Current Issues in Bilingualism: Cognitive and Socio-linguistic Perspectives

Current Issues in Bilingualism LITERACY STUDIES VOLUME 5 Series Editor: R. Malatesha Joshi, Texas A&M University, U...

Author: Mark Leikin | Mila Schwartz | Yishai Tobin

34 downloads 1125 Views 3MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Current Issues in Bilingualism

LITERACY STUDIES

VOLUME 5

Series Editor: R. Malatesha Joshi, Texas A&M University, USA Editorial Board: Linnea Ehri, CUNY Graduate School, USA George Hynd, Purdue University, USA Richard Olson, University of Colorado, USA Pieter Reitsma, Vrije University Amsterdam, the Netherlands Rebecca Treiman, Washington University in St. Louis, USA Usha Goswami, University of Cambridge, UK Jane Oakhill, University of Sussex, Brighton, UK Philip Seymour, University of Dundee, UK Guinevere Eden, Georgetown University Medical Center, USA Catherine McBride Chang, Chinese University of Hong Kong, China

While language defines humanity, literacy defines civilization. Understandably, illiteracy or difficulties in acquiring literacy skills have become a major concern of our technological society. A conservative estimate of the prevalence of literacy problems would put the figure at more than a billion people in the world. Because of the seriousness of the problem, research in literacy acquisition and its breakdown is pursued with enormous vigor and persistence by experts from diverse backgrounds such as cognitive psychology, neuroscience, linguistics and education. This, of course, has resulted in a plethora of data, and consequently it has become difficult to integrate this abundance of information into a coherent body because of the artificial barriers that exist among different professional specialties. The purpose of the proposed series is to bring together the available research studies into a coherent body of knowledge. Publications in this series are intended for use by educators, clinicians and research scientists in the above-mentioned specialties. Some of the titles suitable for the Series are: fMRI, brain imaging techniques and reading skills, orthography and literacy; and research based techniques for improving decoding, vocabulary, spelling, and comprehension skills.

A complete list of titles published in this series can be viewed by going to the following URL: http://www.springer.com/series/7206

Mark Leikin Mila Schwartz Yishai Tobin Editors

Current Issues in Bilingualism Cognitive and Socio-linguistic Perspectives

13

Editors Mark Leikin University of Haifa Faculty of Education Department of Learning Disabilities Haifa Israel [email protected]

Mila Schwartz Research and Evaluation Authority Oranim Academic College of Education Kiryat Tivon Israel Edmond J. Safra Brain Research Center for the Study of Learning Disabilities University of Haifa Haifa Israel [email protected]

Yishai Tobin Ben-Gurion University of the Negev Department of Foreign Literatures & Linguistics Be’er Sheva Israel [email protected]

ISBN 978-94-007-2326-9 e-ISBN 978-94-007-2327-6 DOI 10.1007/978-94-007-2327-6 Springer Dordrecht Heidelberg London New York Library of Congress Control Number: 2011938585 # Springer Science+Business Media B.V. 2012 No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed on acid-free paper Springer is part of Springer ScienceþBusiness Media (www.springer.com)

Abstract

Current Issues in Bilingualism addresses important cognitive and socio-linguistic issues related to the field of bilingualism and multilingualism. The motivation for this anthology has risen from two general phenomena: (1) the growing interest in bilingualism and multicultural societies due to rising worldwide immigration and an increasing interest in ethnic minorities and their language problems, particularly in the field of literacy and education; (2) the unique language situation in Israel, that represents one of the more complex cases of a multilingual and multicultural society, which includes the coexistence of two official languages, Hebrew and Arabic, English that is widely used in numerous contexts, and the native languages of large groups of immigrants such as Russian, Amharic and other languages. Therefore we view Israel as a mosaic of multilingualism and multiculturalism and thus a natural laboratory for studies dealing with these complex topics. This volume discusses such themes as the education of immigrant children and adolescents, the development of bilinguals with specific language and reading impairment, bi-literacy and triliteracy acquisition in the context of a complex multilingual and multicultural society, and neuro-cognitive aspects of bilingual language processing, in particular in the field of reading and reading disabilities.

v

Contents

1

Current Issues in Bilingualism: A Complex Approach to a Multidimensional Phenomenon . . . . . . . . . . . . . . . . . . . . . . . . . . . Mark Leikin, Mila Schwartz, and Yishai Tobin

Part I 2

Language and Literacy in Multilingual Society

Relevance of the Linguistic Coding Difference Hypothesis to English as an Additional Language of Literacy in Israel. . . . . . . . . . . . . . . . . . Janina Kahn-Horwitz, Richard L. Sparks, and Zahava Goldstein

3

Literacy Reflexes of Arabic Diglossia . . . . . . . . . . . . . . . . . . . . . . . . . Elinor Saiegh-Haddad

4

Multilingualism Among Israeli Arabs, and the Neuropsychology of Reading in Different Languages. . . . . . . . . . . . . . . . . . . . . . . . . . . . Zohar Eviatar and Raphiq Ibrahim

Part II

5

6

7

1

21

43

57

Academic Achievement of Children Coming from Immigrant Families

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable School Children – The Case of Ethiopian Israeli Children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Michal Shany and Esther Geva Second Generation Immigrants: A Socio-Linguistic Approach of Linguistic Development Within the Framework of Family Language Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mila Schwartz Understanding Language Achievement of Immigrants in Schools: The Role of Multiple Academic Languages . . . . . . . . . . . . . . . . . . . . . Tamar Levin and Elana Shohamy

77

119

137

vii

viii

Contents

Part III 8

9

10

Multilingual Acquisition and Processing

Adjective Inflection in Hebrew: A Psychollinguistic Study of Speakers of Russian, English and Arabic Compared with Native Hebrew Speakers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Iris Alfi-Shabtay and Dorit Ravid Verb Inflections as Indicators of Bilingual SLI: Qualitative Vs. Quantitative Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sharon Armon-Lotem, Galit Adam, Anat Blass, Jonathan Fine, Efrat Harel, Elinor Saiegh-Haddad, and Joel Walters Procedural and Declarative Memory in the Acquisition of Morphological Knowledge: A Model for Second Language Acquisition in Adults . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sara Ferman and Avi Karni

159

179

201

11

Reading in L1 and L2: Behavioral and Electrophysiological Evidence: A Comparison Between Regular and Dyslexic Readers . . . . . . . . . . . 217 Zvia Breznitz and Liat Fabian

12

Identification of Grammatical Functions in Two Languages . . . . . . . Mark Leikin and Elina Ritvas

241

Contributors

Galit Adam The Open University, Milton Keynes, England, [email protected] Iris Alfi-Shabtay Tel Aviv University, Tel Aviv, Israel, [email protected] Sharon Armon-Lotem Bar-Ilan University, Ramat-Gan, Israel, [email protected] Anat Blass Bar-Ilan University, Ramat-Gan, Israel, [email protected] Zvia Breznitz Edmond J. Safra Brain Research Center for the Study of Learning Disabilities, University of Haifa, Haifa, Israel, [email protected] Zohar Eviatar Department of Psychology, University of Haifa, Haifa, Israel, [email protected] Liat Fabian Edmond J. Safra Brain Research Center for the Study of Learning Disabilities, University of Haifa, Israel, [email protected] Sara Ferman Department of Communication Disorders, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel, [email protected] Jonathan Fine Bar-Ilan University, Ramat-Gan, Israel, [email protected] Esther Geva University of Toronto, Toronto, ON, Canada, [email protected] Zahava Goldstein Faculty of Education, University of Haifa, Haifa, Israel, [email protected] Efrat Harel Bar-Ilan University, Ramat-Gan, Israel, [email protected] Raphiq Ibrahim Department of Learning Disabilities, University of Haifa, Haifa, Israel, [email protected] Janina Kahn-Horwitz Oranim Academic College of Education, Edmond J. Safra Brain Research Center for the Study of Learning Disabilities, University of Haifa, Haifa, Israel, [email protected]

ix

x

Contributors

Avi Karni Laboratory for Functional Brain Imaging and Learning Research, Faculties of Science and Education, University of Haifa, Haifa, Israel, [email protected] Mark Leikin Laboratory for Neurocognitive Research of Giftedness, Department of Learning Disabilities, Faculty of Education, University of Haifa, Haifa, Israel, [email protected] Tamar Levin Tel Aviv University, Tel Aviv, Israel, [email protected] Dorit Ravid Tel Aviv University, Tel Aviv, Israel, [email protected] Elina Ritvas Department of Learning Disabilities, Faculty of Education, University of Haifa, Haifa, Israel, [email protected] Elinor Saiegh-Haddad Department of English Linguistics, Bar-Ilan University, Ramat-Gan, Israel, [email protected] Mila Schwartz Research and Evaluation Authority, Oranim Academic College of Education, Kiryat Tivon, Israel; Edmond J. Safra Brain Research Center for the Study of Learning Disabilities, University of Haifa, Haifa, Israel, [email protected] Michal Shany Department of Learning Disabilities, Faculty of Education, University of Haifa, Haifa, Israel, [email protected] Elana Shohamy Tel Aviv University, Tel Aviv, Israel, [email protected] Richard L. Sparks College of Mount St. Joseph, Delhi Township, OH, USA, [email protected] Yishai Tobin Department of Foreign Literatures & Linguistics, Ben-Gurion University, Be’er Sheva, Israel, [email protected] Joel Walters Bar-Ilan University, Ramat-Gan, Israel, [email protected]

Chapter 1

Current Issues in Bilingualism: A Complex Approach to a Multidimensional Phenomenon Mark Leikin, Mila Schwartz, and Yishai Tobin

1.1 Introduction The past 20 years have seen an unprecedented upsurge of interest in bilingualism and multilingualism. A major reason for this is no doubt the acknowledgement by a growing number of researchers that the use of two or more languages is far more common than was previously thought, and may perhaps even be the norm. There are no exact data on the number of bilinguals throughout the world.1 However, some researchers claim that over 50% of the world’s population is bilingual (Fabbro, 1999). Bilingualism prevails at the societal level as well. With an estimated 6000 languages at this time being spoken in the world (Grimes, 2000), and with only some 200 countries, simple mathematics demonstrates that many countries must, in one way or another, be bi- or multilingual (Dewaele, Housen, & Wei, 2003). An investigation of the phenomenon of bilingualism and second language acquisition, therefore, has not only great theoretical significance, but will also be of great practical importance. In this context, we considered that it would be both interesting and useful to present a volume containing a number of experimental studies focusing on some major aspects of bilingualism and multilingualism from diverse theoretical and methodological points of view which have been carried out recently in Israel. The motivation for this anthology has risen from two general reasons: (1) The growing interest in multicultural societies due to rising worldwide immigration and an increase in attention given to ethnic minorities and their language problems, particularly in the field of literacy and education from cognitive and socio-linguistic points of view; 1 Hereinafter the term ‘‘bilingual’’ is used in the wider sense of the word: people who know more than one language.

M. Leikin (*) Laboratory for Neurocognitive Research of Giftedness, Department of Learning Disabilities, Faculty of Education, University of Haifa, Haifa, Israel e-mail: [email protected] M. Leikin et al. (eds.), Current Issues in Bilingualism, Literacy Studies 5, DOI 10.1007/978-94-007-2327-6_1, Ó Springer ScienceþBusiness Media B.V. 2012

1

2

M. Leikin et al.

(2) The unique language situation in Israel, which represents one of the more complex cases of a multilingual and multicultural society, including the coexistence of two official languages, Hebrew and Arabic (the language of the largest ethnic minority in Israel), English as semi-official language that is widely used in numerous contexts (especially in ‘‘academia’’, economics and politics), and the native languages of large groups of immigrants (e.g., Russian, Amharic, Spanish, French and scores of others). Note, however, that the above described linguistic and cultural situation is not only interesting in itself but also promotes scientific study in the multilingual domain and permits the generalization and possible projection of data to other countries and societies.

1.2 Bilingualism and Multilingualism: Research Directions The growing interest in bilingualism and multilingualism is rooted in globalization and increased immigration on the one hand, and in the strong needs of ethnic minorities to maintain their language and culture on the other. Both of these tendencies are characteristic of Israel as a country, which provides natural conditions for diverse and complex language interactions. These interactions are created by the meeting of typologically different languages and scripts, and such dyads and triads as, for example, Hebrew-Arabic and Hebrew-ArabicEnglish or Russian-Hebrew and Russian-Hebrew-English become more common and thus create a multilingual and multicultural mosaic which typifies Israel today. Current research on multilingualism highlights it as a phenomenon whose nature is to be investigated on its own terms apart from bilingualism (Aronin & Hufeisen, 2010). Recently both the awareness of multilingualism and research in this area have become increasingly relevant and useful in light of above noted tendencies of globalization and increasing immigration rates. A significant amount of research that probes more deeply into various aspects of contemporary multilingualism has appeared along with the continuation of intensive study of different aspects of bilingualism (Aronin & Hufeisen, 2010; Bhatia & Ritchie, 2006; Haznedar & Gavruseva, 2008; Kroll & de Groot, 2005). Accordingly, this volume also addresses not only practical topics in the field of bilingualism but additionally deals with such important issues of multilingualism as, language typology, and multilingual education, with its main focus on multilingual and multicultural societies such as Israel. In this context, Israel may be considered as an example of a natural laboratory of multilingualism. In this anthology we have aimed to address some of the most frequently discussed contemporary questions relating to the nature of bilingualism and multilingualism in the world by focusing on language and literacy in the multilingual mosaic of Israel as a research laboratory. It is important to note that even a cursory examination of these issues reveals the complexity and ambiguity

1

Current Issues in Bilingualism: A Complex Approach to a Multidimensional . . .

3

of the phenomenon of bi- and multilingualism. Not surprisingly, bilingualism (as well as multilingualism) appears as the center of interest of various scientific disciplines including linguistics, psycho- and neurolinguistics, sociolinguistics, cognitive psychology, and education. This anthology therefore presents a variety of language pairs, research approaches and scientific fields. However, before we introduce the studies included in the volume, we would like to briefly characterize the linguistic situation in Israel as a multilingual mosaic.

1.3 The Multilingual Mosaic: The Case of Israel In this section we will relate to the unique language situation in Israel historically and from the perspective of language policy. Our focus will be on Arabic as the language of the largest ethnic minority in Israel, English as a semi-official language (mostly second or even third language) and the native languages of the largest groups of new immigrants (Russian and Amharic).

1.3.1 The Arabic Language As previously mentioned, Arabic (together with Hebrew) is one of the two official national languages in Israel. Arabic is used intensively in diverse social and practical contexts (including academic and civil contexts). The native Palestinian Arabic-speaking population of Israel uses Standard Arabic in its schools. As in other Arabic-speaking countries, the ‘‘purity’’ of the Standard Arabic language was preserved by the rigorous separation of the written and spoken varieties, a phenomenon that Ferguson (1959) labeled diglossia. Ferguson (1959) proposed the classical definition of diglossia as being a stable linguistic state that includes different spoken local dialects (ammia) and a very dissimilar literary standard language [version] (fus: h: a), which is usually more grammatically complex, distinct from the different spoken dialects, and includes a respectable written tradition and standard. The literary language is the language officially taught and studied in school and is usually not acquired naturally but formally. Therefore, Standard Arabic (i.e., the standard literary language) is first encountered in school and can be viewed almost as a second language (Ibrahim & Eviatar, 2007). This diglossic situation of the Arabic language hinders children’s acquisition of literacy in Arabic (Ayari, 1996; Saiegh-Haddad, 2003).

1.3.2 The Languages of Immigrants Historically, Israel may be viewed as a classical case of a country of migration that has absorbed a large immigrant population of multiple and diverse origins. However, until recently, the dominant language policy in Israel was to strongly

4

M. Leikin et al.

and exclusively support only one language: Hebrew (the revived national language of the Jewish people), and there was a purposeful and tendentious lack of interest displayed for the languages spoken by the Jewish immigrants (except, perhaps, for English, because of its international character). As a result, the active use of languages such as German, Polish, Yiddish, Spanish and French was discouraged and ignored, and they were basically overshadowed and/or lost. However, in the last two decades there have been significant changes in language policy and in the attitudes to languages other than Hebrew (Spolsky & Shohamy, 1999). There are two major groups of immigrants that have clearly influenced these changes in the linguistic policy in Israel. The first is the massive arrival of immigrants from the Former Soviet Union (FSU) to Israel (over 835,000 people between the years 1990 and 1999). A comparison of this intensive wave of migration with other migrations to a single country in such a short period as the decade of the 1990s shows that in absolute numbers, it was among the largest; and relative to the size of the receiving population, it far exceeded all other countries that absorbed migrant populations. This mass migration has often been compared to the proportionally equivalent immigration of the entire population of France to the United States. These immigrants from the FSU are currently the largest ethnic immigrant group in Israel, comprising approximately one million Russian-speakers. The influx of immigrants from the FSU in the 1990s occurred against the background of a general decline in the original Hebrew monolingual nationbuilding ideology (Glinert, 1995a), as well as a national identity crisis of the ‘‘post-Zionist’’ Israeli society (e.g., Ben-Rafael, 1994). This wave of massive immigration was one of the major factors that encouraged Israeli society towards multilingualism (Spolsky & Shohamy, 1999); challenged the ‘‘melting pot’’ policy towards ethnic minority groups as well as the dominance of Hebrew as the realization of monolingual ideology in Israel (Leshem & Lissak, 1999; Nudelman, 2000; Kotik-Friedgut, 2000; Shuval, 1999). The large-scale character of this immigration group may explain the avowed policy of this community to retain its language of origin: Russian. This strong commitment to the heritage language can be regarded as an application of Fishmann’s Reversing Language Shift theory (RLS) (1991, 2001), which seeks to serve as a directive for worldwide attempts to support minority languages (such as Russian in Israel). The Russian language has attained the status of one of Israel’s main languages (Olstein, 1995), after Hebrew (the official language) and English. A state-sponsored Russian-language radio network and TV channel, and the publication of some 50 Russian-language newspapers and magazines attest to its importance. Another factor strongly associated with Fishmann’s RLS theory (1991, 2001) is the Russian Jewish (RJ) immigrants’ tendency to appreciate their original culture, which encouraged them to maintain their language of origin and promote its acquisition by their children, including those who were born in Israel (Ben-Rafael, Olshtain, & Geijst, 1997; Donitsa-Schmidt, 1999; Leshem & Lissak, 1999). The assertion of RJ immigrants’ original cultural identity

1

Current Issues in Bilingualism: A Complex Approach to a Multidimensional . . .

5

underlies one of the stated aims of language policy at the community level: the promotion of a supplementary private Russian-language education system (non-formal education) designed to preserve the cultural heritage and the mother tongue in its standard written form among Israeli-born RussianHebrew speaking (RHS) children. The second new large immigrant group are the more than 75,000 immigrants from Ethiopia, mainly speakers of Amharic and Tigrinya, with some literacy in Amharic and Ge’ez. The background of these immigrants is from the Ethiopian rural community, and their culture is predominantly oral. The literacy rate in this community is no more than 40% (Spolsky & Shohamy, 1999). Most Ethiopian immigrants are unskilled workers, and the unemployment rate among this community is higher than in the general population (Svirsky & Svirsky, 2002). It should be further noted that the status of Amharic in Israel is low (Spolsky & Shohamy, 1999). In contrast to the Russian-speaking immigrant community, immigrants from Ethiopia relinquished the maintenance of their first language, its culture and traditions in favor of a second language unrelated to their native ethnic and cultural values. As a result, the second-generation Ethiopian immigrants actually grow up as Hebrew-speaking monolinguals with a rather limited link with a heritage language, traditions and culture. Concerning the mastering of Hebrew (L2), recent reports indicate that about 60% of more recent immigrants from Ethiopia claim that they have only a minimal command of Hebrew. In addition, 55% of the male adults cannot read or write Hebrew, and this percentage is even higher among women (70%) (King, Effrati, & Netzer, 2003). Even so, these two large waves of immigration together with the wide spread of English not only caused significant changes in Israeli society from the viewpoint of its relation to bilingualism, but also presented two radically diverse as well as equally interesting cases of bilingualism.

1.3.3 The Place of English in Israel English has spread throughout the word inevitably capturing the status of the primary word language. Thus, Israel has not been exempt from this massive use of English as a universal lingua franca. English, which does not have the status of an official language of the country (as do Hebrew and Arabic), is still widely used in higher education and in many other fields of life and generally is taught early in elementary school. Historically, the role of English grew after the conquest of Palestine by British troops under General Allenby and the subsequent award of a Mandate for Palestine to the British government (Spolsky & Cooper, 1991). As well as serving as a language for access to business, science, education, and travel, English is a language of major Jewish Diasporas in the US and elsewhere and the language of the large number of English-speaking immigrants (Spolsky,

6

M. Leikin et al.

1996). De facto English has become the second language of academic life as well. For that reason, the importance ascribed to English in schools is in a large measure a reflection of the salience of the fact that all Israeli universities require entering students to have a minimum level of competence in the language (Spolsky & Shohamy, 1999) and maintain an obligatory series of English as a Foreign language (EFL) courses to an advanced level to ensure the students’ ability to read academic texts in English. It is important to note also that the demand for earlier English teaching in the elementary schools has continued to increase and resulted in exposure to English in the second and even in the first grade. Thus, national educational policy supports English as the major language of wider communication for Israelis, and it is taught as a major subject in both elementary and secondary schools. Thus, we hope that the picture of the unique ‘‘language mosaic’’ presented above supports our claim that Israel may be viewed as something like a ‘‘natural laboratory’’ of multilingualism and therefore provides a fitting example for discussing contemporary issues in bilingualism and multilingualism. Accordingly, the aim of our volume is not to review the unique language situation in Israel, but to present experimental research on major aspects of bi- and multilingualism from diverse theoretical and methodological points of view. For that reason, we have chosen not to group the chapters of this volume according to specific languages or language communities but rather by their research approaches to the aforementioned issues in different scientific fields. Therefore this anthology is divided into three sections: Part I deals with subjects related to language and literacy in a multilingual society; Part II addresses issues of the academic performance of children coming from immigrant families, and Part III focuses on multilingual acquisition and processing.

1.4 Part I: Language and Literacy in Multilingual Society There is no doubt that bilingualism presents a highly attractive topic for linguists and psycholinguists who study language acquisition and use (e.g., De Bot & Kroll, 2002; Grosjean, 1998). Recent psycholinguistic research in the bilingual and multilingual domain has focused on a wide range of topics including the following which are within the scope of this volume: (1) crosslinguistic transfer and L2 literacy acquisition and processing; (2) language typology and literacy acquisition. Psycholinguistic research in the last two decades has dealt with the potential educational consequences of bilingualism, and, particularly, has attended to literacy acquisition in L1 and L2 (for a detailed review, see August & Shanahan, 2006). There are currently two very broad areas of research in the domain of literacy acquisition. The first concerns the question of whether the process of acquiring literacy skills is different for bilingual children as compared to monolinguals specifically because they are bilingual: i.e. is there

1

Current Issues in Bilingualism: A Complex Approach to a Multidimensional . . .

7

an effect on L2 literacy acquisition that comes from having oral proficiency in L1 (L1 linguistic background) and/or literacy in L1 (L1 orthographic background) among typical readers (see, for example, Deacon & Cain, 2011). Accordingly, the second direction of the above-mentioned research is linked closely with the first and tries to determine the factors which could account for the difficulty in literacy acquisition in L2, such as specific language impairment and specific reading disability in L1 (e.g., Sparks & Ganschow, 1993b). Therefore, L2 learning difficulties are likely to be based in L1 learning and that faculty with one’s language ‘‘code’’ (e.g., linguistic, orthographic) is likely to play an important causal role in learning L2. In addition, both these research domains address the factor of language typology, more specifically linguistic and orthographic proximity between the languages within the dyads or triads in the process of literacy acquisition (Kahn-Horwitz, Schwartz, & Share, 2011). Evidence for the cross-linguistic transfer of phonological awareness and word identification skills from Hebrew (L1) to English as an additional language was supported by the longitudinal study of Kahn-Horwitz, Sparks and Goldstein (Chapter 2). They focus on the extent to which first language (Hebrew) and additional language (English) linguistic and literary abilities measured 6 years earlier (at 4th Grade) could predict word recognition and reading comprehension in English (in the 9th Grade). This paper broadens and enriches our understanding of the cross-linguistic transfer of phonological and word recognition skills by focusing on both opaque and deep orthographies, using Hebrew unpointed script – a consonant-based syllabary – and the English alphabet, each with different degrees and characteristics of opacity. It was found that underlying L1 abilities still predict English word accuracy 6 years after the beginning of English acquisition. Moreover, the knowledge of Hebrew vocabulary together with Hebrew spelling ability measured in fourth grade accounted for a significant amount of variance of English reading comprehension bearing evidence for a simple model of reading (Koda, 2005). The data were discussed with respect to one of the main hypotheses concerning the interaction between literacy development in L1 and L2, namely, the Linguistic Coding Differences Hypothesis (Sparks & Ganschow, 1993a). The authors underscore the educational implications from their findings raising critical issues concerning the education policy used in teaching English in Israel. The focus on linguistic proximity between the literary and spoken forms of the Arabic language is presented in Haddad’s article (Chapter 3) on diglossia in the Arabic language and its link to the acquisition of basic reading skills. This paper examines the classical issue of diglossia dealing not only with the definition of this complex concept but also with the considerable differences between Standard Arabic and the spoken vernacular that significantly challenge reading acquisition. Accordingly, Haddad’s contribution presents a detailed comparative description of the literary and spoken forms of the Arabic language by stressing the great discrepancies between the two in almost all

8

M. Leikin et al.

language domains: lexical, morphological, syntactic, and phonotactic. Then, the author reviews the research focusing on the effect of the phonological distance between Standard Arabic and the spoken dialect on the acquisition of the basic reading process. In one of the studies reported by the researcher, kindergarten and first grade pupils were tested for their performance on phonemic awareness tasks that manipulated the status of syllabic structure (Standard Arabic CVCC versus spoken vernacular CVCVC) of pseudo-word stimuli. The results show that for both groups of children, the isolation of both the initial and final consonants in the Standard form of Arabic was harder than in the spoken form. The author concludes that novel Standard Arabic linguistic structures that are not available in the children’s spoken language pose a particular challenge in the acquisition of phonological processing and in word decoding for standard language phonological structures. This finding was defined as the linguistic affiliation constraint. In terms of the specific socio-cultural context in which the acquisition of literacy in Arabic is embedded (the diglossic situation), Haddad maintains, ‘‘the complex linguistic reality results in poor language skills in the written code and, in turn, in linguistic insecurity.’’ Eviatar and Ibrahim (Chapter 4) present a scientific review of their recent studies of bilingual reading processing in the context of the theoretical interpretation of the data found in the professional literature. It focuses on the linguistic features of specific languages and their influence on reading acquisition and reading processing in different languages from the neuropsychological point of view. Eviatar and Ibrahim’s study summarizes findings (many of which have been obtained by the authors themselves) with regard to Arabic/Hebrew bilingualism. In particular, the authors concentrate on two major topics. The first examines the complex linguistic situation in the Arabic-speaking population of Israel and has implications for reading acquisition. Israeli Arabs present a complicated case of bilingualism since in addition to the necessity of learning the language of the majority (i.e., Hebrew) and English, they encounter the specific problem of diglossia in the process of literacy acquisition. In this context, the authors propose a comprehensive review of the phenomenon of diglossia, which is a phenomenon similar to bilingualism in spite of the fact that in this case we do not deal with two different languages but with spoken and written forms of the same language (Ayari, 1996, see also above for details). The second topic focuses on the functional architecture of reading in Hebrew (as L1 and L2) and in Arabic (as L1) from the neuropsychological perspective. The review underscores the orthographic and morphological properties of Semitic languages influencing reading acquisition and processing in both target languages. For example, Eviatar and Ibrahim discuss the problems that are caused by the fact that in Arabic and Hebrew orthography, vowels are not a part of the orthography and are not presented in script and appear only in poetry, children’s books and liturgical texts as diacritical marks above, below or within the body of the word.

1

Current Issues in Bilingualism: A Complex Approach to a Multidimensional . . .

9

1.5 Part II: Academic Achievement of Children Coming from Immigrant Families Two of the most discussed educational issues during the past 50 years in the countries receiving immigration have concerned (1) the education of immigrant children and (2) the most effective ways of teaching literacy in L2. Recent research points to a necessity to address these concerns from a socio-cultural perspective. More specifically, it is important to address the heterogeneity of immigrant children with respect to their socio-cultural background, language dominance and preference, the sequence of exposure to both languages, and the type of educational system in which they are enrolled. In a recent meta-analysis, Lesaux and Geva (2006) conclude that a number of variables at the contextual level (e.g., parental education, home literacy, demographics) affect immigrant children and the children of immigrants’ academic progress in such areas as reading comprehension. Parent-related factors (parental education, level of L2 mastery, home literacy, the role that parents undertake in the ‘‘education enterprise’’ of their children, the cultural capital resources, the parents’ ability to support academic progress) have also been consistently shown to be related to these children’s academic success (Lesaux, Koda, Siegel, & Shanahan, 2006). In this section, we extend our understanding of issues related to academic achievements of children coming from immigrant families by stressing the complexity of the sociolinguistic context regarding the acquisition of literacy, lexical knowledge and the proficiency in multiple academic languages. Shany and Geva (Chapter 5) present a thorough analysis of language development and the acquisition of literacy skills in Hebrew among second-generation immigrants from Ethiopia in Israel. The chapter addresses critical questions concerning specific components of language and literacy skills from a developmental perspective (in Grades 1, 2, 4, and 6). The authors provide a comprehensive overview of the main hypotheses concerning the acquisition of basic literacy and reading comprehension skills as a theoretical framework for their study as well as an in-depth description of the socio-cultural context in which Ethiopian Israeli children progress in their schooling. Their approach underscores the fact that in general the Ethiopian community in Israel does not possess the cultural capital resources that can be transmitted to their children’s education within Westernized culture, e.g., traditions of home literacy and learning strategic skills. In this case, the authors argue that the school policy makers and teachers are supposed to adapt their instruction to relevant aspects of the children’s home culture. The comparison between low socio-economic status (SES) Ethiopian and Non-Ethiopian groups on cultural, cognitive, linguistic and literacy dimensions revealed that while in underlying reading skills (e.g., phonemic awareness, word identification) no gaps have been found, the gaps persisted across grades on non-verbal ability, cultural knowledge, and grammatical skills. These findings are discussed in terms of their relevance to other groups migrating from non-literate to ‘‘developed’’ societies and their educational implications.

10

M. Leikin et al.

In her contribution, Schwartz (Chapter 6) claims the necessity to take account of the socio-cultural background of second-generation immigrants while coming to conclusions concerning their developmental trends in vocabulary acquisition in L1 and L2. To support this claim, sound empirical evidence from recent studies conducted in Canada, the Netherlands, the United States and Israel is presented. Regarding the data concerning the socio-cultural profile of immigrant families, Schwartz argues that the traditional description of parental background is limited to reports on income and occupation, without reference to education as a separate factor. Accordingly, it may be misleading to relate only to SES of immigrant families in approaching the educational challenges of second-generation children. This is due to the possible discrepancy between the relatively high socio-cultural level obtained in the country of origin and the relatively low SES characterizing the first years of adaptation in the host country. Regarding L1 development in the context of immigration, this chapter attests to the crucial role of teaching literate L1 in both family and informal educational settings. In addition, this chapter underscores the importance of the children’s positive approach toward home language acquisition. It can be concluded, therefore, that Schwartz’s study moves beyond the traditional approach in sociolinguistic research by extending our knowledge to a range of sociolinguistic factors related to the second-generation immigrants’ progress in L1 and L2. If the two previous chapters (Chapters 5 and 6) concentrate on second generation (Israeli-born elementary school children) of immigrants from Ethiopia and the FSU, the contribution of Levin and Shohamy (Chapter 7) focuses on immigrant children from the same socio-cultural backgrounds previously mentioned but with different periods of residence in Israel and at a more advanced age. Levin and Shohamy explore the link between academic language and the educational achievements of the immigrant students from the FSU and Ethiopia as compared to native-Hebrew speakers in three grade levels (Grades 5, 9, and 11). The authors raise major unresolved questions concerning the education of students with diverse cultural backgrounds within a nonadditive Hebrew-speaking context such as: how do these students progress academically in different subject domains, which domains are characterized not only by specific concepts but also by particular language functions? The authors argue that learning a school subject requires the immigrant children to acquire a new language in addition to Hebrew (L2), because each subject area has a specific register, a specialized lexicon and a diverse set of discourse that are culturally related. Having based their claim, the authors present an in-depth analysis of the multiple meanings of academic language and explore the notion of content domain competence. Their findings point out a significant variability between the students born in Israel and in the FSU and Ethiopia by specific content domains focusing primarily on mathematics. The data also support the link between the socio-cultural dimension of the language of mathematics and the relatively low achievements of the students who have had no Western-based experiences in these domains. However, the results show that even students who

1

Current Issues in Bilingualism: A Complex Approach to a Multidimensional . . .

11

immigrated from the FSU, with a rich tradition in mathematics and science instruction, face difficulties in their exposure to the different dimensions of the new academic language. The data found in response to these basic questions suggest that learning a content domain cannot be considered as a scaffolding strategy for L2 acquisition by immigrant students. This paper provides a significant contribution to the examination of the educational achievements of different ethnic immigrant groups within the non-additive context by adopting a quantitative approach to evaluate students’ performance in diverse academic languages.

1.6 Part III: Multilingual Acquisition and Processing Similar to the psycholinguistic research on language and literacy presented above, in this [chapter]/section we address the question of how language typology may affect second language acquisition specifically with regard to linguistic interference in the case of typologically different languages. What happens when the morphological structures of two typologically different languages are in contact is the topic of Alfi-Shabtay and Ravid’s study (Chapter 8). The authors focus on the four most frequently used languages of Israel (Hebrew, Russian, Arabic and English) distinguishing them in terms of morphological typology and complexity. While Hebrew, Russian and Arabic have a rich synthetic morphology, in which words typically contain more than one morpheme, analytical languages such as English contain a more restricted or limited morphology. [Accordingly]/Therefore, this chapter examines the way the nonnative Hebrew adult speakers with different language backgrounds (i.e., Russian, Arabic and English) process two types of adjectives (attributive vs. predicative) as compared to native speakers of Hebrew. Is the participants’ ability to identify grammatical and ungrammatical attributive and predicative adjectives linked to their language background? It is also of interest whether the participants’ performance on a judgment task, which tested their ability to identify grammatical and ungrammatical attributive and predicative adjectives, may vary in encountering the irregular categories (i.e., in cases of conflict between noun phonology and gender). Stressing similarities and differences between the above-mentioned languages in terms of adjective inflection, this paper contributes to our understanding of the psycholinguistic processes employed by adult bilinguals facing the complexity of Hebrew inflectional morphology. The participants’ achievements in the adjective inflection of L2 Hebrew were interpreted according to the following three forces that affect the performance on adjective inflection: (1) first language constraints; (2) second language constraints; and (3) constraints of processing and memory. In addition, this chapter focuses on two relatively novel research domains in multilingual acquisition and processing: (1) linguistic behavior among bilingual children with impaired language development; (2) neurocognitive research

12

M. Leikin et al.

on multilingual acquisition and processing. Concerning atypical language acquisition, research in the interface of bilingual development and specific language impairment (SLI) has increased greatly in the past decade and raised a number of unresolved questions concerning the linguistic behavior of this population (Genesee, Paradis, & Crago, 2004; Paradis, 2007). For example, some recent psycholinguistic studies have begun to focus on the problem of whether dual language acquisition becomes an additional complication for children with impaired language development in their L1. In addition, this research has addressed the extent of difference in the trajectory and pace of language development among bilingual children with and without language impairment. An important constituent of this noted tendency includes discussion regarding the assessment and treatment of young bilinguals with SLI compared to monolingual bilinguals (see Paradis, 2007). The authors of Chapter 9 (Armon-Lotem, Adam, Blass, Fine, Harel, SaieghHaddad, and Walters) compare English-Hebrew early sequential bilinguals with specific language impairment (SLI) attending language pre-schools to typically developing bilinguals from regular pre-schools on the use of the verbal system in general and the verb tense system in particular in both target languages. In addition, the findings concerning the SLI bilinguals’ performance is compared with existing data on SLI monolingual acquisition in both languages, as well as studies of typical bilingual development, in order to evaluate the nature of the errors in usage of tense morphemes. As for comparison between typical and atypical SLI bilinguals language development, the data point out a qualitative similarity of the usage of inflections and the type of errors with difference in the quantity of errors. Moreover, the patterns of difficulty acknowledged for monolingual SLI children, either in English or in Hebrew, were also found in the SLI bilinguals, although none were equally manifested in both languages. The authors therefore argue that early childhood bilingualism per se could not be considered an inappropriate developmental choice for children with SLI by disproving a widespread belief that these children’s limited capacity for language would be overtaxed by learning two linguistic systems. Thus it might be suggested that children affected with a language-learning disability can be raised bilingually without any serious detriment to their grammatical development. A somewhat different approach to the issue of L2 acquisition can be seen in Karni and Ferman’s study (Chapter 10) which focuses on internal cognitive mechanisms which are involved in the process of second language learning among adults. There is a widespread notion that adults have reduced and less effective language learning ability than children. At the same time, there is ample evidence suggesting that the acquisition and retention (long-term memory) of procedural skills as well as declarative memory (facts, singular events) of non-linguistic origin are robust in healthy adults. The authors examine these issues by the intensive, multi-session training of adults on an artificial morphological rule, and examined whether key characteristics of non-linguistic skills are recognizable in linguistic rule learning with regard to L2 learning. The

1

Current Issues in Bilingualism: A Complex Approach to a Multidimensional . . .

13

results showed that adults were highly competent in acquiring and retaining linguistic knowledge, with both memory systems contributing differentially to the learning of distinct aspects of the morphological rule at different stages along the mastering of skilled linguistic performance. Note that this finding agrees with a growing body of evidence suggesting that adults are as competent as, and often better than, children in acquiring and retaining non-linguistic skills. Thus, the results of Karni and Ferman’s study demonstrate the high potential of adult learners in L2 acquisition and accordingly raise a question as to what prevents adults from realizing this potential practically. Thus one may conclude that the cognitive approach to research of bilingualism appears to be very promising because it allows for investigating the basic mechanisms of underlying processes of L2 acquisition and bilingual language use. In this context, it should be noted that in the last two decades there has been an increase of interest in the neuropsychological and neurophysiological aspects of bilingualism and second language learning (Fabbro, 1999; Li & Green, 2007; Paradis, 2004). This interest in the neurocognitive aspects of bilingualism has been inspired by an expanding effort in the cognitive study of bilingualism (e.g., de Groot & Kroll, 1997), on the one hand, and a dramatic growth of language processing research using neuroimaging methods, on the other hand (e.g., Fabbro, 1999, Proverbio, Cok, & Zani, 2002). Historically, this research domain developed from neuropsychological studies of bilingual language impairments which focused on the representation and processing of multiple languages in the brain (Albert & Obler, 1978; Fabbro, 1999; Paradis, 2004; Perani & Abutalebi, 2005). Therefore, the neurocognitive study of bilingualism also mostly concerns cerebral lateralization and localization (Hull & Vaid, 2005; Paradis, 2004; Zhang & Wang, 2007). A number of previous studies have suggested distinct non-overlapping cortical representations of the two languages in bilinguals (Paradis & Goldblum, 1989; Paradis, 2004). In addition, different ERP patterns have been observed in first language processing of bilinguals and monolinguals (Donald, Meuter, & Ardal, 1986). However, a number of other studies have found evidence for overlapping cortical representations in bilinguals (e.g., Sarfarazi & Sedgwick, 1996; Paradis, 2004; Ullman, 2001). Empirical findings (fMRI and other neuroimaging tools) of the locations of activation sources for the two languages have also been contradictory (Briellman et al., 2004; Illes et al., 1999; Wartenburger et al., 2003). Behavioral studies of bilingual language processing have yielded mixed results as well: some have suggested a separate activation of L1 and L2 (Gerard & Scarborough, 1989) and others indicate the simultaneous activation of the two languages (e.g., Dijkstra, 2001). Thus we can see that neurolinguistic studies related to bilingualism have yielded contradictory results resulting in controversy. These controversies have been explained by differences in methodologies and/or subject populations (e.g., Grosjean, 1998). There have been further suggestions to distinguish between different bilingual processing styles for phonetic and syntactic

14

M. Leikin et al.

information on the one hand, and for lexical-semantic information on the other (e.g., Marian, Spivey, & Hirsch, 2003). Two additional factors may be proposed to account for the above-mentioned contradictions: structural differences between the first and second languages and the variety of bilingualism types (e.g., early vs. late bilingualism). These two factors appear to be very interesting and promising areas for further investigation. Therefore, it is not surprising that in the last decade these two issues have played a key role in the research of bilingualism. For example, there have been a number of studies focusing on different types of bilingual speakers (early versus late bilingualism) and the consequences of these differences for cerebral representation: i.e., how the extent of the reliance on each of the above mechanisms among early vs. late bilinguals may affect language processing in each language (e.g., Birdsong, 2006; Fabbro, 1999). Another issue that has been intensively studied relates to the linguistic properties of different languages and their influence on bilingual language representation and processing (e.g., Fabbro, 1999; Paradis, 2004). Such growing interest was caused not only by developments in linguistics and psycholinguistics but also by the geography of neurolinguistic studies that have dramatically changed in the two past decades. In particular, the studies of bilingualism in Israel that have focused on two Semitic (i.e., Hebrew and Arabic) and a wide range of Indo-European languages (e.g., English and Russian) have provided very interesting data (see for details Breznitz & Lebovitz, 2008; Leikin, 2008) on the relationships between different types of bilingualism and the linguistic properties of different languages. The last two chapters in our volume also represent this direction of research. Reading behavior and brain activity of regular and dyslexic adult bilingual readers within linguistically and orthographically distant dyads, Hebrew (L1) and English (L2), was in the center of Breznitz and Fabian’s study (Chapter 11). More specifically, the focus of this research is on the target groups’ ability to process words in a list and in sentences in their L1 (Hebrew) and L2 (English). This chapter not only exemplifies neurocognitive research of typical and atypical bilingual reading processing, but also attempts to compare the readers’ behavior in light of their exposure to two languages (Semitic and Germanic) which are very distinct both linguistically and orthographically. The authors also provide a comprehensive review of neuropsychological studies in the reading domain and discuss the issues of definitions and theories on bilingualism from the neuropsychological perspective. More specifically, this chapter refers to various aspects of bilingual reading processing in different language domains (i.e., semantics, syntax and orthography), language differences and internal brain mechanisms underlying processes of typical and atypical reading. The findings of Breznitz and Fabian’s study indicate behavioral and electrophysiological differences in the processing profiles of dyslexic and regular bilingual readers in both first and second languages. Leikin and Ritvas (Chapter 12) examine differences in brain activity, specifically in Event-Related Potential (ERP) amplitudes and latencies in bilingual and monolingual (Russian/Hebrew and Hebrew) readers when processing

1

Current Issues in Bilingualism: A Complex Approach to a Multidimensional . . .

15

the grammatical functions of words during the reading of sentences in two languages (Russian and Hebrew). Two groups of adult university students participated in the study: native speakers of Russian and native speakers of Hebrew. Results indicated significant differences in brain activity and its localization between Russian and Hebrew speakers during processes of the grammatical functions of words in first and second (for bilingual group) languages, as reflected by Event-Related Potential (ERP) measures and LORETA-Key analysis. In both groups of participants, P100, P200, P300, N400, and P600 ERP waves were identified for three grammatical functions (subject, predicate, and direct object) in each sentence, for all reading items and in both languages. Analysis of the results showed that participants used different strategies to identify the grammatical functions of words in their native languages. The results appear to confirm the hypothesis (Leikin, 2002) that Hebrew readers identify the grammatical functions of words at least partly through their lexicalmorphological properties, while simultaneously using word order as well. By contrast, native speakers of Russian use morphologically based, noun-oriented strategies. The findings are consistent with results reported in the recent literature showing that grammatical processing in a second language is fundamentally different from grammatical processing in one’s native language. It is our hope that the contents of this volume will integrate the emerging interdisciplinary field of multilingualism studies in multicultural societies by bringing together such diverse areas as linguistics, sociology, education, and cognitive science in general while concentrating on the context of Israel as a mosaic of multilingualism and a natural laboratory in particular.

References Albert, M. L., & Obler, L. K. (1978). The bilingual brain. New York: Academic. Aronin, L., & Hufeisen, B. (2010). The exploration of multilingualism: Development of research on L3, multilingualism and multiple language acquisition. Amsterdam/Philadelphia: John Benjamins. August, D., & Shanahan, T. (Eds.). (2006). Developing literacy in second-language learners: A report of the national literacy panel on language-minority children and youth. Mahwah, NJ: Lawrence Erlbaum Associates. Ayari, S. (1996). Diglossia and illiteracy in the Arab world. Language. Culture and Curriculum, 9, 243–252. Ben-Rafael, E. (1994). Language, identity and social division: The case of Israel. Oxford: Clarendon Press. Ben-Rafael, E., Olshtain, E., & Geijst, I. (1997). Identity and language: The social insertion of Soviet Jews in Israel. In N. Levin-Epstein, Y. Ro’i, & P. Ritterband (Eds.), Russian Jews on three continents: Migration and resettlement (pp. 364–388). London: Frank Cass. Bhatia T. K., & Ritchie, W. C. (2006). The handbook of bilingualism. Oxford: Blackwell. Birdsong, D. (2006). Age and second language acquisition. In M. Gullberg & P. Indefrey (Eds.), The cognitive neuroscience of second language acquisition (pp. 9–49). Malden: Blackwell.

16

M. Leikin et al.

Breznitz, Z., & Lebovitz, L. (2008). Neurobiological correlates of dyslexia (pp. 7–50). In Z. Breznitz (Ed.), Brain Research in Language. New York, USA: Springer Publishing Company. Briellman, R. S., Saling, M. M., Connell, A. B., Waites, A. B., Abbott, D. F., & Jackson, G. D. (2004). A high-field functional MRI study of quadrilingual subjects. Brain and Language, 89, 531–542. De Bot, K., & Kroll, J. F. (2002). Psycholinguistics. In N. Schmitt (Ed.), Introduction to applied linguistics (pp. 133–149). London: Arnold Publishers. Deacon, H., & Cain, K. (2011). What we have learned from ‘learning to read in more than one language’. Journal of Research in Reading, 34(1), 1–5. De Groot, A. M. B., & Kroll, J. F. (Eds.). (1997). Tutorials in bilingualism: Psycholinguistic perspectives. Mahwah, NJ: Lawrence Erlbaum Associates. Dewaele, J. M., Housen, A., & Wei, L. (2003). Bilingualism: Beyond basic principles. Clevedon: Multilingual Matters. Dijkstra, T. (2001). What we know about bilingual word recognition: A review of studies and models. Plenary Address at the International Symposium on Bilingualism, Bristol, UK. Donald, M. W., Meuter, R., & Ardal, S. (1986). Event-related brain potentials in the first and second languages of bilinguals. Society of Neuroscience Abstracts, 12(1), 721. Donitsa-Schmidt, S. (1999). Language maintenance or shift: Determinants of language choice among Soviet immigrants in Israel (Unpublished doctoral thesis). Toronto: University of Toronto. Fabbro, F. (1999). The neurolinguistics of bilingualism: An introduction. Hove: Psychology Press. Ferguson, C. H. (1959). Diglossia. Word, 15, 325–340. Fishmann, J. A. (1991). Language and Ethnicity. Amsterdam & Philadelphia: J. Benjamins Publishing Company. Fishmann, J. A. (Ed.). (2001). Can Threatened Languages Be Saved? Clevedon: Multilingual Matters. Genesee, F., Paradis, J., & Crago, M. (2004). Dual language development and disorders: A handbook on bilingualism and second language learning. Baltimore: Brookes. Gerard, L. D., & Scarborough, D. L. (1989). Language-specific lexical access of homographs by bilinguals. Journal of Experimental Psychology: Learning, Memory, and Cognition, 15(2), 305–315. Glinert, L. (1995a). Inside the language planner’s head: Tactical responses to new immigrants. Journal of Multilingual and Multicultural Development, 16(5), 351–371. Grimes, B. (Eds.). (2000). Ethnology: Volume 1, languages of the word (14th ed.). Dallas: SIL International. Grosjean, F. (1998). Studying bilinguals: Methodological and conceptual issues. Bilingualism: Language and Cognition, 1(2), 131–149. Haznedar, B., & Gavruseva E. (Eds.). (2008). Current trends in child second language acquisition. Amsterdam/Philadelphia: John Benjamins Publishing Company. Hull, R., & Vaid, J. (2005). Clearing the cobwebs from the study of the bilingual brain: Converging evidence from literature and electrophysiological research (pp. 480–496). In J. F. Kroll & A. M. B. de Groot (Eds.), Handbook of bilingualism. Oxford: University Press. Ibrahim, R., & Eviatar, Z. (2007). Meta-linguistic awareness and reading performance: A cross language comparison. Journal of Psycholinguistic Research, 36(4), 297–317. Illes, J., Francis, W. S., Desmond, J. E., Gabrieli, J. D. E., Glover, G. H., Poldrack, R., et al. (1999). Convergent cortical representation of semantic processing in bilinguals. Brain and Language, 70(3), 347–363. Kahn-Horwitz, J., Schwartz, M., & Share, D. (2011). Acquiring the complex English orthography: A triliteracy advantage? Journal of Research in Reading, 34(1), 136–156. King, J., Effrati, R., & Netzer, N. (2003). Select findings of surveys of Ethiopian immigrants in eight urban communities in Israel. Meyers – JDC – Brookdale Institute and the Absorption Ministry, State of Israel, www.jdc.org.il/brookdale

1

Current Issues in Bilingualism: A Complex Approach to a Multidimensional . . .

17

Koda, K. (2005). Insights into second language reading: A cross-linguistic approach. New York: Cambridge University Press. Kotik-Friedgut, B. (2000). Variety of languages and the dynamics of language policy in contemporary Israel. In A. D. Epstein & A. V. Fedorchenko (Eds.), Mass migration and its impact on the Israeli society (pp. 154–192). Moscow: Institute Izucheniya Izrailya i Blizhnego Vostoka [in Russian]. Kroll, J. F., & de Groot, A. M. B. (Eds.). (2005). Handbook of bilingualism. Oxford: University Press. Leikin, M. (2002). Processing words’ syntactic functions in normal and Russian readers. Journal of Psycholinguistic Research, 31(2), 145–163. Leikin, M. (2008). Syntactic processing in two languages by native and bilingual adult readers: An electrophysiological study. Journal of Neurolinguistics, 21, 273–349. Lesaux, N. K., & Geva, E. (2006). Synthesis: Development of literacy in language minority learners. In D. L. August & T. Shanahan (Eds.), Developing literacy in a second language: Report of the national literacy panel. (pp. 53–74). Mahwah, NJ: Lawrence Erlbaum Associates. Lesaux, N. K., with Koda, K., Siegel, L. S., & Shanahan, T. (2006). Development of literacy of language minority learners. In D. L. August & T. Shanahan (Eds.), Developing literacy in a second language: Report of the national literacy panel. (pp. 75–122). Mahwah, NJ: Lawrence Erlbaum Associates. Leshem, E., & Lissak, M. (1999). Development and consolidation of the Russian community in Israel. In S. Weil (Ed.), Roots and routes: Ethnicity and migration in global perspective (pp. 135–171). Jerusalem: The Magnes Press, Hebrew University. Li, P., & Green, D. W. (2007). Neurocognitive approaches to bilingualism: Asian languages. Bilingualism: Language and Cognition, 10, 117–119. Marian, V., Spivey, M., & Hirsch, J. (2003). Shared and separate systems in bilingual language processing: Converging evidence from eye tracking and brain imaging. Brain and Language, 86(1), 70–82. Nudelman, R. (2000). Israeli society and its new search of identity (pp. 221–253). In A. D. Epstein & A. V. Fedorchenko (Eds.), Mass migration and its impact on the Israeli society. Moscow: Institute Izucheniya Izrailya i Blizhnego Vostoka [in Russian]. Olstein, A. (1995). Paper presented at a conference on Ingathering of the exiles? The impact of the 1990s’ wave of immigration on central areas of Israeli life, the Israel Sociological Society Immigrant absorption section, Tel-Aviv University [in Hebrew]. Paradis, M. (2004). Neurolinguistic theory of bilingualism. Amsterdam/Philadelphia: John Benjamins. Paradis, J. (2007). Bilingual children with specific language impairment: Theoretical and applied issues. Applied Psycholinguistics, 28, 512–564. Paradis, M., & Goldblum, M. S. (1989). Selective crossed aphasia in a trilingual aphasic patient followed by reciprocal antagonism. Brain and Language, 36(1), 62–75. Perani, D., & Abutalebi, J. (2005). The neural basis of first and second language processing. Current Opinion in Neurobiology, 15, 202–206. Proverbio, A. M., Cok, B., & Zani, A. (2002). Electrophysiological measures of language processing in bilinguals. Journal of Cognitive Neuroscience, 14, 994–1017. Saiegh-Haddad, E. (2003). Linguistic distance and initial reading acquisition: The case of Arabic diglossia. Applied Psycholinguistics, 24, 115–135. Sarfarazi, M., & Sedgwick, E. M. (1996). Event related potentials (ERP) as a function of language processing in bilinguals. Electroencephalography and Clinical Neurophysiology, 99(4), 347. Shuval, J. T. (1999). Israel in the context of post industrial migration: The mythology of ‘‘Uniqueness’’. In W. Weil (Eds.), Roots and routes: Ethnicity and migration in global perspective (pp. 225–245). Jerusalem: The Magnes Press, Hebrew University.

18

M. Leikin et al.

Sparks, R., & Ganschow, L. (1993a). Searching for the cognitive locus of foreign language learning difficulties: Linking native and foreign language learning. Modern Language Journal, 77, 289–302. Sparks, R., & Ganschow, L. (1993b). The impact of native language learning problems on FL learning: Case study illustrations of the linguistic coding deficit hypothesis. Modern Language Journal, 77, 58–74. Spolsky, B. (1996). English in Israel after independence (pp. 535–555). In J. A. Fishman, A. Rubal-Lopez, & A. W. Conrad (Eds.), Post-imperial English. Berlin: Mouton. Spolsky, B., & Cooper, R. L. (1991). The languages of Jerusalem. Oxford: Clarendon Press. Spolsky, B., & Shohamy, E. (1999). The languages of Israel: Policy, ideology and practice. Clevedon: Multilingual Matters. Svirsky, S., & Svirsky, B. (2002). The Ethiopian Jews in Israel: Housing, employment and education. In ‘‘Meida al Shivion’’ 11, Edva Center, Israel. Ullman, M. T. (2001). The neural basis of lexicon and grammar in first and second language: The declarative/procedural model. Bilingualism: Language and Cognition, 4(1), 105–122. Wartenburger, I., Heekeren, H., Abutalebi, J., Cappa, S., Villringer, A., & Perani, D. (2003). Early setting of grammatical processing in the bilingual brain. Neuron, 37(1), 159–170. Zhang, Y., & Wang, Y. (2007). Neural plasticity in speech acquisition and learning. Bilingualism: Language and Cognition, 10(2), 147–160.

Part I

Language and Literacy in Multilingual Society

Chapter 2

Relevance of the Linguistic Coding Difference Hypothesis to English as an Additional Language of Literacy in Israel Janina Kahn-Horwitz, Richard L. Sparks, and Zahava Goldstein

2.1 Introduction Children studying in Israeli schools, as in most other countries where English is not a first language, begin acquiring English as a second or additional language (EAL) in elementary school. English is considered an important subject of study and a passport to future success in higher education, business and travel (Ministry of Education, 2001). In Israel, the majority of the school-going population is comprised of children for whom Hebrew is their first language (L1). Arabic is the official second language of Israel. The largest minority group in Israel are Arabic L1 speaking children, and alongside this group there are numerous originally immigrant minority groups including Russian, Amharic, French, and Spanish L1 speaking children. For the aforementioned groups, English will be a third or fourth language. In this research, only Hebrew L1 students acquiring EAL literacy were tested. Hebrew L1 speaking children will have, for the most part, acquired at least 2 years of Hebrew literacy before they start acquiring EAL literacy. During the 9 to 12 years of formal EAL studies at school,1 children continue developing their literacy ability so that by the time they complete school and enter tertiary education they are expected to be able to read and comprehend EAL academic articles. The purpose of the present study was to examine processes that contribute to successful reading in an additional language in ninth grade by examining both L1 Hebrew as well as EAL reading components regarding their predictive ability from a longitudinal perspective.

1

Although English literacy instruction begins in third or fourth grade in Israel with instruction in letter sound and name knowledge, some local municipal areas and/or schools may begin oral English instruction from first grade. J. Kahn-Horwitz (*) Oranim Academic College of Education, Edmond J. Safra Brain Research Center for the Study of Learning Disabilities, University of Haifa, Haifa, Israel e-mail: [email protected] M. Leikin et al. (eds.), Current Issues in Bilingualism, Literacy Studies 5, DOI 10.1007/978-94-007-2327-6_2, Ó Springer ScienceþBusiness Media B.V. 2012

21

22

J. Kahn-Horwitz et al.

Models of word recognition take into account orthographic specific characteristics that influence word recognition (Frost, 2005). Hebrew is a Semitic language and it can be read either with or without vowels, which include diacritical marks inserted above or below the letters. This distinguishes Hebrew vowels both orthographically and phonemically from English vowels. As opposed to Hebrew vowels, English vowels have the same orthographic status as other letters in the alphabet and phonemically Hebrew vowels do not have a short long distinction. The voweled version of the Hebrew orthography, which is acquired and used by elementary school children in their first 2 years of reading acquisition, can be decoded by direct grapheme-phoneme translation making it a relatively transparent or regular alphabetic orthography. It is less transparent for encoding as there are numerous graphemes that represent a number of the phonemes such as tet/taf, kuf/kaf, xaf/xet and aleph/ayin (Frost, 2005; Geva, Wade-Woolley, & Shany, 1993). Without vowels, the Hebrew orthography, which historically is essentially a consonant-based syllabary, is significantly more opaque, and accurate decoding depends on utilizing morphological, semantic and contextual information in order to fill in incomplete phonological information. The English orthography is considered opaque for both decoding and encoding for two central reasons. The first reason is that irregular words do not follow orthographic conventions in terms of their grapheme-phoneme correspondences (e.g., who, one), and the second reason is inconsistency in the resulting phonemes that particular orthographic patterns in different words result in (e.g., ough in through, enough and bought) (Frost, 2005; Seymour, Aro, & Erskine, 2003). Experience with specific L1 word recognition processes may impact on EAL word recognition as well as EAL comprehension (Koda, 1995). According to a simple model of reading (Koda, 2005) reading comprehension in a first language is dependent on accurate and fluent word recognition together with listening comprehension. A limited-capacity theory explains how these two automatic lower level processes free cognitive resources to process higher level reading comprehension processes (Koda, 1992). Similar processes are required for second or foreign language (L2) reading comprehension (Koda, 1996). L2 reading acquisition is largely dependent on underlying linguistic abilities as well as processing speed, which first impacted on first language (L1) reading acquisition (Kahn-Horwitz, Shimron, & Sparks, 2005; Sparks & Ganschow, 1993a, 1993b; Sparks, 1995). Underlying linguistic abilities impacting reading acquisition include phonological awareness, which together with knowledge of letter sounds and names, constitute the alphabetic principle. This is a crucial precursor of reading acquisition in first and additional languages (Muter & Diethelm, 2001). Phonological awareness facilitates reading in that it enables readers to associate phonemes with graphemes and synthesize the phonemes to form words, a process referred to as phonological recoding (Adams, 1990; Hagiliassis, Pratt, & Johnston, 2006; Oloffson & Niedersoe, 1997; Share, 1995). Phonological awareness has been

2

Relevance of the Linguistic Coding Difference Hypothesis to English . . .

23

found to transfer across orthographies (Cisero & Royer, 1995; Comeau, Cormier, Grandmaison, & Lacroix, 1999; Dufva & Voeten, 1999; Durgunoglu, Nagy, & Hancin-Bhatt, 1993; Kahn-Horwitz et al., 2005; Wade-Woolley, Chiappe, & Siegel, 1998). Phonological awareness has been found to be the strongest predictor of word reading for EL1 and EAL first to fourth graders (Jongejan, Verhoeven, & Siegel, 2007) as well as showing longitudinal effects for kindergarten to sixth grade EAL students (together with RAN and oral language measures) (Nakamato, Lindsey, & Manis, 2007). Together with oral Hebrew ability, phonological awareness accounted for L2 Hebrew reading amongst Russian-Hebrew bilinguals (Leikin, Share, & Schwartz, 2005) and when measured in Spanish L1 phonological awareness has also been found to predict reading comprehension in English L2 amongst elementary school Spanish-English bilinguals and biliterates (Carlisle, Beeman, Hull Davis, & Sphraim, 1999). The direct relationship between phonology and comprehension is explained by incoming information being stored temporarily in the phonological loop component of working memory (Baddeley, 2006). If there is a disruption in this temporary storage of information, there will be a faulty flow of information to the long term memory function. The phonological loop component of working memory is thus vital to L1 as well as additional language vocabulary acquisition (Gathercole & Thorn, 1998) and subsequently reading comprehension in the respective languages (Koda, 2005). Together with phonological processing, orthographic processing is a crucial component of word decoding and encoding. Orthographic processing entails using visual-orthographic input from words for the purposes of word recognition (Ehri, 1992; Nassaji & Geva, 1999). Novice readers depend more on phonological processing to decode words and as these words become more frequent for these readers, orthographic processing is used to directly access them from the lexicon (Share, 1995; Share & Stanovich, 1995). Lexical access is the ability to automatically access oral vocabulary from memory and is measured by rapidly naming numbers, letters, pictures or colors. There is a strong connection between this ability and reading for younger EL1 students and for younger and older EAL elementary school students (Gholomain & Geva, 1999; Jongejan et al., 2007; Nakamato et al., 2007). Oral vocabulary is stored phonologically in memory making this task a phonological processing task. However, it measures a qualitatively different dimension of phonological processing from phonological awareness tasks which measure the ability to manipulate phonemes thereby facilitating decoding and encoding (Hagiliassis et al., 2006). Morphological awareness requires sensitivity to the meaningful components of words and can assist in word recognition, lexical access, and reading comprehension for L1 (Oloffson & Niedersoe, 1997) and L2 (Parel, 2004) readers. Morphological awareness measured in Hebrew L1 predicted EAL decoding as well as reading comprehension amongst fourth graders after their first year of EAL literacy acquisition (Kahn-Horwitz et al., 2005).

