided new information on different languages. There have been suggestions that the etiology of dyslexia might differ somewhat across languages and writing systems involving not only metaphonological problems but also, especially in more transparent languages, fluency-related skills or other language and nonlanguage factors.Whether such differences reflect only superficial variation around a common (phonological) cause or deeper variation in causes related to linguistic and writing system differences is not yet clear. The time is right to bring together observations on reading problems across languages and writing systems within cross-linguistic and cross-writing system perspectives while taking into account a broad multidisciplinary scope, and this is the primary focus of this special issue. Current theoretical frameworks As noted, the most prominent models of dyslexia have tended to focus on phonological processing deficits, predicated on the findings that for the vast majority of struggling readers, a core difficulty in reading manifests itself as a deficiency within the language system, particularly at the level of phonological processing. To learn to read successfully in an alphabetic writing system, a child must CONTACT Ludo Verhoeven [email protected] Behavioural Science Institute, Radboud University Nijmegen, the Netherlands. © Published with license by Taylor & Francis Group, LLC © 2017 [Kenneth Pugh and Ludo Verhoeven] This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License ((http:// creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way. SCIENTIFIC STUDIES OF READING 2018, VOL. 22, NO. 1, 1–6 https://doi.org/10.1080/10888438.2017.1390668 D ow nl oa de d by [ R ad bo ud U ni ve rs ite it N ijm eg en ] at 0 7: 50 0 3 Ja nu ar y 20 18 develop an appreciation of the segmental nature of speech and come to realize that spoken words are composed of the smallest of these segments, the phonemes. This appreciation of the segmental nature of speech is termed phonemic awareness (PA). It is PA and the understanding that the constituents of a printed word—its letters—bear a relationship to phonemes that allow the beginning readers to connect printed words to the corresponding words in their speech lexicon. There is considerable evidence that PA is characteristically deficient (or lacking) in young readers with reading disorders who, during literacy instruction, will have difficulty developing efficient routines for mapping alphabetic characters into the phonetic constituents they represent (Bradley & Bryant, 1983; Liberman, 1992). Although the underlying mechanisms of the development of PA are not fully understood, there is some support for the claim that at least part of the difficulty resides in the phonological component of a larger specialization for spoken language. If that component is imperfect in an individual, the individual’s perception of phonemes may be less than ideally distinctive. For children with adequate phonological representations, orthographic (O), phonological (P) and semantic (S) codes can become well integrated into high-quality lexical representations (Perfetti et al., 2007) and with practice word decoding becomes more fluent or automatic (Pugh et al., 2013; Verhoeven & Van Leeuwe, 2009). There is also broad agreement that skilled word decoding in any writing system must contain at least two major processing pathways (Harm & Seidenberg, 2004): A phonological pathway by which the phonological representation of a printed word is computed via O to P mapping routines and a lexical/semantic pathway with direct O to S mapping routines. More recently, this phonological deficit framework has framed much of the work on the neurobiological basis of skilled and less skilled reading development where left hemisphere component dorsal and ventral networks that support orthographic to phonologic and/or orthographic to semantic mapping have been shown to be poorly organized in untreated dyslexia (Pugh et al., 2013; Van den Bunt et al., 2017). Although it is generally well accepted that phonological processing is vulnerable in dyslexia, four strains of thought have, in different ways, taken some issue with the canonical account. The first is the attempt to link problems of phonology to nonlanguage mechanisms such as basic sensorimotor processes (Goswami et al., 2011), or other domain general mechanisms like attention (Bosse & Valdois, 2009), or procedural learning (Nicolson & Fawcett, 2007). A second, staying within the language domain, broadens focus beyond phonology alone to include other domains such as morphological awareness that can impede the development of adequate word decoding (Kirby et al., 2012). A third multifactorial framework starts from the observation that dyslexia symptoms can be quite heterogeneous and allows that there may be many paths (or combinations of deficits) that can get to the same end state (Pennington, 2006). Finally, a fourth strain, of particular relevance to the current articles, is the notion that any of these models (from canonical to multifactorial) may be more or less relevant depending on the transparency of the writing system (Daniels and Share), and we turn to this next. Dyslexia across languages and writing systems An important question is how universals and particulars in developmental dyslexia across languages and writing systems can be explained. As noted, dual pathway accounts of word reading are generally adopted by researchers in many languages to account for the processing of visual word forms. However, specific models were generally developed with alphabetic reading in mind, although their principles should be extendable to any writing system with written words that have constituent components that can be assembled to produce word identification as well as a whole word form for direct accesses. In recent years, the research base has broadened and provided new information on different languages, and clearly the time is right to bring together observations across languages and writing systems within cross-linguistic and cross-writing system perspectives. These perspectives allow a focus on those aspects of developmental dyslexia that might be relatively specific to language properties and those aspects that are common to all languages and orthographies. 2 K. PUGH AND L. VERHOEVEN D ow nl oa de d by [ R ad bo ud U ni ve rs ite it N ijm eg en ] at 0 7: 50 0 3 Ja nu ar y 20 18 Thus, in some basic ways fluent word reading must depend both on the quality of phonological knowledge and on the capacity to transition from accurate to fluent decoding. It has been argued by some researchers that the relative contributions to dyslexia from indices of phonological knowledge such as PA or fluency such as rapid automatic naming or decoding fluency may vary across writing systems (cf. Winner et al., 2000; but also see Caravolas et al., 2013, for a contrastive view). Whether differences in the relative predictive utility of one or another cognitive test is actually indicative of fundamental differences in etiology is debatable. Indeed, as PA shares a bidirectional relation to word decoding skills (Wagner, Torgesen, & Rashotte, 1994), and as decoding skills unfold more quickly in transparent orthographies (Seymour, Aro, & Erskine, 2003), it might be that PA distributions are reduced as well and the variance accounted for therefore shits to measures with greater range. We note this here simply to illustrate that speculative can be made for or against notions of universality or language specificity and increased data on transitions to literacy (including neurobiological) will be needed to adjudicate the debates. The questions of universality versus language specificity for typical and atypical reading development become especially relevant when contrasting alphabetic languages (of any degree of transparency) with consonantal root-based writing systems such as Arabic and Hebrew and nonalphabetic morpho-syllabic writing systems such as Chinese. On first consideration, it might seem that parallels with O to P and O to S division of labor that form key elements in alphabetic systems would be unlikely given the grain size of phonological units in Chinese. However, some neurocognitive evidence suggests similarity both for typical development (Rueckl et al., 2015) and in dyslexia (Hu et al., 2010), whereas other studies appear to argue for significant neurocognitive variation along the morpho-phonological dimension (Perfetti et al., 2007; Tan, Laird, Li, & Fox, 2005). Thus, it is crucial at this juncture that we begin to acquire comparable developmental data and (informed by news tools from computational modeling and brain imaging) to identify what is universal and what is not. We see the current collection of articles as making a contribution to all of these cross-language questions.