24

J. Kahn-Horwitz et al.

Theories that explain the impact of underlying abilities on additional reading acquisition and development include the Linguistic Coding Differences Hypothesis (LCDH) (Sparks & Ganschow, 1993a), the Central Processing Hypotheses (Geva & Siegel, 2000) and the Interdependence Hypothesis (Cummins, 1984). These theories explain that linguistic codes, or abilities, measured in L1 explain individual differences in additional language literacy acquisition. In the present study, we question whether after numerous years of additional language learning when the language learner is no longer at the initial stage of reading acquisition, linguistic codes measured in L1 still explain differences in EAL reading. Alternatively, do abilities measured in the EAL itself become predictors of the more skilled EAL reader? If so, this finding would show evidence for a threshold hypothesis whereby EAL abilities predict EAL reading comprehension (Leikin et al., 2005; van Gelderen, Schoonen, Stoel, & de Glopper, 2007). Sparks, Patton, Ganschow, Humbach, and Javorsky (2008) found longitudinal evidence for L1 literacy abilities measured in elementary school successfully predicting second language literacy abilities measured in 9th and 10th grade after one and then 2 years of second language study. In another study with L2 learners in high school, they found that L1 skill differences in reading and spelling emerged early in elementary school (Sparks, Ganschow, & Patton, 2008; Sparks, Patton, Ganschow, & Humbach, 2009). Their studies examined English L1 speakers acquiring additional alphabetic languages (German, Spanish and French). The present study reports on ninth grade Hebrew L1 speakers in their sixth year of English as an additional language study. This lengthier formal exposure to English as an additional language may have changed the relationship between L1 literacy related abilities and EAL literacy abilities.

2.1.1 The Present Study The aim of the present study was to address the relevance of the LCDH after almost 6 years of EAL instruction. In other words, do linguistic abilities measured in L1 Hebrew in fourth grade and again in ninth grade still predict EAL word reading and reading comprehension measured at the end of ninth grade? In the case of the present study, we examined students for whom Hebrew is their first spoken and literate language. After having acquired the transparent voweled Hebrew orthography in first and second grades, students in third grade began reading and spelling the more opaque unvoweled Hebrew consonant-based syllabic orthography with less phonological information. In fourth grade these same students began acquiring the more opaque English orthography with its relative inconsistency and irregularity (Frost, 2005). The present study aimed to examine whether long term transfer of L1 literacy abilities would still be observed when long term EAL predictors were included as predictors.

2

Relevance of the Linguistic Coding Difference Hypothesis to English . . .

25

2.2 Method 2.2.1 Participants Participants were 77 ninth graders from the north of Israel who were part of an original study consisting of 145 children (Kahn-Horwitz et al., 2005; KahnHorwitz, Shimron, & Sparks, 2006) and participated in this study 5 years earlier in fourth grade. The original participants in the study came from schools in three different areas, representing two different socio-economic status backgrounds (SES) (Kahn-Horwitz et al., 2006). At the beginning of seventh grade, participants in the original study started junior high school and participants from two of the schools were dispersed amongst numerous high schools in various areas, making it difficult to locate them towards the end of ninth grade. Most of the participants from one of the three original schools in a middle SES area continued studying at the adjoining junior high school. Of the original 96 participants in this high school, 77 agreed to take part in this follow up study. Their participation was dependent on parental consent. Of the 77 participants in the study, 42 were males and 35 were females. Their mean age was 15 years, 1 month. All participants had been formally studying EAL for 4 weekly hours over a period of 6 years.

2.2.2 Measures The following L1 measures were administered in order to examine whether abilities associated with L1 reading ability will also longitudinally predict EAL reading ability. 2.2.2.1 Independent L1 (Hebrew) Measures 1. Phonological awareness was assessed using a phoneme deletion task (Shany, Zieger, & Ravid, 2001), which is a Hebrew version of the Rosner phoneme deletion task. Participants were required to repeat a word that the tester pronouncedRand then delete an initial, medial or final phoneme from R that word, e.g., / elet/ (Hebrew for ‘‘sign’’) and then delete the phoneme / /. The participant was then required to pronounce the resulting pseudo-word /elet/. The maximum score for this task was 20 (range 0–20). Participants were tested on this task at the beginning and end of fourth grade and at the end of ninth grade. 2. Morphological awareness was assessed using a production task (Ben-Dror, Bentin, & Frost, 1995). The tester pronounced a word root and then participants completed a sentence provided by the tester with the appropriate e.g., lamad (Hebrew for ‘‘studied’’ /masc./sg.) – ‘‘Ha’yeled Rderivation, R eyo ev kan hu ha’_________’’ (Hebrew for ‘‘The boy who is sitting here is the ____’’). The student had to complete the sentence with talmid (Hebrew

26

3.

4.

5.

6.

7.

8.

J. Kahn-Horwitz et al.

for ‘‘pupil’’/masc./sg.). The maximum score for this task was 15 (range 0–15). Participants were tested on this task at the beginning and end of fourth grade. A spelling dictation consisting of single Hebrew words (which were pronounced as single words and then in the context of a sentence) and two sentences were dictated to the fourth graders (see Kahn-Horwitz et al., 2005). The maximum score for this task was 25 (range 0–25). Participants were tested on this task at the beginning and end of fourth grade. The scores of the beginning and end of fourth grade spelling tests were combined based on their correlation (r ¼ 0.80, p < 0.001). A spelling dictation that (Shany et al., 2001) consisted of 43 words was dictated to participants. Participants heard each word pronounced, followed by a contextual sentence and again the key word. They thereafter wrote the word, e.g., nikneta (Hebrew for ‘‘was bought’’ /fem./sg.). The maximum score for this task was 43 (range 0–43). Participants were tested on this task at the end of ninth grade. The fourth and ninth grade spelling dictations were combined based on their correlation (r ¼ 0.61, p < 0.001). A word reading cluster was administered at the beginning and end of fourth grade, which consisted of a word recognition task (Balgur, 1977) testing accuracy and speed (beginning of fourth grade correlation between word recognition accuracy and speed r ¼ 0.36, p < 0.01; end of fourth grade correlation between accuracy and speed r ¼ 0.16, ns), and a word attack task measuring phonological recoding (Greenbaum & Lichter, 1996), which were combined based on their Pearson 2-tailed correlation coefficients (beginning of fourth grade word recognition with word attack accuracy r ¼ 0.71, p < 0.01; end of fourth grade word recognition with word attack accuracy r ¼ 0.70, p < 0.01, and beginning of fourth grade word attack and word recognition speed r ¼ 0.21, p < 0.05; and end of fourth grade word attack and word recognition speed r ¼ 0.27, p < 0.05). At the end of ninth grade, the Greenbaum and Lichter (1996) Hebrew Word Attack task measuring phonological recoding which consists of 25 nonwords following the phonotactic conventions of the Hebrew language, e.g., vasug, was again administered. Both speed (in seconds) and accuracy (maximum score ¼ 25) was checked (range 0–25). At the end of ninth grade a task measuring rapid automatic naming of letters in Hebrew was administered (Shany et al., 2001). Each participant was asked to name accurately and rapidly five Hebrew letters. The letters were: ‫ל‬, ‫ג‬, ‫ד‬, ‫א‬, ‫ – ס‬samex, alef, daled, gimel, lamed. A vocabulary cluster (Glantz, 1991) consisting of an antonyms and synonyms task was administered at the beginning and end of fourth grade and a more advanced version was administered at the end of ninth grade (Pearson 2-tailed correlation coefficients between beginning of fourth grade antonym and synonym tasks r ¼ 0.43, p < 0.001; Pearson 2-tailed correlation coefficients between end of fourth grade antonym and synonym tasks r ¼ 0.27, p < 0.001; Pearson 2-tailed correlation coefficients between end of ninth

2

Relevance of the Linguistic Coding Difference Hypothesis to English . . .

27

grade antonym and synonym tasks r ¼ 0.42, p < 0.001). In the tasks participants saw and heard a key word followed by a series of 4–5 words of which they were required to choose the appropriate synonym or antonym. The maximum score for antonyms and synonyms in fourth grade was 12 (range 0–12). The maximum score for antonyms and synonyms in ninth grade was 20 (range 0–20).

2.2.2.2 Independent EAL (English) Measures 1. An English letter knowledge cluster consisting of a measure of knowledge of English letter names and sounds was administered at the beginning and end of fourth grade. Participants were required to name and pronounce the sound of each of the 26 English letters presented in a random manner (Pearson 2-tailed correlation coefficients between beginning of fourth grade knowledge of sounds and names of English letters r ¼ 0.96, p < 0.01; Pearson 2-tailed correlation coefficients between end of fourth grade knowledge of sounds and names of English letters r ¼ 0.95, p < 0.01). The scores for both sounds of letters and names of letters was 0–26. 2. Word Reading – At the end of fourth grade the following three tasks were administered: a. An informal measure with 20 words that the fourth graders had been exposed to during their first year of EAL studies (e.g., stop, farm, number) was presented to the students (see Kahn-Horwitz et al., 2005). Each student was requested to read all 20 words aloud. The maximum score for this task was 20 (range 0–20). b. The Woodcock Reading Mastery Test-Revised (1987), Form H: Word Recognition sub-test was administered. Students read word after word aloud and stopped reading after six consecutive errors. c. The Woodcock Reading Mastery Test-Revised, Form H: Word Attack sub-test was administered. Students read each of the non-words aloud and stopped reading after six consecutive errors. The above tasks were combined to form a word reading cluster based on their Pearson 2-tailed correlation coefficients, which yielded the following correlations: informal measure and Woodcock Word Recognition (r ¼ 0.87, p < 001), informal measure and Woodcock Word Attack (r ¼ 0.87, p < 001), Woodcock Word Recognition and Woodcock Word Attack (r ¼ 0.85, p < 001). 3. Spelling – At the end of fourth grade, the following two tasks checked basic EAL spelling: a. Students were asked to write graphemes representing eight phonemes. The maximum score was 8 (range 0–8). b. Students were asked to write a short sentence consisting of eight frequent words. The sentence was Look at the black cat in the tree (range 0–8).

28

J. Kahn-Horwitz et al.

These two tasks were combined into a spelling cluster based on Pearson 2-tailed correlation coefficients r ¼ 0.59, p < 0.001. c. At the end of ninth grade, three English sentences were dictated to the participants. The sentences consisted of words taken from ninth grade text books (Assis, 2002; Ezra & Kerman, 1998), e.g., Last year he taught computer technology. Participants first heard the entire sentence read twice and then the tester dictated chunks of two to three words at a time. Participants were required to write the sentences. The maximum score for this sub-test was 25 (range 0–25). The overall fourth and ninth grade spelling tasks were combined into a spelling cluster based on Pearson 2-tailed correlation coefficients r ¼ 0.58, p < 0.001. 4. Vocabulary – At the end of fourth grade the following vocabulary tasks were administered: a. A single word vocabulary identification task was completed by the students. They were presented with a page including ten items. The tester read ten target words aloud and the students circled the appropriate picture (out of 4 possible options) that represented the word they heard, e.g., they heard the word horse and had to circle the appropriate picture amongst pictures of man, bag, horse, table. The maximum score for this task was 10 (range 0–10). b. A simple sentence comprehension task was completed by students. They circled the appropriate picture representing the simple sentence they heard, e.g., Three children are playing. The maximum score for this task was 15 (range 0–15). These two tasks then were combined into a vocabulary cluster based on Pearson 2-tailed correlation coefficients r ¼ 0.68, p < 0.001. At the end of ninth grade, three English vocabulary tasks were administered. They were designed using 60 vocabulary items from two ninth grade EAL text books (Assis, 2002; Ezra & Kerman, 1998). Fifteen ninth grade EAL teachers were asked to grade word frequency of these words amongst ninth graders by giving a grade of one to five, whereby one represented a very infrequent word and five represented a very frequent word. Each word was then given a cumulative score and three separate lists were created to be equivalent from a frequency perspective, with the most frequent words being earlier in the lists progressing towards less frequent words. c. The first sub-test was a passive vocabulary recognition task whereby the student heard and saw a key word followed by five possible Hebrew translations. One of the five options was the accurate translation; the second was a Hebrew foil, which was the same part of speech as the key word; the third had the same opening phoneme as the key word; the fourth option’s translation into English had the same opening phoneme as the key word; and the fifth either rhymed with or was morphologically similar to the Hebrew translation of the key word, e.g., bridge (key word) – a. R delet (English – door: same part of speech) b. ne er (English – eagle

2

Relevance of the Linguistic Coding Difference Hypothesis to English . . .

29

(masc./sg.): morphologically similar to Hebrew translation) c. mevi (English – bringR (masc./sg.): translation has same opening phonemes as key word) d. ge er (English – bridge (masc./sg.): accurate translation) e. brit (fem./sg.)(English – pact: Hebrew foil has same opening phonemes as key word). The maximum score for this measure was 20 (range 0–20). d. The second sub-test was a contextualized vocabulary recognition task. In this sub-test, the student saw and heard a key word followed by a contextual sentence and then heard the key word repeated again. The participant was required to say aloud the Hebrew translation for the English key word e.g., key word: danger, contextual sentence: Fire can be a danger. The maximum score for this sub-test was 20 (range 0–20). e. The third sub-test was a de-contextualized vocabulary recognition task. In this sub-test, the student heard and saw the key word and had to then say aloud the Hebrew translation for the key word, e.g., news, dry, nature. The maximum score for this sub-test was 20 (range 0–20). The three vocabulary measures were combined into a vocabulary cluster. Pearson 2-tailed correlation coefficients between these variables yielded the following results: Passive vocabulary recognition and contextualized vocabulary recognition r ¼ 0.95, p < 0.001; passive vocabulary recognition and de-contextualized vocabulary recognition r ¼ 0.93, p < 0.001; contextualized vocabulary recognition and de-contextualized vocabulary recognition r ¼ 0.96, p < 0.001. 5. Reading Comprehension – at the end of fourth grade, English reading comprehension was tested by two different short narrative texts, each followed by five multiple choice questions presented in Hebrew L1 that checked factual information as well as the overall idea of each text (see Kahn-Horwitz et al., 2005). The maximum score for this task was 10 (range 0–10). 6. Lexical Access – at the end of ninth grade a task measuring rapid automatic naming of letters was administered (Kail & Hall, 1994). Each participant was asked to name as accurately and rapidly as they could five English letters that appeared in ten different sequences. The letters were: a, d, o, p, s. Speed was measured using a stopwatch and accuracy was measured by marking off any letter incorrectly named.

2.2.2.3 Dependent EAL Measures Tested at the End of Ninth Grade 1. Word Reading – The following three tasks were administered: a. A single word recognition task (Stanovich & Siegel, 1994) consisting of 40 frequent one syllable words was presented to students. They read the words aloud. Speed was measured using a stopwatch and accuracy was measured by giving a score of 1 to each correctly read word. Maximum score for accuracy was 40 (range 0–40).

30

J. Kahn-Horwitz et al.

b. The Woodcock Reading Mastery Test-Revised, Form H: Word Recognition sub-test was presented to students and they were required to read each word aloud. After six consecutive errors, the task was ended. c. The Woodcock Reading Mastery Test-Revised, Form H: Word Attack sub-test was presented to students and they were required to read each non-word aloud. The task was discontinued after six consecutive errors. A word reading cluster was combined based on the following Pearson 2-tailed correlation coefficients: Woodcock Word Recognition and Woodcock Word Attack (r ¼ 0.70, p < 0.01); Woodcock Word Recognition and the Stanovich and Siegel Word Recognition Task (accuracy) (r ¼ 0.68, p < 0.01); Woodcock Word Recognition and the Stanovich and Siegel Word Recognition Task (speed) (r ¼ –0.49, p < 0.01); Woodcock Word Attack and the Stanovich and Siegel Word Recognition Task (accuracy) (r ¼ 0.54, p < 0.01); Woodcock Word Attack and the Stanovich and Siegel Word Recognition Task (speed) (r ¼ –0.44, p < 0.01); and the Stanovich and Siegel Word Recognition Task (accuracy and speed) (r ¼ –0.67, p < 0.01). 2. At the end of ninth grade, EAL reading comprehension was measured using three progressively graded texts designed for ninth graders by EAL teachers. Comprehension questions presented in English examined word level (semantic) comprehension, text level comprehension (detecting explicit information), global comprehension (finding the main idea), and conceptual comprehension found within the text as well as the application of background knowledge. Maximum score for the overall test was 50 (range 0–50).

2.2.3 Procedure At the beginning as well as the end of their first year of EAL studies, the first author conducted group as well as individual testing on the original fourth grade participants, which lasted a little over 1 h in a quiet room at the elementary school. The first author together with research assistants (who were all qualified English teachers studying Didactic Assessment and who received special training for using the above measures) tested each of the participants individually during the second last month of their ninth grade year. Testing took place in a quiet room in the school and lasted approximately 1 h. The phonological awareness, vocabulary and spelling tasks were measured in Hebrew (L1) at the beginning and end of fourth grade and at the end of ninth grade. Each of these three L1 measures was combined to create a mean score for phonological awareness, vocabulary knowledge and spelling ability, respectively. The morphological awareness task and the word reading cluster measured in Hebrew (L1) at the beginning and end of fourth grade was combined to create a mean score for morphological awareness and word reading, respectively.

2

Relevance of the Linguistic Coding Difference Hypothesis to English . . .

31

Due to a lack of significant correlations between many of the response time and accuracy measures, a finding that has been supported by the literature whereby speeded as opposed to accuracy tasks are seen as tapping qualitatively different underlying abilities (Hagiliassis et al., 2006), stepwise regression analyses were run separately for 9th grade English word recognition accuracy results and for 9th grade English word recognition accuracy combined with response time results.

2.3 Results Means and standard deviations were calculated for each of the independent and dependent variables (see Tables 2.1 and 2.2). Correlations between the respective variables were calculated (see Table 2.3). Speeded measures often did not correlate significantly with accuracy measures. Correlations were insignificant between lexical access (L1 RAN) and the phoneme deletion measured in Hebrew. Phoneme deletion measured in Hebrew also had a non-significant correlation with the response time for Hebrew word attack. A non-significant correlation was observed between Hebrew spelling and Hebrew word reading.

Table 2.1 Descriptive data for accuracy based tasks: Means (M), standard deviations (SD), minimum (Min. score) and maximum observed scores (Max. score) and observed range of participants on the L1 and EAL measures Min. Max. Measure M SD score score Range Hebrew (L1) Measures L1 Phoneme Deletion* L1 Morphological Awareness* L1 Spelling* L1 Word Reading Cluster* L1 Vocabulary Cluster* English (EAL) Measures EAL Letter Sound and Names* EAL Spelling* EAL Word Reading Cluster – 4th grade* EAL Word Reading Cluster – 9th grade* EAL Vocabulary Cluster* EAL Reading Comprehension 4th grade EAL Reading Comprehension 9th grade * z scores

0.03 0.01 –0.03 –0.01 –0.02

0.83 0.95 0.92 0.62 0.61

–1.57 –2.69 –3.96 –1.43 –1.52

1.82 1.97 1.51 1.05 1.36

3.39 4.66 5.47 2.48 2.88

0.04 0.07 0.23

0.86 0.85 1.00

–2.12 –1.78 –1.22

1.94 1.74 2.83

4.06 3.52 4.05

0.001

0.54

–1.58

1.06

2.63

0.004 4.44

0.82 3.86

–3.15 0

1.58 10

4.73 10

50

50

35.12

12.28

0

32

J. Kahn-Horwitz et al.

Table 2.2 Descriptive data for speed based tasks: Mean response time in seconds (M) and standard deviations (SD) of participants on the L1 and EAL measures Measure M SD L1 Word Attack Speed L1 Rapid Automatic Naming EAL Word Reading Cluster – 9th gradea EAL Word Reading Cluster incl. speed – 9th gradea EAL Rapid Automatic Naming a

25.93 22.41 0.001 0.01 21.18

11.56 3.76 0.50 0.51 4.09

z score

Separate stepwise regression analyses were conducted to determine whether and to what extent L1 variables measured in fourth grade and again in ninth grade (separately and then together) predicted EAL word reading and EAL reading comprehension (measured at the end of ninth grade). The results of these analyses would provide an answer regarding the extent that the Linguistic Coding Differences Hypothesis is relevant to students studying EAL for 6 years. Another set of separate analyses were conducted that combined both L1 and EAL measures tested in fourth and ninth grade to determine whether and to what extent L1 measures predicted EAL word reading and reading comprehension when they were entered together with EAL measures.

2.3.1 L1 Measures in Fourth Grade Predicting EAL Word Reading Accuracy in Ninth Grade The stepwise regression analysis yielded a one-variable model whereby phonological awareness measured in fourth grade accounted for 25% of the variance (b ¼ 0.50), F (1, 76) ¼ 24.61, p < 0.001.

2.3.2 L1 Measures in Fourth Grade Predicting EAL Word Reading in Ninth Grade Including the Speeded Reading Measure The stepwise regression analysis yielded a one-variable model whereby morphological awareness measured in fourth grade accounted for 6% of the variance (b ¼ –0.25), F (1, 76) ¼ 5.20, p < 0.05.

2.3.3 L1 Measures in Fourth and Ninth Grade Predicting EAL Word Reading in Ninth Grade Including the Speeded Reading Measure The stepwise regression analysis yielded a one-variable model whereby rapid automatic naming of Hebrew letters measured in ninth grade accounted for 11% of the variance (b ¼ 0.33), F (1, 75) ¼ 9.30, p < 0.01.

–0.15

–0.07

–0.09

– 0.11 0.49** 0.35** –0.16

5

6

0.06 –0.23* 0.15

0.18

–

7

–0.22

0.57**

–

8

–0.27*

–

9

–

10

11

12

13

–0.16

0.14

0.49** –0.02

0.45** 0.37** –0.12

–0.17

0.48** .37**

0.04

0.13

–0.19

–0.17

–0.15

0.27*

0.47** 0.63** 0.37** 0.57** – –

15

–0.36** 0.61** 0.63** 0.42** 0.74** 0.50** –0.15

0.51**

–0.19

14

–

16

17

–0.30** –0.45** 0.44** 0.62** 0.54** 0.52** 0.44** –0.44** 0.40** –

–0.26*

0.33**

0.43** 0.47** –0.28* 0.01

0.41** 0.37** –0.17

–0.25* –0.14

0.40** 0.29*

0.45** 0.51** 0.52** 0.14

0.02

0.60** 0.19

0.51** 0.50** 0.65** 0.26* 0.60** 0.59** –0.20 –0.30** –0.44** 0.74** – 0.51** 0.65** 0.56** 0.22* 0.66** 0.30** –0.03 –0.20 –0.23* 0.56** 0.55** – 0.54** 0.49** 0.56** 0.33** 0.61** 0.56** –0.24* –0.24* –0.39** 0.82** 0.80** 0.55** –

0.53** 0.53** 0.49** 0.37** 0.59** 0.53** –0.12

–0.17

–0.23* –0.12

– 0.66** – 0.43** 0.44** – –0.26* –0.21 –0.18

4

Table 2.3 Intercorrelations between L1 and EAL measures

–0.23* –0.25* –0.23* 0.11

0.59** 0.10 0.47** 0.10 –0.15

0.50** 0.31** 0.49** 0.38** –0.14

3

Relevance of the Linguistic Coding Difference Hypothesis to English . . .

* p < 0.05; ** p < 0.01

–

1. L1 Phoneme deletion 2. L1 Morphological awareness 3. L1 Spelling 4. L1 Word reading 5. L1 Vocabulary 6. L1 Word attack 7. L1 Word attack speed 8. L1 Rapid automatic naming 9. EAL Rapid automatic naming 10. EAL Letter sounds and names 11. EAL spelling 12. EAL vocabulary 13. EAL word reading 4th grade 14. EAL word reading 9th grade 15. EAL word reading speed 9th grade 16. EAL Reading comprehension 4th grade 17. EAL Reading comprehension 9th grade

2

0.51** –

1

MEASURE

2 33

34

J. Kahn-Horwitz et al.

2.3.4 L1 Measures in Fourth and Ninth Grade Predicting EAL Word Reading Accuracy in Ninth Grade The stepwise regression analysis yielded a two-variable model whereby the overall phonological awareness task and the overall Hebrew word reading task together accounted for 42% of the variance. Phonological awareness (b ¼ 0.60), F (1, 75) ¼ 41.23, p < 0.001 explained 35% of the variance and the Hebrew word reading task (b ¼ 0.28) F change (1, 74) ¼ 8.62, p < 0.01 explained an additional 7% of the variance.

2.3.5 L1 Measures in Fourth Grade Predicting EAL Reading Comprehension in Ninth Grade The stepwise regression analysis yielded a two-variable model whereby the Hebrew fourth grade vocabulary cluster and the Hebrew fourth grade spelling cluster together accounted for 25% of the variance. The Hebrew vocabulary cluster (b ¼ 0.44), F (1, 76) ¼ 18.17, p < 0.001 accounted for 19% of the variance and the Hebrew spelling cluster (b ¼ 0.29), F change (1, 75) ¼ 6.32, p < 0.05 accounted for a further 6% of the variance.

2.3.6 L1 Measures in Fourth and Ninth Grade Predicting EAL Reading Comprehension in Ninth Grade The stepwise regression analysis yielded a three-variable model whereby the Hebrew spelling task, the phonological awareness task and the ninth grade rapid automatic naming task of Hebrew letters together accounted for 37% of the variance. The Hebrew spelling task (b ¼ 0.49), F (1, 75) ¼ 23.62, p < 0.001 explained 24% of the variance. The rapid automatic naming task of Hebrew letters (b ¼0.26), F change (1, 74) ¼ 7.12, p < .01 explained an additional 7% of the variance and the phonological awareness task (b ¼ 0.29), F change (1, 73) ¼ 7.14, p < 0.01 explained a further 4% of the variance.

2.3.7 L1 and EAL Measures in Fourth and Ninth Grade Predicting EAL Word Reading Accuracy in Ninth Grade The stepwise regression analysis yielded a six-variable model whereby the overall English spelling measure, the overall Hebrew phonological awareness measure, the ninth grade Hebrew rapid automatic naming task, the ninth grade Hebrew word attack (speed) task, the overall Hebrew fourth grade morphological awareness task and the English fourth grade reading comprehension task

2

Relevance of the Linguistic Coding Difference Hypothesis to English . . .

35

together accounted for 60% of the variance. The overall English spelling measure (b ¼ 0.62), F (1, 75) ¼ 48.75, p < 0.001 accounted for 39% of the variance, the overall Hebrew phonological awareness measure (b ¼ 0.35), F change (1, 74) ¼ 13.27, p < 0.001 accounted for an additional 9% of the variance. The ninth grade Hebrew rapid automatic naming task (b ¼ 0.22), F change (1, 73) ¼ 4.07, p < 0.05 accounted for an additional 3% of the variance. The ninth grade Hebrew word attack (speed) task (b ¼ –0.18), F change (1, 72) ¼ 4.70, p < 0.05 accounted for a further 3% of the variance. The Hebrew fourth grade morphological awareness measure (b ¼ –0.23), F change (1, 71) ¼ 4.30, p < 0.05 accounted for a further 3% of the variance. Finally, the English fourth grade reading comprehension task (b ¼ 0.23), F change (1, 70) ¼ 4.98, p < 0.05 accounted for an additional 3% of the variance.

2.3.8 L1 and EAL Measures in Fourth and Ninth Grade Predicting EAL Speeded Word Reading in Ninth Grade The stepwise regression analysis yielded a one-variable model whereby rapid automatic naming of English letters measured in ninth grade accounted for 25% of the variance (b ¼ 0.50), F (1, 74) ¼ 25.11, p < 0.001.

2.3.9 L1 and EAL Measures in Fourth and Ninth Grade Predicting EAL Reading Comprehension in Ninth Grade The stepwise regression analysis yielded a three-variable model whereby the overall English vocabulary measure, the ninth grade rapid automatic naming task of English letters and the overall phonological awareness task (measured in Hebrew) together accounted for 62% of the variance. The English vocabulary task (b ¼ 0.54), F (1, 75) ¼ 30.88, p < 0.001 explained 29% of the variance. English word reading speed (b ¼ 0.36), F change (1, 74) ¼ 15.85, p < .001 explained a further 13% of the variance. English word reading accuracy (b ¼ 0.47), F change (1, 73) ¼ 29.00, p < 0.001 explained a further 17% of the variance. Finally, Hebrew word reading (b ¼ 0.19), F change (1, 72) ¼ 6.20, p < 0.01 explained 3% of the variance.

2.4 Discussion The research question in this longitudinal study of ninth grade EAL students examined the extent to which first language (Hebrew) and additional language (English) linguistic and literary abilities measured up to 6 years earlier and concurrently predicted EAL word recognition and reading comprehension.

36

J. Kahn-Horwitz et al.

2.4.1 The Relevance of the LCDH to EAL Word Recognition from a Longitudinal Perspective Two measures were constructed for measuring EAL word recognition at the end of 9th grade. One measure was a cluster of English word reading measures including speeded reading and the other was the cluster of English word reading independent of speeded reading. A rationale for splitting these two abilities is suggested by Hagiliassis et al. (2006), who provide evidence that supports a separate construct being measured when examining task accuracy as opposed to task response time. Support for this rationale in the present research was provided by insignificant correlations between RAN and phonological awareness as well as word attack speed and phonological awareness. Phonological awareness measured in L1 Hebrew in fourth grade accounted for 25% of the explained variance of EAL word recognition at the end of ninth grade. This finding provides evidence for cross-language transfer of phonological awareness skills as predicting word reading accuracy that has been supported by other research (Comeau et al., 1999; Jongejan et al., 2007; Kahn-Horwitz et al., 2005; Nakamato et al., 2007). In the case of the present study there is a longitudinal dimension of almost 6 years to the transfer. Morphological awareness measured in L1 Hebrew in fourth grade accounted for 6% of the explained variance of EAL speeded word reading. In other words, sensitivity towards the morphological structure of the L1 directly contributed almost 6 years later to the speed of English word reading. When ninth grade L1 predictors were added to the fourth grade L1 Hebrew predictors, a combination of phonological awareness measured in L1 Hebrew together with Hebrew word reading accounted for 42% of the explained variance. These results are remarkably similar to Sparks et al. (2008) findings in their longitudinal study that 43% of the explained variance of word reading in Spanish, German and French was explained by English L1 word reading. The finding in the present study that most of the 42% of explained variance comprised of phonological awareness measured in Hebrew may point to the differences in word recognition strategies in two opaque orthographies with different characteristics; that is, the opaque Hebrew orthography lacks complete phonological information as opposed to the opaque English orthography, which shows a relative degree of inconsistency and irregularity (Frost, 2005). It could be that due to the characteristics of the English orthography the ninth graders use a phonological strategy to decode the EAL words that they used when originally decoding voweled Hebrew words with their diacritics appearing above, below or inside the consonant letters. This use of a phonological strategy may occur despite the phonemic differences between English and Hebrew vowels which may require different processing strategies. When ninth grade L1 predictors were added to the fourth grade L1 predictors, RAN of Hebrew letters accounted for 11% of the explained variance of the English speeded word reading measure. This could be evidence of the direct impact of a speeded naming task on a speeded word reading task from a crosslanguage perspective.

2

Relevance of the Linguistic Coding Difference Hypothesis to English . . .

37

When English predictors were added to the fourth and ninth grade Hebrew predictors, 60% of the explained variance of English word reading accuracy was explained by the English spelling cluster, which contributed 39% of the 60% explained variance. This finding reflects the strong connection between spelling and reading in the opaque English orthography. In contrast with the lack of correlation found between Hebrew spelling and Hebrew word recognition in the earlier grades, this finding may provide support for Hebrew spelling being more opaque than Hebrew voweled reading (Geva et al., 1993). Spelling is the more productive expression of decoding. Spelling is also strongly associated with meaning and this is particularly salient in an opaque orthography such as English. English spelling is considered an extremely challenging task for additional language learners. Because good reading comprehension is considered one of the primary tasks of EAL acquisition in Israel, spelling instruction is often considered less of a priority. The above mentioned results confirm the interdependence of word reading and spelling and suggest that these skills are two sides of the same coin and may strengthen one another. Phonological awareness measured in Hebrew once again appeared as a significant contributor of a further 9% of the explained variance. The finding that rapid automatic naming of Hebrew letters and speed of non-word reading in Hebrew contributed small but significant amounts (3% each) of the variance suggests cross language transfer of speeded naming and word reading for accurate English word recognition. Morphological awareness measured in Hebrew in fourth grade contributed a further 3% of the variance, which combines with phonological awareness and the two speeded measures in Hebrew to support the relevance of the LCDH, or cross-language transfer of L1 abilities from a longitudinal and concurrent perspective. In other words, underlying L1 abilities still predict English word reading accuracy 6 years after the beginning of English acquisition. These factors might reflect Koda’s (2005) description of transferred L1 competencies continuing to develop and impacting on L2 word recognition processing with its respective orthographic specifications. The final predictor of English word reading accuracy at the end of 9th grade was English reading comprehension measured at the end of 4th grade. This English reading comprehension task could represent the semantic dimension of word recognition in an opaque orthography whereby correct word identification is directly connected to a lexical association. This finding may be seen as evidence of a bootstrapping process whereby fourth grade English reading comprehension predicted ninth grade English word recognition, which subsequently impacts on higher level English reading comprehension. This aforementioned model possibly illustrates the interface of L1 and EAL components underlying EAL word recognition (Koda, 2005). The analysis of English and Hebrew predictors of the speeded English word reading task yielded a one variable model in which rapid automatic naming of English reading accounted for 25% of the explained variance. In this case because English predictors were added into the regression, their predictability increased and they predicted greater variance than when the Hebrew variables

38

J. Kahn-Horwitz et al.

were entered alone (in that case, rapid automatic naming of Hebrew letters accounted for 11% of the explained variance).

2.4.2 The Relevance of the LCDH to EAL Reading Comprehension from a Longitudinal Perspective Hebrew vocabulary knowledge together with Hebrew spelling ability measured in fourth grade accounted for 25% of the explained variance of EAL reading comprehension. These results show an interesting variation of support for the simple model of reading (Koda, 2005) from a cross language and longitudinal perspective. Hebrew vocabulary knowledge, which accounted for 19 out of the 25% of variance, represents L1 understanding which longitudinally shows transfer effects to the additional language for the purpose of reading comprehension. Hebrew spelling ability, which represents the more opaque dimension of Hebrew literacy (see Geva et al., 1993) may tap a similar cognitive functioning for reading an opaque orthography that is required for automatic English word recognition when reading a text. When ninth grade L1 Hebrew measures were added to the fourth grade L1 Hebrew measures, 37% of the variance was jointly accounted for by Hebrew spelling (24%), phonological awareness (4%) and RAN of Hebrew letters (7%). These results support the relevance of the LCDH in that EAL reading comprehension was predicted by the components that comprise accuracy (phonological awareness), fluency (RAN) and spelling which in Hebrew implicates orthographic, morphological and semantic knowledge, all from a cross linguistic perspective. Sparks et al. (2008) found that reading comprehension in English L1 accounted for 25% of the explained variance of Spanish, German and French reading comprehension. Their model comprised one component whereas the findings in this study comprised various components. When English predictors were added to the fourth and ninth grade Hebrew variables, 62% of the explained variance of EAL reading comprehension was explained by English vocabulary knowledge (29%), English word reading speed (13%), English word reading accuracy (17%) and Hebrew word reading (3%). In other words, when English predictors were added to the original L1 predictors, English vocabulary knowledge together with English word reading (accuracy and speed) were the main predictors of English reading comprehension supporting a simple model of reading (Koda, 2005). The impact of Hebrew remains whereby Hebrew word reading is a significant predictor of EAL reading comprehension. In sum, the present research provides longitudinal support for the Linguistic Coding Differences Hypothesis. Fourth grade Hebrew L1 measures when entered without English measures predicted EAL word recognition or reading comprehension. When entered together with ninth grade Hebrew measures, their prediction value increased.

2

Relevance of the Linguistic Coding Difference Hypothesis to English . . .

39

2.4.3 The Role of Phonological Awareness in Predicting EAL Word Recognition and Reading Comprehension It is of interest that at this more advanced stage of EAL literacy development, a phoneme deletion task measured in L1 should have still remained a predictor. It may be that because we are looking at Hebrew L1 students whose L1 processing experience involves reading an opaque orthography with incomplete phonological information and are acquiring and developing their abilities in an opaque orthography with irregular and inconsistent features, phonological awareness remains a predictor of EAL literacy abilities. More finely grained studies looking at these cross-orthographic differences comparing EAL readers from varied orthographic backgrounds may shed light on this question.

2.5 Conclusion This research supports the LCDH as a viable hypothesis for predicting EAL literacy from a longitudinal perspective. When coupled with EAL predictor variables, there is greater explained variance for word reading and reading comprehension, which is accounted for by a mixture of L1 and EAL component measures. The simple model of reading is supported by this research whereby EAL reading comprehension is predicted by concurrent EAL vocabulary knowledge, EAL word recognition as well as cross-linguistic transfer of Hebrew word recognition. Regarding the connection between spelling and word recognition, in the present research, the students were not first language English speakers but rather EAL students, which underscores the importance of developing good spelling for efficient word recognition and understanding. An important educational implication from this finding is that more emphasis should be placed on developing English spelling in the EAL classroom. In Israel, this skill is often a neglected pedagogical dimension, particularly because reading is considered more of a priority and English writing is considered secondary, a luxury of the more advanced learner. The results of this research demonstrate the strong interdependence between the two.

References Adams, M. J. (1990). Beginning to read. Cambridge: MIT Press. Assis, E. (2002). Horizons. Raanana, Israel: Eric Cohen Books. Baddeley, A. D. (2006). Working memory: An overview. In S. Pickering (Ed.), Working memory and education (pp. 1–31). New York: Academic Press. Balgur, R. (1977). Mivxan kriya diagnosti letalmidey kitot bet-yud [Diagnostic reading test for pupils in second to tenth grade]. Tel Aviv Israel: Shockan.

40

J. Kahn-Horwitz et al.

Ben-Dror, I., Bentin, S., & Frost, R. (1995). Semantic, phonologic and morphologic skills in reading disabled and normal children: Evidence from perception and production of spoken Hebrew. Reading Research Quarterly, 30, 876–893. Carlisle, J. F., Beeman, M., Hull Davis, L., & Sphraim, G. (1999). Relationship of metalinguistic capabilities and reading achievement for children who are becoming bilingual. Applied Psycholinguistics, 20, 459–478. Cisero, C. A., & Royer, J. M. (1995). The development and cross-language transfer of phonological awareness. Contemporary Educational Psychology, 20, 275–303. Comeau, L., Cormier, P., Grandmaison, E., & Lacroix, D. (1999). A longitudinal study of phonological processing skills in children learning to read in a second language. Journal of Educational Psychology, 91, 29–43. Cummins, J. (1984). Bilingualism and special education: Issues in assessment and pedagogy. Austin, TX: Pro-Ed. Dufva, M., & Voeten, M. J. M. (1999). Native language literacy and phonological memory as prerequisites for learning English as a foreign language. Applied Psycholinguistics, 20, 329–348. Durgunoglu, A. Y., Nagy, W. E., & Hancin-Bhatt, B. J. (1993). Cross-language transfer of phonological awareness. Journal of Educational Psychology, 85, 453–465. 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. Ezra, E., & Kerman, S. (1998). Flying colors. Raanana, Israel: Eric Cohen Books. Frost, R. (2005). Orthographic systems and skilled word recognition process in reading. In M. J. Snowling, & C. Hulme (Eds.), The science of reading: A handbook (pp. 272–295). Oxford, UK: Blackwell. Gathercole, S. E., & Thorn, A. S. C. (1998). Phonological short-term memory and foreign language learning. In A. F. Healy & L. E. Bourne Jr. (Eds.), Foreign language learning: Psycholinguistic studies on training and retention (pp. 141–160). Hillsdale, NJ: Erlbaum. Geva, E., & Siegel, L. (2000). Orthographic and cognitive factors in the concurrent development of basic reading skills in two languages. Reading and Writing, 12, 1–30. Geva, E., Wade-Woolley, L., & Shany, M. (1993). The concurrent development of spelling and decoding in two different orthographies. Journal of Reading Behavior, 25, 383–406. Gholomain, M., & Geva, E. (1999). Orthographic and cognitive factors in the concurrent development of basic reading skills in English and Persian. Language Learning, 49, 183–217. Glantz, I. (1991). Xemed, ma’arexet l’bdika m’kifa, [Xemed – A comprehensive testing battery]. Tel Aviv: Barak Information Systems. Greenbaum, N., & Lichter, Y. (1996). Milot tefel [Hebrew word attack sub-test]. In N. Greenbaum & Y. Lichter (Eds.), Ivxun kri’a ktiva [Reading and writing diagnoses] (p. 8). Oranim: Oranim College. Hagiliassis, N., Pratt, C., & Johnston, M. (2006). Orthographic and phonological processes in reading. Reading and Writing, 19, 235–263. Jongejan, W., Verhoeven, L., & Siegel, L. S. (2007). Predictors of reading and spelling in firstand second-language learners. Journal of Educational Psychology, 99, 835–851. Kahn-Horwitz, J., Shimron, J., & Sparks, R. L. (2005). Predicting foreign language reading achievement in elementary school students. Reading and Writing, 18, 527–558. Kahn-Horwitz, J., Shimron, J., & Sparks, R. L. (2006). Weak and strong novice readers of English as a foreign language: Effects of first language and socioeconomic status. Annals of Dyslexia, 56, 161–186. Kail, R., & Hall, L. K. (1994). Processing speed, naming speed, and reading. Developmental Psychology, 30, 949–954. Koda, K. (1992). The effects of lower-level processing skills on FL reading performance: Implications for instruction. The Modern Language Journal, 76, 502–512.

2

Relevance of the Linguistic Coding Difference Hypothesis to English . . .

41

Koda, K. (1995). Cognitive consequences of L1 and L2 orthographies. In I. Taylor & D. R. Olson (Eds.), Scripts and literacy (pp. 311–326). Netherlands: Kluwer. Koda, K. (1996). Second language word recognition research: A critical review. Modern Language Journal, 80, 450–460. Koda, K. (2005). Insights into second language reading: A cross-linguistic approach. New York: Cambridge University Press. Leikin, M., Share, D. L., & Schwartz, M. (2005). Difficulties in L2 Hebrew reading in Russian-speaking second graders. Reading and Writing, 18, 455–472. Ministry of Education, State of Israel. (2001). English curriculum: Principles and standards for learning English as a foreign language in Israeli schools. Jerusalem: Ministry of Education. Muter, V., & Diethelm, K. (2001). The contribution of phonological skills and letter knowledge to early reading development in a multilingual population. Language Learning, 5, 187–219. Nakamato, J., Lindsey, K. A., & Manis, F. R. (2007). A longitudinal analysis of English language learners’ word decoding and reading comprehension. Reading and Writing, 20, 691–719. Nassaji, H., & Geva, E. (1999). The contribution of phonological and orthographic processing skills to adult ESL reading: Evidence from native speakers of Farsi. Applied Psycholinguistics, 20, 241–267. Oloffson, A., & Niedersoe, J. (1997). Early language development and kindergarten phonological awareness as predictors of reading problems: From 3 to 8 years of age. In C. K. Leong & R. M. Joshi (Eds.), Cross-language studies of learning to reading and spell (pp. 289–306). Netherlands: Kluwer. Parel, R. (2004). The impact of lexical inferencing strategies on second language reading proficiency. Reading and Writing, 17, 847–873. Seymour, P. H. K., Aro, M., & Erskine, J. M. (2003). Foundation literacy acquisition in European orthographies. British Journal of Psychology, R94, 143–174. Shany, M., Zieger, T., & Ravid, D. (2001). Pitu’ax v’tikuf el kley ivxun R l’tahalixim bsisim b’kriya uv’xtiv: Mem’tsa’im R al tif’kud kor’im t’kinim v’hatsa’ot l’ iluv kley ha’ivxun b’ha’araxot kor’im mit’ka im [The development and validation of tests for basic processes in reading and spelling: Performance of normally-achieving readers at various grade levels and implications for the assessment of individuals with reading disabilities]. Script: Literacy Research, Theory and Practice, 2, 167–203. Share, D. L. (1995). Phonological recoding and self-teaching: Sine qua non of reading acquisition. Cognition, 55, 151–218. Share, D. L., & Stanovich, K. (1995). Cognitive processes in early reading development: Accommodating individual differences into a model of acquisition. Issues in Education, 1, 1–57. Sparks, R. (1995). Examining the linguistic coding differences hypothesis to explain individual differences in foreign language learning. Annals of Dyslexia, 45, 187–214. Sparks, R., & Ganschow, L. (1993a). Searching for the cognitive locus of foreign language learning difficulties: Linking native and foreign language learning. Modern Language Journal, 77, 289–302. Sparks, R., & Ganschow, L. (1993b). The impact of native language learning problems on FL learning: Case study illustrations of the linguistic coding deficit hypothesis. Modern Language Journal, 77, 58–74. Sparks, R., Ganschow, L., & Patton, J. (2008). L1 and L2 literacy, aptitude, and affective variables as discriminators among high and low-achieving L2 learners. In J. Kormos & E. Kontra (Eds.), Language learners with special needs: An international perspective (pp. 11–35). London: Multilingual Matters. Sparks, R., Patton, J., Ganschow, L., & Humbach, N. (2009). Long-term cross linguistic transfer of skills from L1 to L2. Language Learning, 59, 203–243.

42

J. Kahn-Horwitz et al.

Sparks, R., Patton, J., Ganschow, L., Humbach, N., & Javorsky, J. (2008). Early firstlanguage reading and spelling skills predict later second-language reading and spelling skills. Journal of Educational Psychology, 100, 162–174. 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. van Gelderen, A., Schoonen, R., Stoel, R. D., & de Glopper, K. (2007). Development of adolescent reading comprehension in language 1 and language 2: A longitudinal analysis of constituent components. Journal of Educational Psychology, 99, 477–491. Wade-Woolley, L., Chiappe, P., & Siegel, L. (1998). Learning to read in a second language: Does phonological awareness matter? Paper presented to the fifth annual meeting of the Society for the Scientific Study of Reading, San Diego, CA. Woodcock, R. (1987). Woodcock reading mastery test-revised. Circle Pines, MN: American Guidance.

Chapter 3

Literacy Reflexes of Arabic Diglossia Elinor Saiegh-Haddad

3.1 Introduction The term ‘‘diglossia’’ emerged in sociolinguistic theory to describe a situation where in a given society there is more than one language variety in use. In what is probably the earliest use of the term, Karl Krumbacher (1902, cf. Sotiropoulos, 1977) used the term diglossia, a translation from Greek meaning ‘‘bilingualism’’, to describe the situation of Greek, with Dimotiki (popular language, or people’s language) as the medium for everyday conversation and Katharevusa (pure language) as the written medium. William Marcais (1930, cf. Zughoul, 1980) then used the term to describe the linguistic situation in the Arabic speaking world and the remarkable gulf between spoken Arabic vernaculars and the written standard- Classical Arabic. The term was publicized in a famous 1959 Word article by the linguist Charles Ferguson (1959) who offered the first coherent theory of diglossia referring to Arabic as a typical example. Ferguson proposes a classification of diglossic features that include function, prestige, literary heritage, acquisition, standardization, stability, grammar, lexicon, and phonology. According to Ferguson, a diglossic context is characterized by a stable co-existence of two linguistically-related language varieties, a High variety and a Low variety, which are used for two sets of complementary social functions; Such rigid functional complementarity, it is argued, should give way only to slight and insignificant overlap (Maamouri, 1998). In a diglossic context, no section of the community uses the High variety for ordinary conversation. This is arguably ‘‘the most important factor in a diglossic situation and one that makes for relative stability’’ (Keller, 1982, p. 90). In a diglossic context, the two language varieties are associated with different degrees of prestige. The High variety is considered prestigious and elegant; it is used for formal linguistic functions, such as religious sermons, news broadcasts, speeches, etc, and is the only written code. The Low variety E. Saiegh-Haddad (*) Department of English Linguistics, Bar-Ilan University, Ramat-Gan, Israel e-mail: [email protected] M. Leikin et al. (eds.), Current Issues in Bilingualism, Literacy Studies 5, DOI 10.1007/978-94-007-2327-6_3, Ó Springer ScienceþBusiness Media B.V. 2012

43

44

E. Saiegh-Haddad

is stigmatized as a literary from and is an exclusively spoken code. The High variety is usually associated with a rich literary tradition and is the standardized form that is encoded in dictionaries and books. As it is only a written variety, the High variety is nobody’s mother tongue. Instead, it is the Low code that children acquire as their first language. Stability is another important feature of diglossia as a long-lived phenomenon. Finally, in a typical diglossic context, the two varieties of the language are linguistically related, though the phonology, morphology, and syntax of the Low variety are often simpler than that of the High variety. A diglossic context is also characterized by the existence of a large set of paired lexical items, which are available in both codes but which, though linguistically related, are phonologically distinct. Fishman (1967) has reformulated and extended the term ‘‘diglossia’’. According to Fishman, diglossia does not necessarily encompass two varieties of a language, a High standard and a Low spoken vernacular, but may also be extended to refer to situations where two dialects or registers are used for socially distinct sets of functions. Classical examples include the Standard German-Swiss German context of Switzerland and the Creole-French context in Haiti. The hallmark of this new definition is that the two codes are functionally different. In yet another extension of the term, Fishman (1991), uses the concept diglossia to refer to bilingual communities. Such reformulation encompasses under the term ‘‘diglossia’’ many other bilingual contexts, including the English–Latin context of Medieval England. While the broader meaning of the term ‘‘diglossia’’ has gained widespread currency, scholars warn against such a mix of terms. Hudson (1991), for instance, argues that Fishman’s (1991) reformulation is problematic as the direction of language evolution in a classic diglossic context is opposite to that in the case of widespread bilingualism. In other words, while in a diglossic context, the Low variety, or at least some linguistic aspects of the Low variety take over the outdated High variety, in a bilingual context the Low variety loses ground to the superposed High variety. In a more recent account, Hudson (2002) stresses the need for sociolinguistic theory to distinguish typical diglossic contexts from standard-with-dialect contexts or contexts of societal bilingualism. Hudson points to a central feature unique to diglossia. He argues that, unlike other contexts, societal bilingualism or standard-with-dialects, stratification of variation in diglossia shows sensitivity to differences in situational context (use-oriented) without much sensitivity to differences in social class (user-oriented) (Bell, 1984; Walters, 1994). In the same vein, Anghelescu (1974) had warned that such terminological mix disguises the true nature of diglossia and reduces it to ‘‘stylistic functional variation’’. According to Anghelescu, ‘‘diglossia implies sufficiently similar languages for the speakers to feel that it is the same language, yet remote enough, so that the acquisition of the literary language implies long-term efforts and can never be fully achieved’’ (cf. Maamouri, 1998, p. 30).

3

Literacy Reflexes of Arabic Diglossia

45

3.2 Arabic Diglossia Perhaps one of the most compelling features of Arabic, especially in the context of literacy acquisition, is its diglossic nature (Al-Batal, 1992; Al-Toma, 1969; Haeri, 2000). The most important feature of diglossia, and one that has direct implications for literacy acquisition in Arabic, is the establishment of rigid and complementary functional separation of two linguistic codes: the written and the spoken. According to Hymes (1973), the dominant practical response to linguistic diversity is ‘‘to impose a novel unity in the form of the hegemony of one language or standard’’ (p. 22). The standardization of Arabic, which began in the 8th and 9th centuries A.D., has produced a set of norms that the early grammarians called fusha (Maamouri, 1998). Over the course of many years, the continued use of this favored set of written linguistic norms led to substantial differences between the dynamic spoken vernaculars and the fixed written form, making the two varieties distinct and linguistically distant, and engendered the notion that the written standard was the ‘‘real language’’ while the other varieties of it were ‘‘degenerate’’ and ‘‘corrupt’’ versions (Maamouri, 1998). The term fusha /fusħa/ is currently used to refer both to Classical Arabic (the language of the Qur’an), and to a modern descendant of Classical Arabic, Modern Standard Arabic (hereafter, MSA, or Standard Arabic), which serves as the modern written standard and the variety used in formal oral discourse. In contrast, the term ‘‘The Arabic Language’’, al’arabiya, is an abstraction that refers to the written standard(s), as well as the sum of all the varieties in use in the Arab region (Maamouri, 1998). Another unique feature of diglossia that has direct ramifications for literacy development in Arabic is acquisition. While all Arab children naturally acquire the local spoken vernacular of the speech community to which they belong, learning the written standard, i.e., MSA, comes about mainly as a result of formal instruction in reading. Thus, MSA is nobody’s mother tongue (Coulmas, 1987; Joseph, 1987). It is typically never spoken at home or in the neighborhood. It is argued that the fact that the functional distribution of codes in a diglossic context protects the role of the Low variety as a natively learned variety is what distinguishes diglossia from other interlingual or intralingual situational alternation (Hudson, 2002, p. 7). This, together with the linguistic distance between Spoken Arabic and the linguistic system encoded in print (Standard Arabic), implies a series of challenges to the acquisition of literacy in Arabic. In the next section we provide a linguistic description of Arabic diglossia. Then we review recent psycholinguistic evidence in support of the impact of diglossia on the acquisition of basic literacy processes in Arabic.

3.3 The Linguistic Distance in Diglossic Arabic All spoken Arabic vernaculars are linguistically related to Standard Arabic. Yet, the linguistic relatedness between Spoken and Standard Arabic is ‘‘flexible’’ and ‘‘changeable’’ (Kaye, 1972). Despite linguistic relatedness between the two linguistic codes, a linguistic analysis of the structure of Standard Arabic and

46

E. Saiegh-Haddad

any given spoken Arabic vernacular reveals differences in all language domains of language and most clearly in the lexicon, the phonology and the morphosyntax. In the following section we will focus on the phonological and lexical distance between Spoken and Standard Arabic. A major manifestation of the linguistic distance between Standard and Spoken Arabic exists in the phonemic inventories of the two language varieties. Although Standard Arabic shares most of the phonemes with all spoken vernaculars, no single Spoken Arabic vernacular has the same set of phonemes as Standard Arabic (Maamouri, 1998). For instance, while the phonemic inventory of MSA includes voiced, voiceless, and emphatic interdental fricatives, /ð/, / y/, /ð/, these phonemes are absent from a local dialect of Northern Palestinian Arabic spoken in the north of Israel. In this dialect, then, paired lexical items existing in both Spoken and Standard Arabic have different phonological structures in the two varieties, with Standard phonemes systematically substituted for by Spoken phonemes. For instance, the Standard word /ðakar/, meaning ‘‘male’’, becomes /dakar/ in this dialect of Spoken Arabic. Another phonological disparity between Standard and Spoken Arabic pertains to the internal phonological composition of the syllable. The internal phonological structure of the syllables in Standard and Spoken Arabic is subject to different phonotactic constraints. For instance, while word initial (onset) consonantal clusters are not permissible in Standard Arabic, they are frequent in many Spoken Arabic vernaculars. As a result, the Standard word /tura : b/, meaning ‘‘soil’’, becomes /tra: b/ in some vernaculars. At the same time, while word final (coda) consonantal clusters are rare in spoken Arabic vernaculars, it is a very common monosyllabic word structure in Standard Arabic. This explains the reason that the Standard word /sah|/, meaning ‘‘plain’’, becomes /sahi|/ in Spoken Arabic, or that the Standard word /ya|Z/, meaning ‘‘snow’’, becomes /ta|iZ/, respectively, with both the segmental and the syllabic structure of the word altered in accordance with the phonological constraints of the specific spoken vernacular. The above illustration of some of the phonological distance parameters between Standard and Spoken Arabic (for a more detailed description see Levin, 1994) suggests that even for shared vocabulary items (so-called paired lexical items) there may be a substantial phonological difference between their forms in Standard and Spoken Arabic. Paired lexical items (cognate words) were found by a recent study to make up about 40% of the lexicon of 5-year-old Arabic native speaking children (Saiegh-Haddad, ms.). This study also showed that the next largest category in the lexicon of children is Spoken Arabic unique words. That is, words, which exist in Spoken Arabic only and have a completely different form in Standard Arabic. This category was found to make up about another 40% of the children’s lexicon leaving only 20% overlapping words. These findings support the remarkable linguistic distance between Spoken and Standard Arabic.

3

Literacy Reflexes of Arabic Diglossia

47

3.4 The Acquisition of Basic Reading Processes in Diglossic Arabic Learning to read in Arabic, and owing to the diglossic context explicated earlier, requires the acquisition of two systems: an orthographic system and a linguistic system. Perhaps the first and most basic linguistic system that children must become familiar with in order to embark upon the process of reading acquisition is the phonological system of the language of literacy. Phonological skills are fundamental to the acquisition of reading as they are directly implicated in the execution of two critical word reading processes: phonological segmentation and phonological recoding. To be able to read, children must first become aware of the phonological structure of spoken words (for reviews, see Adams, 1990; Goswami & Bryant, 1990; National Reading Panel, 2000). Phonological awareness underlies the ability to segment words into phonemes. Having segmented the word, the reader is then required to translate the graphemes of written words into the corresponding oral language phonemes. Such a phonological recoding process is the hallmark of word decoding in an alphabetic orthography. If a basic prerequisite to reading is for beginning readers to become aware of the phonological structure of the written language, the question that looms large relates to the acquisition of reading in the absence of such a foundational infrastructure. In other words, how do beginning readers develop reading in a context where reading is first learned in a language that is phonologically distinct from the oral language they have already acquired, as is the case with Arabic native speaking children learning to read in MSA? Does the phonological distance between the two varieties of the language, the spoken vernacular and the written standard, disrupt the natural acquisition of basic reading processes? These questions have guided the research reviewed in this paper. Saiegh-Haddad (2003, 2004, 2007) tested the effect of the linguistic distance (phonological and lexical) between Standard and a local Arabic vernacular spoken in the north of Israel on phonological awareness among kindergarten, 1st, 2nd, and 3rd graders. These studies showed that the linguistic affiliation of the target phoneme (Standard versus Spoken) reliably predicted children’s phoneme isolation performance in kindergarten through the second grade. This effect fell below statistical significance in the third grade, 3 years after extensive formal exposure to Standard Arabic. As such, both kindergarten, first-grade, and second-grade children found Standard phonemes, and despite proper articulation, significantly harder to isolate than Spoken phonemes. This factor was found to interact with grade, with the linguistic affiliation of the syllable in which the phoneme was embedded, and with the lexical status of the word resulting in particular difficulty when kindergarteners were asked to isolate a Standard phoneme from a Standard syllable, or from a Standard word. These results were argued to be attributed to a deficiency in the phonological representations of Standard linguistic structures (phonemes, syllables, and words) which appears to disrupt phonological analysis (Elbro, 1996, 1998; Goswami, 2000, 2002; Katz, 1986).

48

E. Saiegh-Haddad

The finding that the absence of some phonemes, syllabic structures, and words from the spoken vernacular of children remains significant in explaining phoneme isolation performance in Standard Arabic, even at the end of the second grade, has serious implications for the effect of the linguistic profile of the spoken vernacular of children, and of the linguistic distance between this variety and the linguistic system encoded in print, on the development of phonemic awareness in the language of literacy. The results also have serious ramifications for the cross-dialectal validity of phonological sensitivity tasks in MSA, and imply that a thorough analysis of the linguistic distance between the specific spoken vernacular and MSA is warranted before any such task is designed. Though the linguistic affiliation of the phonological structure was found to impact the phoneme isolation performance of preliterate kindergarteners and literate children in the early grades, the latter were found to significantly outperform the former. This implies that though first grade children have somewhat benefited from the increased exposure and experience with Standard Arabic structures that formal literacy instruction allows, they were still facing particular difficulty when confronted with the task of isolating standard phonological structures. This finding demonstrates one of the unique complexities present in learning to read in Standard Arabic. While in other linguistic contexts, beginning readers come to the reading task with ample knowledge and experience with the linguistic structures of the language of literacy to embark upon the task of learning to decode oral language linguistic structures, Arabic native children are required to acquire both the linguistic system of the written language and its orthographic representation. Acquiring the linguistic system of the language of literacy requires the construction of proper phonological representations for the novel linguistic units that the written language encodes. Given consistent evidence in favor of the impact of the diglossic linguistic distance on children’s phonological sensitivity for contrastive Standard linguistic structures, the question that naturally follows relates to the impact of the linguistic distance on the phonological recoding, or the decoding skills of Arabic native children. This question was addressed in Saiegh-Haddad (2003, 2005). These studies tested the pseudoword decoding accuracy and fluency (speed) of first grade children and showed that the linguistic affiliation of the phonemes that make up pseudowords reliably explained the decoding accuracy of children at the end of the first grade. As such, the results showed that, despite generally high rates of decoding accuracy, a large percentage of the decoding errors committed by children were related to Standard phonemes and Standard syllabic structures. Furthermore, children’s errors were primarily phonetic and constituted primarily of phonologically similar pseudowords that reflected difficulty in the phonological recoding of novel phonological structures. In contrast, Spoken Arabic phonemes and Spoken syllabic structures that were available in the oral language of children were almost perfectly decoded. With respect to decoding fluency, Saiegh-Haddad (2005) found that speed of converting letters into sounds (letter recoding speed) along with memory and rapid

3

Literacy Reflexes of Arabic Diglossia

49

automatized naming were the strongest predictors. Given the transparent orthography of voweled Arabic, the high decoding accuracy that children had revealed was not surprising. As such, the regular and consistent orthographic representation of oral language phonemes make the phonological recoding of those graphemes a straightforward task (Seymour et al., 2003) and encourage an analytic, phonological assembly decoding mechanism. Given this decoding strategy, and as some of the graphemes (primarily single letters) map onto novel phonemes, it follows, and as the above study has shown, that the great majority of the children’s decoding errors would be specifically bound to novel phonological units. Hence, two sets of forces operate in the acquisition of reading fluency in voweled Arabic. The first is the shallow orthography. The second is the diglossic context. Arabic pseudoword reading fluency, at the end of the first grade, was best predicted by letter recoding speed, memory, and rapid naming. These findings are consistent with earlier evidence supporting the role of such cognitive skills to reading in general (Abu-Rabia & Siegel, 2002; Bowers & Wolf, 1993; Geva & Siegel, 2000; McBride-Chang, 1995; Wolf & Bowers, 1999; Wolf, Bowers, & Biddle, 2000; Wolf et al., 2002), and to reading in a shallow orthography, in particular (de Jong & van der Leij, 1999; Landerl & Wimmer, 2000; Wimmer, Mayringer, & Landerl, 1998, 2000). Yet, the prediction that phonological processing (phoneme discrimination and phoneme isolation), especially for Standard phonemes, would turn out a significant correlate of reading fluency in diglossic Arabic was not supported. Although phoneme isolation for Standard phonemes was more difficult than that for Spoken phonemes, the study revealed only a moderate correlation between phoneme isolation and reading fluency. Further, this relationship disappeared when rapid naming, or when letter recoding speed, along with memory, were entered into the regression equation. Though inconsistent with the prediction that phonological awareness in Arabic, and because of the phonological distance between Standard and Spoken Arabic, would turn out a strong correlate of reading fluency, this finding is supportive of recently emerging evidence demonstrating significant shrinkage in the power of phonemic awareness in predicting individual differences in reading fluency in a shallow orthography (de Jong & van der Leij, 1999; Landerl & Wimmer, 2000; Wimmer et al., 1998, 2000). In a recent study (Saiegh-Haddad, Levin, Hende, & Ziv, 2011), we tested the effect of the phonological distance in diglossic Arabic on kindergarten children’s phoneme recognition, a task that does not require phonological production. Using a picture selection task and through careful manipulation of the phonological properties of target phonemes and distractors, the study showed that children’s recognition of Standard phonemes was poorer than that of Spoken phonemes. This finding was interpreted as indicating a deficiency in the phonological representations of Standard words. The study also tested two hypotheses regarding the specific consequences of under-specified phonological representations: phonological encoding versus phonological processing. These hypotheses were addressed through an analysis of the relative power of

50

E. Saiegh-Haddad

distractors. The findings revealed that children’s difficulty in accessing Standard Arabic phonemes was due to a difficulty in the phonological encoding of Standard Arabic lexical items and suggested the need for explicit, systematic, and sustained exposure to Standard Arabic phonological structures. The evidence reviewed above shows that diglossia does not support the acquisition of basic literacy processes in Arabic, and may indeed contribute to the widespread low levels of mother tongue reading comprehension scores (Kramarsky & Mevarech, 2003; Olshtain & Rozovsky, 2004). The findings also support the conjecture that diglossia may be to blame for the high rates of illiteracy in the Arabic speaking world (Ayari, 1996). Yet, is the linguistic distance between Spoken and Standard Arabic the only factor to blame? The infrastructure for literacy acquisition is established first in the home of the child. Thus, literacy acquisition is embedded in the socio-cultural contexts of the young children’s families. In effect, the belief system of parents regarding literacy should affect parent-child literacy interactions and activities. The impact of these factors on literacy development in diglossic Arabic remains largely unexplored. Iraqi (1990), in a study of book use habits in Arab Israeli families found that only five (1.8%) of the 290 Arab families investigated reported reading to their children from books. In 58.2% of the families, parents orally recited stories that they remembered from their own childhood, and in the remaining 40.0% of them said that they had used books in storytelling sessions but did not read from them. The two reasons that parents provide for not reading from books were that children do not understand the language of the book (Standard Arabic) and that they do not enjoy being read to from books. The Arab parents’ belief system regarding the importance of early exposure to Standard Arabic and regarding their children’s ability to acquire basic literacy skills at an early age determines parent-child literacy activities and interactions. This may be a very important determinant of children’s later literacy achievement and is a question for future research to explore. Informal experiences and exposure to literacy in the home of the child may be best tapped by preschoolers’ familiarity with letters. In a pilot study of 77 Arab Israeli middle-to-low class children, we tested children’s knowledge about Arabic letters: concept, function, name, and sound (Saiegh-Haddad, 2009). This study showed that, at the age of three, 98.8% of the children did not have a concept of letter. That is, upon presentation on a card, children could not categorize them as letters and only few said that they were numbers. This rate dropped to 57.1% at the age of four. Only at the age of five did all children realize that what they were shown were letters. With regard to the function of letters, 84.4% of the 3-year-olds did not know what letters were used for. At the age of four the percentage dropped to 71.4%. At the age of five, only 1 year before the first grade, 46.2% of all children did not know what the function of letters was. As to the number of letters that children could identify (either by name or sound), 93.8% of the three-year-olds, 75% of 4-year-olds, and 7.7% of 5-year-olds could not identify any of the 28 Arabic letters. Knowledge about the direction of writing and personal-name writing was assessed as well. Here,

3

Literacy Reflexes of Arabic Diglossia

51

65.5% of 3-year-olds did not know the direction of writing (right to left) and 93.8% of them could not write their names. These figures dropped to 41.7% and 75% at the age of four, and 15.4% and 15.4%, at the age of five, respectively. These compelling figures suggest that, though the linguistic distance between Spoken Arabic and Standard Arabic may be one important factor to blame, it is essential that future research investigate the direct effect of the literacy environment in which the child is raised, as well as the effect of the socio-cultural context in which informal literacy is embedded on the acquisition of literacy in Arabic. This is a significant avenue for future research to pursue. In support of the above argument, Levin, Saiegh-Haddad, Hende, and Ziv (2008) tested letter knowledge and phonological awareness in low SES 5-yearold Arabic native speaking children and the impact of an intervention program on children’s acquisition of these skills and showed that 40% of the children could not name any letter in the pretest, and, on average, this group succeeded in naming less than a single letter at that point in time. This performance level is considerably low when compared with middle-class children reared in other countries. For instance, American and Brazilian-Portuguese preschoolers were found to name an average of 64% of the letters in their alphabets (Treiman, Kessler & Pollo, 2006). Similarly, Hebrew speaking preschoolers and kindergartens, in a low-to-middle SES sample, though not taught about letters at school, succeeded in naming 61% of the letters (Treiman, Levin, & Kessler, 2006). The study also showed, however, that the intervention program succeeded in raising the letter knowledge and phonological processing performance of the children in the intervention group to a level comparable to that of their English- and Hebrew-speaking counterparts. The sustained low level of literacy knowledge throughout the school year among kindergartners in the control comparison group suggests that these children were deprived of the education that might have prepared them for learning to read and write in school. This deprivation exacerbates the effect of the linguistic distance between the spoken and the written codes on the acquisition of literacy in young Arabic speaking children and increases their risk for academic failure.

3.5 Conclusion Strong linguistic skills are associated with better reading outcomes. Given the diglossic context of Arabic and the remarkable orality-literacy gap, the question is how the linguistic disparity between Spoken Arabic and the language of literacy might factor into the unfortunate literacy and educational failure of Arab children. Despite frequently articulated concerns over the possibility that this longlived phenomenon of diglossia may be to blame for the widespread illiteracy in the Arab World (Ayari, 1996) and claims, explicit or implicit, by others that the acquisition of reading in Standard Arabic may be comparable to literacy acquisition in a foreign or second language (Abu-Rabia, 2000; Ibrahim, 1977,

52

E. Saiegh-Haddad

1983), the direct impact of the linguistic distance between Spoken and Standard Arabic on the acquisition of literacy has only recently begun to attract empirical attention. The studies reviewed here converge in demonstrating that the linguistic distance between Spoken Arabic, which children acquire as a mother tongue, and Standard Arabic, which is acquired primarily as a result of formal instruction in reading, does not support the natural acquisition of basic reading processes in Arabic (Saiegh-Haddad, 2003, 2004, 2005, 2007, 2008; SaieghHaddad et al., 2011). The linguistic duality problem that diglossia as a sociolinguistic phenomenon presents is supported and perpetuated by a unique socio-cultural milieu in which Arabic literacy acquisition is embedded. The complex linguistic reality results in poor language skills in the written code and, in turn, in linguistic insecurity that curtails language practice and impacts literacy-related practices and outcomes. This perpetuates the gulf between the two codes and stabilizes the diglossic reality. Further, the widespread low levels of proficiency in the written code and the high rates of illiteracy, together with abject poverty and low socio-educational background results in children beginning the process of literacy acquisition with little or no knowledge about the basic prerequisite language and literacy skills that form the infrastructure of reading development. This deprivation brings about more deprivation and less chances of success.

References Abu-Rabia, S. (2000). Effects of exposure to literary Arabic on reading comprehension in a diglossic situation. Reading and Writing: An Interdisciplinary Journal, 13, 147–157. Abu-Rabia, S., & Siegel, L. (2002). Reading, syntactic, orthographic, and working memory skills of bilingual Arabic-English speaking Canadian children. Journal of Psycholinguistic Research, 31, 661–678. Adams, M. (1990). Beginning to read: Thinking and learning about print. Cambridge, MA: MIT Press. Al-Batal, M. (1992). Diglossia proficiency: The need for an alternative approach to teaching. In A. Rouchdy (Ed.), The Arabic language in America (pp. 284–304). Detroit, MI: Wayne State Press. Al-Toma, S. J. (1969). The problem of diglossia in Arabic: A comparative study of classical and Iraqi Arabic. Cambridge, MA: Harvard University Press. Anghelescu, N. (1974). Arabic diglossia and its methodological implications. In M. Anghelescu (Ed.), Romano-Arabica (pp. 81–92). Bucharest: Romanian Association for Oriental Studies. Ayari, S. (1996). Diglossia and illiteracy in the Arab world. Language, Culture and Curriculum, 9, 243–252. Bell, A. (1984). Language style as audience design. Language in Society, 13, 145–204. Bowers, P. O., & Wolf, M. (1993). Theoretical links among naming speed, precise timing mechanisms, and orthographic skill in dyslexia. Reading and Writing: An Interdisciplinary Journal, 5, 69–85. Coulmas, F. (1987). What writing can do to language: Some preliminary remarks. In S. Battestini (Ed.), Developments in linguistics and semiotics, language teaching and learning, communication across cultures (pp. 107–129). Georgetown University Round

3

Literacy Reflexes of Arabic Diglossia

53

Table on Language and Linguistics, 1986. Washington, DC: Georgetown University Press. 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. Elbro, C. (1996). Early linguistic abilities and reading development: A review and a Hypothesis. Reading and Writing: An Interdisciplinary Journal, 8, 453–485. Elbro, C. (1998). When reading is ‘‘readn’’ or somthn. Distinctness of phonological representations of lexical items in normal and disabled readers. Scandinavian Journal of Psychology, 39, 149–153. Ferguson, C. A. (1959). Diglossia. Word, 14, 47–56. Fishman, J. A. (1967). Bilingualism with and without diglossia: Diglossia with and without bilingualism. Journal of Social Issues, 32, 29–38. Fishman, J. A. (1991). Diglossia revisited. Southwest Journal of Linguistics, 10, 214–234. Geva, E., & Siegel, 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. Goswami, U. (2000). Phonological representations, reading development and dyslexia: Towards a cross-linguistic theoretical framework. Dyslexia, 6, 133–151. Goswami, U. (2002). Phonology, reading, and dyslexia: A cross-linguistic perspective. Annals of Dyslexia, 52, 141–164. Goswami, U., & Bryant, P. (1990). Phonological skills and learning to read. United Kingdom: Erlbaum. Haeri, C. (2000). Form and ideology: Arabic sociolinguistics and beyond. Annual Review of Anthropology, 29, 61–87. Hudson, A. (1991). Toward the systematic study of diglossia. Southwest Journal of Linguistics, 7, 5–15. Hudson, A. (2002). Outline of a theory of diglossia. International Journal of the Sociology of Language, 157, 1–48. Hymes, D. (1973). Speech and language: On the origins and foundations of inequality among speakers. In Daedalus (pp. 59–86). Reprinted in Language and education: Ethnolinguistic essays. Language and Ethnography Series. CAL. 1980. Ibrahim, M. H. (1977). Diglossia and foreign language teaching. International Review of Applied Linguistics, 15, 154–163. Ibrahim, M. H. (1983). Linguistic distance and literacy in Arabic. Journal of Pragmatics, 7, 507–515. Iraqi, J. (1990). Reading to Arabic-speaking kindergarten children compared to alternative enrichment activities as means of improving reading comprehension and language skills. Unpublished master’s thesis. University of Haifa, Haifa (In Hebrew). Joseph, J. (1987). Eloquence and power: The rise of language standards and standard languages. New York: Blackwell. Katz, R. B. (1986). Phonological deficiencies in children with reading disability: Evidence from an object-naming task. Cognition, 22, 225–257. Kaye, A. (1972). Remarks on diglossia in Arabic: Well defined vs. ill defined. Linguistics, 81, 32–48. Keller, R. (1982). Diglossia in German-speaking Switzerland. In W. Haas (Ed.), Standard languages: Spoken and written (pp. 70–93). Manchester: Manchester University Press. Kramarsky, B., & Mevarech, Z. (2003). Reading, mathematical, and scientific literacy: PISA research 2002 (Unpublished manuscript). The School of Education, Bar-Ilan University, Israel. (In Hebrew). Landerl, K., & Wimmer, H. (2000). Deficits in phoneme segmentation are not the core problem of dyslexia: Evidence from German and English children. Applied Psycholinguistics, 21, 243–262.

54

E. Saiegh-Haddad

Levin, I. (1994). The grammar of the Jerusalemite dialect of Arabic. Jerusalem, Israel: Magness Press. (In Hebrew). Levin, I., Saiegh-Haddad, E., Hende, N., & Ziv, M. (2008). Early literacy in Arabic: An intervention with Israeli Palestinian kindergarteners. Applied Psycholinguistics, 29, 413–436. Maamouri, M. (1998). Language education and human development: Arabic diglossia and its impact on the quality of education in the Arab region. Discussion paper prepared for The World Bank. The MediterraneanDevelopment Forum. McBride-Chang, C. (1995). What is phonological awareness? Journal of Educational Psychology, 8, 179–192. 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. Olshtain, E., & Rozovsky R. (2004). The acquisition of reading literacy in Israel:Findings from the international reading literacy research PIRLS-2001. The Ministry of Education and the Centre for Scientific and Technological Education at Tel-Aviv University. Tel-Aviv: Ramot Publishing, Israel (In Hebrew). Saiegh-Haddad, E. (2003). Linguistic distance and initial reading acquisition: The case of Arabic diglossia. Applied Psycholinguistics, 24, 431–451. Saiegh-Haddad, E. (2004). The impact of phonemic and lexical distance on the phonological analysis of words and pseudo words in a diglossic context. Applied Psycholinguistics, 25, 495–512. Saiegh-Haddad, E. (2005). Correlates of reading fluency in Arabic: Diglossic and orthographic factors. Reading and Writing: An Interdisciplinary Journal, 18, 559–582. Saiegh-Haddad, E. (2007). Linguistic constraints on children‛s ability to isolate phonemes in diglossic Arabic. Applied Psycholinguistics, 28, 605–625. Saiegh-Haddad, E. (2008). On the challenges that diglossia poses to the acquisition of basic literacy skills in Arabic. Language and Literacy, 1, 105–126. Saiegh-Haddad, E. (2009). The effect of exposure to Standard Arabic and linguistic distance from Spoken Arabic on lexical processing in Standard Arabic. In D. Aram & O. Korat (Eds.), Literacy development: Learning and teaching at home and school (pp. 321–336). Jerusalem: The Magnes Press, Hebrew University. Saiegh-Haddad, E. (ms.). A corpus linguistic study of the lexicon of five-year-old Arabic native speaking children. In Preparation. Saiegh-Haddad, E., Levin, I., Hende, N., & Ziv, M. (2011). The linguistic affiliation constraint and phoneme recognition in diglossic Arabic. Journal of Child Language. In Press. Available on CJO 24 Jun 2010 doi:10.1017/S0305000909990365. Seymour, P. H. K., Aro, M., & Erskine, J. M. (2003). Foundation literacy acquisition in European orthographies. British Journal of Psychology, 94, 143–174. Sotiropoulos, D. (1977). Diglossia and the national language question in Modern Greece. Linguistics: An Interdisciplinary Journal of the Language Sciences, 197, 5–31. Treiman, R., Kessler, B., & Pollo, T. C. (2006). Learning about the letter name subset of the vocabulary: Evidence from U.S. and Brazilian preschoolers. Applied Psycholinguistics, 27, 211–227. Treiman, R., Levin, I., & Kessler, B. (2006). Learning of letter names follows similar principles across languages: Evidence from Hebrew. Journal of Experimental Child Psychology, 93, 139–165. Walters, K. (1994). Diglossia, linguistic variation, and language change in Arabic. In M. Eid (Ed.), Perspectives on Arabic linguistics VIII (pp. 157–197). Amsterdam/ Philadelphia: John Benjamins Publishing Company. Wimmer, H., Mayringer, H., & Landerl, K. (1998). Poor reading: A deficit in skillautomatization or a phonological deficit? Scientific Studies of Reading, 2, 321–340.

3

Literacy Reflexes of Arabic Diglossia

55

Wimmer, H., Mayringer, H., & Landerl, K. (2000). The double-deficit hypothesis and difficulties in learning to read a regular orthography. Journal of Educational Psychology, 91, 415–438. Wolf, M., & Bowers, P. (1999). The question of naming-speed deficits in developmental reading disabilities: An introduction to the double-deficit hypothesis. Journal of Educational Psychology, 19, 1–24. Wolf, M., Bowers, P., & Biddle, K. (2000). Naming-speed processes, timing, and reading: A conceptual review. Journal of Learning Disabilities, 33, 387–407. Wolf, M., O’rourke, A. G., Gidney, C., Lovett, M., Cirino, P., & Morris, R. (2002). The second deficit: An investigation of phonological and naming-speed deficits in developmental dyslexia. Reading and Writing: An Interdisciplinary Journal, 15, 43–72. Zughoul, M. R. (1980). Diglossia in Arabic: Investigating solutions. Anthropological Linguistics, 22, 201–217.

Chapter 4

Multilingualism Among Israeli Arabs, and the Neuropsychology of Reading in Different Languages Zohar Eviatar and Raphiq Ibrahim

In this chapter we focus on two major topics. The first examines the linguistic situation in the Arab population of Israel, which is quite complex, and has implications for reading acquisition. As we show, literate Israeli Arabs are minimally quadralingual, and reading and writing in multiple languages are always acquired in the non-native language. The second topic focuses the psycholinguistic characteristics of Arabic and Hebrew orthography, and on models of differences in the functional architecture of reading in Hebrew and in Arabic.

4.1 Multilingualism in the Arab Context in Israel 4.1.1 Diglossia or Multilingualism? In Arabic, the spoken form (ammia – the local dialect) is used by speakers of the language in a specified geographic area for daily verbal communication, and is the native language of virtually all Arabic speakers. This form is differentiated from the fuṣḥa (the literary form), which is the language in which all speakers of Arabic, from all over the world, read and write. This literary Arabic is universally used in the Arab world for formal communication and is known as ‘‘written Arabic’’ or ‘‘Modern Standard Arabic’’ (MSA). Spoken Arabic (SA) appears entirely in colloquial dialect and has no written form. Hence, from the ecological point of view, SA and MSA could be considered as an instance of ‘‘diglossia’’, that is, a social environment in which a community uses two forms of the same language concomitantly (Ferguson, 1959). However, from a linguistic perspective, literate Arabic speakers could be considered, de facto, bilinguals, and contribute to the debate as to whether the two forms of Arabic represent different languages (Eid, 1990). We take up this debate below. Z. Eviatar (*) Department of Psychology, University of Haifa, Haifa, Israel Institute for Information Processing and Decision Making, University of Haifa, Haifa, Israel e-mail: [email protected] M. Leikin et al. (eds.), Current Issues in Bilingualism, Literacy Studies 5, DOI 10.1007/978-94-007-2327-6_4, Ó Springer ScienceþBusiness Media B.V. 2012

57

58

Z. Eviatar and R. Ibrahim

One of the foci in bilingualism research has been the effect of bilingualism on cognitive and paralinguistic development. Recently we (Eviatar & Ibrahim, 2001) asked how early exposure to the two forms of Arabic can affect the metalinguistic abilities of children. The received view in bilingualism literature is that the challenges posed by two linguistic systems promote the development of cognitive strategies that result in heightened metalinguistic abilities in young children (e.g., Bialystok, 1991). Metalinguistic awareness refers to the ability to think about the linguistic nature of the message and to be aware of certain properties of language such as its arbitrariness and phonological structure (Ben Zeev, 1977; Dash & Mishra, 1992). To be metalinguistically aware is to solve linguistic problems such as the detection of ambiguity and grammaticality (Galambos & Hakuta, 1988; Galambos & Goldin Meadow, 1990), to create new words (Titone, 1994), and to be able to segment words into their constituents (Campbell & Sais, 1995). All of these require an awareness of language as a system and the ability to access and manipulate knowledge about that system (Bialystok & Ryan, 1985). We focused on the finding that exposure to two languages results in a specific pattern of performance on three types of tests (arbitrariness, phonological awareness, and vocabulary size) and asked whether children exposed to the two forms of Arabic would show this pattern. That is, we asked, are the two forms of Arabic different enough from each other to result in the pattern typical of bilingualism? We tested Arabic-speaking children, who are exposed to both SA and MSA and compared them to Russian-Hebrew bilinguals and Hebrewspeaking monolinguals. All of the children were in kindergarten or in first grade. The results of the Russian-Hebrew bilinguals showed the classic pattern resulting from exposure to two languages: higher performance levels in metalinguistic tests and lower performance levels in the vocabulary measure as compared to monolinguals. The Arab children’s performance mimicked those of the bilingual children for the most part, and suggested that exposure to MSA in early childhood affects metalinguistic skills in the same manner as that reported for children exposed to two different languages. Thus, Arab children enter first grade as bilinguals, and those who attend the Arab school system begin to learn to read MSA in first grade, to speak Hebrew in second grade, and to read and write in Hebrew and in English in third grade. At the high-school level, most students are as proficient in Hebrew as they are in MSA. Ibrahim (Ibrahim, 1998; Ibrahim & Aharon-Peretz, 2005) examined the status of MSA and Hebrew in high-school students. The logic was the following: if MSA constitutes a second language for the Arabic speaker, then the results obtained regarding the lexical status of MSA words and their connections to meaning should be similar to patterns found for Hebrew. On the other hand, if the daily interactive use of SA and MSA, along with the sociolinguistic reality in which the two languages may be considered two forms of one language, have led to the combination of both forms of Arabic in a single lexicon, the results of linguistic manipulations between the two forms of Arabic should resemble those known to exist when the same linguistic manipulations are

4

Multilingualism Among Israeli Arabs, and the Neuropsychology of Reading . . .

59

performed within a language. The relations between the two forms of Arabic were compared to the relations existing between Hebrew and SA. Ibrahim compared semantic priming effects within Spoken Arabic, with the effects found across languages, with MSA or Hebrew being the other language. The results in all of the studies were consistent and straightforward: when both the primes and the targets were presented in SA, the semantic priming effect was significantly greater than when the prime and target words were from MSA or Hebrew. Most importantly, the cross-language priming effects on response times were virtually the same, regardless of whether the second language was Hebrew or MSA. As with the young children, these findings suggest that the representation of MSA is that of a second language (L2), similar to Hebrew, and that SA and MSA have the status of two separate languages in the cognitive systems of Arabic speaking adolescents. These findings align nicely with previously reported asymmetry in crosslingual semantic priming (Altarriba, 1992; Keatley & DeGelder, 1992). The interpretation of the difference between the patterns of priming within and across languages is that words in a second language have looser connections with their meanings than do words in the first language. Therefore the semantic priming pattern described above suggests that, at least in regard to their connections with the semantic network, MSA, as well as Hebrew, constitute second languages for the bilingual native speaker of SA. Being Semitic languages, Hebrew and both SA and MSA have similar morphological structure. Ibrahim (2006) examined to what extent the morphophonological similarity among the languages is modulated by their psychological status as a native or a second language. Based on a distributed conceptual representations model (see for example, McClelland, Rumelhart, & Hinton, 1986), de Groot (e.g., 1992, 1995) accounted for the morpho-phonemic similarity effect of cognate translation equivalents by assuming that because they have a common etymology, they also share more meaning features than noncognate translation equivalents. Therefore priming between cognate translations is more effective than between non-cognate translations. In essence, according to her view, repetition priming between translation equivalents, like semantic priming, originates from activation of common semantic features in the conceptual system. If this is the case, then the findings cited above might be due to the morpho-phonemic similarity among SA and MSA and Hebrew and MSA. Ibrahim (2006) asked if the difference in semantic overlap of HebrewMSA and SA-MSA translation equivalants remained constant across repetition lags. The results showed larger priming effects for cognate Hebrew – MSA than SA – MSA pairs at lag 0, on the one hand, and smaller priming effects at lag 4, on the other hand. Ibrahim concluded that non-linguistic factors qualified the influence of the linguistic factors in determining the magnitude of the morphophonemic similarity effects. Specifically, he proposed that among these factors are lexical-episodic associations, which are apparently stronger between translation equivalents in two languages that are interactively and concomitantly used on an everyday basis (such as SA and MSA), than between translation

60

Z. Eviatar and R. Ibrahim

equivalents in languages that are not concomitantly used (such as Hebrew and MSA). In concert with previous findings, this study indicated that despite extremely intensive and concomitant use of Spoken and Literary Arabic, this does not bring about a change in their status as first and second languages, respectively. However, under such circumstances, associative links are formed between translation equivalents at the lexical level. The strength of these associations is determined by the frequency of concomitant use and by the psychological perception of the two languages as being forms of a single language. To summarize, the linguisitic requirements of literate Israeli Arabs are quite complex. Adults can be minimally considered quadrilingual, with SA as L1, and MSA, Hebrew and English as additional languages. Because SA does not have a written form, all reading and writing are carried out in the nonnative language.

4.1.2 Multilingualism and Reading Acquisition Ayari (1996) and Maamouri (1998) have suggested that the diglossic situation in Arabic might be related to some of the hardships that native Arabic beginning readers encounter and might even hinder their acquisition of basic academic skills. Abu-Rabia (2000) directly investigated the contention that reading difficulties in Arabic in elementary school result from the diglossic situation of fuṣḥa, the language of books and school instructions, and its opposition to the spoken dialect of the home. Starting from the belief shared by educators, teachers, and parents that the exposure of young Arabic speakers to fuṣḥa in the preschool period is not useful and a burden to all, Abu-Rabia compared the reading comprehension performance of first and second grade children who had been experimentally exposed to literary Arabic throughout their preschooling period with the reading performance of a parallel control group only exposed to spoken Arabic during that period. He found, contrary to the commonly held belief, that the early exposure of Arab preschool children to fuṣḥa text (stories) enhances their reading comprehension abilities and improves their performance in reading comprehension tests 2 years later. Finally, more than his scientific findings, Abu-Rabia’s conclusions (2000: 155) are worth noting: (a) policy-makers may incorporate this pedagogy in all preschool years, (b) educating elementary-school teachers and kindergarten teachers in diglossic issues, and (c) the recommendation that ‘‘teachers at all levels use literary Arabic as the language of instruction.’’ Elinor Saiegh-Haddad (2003) studied the role of oral language in the acquisition of basic fuṣḥa reading processes in kindergarten and first grade Arabicspeaking children. She hypothesized that because native speakers of Arabic first learn to read in fuṣḥa, a language structurally different from the local dialect they grow up speaking, the linguistic differences between the two Arabic language varieties would interfere with the acquisition of basic reading processes. In a series of studies of the acquisition of reading in diglossic Arabic (Saiegh-Haddad, 2004, 2005) Saiegh-Haddad tried to answer the following

4

Multilingualism Among Israeli Arabs, and the Neuropsychology of Reading . . .

61

questions: first, do diglossic variables or linguistic distance parameters interfere with the acquisition of basic reading processes in fuṣḥa? Second, which diglossic structures interfere with the acquisition of basic reading skills, phonological, syntactic, morphosyntactic or lexical? Third, which reading skills (phonemic awareness, word decoding, reading fluency, or reading comprehension) are sensitive to diglossic variables? Testing phonemic awareness skill and decoding pseudo-word for example, showed that although the first grade children seemed to have benefited from the increased exposure to fuṣḥa structures that formal literacy instruction allowed, they still found the task of isolating standard phonological structures quite difficult, especially for sounds that do not occur in their local dialect. The findings were similar to those reported by Abu-Rabia, as the first grade children seemed to have benefited from the increased exposure to fuṣḥa structures that formal literacy instruction allowed. Saiegh-Haddad concluded that, diglossia (presented by the linguistic disparity between SA of children and the linguistic system encoded in print) does not appear to support the natural acquisition of basic literacy processes in Arabic. In that regard, it could even contribute to the proven low levels of mother tongue literacy achievement in Arabic, as international literacy assessment projects have revealed (Olshtain & Rozovsky, 2004) and may support Ayari (1996) who blames the diglossia for the high rates of illiteracy in the Arabic-speaking world. In the following section we review some of the findings on reading in Hebrew and Arabic from both psycholinguistic and neuropsychological perspectives.

4.2 Reading in Hebrew and in Arabic 4.2.1 The Importance of Vowels Unlike the Latin orthography in which vowels are represented by letters, in Arabic and Hebrew vowels are not part of the alphabet. Vowelization is optional, and most texts do not include them. When vowels do appear (in poetry, children’s books and liturgical texts), they are signified by diacritical marks above, below or within the body of the word. To complicate matters, there are four letters which also specify long vowels, in addition to their role in signifying specific consonants (in Hebrew: ‘alef , vav, yud and he’; in Arabic: ‘alif’ or ‘imaala, waaw, yaa’). However, in some cases it is difficult for the reader to determine whether these dual-function letters represent a vowel or a consonant. Research exploring the processing of vowel diacritics in Hebrew has revealed that naming of unambiguous words was facilitated when these were presented in pointed script (Frost, 1994), while lexical decisions did not benefit from pointing (Bentin & Frost, 1987; Frost, 1994, 1995; Koriat, 1984, 1985a; for an exception see Koriat, 1985b). Only studies that applied an interference paradigm demonstrated automatic processing of vowel information for naming as

62

Z. Eviatar and R. Ibrahim

well as for verbal lexical decisions and semantic judgments (Navon & Shimron, 1981). These researchers used an interference paradigm in which vowel information conflicted with the response suggested by the relevant consonant information. In three experimental groups, adult native Hebrew speakers were presented with correctly voweled words, unvoweled words, and ‘‘phonemically distorted’’ vowels, in which the vowels turned the sound of a word into a nonword. Although participants were instructed to ignore the vowel information, responses for phonemically incorrect pointing were considerably slower, indicating that participants were not able to disregard the presented vowel information. In addition, ‘‘graphemically distorted’’ pointing was applied, in which orthographically incorrect diacritics changed the visual pattern of a word, but nevertheless preserved its correct sound. This kind of pointing was introduced to control for the fact that phonemically incorrect pointing also provided incompatible visual cues. The fact that orthographically incorrect pointing did not interfere with naming latencies suggested that the interference produced by phonemic distortion was due to phonological processing. The authors concluded that vowel graphemes are automatically translated into phonemes. Smolka and Eviatar (2006) showed that this automatic phonological recoding occurs in the left hemisphere (LH) of adult readers, and that diacritics are processed as visual patterns in the right hemisphere (RH). The Arabic script uses diacritical forms (or diacritics) for three short vowels (a-fatha, u-damme, i-kasra) that represent the long vowels: a; o; e (‫ﻱ‬, ‫ﻭ‬, ‫)ﺍ‬. An additional diacritical marking, the shadda, is used for lexical differentiation. Most of the grammatical functions at both the morphological and syntactic levels are represented by the short vowels, which also represent mood and case endings in the Verb-Subject-Object literary (fuṣḥa) syntax. From a psycholinguistic view, inclusion of these marks specifies the phonological form of the orthographic string, making it completely transparent in terms of orthography/ phonology relations. In both Arabic and Hebrew, the majority of written materials do not include the diacritical marks, such that a single printed word is often not only ambiguous between different lexical items (this ambiguity is normally solved by semantic and syntactic processes in text comprehension), but also does not specify the phonological form of the letter string. Thus in their unvoweled form, Hebrew and Arabic orthographies contain a limited amount of vowel information and include a large number of homographs. For example, the bare unvowelized fuṣḥa form TH-K-R (‫ )ﺫﻛﺭ‬has five readings and five corresponding semantic interpretations: (a) thakara ‘‘mentioned’’ (b) thakkara ‘‘remind (singular/masculine)’’ (c) thokera ‘‘was mentioned’’ (d) theker ‘‘learning situation’’ and (e) thakar ‘‘masculine’’. Nevertheless, as mentioned above, it has been found that in both Hebrew (Bentin, Bargai, & Katz, 1984) and Arabic (Roman & Pavard, 1987) the addition of phonological disambiguating vowel points inhibits (rather than facilitates) lexical decision. On the basis of such results, it has been suggested that, at least in Hebrew, correct lexical decisions may be initiated on the basis of orthographic codes, before a particular

4

Multilingualism Among Israeli Arabs, and the Neuropsychology of Reading . . .

63

phonological unit has been accessed (Bentin & Frost, 1987). Evidence for orthographic-based access in Hebrew has been presented by Peleg and Eviatar (2009), who showed that heterophonic homographs in Hebrew (ambiguous words that have an identical graphemic form and different phonological forms, such as ‘‘tear’’) as well as homophonic homographs (ambiguous words with identical graphemic and phonological forms, such as ‘‘bank’’) activate both meanings, when probed very early in the reading process. An important longitudinal study on the acquisition of literacy and Arabic reading skills was done by Wranger in Morocco (Wagner, Spratt, & Ezzaki, 1989; see also, Wagner, 1993). Wranger and his colleagues investigated whether preschool experience with a spoken Moroccan Arabic dialect facilitates literacy acquisition differently than preschool experience with Berber, which is a member of the Hamitic family of languages and has no semantic or syntactic similarity to Arabic. This study compared the performance of monolingual and bilingual Arabic- and Berber-speaking children in learning to read MSA and French. The results showed an advantage for the children whose mother tongue was the Arabic dialect over Berber-speaking children in MSA. In contrast, the preschool experience of the two languages groups had little effect in learning to read French. The researchers concluded that the superiority of the Arabic speaking children in the in the early stages of MSA literacy acquisition is due primarily to the substantial similarity and transfer from spoken Moroccan Arabic to MSA, but not to French. Wagner (1993) also showed that knowledge in first grade of Arabic letters, their graphemic variability and pronunciation, predicted more than 30% of the variance in reading achievement 5 years later. Early decoding skills at the single word level explained an additional 14% of the same variance. Wagner concluded that learning to read in Arabic necessitates an even greater reliance on decoding skills than in other languages. Wagner highlights ‘‘the absence of vocalization diacritics as the main reason behind the growing difficulty of decoding for word recognition and paragraph comprehension, a difficulty which mars advanced Arabic reading stages and requires knowledge of appropriately correct inflectional endings and the ability to place full and correct diacritical marking’’ (p. 240). Azzam (1990) examined the misreadings and misspellings that Arab primary school children make and identified vocalization and its use of diacritical markings as the main culprit. Her research seems to suggest that diacritical markings are significantly important in the process of reading and comprehending written language at all levels of Arabic reading. Abu-Rabia (1998) also investigated the effect of vowels on reading accuracy in Arabic orthography. He presented four kinds of written fuṣḥa Arabic texts (narrative, informative, poetic, and Koranic) to sixty-four native Arabic speakers to be read aloud. Three texts of each kind were presented in three reading conditions: correctly vowelized, unvowelized, and wrongly vowelized. The most important finding of this study is that vowels were found to significantly influence the reading of both poor and skilled readers in the four fuṣḥa writing styles in all three

64

Z. Eviatar and R. Ibrahim

conditions. It was also found that both skilled and poor readers improved their reading accuracy in all writing styles when they read with vowels. This last study reinforces and supports similar previous findings obtained by Abu-Rabia (1997), where it was demonstrated that the vowels and the sentence context were significant factors for word recognition for both skilled and poor fuṣḥa readers. Thus, vowel diacritics are automatically processed in Hebrew, and facilitate naming even in skilled readers. However, word recognition in Hebrew can be orthographically based, with vowels, when they are automatically recoded into phonology, actually slowing the process down. In Arabic, the inclusion of diacritics improves reading accuracy of both beginning and skilled readers.

4.2.2 Orthographic Complexity Although Arab children’s scores on tests of phonological awareness were higher than those of monolingual Hebrew speakers, their scores on tests of reading achievement were lower (Eviatar & Ibrahim, 2001; Ibrahim, Eviatar, & Aharon Peretz, 2002). We suggested that this is due to the complexity of Arabic orthography as compared with Hebrew orthography. Indeed, comparing Arabic and Hebrew orthographic systems with English, additional complexity is found in both orthographies but to a much larger extent in Arabic than in Hebrew. The two orthographies differ in two central aspects. The first aspect is related to shape of some letters that differ depending on their placement in the word. This phenomenon is much less extensive in Hebrew than in Arabic. In Hebrew there are five letters that change shape when they are word final: (‫ם‬-‫מ‬, ‫ן‬-‫נ‬, ‫ץ‬-‫צ‬, ‫ף‬-‫פ‬, ‫ך‬-‫)כ‬. The Arabic writing system is an alphabetic system with twenty eight basic consonant letters. Most of these consonants show a very close resemblance in form, with only additional dots or strokes to distinguish them from each other. They are usually composed of one base form and most of them have up to three or four distinct variant shapes. Graphemic variants differ depending on whether they occur in a word in the initial, middle or final position and whether they follow a connecting or a non-connecting letter (for example, the phoneme /h/ is represented as: ). The second characteristic has to do with diacritics and dots. In Hebrew, dots occur only to mark vowels and as a stress-marking device (dagesh). In the case of three letters, this stress-marking device (which does not appear in unvowelized scripts) changes the phonemic representation of the letters from fricatives (v, x, f) to stops (b, k, p for the letters ‫ פ ק ב‬respectively). In the unvowelized form of the script, these letters can be disambiguated by their place in the word, as only word or syllable initial placement indicate the stop consonant. In Arabic, the use of dots is more extensive: many letters have a similar or even identical structure and are distinguished only on the basis of the existence, location and number of dots (e.g., the Arabic letters representing /t/ and /n ( )

4

Multilingualism Among Israeli Arabs, and the Neuropsychology of Reading . . .

65

become the graphemes representing /th/ and /b/ ( ) by adding or changing the number or location of dots. We tested the hypothesis that the graphemic complexity of Arabic is larger than in Hebrew, and that this results in additional perceptual load (Ibrahim, Aharon-Peretz, & Eviatar, 2002). The subjects tested were all adolescent healthy native Arabic speakers who had mastered Hebrew as a second language. We used oral and visual variants of the trail making test (Reitan & Wolfson, 1993) in both languages. Both versions have two levels of complexity: Level A requires connecting visually numbers or letters in order. Level B in the two modalities requires alternation between letters and numbers. Level B in the oral version requires declamation of the alternation. Performance time was the dependent variable. At the low level of complexity (Level A) there were no differences between performance in Hebrew and in Arabic in either the oral and the visual versions. In the more complex version (Level B), language (Hebrew or Arabic) did not affect speed in the oral version, but in the visual version, Arabic was performed significantly slower than Hebrew. These findings supported our major conclusion that Arabic letters are harder to identify than Hebrew letters as a result of their greater visual complexity. In two divided visual field studies we have shown that Arabic letters are harder to identify than English and Hebrew letters and have suggested that the locus of this difficulty is in the right hemisphere. In the first study (Eviatar & Ibrahim, 2004) we examined directly the effects of grapheme–phoneme conversion in English, Hebrew and Arabic, using a lateralized nonsense syllable identification task. The syllables were constructed as consonant–vowelconsonant (CVC) trigrams (the vowels in Hebrew and Arabic were letters that double as consonants or vowels), and the task of the participants was to identify the three letters. The stimuli were presented vertically in three conditions: left visual field (LVF), right visual field (RVF), and bilaterally (BVF). The participants were university students, native readers of each of the languages. In this study we were interested in errors, so we titrated exposure duration independently for each participant, in order to achieve a duration that resulted in 50% errors. This paradigm allowed us to measure three dependent measures that indexed different aspects of the task. The first measure was the mean exposure duration that was reached in each of the three groups of participants in order to achieve a 50% error rate. This is an index of the speed at which native readers can identify letters in each of the languages. This measure revealed that the Arabic readers required significantly longer exposure durations that the readers of Hebrew and English, and that Hebrew readers required significantly longer exposure durations that the readers of English. Thus, these results suggest that English letters are easier to identify than Hebrew letters, and that Arabic letters are the hardest to identify. The second measure in this study was the total number of errors in each presentation condition (LVF, RVF, and BVF). This measure revealed that all of the participants showed a right visual field advantage (RVFA) that reflects specialization of the left hemisphere (LH) for this linguistic task. This advantage, the difference between performance levels in the

66

Z. Eviatar and R. Ibrahim

LVF and in the RVF was significantly larger in the Arabic speakers than in the other groups, as a result of poorer performance in the LVF in Arabic than in the other languages, while performance in the other presentation conditions (RVF, BVF) was equivalent among the groups. The third measure in this study was the difference between errors on the first letter and errors on the last letter of the trigram, a qualitative measure of sequential processing (Levy, Heller, Banich, & Burton, 1983; Eviatar, Hellige, & Zaidel, 1997). This qualitative measure revealed that Arabic and Hebrew speakers evinced a similar pattern that was different from the one shown by the English speakers. We interpreted this as indicating a different division of labor between the hemispheres while reading English or the Semitic languages, and attributed the difference to the demands made by concatenative versus nonconcatenative word morphology. We discuss this issue in more detail below. Of importance to us here is that it took longer for Arabic speakers to identify Arabic letters than it did for Hebrew speakers to identify Hebrew letters or English speakers to identify English letters. In addition, the large difference between performance levels in the two visual fields of Arabic speakers suggested that there are large differences in the abilities of the hemispheres in letter identification in Arabic, but not in the other languages. In the second divided visual field study (Eviatar, Ibrahim, & Ganayim, 2004) we explored the locus of this difference, and showed that adult native Arabic speakers who can read both Arabic and Hebrew, are better at identifying letters in Hebrew than in Arabic, and that the main disadvantage for Arabic letters is in the left visual field, when they are exposed to the right hemisphere (RH). We asked native Arabic speakers and native Hebrew speakers to perform a lateralized letter matching task in both Arabic and Hebrew, using a physical identity criterion (the Arabic speakers were literate in both languages, but the Hebrew speakers could not read Arabic, and thus performed the task as a pattern matching task). The pattern of results in response times was as we expected, revealing a RVFA in all of the conditions in which the participants could read (all except Hebrew speakers in Arabic, who showed a slight LVFA in this condition). The results of the accuracy measure were quite dramatic. In Hebrew, both groups revealed low error rates and equivalent performance in the two visual fields (both hemispheres are able to match letters quite well). In Arabic, the Hebrew speakers made many errors, with equivalent performance in the two visual fields. Arabic speakers revealed good performance in the RVF (their LH was able to match letters in Arabic as well as in Hebrew). In the LVF, however, Arabic speakers made as many errors as Hebrew speakers (who cannot read the language)! We hypothesized that the reason for this RH disadvantage in letter recognition is the complexity of grapheme-phoneme relations in Arabic. In order to examine our hypothesis that the RH cannot differentiate between very similar different letters in Arabic, whereas the LH can do so, we created a global-local task with two types of incongruent stimuli: one where the two letters on the two levels of the hierarchical stimulus were physically very different from each other: ‫ ﺕ‬and ‫ ;ﻡ‬and another where the two letters were very similar to each other: ‫ ﺕ‬and ‫ﺏ‬. Participants were required to

4

Multilingualism Among Israeli Arabs, and the Neuropsychology of Reading . . .

67

attend to the local or the global levels of these hierarchical stimuli in different blocks. We measured the difference between congruent (where the same letter was used on both levels of the stimuli) versus incongruent conditions (where the letter in the global level was made out of small versions of the other letter). This difference indicated the degree of interference between the levels. In the first condition, where the letters were very different from each other, we replicated the results of other studies in other languages (e.g., Lamb, Robertson, & Knight, 1990). We found a global precedence (responses to the global level were always faster than to the local level (Navon, 1977)), and an asymmetry in the degree of interference between the two visual fields: stronger interference from the global level to the local level in the LVF, and the opposite pattern in the RVF. This pattern has been used to support the hypothesis that the LH is relatively more sensitive to the local aspects of visual stimuli and that the RH is relatively more sensitive to the global aspects of these stimuli. Most interestingly, however, in the second condition, where the two letters differed only in the number and placement of dots, there was no incongruence effect in the LVF at all, while congruent stimuli were faster than incongruent stimuli in the RVF in the local conditions. These results show that the RH cannot discriminate between letters that differ only in the placement or number of dots (e.g., /t/- ‫ﺕ‬ and /b/ - ‫)ﺏ‬, but that the LH can do so. Thus, in the case of Arabic, an important part of the literacy problem is posed by the Arabic orthographic system and its failure to support easy and efficient reading. Previous research on reading acquisition in the Arabic language has revealed that this process is slower than in Hebrew. We have shown that in both beginning and skilled readers, letter discrimination in Arabic is quite difficult. In the next section we examine the effects of word morphology in Arabic and in Hebrew on the process of reading, and on the division of labor between the cerebral hemispheres in the early stages of visual word recognition.

4.2.3 Word Morphology Arabic and Hebrew, as Semitic languages, are characterized by a nonconcatenative, highly productive derivational morphology (Berman, 1978). In both languages words are constructed by combining a consonantal root (that carries most of the semantic information) and a word pattern that includes vowels as well as consonants, and provides information about the word class and its morphological status. Hence, the majority of words in Hebrew and Arabic are at the very least bi-morphemic, but none of the composing morphemes are words by themselves. In most words, the core meaning is conveyed by the root, while the phonological pattern conveys word class information. For example, in Arabic the word (TAKREEM) consists of the root (KRM, whose semantic space includes things having to do with respect) and the phonological pattern TA- -I-. The combination results in the word ‘‘honor.’’ The letters that make up the root may be dispersed across the word, inter-digitated with letters

68

Z. Eviatar and R. Ibrahim

that can double as vowels and other consonants that belong to the morphological pattern. Roman and Pavard (1987) used oculomotor recording techniques to evaluate visual scanning strategies. They found that although mean reading time did not differ between Arabic and French texts (note that for conveying identical content, the number of words needed in Arabic is less than in French because Arabic morphology is more dense), gaze duration per word was significantly longer in Arabic (342 ms) than in French (215 ms). This phenomenon also has been found in comparisons of Hebrew and English text reading, in which the morphology of Hebrew is dense and similar to that of Arabic, and English morphology is concatenative and more similar to French (Shimron & Sivan, 1994). A number of psycholinguistic studies (Frost & Bentin, 1992; Frost, Katz & Bentin, 1987; Deutsch, Frost, & Forster, 1998; Berent, 2002) have explored the effects of the morphology and orthography of Hebrew on lexical access and the structure of the mental lexicon. One of the conclusions from these studies is that the nonconcatenative and agglutinative morphological structure of Hebrew, together with the distributional properties of abstract word forms, results in the inclusion of subword morphological units in the mental lexicon of Hebrew speakers. Similarly, Prunet, Beland, and Idrissi (2000) report a case study of an Arabic-French agrammatic patient, who showed identical deficits in the two languages, except for a specific type of error, metathesis, in which he modified the order of the root consonants, with the vowel patterns remaining intact, only in Arabic, not in French. They interpret this finding as reflecting the manner in which words are stored in the mental lexicon in the two languages: whole words plus affixes in French, and roots plus word patterns in Arabic. These findings converge with the conclusions of Eviatar (1999, Experiment 4) and Eviatar and Ibrahim (2004), who showed that nonsense CVC trigrams are processed sequentially in both visual fields in English, but in neither visual field in Hebrew and in Arabic, and hypothesized that this is because Hebrew and Arabic nonwords cannot be read sequentially. A similar conclusion for words was reached by Farid and Grainger (1996), who showed that the location of initial fixation in a word is different in French and in Arabic words. In French, fixation slightly to the left of the word’s center results in best recognition for both prefixed and suffixed words, while in Arabic, prefixed words result in best recognition from leftward fixations and suffixed words result in best recognition from rightward initial fixations. They suggest that this is due to the greater importance of morphological structure in Arabic, because ‘‘. . .much of the phonological representation of the word can be recovered only after successfully matching the consonant cluster to a lexical representation’’ (p. 364), that is, after extraction of the root. Berent (2002) has also concluded that in Hebrew, ‘‘Speakers decompose the root from the word pattern in on-line word identification. . .’’ (p. 335). Most recently we reported that the different manner in which words are constructed in English and in Hebrew and Arabic has an effect on the division of labor between the cerebral hemispheres in a lateralized lexical decision task

4

Multilingualism Among Israeli Arabs, and the Neuropsychology of Reading . . .

69

(Eviatar & Ibrahim, 2007). We presented native speakers of Arabic, Hebrew, and English with morphologically simple and complex words and nonwords in their native language, and measured indexes of hemispheric integration. Morphological complexity was operationalized differently in English than in the Semitic languages. In English we defined monomorphemic words as morphologically simple, and derivations (e.g., farmer¼farmþer) as morphologically complex. Morphologically complex nonwords were made up of legal morphemes in illegal combinations (e.g., logly). In Arabic and in Hebrew we defined a word as morphologically simple if the rootþwordform structure was not transparent (e.g., the word is not easily divisible into these morphemes or the root is not generative, and appears only in that form), and as morphologically complex if it was easily and transparently divisible into these elements. Morphologically complex nonwords were created by inserting nonexistent roots into legal wordforms. In English, we replicated the findings of previous studies: similarly to Iacoboni and Zaidel (1996), we showed that while the RH is able to independently recognize nonwords; it draws upon resources of the LH when encountering words. Similarly to Burgess and Skodis (1993) in English, and to Koenig, Wetzel, and Caramazza (1992) in French, we showed that for the English speakers, only the LH was sensitive to the morphological complexity of the stimuli. Morphological complexity affected words and nonwords in the same manner, with complex stimuli requiring longer latencies to be identified either as a word or as a nonword only in the RVF. As opposed to the English speakers, both groups of speakers of the Semitic languages showed bilateral sensitivity to morphological complexity. In addition, the Arabic and Hebrew readers showed higher values on our indexes of interhemispheric integration, suggesting more intensive hemispheric cooperation during the reading of Hebrew and Arabic than of English. Interestingly, in both languages, morphological complexity had opposing effects for words and for nonwords. Morphological complexity, or transparency of the root þ word form structure, facilitated the recognition of words and decelerated the rejection of nonwords. We suggested that the nonconcatenative morphology of the Semitic languages, in which words are analyzed into their root and word-form constituents, requires that both hemispheres be sensitive to morphological structure. The automatic analysis of a character string into a recognizable word-form and a root resulted in faster recognition of complex words than of the simple words, which are not divisible in this way. This analysis also resulted in slower responses to complex nonwords than to simple nonwords, which did not contain a recognizable word form. Thus, the word form made complex nonwords more ‘‘word-like’’, requiring a more intensive search before they could be correctly rejected in the lexical decision task. In general, we found that the manner in which words are formed in these different languages resulted in different types of interhemispheric division of labor in the lexical decision task. Specifically, we showed that when languages

70

Z. Eviatar and R. Ibrahim

make different types of demands upon the cognitive system, interhemispheric interaction is dynamic and is suited to these demands. Arabic and Hebrew require a higher level of interhemispheric interaction than does English.

4.3 Reading in the Non-native Language Recently we have examined the interaction of the effects of reading Arabic and other languages (Ibrahim & Eviatar, 2009). We took advantage of the high facility of Arab university students in Hebrew and English, in order to examine the manner in which such multilingual brains deal with word morphology. We used the same paradigm described above, with trilingual Arab participants making lexical decisions on morphologically simple and complex words and nonwords, in Arabic, in Hebrew, and in English. One of the interesting results from this study is that in the RVF/LH there was a significant difference between performance levels in the three languages, reflecting significantly better performance in Arabic, which these participants learned to read first, than in Hebrew and English, which these participants consider their nonnative languages. However, in the LVF/RH, there was no difference between performance levels in the three languages. We interpret this as reflecting the specific RH deficit in reading Arabic, which lowers performance in the LVF for this language, such that it is not better than the second and third languages, in which these participants have lower facility. Another interesting result from this study is that the participants showed the same patterns of interhemispheric cooperation in the three languages, suggesting that they used the same reading strategies in all of the languages. For Hebrew, the patterns are similar to the ones shown by native Hebrew speakers, suggesting that morphological processes are similar in these similar languages. However, the patterns shown by native Arabic speakers in English are different from the patterns shown by native English speakers. Thus, our participants were reading a second (or third) language with the same mechanisms as the first learned language. This type of pattern was also reported by Eviatar (1999) for native Hebrew readers recognizing nonsense syllables in English. Eviatar (1999) suggested that this is due to the demand for morphological decomposition in Hebrew that determines reading strategies for other languages as well. The results reported here suggest that these same demands occur for Arabic readers.

4.4 Concluding Remarks In this chapter we discussed the Arabic literacy problem by taking up an important issue in the study of multilingual processing and languages’ representation. One of the critical issues discussed was whether the representations of native spoken language of Arab speakers (SA) and written Arabic (MSA) by literal

4

Multilingualism Among Israeli Arabs, and the Neuropsychology of Reading . . .

71

Arabic speakers are organized in a typical L2 to L1 relationship. The evidence we presented in our series of studies revealed that spoken and literary forms of Arabic are treated by proficient speakers/readers as different languages. This conclusion, in concert with previous findings (Wagner et al., 1989, Azzam, 1990; Ayari, 1996; Maamouri, 1998; Abu-Rabia, 2000; Saiegh-Haddad, 2003) suggests that the diglossic situation in Arabic might be related to some of the hardships that native Arabic beginning readers encounter and might even hinder their acquisition of basic academic skills. An additional important part of the Arabic literacy problem is posed by the Arabic orthographic system and its failure to support easy and efficient reading (Ibrahim et al., 2002). In subsequent studies we have shown that Arabic letters are harder to identify than English and Hebrew letters and have suggested that the locus of this difficulty is in the right hemisphere (Eviatar & Ibrahim, 2004; Eviatar et al., 2004). From the neuropsychological perspective, we discussed how an individual’s language experience influences the amount of hemispheric interaction during language processing, as indexed by the degree of difference in visual field asymmetry in unilateral versus bilateral presentation conditions during a visual lexical decision task. Under the premise that larger asymmetries during bilateral presentation reflect increased interhemispheric disruption and presumably therefore imply interhemispheric cooperation under normal conditions, we found that when languages make different types of demands upon the cognitive system, interhemispheric interaction is dynamic and is suited to these demands. Arabic and Hebrew require a higher level of interhemispheric interaction than does English (Eviatar & Ibrahim, 2007). In the light of above, we argue for inclusion of the neurofunctional perspective as a comprehensive basis for the discussion of teaching second language (L2). After all, teachers deal every day with the ability of students to learn and they necessitate the knowledge of the structural relationship between languages, of relevant pedagogical methods, to allow them monitoring the learning process, checking if it is optimally effective, and to intervene to shape it toward effectiveness.

References Abu-Rabia, S. (1997). Reading in Arabic orthography: The effect of vowels and context on reading accuracy of poor and skilled native Arabic readers. Reading and Writing: An Interdisciplinary Journal, 9, 65–78. Abu-Rabia, S. (1998). Reading Arabic texts: Effects of text type, reader type and vowelization. Reading and Writing: An Interdisciplinary Journal, 10, 105–119. Abu-Rabia, S. (2000). Effects of exposure to literary Arabic on reading comprehension in a diglossic situation. Reading and Writing: An Interdisciplinary Journal, 13, 147–157. Altarriba, J. (1992). The representation of translation equivalents in bilingual memory. In R. J. Harris (Ed.), Cognitive processing in bilinguals (pp. 157–174). Amsterdam: Elsevier. Ayari, S. (1996). Diglossia and illiteracy in the Arab world. Language, Culture and Curriculum, 9, 243–252.

72

Z. Eviatar and R. Ibrahim

Azzam, R. (1990). The nature of Arabic reading and spelling errors of young children: A descriptive study. Unpublished doctoral dissertation, Teachers College, Columbia University, New York. Bentin, S., Bargai, N., & Katz, L. (1984). Graphemic and phonemic coding for lexical access: Evidence from Hebrew. Journal of Experimental Psychology: Learning, Memory and Cognition, 10, 353–368. Bentin, S., & Frost, R. (1987). Processing lexical ambiguity and visual word recognition in a deep orthography. Memory & Cognition, 15, 13–23. Ben Zeev, S. (1977). The influence of bilingualism on cognitive strategy and cognitive development. Child Development, 48, 1009–1018. Berent, I. (2002). Identity avoidance in the Hebrew lexicon: Implications for symbolic accounts of word formation. Brain and Language, 81(1–3), 326–341. Berman, R. A. (1978). Modern Hebrew structure. Tel Aviv, Israel: University Publishing. Bialystok, E. (1991). Language processing in bilingual children. Cambridge, England: Cambridge University Press. Bialystok, E., & Ryan, E. B. (1985). Toward a definition of metalinguistic skill. MerrillPalmer Quarterly, 31, 229–251. Burgess, C., & Skodis, J. (1993). Lexical representation and morpho-syntactic parallelism in the left hemisphere. Brain and Language, 44, 129–138. Campbell, R., & Sais, E. (1995). Accelerated metalinguistic (phonological) awareness in bilingual children. British Journal of Developmental Psychology, 13(1), 61–68. Dash, U. N., & Mishra, H. C. (1992). Bilingualism and metalinguistic development: Evidence from Kond tribal culture. Psychological Studies, 37(2–3), 81–87. de Groot, M. B. A. (1992). Determinants of word translation. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 1000–1018. de Groot, A. M. B. (1995). Determinants of bilingual lexicosemantic organization. Computer Assisted Language Learning, 8, 151–180. Deutsch, A., Frost, R., & Forster, K. I. (1998). Verbs and nouns are organized and accessed differently in the mental Lexicon: Evidence from Hebrew. Journal of Experimental Psychology: Learning, Memory and Cognition, 24(5), 1238–1255. Eid, M. (1990). Arabic linguistics: The current scene. In M. Eid (Ed.), Perspectives on Arabic linguistics I (pp 3–37). Amsterdam: John Benjamins Publishing Company. Eviatar, Z. (1999). Cross-language tests of hemispheric strategies in reading nonwords. Neuropsychology, 13(4), 1–18. Eviatar, Z., Hellige, J. B., & Zaidel, E. (1997). Individual differences in hemispheric specialization: Effects of gender and handedness. Neuropsychology, 11, 562–576. Eviatar, Z., & Ibrahim, R. (2001). Bilingual is as bilingual does: Metalinguistic abilities of Arabic-speaking children. Applied Psycholinguistics, 21(4), 451–471. Eviatar, Z., & Ibrahim, R. (2004). Morphological and orthographic effects on hemispheric processing of nonwords: A cross-linguistic comparison. Reading and Writing, 17, 691–705. Eviatar, Z., & Ibrahim, R. (2007). Morphological structure and hemispheric functioning: The contribution of the right hemisphere to reading in different languages. Neuropsychology, 21, 470–484. Eviatar, Z., Ibrahim, R., & Ganayim, D. (2004). Orthography and the hemispheres: Visual and linguistic aspects of letter processing. Neuropsychology, 18(1), 174–184. Farid, M., & Grainger, J. (1996). How initial fixation position influences visual word recognition: A comparison of French and Arabic. Brain and Language, 53, 351–361. Ferguson, C. A. (1959). Diglossia. In P. P. Giglioli (Ed.), Language and social context (pp. 232–251). London: Penguin. Frost, R. (1994). Prelexical and postlexical strategies in reading: Evidence from a deep and a shallow orthography. Journal of Experimental Psychology: Learning, Memory, and Cognition, 20, 1–16.

4

Multilingualism Among Israeli Arabs, and the Neuropsychology of Reading . . .

73

Frost, R. (1995). Phonological computation and missing vowels: Mapping lexical involvement in reading. Journal of Experimental Psychology: Learning, Memory and Cognition, 21 (2), 398–408. Frost, R., & Bentin, S. (1992). Processing phonological and semantic ambiguity: Evidence from semantic priming at different SOAs. Journal of Experimental Psychology: Learning Memory and Cognition, 18(1), 58–68. Frost, R., Katz, L., & Bentin, S. (1987). Strategies for visual word recognition and orthographical depth: A multilingual comparison. Journal of Experimental Psychology: Human Perception and Performance, 13, 104–114. Galambos, S. J., & Goldin Meadow, S. (1990). The effects of learning two languages on levels of metalinguistic awareness. Cognition, 34(1), 1–56. Galambos, S. J., & Hakuta, K. (1988). Subject specific and task specific characteristics of metalinguistic awareness in bilingual children. Applied Psycholinguistics, 9(2), 141–162. Iacoboni, M., & Zaidel, E. (1996). Hemispheric independence in word recognition: Evidence from unilateral and bilateral presentations. Brain and Language, 53, 121–140. Ibrahim, R. (1998). The cognitive representation of literary and spoken Arabic languages: A comparison within and between languages. Unpublished Ph.D. dissertation, Hebrew University, Israel. Ibrahim, R. (2006). Morpho-Phonemic similarity within and between languages: A factor to be considered in processing Arabic and Hebrew. Reading and Writing, 19(6), 563–586. Ibrahim, R., & Aharon-Peretz, J. (2002). Text processing and perceptual characteristics of Arabic orthograpy. Neurology Update, 9, 22–24 (Hebrew). Ibrahim, R., & Aharon-Peretz, J. (2005). Is literary Arabic a second language for native Arab speakers? Evidence from a semantic priming study. The Journal of Psycholinguistic Research, 34(1), 51–70. Ibrahim, R., & Eviatar, Z. (2009). Language status and hemispheric involvement in reading: Evidence from trilingual Arabic speakers tested in Arabic, Hebrew, and English. Neuropsychology, 23, 240–254. Ibrahim, R., Eviatar, Z., & Aharon Peretz, J. (2002). The characteristics of the Arabic orthography slow it’s cognitive processing. Neuropsycholgy, 16(3), 322–326. Keatley, C., & DeGelder, B. (1992). The bilingual primed lexical decision task: Cross language priming disappears with speeded responses. European Journal of Cognitive Psychology, 4 (4), 273–292. Koenig, O., Wetzel, C., & Caramazza, A. (1992). Evidence for different types of lexical representations in the cerebral hemispheres. Cognitive Neuropsychology, 9, 33–45. Koriat, A. (1984). Reading without vowels: Lexical access in Hebrew. In H. Bouma & D. G. Bouwhuis (Eds.), Attention and performance X: Control of language processes (pp. 227–242). London, England: Lawrence Erlbaum. Koriat, A. (1985a). Lateralization effects in reading pointed and unpointed Hebrew. The British Journal of Psychology, 76, 161–173. Koriat, A. (1985b). Lexical access for low and high frequency words in Hebrew. Memory & Cognition, 13, 37–44. Lamb, M. R., Robertson, L. C., & Knight, R. T. (1990). Component mechanisms underlying the processing of hierarchically organized patterns: Inferences from patients with unilateral cortical lesions. Journal of Experimental Psychology: Learning, Memory, and Cognition, 16, 471–483. Levy, J., Heller, W., Banich, M. T., & Burton, L. (1983). Are variations among right handed individuals in perceptual asymmetries caused by characteristic arousal differences between hemispheres? Journal of Experimental Psychology: Human Perception and Performance, 9, 329–359. Maamouri, M. (1998). Language education and human development: Arabic diglossia and its impact on the quality of education in the Arab region. The Mediterranean Development Forum. Washington, DC: The World Bank.

74

Z. Eviatar and R. Ibrahim

McClelland, J. L., Rumelhart, D. E., & Hinton, G. E. (1986). The appeal of parallel distributed processing. In D. E. Rumelhart & J. L. McClelland (Eds.), Parallel distributed processing (pp. 3–44). Cambridge, MA: MIT Press. Navon, D. (1977). Forest before trees: The precedence of global features in visual perception. Cognitive Psychology, 9, 353–383. Navon, D., & Shimron, J. (1981). Does word naming involve grapheme-to-phoneme translation? Evidence from Hebrew. Journal of Verbal Learning and Verbal Behavior, 20, 97–109. Olshtain, E., & Rozovsky R. (2004). The acquisition of reading literacy in Israel: Findings from the international reading literacy research PIRLS-2001. The Ministry of Education and the Centre for Scientific and Technological Education at Tel-Aviv University. Tel-Aviv: Ramot Publishing, Israel (In Hebrew). Peleg, O., & Eviatar Z. (2009). Hemispheric sensitivities to lexical and contextual information: Evidence from lexical ambiguity resolution. Brain and Language, 70, 154–162. Prunet, J., Beland, R., & Idrissi, A. (2000). The mental representation of Semitic words. Linguistic Inquiry, 31(4), 609–648. Reitan, R. M., & Wolfson, D. (1993). The Halsted reitan neuropsychological test battery: Theory and clinical interpretation (2nd ed.). Tucson, AZ: Neuropsychology Press. Roman, G., & Pavard, B. (1987). A comparative study: How we read Arabic and French. In J. K. O’Regan & A. Levy-Schoen (Eds.), Eye movement: From physiology to cognition (pp. 431–440). Amsterdam, The Netherlands: North Holland Elsevier. Saiegh-Haddad, E. (2003). Linguistic distance and initial reading acquisition: The case of Arabic diglossia. Applied Psycholinguistics, 24, 431–451. Saiegh-Haddad, E. (2004). The impact of phonemic and lexical distance on the phonological analysis of words and pseudo words in a diglossic context. Applied Psycholinguistics, 25, 495–512. Saiegh-Haddad, E. (2005). Correlates of reading fluency in Arabic: Diglossic and orthographic factors. Reading and Writing: An Interdisciplinary Journal, 18, 559–582. Shimron, J., & Sivan, T. (1994). Reading proficiency and orthography: Evidence from Hebrew and English. Language Learning, 44, 5–27. Smolka, E., & Eviatar, Z. (2006). Phonological and orthographic visual word recognition in the two cerebral hemispheres: Evidence from Hebrew. Cognitive Neuropsychology, 23(6), 972–989(18). Titone, R. (1994). Bilingual education and the development of metalingusitic abilities: A research project. International Journal of Psycholinguistics, 10(1), 5–14. Wagner, D. A. (1993). Literacy, culture, & development: Becoming literate in Morocco. Cambridge: Cambridge University Press. Wagner, D. A., Spratt, J. E., & Ezzaki, A. (1989). Does learning to read in a second language always put the child at a disadvantage? Some counter evidence from Morocco. Applied Psycholinguistics, 10, 31–48.

Part II

Academic Achievement of Children Coming from Immigrant Families

Chapter 5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable School Children – The Case of Ethiopian Israeli Children Michal Shany and Esther Geva

5.1 Introduction Israel is an immigration country. New immigrants are faced with the challenges of acculturation to new norms, values and expectations (Berry, 2003). This chapter is concerned with the children of immigrants who came to Israel from Ethiopia in the last two decades. According to official statistics of the Israeli Municipalities (2006, 2007), children of Ethiopian families constitute 10% of the immigrant children living in Israel. Various reports suggest that the school dropout rates in this community is twice as high as in the general population. The picture emerging from various reports (e.g., Levin et al., 2002) and from the media (e.g., Yediot Axronot, January 1, 2008) is that the academic achievement of children in this group is especially low, and that a disproportionate number of children from this community are placed in special education programs (Berhanu, 2005). Factors associated with different cultural norms and expectations (Levin-Rozalis & Shafran, 2003; Berhanu, 2005; Bar-Yosef, 2001), conformity to strongly entrenched traditional norms (for example, obedience to elders of the family and the entire adult community, being quiet and respectful), different patterns of socialization, and external locus of control, have been offered as processes that contribute to the academic failure of individuals from this group in the host country (Horowitz & Mosher, 1997; Korten, 1972). Concomitant with the absence of literacy skills in their home language, the Ethiopian Israeli students need to develop their language and literacy skills in a language (Hebrew) that is different from the language spoken at home (Amharic). To date, most reports involving students from the Ethiopian immigrant community have focused on gaps in general academic achievement between this group and other immigrant groups, or in comparison to the general population (e.g., Levin et al., 2002; Meitzav, 2006; Brookdale Institute, 2003). Indeed, there is a growing body of research focusing on the language and literacy skills of M. Shany (*) Department of Learning Disabilities, Faculty of Education, University of Haifa, Haifa, Israel e-mail: [email protected] M. Leikin et al. (eds.), Current Issues in Bilingualism, Literacy Studies 5, DOI 10.1007/978-94-007-2327-6_5, Ó Springer ScienceþBusiness Media B.V. 2012

77

78

M. Shany and E. Geva

various groups of immigrants, but this research has involved immigrants who are literate to varying degrees in their first language (L1). To date, however, no studies examined systematically, and in depth the development of components of language and literacy development in a large representative sample of Ethiopian students. This chapter addresses this dearth. Given global demographic immigration trends, a careful analysis of the situation in the Ethiopian immigrant community in Israel is important because it likely mirrors other similar immigrant communities elsewhere. In particular, even though in the last decade the number of studies examining the development of language and literacy skills in second language (L2) learners has mushroomed, very little is known about the development of language and literacy skills in immigrant school children who are expected to transition rapidly from an oral society to a Westernized, literate society. In this chapter, we report on research that targeted development in a number of related domains that underlie literacy development. In particular, the research focused on the cultural, cognitive, language and literacy development of this vulnerable group of minority children, whose parents made an abrupt transition from Ethiopia to Israel in the 1980s and 1990s (Bender, 1985; Shenkut, 1991). This cross-sectional study of Ethiopian Israeli children in grades one-to-six provides a window into development in the cognitive, linguistic, cultural, and literacy domains. We compare the performance of the Ethiopian Israeli children to that of their Non-Ethiopian peers, who live in the same poor communities, and attend the same classes. The following key questions are examined in this chapter: 1. Are there gaps between low SES Ethiopian Israeli and Non-Ethiopian groups on cultural, cognitive, linguistic, and literacy dimensions? 2. In what domains do skill gaps that exist in the primary grades disappear in higher grades? 3. In what domains do skills gaps persist or increase? In order to better understand the issues facing the Ethiopian Israeli community it is essential to consider key aspects of the historical, socio-cultural, economic, and educational context of this community. It is also important to discuss briefly the research literature concerning the development of language and literacy skills in the societal language of children who speak a different language at home. Before we launch into a description of the study, we provide in the next two sections an overview of these two areas.

5.1.1 The Ethiopian Israeli Community: The Socio-cultural Context 5.1.1.1 Transition from Ethiopia to Israel Along with other groups, many Ethiopian Jews fled from Ethiopia to Sudan in the 1970s due to a civil war and famine, and lived in refugee camps. Subsequently, many Ethiopian Jews arrived in Israel in a number of waves in the

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . .

79

mid-1980s and 1990s. The migration experience was especially traumatic (Bar-Yosef, 2001) and arrival in Israel led to major changes in community organization (Berhanu, 2005). The bulk of this community was rural and there is evidence that the communal, patriarchal and interconnected community structure already began to crumble with the transition to refugee camps. This process continued with the abrupt cultural changes that took place as a result of contact with Israeli society (Ben-Ezer, 1992; Berhanu, 2005; Bodovski, David, & Eran, 1994; Kaplan, 1992). The majority of Ethiopian immigrants came from a rural, oral tradition society, where over 90% of the community members were illiterate (Bender, 1985; Shenkut, 1991). In Ethiopia they lived within a traditional structure, consisting of compounds of the larger kinship group. The communities were led by elders and traditional priests who played an important role in maintaining and transmitting communal-cultural knowledge (Bodovski et al., 1994). 5.1.1.2 Employment, Education, Acculturation, and Language Proficiency The declared policy of the Israeli government is to spread the immigrants from Ethiopia throughout the country. However, according to an official report of the Israeli Knesset (Parliament) in Israel (Vertsberger & Noyfeld, 2003), based on the 2002 census, most congregate in a small number of towns in the North and Center of Israel, and live in very poor neighborhoods. Most Ethiopian immigrants are unskilled workers, and the unemployment rate among the Ethiopian community living in Israel is higher than in the general population (Svirsky & Svirsky, 2002). Over 52% of Ethiopian families live below the poverty line, in comparison with 16% in the general population (Bank of Israel, 2006). Wide cultural, social, occupational, and literacy differences between the community of Ethiopian origin and the host Israeli society persist (Bodovski et al., 1994; Levin-Rozalis, 2000; Gaon, King, & Volda-Tzadik, 2006; Offer, 2004). More recent statistics provide a more promising picture. Whereas in the mid-90s, 55% of adults of Ethiopian origin ages 22–64 had no formal education, in recent years this rate has dropped to 42%. Another promising statistic is the fact that the number of individuals who obtained at least high school education has increased from 6% to 15% (Israel Statistical Municipalities, 2006, 2007). As might be expected in any group of immigrants, about 40% of the immigrants who came in the 1980s and 1990s, and 60% of more recent immigrants from Ethiopia indicate that they have only minimal command of Hebrew. In addition, 55% of the male adults cannot read or write Hebrew, and this percentage is even higher among women (70%) (King, Effrati, & Netzer, 2003). In comparison with other immigrant groups, the linguistic integration of immigrants from Ethiopia has been alarmingly slow (Barkon & Avinor, 1995, Bar-Yosef, 2001; Spolsky, 1996). Upon arrival to Israel, Ethiopian adults could attend a 1-year intensive Hebrew language and literacy program, known as Ulpan. In our sample about 35% of the mothers and fathers attended this

80

M. Shany and E. Geva

basic, Hebrew based, adult education program, and for many this was the first time they were exposed to written language of any sort. In Ethiopian culture, the main responsibility of fathers is to ensure that their children behave properly and respect the elders of the family (Berhanu, 2005; Horowitz & Mosher, 1997; Korten, 1972; Yediot Axronot, January 1, 2008). But in Israel the children speak Hebrew better than their parents, and as a result the balance of power in the home has shifted to other family members, and the fathers do not bare the same authority (Segal, 2006). Once their children begin to attend school Ethiopian Israeli parents tend to be less involved, since they feel that the school is an official authority that bears responsibility for their children’s education. However, the schools interpret this lack of involvement as lack of interest (Bar-Yosef, 2001; Stravas & Olshtain, 2006). Most Ethiopian immigrant families do not share with the Israeli school system a language or concepts about the actual role that parents in other segments of Israeli society are likely to undertake in the ‘‘education enterprise’’ of their children (Stravas & Olshtain, 2006). In this regard, it is useful to borrow from sociology the ‘‘cultural capital’’ framework (Bourdieu & Passeron, 1990). Cultural capital involves the maintenance and transmission of forms of knowledge, values, education, and expectations, and includes the interplay of individual values and attributions, cultural goods (e.g., books, computers), and educational institutions such as academic credentials and diplomas. According to Bourdieu and Passeron (1990) disadvantages and inequalities are maintained from one generation to the next, partly through the educational system that reproduces social structures. Using this framework, it is possible to see that the Ethiopian Israeli community does not possess the cultural capital resources that would enable it to transmit to their children the modes of language, literacy, and strategic skills that are crucial for further learning in a westernized culture. From a socio-cultural perspective, it is obvious that there is no ‘‘fit’’ between children’s home culture and the culture of the school and classroom. According to this view, poor academic performance does not necessarily reflect the negative influence of children’s home cultural characteristics on their learning, but rather, the inability of schools to adapt instruction to relevant aspects of children’s home cultures (Rueda, August, & Goldenberg, 2006). In spite of various well-intentioned programs, it has been difficult to bridge the difference between Ethiopian and Israeli cultural values, attributions, and expectations. The prevailing societal view is one of deficit, and as a result the community has been further alienated from the mainstream culture. 5.1.1.3 Academic Achievement A recent, large scale, comprehensive study of academic achievement in mathematics and literacy skills compared the achievement of children of Ethiopian and Russian immigrants to that of non-immigrant Israeli students, in Grades 5, 10 and 11 (Levin et al., 2002). In general, achievement scores in both immigrant groups were lower than those of the non-immigrant comparison group.

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . .

81

However, the researchers found that the literacy skills of immigrant students from Russia catch up after 10–11 years of schooling in the societal language, and already by Grade 9 the gap is rather small. However, the performance of youth of Ethiopian origin is consistently lower than that of the Russian immigrant and non-immigrant students, and the gaps in academic achievement continue to be rather large even for the 10% of Ethiopian students who complete their high school education (Levin et al., 2002). According to Barkon and Avinor (1995), young children growing up in Ethiopian families are exposed to oral Amharic at home, but immigrant adolescents have lost their L1 (Amharic) skills, and they continue to have poor Hebrew language proficiency. Other research has shown that, over and above Socio-Economic Status (SES), parental education correlates with educational outcomes for immigrant children (Hooddinoot, Lethbridge, & Phipps, 2002; Willms, 2003). In addition, research has shown that both home-based and community-based variables mediate language and literacy achievement and are significant predictors of language and school achievement (Leseman & de Jong, 1998; Willms, 2003). In this regard, one would not expect perhaps that children of Ethiopian Israeli immigrants who live in poverty should fare any better than children of other communities who live in poverty. However, as is evident from the discussion above, in this community one notes the interaction of a number of familial and community level risk factors, including the break-up of traditional institutions, and diminishing exposure to traditional modes of literacy (Segal, 2006). Other interacting factors include societal prejudice, lack of fit in cultural resources between the culture of origin and the host country, poverty and living in poor communities, and having parents who come from a culture that is predominantly oral. Poverty interacts with societal power relations that prevent the minority group from ensuring that aspects of the curriculum, assessment, and teacher education reflect and acknowledge the values of the vulnerable community and its modes of transmission (Cummins, in press). On the whole, these factors come together to create a challenging backdrop against which children of immigrants from Ethiopia need to develop language and literacy skills and achieve academically.

5.2 Theoretical Framework: L2 Literacy Development Reading, writing and language skills are complex cognitive activities involving sub-components that interact with each other and affect each other. Even though the Ethiopian Israeli children are known to be underachieving academically, it is not known to what extent they underachieve on all aspects of language and literacy, nor has their performance on various language and literacy skill components been examined developmentally. The recent literature on language and literacy development in second language (L2) learners is of some relevance in the present context.

82

M. Shany and E. Geva

5.2.1 Word Level Skills in L2 Learners Some might expect that due to lack of language proficiency in the target language, students learning to read in L2 might have poorer word level reading skills than monolingual (L1) students. However, the conclusions of a recent comprehensive systematic review conducted by the National Literacy Panel (NLP), that focused on research on students who develop their language and literacy skills in a language different from their home language (August & Shanahan, 2006), are more nuanced. It appears that the difference between language-minority students and their native-speaking counterparts on word level skills such as word recognition, decoding, and spelling is not very robust. This general conclusion is based on studies involving immigrant children learning English (the societal language) as their second language (ESL) with various home languages, including Punjabi (Chiappe & Siegel, 1999), Urdu (Mumtaz & Humphreys, 2001), Italian (D’Angiulli, Siegel, & Maggi, 2004), Portuguese (Da Fontoura & Siegel, 1995), Spanish (Durgunoglu, Nagy, & Hancin-Bhatt, 1993; Lindsey, Manis, & Bailey, 2003) and from various other ESL backgrounds (Lesaux & Siegel, 2003; Chiappe, Siegel, & Wade-Woolley, 2002; Geva, Yaghoub-Zadeh, & Schuster, 2000; WadeWoolley & Siegel, 1997). This conclusion has also been reached with regard to other language combinations such as Turkish-Dutch (e.g., Verhoeven, 1990), Arabic-English (Abu-Rabia & Siegel, 2002), English-Farsi (Arab-Moghaddam & Se´ne´chal, 2001; Gholamain & Geva, 1999) and French-English, (Comeau, Cormier, Grandmaison, & Lacroix, 1999; Lafrance & Gottardo, 2005). The overall conclusion from these studies is that accurate word level skills are not heavily dependent on oral language proficiency. Of particular relevance are a number of studies that focused on the word level skills of English speaking children who were learning Hebrew in bilingual English-Hebrew schools (Geva & Siegel, 2000; Geva & Wade-Woolley, 1998; Geva, Wade-Woolley & Shany, 1997; Wade-Woolley & Geva, 2000). These studies have shown that in spite of limited language proficiency in Hebrew, these middle-class children can learn to decode and spell Hebrew words with accuracy, and decode Hebrew words with efficiency. These outcomes can be understood to some extent by considering typological differences in orthographic and language features (Katz & Frost, 1992). Thanks to the simplicity of the phonological structure of Hebrew (Ziegler & Goswami, 2005; Share, 2008), the lexical outcome of assembling into words a series of matched graphemephonemes in vowelled Hebrew is unequivocal, and can be accomplished with accuracy even in the absence of linguistic proficiency. In this regard, Share (2008) concludes in a recent review that ‘‘These findings converge on the conclusion that both lexical and extra-lexical factors, such as syntax and vocabulary, are superfluous when spelling-sound relations are straightforward, but are indispensable when the orthography is opaque.’’ This observation should apply to immigrant children whose Hebrew language proficiency is underdeveloped, including the Ethiopian Israeli children who were the focus of this study.

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . .

83

5.2.2 Reading Comprehension and L2 Learners Successful reading comprehension requires the integration of various component skills, including accurate and efficient word reading, reading fluency, awareness of text features, language proficiency (including vocabulary, morphological, grammatical, and pragmatic skills), background knowledge, and various reading strategies (Deacon & Kirby, 2004; Mu¨ller & Brady, 2001; Pressley, 1998; National Reading Panel, [NRP], 2000; RAND). Recent studies of language-minority students indicate that the reading comprehension performance of minority students falls well below that of their native-speaking peers (Lesaux & Geva, 2006). As part of the NLP systematic review, the authors of one of the chapters (Lesaux & Geva, 2006) conclude that numerous variables at the individual level (e.g., L2 vocabulary, background knowledge, motivation) and contextual level (e.g., familiarity with text structure conventions, home literacy, demographics) affect the second-language reading comprehension of language-minority students. Various variables contribute to L2 reading comprehension, including word level reading skills such as automatic word recognition and decoding skills (Reese, Garnier, Gallimore, & Goldenberg, 2000; Verhoeven, 1990), metalinguistic awareness (Carlisle, Beeman, Davis, & Spharim, 1999), background and cultural knowledge (Carlisle et al., 1999; Droop & Verhoeven, 2003; Lee & Schallert, 1997), the use of cohesion markers and reading strategies (Padro´n & Waxman, 1988), and language skills, including vocabulary, morphological and syntactic knowledge (Droop & Verhoeven, 2003; Geva & Yaghoub-Zadeh, 2006; Nakamoto, Lindsey, & Manis, 2008; Proctor, Carlo, August, & Snow, 2005; Yaghoub Zadeh, Farnia, & Geva, 2011). As noted above, in spite of poorer L2 language proficiency, L2 learners do not typically differ on basic word-based skills such as word recognition, pseudoword decoding and spelling. However, L2 oral language comprehension skills are strongly and consistently related to text comprehension. As noted in the NLP systematic review, demographic factors such as parental education and poverty indicators have also been consistently shown to be related to reading comprehension of minority students (Lesaux with Koda, Siegel, & Shanahan, 2006). In general, the more home literacy experiences and opportunities children have, the more likely they are to do well on literacy outcomes (Goldenberg et al., 2006). In addition, research on the contributions of home practices has shown that the attitudes, literacy related activities, and the degree of linguistic sophistication that children are exposed to at home from an early age have a long lasting impact on later academic achievement (August & Shanahan, 2006). However, for children who come from cultures that do not promote written language literacy, there may not be a fit between the language, learning and strategies that are utilized and valued in the L2 environment and the skills and values fostered at home.

84

M. Shany and E. Geva

According to the ‘‘Linguistic Interdependence Hypothesis’’ (Cummins, 1991), reading skills in the native language (L1) will transfer to reading in the L2, provided that an adequate level proficiency in the L2 has been achieved. As part of the NLP review Dressler and Kamil (2006) conclude that there is evidence for cross-language transfer of reading comprehension ability in bilinguals speaking different languages (e.g., Lee & Schallert, 1997; Reese et al., 2000; Royer & Carlo, 1991; Verhoeven, 1994). Indeed, transfer effect has been noted in various studies involving minority language students. Importantly, Gersten (1999) cautions, however, that transfer of skills from the L1 to the L2 will occur when well trained teachers deliver adequate L2 instruction. In the present context, where there is no systematic exposure to the language and literacy skills of the home language and where the teachers are less likely to be trained in adequate L2 instruction, such transfer is less likely to occur.

5.2.3 Underlying Cognitive-Linguistic Processes in L1 and L2 Moving from contextual to intra-individual factors, studies examining underlying processing skills have consistently found that cognitive skills such as phonological processing skills, including phonemic awareness, rapid lexical access, and phonological memory, predict word identification skills in English, the L2. (For more details see the conclusion of the NLP report, August & Shanahan, 2006). Another body of research has examined cognitive and neurological processes that underlie the reading process in spite of typological differences in the spoken language or the orthography. The focus of these studies has been on examining ‘‘universal’’ cross-linguistic commonalities and ‘‘typological’’ differences in the role processes such as phonemic awareness and rapid serial naming. Support for universal commonalities comes, for example, from studies comparing good and poor readers across orthographies (e.g., Da Fontoura & Siegel, 1995; Katzir, Shaul, Breznitz, & Wolf, 2004). It suggest that regardless of the orthographies involved, bilinguals or L2 learners who have decoding and spelling problems in their native language will have difficulties in their L2 as well. Evidence for typological effects comes from research on bilingual children (e.g., Geva & Siegel, 2000; Geva, Wade-Woolley & Shany, 1993; Mumtaz & Humphreys, 2001; Wang & Geva, 2003; Wade-Woolley & Geva, 2000). This research suggests that the type of reading and spelling errors observed in novice L2 students reflects typological influences. On the whole, this research suggests that the ‘‘typological’’ and ‘‘universal’’ frameworks are complementary (Geva, 2007). In his review, Share (2008) argues that ‘‘accuracy. . .is largely a non-issue for the majority of the worlds’ (alphabetic) orthographies in which performance levels approach ceiling by the end of Grade 1 (Seymour, Aro, & Erskine, 2003)’’, and that ‘‘when accuracy asymptotes quickly, speed and fluency become the discriminating measures of developmental and individual differences. . .’’. In fact, this point is more or less relevant also when ESL readers are compared to non-ESL readers. For example, Geva and Yaghoub-Zadeh (2006) examined the

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . .

85

extent to which ESL and monolingual English speakers are similar on various aspects of reading fluency, and whether reading fluency in ESL and monolingual children can be understood in terms of similar underlying processing components. In spite of an oral language proficiency advantage in the English monolingual group, they did not have an advantage on any of the cognitive and reading measures. Two robust predictors of word reading fluency in both language groups were rapid automatized naming (RAN) and phonological awareness (PA), though oral sentence comprehension made a modest contribution as well. Of particular relevance here are two additional studies by Geva and her colleagues (Geva et al., 1993, 1997). Geva et al. (1993) provide evidence that phonemic awareness may play a more temporary role in learning to read Hebrew than it does in English. In that study the authors found a correlation of 0.62 between word reading and phonemic awareness in children’s L1 (English), but only 0.32 for Hebrew (the L2). Geva et al. (1997) found that reading efficiency indices (based on accuracy and speed) of word recognition in L1 (English) and L2 (Hebrew) were very similar, though when it came to reading simple narratives children were more efficient in English, their L1, than in Hebrew, a language in which their language proficiency was minimal. On the whole then, it appears that processing components such as phonemic awareness and rapid automatized naming (RAN) underlie word-based skills of L2 learners of alphabetic languages, but language proficiency is important for text fluency and text comprehension. There is ample evidence that rapid automatized naming may be a strong(er) predictor of reading fluency in regular writing systems such as Hebrew and Dutch (Breznitz, 2006; de Jong & van der Leij, 2003; Wimmer, 1993) than in writing systems that are not as regular, such as English. As for L2 children, there is evidence that ESL readers may be slower at naming simple items than their L1 counterparts at the beginning of school, but that RAN speed improves and catches up with their L1 peers (e.g., Gholamain & Geva, 1999; Lesaux & Siegel, 2003; Geva & Yaghoub Zadeh, 2006). RAN predicts reading accuracy and fluency in primary level ESL readers (Geva et al., 2000; Geva & Yaghoub-Zadeh, 2006; Lesaux & Siegel, 2003). Given this literature, there is good reason to expect that the performance of Ethiopian Israeli minority children whose Hebrew language skills are developing at school would improve at the same rate as their monolingual counterparts on underlying cognitive processing components such as naming speed and phonemic awareness. However, given the available literature it may be expected that the performance of Ethiopian Israeli minority children on higher level, more complex language and literacy tasks may fall increasingly behind. On the whole, this brief literature review suggests that Ethiopian Israeli minority children will not display serious difficulties on simple cognitive underlying skills related to word-based reading skills, nor will they experience difficulties in developed word-based reading skills. However, it is reasonable to hypothesize that their Hebrew (L2) language skills will continue to be underdeveloped and that, relatedly, they will be having difficulties with reading comprehension.

86

M. Shany and E. Geva

5.3 Method 5.3.1 Participants 5.3.1.1 Children The sample consisted of 326 children, 175 Ethiopian Israeli and 151 NonEthiopians. The participants were drawn from Grades 1, 2, 4, and 6, in six towns in the North and Center of Israel. The Ethiopian Israeli and NonEthiopian participants within each school in each grade level came from the same classrooms. This design enabled us to minimize confounding due to instructional and neighborhood differences between the two groups. The percent of Ethiopian children in each classroom ranged from 48% to 58%, with an average of 53%. The sample consisted of 152 boys, (88 Ethiopians and 64 NonEthiopians), and 174 girls, (87 Ethiopians and 87 Non-Ethiopians). Table 5.1 presents the distribution of participants by group, gender, and grade level. A chisquare analysis, examining the distribution of males and females was not significant (Chi-square (1,1) ¼ 2.034), p ¼ 0.154). Ethiopian and Non-Ethiopian students lived in the same poor neighborhoods. All the Ethiopian participants were students whose families immigrated to Israel from Ethiopia, who were born in Israel or came to Israel before the age of 2. All children from the comparison group of non-Ethiopians were born in Israel to non-Ethiopian families. 5.3.1.2 Families Home Language and Literacy We first present descriptive information about home language, literacy, and poverty indices in the Ethiopian Israeli in comparison with the non-Ethiopian families, whose children participated in the study. We were unable to interview parents directly due to limited access to bilingual interpreters. Instead, this information was obtained through questionnaires filled out by the classroom teachers and the school-home liaisons. The Teacher Questionnaire included questions about demographic and family language and literacy factors. The questionnaire was divided into two main sections. The first section focused on the children and included items such age of onset of kindergarten education, country of birth, gender, spoken language(s) and level of mastery (for target Table 5.1 Distribution of participants by ethnicity, gender and grade levels Ethiopians Non-Ethiopians Grade level Boys Girls Total Boys Girls Total Total sample Grade 1 Grade 2 Grade 4 Grade 6 Total sample

34 18 19 17 87

24 24 20 19 87

58 42 39 36 175

20 14 14 16 64

22 20 22 23 87

42 34 36 39 151

100 76 75 75 326

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . .

87

children and their parents). The second part focused on family variables such as parental occupation, number of children and other adults living with the family, number of rooms, and level of mastery of spoken and written language skills in Hebrew and Amharic. It should be noted that since all participants were drawn from the same neighborhoods and schools, inspection of the questionnaires revealed that the reported demographic data did not differ across grades. Therefore, we report the aggregate data for the whole sample. Among the Ethiopian mothers, 1.1% spoke only Hebrew, 72.5% spoke a mixture of Hebrew and Amharic, and 26.5% spoke only Amharic. Among the non-Ethiopian mothers, 57.1% spoke only Hebrew at home and 42.9% spoke a mixture of Hebrew and one of 9 other languages (including Russian, Kavkaz, followed by French, Persian, Kurdish, Romanian, Spanish, and Portuguese). The difference in distribution between the Ethiopian and non-Ethiopian groups in terms of maternal literacy (in Hebrew and/or another language) was significant (Chi-square (3,1) 347.16, p < 0.0001). The patterns are highly similar for fathers, except that among Ethiopian fathers (in contrast with Ethiopian mothers), 30.5% speak only Amharic, and have no proficiency in Hebrew. It was impossible to collect reliable information about parental education. Based on national surveys, Gould, Lavy, and Paserman (2004) report that parents of Ethiopian children typically have 1 or 2 years of education, compared to an average of over 11.5 years for parents of Israeli Jewish children. There is no reason to believe that these national patterns do not apply to the Non-Ethiopian parents in our sample. Poverty A series of t-tests was used to compare the groups on available poverty indicators. Significant differences between the Ethiopian and Non-Ethiopian groups were found on: (a) mean number of rooms at home: M ¼ 3.45 (SD ¼ 0.90) in the Ethiopian families versus M ¼ 4.09 (SD ¼ 1.28) in non-Ethiopian families (t (1, 315)¼ 5.07, p < 0.0001); (b) mean number of family members per household: M ¼ 7.41 (SD ¼ 2.21) in the Ethiopian families versus M ¼ 5.54 (SD ¼ 1.26) in non-Ethiopian families (t (1,172) ¼ 7.02, p < 0.0001); and (c) density (mean number of persons per room): M ¼ 4.50 (SD ¼ 2.23) in the Ethiopian families versus M ¼ 2.33 (SD ¼ 1.51) in non-Ethiopian families (t (1, 165) ¼ 7.77, p < 0.0001). On all these indicators the Ethiopian families fared worse than the Non-Ethiopian families. Parent employment was rated according to a three-level scale: 1 ¼ unemployed; 2 ¼ unskilled work; 3 ¼ skilled work. The results indicated that the vocational training level of the NonEthiopians was significantly higher, M ¼ 2.27 (SD ¼ 0.63) versus M ¼ 1.83 (SD ¼ 0.69) among the Ethiopian workers (t (1, 260) ¼ 5.36, p < 0.0001). Overall, 30.5% of the Ethiopian fathers and 54% of the mothers were unemployed, versus 5.9% of the non- Ethiopian fathers and 28% of the non-Ethiopian mothers. Again, these differences were significant for both fathers (Chi-square (6,6) 29.0 p < 0.0001) and mothers (Chi-square (6,6) 36.05 p < 0.0001). These differences are in line with various reports based on national surveys.

88

M. Shany and E. Geva

5.3.2 The Curriculum The Ethiopian and Non-Ethiopian children learn the regular, centralized, curriculum, designed by the Ministry of Education (http://cms.education.gov.il). In elementary schools in Israel, minority children do not receive any formal instruction in their home language.1 In another study based in the same dataset, BenDavid (2005) reported that the Ethiopian children had very little proficiency in Amharic and that very few were able to engage in simple conversation in Amharic. Similar results were reported for Ethiopian children in kindergarten (Shany, Geva, & Melech, 2010). In informal communication (Dec 26, 2007), Ministry of Education supervisors indicated to us that various enrichment programs are available, but implementation is not compulsory or systematic, and that it depends on funding allocation by the municipality and local decisions at the school level. Enrichment programs were more likely to be implemented in the early years following immigration than it has been in more recent years. No formal followup on these programs, their implementation and outcomes is available.

5.3.3 Measures 5.3.3.1 Cognitive Non-verbal Ability The Raven’s Matrices test (Raven & Court, 1976) was administered. This test is designed to assess non-verbal intellectual and reasoning ability and the ability to make sense of increasingly complex visually presented stimuli, to draw meaning out of ambiguity and to perceive and think clearly. The test, considered by some to be relatively culture-free, consists of 5 subtests. Scores can range from 0 to 36, and the analyses were based on raw scores because norms for immigrant children are not available. Age in months was used to control for developmental effects where relevant. Rapid Automatized Naming (RAN) Object Naming Speed (RAN-O): This subtest of the reading and writing diagnostic test ‘‘Alef Ad Taf’’ (Shany, Lachman, Shalem, Bahat, & Zeiger, 2006) consists of pictures of 5 objects: flower, cat, book, watch and flag, each repeated randomly 10 times. The 50 printed objects are printed on a sheet (10 items per line) and the child is asked to name them out loud as fast as s/he can. Practice items ensure that the child is familiar with the names of the objects. Results are reported in terms of time (in seconds) to name the objects. Reliability 1

Beginning in Grade 7, some schools offer weekly, 2–3 h of instruction in the home language (http://cms.education.gov.il).

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . .

89

(correlation with Rapid Alternate Stimulus subtest, Shany et al., 2006) is r ¼ 0.61 for Grade 2, r ¼ 0.59 for Grade 4, and r ¼ 0.63 for Grade 6. Digit Naming Speed (RAN-D): This subtest (Shany et al., 2006) consists of 5 digits: 1, 5, 9, 3 and 7. Each digit is repeated randomly 10 times. The 50 digit, printed on a sheet (10 items per line) are presented to the child, who has to name them as fast as s/he can. Practice items ensure that the child is familiar with the names of the digits. Results are reported in terms of time (in seconds) to name the digits. Reliability (correlation with RAN-L subtest, Shany et al., 2006) is r ¼ 0.74 for Grade 2, r ¼ 0.69 for Grade 4, and r ¼ 0.67 for Grade 6. Letter Naming Speed (RAN-L): This subtest (Shany et al., 2006) consists of 5 Hebrew letters: ‫( ס‬/s/), ‫א‬, (/a/), ‫ ד‬/d/), ‫( ג‬/g/) and ‫ ל‬/l/), each repeated randomly 10 times. The 50 letters, printed on a sheet (10 items per line), are presented to the child, who has to name them aloud as fast as s/he can. Practice items ensure that the child is familiar with the names of the digits. Results are reported in terms of time (in seconds) to name all the letters. Reliability (correlation with RAN-D, Shany et al., 2006) is r ¼ 0.74, r–0.69, and r ¼ 0.67 for Grades 2, 4, and 6 respectively.

Working Memory This subtest (Shany et al., 2006) was adapted from Daneman and Carpenter (1980). In this test the participant hears sentences, is asked to complete the missing word in each, and then repeat the missing words to all the sentences in the same sequence. For example: The hens lay ___(eggs); An airplane flies in the ___(sky). In this example, the child has to provide the words ‘‘eggs’’ and ‘‘sky’’ and then repeat these two words. The number of items in sets increases gradually from 2 to 6. The child receives one point for each correctly repeated word and an extra point for repeating all the words in each set in the correct sequence. The Cronbach Alpha for this test is 0.81.

5.3.3.2 Cultural Literacy Faces and Places This subtest is taken from the standardized Hebrew version of the K-ABC (Fraser, Shimrovski, Wolf, & Hazani, 1994). The subtest includes 33 pictures involving ‘‘faces and places’’ that have to be named by the participants. The faces and places include famous objects (e.g., Noah’s ark), figures in famous stories (e. g., Pinocchio, Little Red Riding Hood), famous monuments (e.g., the Wailing Wall), and symbols (e.g., the Star of David). Scores can range from 0 to 33, and raw scores were used in the analyses. The split-half reliability of this normed test is r ¼ 0.70.

90

M. Shany and E. Geva

5.3.3.3 Language Skills Phonetic Discrimination This subtest was taken from the Diagnostic Battery for Reading Processes in Hebrew (Lachman & Shalem, 1998). Fifteen items were randomly taken from the original test which consists of 30 word pairs. Four items consist of identical pairs and 11 of different pairs. The items may differ on the first, second or the final consonant (e.g., ‘‘safar’’ – tzafar (counted – honked), ‘‘kibel’’ —‘‘kipel’’ (receivedfolded). The tester says each pair one at a time and the child has to say whether the words are ‘‘the same’’ or, ‘‘not the same’’. Scores range from 0 to 15.2 Phonemic Awareness Phonemic awareness was assessed with a subtest of the standardized Alef Ad Taf test (Shany et al., 2006). This subtest focuses on children’s ability to segment spoken real words into phonemes (e.g., ‘‘Say xatul (cat)’’. ‘‘Now say it without the /x/’’ (response: /atul/). This is a more demanding task as the correct response begins with the vowel /a/). None of the outcomes of this phoneme level word manipulation is a real word in Hebrew. National norms are available for this task (Shany et al., 2006). Scores can range from 0 to 16, and Alpha-Cronbach is 0.87 in Grades 2 and 4, and 0.86 in Grade 6. Receptive Vocabulary Children’s command of Hebrew receptive vocabulary was assessed with a standardized Hebrew adaptation of the Peabody Picture Vocabulary Test (PPVT, Dunn, 1965) (Solberg & Nevo, 1979). In this test the child hears a word, is shown four pictures, and is asked to point to the picture that matches the word heard. There are 110 items on the test and testing is stopped when the child makes 6 consecutive errors. National norms are available for this task (Solberg & Nevo, 1979). Split-half reliability for the Hebrew version is 0.90. Syntactic Skills The Test of Receptive Oral Grammar (TROG; Bishop, 1989), adapted to Hebrew from English is a receptive task of syntactic skills. Children listen to progressively longer and increasingly more complex sentences, presented one at a time, and for each item are asked to point to one of four pictures that matches the sentence. Foils include syntactic and semantic distracters. There are 4 items testing each of 20 syntactic categories such as negatives (e.g., ‘‘the boy is not running’’); passives (e.g., ‘‘the elephant is pushed by the child’’); x but not y (e.g., ‘‘The box but not the chair is red’’); and complex sentences (e.g., ‘‘The circle in which there is a star is 2

The authors of this test do not report on reliability.

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . .

91

red’’). Unless all the items within a category are correct, the whole category is discounted. Each correct category receives a score of 1, and test scores can range from 0 to 20. The Cronbach alpha for this test is 0.70. 5.3.3.4 Hebrew Reading and Spelling Word Reading Word reading was assessed with a subtest of the standardized Alef Ad Taf test (Shany et al., 2006). This subtest focuses on children’s ability to read 38 isolated words, representing various morphological structures, and arranged according to length and frequency. Measures of rate (WPM) and accuracy (% of errors) are calculated. Alpha-Cronbach is 0.90 in Grade 2, 0.85 in Grade 4, and 0.75 in Grade 6. Alpha-Cronbach and test norms are unavailable for Grade 1. Spelling This test was developed specifically for the present research and has no national norms. In this task the child has to write 48 dictated words. Each word is presented orally in isolation, in context (within a sentence) and again in isolation. The testers are instructed not to stress homophones (such as the phoneme / X/ in the word ‘‘xalav’’ (translation: milk). The overall Alpha-Cronbach is 0.87. Listening Comprehension This test, developed by Shatil (2002), requires participants to listen to a recorded story about a little boy and his dog. After having heard the story the child hears a series 12 statements related to the story, each presented twice, and has to respond ‘‘correct’’ or ‘‘incorrect’’ to each item. Test-retest reliability is 0.72. Reading Comprehension In this experimental task, six texts were developed altogether, 2 for each grade level (2, 4, and 6).3 For each grade level one text describes how to prepare a popular Ethiopian dish (Injara) and the other describes how to prepare a Middle Eastern dish (Cousous). While the content of the texts is identical across grade levels, the texts become more complex from grade to grade in terms of range of vocabulary and syntactic complexity. Ten questions that address four levels of comprehension follow each text. The items assess vocabulary, main idea in each paragraph, main idea and application of the knowledge and syntactic structures. Two experts (linguists) rated the texts and confirmed the increasing level of linguistic complexity of the stories. The Cronbach alpha was low in Grade 2 (0.42,) but acceptable 0.67, and 0.63, for Grades 4, and 6, respectively. 3

In Israel reading comprehension is not evaluated in Grade 1, and listening comprehension is evaluated instead.

92

M. Shany and E. Geva

5.3.4 Procedures Data collection took place towards the end of the school year, between April and June. Tasks were administered individually, by trained graduate students of education. In Grades 1 and 2 the tasks were administered over two sessions, lasting 75 min each. A 1-week interval separated the two sessions. In Grades 4 and 6 the tasks were administered in one, 2-h, session. The cognitive measures were administered first, followed by the language and literacy measures.

5.4 Results 5.4.1 Do Cognitive, Cultural, Language, and Literacy Skills Correlate with SES? Even though all the participating schools were ranked as low SES, as discussed above, there were significant differences between the Ethiopian and NonEthiopian samples on various demographic variables. Ethiopian children are rarely exposed to Hebrew at home. Instead, they are exposed to Amharic or a mixture of Hebrew and Amharic, and only about a third of their parents are reported to be literate in Hebrew. In addition, Ethiopian Israeli parents who are employed are more likely to hold less skilled jobs, and housing density is higher in Ethiopian families than in non-Ethiopian. We first examined whether differences between poverty and ‘‘extreme’’ poverty might be related in a systematic manner to various cognitive, language and literacy skills. Correlations indicated that housing density and parental occupation rank correlated significantly with group (r ¼ 0.49, p < 0.01) and with each other (r ¼ 0.31, p < 0.01), indicating that in the Ethiopian group more people live in each household than in the Non-Ethiopian group. These correlations suggest that, in contrast with the Non-Ethiopian group, Ethiopian parents were more likely to be unemployed or to be unskilled laborers than skilled workers. Finally, with two exceptions (Faces and Places, and Phonemic Awareness), neither household density nor parental occupation correlated with any of the child-level measures. The correlation of occupation rank with Phonemic Awareness and Faces and Places was low butt significant (r ¼ 0.13, p < 0.05, and r ¼ 0.12, p < 0.05, respectively). That is, in both groups a relatively higher occupation rank was associated with somewhat better performance on these two child-level measures. In other words, in this study, ethnicity, poverty indicators, and cultural differences appear to form one complex cluster, but the variance on the available poverty and low SES indicators did not correlate in a systematic and direct manner with the bulk of child level cognitive, language, and literacy measures. A multiple analysis of variance (MANOVA) was run on the cognitive and language measures. Another MANOVA was run with the literacy measures. In these analyses, the effects of grade and group and the interaction of grade by group were examined. Where a significant interaction was found, a Bonferroni

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . .

93

post-hoc test was applied to clarify the nature of the interaction. Tables 5.3, 5.4, and 5.5 present, by grade and group, summary statistics and the MANOVA results for cognitive skills, cultural literacy and language skills, and literacy measures in the Ethiopian and Non-Ethiopian groups. For comparison purposes, depending on availability, the equivalent to group means are presented also in terms of performance rank, percentile scores, percentile ranges, or standard score equivalents. The overall MANOVA results on the cognitive and language measures indicated significant effect of grade (F3,325 ¼ 16.02, Z2 ¼ 0.26, p < 0.001), group (F3,325 ¼ 5.55, Z2 ¼ 0.15, p < 0.001), and a significant interaction of grade and group (F3,325 ¼ 1.51, Z2 ¼ 0.05, p < 0.05).

5.4.1.1 Non-verbal Ability As can be seen in Table 5.2, there was a significant grade and group effect on the Raven’s matrices test. Within each group there was significant improvement across grades, and the Ethiopian group means were significantly lower than the means for the Non-Ethiopian groups at each grade level. The interaction of grade by group was not significant. A comparison of these statistics with the equivalent percentile scores published in the Raven’s Matrices manual indicates that, in comparison to the norming population, the mean percentile scores in both Ethiopian and Non-Ethiopian groups in Grades 1 and 2 are within the average range, but they decrease considerably in both groups in Grades 4 and 6.

5.4.1.2 Speed of Lexical Access – (RAN) As can be seen in Table 5.2 on RAN-Objects there was a significant grade and group effect. In both groups there was a significant improvement from grade to grade, and at each grade level the Ethiopian group was significantly slower in naming objects than the Non-Ethiopian group. As noted in Table 5.2, in comparison to the published norms (Shany et al., 2006), the mean equivalent in Grades 2 and 4 are in the Low to Low Average percentile ranges, but by Grade 6 the mean equivalent percentile score is in the Average range in the case of the Ethiopian group, and in the High Average range in the Non-Ethiopian group. In other words, the gaps decrease in relation to the norming population. A different pattern was noted with regard to RAN-Digits and RAN-Letters. There was a significant grade and a significant group effect. In both groups there was a significant improvement in naming speed from grade to grade. There was also a significant interaction of grade by group for each measure, indicating that the gap between the Ethiopian and Non-Ethiopian groups decreased gradually. In particular, on RAN-Digits the group effect was significant in Grades 1 and 2, but was no longer significant in Grades 4 and 6. The gap between the naming time on RAN-Digits in the Ethiopian and Non-Ethiopian groups decreased gradually from a significant gap of 9.87 seconds in Grade 1,

48.90 (10.66) NA

9.58 (6.81) (n ¼ 38) NA

Performance Rank RAN–L (sec.) 60.66 (20.51) Percentile Range NA

Performance Rank Working Memoryb 8.22 (5.52) (0–45)a (n ¼ 55) Percentiles Range NA

Average

Low av. 9.66 (4.78) (n ¼ 29) 35.1–65

Very Low 49.90 (16.14) 25.1–35

Low 40.84 (11.10) 7.1–16

20.62 (5.01) 44 61.19 (11.23) 16.1–25

Grade 2 Eth (n ¼ 42)

High

Low 14.71 (7.39) (n ¼ 21) 80.1–90

Low 42.50 (8.70) 16.1–25

Average 32.88 (6.25) 16.1–25

21.85 (4.11) 50 54.94 (11.21) 35.1–65

NEth (n ¼ 34)

High av.

Very Low 15.85 (6.30) (n ¼ 26) 65.1–80

Low 35.51 (8.48) 7.1–16

Low av. 28.79 (7.50) 16.1–25

24.46 (4.94) 22 49.46 (10.77) 25.1–35

Grade 4 Eth (n ¼ 39)

High av.

Low 17.06 (8.25) (n ¼ 35) 65.1–80

Low av. 34.00 (5.88) 16.1–25

Average 28.11 (5.09) 25.1–35

26.08 (5.01) 30 46.25 (7.18) 35.1–65

NEth (n ¼ 36)

High av.

Average 18.87 (9.04) (n ¼ 15) 65.1–80

Average 28.75 (5.56) 35.1–65

Average 24.44 (5.27) 35.1–65

26.03 (4.82) 13 42.22 (10.96) 35.1–65

Grade 6 Eth (n ¼ 36)

High

Average 20.67 (7.33) (n ¼ 9) 80.1–90

High av. 27.38 (5.76) 35.1–65

High av. 21.97 (5.62) 65.1–80

29.12 (4.72) 22 37.41 (7.47) 65.1–80

NEth (n ¼ 39)

24.08***

78.85*** Z2 ¼ 0.43

93.37*** Z2 ¼ 0.47

73.66*** Z2 ¼ 0.41

60.26*** Z2 ¼ 0.36

Grade effect F(3, 323)

Range of scores; A separate ANOVA due to missing data on this measure – F(3, 228); *** p < 0.001; ** p < 0.01; * p < 0.05

Performance Rank

a

b

40.45 (8.61) NA

Performance Rank RAN–D (sec.) 50.32 (15.05) Percentiles Range NA

Percentiles Range

Percentiles RAN–O (sec.)

19.67 (4.89) 67 58.74 (13.07) NA

NEth (n ¼ 42)

17.34 (4.64) 37 68.41 (13.99) NA

Raven’s (0–36)a

Tasks

Grade 1 Eth (n ¼ 58)

5.28*

16.00*** Z2 ¼ 0.05

25.31*** Z2 ¼ 0.07

22.62*** Z2 ¼ 0.07

14.85*** Z2 ¼ 0.05

Group effect F(1, 325)

0.96

3.58* Z2 ¼ 0.03

4.57** Z2 ¼ 0.04

1.31

0.550

Grade group F(3, 322)

Table 5.2 Performance of Ethiopian and non-Ethiopian children in grades 1, 2, 4, and 6 on cognitive components: Descriptive statistics and MANOVA results

94 M. Shany and E. Geva

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . .

95

to a non-significant difference of 0.98 and 2/47 seconds in Grades 4 and 6, respectively. As for a comparison to the norms published by Shany et al. (2006) one notes in Table 5.2 that whereas the mean equivalent scores were in the Very Low to Low percentile ranges in Grades 2 and 4, by Grade 6, the equivalent percentile scores were in the Average to High Average ranges. Likewise, on RAN-Letters the mean gap between Ethiopian and NonEthiopian groups decreased from a significant difference in favor of the NonEthiopian group in Grades 1 and 2 to a non-significant difference in Grades 4 and 6. Relatedly, with regard to the published norms (Shany et al., 2006), whereas the mean equivalent scores were in the Very Low to Low percentile ranges in Grades 2 and 4, by Grade 6, the equivalent scores were in the Average percentile ranges.

5.4.1.3 Working Memory On Working Memory there was a significant grade and group effect, and the interaction of group by grade was not significant. Performance improved significantly from grade to grade, and at each grade level the Ethiopian group was consistently lower than the Non-Ethiopian group. In comparison to the published norms (Shany et al., 2006), the performance of the Ethiopian and NonEthiopian groups is not lower than in the norming population, and the equivalents of the mean percentages vary from Average to High.

5.4.1.4 Cultural Literacy As can be seen in Table 5.3, there was a significant grade and group effect on this measure. Mean scores increased significantly from grade to grade, and the Ethiopian group performed consistently below the Non-Ethiopian group at each grade level. The means on the K-ABC for the Ethiopian group were 1 standard deviation below the normed mean at each grade level, whereas in the Non-Ethiopian group the standard score equivalents at each grade level were 0.5 standard deviations below the normed mean.

5.4.2 The Development of Language Skills 5.4.2.1 Phoneme Discrimination On Phone Discrimination there was a significant grade effect. Students in both groups improved significantly their performance from grade to grade. Neither the effect of group nor the interaction of grade by group were significant. In comparison with a large study (Lachman& Shalem, 1998) it appears that the two groups performed within developmental norms.

a

52.36 (7.42) 65

14.21 (3.75) 38

12.83 (3.69) 25

10.12 (3.95)

11.80 (2.22) (n ¼ 41)

8.28 (2.68) 92

49.77 (8.52) 63

10.24 (4.22)

11.88 (1.94) (n ¼ 57)

6.98 (2.98) 88

NEth (n ¼ 42)

13.79 (3.63) 25

Average 52.26 (11.45) 51

9.72 (3.71) 35.1–65

12.56 (2.07) (n ¼ 41)

8.44 (3.39) 85

Grade 2 Eth (n ¼ 42)

14.24 (1.97) 38

Average 52.44 (12.61) 51

8.60 (3.12) 35.1–65

12.73 (1.84) (n ¼ 33)

9.73 (3.26) 92

NEth (n ¼ 34) 13.72 (3.02) 93

NEth (n ¼ 36)

13.58 (1.38) (n ¼ 36)

15.14 (3.79) 86

Grade 6 Eth (n ¼ 36)

13.51 (1.25) (n ¼ 39)

18.05 (4.44) 93

NEth (n ¼ 39)

15.97 (1.94) 25

Low-av. 59.69 (11.52) 45 17.00 (1.59) 38

Low-av. 63.78 (10.26) 50 17.22 (1.91) 25

Average 66.61 (10.38) 45

17.28 (3.16) 25

Average 78.79 (11.78) 58

Scores are equivalent to percentile 30-40 7.02 6.25 4.62 4.30 (4.10) (4.26) (4.14) (3.68) 16.1–25 25.1–35 35.1–65 35.1–65

Language Skills 12.72 13.50 (2.11) (1.08) (n ¼ 39) (n ¼ 36)

11.44 (3.44) 84

Grade 4 Eth (n ¼ 39)

Range of scores; b A separate ANOVA due to missing data on this measure; *** p < 0.001; ** p < 0.01; * p < 0.05

Performance Rank Vocabulary: PPVT (0–110)a Standard Scores Equivalent Grammatical skills TROG (0–20)a Percentile

Percentile Phonemic Awareness No. Errors (0–16)a Percentiles Range

Phoneme Discriminationb No. correct (0–15)a

Cultural Literacy Faces and Places: KAB-C (0–33)a Standard Scores Equivalent

Tasks

Grade 1 Eth (n ¼ 58)

31.51*** Z2 ¼ 0.23

73.34*** Z2 ¼ 0.41

34.98*** Z2 ¼ 0.25

13.94*** Z2 ¼ 0.12

113.41*** Z2 ¼ 0.52

Grade effect F(3, 323)

4.89* Z2 ¼ 0.02

16.46*** Z2 ¼ 0.05

1.75

0.975

26.23*** Z2 ¼ 0.08

Group effect F(1, 325)

0.84

4.72** Z2 ¼ 0.05

0.26

0.970

1.09

Grade group F(3, 322)

Table 5.3 Performance of Ethiopian and non-Ethiopian children in grades 1, 2, 4, and 6 on cultural and language components: Descriptive statistics and MANOVA results

96 M. Shany and E. Geva

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . .

97

5.4.2.2 Phonemic Awareness There was a significant grade effect on this task. Performance improved significantly between Grade 2 and Grade 4, and between Grade 4 and Grade 6, in both groups. The group and interaction effects were not significant. The equivalents of the means in the Ethiopian group are in the Average range in Grade 2, in the Low-Average range in Grade 4, and again in the Average range in Grade 6. The equivalents of the means in the Non-Ethiopian group are identical to those noted in the Ethiopian samples (Shany et al., 2006) (Comparison statistics are unavailable for Grade 1). 5.4.2.3 Vocabulary As can be seen in Table 5.3, there was a significant grade effect on the vocabulary test. Post-hoc tests indicated that there was no significant difference between Grades 1 and 2, but the increase from Grade 2 to Grade 4 and from Grade 4 to Grade 6 was significant. There was also group effect, and a significant grade by group effect. The mean in the Ethiopian group was significantly lower than in the Non-Ethiopian group only in Grade 6. Notably, in relation to the norming population, the mean equivalent vocabulary standard scores decrease over time in both groups. 5.4.2.4 Grammatical Skills As can be seen in Table 5.3, there was a significant grade effect. Growth from Grade 2 to 4 was statistically significant, but the mean differences in the number of blocks mastered in Grades 1 vs. 2 and in Grades 4 vs. 6 were not statistically different from each other within each group. At each grade level the mean in the Ethiopian group was statistically lower than in the Non-Ethiopian group. The equivalent blocks in the norming population are 15, 16, 18, and 18 in Grades 1, 2, 3, and 6, respectively. That is, performance in both groups is below the comparable published norms for these age groups, and in fact, relative to the population, level of performance is constant in the Ethiopian group, and decreases in the Non-Ethiopian group.

5.4.3 The Development of Reading and Writing Skills 5.4.3.1 Word Reading Accuracy As can be seen in Table 5.4, there was a significant grade effect, but the group and interaction of grade by group effects were not significant. There was a significant increase in accuracy between Grades 1 and 2 and between Grades 2 and 4, but the increase between Grades 4 and 6 was not statistically significant. The equivalent word reading accuracy mean scores (50th percentile) reported by Shany et al. (2006) for the norming population are M ¼ 22.32 (SD ¼ 18.10),

a

20.25 (12.28) NA NA 33.99 (22.86) No norms No norms 27.92 (11.06) 10.62 (1.40)

F ¼ 1.50; (1, 93), p ¼ 0.223

16.93 (8.48) NA NA 43.26 (27.39) No norms No norms 20.52 (3.30) 10.18 (1.90)

NEth (n ¼ 42)

Range of scores; *** p < 0.001; * p < 0.05

Words –WPM (38 words) Percentiles Range Performance Rank Words – Accuracy % Errors Percentiles Range Performance Rank Spelling (0–48)# correcta Discourse/ Reading Comprehensiona

Tasks

Grade 1 Eth (n ¼ 58) 36.27 (20.03) 65.1–80 High-average 34.44 (24.09) 16.1–25 low 37.25 (8.59) 7.14 (3.62)

NEth (n ¼ 34) 48.88 (24.21) 35.1–65 Average 25.44 (21.09) 7.1–16 Very-Low 40.78 (8.52) 17.41 (6.37)

Grade 4 Eth (n ¼ 39) 43.33 (15.55) 25.1–35 Low-average 19.66 (13.17) 16.1–25 low 42.81 (5.08) 19.48 (6.16)

NEth (n ¼ 36)

F ¼ 0.565; (1, 63), p < 0.555 F ¼ 1.89; (1, 69), p < 0.174

26.07 (15.21) 25.1–35 Low-average 32.78 (23.31) 25.1–35 Low-average 32.06 (12.10) 6.93 (3.70)

Grade 2 Eth (n ¼ 42) 67.91 (22.00) 35.1–65 Average 14.04 (13.54) 16.1–25 low 44.69 (5.74) 19.53 (4.02) F ¼ 7.16; (1, 74), p < 0.009

60.71 (17.12) 35.1–65 Average 16.45 (12.46) 16.1–25 low 45.33 (5.27) 17.14 (3.63)

1.04

2.78

10.48*** Z2 ¼ 0.03

20.86*** Z2 ¼ 0.16

79.89*** Z2 ¼ 0.43

2.81* Z2 ¼ 0.03

3.01* Z2 ¼ 0.03

Grade group F(3, 322) 3.94* Z2 ¼ 0.01

Group effect F(1, 325)

111.15*** Z2 ¼ 0.51

Grade 6 Eth NEth Grade effect (n ¼ 36) (n ¼ 39) F(3, 323)

Table 5.4 Performance of Ethiopian children and non-Ethiopian children in grades 1, 2, 4, and 6 on word level reading components: Descriptive statistics and MANOVA results

98 M. Shany and E. Geva

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . .

99

M ¼ 14.14 (SD ¼ 12.94), and M ¼ 10.02 (SD ¼ 8.93) for Grades 2, 4 and 6, respectively. (Comparison statistics are not available for Grade 1). As can be seen in Table 5.3, both groups perform consistently in the low range at each grade level. 5.4.3.2 Word Reading Rate As can be seen in Table 5.4, there was a significant grade effect, a significant group effect, and a significant interaction of grade by group. Increase in word reading rate was significant from grade to grade in both groups. In addition, the group difference in word reading rate was significant only in Grade 2. (Comparison statistics are not available for Grade 1). In comparison to the norming population, performance in both the Ethiopian and Non-Ethiopian groups at each grade level was in the average range. 5.4.3.3 Spelling As noted in Table 5.4, there was a significant grade effect, a significant group effect, and a significant grade by group effect. The increase from Grade 1 to 2 and from Grade 2 to 4 was significant, but Grades 4 and 6 were not statistically different from each other. The Ethiopian group was lower than the NonEthiopian group in Grades 1 and 2, but not in Grades 4 and 6. 5.4.3.4 Discourse Comprehension Listening Comprehension Data for this task are available only for Grade 1. A one-way ANOVA indicated that the groups did not differ from each other. The group means are equivalent to the 75th percentile, as reported by Shatil (2002) in a large representative sample (n ¼ 349). Reading Comprehension There is no standardized reading comprehension test in Hebrew. As noted above, the content of the tests remained constant across grade levels, but the texts administered in Grades 2, 4, and 6 varied as a function of length, vocabulary, and language complexity. One-way ANOVAs, with group as the independent variables (summarized in Table 5.4), revealed that in Grades 2 and 4 the groups did not differ from each other, but by Grade 6, the Non-Ethiopian group outperformed the Ethiopian group. Table 5.5 summarizes the general trends that emerged in the results shown in Table 5.4. The columns on the left summarize the group comparisons and the columns on the right summarize comparisons with various norming populations. These trends are addressed in the Discussion section below.

Cognitive skills Raven’s RAN–Oa RAN–D RAN– L Working Memoryb Cultural literacy Faces and Places: KAB-C Language skills Phoneme Discrimination Phonemic Awareness Vocabulary: PPVT Grammatical skills

Domains

yes

yes

yes

yes

yes

yes

no

yes

no

yes

yes yes Yes yes no

no

yes

yes

ys yes Yes yes no

no

yes

yes yes

yes yes yes yes yes

yes yes

yes

yes

no

no

yes yes yes

yes yes yes

yes

yes

yes

yes

yes

yes

yes

yes

Table 5.5 Summary of results: Gap patterns from early grades (1 and 2) to higher grades (4 and 6) Comparison between the research groups Comparison of research groups with national norms Do the gaps Do the gaps Do the gaps decrease/ increase in persist in Are there disappear in higher higher higher gaps in early grades? grades? grades? grades? Do the gaps Do the Do the gaps Are there gaps persist increase in decrease/ gaps in higher in higher disappear in early grades? Eth. NEth. Eth. NEth. Eth. NEth. Eth. NEth. grades? higher grades? grades?

100 M. Shany and E. Geva

Domains

b

a

yes

yes

yes no

Do the gaps decrease/ disappear in higher grades?

In Gr. 2 no

Are there gaps in early grades?

In Gr. 6

no yes

No norms No norms

no yes

yes

yes

Even though there was a difference between the groups, both performed in the average range in comparison to the norming population The same comment

Cognitive skills Words –WPM Words – Accuracy Spelling Discourse/ Reading Comp.

Table 5.5 (continued) Comparison of research groups with national norms Do the gaps Do the gaps Do the gaps decrease/ increase in persist in Are there disappear in higher higher gaps in early higher grades? grades? grades? grades? Do the gaps Do the gaps persist increase in higher in higher grades? Eth. NEth. Eth. NEth. Eth. NEth. Eth. NEth. grades?

Comparison between the research groups

5 Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . . 101

102

M. Shany and E. Geva

5.5 Discussion To date, most of the official and non-official available reports on Ethiopian Israeli children that focused on academic achievement focused on older learners. These reports consistently highlighted gaps between the performance of these students and other immigrant groups and the general population. The present research examined in depth various component processes that underlie academic achievement of Ethiopian children and compared their performance to that of their Non-Ethiopian peers, who live in the same poor communities, attend the same classes, and are exposed to the same national curriculum. One general observation that emerges from the current study is that the Ethiopian children are not consistently low on all aspects of cognitive, language and literacy skills; on some components they are not different from their counterparts, and on some they begin with very low skills but they close the gaps over time. Importantly, in spite of linguistic, economic, and cultural differences, on certain components children of Ethiopian families are inherently similar to the general population. At the same time, on certain components of language and literacy the gaps increase over time. In what follows we discuss these various trends. The two levels of comparison shown in Table 5.5 are highly informative: the Ethiopian and non-Ethiopian groups, coming from a number of towns and schools in Israel, all lived in depressed neighborhoods, and in each case the Ethiopian and non-Ethiopian students received identical instruction. The comparison between Ethiopian and non-Ethiopian students therefore underscores the cultural differences that exert their effect over and above the influence of what happens, at least formally, at school. The comparison with the general norms, on the other hand, underscores the effects of living in poor neighborhoods in relation to general population trends.

5.5.1 No Gaps Between the Ethiopian and Non-Ethiopian Groups Beginning in Grade 1, the Ethiopian and Non-Ethiopian groups did not differ from each other on phoneme discrimination and on phonemic awareness. In addition, overall, both groups were not different from each other or from the general population on these skills. In this regard, it is interesting to note that in a study that compared Ethiopian and Non-Ethiopian children in kindergarten (based on the same sample), Shany et al. (2010) found differences between the groups on phoneme awareness and phoneme discrimination. In other words, with the onset of formal education and learning to read Hebrew, the differences between the Ethiopian and Non-Ethiopian children on these basic processing skills gradually disappeared. As noted in Table 5.5, even though their performance in Grade 4 is better than in Grade 2, the rate of improvement of phoneme awareness in the Ethiopian group is not as steep as in the general population.

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . . 103

On word reading accuracy there were no differences between the groups across grades. However, in comparison with the norming sample, both groups performed consistently in the low range. This finding is in line with numerous other studies involving L2 students that have shown that with the onset of formal education, there are no differences between L1 and ESL learners on phonemic awareness and word level reading skills. At the same time, these observations suggest that regardless of home language, the performance of children who come from low SES backgrounds may lag behind the general population (Oller & Eilers, 2002; Willms, 2003). In line with the general literature concerning the development of word-level skills in ESL learners, this lag may be attributed therefore less to home language than to insufficient exposure to extended reading and rich language and literacy contexts, as reflected in the low familiarity of children in both groups with popular children’s books (Faces and Places). On another basic underlying cognitive measure, working memory, while there was a difference between the two groups, in fact, both performed within the average range in relation to the norming population. It appears that on basic underlying cognitive measures that are less susceptible to cultural, linguistic, or educational influences, the Ethiopian group was not inherently different from the general population (nor was the comparison group of Non-Ethiopian children).

5.5.2 Closing Gaps Between the Ethiopian and Non-Ethiopian Groups On certain measures gaps existed between the Ethiopian and Non-Ethiopian students in the early school years (Grades 1 and 2), but disappeared by Grade 4. This trend was noted with regard to alphanumeric naming speed (letter and digits), word reading rate, and spelling. Moreover, when the gaps close between the Ethiopian and Non-Ethiopian groups on these basic processing components, both groups become similar to the norming population. The reduction over time of gaps between language minority children and their L1 counterparts on tests of speed of lexical retrieval of basic units such as digits and letters is well documented in the literature (Gholamain & Geva, 1999; Lesaux & Siegel, 2003; Geva & Yaghoub-Zadeh, 2006). The interesting question we need to explain is why in this study the gap in object naming persists, even though it disappears on the alphanumeric naming tasks. A possible explanation comes from the developmental literature on naming tasks. For example, cognitive researchers have shown that object naming is slower than letter and digit naming, and have explained that object naming is semantically more complex, and involves higher level retrieval processes than letter or digit naming, probably because it involves an open and less rehearsed class (Wolf & Goodglass, 1986; Southwood & Chatterjee, 1999, 2000). In this respect it is useful to consider the results in the Ehtiopian and Non-Ethiopian groups in

104

M. Shany and E. Geva

comparison to the norming data. Up to Grade 4, both the Ehtiopian and NonEthiopian groups are slower in alphanumeric naming tasks than the norming population, and one would argue that this might be a consequence of their low SES (Lucchese & Tamis-LeMonda, 2007). However, on object naming, up to Grade 4 only the Ethiopian children are slower in comparison to the norms. It is possible that this pattern reflects generally less developed semantic networks. According to the literature (Hedden, Lautenschlager, & Park, 2005; Salthouse, 1993), the object naming speed task is related to language proficiency, specifically, to vocabulary size, an-area on which Ethiopian children are lower in comparison to Non-Ethiopians. Even though in Grade 6 Ethiopian children performed on the object naming task within the average range, a significant gap in comparison to the Non-Ethiopian group was still apparent. In other words, as in the general literature, it is possible that the pattern noted on rapid object naming is related more generally to language development. However, as they move to higher grades, the Ethiopian Israeli children are not slower, relative to their counterparts, on rapid naming of simple alphanumeric items such as letters and digits that have been shown to be related to decoding skills (Sunseth & Bowers, 2002). These tasks are not related to vocabulary knowledge in L2 learners either (Geva, 2006). In line with this literature, the performance of the Ethiopian group on the alphanumeric tasks is not problematic in either of the groups targeted in this study. The results regarding spelling show that the Grade 1 gaps between the Ethiopian and Non-Ethiopian groups disappear by Grade 2. Similarly, other research has shown that the spelling gaps between monolingual and ESL learners coming from somewhat economically less challenged neighborhoods either close over time (Fashola, Drum, Mayer, & Kang, 1996; Wang & Geva, 2003) or are non-existent (Geva & Lafrance, 2011). These results are in line with the general argument that decoding and spelling skills are less susceptible to the influences of language proficiency, and that once the basic tools of spelling have been acquired, the L2 or minority learners do not necessarily lag in terms of their spelling skills (Geva & Lafrance, 2011; Wade-Woolley & Siegel, 1997).

5.5.3 Persistent Gaps Between the Ethiopian and Non-Ethiopian Groups Gaps persisted between the groups across grades on non-verbal ability, cultural knowledge, and grammatical skills. Even though the Ethiopian and NonEthiopian groups alike can be considered as low SES, the gap between them on non-verbal ability persists. Moreover, the gap between both groups and the national norms also increases systematically across grades. Even though the Raven’s test used in this study to evaluate non-verbal ability is considered to be a relatively culture-free task, we concur with Kozulin (1999) and Tzuriel and Kaufman (1994) who maintain that poor performance on such a task does reflect cultural differences, lack of exposure, and lack of opportunities to

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . . 105

develop the underlying problem solving skills needed to do well on this task. Indeed, there is scientific evidence (e.g., Kozulin, 1999; Tzuriel & Kaufman, 1994) that it is possible to improve the performance of Ethiopian children on the Raven’s Matrices with intensive and targeted intervention. We would argue that at least in regard to the Ethiopian sample, the Raven’s Matrices should be seen as reflecting to some extent cultural knowledge and experience, rather than intellectual, problem-solving ability. King et al. (2003) report that 60% of Ethiopian homes do not have any toys or books, and in another 20–25% of the families, up to 5 toys and 10 books were found on average. Ethiopian children have very few opportunities to play with puzzles, and there is often a dearth of toys in their homes. This supports the conjecture that at least to some extent their poor performance on the non-verbal task reflects lack of exposure and opportunities to develop relevant skills involving the manipulation of visual arrays that are not necessarily representational. In a similar vein, the gap between the Ethiopian and Non-Ethiopian groups persisted on Faces and Places, a task that also captures cultural knowledge and exposure. Another area in which the differences between the groups persist involved grammatical skills. Our results indicated that the relative position in relation to the population norms is constant in the Ethiopian group (25th percentile), and that it decreases in the Non-Ethiopian group (from the 35th to the 25th percentile). The low performance of both groups on this aspect of oral language proficiency is again likely related to their low socio-economic status (SES), and more specifically to insufficient exposure to rich language in oral and written contexts. Ample studies have shown that low SES is associated with fewer language-promoting experiences during the preschool years, and the latter, in turn, are related to lower receptive language abilities (Arriaga, Fenson, Cronan, & Pethick, 1998; Hart & Risley, 1995; Walker, Greenwood, Hart, & Carta, 1994). From a deficit perspective, some might attribute this persistent gap between the Ethiopian children and the Non-Ethiopians to low parental command of Hebrew. However, research on bilingual children has shown that children exposed simultaneously to two languages may, in fact, have cognitive advantages thanks to the fact that they are exposed to two linguistic systems (Bialystok, 2001). Children can actually develop high levels of competence when they are exposed to the L1 and L2 in formal educational contexts (Genoz & Genessee, 1998) that consider bilingualism as additive and not as subtractive (Cummins, 1989). The key here is the extent to which the L2 and L1 are valued, and fostered within rich language and literacy contexts in both languages (Schwartz, Leikin, & Share, 2005). In studies conducted within additive contexts, Bialystok and Feng (2011) did not find evidence for bilingual deficits on various cognitive tasks, even though consistently the bilinguals had lower vocabulary scores than their monolingual counterparts. Clearly, early language development is rooted in the interactions children have with their parents and other care takers, and these interactions both foster developing language skills and provide a vital foundation for children’s academic performance (Lucchese & Tamis-LeMonda, 2007). What may matter is not merely the

106

M. Shany and E. Geva

fact that the parents either do not speak Hebrew (the societal language) or do not speak it well, but that the parents have very low written language literacy skills in either language, and that the range and quality of language used at home, whether in Hebrew or Amharic, does not prepare the children for the language and concepts that are needed to excel in Israeli schools. It is important also to acknowledge that the school curriculum does not address adequately and early enough these differences and needs and that it treats children’s exposure to the home language within a subtractive framework which discourages the development of skills in both the societal and the home language.

5.5.4 Gaps Between the Ethiopian and Non-Ethiopian Groups That Increase An alarming finding is the observation that in the case of vocabulary and reading comprehension the gaps between Ethiopian and Non-Ethiopian Israeli children increased by Grade 6.4 In addition, and as reflected in the standard scores equivalents of the vocabulary test, across grades, the vocabulary size of both groups is below population norms. As many researchers have noted, semantic knowledge develops through exposure to both oral and written language, and ample research indicates the serious risk that factors such as poverty and low parental education pose to children’s early language development (Hoff, 2003; Walker et al., 1994). The findings indicate however, that from the onset children in the Ethiopian group achieve even worse than the comparison group. In other words, the results suggest that while some improvement takes place in the Non-Ethiopian group, in the Ethiopian group there may not be much improvement. The significant interaction between groups and grades on vocabulary deserves attention by policy makers and educators. Why does the gap between the Ethiopian Israeli children and their counterparts increase significantly by Grade 6? It might be suggested that this increase is associated with restricted and low quality of interactions in the Hebrew language which children from the Ethiopian community experience with their parents, their peers, and their teachers. Moreover, from the ‘‘cultural capital’’ perspective (Bourdieu & Passeron, 1990), it seems that disadvantages are maintained from one generation to the next. In that respect, the low linguistic performance of the Ethiopian children on both vocabulary (and grammatical skills), might reflect the failure of schools to adapt instruction to the needs of this population (Rueda et al., 2006). Citing the National Reading Panel (2000) Biemiller (2006) wrote that ‘‘Current reading instruction is apparently premised on the view that children 4

We checked to see whether the increase in the gaps in Grade 6 relates to sampling issues – however, in Grade 6 the gap between Ethiopian and Non-Ethiopian samples on the Raven’s does not increase, and on some indices, such as phonological discrimination, the gap disappears.

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . . 107

can build the vocabulary they need after learning to read (decode) fluently, as little or no vocabulary instruction occurs during the primary grades’’ (p. 44). Indeed, according to Israel Ministry of Education supervisors (personal communication, December 26, 2007), language enrichment programs are more likely to be implemented in the early years, following immigration and less likely to occur in higher grades. Vocabulary is a powerful predictor of reading comprehension (Biemiller, 2006), and a direct link seems to exists between an increase in the gap between the groups on vocabulary and on reading comprehension. Chall and colleagues (Chall, Jacobs, & Baldwin, 1990) have shown that in the lower grades the focus of instruction is on learning to read, and that regardless of SES (monolingual) children are less likely to be challenged by unfamiliar vocabulary and therefore the gap in language and reading skills in the lower grades is not as striking. However, many children coming from low SES families experience a ‘‘slump’’ in reading comprehension in grade 4, caused by below-grade vocabulary levels, and, while word recognition appears to be a major hurdle in the primary grades, word meaning becomes a major hurdle in Grade 4 and above (Chall, 1987). Studies focusing on minority students learning to read in the societal language demonstrate likewise that the reading comprehension of minority students falls well below that of their native-speaking peers (for a systematic review, see Lesaux & Geva, 2006). L2 students who have better developed L2 oral language comprehension skills are better able to comprehend texts (e.g., August, Carlo, Dressler, & Snow, 2005; Royer & Carlo, 1991). Of relevance here is the study of Proctor et al. (2005) who tested a structural equation model of second language (English as L2) reading comprehension and found that over and above adequate L2 decoding ability, L2 vocabulary knowledge is crucial for improved English reading comprehension outcomes for children of Spanish-speaking immigrants.

5.5.5 Language and Literacy Skills Among Children of Ethiopian Immigrants – What Develops? Overall, the results presented in this chapter indicate that on relatively simple language components, such as auditory discrimination and phonemic awareness, the Ethiopian and Non-Ethiopian groups perform in a similar way. On such simple language components, there is steady development from year to year and even though in the early years the group means are in the low-average by comparison to population mean, the gaps between the group means and the population mean appear to be closing sometime between Grades 4 and 6. In other words, the basic language components of children who come from low SES backgrounds emerge slowly, but with schooling and learning to read and write they master these basic components, and by the time they reach middle school these groups are not different from what one might expect in the population at large. While in and of itself this may be an encouraging trend, this slow emergence

108

M. Shany and E. Geva

may have negative effects on the development of more complex and more demanding aspects of language, literacy, and academic achievement. A different picture emerges, however, with regards to more complex aspects of language skills, and in particular vocabulary and grammar. Three observations are particularly noteworthy with regard to vocabulary skills. First, the gap between the Ethiopian and Non-Ethiopian groups increases in the upper elementary years, even though the two groups are not that different from each other in the lower grades. Second, the performance of both groups is considerably below the national norms at each grade level, with a gap of at least 3 standard deviations. Third, the vocabulary skills of the Ethiopian group slide further away in relation to national norms. It may be argued that comparing the performance of the Ethiopian group to the national norms is inappropriate due to their different home language background. This may be a legitimate argument but it is less relevant in this context, as this study shows that regardless of home language, the vocabulary skills of school children who live in very poor neighborhoods are extremely poor and that they do not develop at an acceptable rate. This observation has been made in other linguistic contexts (e.g., Hart & Risley, 1995). The fact that with time the Ethiopian Israeli children fall further behind is especially worrisome. Theoretically, without a special intervention program one would expect a steady rate of growth over time. That is, one would expect that the relative standing of the Ethiopian Israeli group would be maintained relative to the population or at least in relation to the Non-Ethiopian comparison group consisting of children who attend the same schools and live in the same poor neighbourhoods. The fact that the Ethiopian Israeli children fall even further behind suggests that their language skills do not develop at a steady rate. This may not be attributed merely to their second language status, as studies of vocabulary development in L2 learners (e.g., Farnia & Geva, 2011) have shown that even after 6 years of instruction in the societal language, the vocabulary skills of ESL children do not catch up with those of their first language (L1) counterparts, but that in the ESL group there is steeper growth in vocabulary learning in the early school years than in the non ESL group. However, in this project, instead of seeing an upward trajectory we see a downwards trajectory. Vocabulary knowledge is strongly related to reading comprehension in both monolingual and L2 learners, but is not as strongly implicated in word-level skills (Bialystok & Feng, 2011; Geva, 2006). This is also the case in this study of Ethiopian Israeli children. Even though these children have rather limited vocabulary knowledge, their phonological awareness, rapid naming, and word-level skills are acceptable. That is – they have the cognitive underpinnings needed to learn to read and achieve academically. As has been shown elsewhere and in this study, large language and vocabulary differences and different developmental trajectories distinguish low and middle class children even before kindergarten (Hart & Risley, 1995, 1999, 2003). These differences are probably compounded by current school practices that allow further widening of vocabulary gaps during the primary years (Biemiller & Slonim, 2001).

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . . 109

In sum, the children of Ethiopian immigrants are born into a culture of poverty, profound cultural differences, and lack of fit between home practice and beliefs on the one hand and the school system on the other. Our data suggest that less complex skills develop appropriately, as perhaps these skills are less susceptible to the cumulative effects of SES, cultural and school practices.. On these skills the Ethiopian Israeli children are not different from their nonEthiopian counterparts. However, on more complex skills, these children are not thriving because, in spite of their cognitive potential, they do not have sustained instructional and environmental exposure to cultural experiences that are needed in order to become more attuned to what it takes to become good readers, good writers, and sophisticated users of the language.

5.5.6 Summary and Implications Several trends and conclusions, relevant to the education of children of other groups migrating from non-literate to ‘‘developed’’ societies, should be stressed: Children who come from low SES backgrounds, from an oral tradition, and whose parents are illiterate in their L1 have more challenges in acquiring higher level language and literacy skills than low SES children whose parents are have gained some literacy skills. When families and/or a community have different cultural capital resources, they are limited in their ability to prepare and support their children in academic domains. In addition, when parents are not literate in their first language, their children cannot draw on their home language and L1 literacy in order to develop language skills. This is especially the case when the cultural institutions of the country of origin are not sustained in the new country. With development and schooling, gaps involving language and literacy components that are cognitively less complex and modularized (e.g., rapid retrieval of visual stimulus and phonological abilities) and word-based skills such as decoding, word reading efficiency, and spelling diminish or disappear. On the other hand the gaps persist or increase on more complex cognitive components that become increasingly more strategic and multi-componential, including non-verbal problem solving skills, language components such as vocabulary and grammatical knowledge, familiarity with the cultural symbols of the new country, and reading comprehension. As to the question of when can one expect low SES groups to perform more poorly than the population mean, four trends emerge. The first two trends characterize simple language and processing components: (a) when gaps do not exist between the two low SES groups beyond Grade 2, no gaps are noted between these groups and typical performance in the norming population either; (b) when gaps between the Ethiopian and Non-Ethiopian groups decrease, the gaps with national norms either decrease or stay stable. As for more complex aspects of language and literacy skills two trends appear: (c) when the gaps between the Ethiopian and Non-Ethiopian groups increase, both groups also decrease relative to the national norms; and d) when the gaps

110

M. Shany and E. Geva

between the Ethiopian and Non-Ethiopian groups are stable, both groups either decrease relative to norm or they maintain their relative position. The comparison of the mean performance in the Ethiopian and NonEthiopian groups with national norms underscores the joint effects of home and community factors associated with low SES and lack of cultural fit on various cognitive, linguistic and literacy skills.5 It is evident that low SES and cultural differences are associated with a wide array of cognitive and academic outcomes, with long term effects. These differences emerge early and are amplified in later years (Bradley & Corwyn, 2002). These trends underscore the need for the introduction of sustained national policies concerning the introduction of very early prevention programs targeting children, families and the training of educators. Countries where immigrants do well have well-established language support programs in early childhood education and primary school, with clearly defined goals, standards and evaluation systems (CERI, 2006). Countries where immigrant and minority children do not do well should develop and undertake policies such as the recent introduction of the ‘‘Early Years’’ policy in Ontario and translate these recommendations into specific policy priorities, professional training, curriculum development, and follow up research. The Ontario plan – With Our Best Future in Mind: Implementing Early Learning in Ontario (Ontario Government, June 2009) – is a totally integrated pre-natal to 12 years plan that when fully implemented, will provide a universal platform for healthy child and family development. Acknowledgements The research reported in this paper was partially supported by a grant from the Ministry of Education, Chief Scientist’s Office, Israel.

References Abu-Rabia, S., & Siegel, L. S. (2002). Reading, syntactic, orthographic, and working memory skills of bilingual Arabic-English speaking Canadian children. Journal of Psycholinguistic Research, 31, 661–678. Arab-Moghaddam, N., & Se´ne´chal, M. (2001). Orthographic and phonological processing skills in reading and spelling in Persian/English bilinguals. International Journal of Behavioral Development, 25, 140–147. Arriaga, R. I., Fenson, L., Cronan, T., & Pethick, S. J. (1998). Scores on the MacArthur communicative development inventory of children from low- and middle-income families. Applied Psycholinguistics, 19, 209–223. August, D., Carlo, M., Dressler, C., & Snow, C. (2005). The critical role of vocabulary development for English language learners. Learning Disabilities Research & Practice, 20, 50–57. August, D., & Shanahan, T. (2006). Developing literacy in second-language learners: A report of the National Literacy Panel on language-minority children and youth. Mahwah, NJ: Erlbaum. 5

In the present research there was no control group of middle class and we used national norms.

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . . 111

Bank of Israel. (2006). Annual Report (2007). Cited in the Ethiopian immigrants in the labor market: A report of the Israeli Association of Ethiopian Jews. Barkon, E., & Avinor, A. (1995). Academic difficulties and early literacy deprivation: The case of Ethiopians in Israel. Language, Culture and Curriculum, 8, 201–210. Bar-Yosef, R. (2001). Children of two cultures: Immigrant children from Ethiopia in Israel. Journal of Comparative Family Studies, 32, 231–246. Ben-David, I. (2005). The level of achievement among Ethiopian children from Grades 1 & 2 in Amharic and Hebrew: Oral language awareness and the contribution of this level to the acquisition of the Hebrew written language. An M.A Thesis submitted to the Faculty of Education, University of Haifa. Bender, M. L. (1985). Ethiopian language policy 1974–1981. Anthropological Linguistics, 27, 273–279. Ben-Ezer, G. (1992). Like a light in the vessel: The immigration and the absorption of Ethiopian Jews. Jerusalem: Rueben Mass Publishers. Berhanu, G. (2005). Normality, deviance, identity, cultural tracking and school achievement: The case of Ethiopian Jews in Israel. Scandinavian Journal of Educational Research, 49(1), 51–82. Retrieved from http://www.informaworld.com/smpp/titlecontent=t713445993 db=alltab=issueslistbranches49 -v49 Berry, J. W. (2003). Conceptual approaches to acculturation. In K. M. Chun, P. B. Organista, & G. Marin (Eds.), Acculturation: Advances in theory, measurement, and applied research (pp. 17–37). Washington, DC: American Psychological Association. Bialystok, E. (2001). Bilingualism in development: Language, literacy, and cognition. New York: Cambridge University Press. Bialystok, E., & Feng, X. (2011). Language proficiency and its implications for monolingual and bilingual children. In A. Durgunoglu & C. Goldenberg (Eds.), Language and literacy development in bilingual settings (pp. 121–138). New York: Guilford Press. Biemiller, A. (2006). Vocabulary: The missing link between reading and literacy. A lecture given at the World Congress on literacy. Budapest, Hungary, August 10. Biemiller, A., & Slonim, N. (2001). Estimating root word vocabulary growth in normative and advantaged populations: Evidence for a common sequence of vocabulary acquisition. Journal of Educational Psychology, 93, 498–520. Bishop, D. V. M. (1989). Receptive oral grammar test T.R.O.G. manual (2nd ed.). London: Medical Research Council. Bodovski, D., David, J., & Eran, Y. (1994). Issues in the treatment of families from different cultures. Jerusalem, Beithachin. Bourdieu, P., & Passeron, J. C. (1990). Reproduction in education, society and culture (R. Nice, Trans.). London: Sage. Bradley, R. H., & Corwyn, R. F. (2002). Socioeconomic status and child development. Annual Review of Psychology, 53, 371–399. Breznitz, Z. (2006). Fluency in reading: Synchronization of processes. Mahwah, NJ: Lawrence Erlbaum and Associates. Brookdale Institute. (2003). The immersion of the Ethiopian community in education and employment: New findings. brookdale.jdc.org.il/_. . ./22es-ethiopianemployment-eng.pd Carlisle, J. F., Beeman, M., Davis, Y. H., & Spharim, G. (1999). Relationship of metalinguistic capabilities and reading achievement for children who are becoming bilingual. Applied Psycholinguistics, 20, 459–478. Centre for Educational Research and Innovation (CERI). (2006). Where immigrant students succeed – A comparative review of performance and engagement in PISA – Programme for International Student Assessment (PISA 2003). OECD’s Media Division. Accessible through the website. Release date: 15 May 2006 – ISBN 91-64-023607 (downloaded, December 23, 2007).

112

M. Shany and E. Geva

Chall, J. S. (1987). Two vocabularies for reading: Recognition and meaning. In M. G. McKeown & M. E. Curtise (Eds.), The nature of vocabulary acquisition (pp. 7–17). Hillsdale, NJ: Lawrence Erlbaum Associates. Chall, J. S., Jacobs, V. A., & Baldwin, L. C. (1990). The reading crisis: Why poor children fall behind. Cambridge, MA: Harvard University Press. Chiappe, P., & Siegel, L. S. (1999). Phonological awareness and reading acquisition in English-and Punjabi-speaking Canadian children. Journal of Educational Psychology, 91, 20–28. 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. Comeau, L., Cormier, P., Grandmaison, E., & Lacroix, D. (1999). A longitudinal study of phonological processing skills in children learning to read in a second language. Journal of Educational Psychology, 91, 29–43. Cummins, J. (1989). Empowerment of minority children. Sacramento, CA: California Association for Bilingual Education. Cummins, J. (1991). Interdependence of first and second-language proficiency in bilingual children. In E. Bialystok (Ed.), Language processing in bilingual children (pp. 70–89). UK: Cambridge University Press. Cummins, J. (in press). The intersection of cognitive and sociocultural factors in the development of reading comprehension among immigrant students. In X. Chen & E. Geva (Ed.), Reading and writing – Special issue 2: Socio-cultural factors. Da Fontoura, H. A., & Siegel, L. S. (1995). Reading, syntactic, and working memory skills of bilingual Portuguese-English Canadian children. Reading and Writing, 7(1), 139–153. D’Angiulli, A., Siegel, L. S., & Maggi, S. (2004). Literacy instruction, SES, and word-reading achievement in English-language learners and children with English as a first language: A longitudinal study. Learning Disabilities Research & Practice, 19(4), 202–213. Daneman, M., & Carpenter, P. A. (1980). Individual differences in working memory and reading. Journal of Verbal Learning and Verbal Behavior, 19, 450–466. 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. 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. Dressler, C., & Kamil, M. (2006). First-and second-language literacy. In D. August & T. Shanahan (Eds.), Developing literacy in second-language learners: A report of the National Literacy Panel on language-minority children and youth (pp. 197–238). Mahwah, NJ: Erlbaum. Droop, M., & Verhoeven, L. (2003). Language proficiency and reading ability in first- and second-language learners. Reading Research Quarterly, 38, 78–103. Dunn, L. M. (1965). Peabody picture vocabulary test. Minnesota: American Guidance. Durgunoglu, A. Y., Nagy, W. E., & Hancin-Bhatt, B. J. (1993). Cross language transfer of phonological awareness. Journal of Educational Psychology, 85, 453–465. Farnia, F., & Geva, E. (2011). Cognitive correlates of vocabulary growth in English language learners. Applied Psycholinguistics, 32(4), 711–738. Fashola, O. S., Drum, P. A., Mayer, R. E., & Kang, J. S. (1996). A cognitive theory of orthographic transitioning: Predictable errors in how Spanish-speaking children spell English words. American Educational Research Journal, 33, 825–843. Fraser, M., Shimrovski, G., Wolf, N., & Hazani, A. (1994). The Kaufman assessment battery for children – K-ABC (Hebrew version). Jerusalem: Ministry of Education, Psychological and Counseling Service (SHEFI) and Henrietta Szold Institute.

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . . 113

Gaon, D., King, Y., & Volda-Tzadik, A. (2006). Populations which do not participate in the labor market: Characteristics, and programs to enhance them. A report submitted to “TABAT” (pp. 1–100). Jerusalem: Myers-JDC-Brookdale Institute. Genoz, J., & Genessee, F. (1998). Psycholinguistic perspectives on multilingualism and multilingual education. In J. Genoz & F. Genesee (Eds.), Beyond bilingualism: Multilingualism and multilingual education (pp. 16–32). Clevedon: Multilingual Matters. Gersten, R. (1999). Lost opportunities: Challenges confronting four teachers of Englishlanguage learners. The Elementary School Journal, 100, 37–56. Geva, E. (2006). Second-language oral proficiency and second-language literacy. In D. August & T. Shanahan (Eds.), Developing literacy in second-language learners: A Report of the National Literacy Panel on language-minority children and youth (pp. 123–139). Mahwah, NJ: Erlbaum. Geva, E. (2007). Facets of metalinguistic awareness related to reading acquisition in Hebrew: Evidence from monolingual and bilingual children. In K. Koda & A. Zehler (Eds.), Learning to read across languages (pp. 154–187). New York: Routledge. Geva, E., & Lafrance, L. (2011). Linguistic and cognitive processes in the development of spelling in ELLs: L1 transfer, language proficiency, or cognitive processes? In A. Durgunoglu & C. Goldenberg (Eds.), Language and literacy development in bilingual settings (pp. 245–279). Guilford. Geva, E., & Siegel, L. S. (2000). Orthographic and cognitive factors in the concurrent development of basic reading skills in two languages. Reading and Writing, 12(1–2), 1–30. Geva, E., & Wade-Woolley, L. (1998). Component processes in becoming English-Hebrew bi-literate. In A. Y. Durgunog˘lu & L. Verhoeven (Eds.), Literacy development in a multilingual context: Cross-cultural (pp. 85–110). Mahwah, NJ: Lawrence Erlbaum Associates. Geva, E., Wade-Woolley, L., & Shany, M. (1993). The concurrent development of spelling and decoding in two different orthographies. Journal of Reading Behavior, 25(4), 383–405. Geva, E., Wade-Woolley, L., & Shany, M. (1997). Development of reading efficiency in first and second language. Scientific Studies of Reading, 1(2), 119–144. Geva, E., & Yaghoub-Zadeh, Z. (2006). Reading efficiency in native English-speaking and English-as-a-second-language children: The role of oral proficiency and underlying cognitive-linguistic processes. Scientific Studies of Reading, 10(1), 31–57. Geva, E., Yaghoub-Zadeh, Z., & Schuster, B. (2000). Part IV: Reading and foreign language learning: Individual differences in word recognition skills of ESL children. Annals of Dyslexia, 50, 121–154. Gholamain, M., & Geva, E. (1999). Orthographic and cognitive factors in the concurrent development of basic reading skills in English and Persian. Language Learning, 49(2), 83–217. Goldenberg, C., Rueda, R., & August, D. (2006). Socio-cultural influences on literacy attainment of language-minority children and youth. In D. August & T. Shanahan, (Eds.), Developing literacy in second-language learners: A report of the National Literacy Panel on language-minority children and youth (pp. 269–318). Mahway, NJ: Erlbaum. Gould, E.D., Lavy, V., & Paserman, M. D. (2004). Immigrating to opportunity: Estimating the effects of school quality using a natural experiment on Ethiopians in Israel. Quarterly Journal of Economics, 119, 489–526. Hart, B., & Risley, T. R. (1995). Meaningful differences in the everyday experience of young American children. ERIC #:ED387210. Baltimore, MD: Brookes Publishing Company. Hart, B., & Risley, T. (1999). The social world of children learning to talk. Baltimore, MD: Paul Brookes Publishing Co. Hart, B., & Risley, T. R. (2003). The early catastrophe: The 30 million word gap by age 3 – all 8 versions. Retrieved September 23, 2005, from http://www.aft.org/american_educator/ index.html Hedden, T., Lautenschlager, G., & Park, D. (2005). Contributions of processing ability and knowledge to verbal memory tasks across the adult life-span. The Quarterly Journal of Experimental Psychology, 58(1), 169–190.

114

M. Shany and E. Geva

Hoff, E. (2003). The specificity of environmental influence: Socioeconomic status affects early vocabulary development via maternal speech. Child Development, 74(5), 1368–1378. Hooddinoot, J., Lethbridge, L., & Phipps, S. (2002). Is history destiny? Resources, transitions and child education attainment in Canada. Ottawa: Human Resources Development Canada, Applied Research Branch. Retrieved from www.drhc-hrdc.gc.ca/spps/arb.dgra/ publicatins/research/2002.docs Horowitz, T. R., & Mosher, N. (1997). Achievement motivation and level of aspiration: Adolescent Ethiopian immigrants in the Israeli education system. Adolescence, 32, 169–180. Kaplan, R. (1992). The psychological benefits of nearby nature. In D. Relf (Ed.), The role of horticulture in human well-being and social development: A national symposium (pp. 125–133). Portland, OR: Timber Press. Katz, L., & Frost, R. (1992). The reading process is different for different orthographies: The orthographic depth hypothesis. Advances in Psychology, 94, 67–84. Katzir, T., Shaul, S., Breznitz, Z., & Wolf, M. (2004). The Universal and the unique in dyslexia: A cross-linguistic investigation of reading and reading fluency in Hebrew-and English-speaking children with reading disorders. Reading and Writing, 17, 739–768. King, J., Effrati, R., & Netzer, N. (2003). Select findings of surveys of Ethiopian immigrants in eight urban communities in Israel. Meyers – JDC – Brookdale Institute and the Absorption Ministry, State of Israel. Retrieved from www.jdc.org.il/brookdale Korten, F. F. (1972). The influence of culture and sex on the perception of persons. International Journal of Psychology, 9, 31–44. Retrieved from http://www.informaworld.com/ smpp/titlecontent=t713659663db=alltab=issueslistbranches=9 - v9 Kozulin, A. (1999). Cognitive learning in younger and older immigrant students. School Psychology International, 20, 177–190. Lachman, D., & Shalem, T. (1998). A battery for the diagnoses of reading disability. Tel-Aviv: Nitzan Organization. Lafrance, A., & Gottardo, A. (2005). A longitudinal study of phonological processing skills and reading in bilingual children. Applied Psycholinguistics, 26, 559–578. Lee, J. W., & Schallert, L. S. (1997). The relative contribution of L2 language proficiency and L1 reading ability to L2 reading performance: A test of the threshold hypothesis in an EFL context. TESOL Quarterly, 31, 713–739. Lesaux, N. K., & Geva, E. (2006). Synthesis: Development of literacy in language minority learners. In D. L. August & T. Shanahan (Eds.), Developing literacy in a second language: Report of the national literacy panel (pp. 53–74). Mahwah, NJ: Lawrence Erlbaum Associates. Lesaux, N. K., with Koda, K., Siegel, L. S., & Shanahan, T. (2006). Development of literacy of language minority learners. In D. L. August & T. Shanahan (Eds.), Developing literacy in a second language: Report of the national literacy panel (pp.75–122). Mahwah, NJ: Lawrence Erlbaum Associates. Lesaux, N. K., & Siegel, L. S. (2003). The development of reading in children who speak English as a second language. Developmental Psychology, 39, 1005–1019. Leseman, P. P. M., & de Jong, P. F. (1998). Home literacy: Opportunity, instruction, cooperation and social-emotional quality predicting early reading achievement. Reading Research Quarterly, 33, 294–318. Levin, T., Shohamy A., Spolsky, D., Levi-Keren, M., Inbar, I., & Shemesh, M. (2002). Educational status of immigrant students. Report to the Chief Scientist of the Ministry of Education. School of Education, University of Tel-Aviv. Levin-Rozalis, M. (2000). Social representations as emerging from social structure: The case of the Ethiopian immigrants to Israel. Papers on Social Representations, 9, 1–22. Levin-Rozalis, M., & Shafran, N. (2003). A sense of belonging: A tracer study of ALMAYA’s parents cooperative kindergarten, Israel. Early childhood development: Practice and reflections. Bernard van Leer Foundation. Retrieved December 20, 2007, from http:// www.bernardvanleer.org

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . . 115

Lindsey, K. A., Manis, F., & Bailey, C. (2003). Development of reading in grades K–2 in Spanish-speaking English-language learners. Learning Disabilities Research & Practice, 19, 214–224. Lucchese, F., & Tamis-LeMonda, C. S. (2007). Fostering language development in children from disadvantaged backgrounds. In Encyclopedia of language and literacy development (pp. 1–11). London, ON: Canadian Language and Literacy Research Network. Retrieved December 2007, from http://www.literacyencyclopedia.ca/pdfs/topic.php?topId¼229 Meitzav. (2006). Retrieved December 22, 2007, from http://cms.education.gov.il/Educa tionCMS/Units/Rama/MivchaneiHameitzavKlali/DochotMeitzav1. Mu¨ller, K., & Brady, S. (2001). Correlates of early reading performance in a transparent orthography. Reading and Writing, 14(7–8), 757–799. Mumtaz, S., & Humphreys, G. W. (2001). The effects of bilingualism on learning to read English: Evidence from the contrast between Urdu-English bilingual and English monolingual children. Journal of Research in Reading, 24, 113–134. 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. 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 Institutes of Child Health. & Human Development. Retrieved form http://www.nichd.nih.gov/publications/nrp/small book.pdf Offer, S. (2004). The socio-economic integration of the Ethiopian community in Israel. International Migration, 42, 29–55. Oller, D. K., & Eilers, R. E. (2002). Language and literacy in bilingual children. Clevedon, UK: Multilingual Matters. Ontario Government. (2009). Retrieved from http://www.ontario.ca/en/initiatives/early_ learning/ONT06_018865 Padro´n, Y. M., & Waxman, C. (1988). The effect of ESL students’ perceptions of their cognitive strategies on reading achievement. TESOL Quarterly, 22(1), 146–150. Pressley, M. (1998). Reading instruction that works: The case for balanced teaching. New York: Guilford Press. Proctor, C. P., Carlo, M., August, D., & Snow, C. (2005). Native Spanish-speaking children reading in English: Toward a model of comprehension. Journal of Educational Psychology, 97(2), 246–256. Raven, J., & Court, J. (1976). Raven’s progressive matrices test. Oxford: I.C. Raven Ltd., Oxford Psychologists Press. Reese, L., Garnier, H., Gallimore, R., & Goldenberg, C. (2000). Longitudinal analysis of the antecedents of emergent Spanish literacy and middle-school English reading achievement of Spanish-speaking students. American Educational Research Journal, 37, 633–662. Royer, J. M., & Carlo, M. S. (1991). Transfer of comprehension skills from native to second language. Journal of Reading, 34(6), 450–455. Rueda, R. S., August, D., & Goldenberg, C. (2006). The social context in which children acquire literacy. In D. August & T. Shanahan (Eds.), Developing literacy in secondlanguage learners: A report of the National Literacy Panel on language-minority children and youth (pp. 319–330). Mahwah, NJ: Erlbaum. Salthouse, T. A. (1993). Speed mediation of adult age differences in cognition. Developmental Psychology, 29(4), 722–738. Schwartz, M., Leikin, M., & Share, D. L. (2005). Bi-literate bilingualism versus mono-literate bilingualism: A longitudinal study of reading acquisition in Hebrew (L2) among Russianspeaking (L1) children. Written Language and Literacy, 8, 179–206. Segal, B. (2006). Literacy indices in Ethiopian families in Israel and their contribution to the literacy development of their children. Unpublished M.A thesis, submitted to the Faculty of Education, University of Haifa.

116

M. Shany and E. Geva

Seymour, P. H. K., Aro, M., & Erskine, J. M. (2003). Foundation literacy acquisition in European orthographies. British Journal of Psychology, 94(2), 143–174. Shany, M., Geva, E., & Melech, L. (2010). Emergent literacy in children of immigrants coming from a primarily oral literacy culture. Journal of Written Language and Literacy, 13, 24–60. Shany, M., Lachman, D., Shalem, Z., Bahat, A., & Zeiger, T. (2006). ‘‘Aleph – Taph’’ – An assessment system for reading and writing disabilities. Tel Aviv: Yesod Publishing. Share, D. (2008). On the Anglocentricities of current reading research and practice: The perils of overreliance on an ‘‘outlier’’ orthography. Psychological Bulletin, 134, 584–615. Shatil, E. (2002). Shatil test to the assessment of at risk preschool children. Haifa: Ach Publishing. Shenkut, M. K. (1991). Mobilizing for literacy: The Ethiopian experience. L’Alphabetisation et la Formation des Adults, 5(3), 12–16. Solberg, A., & Nevo, B. (1979). The Peabody picture vocabulary test (PPVT). (Hebrew version). Southwood, M. H., & Chatterjee, A. (1999). Simultaneous activation of reading mechanisms: Evidence from a case of deep dyslexia. Brain and Language, 67, 1–29. Southwood, M. H & Chatterjee, A. (2000). The interaction of multiple routes in oral reading: Evidence from dissociations in naming and oral reading in phonological dyslexia. Brain and language, 72, 14–39. Spolsky, B. (1996). Multilingualism in Israel. Annual Review of Applied Linguistics, 17, William Grabe (Ed.). Stravas, A., & Olshtain, E. (2006). Literacy patterns in bilingual Ethiopian immigrant parents in Israel: A developmental perspective. Poster presentation at the BAL conference, May 2006, Toronto. Sunseth, K., & Bowers, P. G. (2002). Rapid naming and phonemic awareness: Contributions to reading, spelling, and orthographic knowledge. Scientific Studies of Reading, 6, 401–429. Svirsky, S., & Svirsky, B. (2002). The Ethiopian Jews in Israel: Housing, employment and education. In ‘‘Meida al Shivion’’ 11, Edva Center, Israel. The Israeli Municipalities. (2006, 2007). Physical data. Jerusalem, Israel. Tzuriel, D., & Kaufman, R. (1994). Mediated learning and cognitive modifiability: Dynamic assessment of young Ethiopian immigrant children to Israel. Journal of Cross-Cultural Psychology, 30, 359–380. Verhoeven, L. T. (1990). Acquisition of reading in a second language. Reading Research Quarterly, 25(2), 90–114. Verhoeven, L. T. (1994). Transfer in bilingual development: The linguistic interdependence hypothesis revisited. Language Learning, 44, 381–415. Vertsberger, R., & Noyfeld, D. (2003). The Ethiopian community in Israel. A report to the Knesset, Israel. Wade-Woolley, L., & Geva, E. (2000). Processing novel phonemic contrasts in the acquisition of L2 word reading. Scientific Studies of Reading, 4, 295–311. Wade-Woolley, L., & Siegel, L. S. (1997). The spelling performance of ESL and native speakers of English as a function of reading skill. Reading and Writing: An Interdisciplinary Journal, 9, 387–406. Walker, D., Greenwood, C., Hart, B., & Carta, J. (1994). Prediction of school outcomes based on early language production and socioeconomic factors. Child Development, 65(2), 606–621. Wang, M., & Geva, E. (2003). Spelling acquisition of novel English phonemes in Chinese children. Reading and Writing: An Interdisciplinary Journal, 16, 325–348. Willms, J. D. (2003). Ten hypotheses about socioeconomic gradients and community differences in children’s developmental outcomes. Final Report, Applied Research Branch, Strategic Policy, Human Resources Development, Canada. Wimmer, H. (1993). Characteristics of developmental dyslexia in a regular writing system. Applied Psycholinguistics, 14, 1–13.

5

Cognitive, Language, and Literacy Development in Socio-culturally Vulnerable . . . 117

Wolf, M., & Goodglass, H. (1986). Dyslexia, dysnomia, and lexical retrieval: A longitudinal investigation. Brain and language, 28, 154–168. Yaghoub Zadeh, Z., Farnia, F., & Geva, E. (2011). Toward modeling reading development in young English as second language learners. Reading and Writing: An Interdisciplinary Journal. doi:10.1007/s11145-010-9252-0. Yediot Axronot, January 1, 2008. 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.

Chapter 6

Second Generation Immigrants: A Socio-Linguistic Approach of Linguistic Development Within the Framework of Family Language Policy Mila Schwartz

6.1 Introduction Immigrant families overcome the process of ‘‘insertion,’’ in which they must find their way into host country society, sometimes being pressured to abandon their home country language or detach themselves from their country of origin (BenRafael et al., 2006). For example, Pease-Alvarez (2003) conducted in-depth interviews with 63 parents who were first and second generation immigrants from Mexico in California. The findings point out a strong tendency of some participants to abandon the use of Spanish with their children and adopt English monolingual norms and Anglo values in an effort to improve their social status and new cultural identity, and to enjoy the benefits associated with becoming Americans. On the other hand, the insertion in a new society might be characterized by a strong tendency to integrate by combining both the maintenance of the home country language and the acquisition of the host country language, as in the case of the last wave of Russian-Jewish immigrants from the former Soviet Union (FSU) in Israel (Horenczyk & Ben-Shalom, 2006; Lissitsa, 2007). Accordingly, the focus on the case of Russian-Jewish immigrants from the FSU and their Israeli-born children might provide a unique and intriguing case study of how family language policy (FLP) and family background can enhance first language (L1; the heritage language) inter-generational transmission and second language (L2) acquisition. This paper presents an analysis of studies focusing on the process of vocabulary acquisition among second generation immigrants from the FSU in Israel. As presented in the Introduction to this volume, this immigrant community is characterized by high education levels. Approximately 60% of the FSU immigrants in the workforce have academic professions, compared with 30% M. Schwartz (*) Research and Evaluation Authority, Oranim Academic College of Education, Kiryat Tivon, Israel Edmond J. Safra Brain Research Center for the Study of Learning Disabilities, University of Haifa, Haifa, Israel e-mail: [email protected] M. Leikin et al. (eds.), Current Issues in Bilingualism, Literacy Studies 5, DOI 10.1007/978-94-007-2327-6_6, Ó Springer ScienceþBusiness Media B.V. 2012

119

120

M. Schwartz

of the veteran Israeli population (Leshem & Lissak, 1999). At the same time, many individuals from this wave of immigration were obliged to seek alternative blue collar employment. Thus, one of the features of this wave of immigration is the discrepancy between the relatively low-middle socioeconomic status of the community and its high education level. In general, the first generation of immigrants from the FSU in Israel is distinguished by a relatively rapid and successful acquisition of Hebrew, the majority language, compared to other waves of immigration to the country (Olshtain & Kotik, 2000; Stavans & Goldzweig, 2008). Stavans and Goldzweig (2008) found that the Russian-Jewish immigrants’ increased proficiency in L2 does not pose a threat or a need to alter their perception of their L1 community or language as the cultural capital that is complete and satisfying. Note also that the promotion of supplementary after-school educational settings designed to preserve the cultural heritage and L1 in its literate form among second generation is one of the main target of this immigrant community. In addition, it was found that for these immigrants as parents, their child’s academic success is a very important personal objective in the country of origin, as well as in Israel as the host country (Horowitz, 1986). Focusing on this community, this chapter examines what socio-cultural and socio-linguistic factors are related to vocabulary in L1 and L2 among second generation immigrants within the context of the family language policy and its background (i.e., parental education, socio-economic status, and length of parents’ residence in the host country and social milieu).

6.2 Family Language Policy and Vocabulary in L1 and L2 Research on family language policy (FLP) incorporates analysis of language ideology, practice, and management, which were classified by Spolsky (2004) as components of the language policy model with respect to the speech community. In distinguishing these three components, Spolsky (2004:5) notes ‘‘. . .language practices – the habitual pattern of selecting among the varieties that make up its linguistic repertoire; its language beliefs or ideology – the beliefs about language and language use; and any specific efforts to modify or influence that practice by any kind of language intervention, planning or management.’’ Using this model at the family level enables us to integrate the separate components within a structural, flexible, and expandable framework. Developmental psychologists suggest that parents’ beliefs motivate their practices, which in turn are strong determinants of the children’s development (De Houwer, 1999; Johnson & Martin, 1985). For example, Barkhuizen (2006) found that Afrikaans-speaking South African immigrant parents in New Zealand believed that their children’s transition to English would be smooth if they were socialized in English during the pre-immigration stage. At the same time, there is a growing evidence for full complexity and non-linearity of relationships between parental language ideology and actual language and

6

Second Generation Immigrants: A Socio-Linguistic Approach. . .

121

practice at home. More specifically, the declared language ideology of one or both parents does not necessarily coincide with the strategies followed consciously or unconsciously in language practice with children (De Houwer, 1999; Goodz, 1994; King, 2000; Kopeliovich, 2006; Schwartz, 2008; Spolsky, 2004). For example, Eilers, Pearson, and Cobo-Lewis’s (2006) study focused on the link between the Spanish-English speaking parents’ positive attitudes to their children balanced bilingual development (i.e., language ideology) and real language practice at home. The findings showed a clear-cut inconsistency between the parents’ ‘‘firm plans’’ to provide balanced bilingual development in Spanish and English for their newborns through the first 3 years of life and these plans actual realization. Thus, only one family out of 24 was persistent in providing equal exposure to both languages during the 3 years of the study. The parents’ beliefs might be realized by means of planned or non-planned regulation (i.e., management) of L1 vocabulary development. For example, parents might control L1/L2 vocabulary development by using enjoyable tactics such as joint reading and learning illustrated rhymes and poems by heart (Gregory, Arju, Jessel, Kenner, & Ruby, 2007; Kopeliovich, 2006). This family language strategy was found to be rather fruitful for L1 pronunciation improvement and understanding of culturally unique idioms in L1 as well. Through poetry recitation in L1 the family also enhances cultural identity of the second generation children. Note also that the acquisition, development, and maintenance of L1 across generations are encouraged and stimulated by family efforts to transmit its written form (Cummins, 2000; Pearson, 2007; Seville-Troike, 2000). It appears that within the framework of FLP, child language policy has received rather minor attention (Okita, 2002). This study tried to fill this void by exploring whether parental language ideology was indeed linked to the family’s language management strategies and contributed to the children’s L1 vocabulary by collecting both the parents’ reports and children’s reports on their attitudes toward the L1 maintenance. Concerning FLP and L2 vocabulary skills among the second generation, the existing data relate mostly to patterns of language input at home. The child’s language input refers to language interaction in the family milieu (parents, grandparents, siblings) (De Houwer, 1995). Note that findings regarding the link between L2 inputs at home and children’s L2 vocabulary are not conclusive. For example, addressing the role of dual input (L1þL2) at home, CoboLewis, Pearson, Eilers, and Umbel (2002) used English (L2) literacy and oral proficiency measures to compare bilingual children equally exposed to English (L2) and Spanish (L1) at home (ESH) with their bilingual peers exposed only to L1 Spanish at home (OSH) and with monolingual English-speaking children. Group selection was based on parental reports of family language practice in the family. The researchers reported a strong effect of ESH, especially on oral English (L2) proficiency. However, the authors concluded that L2 exposure was important mainly in the first years of life. The present project continues to study this issue.

122

M. Schwartz

In addition to family language input, child’s social milieu (e.g., friends, classmates) might be another important factor linked to L2 vocabulary. Cobo-Lewis et al. (2002) concluded that since over 90% of children in the schools participating in the study were bilingual, there was an apparent effect of limited L2 native proficiency. This suggests the need to examine further the role that friends and classmates play in the attainment of L1/L2 vocabulary among second generation immigrants over and above family language practice. Note that when children are young, their parents might influence their childrens’ social milieu as a part of their language management. The parents might search for a supporting socio-linguistic environment, which provides L1 input and vice versa. For example, in his study of FLP among Afrikaans-speaking South African immigrants living in New Zealand, Barkhuizen (2006) found a tendency among parents to plan for their children’s maximum exposure to L1, including the choice of suburbs with a high concentration of South African immigrants. In contrast, Okita (2002) described the phenomenon of L1 avoidance among Japanese mothers in the UK married to English men by creating L2 English social milieu for themselves and their children.

6.3 Background Factors and Vocabulary in L1 and L2 In the present study, the focus was on several factors of family background that appear to be linked to FLP, such as parental education, socio-economic status (SES), and length of residence in the host country. Findings regarding parental education and children’s L1 vocabulary maintenance are inconsistent. It has often been claimed that ethno-linguistic minorities need a strong educational experience in their own language and traditions in order to maintain their mother tongue and ethnic identity (Kloss, 1966; Allard & Landry, 1992). For example, Guardado (2002) found that SpanishCanadian families in Vancouver children who maintained L1 (Spanish) came from families headed by parents with higher levels of education. At the same time, Doucet (1991), Harres (1989) and Bill, Hernandez-Chavez, and Hudson (2000) found that the opposite was true, so that the higher the educational level of the informants, the lower loyalty and retention rates for L1 and greater L1/L2 shift. Results concerning socio-economic status (SES) were also contradictory: some families of lower SES were favorably disposed toward language shift, whereas families from higher SES favored language maintenance (Williams, 1987; Lambert & Taylor, 1996). With regard to L2 vocabulary, the SES of immigrant families was found to be an important factor, but research results concerning its role are inconclusive. Most of the existing data on this factor come from a Miami study (Oller & Eilers, 2002, Pearson, 2007) where the parents were Spanish (L1) speaking immigrants from the Cuban elite escaping Castro who arrived in Miami in the early 1960s and were characterized by relatively high social status. Another group of studies focused on Spanish (L1)

6

Second Generation Immigrants: A Socio-Linguistic Approach. . .

123

children of Mexican descent in US, with poor SES, low parental education, very low income and low educational level. Oller and Pearson (2002) argued that the apparent ability of children to profit from the bilingual experience might result from the advantage associated with their socio-economic background, which provides access to more reading material in the home, additional parental involvement in the education process, and consequently greater academic support. In contrast, bilingual children born into poverty have been exposed to fewer books and have had fewer opportunities to increase both their L1 and L2 vocabulary. This argument was examined in the Cobo-Lewis et al.’s (2002) study mentioned above. The researchers distinguished between the effects of SES and bilingualism, and demonstrated that SES had an especially large effect on English (L2) oral proficiency. Note, however, that to date the standard approach to the description of bilingual parental background has been limited to reports on income and occupation, without reference to education as a separate factor. This approach may be attributed to the fact that in most immigrant communities parental educational level is highly correlated with SES (Kosmin, 1990), with no discrepancy between the former and the latter. At the same time, it may be misleading to treat these factors as indistinguishable from one another for some immigrant communities, as for example the Russian-Jewish immigration in Israel, which is the focus of the present study. Concerning the length of the immigrant families’ residence in the host country, the literature suggests that it is strongly associated with both L2 proficiency and L1 attrition among immigrant children (Baker, 2001). The longer the families have resided in the country, the better their command of L2 and the greater the language shift. To what extent this is evident in our case of Russian-Jewish immigrants with a tendency for integration into host society? And does the role of the length of the parents’ residence in the host country retain its significance for L2 knowledge of their children as well? The present study seeks to answer these questions. In sum, despite obvious commonalities among factors affecting the lexical knowledge of the immigrants’ children, it is important to stress that the magnitude and strength of these factors tend to vary from one immigrant population to another. In most situations, the extent of L1/L2 knowledge is likely to be related to a complex interaction of many socio-cultural and socio-linguistic variables reflecting the unique context of each immigrant population (Edwards, 1992). It is important, therefore, to examine each ethno-linguistic group individually. Note, however, that no research can address all factors related to the role of socio-cultural and socio-linguistic conditions in children’s L1/L2 vocabulary knowledge (Baker, 2001; Ellis, 1985; Spolsky, 1989). Concerning L1 vocabulary knowledge as a dependent measure, the focus was, therefore, on the following independent variables of parents’ and children’s language ideology (i.e., language attitudes), parents’ language practice (i.e., language choice with children), parents’ language management (i.e., teaching L1 literacy to their children), background of FLP (i.e., length of parents’ residence in Israel,

124

M. Schwartz

parental education level) and socio-economic index of the school as a measure of parents’ SES. Regarding L2 vocabulary knowledge as a dependent measure, the independent variables included were parents’ language practice (i.e., parentchild language and parents’ self-reported L2 proficiency), background of FLP (i.e., length of parents’ residence in Israel, parental education level, parental education in Israel), and children’s social milieu (percent of Russian-speaking classmates and percent of Russian-speaking friends) and socio-economic index of the school as a measure of parents’ SES.

6.4 Method 6.4.1 Participants The sample consisted of 70 second grade Russian-Hebrew speaking children and 70 of their parents (73% of the parents who participated were mothers). Children were selected from 12 elementary schools situated in the northern region of Israel. The children were speaking Russian as the first and dominant language at home, most of them were born in Israel (n ¼ 65) or immigrated to Israel at least 5 years before the start of schooling.

6.4.2 Measures 6.4.2.1 Socio-Economic Index of the Schools Because information concerning parental SES was not available, data on the socio-economic index (SEI) of each school involved in the study was collected. This index is calculated based on the parents’ reports regarding their income and occupation, and on a ranking of the families’ residential area, but not on their educational level. This index may therefore be considered as an equivalent of the SES of the families participating in the study. The socio-economic index of schools in Israel is measured on a 10-point scale ranging from 1 to 10, with higher scores indicating lower SES. 6.4.2.2 Parents’ Questionnaire The questionnaire was based in part on an existing questionnaire (DonitsaSchmidt, 1999) and was developed in part specifically for this study. The questionnaire took approximately 30 min to complete and included the following 4 sections (lettered A to D): (a) Background information: The parents were asked to provide information about the year of immigration, educational level of both parents, and schooling in Israel. (b) Language practice and Hebrew (L2) proficiency: Parents were asked to state their language of choice: only Russian (1), only Hebrew (2), or both

6

Second Generation Immigrants: A Socio-Linguistic Approach. . .

125

Russian and Hebrew (3), with spouse, parents, and offspring and in various leisure time situations that necessitate the use of language (e.g., reading books for pleasure, reading newspapers, listening to radio programs, and watching TV). In addition, parents were asked to rate their language proficiency in Hebrew (L2) using ‘‘can-do’’ items in different skills (speaking, comprehending, writing, and reading), in both formal and informal situations (e.g., reading a book vs. reading signs). There were 8 items for each language. To increase statistical power and for the purpose of data reduction, a composite measure of parents’ Hebrew language proficiency (M ¼ 4.1; SD ¼ 0.81) has been constructed based on principal component analysis. (c) Language management: The parents were asked whether their child has learned to read Russian, and who exposed the child to Russian literacy (parents, grandparents, caregiver, or teacher within the framework of a community school). (d) Language ideology: Eight items concerning attitudes toward L1 maintenance and two items regarding myths about language, formulated as statements (see details in Schwartz, 2008). Responses were provided on a 5-point Likert-scale, ranging from ‘‘completely agree’’ to ‘‘completely disagree.’’ To increase statistical power and for the purpose of data reduction, principal component analysis was used to extract three factors: (1) ‘‘positive attitude toward development of L1 and L2 at home and L1 literacy acquisition at school’’ (Parents’ attitudes L1þL2); (2) ‘‘attitudes toward developing only L1 at home’’ (Parents’ attitudes L1); (3) ‘‘attitudes toward L1–L2 shifting’’ (Parents’ attitudes L2). 6.4.2.3 Children’s Questionnaire The questionnaire took approximately 20 min to complete and included the following three sections (lettered a to c): (a) Background information: The children were asked about number of close friends, and the number of Russian-speaking friends. (b) Language management and practice: Nine items requesting information about the children’s linguistic choice and use: (1) Russian only, (2) Hebrew only, or (3) both Russian and Hebrew, with father, mother, grandparents, siblings and Russian-speaking friends. This section also provided information about the children’s active practice of reading in Russian, that is, whether the children read frequently in Russian. (c) Language ideology: Eight yes-no questions were asked to determine the attitudes of the children toward L1 maintenance and L1–L2 shifting (see details in Schwartz, 2008). Because of the children’s young age, 3-point scale has been used to score responses, with the highest score (3) representing a positive answer and the lowest score (1) a negative one. Point 2 signified a neutral attitude. The principal component analysis has been applied to

126

M. Schwartz

construct the following three factors: (1) ‘‘positive attitude toward development of L1 and L2 at home and L1 literacy acquisition at school’’ (Children’s attitudes L1þL2); (2) ‘‘attitudes toward developing only L1 at home’’ (Children’s attitudes L1); (3) ‘‘positive attitude toward only Hebrew at home’’ (Children’s attitudes L2). 6.4.2.4 Composite Measures of Russian and Hebrew Vocabulary Two composite measures of Russian (L1) and Hebrew (L2) vocabulary were created by extracting the first principal component from a set of five vocabulary measures (semantic categories, word descriptions with and without stimulus questions, antonyms, and receptive vocabulary) (see details in Schwartz, 2008; Schwartz, Kozminsky, & Leikin, 2009).

6.4.3 Procedure The study was carried out in the first 2 months of the school year. The children’s tasks and questionnaire were completed individually in two separate sessions lasting about 40 min each during school time. The questionnaires for parents were completed at home.

6.5 Results 6.5.1 Background of the Family Language Policy The SEI of the children’s schools ranged from 4.25 to 7.40, indicating a middlelow range. At the same time, the data obtained from the parents’ questionnaires showed that the level of education reported by parents was relatively high (M ¼ 14.0; SD ¼ 1.53, in years). All parents had at least a high school diploma (10 years of school in the FSU, where compulsory education is from age seven to seventeen. Forty-three percent of the women and 47% of the men had earned a university degree (at least 5 years of post–secondary education in the FSU). The correlation between maternal and paternal education was r ¼ 0.41, p < 0.01. The average score of the mother’s and the father’s education served as a measure of parental education. Furthermore, 60% of the parents were educated primarily in Israel: 69% participated in 1-year professional courses and 31% received conventional higher education. The correlation between parental educational level and their educational experience in Israel was reliable but rather low (r ¼ 0.26, p < 0.05). Therefore, parental educational level and their educational experience in Israel were treated as two separate and independent variables. As for length of the families’ residence in Israel, it was relatively long (M ¼ 11.6; SD ¼ 3.93).

6

Second Generation Immigrants: A Socio-Linguistic Approach. . .

127

6.5.2 Family Language Practice and Management The parents’ questionnaire shows that Russian was the primary language used by parents with their spouses (72%), with their parents (100%), and when reading for pleasure (82%). At the same time, a strong tendency to permit a mixture of L1 and L2 in communication with the children was found. 69% of the parents noted that using Hebrew with the children was a choice conditioned mainly by linguistic circumstances, when it was impossible to find the right word or expression in Russian. As a rule, however, parents reported placing emphasis on maintaining L1. Language choice in other personal domains (reading newspapers, listening to radio, and watching TV) tended to include both Russian and Hebrew. As for social environment, 71% of the parents reported being part of a Russian-onlyspeaking social interaction. The children also reported frequent use of Russian with their parents (59% with fathers, 56% with mothers, 96% with grandparents), but some were more likely to communicate with their parents in Hebrew only (9% with fathers, and 11% with mothers). As expected, the children commonly used Hebrew with their siblings (43%) and with Russian-speaking friends (70%). Concerning family language management, 61% of the parents stated that their children have been taught literacy in L1 and in most cases within a supplementary education system. It was further found that 41% of children reported active involvement in reading in L1 and 70% stated that during childhood they were read to in the heritage language only. There was significant correlation between the parental report on children’s L1 literacy and the information provided by the children about active involvement in reading in L1 (r ¼ 0.45; p < 0.01). The absence of a stronger relationship between the former and the latter may be attributed to the fact that some children were not enthusiastic about reading in the heritage language. As for the Hebrew language, the parents reported a relatively high level of proficiency (M ¼ 4.1; SD ¼ 0.81), clearly indicated weaker ability in written domains (i.e., writing a formal letter, reading a book).

6.5.3 Family Language Ideology The majority of the parents had a positive attitude toward their children’s progress in both languages, with emphasis on the maintenance of the heritage language. Parents tended to support dual language development and valued mastery of both languages. In contrast, parents appeared to downplay widespread myths about linguistic handicaps attributed to bilingualism (Cummins, 1984). Only 3% of parents were unequivocal about the negative effect of a bilingual upbringing with L1 dominance at home. At the same time, the parents were decisive about the notion that the language of the host country can be learned easily in school.

128

M. Schwartz

Despite being educated in a non-additive Hebrew context and with the notion that Hebrew is the language of national identity, 60% of bilingual children were agreeable to learning Russian in school. Only 30% of the children supported the idea of a complete L1–L2 shift at home. Finally, the children conclusively voiced the need for being more proficient in both Hebrew and Russian.

6.5.4 Correlational and Regression Analysis and First Language Vocabulary Knowledge A series of correlational analyses was conducted to examine the relation between the children’s vocabulary knowledge in L1 and the independent factors that may be connected to it: parents’ and children’s language ideology (i.e., language attitudes), parents’ language practice (i.e., language choice with children), parents’ language management (i.e., teaching L1 literacy to their children), background of FLP (i.e., length of parents’ residence in Israel, parental education level) and socio-economic index of the school as a measure of parents’ SES. Concerning FLP factors, significant positive relationships were found between the composite measure of vocabulary knowledge in L1 and the parents’ language management, i.e., children’s L1 literacy knowledge, (r ¼ 0.47, p < 0.001). In addition, a relatively weak but still significant negative correlation was found between vocabulary knowledge and the parent’s language use indicating that the more the co-existence of both languages was permitted at home, the weaker the L1 vocabulary knowledge was. Furthermore, a weak and insignificant association between all factors relating to the parents’ ideology regarding language and the dependent variable was found. The parents’ ideology concerning language did not correlate significantly with their reports about language management and practice at home, nor with any of the measures of the children’s attitudes about language. By contrast, there was a significant correlation between the composite measure of vocabulary knowledge in Russian and the children’s positive attitude toward the development of both languages at home and toward L1 literacy acquisition at school (r ¼ 0.39, p < 0.001). The correlational analysis revealed a weak but significant negative relationship between the children’s mastery of L1 vocabulary and their desire to speak only Hebrew at home (r ¼ –0.25, p < 0.05). Finally, children’s reports on active practice of reading in Russian were related considerably with their actual command of the heritage vocabulary (r ¼ 0.40; p < 0.01). Finally, concerning background factors of FLP (length of parents’ residence in Israel, parental education level and SEI of the school as a measure of parents’ SES) no significant correlations were obtained. Furthermore, a multiple regression analysis was prepared (see Table 6.1). The composite measure of L1 vocabulary knowledge was entered as a

6

Second Generation Immigrants: A Socio-Linguistic Approach. . .

129

Table 6.1 Summary of multiple regression analysis for family policy variables contributing to the composite vocabulary knowledge in L1 Variable B SE B b R2 ¼ 0.39 SEI of the school Length of parents’ residence in Israel Parental education Parents’ attitudes L1þL2 Parents’ attitudes L1 Parents’ attitudes L2 Children’s attitudes L1þL2 Children’s attitudes L1 Children’s attitudes L2 Parent’s language practices (language choice with children) Parents’ language management (teaching children to read in L1)

0.12 0.01 0.03 –0.10 –0.18 –0.05 0.40 –0.17 –0.10 –0.20 0.24

0.10 0.03 0.07 0.11 0.11 0.08 0.16 0.11 0.09 0.08 0.08

0.15 0.06 0.05 –0.09 –0.17 –0.06 0.37* –0.22 –0.12 –0.25* 0.31**

* p<0.05; ** p<0.001

dependent variable and the independent variables of FLP and its’ background factors were entered simultaneously as explanatory variables. The explanatory variables accounted for a significant amount of the variance in the bilingual children’s performance on the composite measure of vocabulary proficiency in Russian, F(8, 61) ¼ 5.01, p < 0.001, R2 ¼ 0.39. Two variables that substantially contributed to the bilingual children’s L1 vocabulary were, in order of importance, the parents’ language management (teaching children to read in L1) and the children’s positive attitude toward the development of both languages at home and toward the acquisition of L1 literacy in school. The parents’ language practice (the tendency to use both languages with children) affected negatively on vocabulary mastery in the heritage language. Finally, consistent with the correlation results, the parents’ ideology concerning language and its’ background variables did not contribute significantly to the dependent variable.

6.5.5 Correlational and Regression Analysis and Second Language Vocabulary Knowledge To investigate the relation between the children’s vocabulary knowledge in L2 and the independent factors that may be connected with it, a series of correlation analyses was performed. Significant positive relationships were found between the children’s vocabulary knowledge in L2 and such background factors as parental length of residence in the host country, education level, and education in Israel (r ¼ 0.35, p < 0.01; r ¼ 0.44, p < 0.01; r ¼ 0.37, p < 0.01). At the same time, the correlation analysis showed a rather weak and non-significant association between the SEI of the school, the children’s social milieu and the parents’ linguistic input variables on the one hand, and the children’s Hebrew vocabulary knowledge on the other.

130

M. Schwartz

Table 6.2 Summary of multiple regression analysis for variables contributing to the children’s Hebrew vocabulary knowledge Variable B SE B b R2 ¼ 0.38 SEI of the school Length of parents’ residence in Israel Parental education Parental education in Israel Parent-child language choice Parental Hebrew (L2) proficiency Percent of Russian-speaking classmates Percent of Russian-speaking friends

–0.15 0.22 0.38 0.32 0.00 –0.27 0.00 0.18

0.12 0.11 0.12 0.12 0.12 0.14 0.12 0.11

–0.19 0.25* 0.43** 0.33* 0.00 –0.32 0.00 0.19

* p<0.05; ** p<0.001

Although the children’s L2 vocabulary knowledge and their parents’ L2 proficiency are not correlated, they are strongly related to the same factors: length of residence in Israel, parents’ educational level, and parents’ educational experience in Israel. Correlation analysis also revealed a large coherence between the measures of family language practice (parent-child language choice and parental Hebrew (L2) proficiency). It seems, therefore, that in families that permitted the co-existence of both languages in communication with the children, the reported level of Hebrew mastery was higher than in the families that preferred to use mainly Russian. Next, a multiple regression analysis was performed (Table 6.2). Hebrew vocabulary knowledge was entered as a dependent variable and parental background, language input, and the children’s social milieu were entered simultaneously as explanatory variables. As shown in Table 6.2, the explanatory variables accounted for a considerable amount of the variance in the children’s Hebrew lexical knowledge, F(8, 61) ¼ 4.12, p < 0.01, R2 ¼ 0.38. Three variables that substantially contributed to the bilinguals’ L2 lexical competence were, in order of importance, parents’ educational level, parents’ education in Israel, and length of residence in Israel. Consistent with the correlation results, the relative contribution of the SEI of the school, parent-child language choice, parental Hebrew (L2) proficiency, and children’s social milieu to the dependent variable was not significant.

6.6 Discussion Our knowledge of socio-cultural and socio-linguistic factors in vocabulary development have been constructed mostly on data collected in the Miami study (Cobo-Lewis et al., 2002; Pearson, 2007) and other studies focusing on Spanish (L1) children of Mexican descent in the US (e.g., Hakuta & D’Andrea, 1992). Although this research contributed a great deal to our understanding, it is quite clear we should not generalize about the facilitative or hindering effect

6

Second Generation Immigrants: A Socio-Linguistic Approach. . .

131

of particular factors of the socio-linguistic environment from studies of a number of Spanish-speaking immigrant communities in the US. An important inspiration for the present paper was the notion that in the case of RussianJewish immigration in Israel, there may be similarities and differences with previously-collected data, due to specific features of this novel theoretical approach to analyze a set of separated socio-cultural and socio-linguistic factors within the framework of family language policy.

6.6.1 First Language Vocabulary Knowledge The first interesting result was that the family background factors do not account for the variance in children’s L1 proficiency in the case of second generation Russian-Hebrew speaking immigrants. It may be that the parents’ internal FLP is more influential in light of relatively high educational level and long length in the host country. Further, the results indicate that literacy in the heritage language accounted for a large portion of the variance in Russian vocabulary among the second generation. Note, however, that according to the parents’ and children’s reports, most of the participants received only basic literacy in L1, which was insufficient for substantial vocabulary enrichment. It is reasonable to assume that the children’s exposure to basic Russian literacy has been combined with overt or covert practical work directed at enhancing vocabulary growth. It is interesting to note that these successful parents’ management was conducted frequently is based on the existence of the communitybased after-school educational settings aimed at arousing the children’s interest in the original cultural heritage The data also point to several lacks of connection between the components of family language policy. A significant contribution of the children’s positive attitude toward development of both languages at home and toward heritage literacy acquisition at school to their lexical knowledge of L1 was found. Whereas the parents’ positive attitudes toward home language retention did not contribute reliably to the children’s competence in L1 vocabulary, the children’s attitudes did. Moreover, the results suggest a strong relationships between the children’s positive attitude toward L1 maintenance, active L1 practice (e.g., reading in Russian), and higher scores on L1 vocabulary measures. These lack of connections provide evidence for the relatively high construct validity of the children’s reports. Even though the children’s reports might be relatively unstable and changeable over time, in contrast to those of the parents; they seem to reflect the real family language policy. The results indicate that these reports may be considered an indicator of the effectiveness of their parents’ efforts to preserve the Russian language at home. At the same time, the parents’ language ideology did not have even a minor impact on the children’s command of the heritage language and did not correlate with their reports on their actual language choice with the children and on Russian literacy management at home. Furthermore, the results point out a

132

M. Schwartz

disassociation between parents’ and children’s positive attitudes toward Russian language preservation: close to a ceiling effect was obtained on the parents’ reports but not on the children’s. Placing the findings within the theoretical framework of Spolsky’s (2004) model of language policy reveals a clear discrepancy between the parents’ declared commitment to L1 maintenance and their reports on actual language practice with their children and on written language management at home. One possible interpretation of these finding is in conflict with language ideology and practice which has been observed also in other immigrant communities (Eilers et al., 2006). This conflict might be attributed to the emotional and timeconsuming demands involved in raising bilingual children in immigrant families. These demands might be manifested in balancing the various needs of the child and the need for continuous monitoring of child’s language development in both languages and negotiation between languages (Schwartz & Katzir, 2010).

6.6.2 Second Language Vocabulary Knowledge Concerning L2 vocabulary knowledge, similar to previous data reported by Pearson (2007) in the Miami study, our findings confirmed that parental education in the host country (even if limited to a brief professional course) and the length of residence can advance the second generation’s ability to progress in L2 vocabulary. At the same time, the present study provided unequivocal evidence of the critical role that the parents’ educational level acquired in the country of origin plays in their children’s vocabulary mastery of L2. This suggests that young Russian-speaking bilinguals of even middle-low SES can enter a non-additive Hebrew educational context at an advantage and profit from their parents’ cultural support. More specifically, immigrant families in which parents have a higher level of education also demonstrate a higher level of L2 proficiency, are likely to pay more attention to their children’s academic progress, have more books to read at home in both languages, and act to further their children’s cognitive and linguistic development. This interpretation is supported by the strong correlation between the parents’ educational level and their self-reported L2 proficiency, and by the significant connections between parents’ education and the number of children’s books in Hebrew that are present in the home (r ¼ 0.36, p < 0.05). These results underscore the important factor of the immigrants’ educational level, which can account for the variance in children’s L2 proficiency independently of their SES characteristics, and in our case in contrast to them. Another noteworthy result is that no significant interaction was found between the Russian-speaking peer environment of bilingual children and their L2 vocabulary knowledge. These results suggest this to be true even in classrooms in which the percentage of Russian-speaking children is lowest (4%), the children had 100% Russian-speaking friends. This implies that young bilinguals tend to choose a Russian-speaking socio-linguistic milieu,

6

Second Generation Immigrants: A Socio-Linguistic Approach. . .

133

which may be imposed by parents’ language management. In general, this immigrant population has a tendency for relatively high concentration in certain neighborhoods, which in turn creates a demographically Russianspeaking socio-linguistic milieu. Note, however, that even though these children and their parents prefer Russian-speaking social environment, this appears not to hamper the children’s process of L2 vocabulary acquisition (see data in Schwartz & Katzir, 2010). Finally, the findings point out the non-significant role of the family language practice (parent-child language choice) in explaining the variance in children’s Hebrew vocabulary knowledge. These data can be attributed to the fact that even in families that reported co-existence of both languages, Russian remains the primary language for transmitting traditions and cultural values to the next generation. The extent to which family language practice may change in the course of the children’s education within a non-additive L2 environment merits further longitudinal study.

6.6.3 Conclusions The data obtained from this study extend our knowledge on how the factors concerning family language policy their educational and acculturation background could be influential for mastering both L1 and L2 vocabulary. Over and above the link between specific features of the Russian-Jewish immigrant community and the second generation L1/L2 vocabulary knowledge, these results may be generalized with regard to other immigrant communities (e.g., see discussion on the second and third generations of Spanish (L1) speaking immigrants from Cuba in Miami in Pearson, 2007). When immigrant families are proactive in their language policy and provide activities in written form of L1, the children respond by learning the language. In addition, when immigrant families have favorable educational background obtained in the country of origin, it is shown to benefit second generation children’s level of L2 vocabulary without substantial economic support. Furthermore, the more families permit the co-existence of both languages in family language practice with the children, the higher level of L2 vocabulary mastery and the more endangered is L1 vocabulary. Finally, children’s report on FLP might become a real test of the effectiveness of this policy.

References Allard, R., & Landry, R. (1992). Ethnolinguistic vitality beliefs and language maintenance and loss. In W. Fase, K. Jaspaert, & S. Kroon (Eds.), Maintenance and loss of minority languages (pp. 173–195). Amsterdam: John Benjamins. Baker, C. (2001). Foundations of bilingual education and bilingualism. Clevedon: Multilingual Matters Ltd. Barkhuizen, G. (2006). Immigrant parents’ perceptions of their children’s language practices: Afrikaans speakers living in New Zealand. Language Awareness, 15, 63–77.

134

M. Schwartz

Ben-Rafael, E., Lyubansky, M., Glo¨ckner, O., Harris, P., Israel, Y., Jasper, W., et al. (2006). Building a Diaspora: Russian Jews in Israel, Germany and the USA. Brill: Leiden. Bill, G., Hernandez-Chavez, E., & Hudson, A. (2000). Spanish home language use and English proficiency as differential measures of language maintenance and shift. Southwest Journal of Linguistics, 19, 11–27. Cobo-Lewis, A. B., Pearson, B. Z., Eilers, R. E., & Umbel, V. C. (2002). Effects of bilingualism and bilingual education on oral and written Spanish skills: A multifactor study of standardized test outcomes. In D. K. Oller & R. E. Eilers (Eds.), Language and literacy in bilingual children (pp. 98–117). Clevedon: Multilingual Matters. Cummins, J. (1984). Bilingualism and special education: Issues in assessment and pedagogy. Clevedon: Multilingual Matters. Cummins, J. (2000). Language, power, and pedagogy: Bilingual children in the crossfire. Clevedon: Multilingual Matters. De Houwer, A. (1995). Bilingual language acquisition. In P. Fletcher & B. McWhinney (Eds.), The handbook of child language (pp. 219–250). Oxford: Blackwell. De Houwer, A. (1999). Environmental factors in early bilingual development: the role of parental beliefs and attitudes. In G. Extra & L. Verhoeven (Eds.), Bilingualism and migration (pp. 75–96). Berlin: Mouton de Gruyter. Donitsa-Schmidt, S. (1999). Language maintenance or shift – determinants of language choice among Soviet immigrants in Israel (Unpublished doctoral thesis). Toronto: University of Toronto. Doucet, J. (1991). First generation of Serbo-Croatian speakers in Queensland: Language maintenance and language shift. In S. Romaine (Ed.), Language in Australia (pp. 270–284). New York: Cambridge University Press. Edwards, J. (1992). Sociopolitical aspects of language maintenance and loss: Towards a typology of minority language situations. In W. Fase, K. Jaspaert, & S. Kroon (Eds.), Maintenance and loss of minority languages (pp. 37–55). Amsterdam: John Benjamins. Eilers, R. E., Pearson, B. Z., & Cobo-Lewis, A. B. (2006). The social circumstances of childhood bilingualism: The Miami experience. In P. McCardle & E. Hoff (Eds.), Childhood bilingualism (pp. 68–90). Clevedon: Multilingual Matters. Ellis, R. (1985). Understanding second language acquisition. Oxford: Basil Blackwell. Goodz, N. (1994). Interactions between parents and children in bilingual families. In F. Genesee (Ed.), Educating second language children (pp. 61–81). Cambridge: Cambridge University Press. Gregory, E., Arju, T., Jessel, J., Kenner, C., & Ruby, M. (2007). Snow White in different guises: Interlingual and intercultural exchanges between grandparents and young children at home in East London. Journal of Early Childhood Literacy, 7(1), 5–25. Guardado, M. (2002). Loss and maintenance of first language skills: Case studies of Hispanic families in Vancouver. The Canadian Modern Language Review, 58(3), 341–363. Hakuta, K., & D´’Andrea, D. (1992). Some properties of bilingual maintenance and loss in Mexican background high school students. Applied Linguistics, 13, 72–99. Harres, R. (1989, September 27). The search for minority teachers. Los Angeles Times, pp. A1, A18. Horenczyk, G., & Ben-Shalom, U. (2006). Acculturation in Israel. In D. Sam & J. Berry (Eds.), A Cambridge handbook of acculturation psychology (pp. 294–310). Cambridge: Cambridge University Press. Horowitz, T. (1986). The absorption of the Soviet Jews in Israel, 1969–1984: Integration without acculturation. In T. Horowitz (Ed.), Between two words: Children from the Soviet Union in Israel (pp. 33–49). Lanham, MD: University Press of America. Johnson, J., & Martin, C. (1985). Parents’ beliefs and home language environments: effects on cognitive development. In I. Sigel (Ed.), Parental belief systems. The psychological consequences for children. Hillsdale, NJ: Lawrence Erlbaum Associates. King, K. (2000). Language ideologies and heritage language education. The International Journal of Bilingual Education and Bilingualism, 3(3), 167–184.

6

Second Generation Immigrants: A Socio-Linguistic Approach. . .

135

Kloss, H. (1966). German-American language maintenance efforts. In J. A. Fishman (Ed.), Language loyalty in the United States: The maintenance and perpetuation of non-English mother tongues by American ethnic and religious groups (pp. 206–252). The Hague: Mouton. Kopeliovich, S. (2006). Reversing Language Shift in the immigrant family: A case-study of a Russian-speaking community in Israel. Unpublished PhD dissertation, Bar-Ilan University, Ramat Gan. Kosmin, B. A. (1990). The class of 1979: The ‘‘acculturation’’ of Jewish immigrants from the Soviet Union (Occasional Papers 5). New York: City University of New York, Graduate Center, North American Jewish Data Bank. Lambert, W. E., & Taylor D. M. (1996). Language in the lives of ethnic minorities: Cuban American families in Miami. Applied Linguistics, 17(4), 477–500. Leshem, E., & Lissak, M. (1999). Development and consolidation of the Russian community in Israel. In S. Weil (Ed.), Roots and routes: Ethnicity and migration in global perspective (pp. 135–171). Hebrew University, Jerusalem, Israel: The Magnes Press. Lissitsa, S. (2007). Index of immigrants’ integration in Israel. In M. Kenigshtein (Ed.), ‘‘Russian’’ face of Israel: the features of social portrait (pp. 141–164). Jerusalem, Moscow: Gesharim. (In Russian.) Okita, T. (2002). Invisible work: Bilingualism, language choice and childrearing in intermarried families. Amsterdam: John Benjamins. Oller, D. K., & Eilers, R. E. (2002). An integrated approach to evaluating effects of bilingualism in Miami school children: The study design. In D. K. Oller & R. E. Eilers (Eds.), Language and literacy in bilingual children (pp. 22–40). Clevedon, England: Multilingual Matters. Oller, D. K., & Pearson, B. Z. (2002). The effects of bilingualism: A background. In D. K. Oller & R. E. Eilers (Eds.), Language and literacy in bilingual children (pp. 3–22). Clevedon, UK: Multilingual Matters. Olshtain, E., & Kotik, B. (2000). The development of bilingualism in an immigrant community. In E. Olshtein & G. Hornczyk (Eds.), Language, identity, and immigration (pp. 210–217). Hebrew University, Jerusalem, Israel: The Magnes Press. Pearson, B. Z. (2007). Social factors in childhood bilingualism in the United States. Applied Psycholinguistics, 28, 399–410. Pease-Alvarez, L. (2003). Transforming perspectives on bilingual language socialization. In R. Bayley & S. Schecter (Eds.), Language socialization in bilingual and multilingual societies (pp. 9–24). Clevendon: Multilingual Matters. Schwartz, M. (2008). Exploring the relationship between family language policy and heritage language knowledge among second generation Russian-Jewish immigrants in Israel. Journal of Multilingual and Multicultural Development, 29(5), 400–418. Schwartz, M. & Katzir, T. (2010, in press). Depth of lexical knowledge among bilingual children: The impact of schooling. Reading and Writing. Schwartz, M., Kozminsky, E., & Leikin, M. (2009). Socio-linguistic factors in second language lexical knowledge: The case of second generation children of Russian-Jewish immigrants in Israel. Language, Culture and Curriculum, 22(1), 14–27. Seville-Troike, M. (2000). Causes and consequences of language maintenance/shift. In E. Olshtein & G. Hornczyk (Eds.), Language, identity, and immigration (pp. 159–171). Jerusalem: The Magnes press, Hebrew University. Spolsky, B. (1989). Conditions for second language learning. Oxford: Oxford University Press. Spolsky, B. (2004). Language policy. Cambridge: Cambridge University Press. Stavans, A., & Goldzweig, G. (2008). Learning Hebrew as a second language by Ethiopian and Russian immigrants in Israel: ‘‘Must’’ or ‘‘Have’’. Israeli Studies in Language and Society, 1(2), 59–85. Williams, G. (1987). Bilingualism, class dialect and social reproduction. International Journal of the Sociology of Language, 66, 85–98.

Chapter 7

Understanding Language Achievement of Immigrants in Schools: The Role of Multiple Academic Languages Tamar Levin and Elana Shohamy

7.1 Introduction Linguistic alnd cultural differences have now become central to the pragmatics of schooling. Immigration requires that in order to effectively interact and adjust to new school environments, we need to use multiple languages and communication patterns that more frequently cross cultural, community, and national boundaries. Multicultural and multilingual approaches to school learning challenge therefore conventional notions of content, learning, teaching, and assessment. They recognize that school learning calls for crossing boundaries of languages, cultures and especially learning patterns for all students regardless of their backgrounds, i.e., immigrants and non-immigrants alike. Indeed, in a number of academic contexts, many students are currently learning different subject areas in their second or third languages (Ellerton & Clarkson, 1996), as is the case in ‘‘content based’’, ‘‘immersion’’ and ‘‘heritage’’ types program. These students, who come from diverse languages and cultures, bring into the classroom their own ways of constructing knowledge based on their homes and communities. These approaches are of particular relevance when students’ home cultures include practices, speaking genres, and interactional norms that are not aligned with the culture and discourse of the school and even with the culture and discourse of particular subject areas. Such misalignments require students to shift between different types of knowledge, practices, and discourses if they are to have access to school learning, without abandoning their home cultures and languages. In other words, students move from their everyday cultures associated with their homes to the culture of school and the subject matters learned, using transitions that are referred to by Giroux (2005) as ‘‘border crossing.’’ Such a process is often accompanied with children’s feeling that their home language, the one they know, is of little value (Sanchez & Thorp, 1998).

T. Levin (*) Tel Aviv University, Tel Aviv, Israel e-mail: [email protected] M. Leikin et al. (eds.), Current Issues in Bilingualism, Literacy Studies 5, DOI 10.1007/978-94-007-2327-6_7, Ó Springer ScienceþBusiness Media B.V. 2012

137

138

T. Levin and E. Shohamy

Due to such disparities between school languages and culture versus students’ home languages and cultures, the quality of educational experiences and outcomes may suffer if and when contents, languages and norms are imposed on students who do not share mutual systems of meanings, symbols, and practices. Equitable learning, teaching and assessment practices for diverse students involve therefore consideration of their cultural and linguistic experiences, in preparing them to function competently in the schools as well as in their homes and communities. It is thus generally accepted that both assessment and teaching should enable diverse students to connect their cultural norms with mainstream expectations, and vice versa. It is also expected that such teaching and assessment practices in the classrooms should enable diverse students to meet national standards as well as to maintain their cultural and linguistic identities all at the same time. This also implies that both teaching and assessment practices should maximize opportunities to demonstrate diverse students’ knowledge and abilities in ways, which are compatible with their cultural and language backgrounds. Schooling therefore needs to mesh with different subjectivities, and with their attendant languages, discourses, and registers, and use these as resources for learning (Johns, 1998). Too often, however, students, immigrants and nonimmigrants alike, indigenous groups, and any other groups with varied backgrounds, are left to manage border crossings on their own. Aikenhead (1997), who studies the cultural aspects of science learning, proposes to view the role of a science teacher as a ‘‘tour-guide’’ or a ‘‘culture broker’’ because most students require assistance from a teacher, similar to a tourist in a foreign land requiring the help of a tour guide. Learning processes need to recruit, rather than to ignore and erase, the different subjectivities including – interests, commitments, desires, capabilities, thinking routines, and aims – in order to ‘‘bring’’ students into learning. Success at learning the nature and knowledge of different schoolbased languages and cultures of multiple subject areas depends, in part, on how smoothly and effectively one crosses these very cultural borders. One such case, frequently referred to in the literature, is in the domain of school science. Although science inquiry is a challenge for most students, it may present additional challenges to students whose background cultures do not encourage students to engage in the practice of science inquiry by asking questions, designing and implementing investigations, forming hypotheses and finding answers independently. Certain cultural values and practices may dispose students to accept teachers’ authority unquestioningly rather than explore or seek alternative solutions. That is, there are basic differences between the underlying assumptions and ways of knowing traditional knowledge about nature and western science and religion-based knowledge about nature and naturally occurring events. These differences place sometimes students from cultures with traditional indigenous knowledge at odds with modern western science tradition (Arellano et al., 2001). Consequently, there are often cultural clashes in science classrooms. These clashes tend to be more extensive, the more a student’s home culture differs from the culture of Western nations, for

7

Understanding Language Achievement of Immigrants in Schools . . .

139

instance, as is the case with aboriginal students worldwide (Aikenhead, 1997). Furthermore, science and mathematics’ terminology is often complex and words used therein are endowed with meanings, which in most cases are completely different from their normal daily usages, and are therefore not totally familiar to immigrant students especially in the new uses that are given to these words in academic learning. While a number of studies investigated the academic proficiency of immigrant students, especially in the context of bilingualism, and while most acknowledge a focus of ‘‘language as a barrier’’ for immigrants, there is a limited understanding about the specific role that school/academic language plays in assessing and promoting immigrants’ achievements, and what this consists of. Cummins’ ‘‘threshold hypotheses’’ (2001) advocates that for students who speak two or more languages, the interplay in the learning process between the language codes may either assist or detract them from learning. On the one hand, if a bilingual or multilingual student has reached a ‘‘threshold’’ of competence in the two or more languages, then the learner may have a cognitive advantage. On the other hand, those bilingual or multilingual students who are not completely proficient and fluent in either of the two (or more) languages tend to experience difficulties in specific school content areas. Indeed, the language of schooling (Schleppegrell, 2004), or the academic language, is a notion that has been around for some time (Halliday, 1989), and researchers in the areas of language minority and multicultural education have suggested different dimensions and meanings of academic language proficiency. Each dimension of academic language proficiency on its own falls short of being all encompassing; yet, collectively they provide a deep insight into the complexity of this construct (Guerrero, 1997). In view of Valdez’s (2004) claims concerning the lack of clear definition of what constitutes academic language and the types of academic language required for performing successfully in academic contexts, this chapter expands the notion of academic language with regard to the subject areas learned in schools. Furthermore, based on the assumption that different subject domains are characterized not only by unique concepts but also by linguistic features and by particular language functions, we provide evidence about the classrooms in Israel as multilingual and multicultural contexts. We do this not only because in most situations in Israel students come from diverse languages and cultures, but also because of the different language functions that are relevant for distinct domains, namely, the existing diversity of subject-matters in specific academic languages. Our current research is informed by empirical results that emerged from socio-cultural theories of learning and from views on language socialization. It elaborates on a socio-cultural theory of human learning and development advanced by Vygotsky (1986) to explain learning achievements of immigrant and non-immigrant students. According to this perspective, all human activity, including the learning of language, math, science and literature cannot be dissociated from the individual, social, cultural and historical contexts in which it is embedded. In particular, language learning is described as a complex

140

T. Levin and E. Shohamy

cultural process of integration into new society whereby the performance and identity of students is shaped as they adopt the socio-cognitive norms and shared meanings of their subject area groups. According to this framework and taking an example from the learning of mathematics, ‘‘mathematics concepts only have meaning within the linguistic and social contexts from which they were derived’’ (Ellerton & Clarkson, 1996, p. 994). Drawing on this perspective, we maintain that the language context of schools are highly varied in terms of the subject areas learned and their specific epistemologies, structure, nature of activities, functions, styles of interaction, discourse register and degrees of students’ involvement. Yet, it is not only science and mathematics. Within all academic disciplines there are a range of terms and words that make up the language of that academic community. This language evolved and developed within disciplines to communicate in particular ways and imply specific ways of seeing and thinking, in relation to specific subjects. This discipline-specific type of communication is not familiar to those outside of the discipline who attempt to enter this academic community. This study therefore views school learning as a process that intrinsically involves the use of a variety of specific languages, which are representative of a number of different disciplines learned in the school. It acknowledges that in bilingual and multilingual settings, the challenge of learning different school subject areas encompasses at least three dimensions, since it simultaneously entails three components: a. access to language of learning (e.g., Hebrew in Israel); b. access to the subject area discourse; and c. access to classroom discourse, namely, dimensions resembling those involved in language learning.

7.2 Literature Review 7.2.1 Multiple Meanings of Academic Language There is general agreement among educators and researchers that the distinct type of language used in classrooms for learning and teaching, referred to as academic language, often hinders the academic achievement of some student groups even though they may be proficient in varieties of the dominant language in non-academic contexts (Cummins, 1981). However, there are conflicting views regarding what exactly constitutes ‘‘academic language.’’ Academic language is generally viewed as the language used in schools, by teachers in classroom talk, in conversations between teacher and students, and especially in textbooks, class discussions about abstract ideas, tests, computer courseware or internet-based practices. According to Solomon and Rhodes (1995), it is the language used in spoken and written academic discourse, where it serves to acquire knowledge and is used for academic functions such as analyzing, explaining, justifying, and evaluating content, ideas, and

7

Understanding Language Achievement of Immigrants in Schools . . .

141

processes. As a result, academic language includes more than just vocabulary. It is complex in that it relates to communicative skills that are part of the curriculum and involves organizational and textual competencies such as the use of cohesive markers (Bachman, 1990). Three distinct views dominate the body of research literature on academic language. The first builds on the idea of a collection of specific and unique features and language structures, the second relates to both cognitive and daily functions and language context and the third refers to a range of contextualized dimensions related to language structure, modes and socio-cultural components. Thus the three views reflect correspondingly reductive-functional, cognitive and social, and contextual-situational approaches. More specifically, the first view suggests that academic language is a collection of unique language functions and structures that are especially difficult for English Language Learners (ELL) to master (Hamayan & Perlman, 1990). Valdez Pierce and O’Malley (1992) and O’Malley and Valdez Pierce (1996) propose a handful of academic language functions that are characteristic of classrooms in general such as seeking information, informing, analyzing, comparing, classifying, predicting, hypothesizing, justifying, persuading, solving problems, synthesizing, and evaluating. In a study performed by the National Center for Research on Cultural Diversity and Second Language Learning (Short, Mahrer, Elfin, Liten-Tejada, & Montone, 1994), academic language was defined broadly to include: (1) semantic and syntactic features such as vocabulary items, sentence structure, transition markers, and cohesive ties; and (2) language functions and tasks that are part of the social studies classroom routine, such as defining terms, explaining historical significance, reading expository text, and preparing research reports, which are considered as ‘‘non-language’’ subjects. Furthermore, based on the assumption that different subject domains are characterized not only by unique concepts but also by linguistic features and by particular language functions, researchers have noted different language functions that are relevant for distinct domains. For example, in order to manage the language of social studies, students are first required to internalize abstract concepts such as democracy, which are presented within long sentences using multiple embedded clauses, and then they are expected to use adequate language to express their points of view on the matter (Renninger, 1992). The academic language of history on the other hand, often uses time-specific language, signaling of cause and effect, hypothesizing, generalizing, comparing and contrasting, and adopting specialized vocabulary (Coelho, 1982). In the sciences, according to Halliday (1989), the following academic language features are involved and used: interlocking definitions, technical taxonomies, special expressions, lexical density, syntactic ambiguity, grammatical metaphor, and semantic discontinuity. However, Halliday also emphasizes that these features do not occur in isolation but rather overlap with one another, particularly in text passages. Chamot and O’Malley (1986), who also describe the functions of scientific academic language, note that the field of science also

142

T. Levin and E. Shohamy

uses technical and non-technical terms that have unique meanings in a scientific context (e.g., ‘‘table’’, ‘‘energy’’), and that scientific discourse is characterized by a particular sequence of steps with a heavy reliance on passive constructions and long noun phrases. Concerning the specific language used in mathematics classes, Spanos, Rhodes, Dale, and Crandall (1988) observe the following syntactic features: comparisons (greater than/less than), logical connectors (if . . . then, given that), reliance on passive constructions, and various uses of prepositions. They also identify several semantic features of mathematics language: technical vocabulary (e.g., additive inverse, coefficient), ordinary vocabulary that has different meanings in math (e.g., square, power), complex combinations and strings of words (e.g., negative exponent, least common denominator), synonymous words and phrases (e.g., add, plus, and combine), and various mathematical symbols and notations. More generally, Ilany and Margolin (2010) also demonstrate the need to bridge between mathematical language that requires an awareness of the mathematical components, and natural language that requires a literacy approach to the whole text, thereby revealing correspondence between the linguistic situation and the appropriate mathematical model. Language in subject-specific contexts however is not limited to learning new concepts through new words or a new system of terminology, which reflects the knowledge structures of a particular subject. It goes far beyond that as it refers also to new ways of thinking within the framework of the particular domain, thereby reflecting their specific world views. It also requires new ways of communicating, of understanding and of producing a variety of text types or genres, and of engaging into specialized forms of discourse, which follow certain traditions, conventions and expectations. The second dominant view differentiates between cognitive academic language and social language (Cummins, 1984, 1989, 1991). Cummins conceptualized Cognitive Academic Language Proficiency (CALP) as consisting of two intersecting continuums that characterize academic language: the degree to which a communication task is cognitively demanding and the degree of contextual support available for the construction of meaning. These two basic distinctions characterizing academic language are important, but it is perhaps even more critical to understanding academic language proficiency. According to Cummins (1984), the more context-reduced and cognitively demanding the language task, the more it will be related to achievement. In addition, Cummins claims that as students progress through the grades, they are increasingly required to manipulate language in cognitively demanding and context-reduced situations that differ significantly from everyday conversational interactions. More specifically, in distinguishing the two kinds of language types, Cummins (2000) asserts that CALP is characterized by an expanded range of vocabulary and complex grammatical structures, an ability to make complex meanings explicit, and greater demand on memory, analysis, and other cognitive processes. Cummins thus emphasizes that the concept of CALP is not

7

Understanding Language Achievement of Immigrants in Schools . . .

143

restricted to syntactic features of language but includes semantic and structural aspects that refer to the social and institutional particularities of classrooms and schools – a perspective which has been developed also by Halliday (e.g., 1978). In contrast, Basic Interpersonal Communicative Skills (BICS), is characterized by language skills that are needed in social conversations. Therefore, it is used in contexts that are rich in visual and physical cues and refers often to topics familiar to the learner. While these two constructs have been questioned and criticized (Valdez, 2004), they still continue to be used extensively in research that is trying to understand the language components of the difficulties immigrant students face in schools. Cummins (2000) provides a multidimensional view of language proficiency by including cognitive demands alongside the formal/informal distinction in his characterization of oral language. Finally, the third view proposed by Solomon and Rhodes (1995) reflects a socio-linguistic perspective suggesting that academic language is a register that includes task-specific stylistic registers. Halliday (1978) explained the concept of register by stating that ‘‘The language we speak or write varies according to the type of situation’’ (pp. 31–32). Solomon and Rhodes, relying on this definition, view academic language registers not only as specific to each domain area but also as shaped by the particular academic task at hand. That is, teachers choose the specific styles of language they think are required of their students for the academic tasks at hand. Accordingly, academic language implies the ability to express knowledge by using recognizable verbal and written academic formats. For example, students must learn acceptable, shared ways of presenting information to the teacher. These formats may or may not be explicitly taught as part of a curriculum, but usage is expected of the students. On the basis of sociolinguistic views, academic language includes the following: knowledge of appropriate language production associated with different contexts and for different purposes, understanding of language functions and genres, ability to write cohesively, and knowledge of the various rhetorical modes, which commonly appear in academic fields. When integrating these three conceptualizations of academic language, according to Scarcella (2003), academic language refers to multiple, dynamic, inter-related competencies. Academic language is, in essence, a variety or a register of the language used in professional books and is characterized by specific linguistic features associated with academic disciplines. Thus, academic language is not a static entity but rather evolves along with technological advances and research discoveries within each discipline or area of study (Schleppegrell & Colombi, 2002). The mastery of academic literacy is disciplinespecific. It entails linguistic, cognitive, and socio-cultural components. Academic language also includes a discourse component that enables students to use linguistic forms and meanings to communicate coherently in organized ways (Canale & Swain, 1980). Finally, academic language leads to basic forms of classroom-discourse to which all students are entitled and which are fundamental prerequisites to survive in schools and to learn efficiently.

144

T. Levin and E. Shohamy

7.2.2 Academic Language and Educational Achievement Language proficiency is considered a powerful factor that affects the differences in the outcomes of educational achievement between immigrant groups and those born in the country and who use the majority language from birth. Nevertheless, the relationship between language background and achievement is not so simple or straightforward. On the one hand, immigrant students, often referred to in the US as English language learners (ELL) have been found to lag significantly behind non-English language learners and to perform worse at all grade levels and in all content areas. However, there is also a mounting body of evidence suggesting that some minority groups do well in schools even though they do not share the home language and culture of the dominant group that is part of the school system (Rumbaut, 1996). In fact, some of those who are doing exceptionally well in US American schools differ more from the dominant group in terms of language barriers and culture than those who are doing less well. Language barriers were found to be more detrimental for students of low socioeconomic background, leading Rumbaut (1996) to suggest that knowledge of English (as a second language) is positively related to test scores when ethnicity, SES and family structure are controlled. There are also indications that immigrant students’ bilingual background is positively associated with school performance (Stanton-Salazar & Dornbusch, 1996). However, there are also data showing that immigrant youth’s second language spoken at home is not a significant predictor of school progress and school completion, when compared to other controls groups (Vernez & Abrahamse, 1996). In contrast, mastery of both mother tongue and the new language is found to be meaningful for immigrant children’s cognitive growth (e.g., Rumberger & Larson, 1998). Research has also shown that being schooled in one’s second language is not a quick and easy process. Moreover, many longitudinal studies examining this process have concluded that only enrichment schooling through students’ two languages provides the conditions needed to eventually reach grade-level performance in second language within 4–7 years (e.g., Cummins, 1981; LindholmLeary, 2001; Thomas & Collier, 2001). The integration of language, content, and culture into the school curriculum has been found to be a promising instructional approach that addresses the needs of an increasingly diverse, multicultural society. Other factors, such as length of residency in the host country, generation, and country of origin or ethnicity, do not show consistent results either. For example, while some research shows that length of residence in the host country tends to be positively related to academic achievement (e.g., Vermeulen & Perlmann, 2000; Fan & Karnilowicz, 2000); according to Portes and Hao (1998) longer periods of U.S. residence lower academic performance for second generation immigrants, and this is true regardless of school context. Rumbaut

7

Understanding Language Achievement of Immigrants in Schools . . .

145

(1999) also shows that the level of school engagement and effort decrease over length of time in the US, and that duration of residency – a proxy for degree of acculturation – retains a strong and negative influence on academic grade point averages. Based on the review above, three research questions were posed: 1. Does student achievement vary by specific content domains and is the pattern of achievements across domains similar for the three groups of students: Israeli born, USSR born and those born in Ethiopia, 2. What is the level of students’ achievements as a function of its specific language and modes of knowledge representations, within a particular content domain (in this case the domain of mathematics)? 3. What role do the modes of knowledge representation play in explaining student achievement in mathematics?

7.3 Method 7.3.1 Sample A stratified sampling procedure was used on the entire student school population of Israel for obtaining a valid national representation according to the following parameters. The students’ country of origin – FSU (Former Soviet Union), Ethiopia; the number of immigrants per school relative to nonimmigrants (schools with large and small number of immigrants) and school’s SES, as determined by the criteria used by the Ministry of Education. Immigrant students were proportionally sampled according to country of origin and the number of students per school from FSU and Ethiopia. It is important to note that the sampling procedure used whole classes for testing since pulling out individual students from classes could create a bias. However, considering the number of students that would enable to meaningfully infer statistical conclusions, a sample of 1000 students was selected for each grade level: 500 immigrant students and 500 Israeli born. Using a stratified cluster (classes) sample, the sampling procedure was performed separately for each grade level and distinctly for FSU and students from Ethiopia. The main variables that affected the cluster definition were the percent of immigrants in each grade level and a school index representing its need for support based on the distribution of students with low SES. Since the sampling unit was based on classrooms and in order to get true estimate values of the real (true) sample variance, a correction named ‘‘the structure effect’’ was made using the estimated value. Thus, the total sample of the study consisted of 299 schools, 100 elementary schools (Grade 5), 100 secondary schools (Grade 11), and 99 Junior high (Grade 9). Table 7.1 displays the student sample in each of the grade levels.

146

T. Levin and E. Shohamy

Table 7.1 Distribution of the number of students across grade levels and country of birth Number of students Group 5th grade 9th grade 11th grade Israeli Native-born FSU Ethiopia

512 421 156

443 348 110

366 297 108

7.3.2 Instruments Two instruments were used in the study: achievement tests in mathematics and achievement tests in academic Hebrew. The instruments were developed specifically for the study since no other standardized instruments that tested academic Hebrew or mathematics literacy, nor instruments that provided accommodations for immigrants from the former USSR and Ethiopia, were available. The instruments were developed by two teams, one for mathematics and one for academic Hebrew, each including an expert in teaching and assessment of the very subject area, two practicing teachers of the subject who were also graduate students, one from the elementary school level and one from the secondary school, and a language expert whose role was to examine how the language was used in the tests, mainly in relation to concepts and terms as they are (being) used in the first language of each immigrant groups. In the development process of the instruments immigrant teachers from former USSR and Ethiopia were consulted on the presentation modes, language, and genres of the test items as well as on their contents, alerting the team as to possible difficulties and/or suggesting accommodations. The national supervisors who are responsible for the curriculum of each subject examined both tests and their suggestions were incorporated. The tests were then piloted in five classrooms consisting of a large numbers of immigrant students. During the study a professional company used standardized administration processes of the tests and the teacher along with or a representative of the research team or the company were present in the classroom. Two parallel forms were used for each grade level, so that for each subtest and measure there are two (Cronbach a) coefficients described below for each; these instruments will be described next. 7.3.2.1 The Language (Hebrew) Achievement Test Since Hebrew is used as the only medium of instruction in all schools; two main factors guided the structure and the construction of the Hebrew tests. The first was that each subject area learned in school was characterized by specific academic language in terms of lexicon, rhetorical structure, genre, and linguistic complexity. The second was the reference to the specific grade levels. Thus the tests were constructed in relation to the language of the specific domains in Grades 5, 9 and 11 and the academic content taught in these grades. The specific

7

Understanding Language Achievement of Immigrants in Schools . . .

147

domains selected were: history, literature, Bible studies, civics, science, earth science, depending on whether these areas were studied at the specific grade level. Within these areas, reading comprehension and writing were the focus on the tests. The choice of the specific Hebrew needed for each of the academic subjects learned in schools, for each of the grade levels, was deduced from careful and systematic examination of the national school curriculum, specific school practices, textbooks and intensive meetings with school experts and guidance from the national Hebrew language supervisors. The total score of the Hebrew test comprised of 80% from the reading comprehension component and 20% from the writing. The rationale for the decision to base the Hebrew test on academic school content was based on the assumption that for immigrants, Hebrew language serves as the primary means for academic functioning in all the school subjects. These considerations also led to the decision to test reading comprehension and writing as these are the most essential areas for academic functioning in schools (Cummins, 2002). Within the academic language, the reading comprehension part focused on two areas: (a) school topic and content learned in specific grade levels, and (b) levels of comprehension processing manifested in three levels: lexical, sentence level (local interpretation) and global/inferential level, referring to the comprehension of the text as a whole. Thus, each of the texts included was based on content areas appropriate for the different grade levels, accompanied by questions of different text types and genres such as multiple-choice, open ended, table completions, flow charts, graphs as well as short performance tasks. Short texts were selected so that a variety of text types and genres could be represented. The Hebrew literature part, for example, included excerpts of poems, fables and prose narratives. In addition, the test included texts related to school contexts such as school brochures, announcements, instructions and codes of conduct. For each grade level in several of the selected content domains separate items were generated: those examining achievement of a domain’s lexicon and those examining integration of concepts and contents understanding. The total test score labeled Total Hebrew included 41 items in one form and 36 items in the second, and the reliabilities were .93 for the two forms. For the 9th grade the total test score included 53 items in one form and 64 items in the second form and the reliabilities were .96 and .95 for the first and the second forms, respectively. For Grade 11 the total test score was based on 49 items in one form and 53 items in the second and the reliabilities were .96 and .97 for the first and the second, respectively. 7.3.2.2 The Mathematics Test Four general dimensions of mathematics literacy defined the four major content-based subtests: (a) arithmetic or algebraic procedures; (b) numeric, geometric, or algebraic reasoning; (c) problem solving, and (d) mathematics communication. In addition however in order to examine the issues related to

148

T. Levin and E. Shohamy

academic language in mathematics we defined measures relating to mathematics communication using Brenner’s (1998) classification of communication in, about, and with mathematics. Specifically, communication in mathematics implied using the language and symbols of math (i.e., special vocabulary; modified everyday vocabulary; the syntax particular to expressions of mathematical relationships). Communication about mathematics meant describing problem-solving processes and reflecting upon these processes (i.e., describing procedures, reasoning, justifying decisions, explaining point of view, and reconciling differences). Communication with math referred to using math to deal with meaningful problems (i.e., interpreting arguments using math concepts, providing alternative solutions). Also to examine the role of modes of knowledge presentation in mathematics achievement, we used items identical in their contents and goals but different in their presentation modalities (such as verbal, visual, and symbolic). This manipulation reflects an assumption: Although these different modes are considered similar in terms of problem understanding, they actually are complementary as each one reflects a different dimension of mathematics literacy. Such a manipulation can be considered a reflection of code switching, requiring flexibility in transition between modes within mathematics. The test reliabilities were 0.94 to 0.96 for the first and the second forms, respectively. For the 9th grade the total test score was based on 67 items in one form and 71 items in the second and the reliabilities ranged from 0.97 and 0.98 for the first and the second forms, respectively. For Grade 11 the total test included 67 items in one form and 62 items in the second form and the reliabilities were 0.97 each for the two forms of the total test.

7.4 Results Concerning the first question: Does student achievement vary by specific content domains and is the pattern of achievements across domains similar for the three groups of students: Israeli born, born in the USSR and in Ethiopia, mean and standard deviations of student achievements are described in Table 7.2. The table demonstrates the profiles of student achievements in the different content domains, by grade levels and groups, as well as by achievement level in the subtests of each domain’s lexicon and contents. The results show differences in the achievement levels across content domains. The results show that less differentiation exists in the fifth grade, between the achievement levels of students in the three groups, within the domains and between the different domains. Nevertheless, in Grade 9 the results demonstrate a relatively consistent pattern in the student achievements on the different dimensions both within the specific learning domains and between the different learning domains, for each of the groups. This pattern shows that student achievements in the lexicons are always higher than their achievements on the contents of each domain. The results further show that student achievements in

7

Understanding Language Achievement of Immigrants in Schools . . .

149

Table 7.2 Student achievements by learning domains and groups and rank order of difficulty levels of the lexicon ISR FSU ETH Fifth Grade Lexicon Science Science Lexicon Bible Bible Lexicon Math Mathematics Lexicon Literature Lexicon Language Literature

0.59 (0.37) 3* 0.55 (0.33) 0.55 (0.36) 4 0.58 (0.33) 0.64 (0.27) 2 0.60 (0.22) 0.69 (0.35) 1 0.55 (0.36) 0.63 (0.26)

0.45 (0.39) 3 0.45 (0.36) 0.44 (0.37) 4 0.47 (0.36) 0.59 (0.30) 1 0.52 (0.24) 0.55 (0.50) 2 0.45 (0.38) 0.51 (0.30)

0.36 (0.39) 3 0.33 (0.32) 0.33 (0.36) 4 0.40 (0.34) 0.35 (0.21) 2 0.36 (0.21) 0.43 (0.40) 1 0.33 (0.38) 0.42 (0.26)

Ninth Grade Lexicon History History Lexicon Bible Bible Lexicon Math Mathematics Lexicon Literature Lexicon Language Literature

0.89 (0.23) 1 0.66 (0.26) 0.71 (0.28) 2 0.58 (0.31) 0.70 (0.28) 3.5 0.53 (0.27) 0.70 (0.36) 3.5 0.67 (0.27) 0.62 (0.27)

0.80 (0.35) 1 0.64 (0.29) 0.61 (0.33) 3 0.46 (0.32) 0.68 (0.29) 2 0.52 (0.28) 0.56 (0.41) 4 0.60 (0.42) 0.55 (0.30)

0.68 (0.38) 1 0.50 (0.27) 0.51 (0.32) 2 0.39 (0.31) 0.45 (0.29) 3 0.28 (0.17) 0.44 (0.40) 4 0.46 (0.38) 0.41 (0.26)

Eleventh Grade Lexicon History Science Lexicon Bible Bible Lexicon Math Mathematics

0.78 (0.39) 1 0.74 (0.27) 0.67 (0.40) 3 0.64 (0.32) 0.69 (0.30) 2 0.59 (0.27)

0.35 (0.45) 2 0.49 (0.35) 0.26 (0.39) 3 0.34 (0.35) 0.58 (0.33) 1 0.49 (0.28)

0.41 (0.46) 1 0.50 (0.32) 0.27 (0.37) 3 0.40 (0.34) 0.29 (0.27) 2 0.22 (0.17)

* Rank order of difficulty level based on the lexicon achievements IS Israeli born; FSU born in former Soviet Union; ETH born in Ethiopia

both the History lexicon and History are the highest of all. The lowest scores were reached in mathematics and literature, depending on the different groups. The rank order levels of achievements between the different domains, where 1 indicates the easiest subtest and 4 the hardest, are similar between the group of students who were born in Israel and those who were born in Ethiopia. Slight differences exist for students from the former USSR. In Grade 11 however, the results show differences in achievement levels between different domains mainly in the groups of Israeli born and students who were born in the former soviet union. Also in the 11th grade differences in the degree of discrimination between achievements in different learning domains differ between the groups: it is smallest in the group of Israeli-born students and largest among students who were born in the former USSR. The results further show that with the exception of achievement in ninth grade mathematics for FSU and Israeli born students, large and significant differences exist between both groups of immigrants and native Israelis, and

150

T. Levin and E. Shohamy

these gaps exist for a considerable number of years of residence in the country (Levin & Shohamy, 2007, 2008). The second question concerns an in depth view of students achievements as a function of its specific language and modes of knowledge representations, within a particular content domain, in this case the domain of mathematics. Table 7.2 provides a view of the pattern of students’ mathematics achievement as a function of the different measures, which reflects three different types of language use: (1) mathematics register, expressing mainly the use of conventional language and symbols of mathematics. i.e., special vocabulary; modified everyday vocabulary, and the syntax of particular expressions of mathematics relationships, referred to by Brenner as ‘‘communication in mathematics’’; (2) mathematics reasoning which represents mainly ‘‘communication about mathematics,’’ describing procedures, reasoning, justifying decisions, explaining point of view, and reconciling differences, and (3) mathematics problem solving, which characterizes ‘‘communication with mathematics’’ expressing the use of mathematics to deal with meaningful problems. This includes interpreting arguments using mathematics concepts and providing alternative solutions. It is evident from Table 7.3 that communication in mathematics, involving mainly the use of mathematics language, is easier than mathematics communication with and about mathematics, which integrates mathematics language, day-to-day language, and complex cognitive functioning, using the two languages in the context of reasoning or arriving at problem solutions. These results indicate that the interaction between components of ‘‘specific mathematics language’’, ‘‘mathematics discourse’’, and the ‘‘cognitive processes’’ required for a solution process is more difficult than non-contextual measures of mathematics language. Interestingly, and important to note, is the fact that this pattern of difficulty concerning the use of language in mathematics tasks is similar for all three group levels, including immigrants as well as natives. Relying on the potential effects of different forms of external representation of a content domain such as text and graphics on comprehension and learning, the third question examines the role that the modes of knowledge representation Table 7.3 Performance measures on types of mathematics communication by ethnic groups and grade levels 5th grade 9th grade 111th grade IS FSU ETH IS FSU ETH IS FSU ETH Communication 0.72 0.67 in math (math (0.27) (0.30) register) Communication 0.55 0.50 about math (0.22) (0.21) (math reasoning) Communication 0.56 0.50 with math (math (0.25) (0.27) problem solving)

0.49 0.68 0.70 0.44 0.64 0.55 0.28 (0.27) (0.28) (0.29) (0.29) (0.30) (0.33) (0.27) 0.35 0.53 0.57 0.28 0.56 0.53 0.21 (0.21) (0.32) (0.32) (0.24) (0.32) (0.31) (0.19) 0.33 0.48 0.49 0.22 0.55 0.44 0.24 (0.25) (0.29) (0.30) (0.16) (0.29) (0.30) (0.22)

7

Understanding Language Achievement of Immigrants in Schools . . .

151

Table 7.4 Performance measures on mathematics by modes of task presentation, by groups and grade levels 5th grade 9th grade 11th grade IS FSU ETH IS FSU ETH IS FSU ETH Symbolic (model) Visual (graphical) Verbal

0.71 (0.38) 0.62 (0.32) 0.53 (0.25)

0.65 (0.40) 0.56 (0.36) 0.47 (0.28)

0.45 (0.41) 0.39 (0.35) 0.30 (0.23)

0.61 (0.29) 0.51 (0.42) 0.43 (0.36)

0.63 (0.27) 0.51 (0.41) 0.42 (0.37)

0.34 (0.27) 0.21 (0.28) 0.15 (0.20)

0.67 (0.40) 0.48 (0.30) 0.57 (0.32)

0.50 (0.45) 0.40 (0.31) 0.44 (0.35)

0.32 (0.41) 0.25 (0.25) 0.22 (0.21)

IS Israeli born; FSU born in former Soviet Union; ETH born in Ethiopia

played in explaining student achievement in mathematics. We therefore used items identical in their contents and goals but different in their presentation modalities (such as verbal, visual, and symbolic). This manipulation reflects the following assumption: although these different modalities are considered similar in terms of problem understanding, they actually are complementary as each one reflects a different dimension of mathematics literacy, modes of representation of a knowledge domain and an individual’s comprehension of content domain. Such a manipulation can be considered a reflection of code switching, requiring flexibility in transition between modalities within mathematics. Table 7.4 demonstrates the similarities in the pattern of mathematics achievement as a function of the modes of knowledge representation among the three groups of students (Israeli born, FSU, and ETH). It also shows the differences in the level of mathematics achievement among the three groups. The data demonstrates that achievement on questions that use a symbolic mode or mathematics models are higher than on questions requiring the use of visual and verbal modes. Furthermore, this pattern of difficulty with the mode of math presentation in mathematics tasks is similar for all the three group levels, immigrants as well as natives. These results indicate that it is not only the content-specific language that affects student achievements but also contentspecific literacy.

7.5 Discussion This study provides a new level of knowledge concerning academic language in multilingual/multicultural classrooms. The major contribution of this aspect of the study is that the findings highlight the significant role that different dimensions of academic language have in affecting students’ achievement, both for immigrant as well as for native-born students. In particular the chapter extends the language dimensions that characterize learning and assessment of mathematics. Some dimensions are reflected in the form of the mathematics questions, and some by the interaction between form and content.

152

T. Levin and E. Shohamy

This expanded construct of academic language presented in this study is not generic but rather depends on the content areas studied, whereby the language of the subject area provides a very crucial element of success. The study thus magnifies the importance of exploring immigrant students’ achievement by using specific measurements that include different dimensions of their academic language within different content domains. Not only can such achievement measures assist in diagnosing students’ strength and weaknesses, they also can direct curriculum planners, teachers, and evaluators to ways of structuring more effective educational interventions or learning and assessment activities. They also can guide evaluators on how to conceptualize tests or other assessment means that fit various immigrant groups. The study therefore strengthens the power of the language component in domain-specific contexts, which has a strong presence. It also shows that without the adequate language competences a student, immigrant and nonimmigrant as well, cannot really follow classroom practices and is deficient in both learning and communicating what has been learned. Although some researchers suggest that learning a content domain may assist immigrant students in acquiring the dominant language, this study does not necessarily support such a view. In contrast, we emphasize the value of considering the differences between the natural-spoken language used in a content domain and the domain-specific language, and mainly the interaction between the two, noting that each has its own role in the learning of mathematics. Since language problems associated with learning mathematics become more conspicuous with immigrant students who have had no Western-based experiences in mathematics and science learning, the study supports the effects of the socio-cultural dimension of the language of mathematics and science. However, this should not imply that immigrant students who are well prepared in terms of mathematics and science skills do not face difficulties: The latter are due to the different dimensions of the academic language of mathematics and science. The results of the study concur with Halliday’s views (1989) concerning the place of lexicon within comprehension, but we further suggest that it is actually literacy rather than language per se that plays the major role, since it includes not only the vocabulary and the symbols that characterize a content domain but also the norms, values, and conventions characteristic of the discipline. It is mainly the interaction among all of these components that aids the conceptualizing of academic language, and the conceptualization enables meaning construction. It is also mainly the academic language with its specific distinctions according to the content domain it represent. Concluding that academic languages reflect the values, beliefs, and assumptions of each specific subject area and the culture from which it comes, we believe that further research is needed before coherent strategies can be developed for helping students in general and particularly immigrant students. As long as we view academic language not only as incorporating a wider social and culture perspective, but also as a product of the subject matter learned and the specific learning context or the communities of practice in which students were

7

Understanding Language Achievement of Immigrants in Schools . . .

153

engaged, both in the past and the present, the exact nature of their difficulties in terms of academic language needs to be further understood. The results of this study have contributed to the beginning of a long venture into a deeper understanding of what constitutes a domain-specific academic language, a topic that needs to be expanded and researched much more extensively. It is specifically in the current era where students need to be able work and function across multiple tools, multiple forms of text and multiple languages, (Sturtevant et al., 2007), it is of further importance to remember that it is the domain specific competencies that would serve as the key sources to ‘‘bridge’’ students’ rich, out-of-school literacy practices (Kajder, 2006; Kress, 2006; Macken-Horarik, 2009; New London Group, 1996), with the specific academic skills or domain-specific literacies. As our findings are seem to suggest, multilingual and multicultural classrooms entail not only diverse languages that the students need to master in each of the specific content domains but also diverse types and forms of literacy, i.e., multi-literacies, entailing also socio-cultural aspect of multi-literacies, which enable students to become fluent and proficient in the specific practices of each content domain.

References Aikenhead, G. S. (1997). Toward a first nations cross-cultural science and technology curriculum. Science Education, 81, 217–238. Arellano, E. L., Barcenal, T. L., Bilbao, P. P., Castellano, M. A., Nichols, S., & Tippins, D. J. (2001). Case based pedagogy as a context for collaborative inquiry in the Philippines. Journal of Research in Science Teaching, 38(5), 502–528. Bachman, L. F. (1990). Fundamental considerations in language testing. New York: Oxford University Press. Brenner, M. E. (1998). Development of mathematical communication in problem solving groups by language minority students. Bilingual Research Journal, 22(2, 3, and 4), 149–174. Canale, M., & Swain, M. (1980). Theoretical bases of communicative approaches to second language teaching and testing. Applied Linguistics, 1, 1–47. Chamot, A. U., & O’Malley, J. M. (1986). A cognitive academic language learning approach: An ESL content-based curriculum. Wheaton, MD: National Clearinghouse for Bilingual Education. Coelho, E. (1982). Language across the curriculum. TESL Talk, 13(3). Cummins, J. (1981). The role of primary language development in promoting educational success for language minority students. In J. Cummins (Ed.) Schooling and language minority students: A theoretical framework (pp. 3–50). Los Angeles: California State University. Cummins, J. (1984). Bilingualism and special education: Issues in assessment and pedagogy. San Diego, CA: College-Hill Press. Cummins, J. (1989). Empowering minority students. Sacramento, CA: Association for Bilingual Education. Cummins, J. (1991). Language development and academic learning. In L. Makave & G. Duquette (Eds.), Language culture and cognition. Clevedon: Multilingual Matters.

154

T. Levin and E. Shohamy

Cummins, J. (2000). Language, power and pedagogy: Bilingual children in the crossfire. In J. Cummins (Ed.), Language, power, and pedagogy: Bilingual children in the crossfire. Clevedon: Multilingual Matters. Cummins, J. (2001). Assessment and intervention with culturally and linguistically diverse learners. In S. R. Hurley & J. V. Tinajero (Eds.), Literacy assessment of second language learners (pp. 126–145). Boston: Allyn and Bacon. Cummins, J. (2002). Beyond instructional techniques and standardized assessment: Implementing classroom interactions that foster power, identity, imagination and intellect among culturally diverse students. Contact, 28(2), 21–31. Ellerton, N. F., & Clarkson, P. C. (1996). Language factors in mathematics teaching. In A. J. Bishop (Ed.), International handbook of mathematics education. Netherlands: Kluwer. Fan, C., & Karnilowicz, W. (2000). The mathematics achievement and aspirations of ChineseAustralian girls and Anglo-Australian girls in Australia. Current Research in Social Psychology, 5, 170–190. Giroux, H. (2005). Border crossings: Cultural workers and the politics of education New York: Routledge. Guerrero, M. D. (1997). Spanish academic language proficiency: The case of bilingual education teachers in the U.S. Bilingual Research Journal, 21(1), 65–84. Halliday, M. A. K. (1978). Language and social semiotic: The social interpretation of language and meaning. London: Arnold. Halliday, M. A. K. (1989). Some grammatical problems in scientific English. Paper presented to the Society of Pakistani English Language Teachers, Karachi, Pakistan. Hamayan, E., & Perlman, R. (1990). Helping language minority students after they exit from bilingual/ESL programs. Washington, DC: National Clearinghouse for Bilingual Education. Ilany, B., & Margolin, B. (2010). Language and mathematics: Bridging between natural language and mathematical language in solving problems in mathematics, Creative Education, 1(3), 138–148. Johns, A. (1998). Languages for specific purposes: Pedagogy. In J. B. Mey (Ed.) Concise Encyclopedia of pragmatics (pp. 474–480). Oxford: Elsevier Science. Kajder, S. (2006). Bringing the outside in: Visual ways to engage struggling readers. Portland, ME: Stenhouse. Kress, G. (2006). A continuing project: English for a period of radical instability only connect: English teaching, schooling and community (pp. 27–30). The Australian association for the Teaching of English, Kent Town, South Australia. Levin, T., & Shohamy, E. (2007). The role of academic language in understanding the mathematics achievements of immigrant students in Israel. In C. S. Sunal & K. Mutua (Eds.), The enterprise of education: Research on education in Africa, the Caribbean, and the Middle East (pp. 313–336). Tuscaloosa, AL: Information Age Publishing. Levin, T., & Shohamy, E. (2008). Achievement of immigrant students in mathematics and academic Hebrew in Israeli school: A large scale evaluation study. Studies in Educational Evaluation, 34, 1–14. Lindholm-Leary, K. (2001). Dual language education. Clevedon: Multilingual Matters. Macken-Horarik, M. (2009). Multiliteracies, Metalanguage and the Protean Mind: Navigating school English in a sea of change. English in Australia, 44(1), 33–43. New London Group. (1996). A pedagogy of multiliteracies: Designing social futures. Harvard Educational Review, 66, 60–92. O’Malley, J. M., & Valdez Pierce, L. (1996). Authentic assessment for English language learners: Practical approaches for teachers. Reading, MA: Addison-Wesley Publishing. Portes, A., & Hao, L. (1998). E pluribus unum: Bilingualism and loss of language in the second generation. Sociology of Education, 71, 269–294.

7

Understanding Language Achievement of Immigrants in Schools . . .

155

Renninger, K. A. (1992). Individual interest and development: Implications for theory and practice. In K. A. Renninger, S. Hidi, & A. Krapp (Eds.), The role of interest in learning and development (pp. 361–395). Hillsdale, NJ: Erlbaum. Rumbaut, R. G. (1996). The new Californians: Assessing the educational progress of children of immigrants. CPS Brief, 8(3), 1–14. Rumbaut, R. G. (1999). Assimilation and its discontents: Ironies and paradoxes. In J. DeWind, C. Hirschman, & P. Kasnitz (Eds.), The handbook of international migration: The American experience (pp. 172–195). New York: Russell Sage Foundation. Rumberger, R. W., & Larson, K. A. (1998). Toward explaining differences in educational achievement among Mexican American language-minority students. Sociology of Education, 71(1), 69–93. Sanchez, S., & Thorp, E. (1998). Policies on linguistic continuity: A family’s right, a practitioner’s choice, or an opportunity to create shared meaning and a more equitable relationship? Zero to Three, 18(6), 12–20. Scarcella, R. (2003). Academic English: A conceptual framework (Technical Rep. 2003-1). Santa Barbara, CA: University of California Linguistic Minority Research Institute. Schleppegrell, M. J. (2004). The language of schooling: A functional linguistics perspective. Mahwah, NJ: Lawrence Erlbaum Associates. Schleppegrell, M., & Colombi, M. C. (Eds.). (2002). Developing advanced literacy in first and second languages. Mahwah, NJ: Erlbaum. Short, D. J., Mahrer, C. A., Elfin, A. M., Liten-Tejada, R. A., & Montone, C. L. (1994). Protest and the American Revolution. Washington, DC: National Center for Research on Cultural Diversity and Second Language Learning & the Center for Applied Linguistics. Solomon, J., & Rhodes, N. C. (1995). Conceptualizing academic language (OERI Publication No. R117G10022). Washington, DC: U.S. Department of Education. Spanos, G., Rhodes, N., Dale, T., & Crandall, J. (1988). Linguistic features of mathematical problem solving: Insights and applications. In R. R. Cocking & J. P. Mestre (Eds.), Linguistic and cultural influences on mathematics learning (pp. 221–290). Hillsdale, NJ: Lawrence Erlbaum. Stanton-Salazar, R. D., & Dornbusch, S. M. (1996). Social capital and the reproduction of inequality: Information networks among Mexican-origin high school students. Sociology of Education, 68, 116–135. Sturtevant, E., Boyd, F., Brozo, W., Hinchman, K., Moore, D., & Alvermann, D. (2007). Principled practices for adolescent literacy: A framework for instruction and policy. Mahwah, NJ: Erlbaum. Thomas, W. P., & Collier, V. P. (2001). A national study of school effectiveness for language minority students’ long-term academic achievement. Final report (Project 1.1). Santa Cruz, CA: Center for Research on Education, Diversity, and Excellence. Valdez, G. (2004). Between support and marginalisation: The development of academic language in linguistic minority children. Bilingual Education and Bilingualism, 7(2 and 3), 102–132. Valdez, G., Pierce, L., & O’Malley, J. M. (1992). Performance and portfolio assessment for language minority students. Washington, DC: National Clearinghouse for Bilingual Education. Vermeulen, H., & Perlmann, J. (2000). Immigrants, schooling and social mobility. Does culture make a difference? London: Macmillan. Vernez, G., & Abrahamse, A. (1996). How immigrants fare in U.S. education. Santa Monica, CA: RAND Center for Research on Immigration Policy. Vygotsky, L. (1986). Thought and language. Cambridge, MA: MIT Press.

Part III

Multilingual Acquisition and Processing

Chapter 8

Adjective Inflection in Hebrew: A Psychollinguistic Study of Speakers of Russian, English and Arabic Compared with Native Hebrew Speakers Iris Alfi-Shabtay and Dorit Ravid

8.1 Introduction World languages share different morphological typologies (Bybee, 1985). While Hebrew, Russian and Arabic have a rich morphology, in which words typically contain more than one morpheme (traditionally called synthetic languages), languages such as English, Chinese contain a restricted or ‘sparse’ morphology (traditionally called analytic languages). The typological differences between Hebrew, Arabic, Russian and English make them ideally suited for this type of comparative research. This study aims to evaluate the way proficient Hebrew learners who are native speakers of languages of distinct morphological typologies perceive the essence of adjective inflection in Hebrew, taking into account syntactic construction, source language typology and the nature of the categories. The examination of non-native speakers as compared to native speakers is rewarding to the language researcher, since each population provides insights of a different kind on L1 and L2 knowledge. It also demonstrates the role that first language plays on success in the second language. Hebrew is a Semitic language, rich in both inflectional and derivational morphology. Inflection is widespread in Hebrew. Nouns are inflected for gender, e.g., masculine shulxan ‘table’, feminine mita ‘bed’; and also for number, e.g., masculine singular kadur, plural kadurim ‘ball/ ball-s’; singular feminine buba, plural bubot ‘doll/doll-s’. There is also a marginally productive dual form with the suffix –ayim e.g., she’atayim ‘two hours’. Verbs inflect for tense (present, past or future), and for mood, taking special forms in imperative and infinitive. Pronouns take a free form only as surface subject, and in all other contexts – are suffixed to prepositions, and inflect for number, gender and person (Berman, 1978, 1981, 1985; Ravid, 1995; Ravid & Malenky, 2001). Hebrew adjectives are inflected in agreement with nouns, and follow the agreement-assigning noun. Adjectives within the NP agree with the agreementassigning nouns in gender, number and definiteness, e.g., sus gadol ‘horse I. Alfi-Shabtay (*) Tel Aviv University, Tel Aviv, Israel e-mail: [email protected] M. Leikin et al. (eds.), Current Issues in Bilingualism, Literacy Studies 5, DOI 10.1007/978-94-007-2327-6_8, Ó Springer ScienceþBusiness Media B.V. 2012

159

160

I. Alfi-Shabtay and D. Ravid

big’/ha-sus ha-gadol ‘the-horse the-big’/susa gdola ‘mare big,Fm’/ha-susa hagdola ‘the-mare the-big,Fm’. Adjectives in predicative construction agree with their subject only in number and gender, e.g., ha-sus hu gadol ‘the-horse is big’/ ha-susa hi gdola ‘the-mare,Fm is big,Fm’ (Berman, 1978, 1981). Mastery of adjective inflection requires wide knowledge of Hebrew (Berman, 1981; Ravid, 1995). Forming a grammatical construction demands that speakers abstract out gender and number marking, where these inflectional marking are conflated within a single morpheme, -im for masculine and plural, -ot for feminine and plural. It also requires the speaker to control the common phenomena of ‘suffix harmony’ in the language, e.g., yeled gadol ‘a big boy’ vs. yald-a gdol-a ‘a big girl’ vs. yelad-im gdol-im ‘big children’ (Berman, 1985). This complex adjectival inflection system also involves many idiosyncratic forms that require rote learning. For example, the singular feminine noun beyca is pluralized as beycim ‘eggs’ which ends with the suffix –im, marked for plural masculine, may lead to a misinterpretation of the form as a marked for masculine gender in adjective inflection, such as *beycim kashim ‘eggs’ [Fm] ‘boiled’, when the feminine suffix –ot ‘beycim kashot’ is required (Berman, 1981; Ravid, 1995). Despite this apparent complexity, initial acquisition of the inflection system takes place in the speech of normally developing Israeli children along with the bulk of other morpho-syntactic development during the third and the early part of the fourth year of life (Berman, 1983, 1985; Dromi, 1987; Dromi & Berman, 1986; Kaplan, 1983; Levy, 1983; Ravid, 1997), with verb inflection being controlled earlier than adjective inflection, and number before gender inflection (Kaplan, 1983; Levy, 1983). Further difficulty is encountered with the inflection of nouns which are masculine in form, lacking, for instance, typical feminine ending such as stressed –a or unstressed –et, but which happen to be feminine in gender assignment. This confusion leads children to say * esh xazak ‘fire strong = strong fire’ when esh xazak-a is required; or * even gadol ‘stone big’ = ‘a big stone’ in place of even gdol-a. Similarly, difficulties occur with singular masculine nouns with a feminine -a or –t ending, e.g., yalkut ‘schoolbag’, resulting in formations such as *yalkut gdol-a ‘schoolbag big = a big schoolbag’, when yalkut gadol is required. An even more complicated type of agreement conflict is incurred by irregular nouns which take opposite plural markers. As a way of resolving the complex grammatical requirement of going from singular masculine noun with feminine plural ending to a masculine plural adjective, or the other way around for singular feminine nouns that are irregular in plural formation, children adopt the following strategy: if the plural noun ends in –im, so does the adjective, if the plural noun ends in –ot, so does the adjective, even though in fact agreement is dictated by the singular gender of the noun. Such errors are corrected, once the specific nouns are learned as isolated lexical items, on the basis of a wider exposure to more examples, and possibly also familiarity with very common, almost formulaic noun-adjective combinations, such as ir gdol-a ‘a big city’. Complete mastery of all allomorphic

8

Adjective Inflection in Hebrew: A Psychollinguistic Study of Speakers of . . .

161

morpho-phonological forms and grammatical agreement systems thus takes years, often until adolescence (Berman, 1985; Levy, 1983). Hence, agreement phenomena, which are widespread in Hebrew, may provide an important source of insights into how proficient native and non-native adult Hebrew speakers tackle a system of marking which is formally quite complex, and which is semantically largely unmotivated, or at best redundant. The current study focuses exclusively on adjective inflection, and aims to pinpoint those precise categories which place a cognitive burden on advanced Hebrew learners, examining the way Russian, Arabic and English speakers judge and process adjective-inflection sentences in two syntactic constructions as compared to Hebrew speakers.

8.1.1 The Category of Adjectives Adjectives constitute a less primary class of content words than nouns and verbs. They are more marked and less dense in meaning, and represent what might be construed as ‘non-essential’ information. Adjectives even comprise an absent category in many languages (Dixon, 1977; Schachter, 1985). The less primary status of adjectives is expressed in several senses. Semantically, adjectives denote attributes or properties of nouns, that is, they serve a secondary function in relation to a primary class (Keenan, 1976). Syntactically, adjectives fulfill two functions in relation to nouns: predicative adjectives have the function of predicate heads (e.g., Mary is smart); attributive adjectives have the function of NP modifiers (e.g., the smart student). In both cases, the adjective denotes a property attributed to a noun – either the subject of the sentence or the NP head (Ferris, 1993). Psycholinguistically, adjectives appear later in child speech (Casseli, Bates, Casadio, & Fenson, 1995; Rice, 1990; Sommers, Kozarevich, & Michaels, 1994), and constitute a low-frequency class when compared to other content words in children’s early lexicons in various languages (Dromi, 1987; Marvin, Beukelman, & Bilyeu, 1994; Ravid & Nir, 2000; Valian, 1986). The distribution of adjectives has proved diagnostic in studies of later language development during the school years (Levie, 2002; Ravid, Levie, & Avivi-Ben Zvi, 2003); and these findings are confirmed by analyses of extended texts (Ravid, 2004a). Adjectives show an increase in Hebrew in more structural classes including denominals, resultatives (passives), and diminutives, and those denoting cognitive and affective states (Ravid & Zilberbuch, 2003a). The interface of adjectives with nouns in the two canonical syntactic constructions – predicative and attributive – has implications for syntactic complexity and for the information structure of discourse (Englebretson, 1997). There is evidence that this distinction is learned by the age of 2;6 (Nelson, 1976). However, interpreting adjective-noun phrases requires a two-step process of first identifying the noun category, then subdividing it based on the adjective. Attributive adjectives thus emerge later than predicative adjectives. Ninio

162

I. Alfi-Shabtay and D. Ravid

(2004), for example, showed that when Hebrew-learning children aged 18 to 52 months interpreted adjective-noun combinations by choosing a picture depicting the correct referent, they tended to ignore the adjective and based responses on the noun. Diesendruck, Hall, and Graham’s (2006) study of English- and Hebrew-speaking preschoolers provides further evidence that they understood that adjectives in the attributive construction label values on dimensions along which objects of a particular kind may vary, whereas predicate adjectives did not appear to carry the same semantic implication. Nevertheless, the proportion of predicative adjectives decreases, while attributive adjectives rise with age and schooling (Ninio, 2004; Ravid & Nir, 2000). This was apparent in Ravid and Levie’s (2006) study of the distribution and use of adjectives in 252 narrative and expository texts produced by 63 Hebrewspeaking children, adolescents and adults. They showed that noun modification is the major function of adjectives in older and more literate Hebrew speaker/ writers. Attributive adjectives participate in greater quantity in the rich nominal and informative structure of written and expository texts. In general, noun phrases with abstract nominal heads tend to attract more morphologically and semantically complex attributive adjectives (Ravid & Cahana-Amitay, 2005), so that semantic and syntactic complexity work in tandem. Thus, the category of adjective can be taken to indicate the emergence of a richer and more complex lexicon compared with nouns and verbs. A combination of lexical, morphological, and syntactic analyses reveal that with age and schooling, adjectives become lexically more diverse, morphologically more complex, and semantically richer, and that they also participate in increasingly more complex syntactic constructions (Ravid & Levie, 2006). Adjectives also provide a window on L2 abilities. Polinsky (2004) showed that less proficient adult learners of L2 Russian do most poorly on comprehension and translation of adjectives compared with nouns and verbs, leading her to suggest that ‘the knowledge of adjectives is more or less a luxury’ given that adjectives are ‘rhetorical devices’ whose comprehension is not necessary for identifying the referent of a given noun phrase. Alfi-Shabtay’s (2006) study of Hebrew knowledge of proficient Russian speakers showed a relatively low production of adjectives in learners’ narrative texts compared with nouns and verbs. The study also demonstrated how longer exposure to Hebrew affected the production of adjective. Long-term resident learners who had over 20 years of Hebrew exposure used more adjectives in their texts than learners with shorter residence. They were also less dependent on basic adjectives with a simple CVC structure such as ‘big’, ‘good’, and were capable of using more advanced constructions such as resultative (passive) adjectives which denote a state by an adjective (Berman, 1979), as well as denominal i-suffixed adjectives, typical of higher-register, written Hebrew (Ravid & Zilberbuch, 2003b). The current study, too, assesses performance on adjectives in L2 Hebrew learners. Adjective inflection has been selected as a central research category

8

Adjective Inflection in Hebrew: A Psychollinguistic Study of Speakers of . . .

163

since it requires wide knowledge in Hebrew. It constitutes one of the most complex morpho-syntactic tasks for the Hebrew-speaking child (Berman, 1985; Kaplan, 1983; Levy, 1983), and one which is as crucial as a developmental criterion for Hebrew as is the acquisition of auxiliary pattering in English (Ingram, 1989; Owens, 2001). Specifically, adjective inflection is considered difficult for L2 Hebrew learners even after gaining fluency in the language (Berman, 1997). Most studies concerning morpho-syntax in the second language have found different performance even in the highly advanced learners compared to those of native speakers. These studies have largely dealt with age effects via grammaticality judgments, and generally concluded that what differentiates a younger learner from an older learner of a language is accuracy of their grammatical knowledge: the younger the learner’s age of exposure – the better chances s/he has to perform like a native speaker (e.g., Alfi-Shabtay, 2006; Coppieters, 1987; DeKeyser, 2000; DeKeyser, Alfi-Shabtay, Ravid, & Shi, 2005; DeKeyser, Alfi-Shabtay, & Ravid, 2010; Johnson & Newport, 1989, 1991; Johnson, 1992; Patkowski, 1994; Schachter, 1990; Sorace, 1993). To eliminate age effects, the current study is exclusively focusing on proficient adult learners who were first exposed to Hebrew during adolescence. This should provide an accurate picture of the performance on adjective inflection in advanced learners.

8.1.2 Hebrew Adjectives Modern Hebrew has four major structural classes of adjectives. The first is an essentially closed class of primary CVC adjectives originating in Biblical present-tense participles (e.g., xam ‘hot’), are morphologically simple words, since despite their verbal origin, they are monomorphemic as well as monosyllabic, having lexicalized into a single unit. These primary CVC adjectives typically designate basic semantic relations such as good, bad, hot, cold and so are very early acquisitions (Ravid & Nir, 2000). A second class contains a variety of nonlinear root + pattern structures (e.g., mahir ‘quick’ from the root m-h-r, in the agentive noun pattern CaCiC).1 Except for certain color terms, which are inherently adjectival (e.g., kaxol ‘blue’, sagol ‘purple’), almost all of these nonlinear root + pattern structure adjectives are appropriated from either verbal or nominal patterns, For example, mahir ‘quick’, and axil ‘edible’ are in the agentive noun pattern CaCiC (cf. pakid ‘clerk’); while mafxid ‘scary’ and mevusˇal ‘cooked, not raw’ use participial maCCiC and meCuCaC verbal patterns. The semantic content of these verbal/nominal adjectives varies across structural categories, covering a range of general and specific properties, 1

Semitic morphological patterns are presented as templates with slots for consonantal root radicals.

164

I. Alfi-Shabtay and D. Ravid

attributes, and states. Time of acquisition also varies by semantic content; for example, even though color terms are structurally complex, they are acquired and conjugated correctly early on because of their high frequency in childdirected speech (Ravid, 1995). Resultative adjectives, in contrast, emerge and consolidate later, between age 4 and 6 (Berman, 1994; Yagev, 2000). A third structural device is reduplication, which is used mostly for adjective (and noun) diminutivization, e.g., vradrad ‘light pink’ from varod ‘pink’. This is a late-emerging device in Hebrew (Hora, Avivi-Ben Zvi, Levie, & Ravid, 2006). The fourth and most productive class of adjectives in Modern Hebrew is a late historical development deriving from Biblical ethnic nouns which in Medieval Hebrew evolved into a full-fledged class of denominal adjectives, e.g., xasˇmal-i ‘electr-ic’, cibur-i2 ‘publ-ic’. Structurally, denominal adjectives are simpler than the root-and-pattern class, since they involve linear formation of a nominal stem and the addition of the adjectival suffix –i. However, they are typical of higher-register, written Hebrew like literary prose or journalistic and expository texts, and their meaning is quite complex (Ravid, 2004b). About one third of the adjective types occurring in newspaper Hebrew is denominal (Ravid & Shlesinger, 1987). Apart from lexicalized forms such as xagigi ‘festive’ and the original Biblical ethnic-attributive meaning (e.g., dati ‘religious’, rusi ‘Russian’), denominal adjectives are completely absent from child-directed speech. Denominal i-suffixed adjectives are the latest to emerge in Hebrew child language, from around age 6, and they surface in text production only from high school up (Levin, Ravid, & Rappaport, 2001; Ravid & Zilberbuch, 2003b). The current study is confined to the first and the second basic structural classes of adjectives: primary CVC adjectives and the nonlinear root + pattern adjectives.

8.1.3 Adjective Inflection in Russian, Arabic and English Russian adjectives are marked for three genders (masculine, feminine, neuter), two numbers (singular, plural), and six primary cases (nominative, accusative, genitive, dative, prepositional, instrumental). They precede the nouns they modify, and agree with them in number, gender and case. While an inanimate singular masculine noun and its attribute adjective are inflected with the same marker in both accusative and nominative cases, e.g., balsho’y avto’bus ‘bus,Sg, Masc big,Sg,Masc = a big bus’, an animate singular masculine noun and its adjective are inflected with different markers in the nominative, e.g., balsho’y avto’bus, and in the accusative cases e.g., balsho’vo ma’lchika ‘a big boy’. In plural, the endings of the plural adjective attributed to noun declensions are syncretized, i.e., are identical for all genders, e.g., balshi’ye ma’lchiki ‘big boys’;

2

We use c to denote the unvoiced coronal affricate.

8

Adjective Inflection in Hebrew: A Psychollinguistic Study of Speakers of . . .

165

balshi’ye devo’chki ‘big girls’; balshi’ye pla’tya ‘big dresses’ (Maltzoff, 1987; Timberlake, 1993). Hebrew and Arabic articles are prefixed not only to the head noun, but also to its various non-sentential modifiers, such as demonstratives and adjectives, e.g., yeled gadol ‘a big boy’; ha-yeled ha-gadol ‘the big boy’ (Berman, 1978). However, while articles constitute part of the adjectival inflection in Hebrew and Arabic, Russian has no articles (Chesterman, 1991; Lyons, 1999; Wexler, 1976), and English has articles, however the definite article is not part of the adjective-inflection system (Conrad, Biber, & Leech, 2002). The adjective inflection system in Russian is often divided into three classes. These include: (1) tautological pairs, in which the inflection of the noun and adjective is the same, e.g., Sg,Fm Instrumental bol’shOJ nogOJ ‘big foot’; (2) reduplicative pairs, in which the inflection of the adjective presents the reduplicated inflection of the noun. This inflection type is usually triphonemic with two similar vowels divided by [m, v, j, h] consonant e.g., Feminine Accusative be’lUJU koftU ‘white shirt’; and (3) contrastive inflection, in which the inflections of the nouns and adjectives have nothing in common, e.g., feminine Genitive: belAJ koftY ‘white shirt’ (Slobin, 1966; Voeykova, 1997). Modern Arabic also contains a rich inflectional system. Attributive adjectives generally follow the agreement-assigning nouns, and agree with them in number, gender and definiteness. Grammatically definite nouns require their attributive adjectives to be overtly marked as definite by the prefixing of al ‘the’. Singular nouns are either masculine or feminine and require masculine or feminine adjectives respectively, whether the noun is part of a noun-adjective phrase or a predicate. For example, ibn kabir/al-ibn al-kabir ‘son big = a big son’/‘the-son the-big’ = ‘the big son’; ibn-a kabir-a/al-ibn-a al-kabir-a ‘daughter big = a big daughter’/‘the-big the-daughter = the big daughter’. Singular feminine nouns often end with –a, e.g., madras-a ‘school’, and there are a few singular nouns which are feminine by form (ending –a), but have exclusively male reference, such as xalifa ‘caliph’ which requires masculine agreement, and a few which are masculine by form but female in reference, e.g., hamil ‘pregnant woman’ which require feminine agreement following natural or biological gender. As far as plural nouns are concerned, the prime factor which determines agreement is whether the noun’s reference is human or non-human. If the modified noun is human, the adjective is marked for both gender and number (‘strict’ or natural-biological gender agreement), e.g., awlad kibar ‘boys big, Masc,Pl = big boys’; banat kabirat ‘girls big,Fm,Pl = big girls’. However, if the noun is non-human, its plural form, regardless of the gender of its singular, requires a feminine singular adjective (‘deflected’ agreement), e.g., kutub qadima ‘books,Masc,Pl old,Fm,Sg’ = ‘old books’; madaris qadima ‘schools,Fm,Pl old, Fm,Sg’ = ‘old schools’. Nevertheless, variations among speakers in adjective agreement with non-human nouns are apparent in many Arabic dialects (Holes, 1995). It has also been accounted within the dialect spoken in Israel (Ravid, Abu-Nofel, & Huri, 1998). English, in contrast to Hebrew, Russian and Arabic, has a restricted inflectional system. It has almost no marking of agreement for gender, number, or

166

I. Alfi-Shabtay and D. Ravid

person, and specifically expresses no agreement between adjectives and the nouns they modify. Attributive adjectives in English precede the nouns they modify. Singular adjective is used in singular, e.g., a beautiful boy/a beautiful girl, and in plural beautiful boys/beautiful girls (Conrad et al., 2002; Quirk, Greenbaum, Leech, & Svartvik, 1985).

8.1.4 Adjective Inflection Categories and Items Based on the unique nature of Hebrew inflection, the current study assessed performance of eight major adjective inflection categories in attributive and in predicative constructions. The categories are inflected in agreement with regular and irregular noun suffixes. Following below is a description and examples of the eight categories of the grammar test (32 item pairs altogether); incorrect items are marked with an asterisk, abbreviations used in the sequel are in brackets. (i) Adjective inflection with regular noun suffixes: (a) Singular masculine nouns (SM), e.g., shaon shaxor/*shxora ‘black, Masc/*Fm watch,Masc’ [4 items] (b) Singular feminine nouns (SF), e.g., rakevet mehira/*mahir ‘speedy,Fm/ *Masc train,Fm’ [4 items] (c) Plural masculine nouns (PM), e.g., sdinim xadashim/*xadashot ‘new,Pl, Masc/*Pl,Fm sheets,Masc’ [4 items] (d) Plural feminine nouns (PF), e.g., megerot mesudarot/*mesudarim ‘tidy, Pl,Fm/*Pl,Masc drawers,Fm’ [4 items] (ii) Adjective inflection with irregular noun suffixes: (a) Mass nouns (MSS), e.g., shamayim behirim/*behirot/‘clear,Pl,Masc/ *Pl,Fm sky,Pl,Masc’ [4 items] (b) Irregular singular masculine nouns (XM), e.g., mivca mishtalem/*mishtalemet ‘good,Masc /*Fm buy,Masc’ [4 items] (c) Irregular singular feminine nouns (XF), e.g., kos nekiya/*naki ‘clean, Fm/*Masc glass,Fm’ [4 items] (d) Plural feminine nouns, irregular suffix (IRG), e.g., avanim xadot/ *xadim ‘sharp,Pl,Fm /*Pl,Masc stones,Pl,Fm’ [4 items]

8.2 Method The current study examined performance on Hebrew adjective inflection in adult speakers of Russian, English, and Arabic, compared with a group of native Hebrew speakers.

8

Adjective Inflection in Hebrew: A Psychollinguistic Study of Speakers of . . .

167

8.2.1 Population The participants in this study were 64 adult participants from the center of Israel. They were all active college students, studying towards a teaching certificate in a Hebrew-speaking institute. They included: 16 Russian speakers, 16 English speakers, 16 Arabic speakers, and 16 Hebrew speakers. They were between 19 and 28 years old at testing. Except for the Hebrew native speakers, all groups had Hebrew as a second language; they were first exposed to Hebrew during adolescence, between ages 10 and 14, and have been exposed to Hebrew for at least 10 years (with an average mean of 16 years), to make sure ultimate attainment levels have been reached (DeKeyser, 2000; Johnson & Newport, 1989).

8.2.2 Instruments Grammaticality judgment test: All participants took a grammaticality judgment test in Hebrew developed by the authors specifically for the purpose of this study. The test consisted of 104 items representing eight major adjectiveinflection categories in Hebrew (see Section 8.1.4). Each category contained three sentence pairs (grammatical and ungrammatical), and was given in two syntactic constructions (attributive and predicative). Each sentence was contextualized with 6–8 frequent everyday words. Biographical questionnaire: All participants filled out a short biographical questionnaire which provided information concerning their native language, age of arrival and age at testing.

8.2.3 Procedures Participants were recruited as part of course requirements. They were tested individually, in a quiet room. They filled out the background questionnaire first, and then took the grammar test. The items on the grammar test were presented auditorily by playing a digitized recording of all sentences, each read twice in a row, with a 3-second interval between the two readings and a 6-second interval between sentence pairs. The sentences were recorded by the first author, a linguist and proficient speaker of Hebrew, in a fixed random order, with one of the two sentences of each grammatical/ungrammatical pair in the first 48 sentences, and the others in the second 48 sentences. No two sentences of the same category or the same syntactic construction ever occurred consecutively. After every eighth sentence, a transplant sentence in Hebrew with no inflection form was entered (altogether there were 8 transplant sentences). This was done to avoid bias to adjective constructions. Participants indicated whether they considered a sentence to be correct or incorrect in Hebrew in an answer sheet. Four practice sentences were given before the start to ensure the testing procedure is understood by all participants.

168

I. Alfi-Shabtay and D. Ravid

At any time participants were able to ask the experimenter to hear a sentence again and/or stop the tape in case they were tired. They were instructed to carry on even if they were not sure about the answer. The entire test took about 25 min; there was a 5-min break halfway.

8.3 Results In preparation for statistical analysis, we classified four independent variables. These were: (1) syntactic construction (i.e., attributive and predicative); (2) group (i.e., English, Arabic, Russian, Hebrew speakers); (3) category (i.e., singular masculine nouns; singular feminine nouns; plural masculine; plural feminine; mass nouns; Irregular singular masculine nouns; Irregular singular feminine nouns; plural feminine nouns); and (4) sentence grammaticality (i.e., grammatical and ungrammatical) on the grammar test. Two major statistical analyses were used in this study. These were: the posthoc Scheffe’ analysis which investigated the sources of differences between the groups at the 0.05 level; and the post-hoc Bonferroni analysis which investigated the sources of each interaction at the 0.05 level. Following is a description of the analyses based on the independent variables.

8.3.1 Analysis by Syntactic Construction and Group We first examined performance on the grammaticality judgment test by syntactic construction and group. We expected attributive adjectives to score lower than predicative adjectives (Ninio, 2004; Ravid & Nir, 2000; Ravid & Levie, 2006); and the English speakers whose L1 does not mark inflection for adjectives (Conrad et al., 2002; Quirk et al., 1985) to perform worst compared to the other groups. Table 8.1 presents error scores in attributive and in predicative adjectives in the four groups. A two-way ANOVA with repeated measures (group (4) X syntactic construction (2)) was conducted on the data on Table 8.1. An effect for syntactic construction emerged (F(1,60) = 19.79, p < 0.001). There were more errors on Table 8.1 Mean percentages and standard deviation (in brackets) of error scores by syntactic construction and group Total error Hebrew Russian Arabic English average speakers speakers speakers speakers (N=64) (N=16) (N=16) (N=16) (N=16) Attributive adjectives Predicative adjectives Total error average

25.26 (17.49)

17.31 (7.17)

12.63 (14.11)

1.17 (1.31)

14.09 (10.01)

27.34 (15.79)

22.52 (9.41)

15.62 (14.80)

3.51 (2.71)

17.24 (10.67)

26.29 (17.97)

19.92 (9.84)

14.12 (15.35)

2.34 (2.75)

15.66 (11.43)

8

Adjective Inflection in Hebrew: A Psychollinguistic Study of Speakers of . . .

169

Table 8.2 Mean percentages and standard deviation (in brackets) of error scores by sentence grammaticality and group Total error Hebrew Russian Arabic English average speakers speakers speakers speakers (N=64) (N=16) (N=16) (N=16) (N=16) Grammatical sentences Ungrammatical sentences Total error average

20.57 (12.99)

9.89 (6.20)

9.50 (11.78)

1.30 (1.49)

10.31 (8.11)

32.03 (21.54)

29.94 (11.60)

18.75 (17.79)

3.38 (2.83)

21.02 (13.44)

26.29 (17.97)

19.92 (9.84)

14.12 (15.35)

2.34 (2.75)

15.66 (11.43)

predicative adjectives than on attributive adjectives. There was also an effect for group (F(1,60) = 29.94, p < 0.001). The English speakers performed worse than the Russian speakers and the Hebrew speakers.

8.3.2 Analysis by Sentence Grammaticality and Group Next, we examined performance on the grammar test by group and sentence grammaticality. We expected the ungrammatical sentences to score lower than the grammatical sentences (Alfi-Shabtay, 2006; DeKeyser, 2000; Johnson & Newport, 1989). Table 8.2 presents error scores on grammatical and ungrammatical sentences in the four groups. A two-way ANOVA with repeated measures (group (4) sentence grammaticality (2)) was conducted on the data on Table 8.2. An effect for sentence grammaticality emerged (F(1,60) = 68.85, p < 0.001). All groups scored worse on the ungrammatical sentences. Group and sentence grammaticality interacted (F(3,60) = 8.22, p < 0.001): the Arabic speakers were most affected by sentence grammaticality than the other groups, and the Hebrew speakers were least affected.

8.3.3 Analysis by Category, Sentence Grammaticality and Group We examined performance on the grammaticality judgment test by category, sentence grammaticality, and group. We expected the irregular categories to score worse than the regular categories (Alfi-Shabtay, 2006). In preparation for analysis we calculated a mean average of the three sentence pairs which comprise each category. Table 8.3 presents error scores on the eight adjective inflection categories by sentence grammaticality: grammatical (henceforth: Gr) and ungrammatical (henceforth: Ung), and group: English speakers (henceforth: ES), Arabic speakers (henceforth: AS), Russian speakers (henceforth: RS) and Hebrew speakers (henceforth: HS).

8.3 (14.0)

1.0 (4.1)

4.1 (10.8)

10.4 (21.7) 3.1 (7.6)

4.1 (10.8) 1.0 (4.1)

13.5 (21.7)

25.0 (25.6)

0.0 (0.0)

5.2 (12.4)

9.3 (14.2)

6.2 (10.3)

13.5 (20.0)

40.6 (33.1)

0.0 (0.0)

12.5 (21.3)

2.1 (8.3)

2.1 (8.3)

8.3 (14.6)

28.1 (28.3)

10.4 (18.0)

2.1 (8.3)

17.7 (28.2)

7.2 (16.5)

32.2 (28.9)

1.0 (4.1)

25 (29.7)

33.3 (29.4)

39.5 (34.6)

1.0 (4.1)

12.5 (26.6)

14.5 (22.1)

34.3 (31.3)

1.0 (4.1)

26.0 (30.0)

36.4 (26.0)

39.5 (34.8)

3.1 (9.8)

14.5 (21.5)

17.6 (22.6)

30.2 (30.2)

3.1 (6.7)

33.3 (34.3)

31.2 (25.5)

0.0 (0.0)

6.2 (15.3)

16.6 (20.4)

15.6 (25.3)

HS (N=16)

19.8 (22.3)

22.8 (25.3)

24.9 (24.9)

3.1 (6.7)

6.25 (12.1)

11.6 (17.9)

RS (N=16)

21.8 (19.5)

29.1 (28.8)

5.2 (11.6)

16.8 (23.5)

AS (N=16)

34.3 (30.3)

43.7 (33.7)

13.5 (18.6)

10.4 (21.7)

ES (N=16)

34.3 (33.0)

IRG Ung

Table 8.3 Mean percentages and standard deviation (in brackets) of error scores by category, sentence grammaticality and group Categories SM SF PM PF MSS XM XF Gr Ung Gr Ung Gr Ung Gr Ung Gr Ung Gr Ung Gr Ung Gr

170 I. Alfi-Shabtay and D. Ravid

% Error Scores

8

Adjective Inflection in Hebrew: A Psychollinguistic Study of Speakers of . . . 100 90 80 70 60 50 40 30 20 10 0

11.3

12.3

13.7

13.9

9.9

Singular masculine

Singular feminine

Plural masculine

Plural feminine

Mass noun

171

19.7

20.7

22.1

Irregular masculine suffix

Irregular feminine suffix

Irregular feminine plural suffix

Adjective Inflection Categories

Fig. 8.1 Ranking order of % error scores on adjective inflection categories

A three-way ANOVA with repeated measures (group (4) X sentence grammaticality (2)) X category (8)) was conducted on the data on Table 8.3. An effect for category emerged (F(7, 54) = 11.71, p < 0.001). It is depicted by ranking order on error scores in Fig. 8.1. Significant differences were found between the categories. Plural feminine nouns, irregular suffix; irregular singular masculine nouns; and irregular singular feminine nouns – scored worse than the regular singular-masculine category; they also scored worse than the mass-noun category. Group and category interacted (F(7, 6) = 6.03, p < 0.001): the English speakers performed worst on the three most difficult categories: Plural feminine nouns, irregular suffix; irregular singular masculine nouns; and irregular singular feminine nouns, and the Hebrew speakers had the fewest errors on all categories. The Arabic speakers scored worse than the Russian and the Hebrew speakers on the singular masculine, singular feminine and plural-masculine categories. There was also an interaction of category and sentence grammaticality (F(7, 54) = 2.91, p < 0.001): more errors were produced on the ungrammatical sentences of singular masculine, singular feminine, plural feminine, irregular singular feminine-nouns compared with their parallel grammatical sentences.

8.3.4 Analysis by Group and the Combined Variable of Sentence Grammaticality and Syntactic Construction Finally, we examined performance on the grammar text by the combined variable of sentence grammaticality and syntactic construction. In preparation for statistical analysis, we combined syntactic construction and sentence grammaticality into four new variables. These were: grammatical attributive; grammatical predicative; ungrammatical attributive; and ungrammatical predicative. This was done to measure their effect on success on the grammaticality judgment test in the four groups. Table 8.4 presents error scores on grammatical and ungrammatical attributive and predicative sentences by group.

172

I. Alfi-Shabtay and D. Ravid

Table 8.4 Mean percentages and standard deviation (in brackets) of error scores by group, sentence grammaticality and construction Total Hebrew Russian Arabic English average speakers speakers speakers speakers (N=64) (N=16) (N=16) (N=16) (N=16) Grammatical attributive Grammatical predicative Ungrammatical attributive Ungrammatical predicative Total average

19.79 (13.48)

7.03 (6.58)

8.85 (10.74)

0.78 (1.67)

9.11 (8.11)

21.35 (14.09)

12.76 (7.37)

10.15 (13.85)

1.82 (2.13)

11.52 (9.36)

30.72 (24.71)

27.60 (11.27)

16.40 (18.72)

1.56 (2.08)

19.07 (14.19)

33.33 (19.60)

32.29 (14.15)

21.09 (18.09)

5.20 (5.15)

22.97 (14.24)

26.29 (17.97)

19.92 (9.84)

14.12 (15.35)

2.34 (2.75)

15.66 (11.43)

A two-way ANOVA with repeated measures (group (4) X the combined variable of sentence grammaticality and construction (4)) was conducted on the data on Table 8.4. An effect for sentence grammaticality and syntactic construction emerged (F(7, 54) = 63.22, p < 0.001). All groups scored worse on the ungrammatical predicative sentences than on the grammatical predicative and attributive sentences.

8.4 Discussion The comparative study described here has focused on the adjective inflection system in Hebrew in 64 proficient speakers of Russian, English, and Arabic, and native Hebrew speakers with whom they were compared. The four languages have distinct morphological typology. While Hebrew and Arabic share a common Semitic origin, yet follow different agreement rules especially in plural, Russian, a Slavic language, is the only language with three genders and case agreement, but without definiteness agreement as exists in Hebrew and Arabic (Chesterman, 1991; Lyons, 1999; Wexler, 1976). English, a morphologically sparse genealogically classified Germanic language with a large lexicon of borrowed words mostly of French-Latinate origin, is the only language whose system expresses no gender and/or number agreement (Conrad et al., 2002; Quirk et al., 1985). The study compared performance on an original auditory grammaticality judgment test in Hebrew of two syntactic constructions, attributive and predicative adjectives, and of eight adjective categories inflected in agreement with nouns. The concern of the study was with the proficiency of adjective inflection in Hebrew in the four populations, as well as with the extent to which they are affected by syntactic construction, category, and sentence grammaticality. Attributive adjectives were expected to score lower than predicative adjectives (Ninio, 2004; Ravid & Nir, 2000; Ravid & Levie, 2006); the English speakers whose L1 does not mark inflection for adjectives (Conrad et al., 2002; Quirk

8

Adjective Inflection in Hebrew: A Psychollinguistic Study of Speakers of . . .

173

et al., 1985) were expected to perform worse than all other speakers; and irregular categories were expected to score worse than the other categories. Participants showed high performance on Hebrew adjectives. Performance was worst among the English speakers, and best in the Hebrew speakers. Contrary to the results of L1 studies of the later acquisition of attributive adjectives (Ninio, 2004; Ravid & Nir, 2000; Ravid & Levie, 2006; Ravid & Cahana-Amitay, 2005), the current study showed that attributive adjectives performed better compared to predicative adjective. As expected, ungrammatical sentences scored lower than the parallel grammatical sentences, and ungrammatical sentences with predicative adjectives scored worse than grammatical sentences with predicative and attributive adjectives. The irregular adjective categories, inflected in agreement with plural feminine nouns, singular feminine nouns, and singular masculine nouns, scored lower than the regular singular masculine category. However, they also scored lower than the irregular mass-noun category. Moreover, the English speakers who performed worse than the Russian and the Hebrew speakers, had low scores also on the three most difficult categories: plural feminine nouns, singular feminine nouns, and singular masculine nouns. The Arabic speakers scored worse than the Russian and the Hebrew speakers on the plural masculine, singular masculine, and singular feminine categories; and the Hebrew speakers had the fewest errors on all categories. The Arabic speakers had more errors on the ungrammatical sentences than the other groups. The results of this study can be interpreted via three forces that affect the performance on adjective inflection in study participants. These are: (1) first language constraints; (2) second language constraints; and (3) constraints of processing and memory. First language constraints are demonstrated in the English-speaking participants, whose first language does not mark inflection for adjectives (Conrad et al., 2002; Quirk et al., 1985). This led to difficulties with judging adjective inflection sentences in Hebrew. Similarly, the Arabic speakers, whose L1 follows agreement rules different than that of Hebrew, especially in plural, (Holes, 1995), encountered many problems with the Hebrew plural categories. These findings can be clarified using DeKeyser’s model of what makes L2 grammar difficult (2005). The findings may reflect problems of form, which refer to the number of choices involved in picking all the appropriate morphemes and allomorphs to express meanings and placing them properly within the system. The acquisition of morphology in L2 is often shakily represented in learners’ intuitions, even after many years of exposure to the target language (e.g., Alfi-Shabtay, 2006; DeKeyser, 2000; DeKeyser et al., 2010; Johnson & Newport, 1989). This problem is most complex in the mastery of richly inflected or what are traditionally called highly synthetic languages such as Hebrew, especially in learners whose L1 contains sparse morphology what are traditionally called analytic languages such as English. Elsewhere, Jiang (2004) showed that errors of verb agreement with complex noun phrases in ESL were due to

174

I. Alfi-Shabtay and D. Ravid

lack of sensitivity to plural marking on the noun, not to problems with processing agreement. In contrast, transparency of a specific form may explain the high performance of study participants on the mass-noun category, despite its opaque meaning. This could have been derived from the ‘final suffix–im harmony’ (Berman, 1985) which is typical of this category e.g., shamay-im behir-im/ ‘clear,Pl,Masc/sky,Pl,Masc’, and allowed facile acquisition (DeKeyser, 2005). Problems of meaning (DeKeyser, 2005) may also explain our findings, and point at L1 constraints. This is the case where the semantic system of the L1 is different from that of the L2, as inflection for English-speaking learners of Hebrew, or where equivalent notions do not get expressed overtly in L1, except through discourse pattern, as is the case of Hebrew and Arabic adjective inflection systems. Following DeKeyser, complexity of form can constitute a source of difficulty, because of novelty, abstractness or a combination of both. Articles, classifiers, grammatical gender, and verbal aspect express highly abstract notions that are extremely hard to infer, implicitly or explicitly, from the input. These elements of grammar are notoriously hard to acquire for native speakers of L1s that do not have them or that use a very different system. Second language constraint constitutes the second factor that may affect the mastery of adjective inflection in Hebrew. This refers to the complex nature (DeKeyser, 2005) of adjective inflection in Hebrew (Berman, 1981; Ravid, 1995), and may indicate the difficulty of the non-native participants to internalize irregular adjective-inflection categories. This difficulty does not seem to derive from either problems of form or meaning, for it was also found in the Russian-speaking participants who already had the concept of inflection in their L1. It does, however, seem to be the result of difficulty in opacity, a complex form of the problem of low form-meaning correlation (DeKeyser, 2005), which represents the adjective inflection system in Hebrew. For example, when a singular feminine noun milim ‘words’ ends with a final suffix –im, marked for plural masculine, it may lead to a misinterpretation of the form as a marked masculine gender in adjective inflection, such as *mil-im kash-im ‘words[Fm] difficult’, when the feminine suffix –ot ‘mil-im kash-ot’ is required (Ravid, 1995). In this case, the correlation between form and meaning becomes very hard to detect: different forms stand for the same meaning, and the same form stands for different meanings. The third factor addresses constraints that are beyond the influence of the first language, or the characteristics of the second language. It refers to constraints of processing and memory that derive from the nature of the syntactic construction as well as the language test. Memory becomes an issue when adjective is far from its subject. This is especially the case with auditory presentations, as given in this study. It may be too hard for subjects to remember all the relevant earlier parts of a sentence to process agreement and to detect errors later. Hence, they may start substituting something generated by their own competence rather than something remembered (Kellerman, 1995). This may explain the current study findings of the lowest performance on predicative

8

Adjective Inflection in Hebrew: A Psychollinguistic Study of Speakers of . . .

175

than attributive sentences, and imply why native Hebrew speakers, too, find adjective inflection difficult on ungrammatical sentences when noun and adjective are far apart. Constraints of processing and memory which constitute the essence of grammaticality judgments may also explain the contradiction found in this study as to the low performance on predicative adjectives when the noun and adjective were far apart. These findings are in contrast to previous developmental studies which showed that predicative adjectives emerge earlier on than attributive adjectives, but the proportion of predicative adjectives decrease while attributive adjectives rise with age and schooling (Ninio, 2004; Ravid & Levie, 2006; Ravid & Nir, 2000). Finally, the worst scores on the ungrammatical sentences are common on grammaticality judgments (Alfi-Shabtay, 2006; DeKeyser, 2000; Johnson & Newport, 1989). This is generally attributed to the strong bias towards accepting, which explains the better percentage correct on the grammatical sentences (DeKeyser, 2000). It further supports the diagnostic nature of ungrammatical sentences found in Alfi-Shabtay (2006). The effect of sentence grammaticality in that study was noticeable given these two conditions. One, the better the performance, the lower the effect of sentence grammaticality; and two, the higher the opacity of the category (DeKeyser, 2005), the higher the effect of sentence grammaticality. The current study also points at the interaction between syntactic construction, sentence grammaticality, category and group. The findings indicate that low performance on adjective inflection strengthens the role of syntactic construction and sentence grammaticality. Hence, the English speakers were especially affected by syntactic construction and sentence grammaticality. In contrast, better scores on the category weaken the role of sentence grammaticality: the mass noun category which scored the best was least affected. In conclusion, the findings of this study shed light on the complexity of adjective inflection in Hebrew, and demonstrate the trio effects for L1, L2, and for memory and processing in proficient native and non-native Hebrew speakers of four typologically distinct L1s. The findings may help develop appropriate curricula and instructional strategies for adult Hebrew learners. Performance on attributive and predicative adjectives in L2 Hebrew should be further tested using semi-constructed instruments (i.e., production tasks). Similar experimental studies should be conducted on other rich categories in learners whose L1 morphological typology is different from that of Hebrew.

References Alfi-Shabtay, I. (2006). Hebrew knowledge of Veteran Russian-Speaking immigrants in Israel. Unpublished doctoral dissertation, Tel Aviv University, Tel Aviv. Berman, R. A. (1978). Modern Hebrew structure. Tel Aviv: Tel Aviv University Press.

176

I. Alfi-Shabtay and D. Ravid

Berman, R. A. (1979). Form and function: Passives, middles and impersonals in Modern Hebrew. Berkeley. Linguistic Society, 5, 1–27. Berman, R. A. (1981). Regularity vs. anomaly: The acquisition of inflectional morphology. Journal of Child Language, 8, 265–282. Berman, R A. (1983). Establishing a schema: Children’s construal of verb-tense marking. Language Sciences, 5, 61–68. Berman, R. A. (1985). Acquisition of Hebrew. In D. I. Slobin (Ed.), The crosslinguistic study of language acquisition (Vol. I, pp. 255–371). Hillsdale, NJ: Erlbaum. Berman, R. A. (1994). Formal, lexical, and semantic factors in the acquisition of Hebrew resultative particles. In S. Gahl, A. Dolbey, & C. Johnson (Eds.), BLS (Vol. 20, pp. 82–92). Berman, R. A. (1997). Discussion and research on acquiring Hebrew as a mother tongue: The beginning acquisition of syntax and discourse formation. In Y. Shimron (Ed.), Studies in the psychology of language in Israel: Language acquisition, reading and writing (pp. 37–69). Jerusalem: Magnes Press [In Hebrew]. Bybee, J. L. (1985). Morphology: A study of the relation between meaning and form. Amsterdam: John Benjamins. Casseli, M. C., Bates, E., Casadio, P., & Fenson, J. (1995). A cross-linguistic study of early lexical development. Cognitive Development, 10, 159–199. Chesterman, A. (1991). On definition: A study on special reference to English and Finnish. Cambridge: Cambridge University Press. Conrad, S. Biber, D., & Leech, G. (2002). Longman student grammar of spoken and written English: Workbook. London: Longman. Coppieters, R. (1987). Competence differences between native and near-native speakers. Language, 63, 544–573. DeKeyser, R. M. (2000). The robustness of the critical period in second language acquisition. Studies in Second Language Acquisition, 22, 499–533. DeKeyser, R. M. (2005). What makes learning second-language grammar difficult? A review of Issues. Language Learning, 55, 1–25. DeKeyser, R. M., Alfi-Shabtay, I., & Ravid, D. (2010). Cross-linguistic evidence for the nature of age effects in second language acquisition. Applied Psycholinguistics, 31(3), 413–438. DeKeyser, R. M, Alfi-Shabtay, I., Ravid, D., & Shi, M. (2005). The role of salience in the acquisition of Hebrew as a second language. International symposium on bilingualism (Vol. 5). Barcelona: Spain. Diesendruck, D., Hall, G., & Graham, S. A. (2006). Children’s use of syntactic and pragmatic knowledge in the interpretation of novel adjectives. Child Development, 77, 16–30. Dixon, R. M. W. (1977). A grammar of Yidin. Cambridge: Cambridge University Press. Dromi, E. (1987). Early lexical development. Cambridge: Cambridge University Press. Dromi, E., & Berman, R. A. (1986). Language-specific and language-general in developing syntax. Journal of Child Language, 13, 371–387. Englebretson, R. (1997). Genre and grammar: Predicative and attributive adjectives in spoken English. Berkeley Linguistics Society, 23, 411–421. Ferris, C. (1993). The meaning of syntax: A study in the adjectives of English. Longman: London. Holes, C. (1995). Modern Arabic: Structures, functions and varieties. London: Longman. Hora, A., Avivi-Ben Zvi, G., Levie, R., & Ravid, D. (2006). Acquiring diminutive structures and meanings in Hebrew: An experimental study. In I. Savickiene & W. U. Dressler (Eds.), The acquisition of diminutives (pp. 295–317). Amsterdam: John Benjamins. Ingram, D. (1989). First language acquisition: Method, description and explanation. Cambridge: Cambridge University Press. Jiang, N. (2004). Morphological insensitivity in second language processing. Applied Psycholinguistics, 25, 603–634.

8

Adjective Inflection in Hebrew: A Psychollinguistic Study of Speakers of . . .

177

Johnson, J. S. (1992). Critical period effects in language acquisition: The effect of written versus auditory materials on the assessment of grammatical competence. Language Learning, 42, 217–248. Johnson, J. S., & Newport, E. L. (1989). Critical period effects in second language learning: The influence of maturational state on learning English as a second language. Cognitive Psychology, 21, 60–99. Johnson, J. S., & Newport, E. L. (1991). Critical period effects on universal properties of language. The status of subjacency in the acquisition of a second language. Cognition, 39, 215–258. Kaplan, D. (1983). Order of acquisition of morpho-syntactic elements in Hebrew-Speaking nursery schoolers. Unpublished Master’s thesis, Tel Aviv University, Tel Aviv [in Hebrew]. Keenan, E. L. (1976). Towards a universal definition of ‘‘subject’’. In C. N. Li (Ed.), Subject and topic (pp. 303–333). New York: Academic. Kellerman, E. (1995). Age before beauty: Johnson and Newport revised. In E. L. Eubank, L. Selinker, & M. S. Smoth (Eds.), The current state of interlanguage: Studies in honor of William (pp. 219–231). Rutherford, Amsterdam: John Benjamins. Levie, R. (2002). Adjective development in schoolchildren. MA thesis, School of Education, Tel Aviv University, Tel Aviv. [in Hebrew]. Levin, I., Ravid, D., & Rappaport, S. (2001). Morphology and spelling among Hebrewspeaking children: From kindergarten to first grade. Journal of Child Language, 28, 741–769. Levy, Y. (1983). The acquisition of Hebrew plurals: The case of the missing gender category. Journal of Child Language, 10, 107–121. Lyons, C. (1999). Definiteness. Cambridge: Cambridge University Press. Maltzoff, N. (1987). Essentials of Russian grammar. A complete guide for students and professionals. Passport Books. Lincolnwood, IL: National Textbook Company. Marvin, C. A., Beukelman, D. R., & Bilyeu, D. (1994). Vocabulary-use patterns in preschool children: Effects of context and time sampling. Augmentative and Alternative Communication, 10, 224–237. Nelson, K. (1976). Some attributes of adjectives used by young children. Cognition, 4, 13–30. Ninio, A. (2004). Young children’s difficulty with adjectives modifying nouns. Journal of Child Language, 31, 255–285. Owens, R. (2001). Language development: An introduction (5th ed.). Boston: Allyn and Bacon. Patkowski, M. S. (1994). The critical age hypothesis and interlanguage phonology. In M. Yavas (Ed.), First and second language phonology (pp. 205–221). San Diego, CA: Singular. Polinsky, M. (2004). Word class distinctions in an incomplete grammar. In D. Ravid & H. Bat Zeev Shyldkrot (Eds.), Perspectives on language and language development (pp. 419–434). Dordrecht: Kluwer. Quirk, R., Greenbaum, S., Leech, G., & Svartvik, J. (1985). A comprehensive grammar of the English language. London: Longman. Ravid, D. (1995). Language change in child and adult Hebrew: A psycholinguistic perspective. New York: Oxford University Press. Ravid, D. (1997). Morphological development a duo: Pre- and proto-morphology in the language of Hebrew-speaking twins. Papers and Studies in Contrastive Linguistics, 33, 79–102. Ravid, D. (2004a). Emergence of linguistic complexity in written expository texts: Evidence from later language acquisition. In D. Ravid & H. Bat-Zeev Shyldkrot (Eds.) Perspectives on language and language development (pp. 337–355). Dordrecht: Kluwer. Ravid, D. (2004b). Later lexical development in Hebrew: Derivational morphology revisited. In R. A. Berman (Ed.), Language development across childhood and adolescence: Psycholinguistic and crosslinguistic perspectives (pp. 53–82). Amsterdam: Benjamins.

178

I. Alfi-Shabtay and D. Ravid

Ravid, D., Abu-Nofel, H., & Huri, R. (1998). In the presence of multiple cues: Learning to inflect plural adjectives in Palestinian Arabic. The Utrecht congress on storage and computation in linguistics, Utrecht Institute of Linguistics, The Netherlands. Ravid, D., & Cahana-Amitay, D. (2005). Verbal and nominal expression in narrating conflict situations in Hebrew. Journal of Pragmatics, 37, 157–183. Ravid, D., & Levie, R. (2010). Hebrew adjectives in later language text production. First Language, 30(1), 27–55. Ravid, D., Levie, R., & Avivi-Ben Zvi, G. (2003). Morphological disorders. In L. Verhoeven & H. van Balkom (Eds.), Classification of developmental language disorders: Theoretical issues and clinical implications (pp. 235–260). Mahwah, NJ: Erlbaum. Ravid, D., & Malenky, A. (2001). Awareness of linear and nonlinear morphology in Hebrew: A developmental study. First Language, 21, 25–56. Ravid, D., & Nir, M. (2000). On the development of the category of adjective in Hebrew. In M. Beers, B. van den Bogaerde, G. Bol, J. de Jong, & C. Rooijmans (Eds.), From sound to sentence: Studies on first language acquisition (pp. 113–124). Groningen: Center for Language and Cognition. Ravid, D., & Shlesinger, Y. (1987). On the classification and structure of –i-suffixed adjectives. Hebrew Linguistics, 25, 59–70 [in Hebrew]. Ravid, D., & Zilberbuch, S. (2003a). Morpho-syntactic constructs in the development of spoken and written Hebrew text production. Journal of Child Language, 30, 395–418. Ravid, D., & Zilberbuch, S. (2003b). The development of complex nominals in expert and non-expert writing: A comparative study. Cognition and Pragmatics, 11, 267–297. Rice, M. L. (1990). Preschoolers’ QUIL: Quick incidental learning of words. In G. ContiniRamsden & C. Snow (Eds.), Children’s language (Vol. 7) (pp. 171–195). Hillsdale, NJ: Erlbaum. Schachter, P. (1985). Parts-of-speech systems. In T. Shopen (Ed.), Language typology and syntactic description, Vol. I: Clause structure (pp. 3–61). Cambridge: Cambridge University Press. Schachter, J. (1990). On the issue of completeness in second language acquisition. Second Language Research, 6, 93–124. Sommers, R. K., Kozarevich, M., & Michaels, C. (1994). Word skills of children normal and impaired in communication skills and measures of language and speech development. Journal of Communications Disorders, 27, 223–240. Sorace, A. (1993). Incomplete vs. divergent representations of unaccusativity in non-native grammars of Italian. Second Language Research, 9, 22–47. Slobin, D. I. (1966). The acquisition of Russian as a native language. In F. Smith & G. Miller (Eds.), The genesis of language: A psycholinguistic approach (pp. 129–148). Cambridge: MIT Press. Timberlake, A. (1993). Russian. In B. Comrie & G. Corbett (Eds.), The Slavonic languages (pp. 827–886). London: Routledge. Valian, V. (1986). Syntactic categories in the speech of young children. Developmental Psychology, 22, 562–579. Voeykova, M. (1997). Acquisition of adjectival inflections (secondary paradigms in child Russian). Papers and Studies in Contrastive Linguistics, 33, 141–151. Wexler, P. (1976). On the non-lexical expression of determinedness. (With special reference to Russian and Finnish). Studia Linguistica, 30, 34–67. Yagev, I. (2000). Language skills in cleft-palate and healthy preschool children. Unpublished MA thesis, Tel Aviv University, Israel.

Chapter 9

Verb Inflections as Indicators of Bilingual SLI: Qualitative Vs. Quantitative Measurements Sharon Armon-Lotem, Galit Adam, Anat Blass, Jonathan Fine, Efrat Harel, Elinor Saiegh-Haddad, and Joel Walters

9.1 Introduction In recent years there has been an increase in the study of Specific Language Impairment (SLI), focusing on a wide variety of languages (e.g. English, Greek, Hebrew, Italian, Spanish, and Swedish) and on an extensive range of linguistic indicators, including verb morphology, definiteness, passives, wh-questions, and relative clauses. These studies are focused primarily on SLI in a single language. Only a few studies have specifically explored the representation and processing of language in bilinguals with SLI, collecting and analyzing data in both of a child’s developing languages (e.g., Paradis, 2007). These studies, which mostly focus on morphosyntactic abilities, suggest that errors typical of SLI within the verbal system in general and in the use of inflections in particular, are also found in the language of bilinguals. The present paper examines the linguistic production of bilingual EnglishHebrew preschool children, in order to assess the relative contribution of each language to the child’s linguistic representations and underlying processes. The children were recruited from ‘‘language preschools’’ in English-speaking neighborhoods and diagnosed for SLI. Findings are compared with those of bilingual children in regular preschools from the same neighborhood. This research focuses on the typological differences between languages and implications for diagnosis and treatment of language-impaired bilingual children. The use of verb morphology in general and the inflectional verbal system in particular, in both Hebrew and English, suggests that despite a similarity in error type, some errors are indicative of SLI. Definitions of bilingualism include social, cognitive, and linguistic criteria, sometimes confounded and overlapping. Most bilingual children in Israel are children of first generation immigrants. This population is highly heterogeneous, and this variability is strongly influenced by the order and ages in which the two languages have been acquired (simultaneous/sequential), as S. Armon-Lotem (*) Bar-Ilan University, Ramat-Gan, Israel e-mail: [email protected] M. Leikin et al. (eds.), Current Issues in Bilingualism, Literacy Studies 5, DOI 10.1007/978-94-007-2327-6_9, Ó Springer ScienceþBusiness Media B.V. 2012

179

180

S. Armon-Lotem et al.

well as by acquisition context (one parent-one language vs. home languageschool language), family size and birth order. For the present study, a bilingual child is a child with bilingual background who is able to function in two languages (carry a conversation and understand) at a near native level (whether typical or impaired). Heterogeneity also characterizes the population which is diagnosed for SLI. Specific Language Impairment (SLI) is first and foremost a significant deficit in linguistic abilities, a ‘‘primary language disorder’’ (Bishop & Snowling, 2004). SLI has been defined by both exclusionary and inclusionary criteria. Among the exclusionary criteria, Tallal and Stark (1981) list: no hearing loss (no history of otitis media), no emotional and behavioral problems, no below average non-verbal IQ (>¼85), no neurological problems, and no severe articulation/phonological deficit. Some inclusionary criteria [are]/include: specific language deficits, but no general communication difficulties; a language deficit more severe than an articulatory disorder; language difficulties greater than reading problems; language deficit for both receptive and expressive skills. Very often, SLI linguistic behavior resembles the linguistic production of children with normal language development at earlier stages of acquisition. For the present study, SLI children will be defined as children who have normal performance IQ, meet the exclusionary criteria for SLI, and are below the norms set for typical development by standardized language tests. The research for this paper focused on nine bilingual English-Hebrew children, ages 5–7, attending language preschools. Data were collected in both languages in separate sessions by native speakers of the target language, using naturalistic procedures and three elicitation tasks. These data are compared with three typically developing (TD) bilingual speakers from regular preschools. All bilingual children, whether TD or SLI, showed omission of past tense –ed and third person –s morphology in English, but the ratio of errors was higher among those children with the lowest scores on the standardized tests in both languages. These children also had many errors in their use of 2nd person morphology in Hebrew. The findings in this paper are compared with studies of SLI monolingual acquisition in both languages, as well as with studies of typical bilingual development, in order to be able to evaluate the nature of the errors. The differences between the tasks will be discussed in the context of processing limitations, and the imitation task will be discussed as a potential basis for a diagnostic tool for bilingual SLI.

9.2 Monolingual SLI SLI is manifested at different levels of representation. Delayed or impaired morpho-phonological and prosodic processes (Owen 2001; Leonard, 1998), delays in acquisition of words and word retrieval (Leonard, 1998), difficulties with morpho-syntactic representations (e.g., Rice & Wexler, 1996), syntactic structures (e.g., Van der Lely, 1998; Fridmann & Novogrodsky, 2004;

9

Verb Inflections as Indicators of Bilingual SLI: Qualitative Vs. Quantitative . . .

181

Novogrodsky & Fridmann, 2006) as well as certain narrative, discourse and pragmatic skills (e.g., Boscolo, Ratner, & Rescorla, 2002) are manifested differently across languages, reflecting typological differences. The present review of manifestations of SLI focuses on morphosyntactic representation. Morphological and syntactic features are primary linguistic indicators of SLI (Crago & Gopnik, 1994; Leonard, Sabbadini, Volterra, & Leonard, 1988). Native English-speaking SLI children show impaired production of verbal inflections, omitting these linguistic structures in obligatory contexts (Bishop, 1994; Clahsen, 1991; Gopnik & Crago, 1991; Rice & Wexler, 1996), resulting in the use of infinitival forms for matrix verbs (Rice & Wexler, 1996). They also omit auxiliary verbs (Hadley & Rice, 1996), e.g., auxiliaries in negation (Eyer & Leonard, 1995). There is little evidence of embedding and relative clauses; when embedded clauses are used, they lack function words, e.g., Me know that my guy right there or Me know what in these (Eyer & Leonard, 1995:186). In addition, there are omissions and deviations in argument structure (Leonard, 1998), as well as adjunct omissions, e.g., temporal adverbs (Fletcher & Garman, 1988). Children with SLI tend to omit plural marking on nouns, e.g. two book (Rice & O¨tting, 1993), determiners, e.g., a, the in obligatory contexts (Beastrom & Rice, 1986) but not genitive ‘s (Leonard, 1998; Steckol & Leonard, 1979; Albertini, 1980). They use simplified noun phrases and sentence structure (the small dog à small dog or bare NPs, e.g., dog), and avoid complex PPs, such as on the rock, preferring bare nouns (Van Der Lely, 1998). Leonard (1998) concludes that the most consistent differences between SLI children and language-matched counterparts are for finite verb inflections, copular forms, and auxiliaries which require agreement. SLI children were also found to be insensitive to derivational relations (Moats & Smith, 1992) showing difficulties in applying derivational rules. Published studies on Hebrew SLI, conducted by Dromi, Leonard and colleagues (e.g., Dromi, Leonard, & Shteiman, 1993; Leonard & Dromi, 1994; Dromi, Leonard, & Adam, 1997; Dromi, Leonard, Adam, & ZadunaiskyEhrlich, 1999), focus on the use of finite verb morphology. Findings in Hebrew indicated that ‘‘Hebrew-speaking children with SLI resembled MLU controls in their use of both present and past tense inflections requiring agreement with the subject’’ (Dromi et al., 1997:1415). A further study (Dromi et al., 1999), however, showed that the more morphologically complex past tense does pose a problem for Hebrew-speaking SLI children, whereas difficulties with the less morphologically complex present tense are less pronounced. Most errors were related to the use of tense or person, but usually not both. Dromi et al. (1997) argue that their findings from Hebrew support a limited processing capacity framework, since more complex structures, which place more demands on the system, seem to be more impaired. For nominal syntax, Dromi et al. (1993) found plural formation, adjectival agreement, and use of the accusative case marker (et) all to be delayed, but not different from language matched controls. Derivational morphology seems to be more challenging in Hebrew, as shown by Ravid, Levie, and Avivi-Ben Zvi (2003) who reported differences between

182

S. Armon-Lotem et al.

children with SLI and a ‘‘language matched’’ group in the use of derived adjectives and derived nominals. To conclude, both languages show a reduction in the use of verbal inflectional morphology. In English, SLI children use root infinitives. In Hebrew, however, due to the salience of the inflectional system, this simplification is more limited and many aspects of inflectional morphology are retained. Nominal morpho-syntax in Hebrew is typologically very different from English, not lending itself easily to comparison. In English, difficulties are more at the functional level, whereas in Hebrew they are found at the word-formation level, due to the derivational complexity of Hebrew words. More specifically, Hebrew is morphologically marked for number and gender in the nominal system and for person, number and gender in verbs; English is morphogically less marked, i.e., for number on nouns and for third person singular on verbs. Further relevant differences are explained later.

9.3 Bilingual SLI Studies comparing bilingual SLI children with other TD bilingual children give us some data on the linguistic abilities of SLI bilinguals compared to monolinguals of the same language. Paradis, Crago, Genesee, and Rice (2003) compared French-English bilingual SLI children with their monolingual SLI peers. This was done with respect to their morphosyntax in language production. Children’s use of tense-bearing and non- tense-bearing morphemes was examined in obligatory contexts in spontaneous speech. Bilingual and monolingual children with SLI showed greater accuracy with non-tense than with tense morphemes. However, bilingual SLI children and their monolingual peers showed no difference in their use of tense morphemes. Bilingual SLI children appeared similar to their monolingual peers in certain aspects of grammatical morphology in each language tested. This is further supported by Paradis, Crago, and Genesee (2005/2006) who found that while bilingual children had more difficulty with clitics than articles in French and with pronouns in English, bilingual children with SLI performed like their younger unaffected bilingual peers and like monolinguals with SLI. Paradis (2005) examined whether the expressive language characteristics of typically developing children (TD) learning English as a second language have similarities to the characteristics of the English that spoken by monolingual children with specific language impairment (SLI) and whether this could result in the erroneous assessment of TD English language learners (ELL) as language impaired. Paradis tested the children’s accuracy and error types in production of grammatical morphemes through spontaneous and elicited speech. The grammatical morphemes were third person singular (-s), past tense (-ed), irregular past tense, copular BE, the DO auxiliary verb, progressive (-ing), prepositions in and on, plural (-s) and determiners a and the. The elicitation probes were part of the TEGI-test of early grammatical impairment (Rice &

9

Verb Inflections as Indicators of Bilingual SLI: Qualitative Vs. Quantitative . . .

183

Wexler 2001). The ELL’s accuracy rates and error patterns with the grammatical morphemes were similar to those that have been reported for same aged monolingual children with SLI, in both spontaneous and elicited speech. The ELL’s elicitation probe scores were compared to criterion scores and group means from the sample of monolingual children used to develop the TEGI, and their performance on the TEGI was in the range of the clinical population even though there is no reason to suspect that any of these children is language impaired. Both analyses point to the possibility that TD ELL’s could be mistaken as language impaired. For other language combinations, Jacobson and Schwartz (2005) compared incipient bilingual Spanish 4 and 5 year old children with and without SLI. They found that children with SLI used clitic pronouns less frequently than their age matched peers and were less accurate in their use of gender agreement for clitics. No group differences were found for third person singular and plural verb inflections in the preterite tense. Ha˚kansson, Salameh, and Nettelbladt (2003) compared Swedish-Arabic bilingual children, and their grammatical development was tested. They found out that bilingual children with language impairment tend to have a balanced low level of language development in both languages, whereas the bilingual children with normal language development show a higher level of language development in at least one language. Steenge (2006) studied 54 Turkish-Dutch bilingual SLI children ages 6–11 for 3 years, administering 12 standardized tasks, the most relevant ones here being comprehension of function words, sentence reproduction, and story comprehension. Using repeated measures and structural equation modeling analyses, she found better performance in L1/Turkish and concluded that metalinguistic skills in L1/Turkish facilitate development of metalinguistic skills in L2/ Dutch, maintaining that bilingual SLI children should be treated in L1.

9.4 Research Questions and Hypotheses Against this background, the present study aims to investigate the use of the verbal system and in particular tense morphemes, e.g., walk vs. walked in English and holex vs. halax in Hebrew, as well as agreement morphemes, e.g. walk vs. walks in English and halaxti vs. halaxta in Hebrew. Further distinctions are illustrated later in the context of the findings. Comparing the performance of bilingual SLI children to that of bilingual TD children can identify relevant diagnostic measures. The paper will address the following questions: 1. How accurate is the use of tense morphemes by bilingual SLI children compared with bilingual TD children? 2. Which errors characterize each of these populations, and how similar are they? 3. Are there differences in performance for the two languages of the same child?

184

S. Armon-Lotem et al.

4. Are there differences for the various tasks within subjects and across populations? Since the children in the present research have been exposed to their second language for over 2 years, we predict based on the research reviewed that: 1. Bilingual SLI children will perform less accurately in their use of tense morphemes than bilingual TD children. 2. SLI and TD children will have similar error patterns. 3. For both SLI and TD children, we may find a gap between the two languages which reflects the dominance of one of the languages. 4. For TD children we expect to find no significant differences between the spontaneous speech and sentence repetition tasks, whereas for the impaired population we expect to find more difficulties in tasks which require more processing (e.g., sentence repetition, which requires both syntactic processing and short term memory skills).

9.5 Method To test the above hypotheses, data in both languages were collected from bilingual children in special language preschools and in regular preschools. Native speakers of the target languages did the testing using spontaneous speech and three elicitation tasks. All children were also tested with standardized language tests in both languages. In addition, the parents of the children in the language programs were asked to complete a questionnaire evaluating the difficulties of their child and explaining why their child attends a special language program. The findings were analyzed separately for each child, generating individual profiles, which together yielded a broader picture. These findings were compared with those of the bilingual children attending regular preschools.

9.5.1 Subjects Subjects were nine preschool children from bilingual or monolingual English homes, who had acquired Hebrew as their second language in the preschool environment for more than 2 years. All of them attended special ‘‘language preschools’’, after being diagnosed as language impaired (LI) by a speech clinician. All children, 3 girls and 6 boys, ages 5;6–7;4, came from the same bilingual neighborhood and same (middle-high) SES. All children were screened again, at time of study, for both languages using standardized tests (CELF Preschool 2 for English (Wiig, Secord, & Semel, 2004), Goralnik for Hebrew (Goralnik, 1995)). Table 9.1 presents the LI children’s length of exposure to Hebrew and scores on these tests, including age equivalent and

9

Verb Inflections as Indicators of Bilingual SLI: Qualitative Vs. Quantitative . . .

185

Table 9.1 Participants from language preschools, length of exposure to L2/Hebrew, their standardized test scores, and age equivalent CELF Goralnik Length of Age Age Name exposure Score equivalent Percentile Score equivalent Dm606 Sm506 Am506 Ym704 Af609 Of603 Rf606 Bm508 Cm506

2;0 4;0 4;0 3;0 2;0 3;5 4;0 3;0 2;0

73 69 67 70 63 80 100 91 90

4;3 3;1 3;1 3;9 3;5 4;2 6;2 4;2 4;7

4 2 1 2 1 9 50 27 25

130 135 121 142 148 147 140 99 122

3;7 4;1 3;7 4;7 5;6 5;1 4;7 3;1 3;7

percentile rank, where available (the pseudonyms contain the gender (m/f) and age information). As seen in Table 9.1, all nine children scored more than 1.5 SD below the norm on the Goralnik, but only six of them also scored more than 1 SD below norm on the CELF. Thus, in addition to the individual analysis, the standardized screening yielded a bipartite classification of the children, which reflects their linguistic abilities as diagnosed by the standardized tests: 1. Children with atypical development in both languages (A-TD) – Dm606, Sm506, Am506 Ym704, Af609, Of602 2. Children with atypical development in Hebrew but typical development in English (ETD) – Rf606, Bm508, Cm506 These data were compared to data from three bilingual children, one boy and two girls, ages 5;9–6;3 from a nearby regular preschool. Table 9.2 presents these children’s length of exposure and scores on the standardized tests, including age equivalent and percentile where available (the names contain the gender and age information). As seen in Table 9.2, one of these children (Mf601) shows typical development in both languages, one shows typical development in Hebrew and almost typical development in English (14th percentile), which is slightly more than 1 SD below the norm on the CELF Preschool 2), and one, who was exposed to Hebrew for only 2 years, shows typical development only in English, with Hebrew just 1.5 SD below the norm on the Goralnik. One of these children Table 9.2 Participants from regular preschools, their length of exposure to Hebrew, standardized test scores, and age equivalent CELF Goralnik Length of Age Age Name exposure Score equivalent Percentile Score equivalent Mf601 Ef603 Em509

3;0 sim 2;0

90 84 96

5;6 5;0 5;5

25 14 39

155 163 137

6;0 6;3 4;3

186

S. Armon-Lotem et al.

(EF603) is simultaneous bilingual and was included in the typically developing group in order to make it as comparable as possible to the target group which included two children who were exposed to Hebrew for 4 years from the age of 18 months. The three children in this group were never suspected of being language impaired and represent the typically developing (TD) population in regular preschools, where some children are balanced bilinguals, while others show dominance of one language over the other either due to attrition of L1 or incomplete learning of L2 due to shorter length of exposure.

9.5.2 Linguistic Measures: Inflections While English inflects its verbs for past tense and for third person singular in the present, Hebrew uses a wide array of verb inflections. In addition to tense, verbs are inflected for gender and number in the present tense and for gender, number and person in the past and future tenses. While tense differences (with the exception of the present tense of some conjugations characterized by a prefix) are marked by changes to the interdigited vowels, e.g., halax ‘‘walked’’ – holex ‘‘walks’’, gender, number and person are marked by suffixation (except for a person prefix in the future), e.g., halax-ti (1st. sg.), halax-t (2nd.fm. sg.), halx-u (3rd. pl.). This variety of inflections yields over twenty different inflected forms for each lexeme. Studies of monolingual SLI and typically developing bilinguals show that both populations use root infinitives (RIs) in English, e.g., David play ball. In Hebrew, however, SLI children find past tense 2nd person inflection more difficult. The present study focuses on past and present tenses only. Future tense, which is considered a feature of later acquisition and is marked in English by a modal rather than an inflection, was not studied. In English, the present study targets the production of past tense –ed and present tense 3rd person –s. In Hebrew, we look at the use of the four present tense forms (singular masculine, which has no overt inflection, singular feminine, plural masculine and plural feminine), as well as eight past tense forms (including the 3rd person singular which has no overt inflection).

9.5.3 Tasks and Procedures Data were collected using naturalistic speech samples (interview, story telling, free play), as well as three controlled tasks: a sentence completion task within a story context and an enactment task (both based on Dromi et al., 1999), and an elicited imitation task. This variety of tasks made it possible to capture a more reliable picture of the children’s knowledge of the system. The enactment task was used only in Hebrew, since it targets the use of past tense and person morphology, which do not have any overt forms in English. In addition, the parents of the children in the language preschools were asked to fill in a

9

Verb Inflections as Indicators of Bilingual SLI: Qualitative Vs. Quantitative . . .

187

questionnaire in which they evaluated their child’s difficulties and were asked to explain why the child was placed in this environment. The present paper reports the findings of the three controlled tasks. 9.5.3.1 Sentence Completion Within a Story Context This task (based on Dromi et al., 1999) serves to elicit production of 3rd person singular and plural, in present and past tense and examine children’s ability to mark agreement. Three story books were used in Hebrew and two in English. Each book contains a sequence of drawings, which the experimenter uses to tell the story, pausing at preplanned points to allow the child to complete a sentence with the target verb form. To facilitate use of the target forms, the sentence prior to the one the child was expected to produce contains a target verb in different, often the infinitival, form. In English the stories tested the use of past tense –ed (as in ‘‘Here grandpa helps; yesterday grandpa helped ’’) and present tense –s (as in ‘‘Here the cats eat, and here the dog eats’’). In Hebrew, the stories tested the use of gender and number in the third person in past tense (e.g. ‘‘ha-yeled raca likfoc; az hu kafac’’ the boy wanted to jump; so he jumped) and present tense (e.g., ‘‘kan ha-kelev mexapes oxel, ve-kan ha-xatulot ‘mexapsot’ oxel’’ here the dog [masc. sing.] looks for [masc. sing bare form] food, and here the cats [fem.] look for [fem. pl] food). A variety of conjugations [binyanim] was represented in the Hebrew stories. 9.5.3.2 Elicited Imitation This task tested for the use of 3rd person singular and plural, in present and past tense in English and for person inflections in the past tense in Hebrew, in order to tap children’s knowledge of agreement. Twenty-four sentences were used in English (3 verb forms: 2 inflected – for past and 3rd person present – with singular subjects and the bare form with plural subjects with eight sentences for each verb form; 40 sentences were used in Hebrew (5 inflections: 1st singular, 2nd singular masculine, 2nd singular feminine, 1st plural, and 2nd plural, again with eight stimulus sentences for each form. Different conjugations were presented in the Hebrew sentences. Each sentence contained 4–8 words. The child was asked to repeat sentences presented by the experimenter and was rewarded with blocks to build a tower for every sentence produced, regardless of its correctness. 9.5.3.3 Enactment This task (based on Dromi et al., 1999) elicits person inflections in Hebrew only and tested the same three singular forms examined by the imitation task in Hebrew: 1st person singular and 2nd person singular feminine and masculine. In order to elicit 1st person singular, one experimenter leaves the room, while the other experimenter and the child are engaged in two different activities.

188

S. Armon-Lotem et al.

When the first experimenter comes back in, the child is asked to tell her who did what. For example, the experimenter draws and the child jumps. When the absent experimenter comes in, the child is asked to tell her who drew and who jumped, targeting the sentence hi ciyra ve-ani kafacti ‘‘she drew [3rd fem. sing.] and I jumped [1st sing.].’’ In order to elicit 2nd person singular, masculine and feminine, another child is invited into the room, and the subject leaves the room with one experimenter. The experimenter and the other child in the room are engaged in two different activities which they stop when the subject comes back in. When they stop, however, they leave the toys they were using nearby, so the subject can guess and tell them who performed which activity. For example, the experimenter tells the child: ‘‘One of us drew and one of us dressed up. Guess who jumped and who dressed up.’’ The target here is at ciyart ve-ata hitxapasta/ hu hitxapes ‘‘you [fem.] drew [2nd fem] and you [masc] dressed-up [2nd masc]/he dressed-up [masc].’’

9.5.4 Categories of Analysis Data analysis was conducted for each child separately looking first at the frequency and percentage of correct responses in both languages, and then comparing the success on the different tasks within each language and across the two languages. This made it possible to see individual differences between the two languages and across tasks. This quantitative analysis was followed by a qualitative analysis looking at the type of errors produced by the different children and their frequency. The individual analysis was followed by an attempt at a group profile for the 6 ATD (atypically developing in both languages) children in comparison with the 3 ETD (typically developing in English and atypically developing in Hebrew) children and finally with the 3 children who attended regular preschools. T-tests were used when applicable in the quantitative analysis and the group profiling, as indicated in the results section.

9.6 Results The major finding is that children in the ATD group are not different from monolingual SLI children. Most of their errors in English are characterized by the use of Root Infinitives instead of inflected forms, and in Hebrew they substituted 1st person for 2nd person forms, and showed a marginal use of verb forms which are atypical of the target language. The ETD children performed generally better than the ATD children and often similar to TD bilinguals, apart from the Hebrew enactment task, where the ETD children tended to omit person morphology. Figure 9.1 presents a comparison of percent success in English and Hebrew combining performance on all the tasks, for each of the nine children from the

9

Verb Inflections as Indicators of Bilingual SLI: Qualitative Vs. Quantitative . . .

189

100% 80% 60% 40% 20% 0% Em505 Mf602 Ef601 Cm506 Bm508 Rf606 Ym704 Of602 Af609 Dm606 Am506 Sm506 TD

E-TD

A-TD

English

Hebrew

Fig. 9.1 Percentage correct scores for individuals across all tasks for English and Hebrew

language preschools and for the three children from the regular preschools. These findings for individual children also indicate the group to which each one has been assigned, based on standardized test scores, a practice followed in other figures as well. Figure 9.1 shows that the TD children, as well as two of the ETD children and two of the ATD children exhibit similar performance in both languages. One ETD child shows preference for his L1 English, and the remaining four ATD children show a higher success level in their L2 Hebrew, which is the language used at school and for clinical treatment. Moreover, only the TD children showed similar success in both languages in over 90% of their responses. The one ETD child who nears this level of success is Rf606, a girl who at the age of 6;06, after 4 years of exposure to Hebrew, three in a language preschool, barely comes to the 80% level of correct responses. The other children who showed balanced performance were doing this at a success level of 40–60%. No correspondence was found between total percent correct in either language and length of exposure to Hebrew (2–4 years), nor between age and total level of success in either language. This finding, combined with the fact that the inflections investigated here are acquired by the age of three in monolingual children and pose no problem for typically developing bilinguals after 2 years of exposure, reflects a difference for bilingual SLI children even after 4 years of exposure. This difference for bilingual SLI children is in fact parallel to the data from monolingual SLI children of the same age, who show the similar difficulties for these inflections. Thus, the level of success on these tasks reflects the degree of severity of the impairment rather than chronological age or length of exposure to L2. In order to clarify the difference between the children who attended regular preschool (TD children) and the children in the language preschools (ATD and ETD children), more detailed findings for each of the tasks are presented, first for the TD children in order to show the pattern of typical acquisition. These data will be followed by a presentation of the data of the children in the

190

S. Armon-Lotem et al.

language preschools. Findings are presented for each child separately giving the percent of correct responses for each of the tasks in the different languages. The quantitative data are followed by a qualitative presentation of the types of errors and their frequency.

9.6.1 Typically Developing Bilinguals Overall TD bilingual children performed at a level of 80–100% on all five tasks. Figure 9.2 presents the percent of correct responses out of the total number of responses for each of the tasks for the three TD children. The total number of target responses appears below the label for each of the tasks. Figure 9.2 shows that the TD children scored at ceiling on the imitation task in English and on the enactment task in Hebrew, and achieved an 80–90% success level in the other tasks. Moreover, the data also suggest that among TD bilinguals length of exposure may be relevant in the Hebrew sentence completion task, since the lowest scores belong to Em505 who had only 2 years of exposure to Hebrew at the time of testing. The distribution of error types varied from one task to the other and from one language to the other. While imitation in English yielded no errors, the sentence completion task yielded a small number of tense errors (in 5 of the 48 forms which were inflected for tense), and three (3) omissions of tense or 3rd person inflections, resulting in root infinitives. There was also one (1) instance of overgeneralization of 3rd person morphology to the plural. In Hebrew, on the other hand, there were only three (3) instances of tense errors out of 126 inflected verbs and only one use of an infinitival form. A more frequent error was the use of masculine plural instead of feminine plural in 5 of the 12 cases in which the sentential subject was a feminine plural. It is worth noting, however, 100% 80% 60% Em505

40%

Mf602 Ef601

20% 0% Sentence completion English N = 21

Imitation English N = 26

Sentence completion Hebrew N = 42

Imitation Hebrew N = 40

Enactment Hebrew N = 24

Fig. 9.2 Correct responses as a percentage of the total number of responses for individuals

9

Verb Inflections as Indicators of Bilingual SLI: Qualitative Vs. Quantitative . . .

191

that in Hebrew the masculine plural form of the verb often serves as the neutral unmarked form both for masculine and feminine subjects. Though there were no errors in Hebrew person inflection in the enactment task, substitutions of 1st and 2nd person morphology were found in the imitation task. These errors showed the following distribution: 1 error out of 24 stimulus sentences for 1st person singular, 3 errors out of 24 sentences for 2nd person masculine singular, 2 errors out of 24 sentences for 2nd person feminine singular, and 7 errors out of 24 sentences for 2nd person plural. No errors were found for 1st person plural. In sum, the TD data shows very few errors which involve tense substitutions or omissions in both languages, but difficulty emerged in the use of 2nd person morphology in the Hebrew imitation task.

9.6.2 Language Impaired Children: Quantitative Analysis The bilingual children in the language preschools showed variable performance on the different tasks, within each language and across the two languages, with up to an 85% error rate on some of the tasks for some of the children, regardless of the length of exposure to Hebrew. Figure 9.3 presents the percent correct responses out of the total number of responses for each of the tasks for these nine children. The total number of responses appears below the label for each of the tasks. Figure 9.3 shows first and foremost the variability across languages and tasks, as evidenced by the relatively better performance in Hebrew and in the enactment task in that language. In the Hebrew enactment task, six children reached a success level of 80% or higher, but only three of these children scored that high in the Hebrew sentence completion task. In the English imitation task, only two children scored this high, and only one child reached this level for the English sentence completion task. 100%

Cm506 Bm508

80%

Rf606 60%

Ym704 Of602

40%

Af609 Dm606

20%

Am506 0% E Sentence completion N = 21

E Imitation N = 26

H Sentence completion N = 42

H Imitation N = 40

H Enactment N = 24

Sm506

Fig. 9.3 Correct responses as a percentage of the total number of responses for individuals

192

S. Armon-Lotem et al. Sentence completion English N = 21

100% 80%

Imitation English N = 26

60%

Sentence completion Hebrew N = 42

40% 20%

Imitation Hebrew N = 40

50 6

Enactment Hebrew N = 24

Sm

50 6 Am

Af 60 9 D m 60 6

f6 02 O

70 4 Ym

f6 06 R

50 8 Bm

C

m

50 6

0%

Fig. 9.4 Correct responses as a percentage of the total number of responses for tasks

Figure 9.4 presents the same results for each child separately. Figure 9.4 shows the within subject variability for the language impaired children, as reflected in the different height of the bars for each child. Only one girl (Rf606, age 6;06 with 4 years of exposure to Hebrew, three in a language preschool) achieved an 80% success level on all the tasks. Another girl (Af609, age 6;09 with 2 years of exposure to Hebrew in a language preschool) achieved this level in the three Hebrew tasks, and the oldest boy (Ym704, age 7;04 with 3 years of exposure to Hebrew, two in a language preschool) achieved this level in two of the three Hebrew tasks. Five additional children achieved this level in one task only, which was enactment in four of five cases. All ETD children scored notably better on the English imitation task (two achieving over 80% success level), and were on average better on the English sentence completion task. On the other hand, two of the three received the lowest scores on the Hebrew enactment task.

9.6.3 Language Impaired Children: Qualitative Analysis – Type and Number of Errors In addition to the frequency of errors which distinguishes typically developing from language impaired children, the type and distribution of errors also characterizes the differences between these groups. Two types of errors were found in both languages (1–2), two were unique to English (3–4) and three were unique to Hebrew (5–7), as follows: Errors found in both languages 1. Wrong tense (past for present or present for past): Here the boy jumps and the girl jumped [ jumped instead of jumps]. Kan ha-yeled kofec ve-kan hayeladot kafcu. [kafcu for kofcot]. 2. Root infinitives: The cat hops and the dog hop (hop instead of hops). ha-xatula kfca ve ha-kelev likfoc [likfoc for kacaf]. Errors found in English only

Verb Inflections as Indicators of Bilingual SLI: Qualitative Vs. Quantitative . . .

9

193

3. Error in 3rd person morphology with a plural subject: The cat hops and the dogs hops [hop]. 4. V-ing without an auxiliary: The cat hops and the dog hopping [hops]. Errors found in Hebrew only 5. Gender error (mostly masculine substituted for feminine): Ha-yeled kafac ve ha-yalda kafac [kafac for kafca]. Ha-kelev mexapes oxel ve ha-xatulot mexapsim oxel [mexapsim for mexapsot]. 6. Number error (mostly singular for plural): Ha-yeled kafac, ha-yalda kafca, kulam kafac. [kafac for kafcu]. b’yom ha-huledet cilamta et ha-buba. [cilamta for cilamtem] 7. Person error (1st person for 2nd person or vice-versa): etmol kafacti baxevel harbe zman [kafacti for kafact]. b’yom ha-huledet cilamta et ha-buba [cilamta for cilamti] Figure 9.5 presents frequencies and distribution of errors for each child in the sentence completion task in English, and Fig. 9.6 presents the distribution of errors for each child in the sentence imitation task in English. Both tasks test the use of past tense morphology and 3rd person morphology in the present tense. Figures 9.5 and 9.6 show that Root Infinitives are the major source of errors in English, and they are mostly concentrated among the 6 ATD children. Two of these six children also make use of verbs inflected with –ing but without the tense carrying auxiliary. Both forms are non-finite. Additional errors which were found in the English of the TD bilinguals are the use of 3rd person morphology with plural subjects and wrong use of tenses. Figure 9.7 presents in raw numbers the distribution of errors for each child in the sentence completion task in Hebrew, which tests the use of 3rd person masculine and feminine, singular and plural in the past and the present tenses. Figure 9.7 shows that in Hebrew, unlike English, the infinitival form of the verb was used only one time, while all other forms used by the children were finite and inflected, with no difference between the ATD and ETD children. Wrong use of inflection was found for all children, and the most frequent errors

16 14 12 10 8 6 4 2 0

Root Infinitives [N = 15] Wrong 3rd person [N = 6] Wrong tense [N = 21]

6 50 Sm

50 6 Am

60 6 m D

60 9 Af

02 f6 O

70 4 Ym

f6 06 R

50 8 Bm

C

m

50 6

V+ing [N = 21]

Fig. 9.5 Distribution of errors in English sentence completion per individual

194

S. Armon-Lotem et al.

16 14 12 Root Infinitives [N = 16] Wrong 3rd person [N = 10]

10 8

Wrong tense [N = 26]

6

V+ing [N = 26]

4 2 0

Cm506 Bm508 Rf606 Ym704 Of602 Af609 Dm606 Am506 Sm506

Fig. 9.6 Distribution of errors in English sentence imitation per individual

12 10 8

Wrong number [N = 42]

6

Wrong gender [N = 42] Wrong tense [N = 42]

4

Root Infinitive [N = 42]

2 0 Cm506 Bm508 Rf606 Ym704 Of602

Af609 Dm606 Am506 Sm506

Fig. 9.7 Distribution of errors in Hebrew sentence completion per individual

were the use of wrong tense and the use of the plural masculine rather than the plural feminine form of the verb in the present tense. Both errors, however, were also found among the typically developing children. A unique finding in this task was combinations of inflections (and conjugations, but this is beyond the scope of the present paper) which are impossible in Hebrew and are rarely found in typical development, but are reported for monolingual SLI (Blass, 2000, Dromi & Blass, 2002). Such errors demonstrate a mixing of a present prefix m- with the number suffix of the past tense -u as in mitgalshu instead of mitgalshim, use of past form of the verb with present tense gender suffix -et hisreket instead of histarka, or even an inflection which does not exist in Hebrew at all as the –e in hitlabshe instead of helbish. Errors of this type were restricted to the more complex conjugation hif’il and hitpa’el and some of these forms, e.g., the two latter ones, also involved use of the wrong conjugation, which might hint at a problem at the derivational level, in addition to the inflectional error. Figure 9.8 presents in raw numbers the distribution of errors for each child in the imitation task in Hebrew, and Fig. 9.9 presents in raw numbers the distribution of errors for each child in the enactment task in Hebrew. Both tasks test the use of 1st and 2nd (masculine and feminine) person morphology in the

Verb Inflections as Indicators of Bilingual SLI: Qualitative Vs. Quantitative . . .

9

195

25 20 2nd plural 1st plural 2nd feminine 2nd masculine

15 10

1st singular 5 0 Cm506 Bm508

Rf606

Ym704

Of602

Af609

Dm606 Am506 Sm506

Fig. 9.8 Distribution of errors in Hebrew sentence imitation by target morpheme [N¼8] per individual

25 20 15

2nd feminine 2nd masculine

10

1st singular

5 0 Cm506 Bm508

Rf606

Ym704

Of602

Af609

Dm606 Am506 Sm506

Fig. 9.9 Distribution of errors in Hebrew enactment by target morpheme [N¼8] per individual

past tense and the imitation task also tests the use of singular vs. plural forms in 1st and 2nd person. Both figures present the number of times the child did not use the target form. The type of errors is discussed separately. Figure 9.8 demonstrates that there are errors in all person inflections in the past tense. Similar to the TD children, most errors are in 2nd person singular and plural. Nonetheless, the number of errors is much higher than is found for the TD children, with 4 of the children not using the 2nd person plural at all. In most of the errors, the children use 1st person singular instead of 2nd person, while a fewer errors, which were mostly found among the ETD children, involved the use of the 3rd person masculine form which has no overt person morphology. Figure 9.8 also shows the variability across subjects in the number of errors. Though more errors were found amongst the ATD children, the ETD children do not score at ceiling either. Figure 9.9 presents the number of times the child did not use the target form in the Hebrew enactment task.

196

S. Armon-Lotem et al.

While in previous tasks the ETD children showed better or similar performance when compared to the ATD children, the enactment task was most difficult for the ETD children, with almost no errors among the ATD children (except Dm606). The ETD children more often than not used the past 3rd person masculine singular form which is not inflected overtly, instead of the forms inflected for 2nd person, masculine and feminine, and even, in the case of Cm506 instead of 1st person. To conclude, our findings show that most errors in English are found among the ATD children and are characterized by the use of Root Infinitives instead of inflected forms. In Hebrew, on the other hand, there was almost no use of nonfinite uninflected forms. In the sentence completion task where person morphology was not targeted, there was no clear difference among the different subjects, nor between the TD children and the ATD and ETD children, apart from a marginal use of nonexistent verb forms. Errors were found however in the other two tasks which targeted the use of 1st and 2nd person morphology. In the imitation task, most subjects substituted 2nd person by 1st person, and in the enactment task, the ETD children omitted the person morphology by using the past 3rd person singular form of the verb which has no overt person morphology.

9.7 Discussion Studying the use of verb inflections by bilingual language impaired children who attend language preschool compared with TD bilingual children, the present study found that language impaired children tend to omit tense and person morphology in English, while in Hebrew omission is less frequent, and these same children opt for substitution of inflections, in particular in 2nd person in the past tense. This finding is not different from what we know about monolingual SLI children neither in the type of errors nor in the quantity of errors typical for 5–6 years old. This seems to suggest that being bilingual has no negative impact on the acquisition of the inflectional system. These same errors are also characteristic of typically developing bilingual children, as has been noted in the literature, but there is a quantitative difference between the two populations which can be indicative of SLI among bilinguals. Comparing the children’s two languages shows that while TD children show similar high proficiency in both languages, only some of the children in the language program showed lowered performance in both languages. Of the 6 ATD children who had low scores on the standardized tests in both languages, only two showed similar overall performance in the two languages, whereas the other four were much better in Hebrew, their L2. Of the 3 ETD children, who seem to have unimpaired knowledge of Hebrew, only one showed better knowledge of Hebrew, while another one showed balance low performance in both languages and one showed better performance in his L1 English. This finding diverges from Ha˚kansson et al. (2003) who found balanced lowered performance in both languages to be typical of language impaired children, while

9

Verb Inflections as Indicators of Bilingual SLI: Qualitative Vs. Quantitative . . .

197

typically developing children showed typical development at least in one language. Our findings suggest that even among the ATD children, one language can be more dominant than the other, resulting in an unbalanced performance, while even some of the ETD children, who by Ha˚kansson et al. might not be considered impaired, might show low balanced performance in both languages making errors which are uncharacteristic of typical bilingual children. This unbalanced performance in the two languages of the children in the language preschools raises the question whether these children, and in particular those who scored within the norm in their L1, can really be considered SLI. While one of the nine children, Rf606, showed balanced performance and level of success which is close to that of the TD children, the other 8 children seem to struggle with the use of the inflectional system either in both languages or in one of them, and their errors are often characteristic of SLI monolingual children. This seems to suggest that having one dominant language does not rule out the possibility of language impairment. Nonetheless, a difference does emerge between the ATD children and the ETD children in the nature of the errors in Hebrew, questioning the nature of the impairment in the ETD population. The nature of the impairment in the ETD population is further questioned by their variable performance on the different tasks. All of the children’s performance was tested by different tasks which engage somewhat different processing abilities. Among the TD children, there was no clear difference among the tasks. In Hebrew, the sentence completion task was slightly more difficult compared to the other two tasks. In this task, there seemed to be a relation between the age and length of exposure and the level of success, as evident from the score of Em505. In English, there was no difference among the tasks, though again sentence completion was more difficult for the younger child. Nonetheless, the rate of success in all tasked was over 80% for the TD children. For the ATD children, there was almost no difference between the two English tasks, both in the rate of errors and type of errors, while the ETD showed better performance on the imitation task than on the sentence completion task, with very few errors on the former. In Hebrew, on the other hand, the sentence completion task yielded very little variability across children, with ATD and ETD performing at a similar level. The other two tasks, imitation and enactment, showed more variability across children. Four of the ATD children found the imitation task, which requires both syntactic processing and the use of echoic memory, more difficult, while 2 children (one ATD and one ETD) found both tasks as challenging, 2 children (one ATD and one ETD) scored over 80% on both tasks, and only one ETD child found the enactment task, which requires generation of the person inflection with no previous modeling, more difficult. These findings do not support one of the tasks as a better diagnostic tool, but point out the heterogeneity of the impaired population and the possible different causes of the impairment. In particular, for the ETD children, the difficulty found in enactment for two of the children, and the difficulty in imitation for one of them, seem to suggest that they might have

198

S. Armon-Lotem et al.

some language related impairment, even if it might not be labeled SLI. Clearly, more subjects should be studied to test this hypothesis.

9.8 Conclusion This paper studied the way English-Hebrew sequential bilingual children in language preschools use the verb inflectional system in Hebrew and English, as part of a larger project to assess the relative contribution of the bilingual representation and the SLI impaired processing to the linguistic performance of bilingual SLI children. We found that the bilingual representation is not manifested equally in the two languages, and has no negative influence on performance in either language. The difficulty found within the inflectional system, which is typical of monolingual SLI children was also found in the bilingual ATD children at a degree which is typical of the SLI population. Conspicuously, in English there was no qualitative difference between the typically developing population and the population in the use of the inflectional system in terms of type of errors. However there was a difference in the quantity of errors. In Hebrew on the other hand, there was also a qualitative difference in the type of errors, since some children who had typically developing English (ETD) omitted person inflections, while some atypically developing children produced non-existing verbal morphology. One last issue which should be addressed here is the question of wrong diagnosis. Of the 9 children studied here, 3 children scored within the norm in their L1, but were nonetheless placed in a language program, probably due to their scores in their L2 which were below the norm. Two of these 3 children present a linguistic profile which is distinct from the TD children, while 1 child is similar. This issue can not be sorted out by a set of case studies and calls for further studies of more children in language preschools and regular preschools to assess the validity of our findings.

References Albertini, J. (1980). The acquisition of five grammatical morphemes: Deviance or delay? Proceedings from the Wisconsin Symposium on Research in Child Language Disorders, 1, 94–111. Beastrom, S., & Rice, M. (1986, November). Comprehension and production of the articles a and the. Paper presented at the Convention of the American Speech-Language-Hearing Association, Detroit. Bishop, D. V. M. (1994). Grammatical errors in specific language impairment: Competence or performance. Applied Psycholinguistics, 15(4), 507–550. Bishop, D. V. M., & Snowling, M. (2004). Developmental dyslexia and specific language impairment: Same or different? Psychological Bulletin, 130(6), 858–886. Blass, A. (2000). The inflectional system of Hebrew speaking SLI children. MA Thesis, Tel Aviv University (in Hebrew).

9

Verb Inflections as Indicators of Bilingual SLI: Qualitative Vs. Quantitative . . .

199

Boscolo, B., Ratner, N. B., & Rescorla, L. (2002). Fluency of school-aged children with a history of specific expressive language impairment. American Journal of Speech-Language Pathology, 11, 41–49. Clahsen, H. (1991). Child language and developmental dysphasia. Linguistic studies of the acquisition of German. Amsterdam: Benjamins. Crago, M., & Gopnik, M. (1994). From families to phenotypes: Theoretical and clinical implications of research into the genetic basis of specific language impairment. In R. Watkins & M. Rice (Eds.), Specific language impairments in children (pp. 35–51). Baltimore, MD: Paul H. Brookes. Dromi, E., & Blass, A. (2002). The spontaneous use of Hebrew verb forms by Israeli preschool children with and without SLI. Special issue on Unity and Diversity in Communication, L. Scliar-Cabral (Ed.). Journal Ilha do Desterro. A Journal of English Language, Literature in English and Cultural Studies. Dromi, E., Leonard, L., & Shteiman, M. (1993). The grammatical morphology of Hebrewseaking children with specific language impairment. Journal of Speech and Hearing Research, 36, 760–771. Dromi, E., Leonard, L. B., & Adam, G. (1997). Verb agreement morphology in Hebrewspeaking children with specific language impairment. Journal of Speech, Language, and Hearing Research, 42, 1414–1431. Dromi, E., Leonard, L. B., Adam, G., & Zadunaisky-Ehrlich, S. (1999). Verb agreement morphology in Hebrew-speaking children with specific language impairment. Journal of Speech, Language and Hearing Research, 42, 1414–1431. Eyer, J., & Leonard, L. (1995). Functional categories and speci1 c language impairment: a case study. Language Acquisition, 4, 177–203. Fletcher, P., & Garman, M. (1988). Normal language development and language impairment: syntax and beyond. Clinical Linguistics & Phonetics, 2, 97–113. Fridmann, N., & Novogrodsky, R. (2004). The acquisition of relative clause comprehension in Hebrew: A study of SLI and normal development. Journal of Child Language, 31(3), 661–681. Genesee, F., Paradis, J., & Crago, M. (2004). Dual language development and disorders: A handbook on bilingualism and second language learning. Baltimore, MD: Brookes. Gopnik, M., & Crago, M. (1991). Familial aggregation of a developmental language disorders. Cognition, 39, 1–50. Goralnik, E. (1995). Goralnik diagnostic test. Matan, Even Yehuda, Israel (in Hebrew). Hadley, P. A., & Rice, M. L. (1996). Emergent uses of BE and DO: Evidence from children with specific language impairment. Language Acquisition, 5(3), 209–243. Ha˚kansson, G., Salameh, E.-K., & Nettelbladt, U. (2003). Measuring language proficiency in bilingual children. Swedish-Arabic bilingual children with and without language impairment. Linguistics, 41(2), 255–288. Jacobsen, P. F., & Schwartz, R. G. (2005). English past tense use in bilingual children with language impairment. American Journal of Speech Language Pathology, 14(4), 313–323. Leonard, L. B. (1998). Children with specific language impairment. Cambridge, MA: MIT Press. Leonard, L., & Dromi, E. (1994). The use of Hebrew verb morphology by children with specific language impairment and by children developing language normally. First Language, 14, 283–304. Leonard, L. B., Sabbadini, L., Volterra, V., & Leonard, J. S. (1988). Some influences on the grammar of English- and Italian-speaking children with specific language impairment. Applied Psycholinguistics, 9(1), 39–57. Moats, L. C., & Smith, C. (1992). Derivational morphology: Why it should be included in language assessment and instruction. Language, Speech, and Hearing Services in Schools, 23, 312–319.

200

S. Armon-Lotem et al.

Novogrodsky, R., & Fridmann, N. (2006). The production of relative clauses in syntactic SLI: A window to the nature of the impairment. International Journal of Speech-Language Pathology, 8(4), 364–375. Owens, R. (2001). Language development: An introduction (5th ed.). Boston, MA: Allyn and Bacon. Paradis, J. (2005). Grammatical morphology in children learning English as a second language: Implications of similarities with specific language impairment. Language, Speech and Hearing Services in the Schools, 36, 172–187. Paradis, J. (2007). Bilingual children with specific language impairment: Theoretical and applied issues. Applied Psycholinguistics, 28(3), 551–564. Paradis, J., Crago, M., & Genesee, F. (2005/2006). Domain-general versus domain-specific accounts of specific language impairment: Evidence from bilingual children’ acquisition of object pronouns. Language Acquisition, 13(1), 33–62. Paradis, J., Crago, M., Genesee, F., & Rice, M. (2003). Bilingual children with specific language impairment: How do they compare with their monolingual peers? Journal of Speech, Language and Hearing Research, 46, 1–15. Ravid, D., Levie, R., & Avivi-Ben Zvi, G. (2003). Morphological disorders. In L. Verhoeven & Hans van Balkom (Eds.), Classification of developmental language disorders: Theoretical issues and clinical implications (pp. 235–260). Mahwah, NJ: Erlbaum. Rice, M., & O¨tting, J. (1993). Morphological deficits of SLI children: Evaluation of number marking and agreement. Journal of Speech and Hearing Research, 37, 1249–1257. Rice, M. L., & Wexler, K. (1996). Toward tense as a clinical marker of specific language impairment in English-speaking children. Journal of Speech and Hearing Research, 39, 1239–1257. Rice, M. L., & Wexler, K. (2001). Rice/Wexler test of early grammatical impairment. San Antonio: The Psychological Corporation. Steckol, K., & Leonard, L. (1979). The use of grammatical morphemes by normal and language-impaired children. Journal of Communication Disorders, 12, 291–301. Steenge, J. (2006). Bilingual children with specific language impairment: Additionally disadvantaged? PhD dissertation, Radboud University, Nijmegen. Tallal, P., & Stark, R. E. (1981). Speech acoustic cue discrimination abilities of normally developing and language impaired children. Journal of Acoustic Society of America, 69(2), 568–574. van der Lely, H. K. J. (1998). SLI in children: Movement, economy and deficits in the computational-syntactic system. Language Acquisition, 7, 161–192. Wiig, E. H., Secord, W. A., & Semel, E. M. (2004). Clinical evaluation of language fundamentals – preschool 2. San Antonio, TX: Harcourt/Psych Corp.

Chapter 10

Procedural and Declarative Memory in the Acquisition of Morphological Knowledge: A Model for Second Language Acquisition in Adults Sara Ferman and Avi Karni

10.1 Introduction A leading tenet in neurobiological theories of learning and memory is that two independent neural systems subserve long-term memory: the declarative and procedural memory systems (e.g., Mishkin, Malamut, & Bachevalier, 1988; Squire & Zola, 1996). The declarative system has been implicated in the learning and subsequent use of knowledge about novel events and facts (‘‘what’’). This type of memory can be established even following a single exposure, can be explicitly recollected, but may be rapidly degraded. The procedural memory system has been implicated in the learning and retention of skills (‘‘how to’’) and habits and its establishment necessitates a critical amount of repetitions (practice) and time (e.g., Hauptmann & Karni, 2002; Karni, 1996). The establishment of procedural memory is sometimes conceptualized as implicit learning, the acquisition of complex structured knowledge independently, to a large degree, of awareness of both the processes and products of acquisition. Explicit learning, in this view, is a more conscious and intentional operation whereby the individual remembers the making and testing of hypotheses concerning ongoing events (Cleeremans, Destrebecqz, & Boyer, 1998; Cohen & Poldrack, 1997; Reber, 1989). Although the two systems are largely independent of each other, they may interact in a number of ways. For example, recent studies suggest that declarative knowledge may turn into procedural knowledge (proceduralization of declarative knowledge) and procedural (implicit) knowledge may be converted into declarative knowledge as a result of accumulating experience (e.g., Anderson et al., 2004; Bitan & Karni, 2003; Logan, 1988; Sun & Zhang, 2004). A growing body of evidence suggests that language learning and language use involve cognitive processes that are not unique to language (e.g., Bates, Devescovi, & Wulfeck, 2001; Elman et al., 1996; Hauser, Chomsky, & Techman Fitch, 2002) and, specifically, that linguistic learning may share some features S. Ferman (*) Department of Communication Disorders, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel e-mail: [email protected] M. Leikin et al. (eds.), Current Issues in Bilingualism, Literacy Studies 5, DOI 10.1007/978-94-007-2327-6_10, Ó Springer ScienceþBusiness Media B.V. 2012

201

202

S. Ferman and A. Karni

with non-linguistic skills acquisition (DeKeyser, 2001; Robinson, 1996; Robinson & Ha, 1993; Schmidt, 1993; Segalowitz, 2003). There are, however, very few studies of language skill learning that were explicitly designed to investigate whether and which of the key features of non-linguistic procedural memory generation are manifest in language skill acquisition. Several second or artificial language acquisition studies showed that fluency in linguistic usage can be achieved in a manner akin to the proceduralization of non-linguistic skills (Johnson, 1996; Robinson, 1996; Robinson & Ha, 1993; Schmidt, 1993; Segalowitz, 2003). For example, in a seminal study, DeKeyser (1997) reported that (artificial) morphological learning performance improved gradually over time, with the group-average learning curves characterized by a power function, very similar to that characterizing the acquisition of non-linguistic skills (DeKeyser 1997; Ellis & Schmidt, 1998). Linguistic skill learning was sometimes described either as the top-down proceduralization of linguistic declarative knowledge (that was explicitly taught in class) – in line with general models of skill learning, or as a bottom-up process wherein implicit language knowledge gradually becomes explicit (e.g., Bialystok, 1982). Ullman (2001, 2004) has recently proposed a dichotomous declarativeprocedural model of language learning and retention. Ullman’s model is in line with dual models which posit that the lexicon and grammar are separable and subserved by distinct cognitive systems. According to these models, lexical forms are stored in a rote memory store (the mental lexicon), and grammar is a set of abstract rule-like mental operations (e.g., Chomsky, 1995; Pinker, 1994, 1999). Ullman’s model posits that the mental dictionary is part of the declarative memory system, while the learning, retention, and usage of grammatical rules are dependent upon procedural memory. The parsing of linguistic knowledge retention between the two memory systems may, however, be more complex than a one-to-one linkage of lexical acquisition to declarative memory and grammar learning to procedural memory. The lexicon and morphology, for example, may interact with each other rather than act independently (Bates & Goodman, 1997; Dionne, Dale, Boivin, & Plomin, 2003; Devescovi et al., 2005). In addition, given that words, as well as morphological knowledge, convey meaning through arbitrary phonological patterns (i.e., articulatory or sensory sequences), there is no a priori reason to discount the possibility that the retention of phonological aspects of both words and morphological rules may depend upon procedural memory. Ullman’s model would seem to ascribe the retention of meaningful sequences (lexicality – in Ullman’s example, ‘‘words, including their sounds’’) to declarative memory. However, a well-rehearsed sound pattern can become represented in procedural memory (Ari-Even Roth, Kishon-Rabin, Hildesheimer, & Karni, 2005). Moreover, several studies of artificial grammar learning indicate that lexicality judgment can be based on implicit learning from a rather limited set of exemplars (e.g., Reber, 1989). Finally, declarative knowledge may turn into procedural knowledge and viceversa as a result of accumulating experience (e.g., Anderson et al., 2004; Bitan & Karni, 2003; Logan, 1988; Sun & Zhang, 2004).

10

Procedural and Declarative Memory in the Acquisition . . .

203

10.2 Method In our study (Ferman, Olshtain, Schechtman, & Karni, 2009) we followed, in a fine-grained manner, the evolution of the knowledge of an artificial morphological rule (AMR) specifically, the memory systems sub-serving the learning of a linguistic skill by testing whether linguistic skill learning share common characteristics with non-linguistic learning. Artificial grammar learning paradigms have been used to investigate the contribution of procedural (implicit) and declarative (explicit) memory systems to the acquisition of specific items and to the generalization of ‘‘rules’’ to new exemplars (e.g., Gomez, 1997; Knowlton & Squire, 1996; Mathews et al., 1989; Pothos, 2007; Reber, 1989). Others have used miniature artificial language paradigms to investigate the cognitive and language processes underlying language learning (e.g., Braine, Brody, & Brooks, 1990; Dahan & Brent, 1999; DeKeyser, 1997; Ellis & Schmidt, 1998; MacWhinney, 1983; Robinson, 1996). Artificial language resembles natural languages in that it contains artificial linguistic rules, artificial words or both, which obey natural language constraints. The advantages of using an artificial language lie in ensuring that the learned knowledge and the exposure to input, instructions, and tasks are equal for all learners, as well as in affording fine-grained data collection that cannot be achieved in real life language learning situations. On the other hand, one may argue that the simplified language and laboratory environment in artificial language paradigms may not express the complexity of natural language or of real-life learning conditions. These arguments, however, express a classic dilemma that inevitably arises as the price of experimental control in laboratory research (e.g., DeKeyser, 1997; Hulstijn, 1997). The artificial morphological rule (AMR) was designed to be analogous to the morphological rules of Hebrew grammar (the participants’ native language) by using a phonological change that corresponded to the phonological constraints of Hebrew. The AMR expressed a semantic distinction that the Hebrew language does not contain but that, nevertheless, exists in some manner in other natural languages. The AMR required a specific differential phonological marking for Hebrew verbs depending on whether the preceding noun (the subject) was animate or inanimate. The suffix ev was to be added to verbs used with animate nouns, and the suffix ar was to be added to verbs used with inanimate nouns. Also in accordance with the phonological rules of Hebrew, all transformed verbs underwent omission of the vowel that preceded the added suffix and hence the stress shifted to the added suffix. The AMR was applied to grammatical Hebrew noun-verb sentences (items) so that the meaning of the phrases could be easily understood. Repeated items of the AMR were used as the means to explore the development of specific item knowledge, and non-repeated, novel, items as the means to investigate the evolution of the knowledge required for the generalization of the AMR to items not previously encountered. We studied the evolution of the knowledge of the AMR across multiple learning sessions using a training

204

S. Ferman and A. Karni

paradigm similar in its intensity and design to those recently used for investigating the acquisition of perceptual, cognitive and motor skills. Eight healthy youngadult volunteers (four females, four males), high-school and college graduates of middle-class background, 18–25 years of age (mean, 21:5 years) participated in this study. All were native speakers of Hebrew with no reported history of speech, language, learning, or hearing difficulties. Each participant was trained individually and took part in 10 consecutive training sessions 1–3 days apart, and again in a retention test session about 2 months after the termination of training to assess the long-term retention of the achievements. Training was through exposure to and use of the AMR in the performance of two tasks: a judgment task and a production task. There was no explicit instruction on the nature of the AMR at any point during the training. At the beginning of the first session, using a pre-recorded introduction, each participant was told that he/she was going to learn a new language, similar to Hebrew and then instructed to listen to a list of sentences in which the AMR was applied (the modeling list) in order to learn the ‘‘new language.’’ Instructions on how to respond in each of the two tasks then followed. In each trial of the judgment task, the participant heard: In the new language, is it correct to say . . .? followed by a noun-verb pair, either well-constructed or not. The incorrect option was always a phonologically correct production that was semantically incorrect. The participant was instructed to make a two-alternative forcedchoice response: correct-incorrect, by pressing one of two buttons. In each trial of the production task, the participant heard: In the new language, how should one say . . . followed by a noun-verb pair in standard Hebrew, and then the noun was repeated and a verbal response – speech production of the transformed verb in accordance with the AMR – was required. Answers were written down by the researcher on a specially designed form and an audio taperecording of speech productions was made for off-line analysis. Any deviation from the precise form of the correct transformed target item was counted as an error. The errors in production were sorted according to the locus within the composite ‘‘verb stem–suffix’’ produced: (1) errors in the artificial suffix; (2) errors in the verb stem; (3) errors in both the artificial suffix and the verb stem. In each session, following the presentation of the modeling list, the participants performed four blocks of repeated-item lists (two for each task) and two blocks of new-item lists (one for each task), which were used to test for the ability to generalize the AMR to previously unencountered items. Rest intervals of 2–3 min were given between blocks. In the second and all subsequent sessions, two pre-test lists, one for each task, were administered at the beginning of the session in order to test between-session gains. Each session lasted about 1 h and included 176 repeated items (80 in each task and 16 in the model list), and 32 new items (16 in each task). Because there were no pre-test lists, the first session included the same number of new items but only 144 repeated items. Altogether, there were 2,080 trials per individual participant. Participants were instructed to answer as fast as possible. A recorded auditory feedback signal comprising the word error was given in both tasks following an incorrect answer.

10

Procedural and Declarative Memory in the Acquisition . . .

205

All stimuli and instructions were recorded by a professional radio announcer using the Goldwave software (www.goldwave.com). The length of the items (stimulus duration) was equalized. The SuperLab software package (Cedrus Corporation, www.cedrus.com) was used to run the experimental trials and to time and log responses for each trial using a PC laptop. Stimuli were presented through a headphone set. Speech production answers were recorded by a microphone placed in front of the participant on an adjustable stand and voice onset measured using a Mel Response Box (Psychology Software Tools, Inc). Both accuracy (percent correct) and speed (Reaction Time [RT] in ms; for both manual [judgment task] and voiced [production task] responses) were logged for each single response. To establish whether the learning of the AMR continued between sessions, beyond the within-session improvement, a measure of between-sessions gains and a measure for within-session gains were computed. The between-session gains were calculated as the difference between the final block of a given session and the first block of the following session (delayed gains). The within-session (early) gains were calculated as the difference between the first block and the final block of the session. A solicited verbal report concerning each participant’s insights on the required transformation was used to assess explicit knowledge of the AMR. At the end of each learning session, the participants heard: You are doing very well. How did you get to your answer? Verbal reports were written down on a pre-prepared form as well as recorded using an audio tape-recording system for off-line analysis. Although the introduction of the explicit verbal reports might have affected the learning process in calling for self-awareness and encouraging introspection on the nature of the task, it is an accepted procedure in investigating the involvement of implicit versus explicit knowledge.

10.3 Results Our results showed that practicing the repeated noun-verb items resulted in significant incremental performance gains in both the judgment and production tasks with respect to accuracy and speed. There were large gains in accuracy: from a group mean of 71% and 51% in the first session up to 97% and 99% in the final session for the judgment and production tasks, respectively. As expected based on studies of non-linguistic skill (procedural memory) acquisition (e.g., Dorfberger, Adi-Japha, & Karni, 2007; Karni & Sagi, 1993; Korman, Raz, Flash, & Karni, 2003) there were also large gains in speed: from a group mean of 983 and 1512 ms in the first session down to 73 and 85 ms in the final training session for the judgment and production tasks, respectively (a more than tenfold improvement in both tasks). In line with recent studies of the time-course of procedural memory consolidation in adults and children learning non-linguistic skills (e.g., Ari-Even Roth et al., 2005; Dorfberger et al., 2007; Karni & Sagi, 1993; Korman et al., 2003; Robertson & Cohen, 2006) there were considerable contributions of the

206

S. Ferman and A. Karni

interval between the training sessions to the performance gains (delayed gains). On average, the delayed gains comprised 21% and 54% (in the judgment and production tasks, respectively) of the accuracy gains, and 48% and 19% (in the judgment and production tasks, respectively) of the gains in speed. Thus, a given session of training could induce not only within-session performance gains but also resulted in delayed, between-sessions gains in both tasks, both in terms of accuracy and speed. Our results also clearly showed that there was robust retention of the gains after an interval of 2 months. Indeed, group average performance in the last session did not differ significantly from the group average performance in the retention session both in terms of accuracy in the two tasks and in terms of speed in the judgment task. There was only a minimal (on average a loss of 0.8% of the total gains) difference between the speed performance in the production task in the retention test compared to the one attained in the final training session. Moreover, within 2–3 blocks of the retention test, those participants who did show a drop in performance during the 2-month interval were able to regain the speed (and accuracy) levels of the end of the training set. Also in line with non-linguistic skill learning paradigms, no speed-accuracy trade-off was evident in any of the test conditions. Moreover, a negative correlation between RTs and correct performance percentage was evident and highly significant, in both tasks. The gains in speed developed more slowly than the gains in accuracy, with the slope of the regression line for accuracy (as a function of session number) significantly steeper than the slope of the corresponding regression line for speed. Thus, whereas accuracy reached the ceiling score in the fourth and second sessions for the judgment and production tasks, respectively, speed continued to improve up to the ninth and eighth sessions in these two tasks. The latter part of the learning was characterized by very short reaction times, indicating that participants were able to respond (correctly) even before the target noun-verb pair was completed. The participants’ ability to generalize the AMR to new noun-verb pairs increased significantly as reflected in both accuracy and speed in both tasks. There were large gains in accuracy as well as a more than tenfold improvement for speed in both tasks. A most consistent finding was that all participants were able to pronounce the target items correctly, i.e., they acquired the phonological aspect of the AMR very early on in training: at the end of the second learning session. However, at the same time, when phonological performance was at ceiling, the majority of participants (5/8) failed to discover the semantic distinction and therefore experienced marked difficulties (<60% accuracy) with the semantic–phonological correspondence. The correct pronunciation of the target verbs, irrespective of semantic accuracy, can be considered as a measure for phonological competence. Using this measure, our results clearly indicated that all the participants were able to correctly apply and correctly produce the phonological aspect of the AMR; moreover, this fluency was acquired in a manner very much indicative of procedural memory (skill learning). Even when the participants pronounced the new verbs correctly, explicit reports on the

10

Procedural and Declarative Memory in the Acquisition . . .

207

source of their success were at best partial, i.e., the production was better than the explicit knowledge thereof. Moreover, even after the discovery of the semantic aspect, and despite correct pronunciation, none of the participants explicitly reported the vowel omission in the basic verb and the stress shift to the artificial suffix. At different time-points between the first and the eighth learning sessions, 7/8 participants spontaneously (explicitly) reported the role of the semantic distinction between animate and inanimate causative nouns as the basis of the AMR. The discovery of the semantic distinction was often very abrupt. For example, during the eighth session, participant Z.A. suddenly said: Oh, now I know! For every inanimate noun the verb terminates in r and for every human one (the verb) terminates in v. At about the same time, the gains in accuracy on new items in both the judgment and production tasks increased abruptly in all seven participants who were able to report the semantic aspect of the AMR. The performance of the one participant who was unable to explicitly report the semantic distinction did not exceed 62% accuracy on the new items in both tasks throughout the training period although significant gains in speed were found and phonological production was perfect. Further training after the moment of the semantic discovery resulted in additional accuracy gains (7/8 participants reached >90% accuracy) and large gains in speed on new items in both the judgment and production tasks, indicating a clear and highly effective proceduralization phase. Thus, by the end of training, the level of accuracy and fluency attained in the performance of new items in the judgment and production tasks by the seven participants who were able to discover the semantic aspect of the AMR was comparable to that achieved for repeated items. These gains too were robustly retained over an interval of 2 months without training.

10.4 Discussion Our findings suggest that the ability to judge and produce repeated items of an AMR demonstrated, in adults, critical characteristics of procedural learning and developed in a manner very similar to the one recently described for the acquisition of non-linguistic skills (Ari-Even Roth et al., 2005; Hauptmann & Karni, 2002; Karni, 1996; Korman et al., 2003). The group average performance on repeated items of the AMR, as a function of the number of repetitions (sessions), was found to be well characterized by a power function model for both speed and accuracy measures in both tasks, closely resembling the learning curves characterizing a wide range of non-linguistic skill learning. Moreover, as in non-linguistic skill acquisition, our data indicated that some of the gains accrued in training evolved between sessions, and not concurrently with the training experience (within session). Such time-related gains in performance (occurring after practice has terminated) have been interpreted, in the context of perceptual and motor skills learning, as reflecting memory consolidation

208

S. Ferman and A. Karni

processes in the neuronal system (e.g., Ari-Even Roth et al., 2005; Karni & Sagi, 1993; Maquet et al., 2003; Walker & Stickgold, 2006). Practicing repeated items resulted in large gains in accuracy and speed in both tasks with no speed-accuracy trade-off. Participants attained almost perfect accuracy scores in the third session and their fluency (speed) increased more than tenfold over the study sessions. These robust gains in both speed and accuracy fit well with the results of studies on non-linguistic skill learning as well as other studies on second language acquisition (Johnson, 1996; Robinson, 1996; Robinson & Ha, 1993; Schmidt, 1993; Segalowitz, 2003), suggesting that extensive experience can result in increasing ‘‘automatization’’. The finding of no speed-accuracy trade-off for the repeated items is a recognized hallmark of skill learning (MacKay, 1982). Furthermore, our data support the view that both accuracy and speed (fluency) need to be taken into account for the effective description of the evolution of linguistic competence. Both measures were shown to be effective for performance assessment in linguistic skills including speech production (e.g., MacKay, 1982), reading (e.g., McLeod & McLaughlin, 1986) and second language acquisition (e.g., DeKeyser, 1997). Nevertheless, because accuracy reached ceiling levels early on, it may constitute a less sensitive measure for determining the finegrained time-course of development of the morphological skill and therefore may not suffice as a criterion for automatization. As in non-linguistic skills (e.g., Korman et al., 2003), performance speed may serve as a more sensitive measure for the progression of skilled performance. Our results showed that although the AMR was never explicitly instructed at any stage of training, the training conditions enabled its generalization to items not previously encountered. The individual participants’ and the group-average performance curves show that the ability to generalize the AMR to novel items often evolved over multiple daily sessions. Thus, even in those individuals who were able to gain explicit knowledge of the AMR as early as the initial practice session, the ability to generalize the AMR accurately and fluently continued to improve with accumulating experience. Our results also suggest that learning to generalize the AMR to new items reflects complex processes whereby the phonological and the semantic aspects of the AMR could be learned separately, differently, and either successively or in parallel. The rapid increase in phonological accuracy in the production of the new items, irrespective of the semantic accuracy, suggests that the phonological aspect of the AMR could be effectively applied to items not previously encountered – even after a rather limited amount of exposure and training. All participants were able to correctly (>75% correct) produce the new verbs after two practice sessions. The finding that at the same time, 5/8 participants were struggling (i.e., at chance level performance) with the application of the semantic aspect of the AMR, producing a correct (verb-suffix) transformation but with no correspondence to the semantic distinction, suggests that the phonological aspect of the AMR could be mastered independently of the semantic aspect.

10

Procedural and Declarative Memory in the Acquisition . . .

209

The explicit verbal report demonstrated that even when participants were accurately generating the phonological transformation of the task in the new items, they often remained unaware of at least some components of the transformation. This finding is compatible with the idea that some types of knowledge are more accessible to awareness than others, and with the notion that salience may play an important role in language (specifically, phonological) acquisition (e.g., DeKeyser, 2005). Thus, the current data indicate that, at least in the initial phase of training, part of the phonological knowledge was implicit. Altogether, our data are in line with the notion that complex (multi-component) knowledge may be learned and generalized without awareness of key aspects of the task (Reber, Knowlton, & Squire, 1996; Robinson, 1996; Stadler, 1989). It is noteworthy, however, that the explicit verbal report in our study could have encouraged participants to become aware of their knowledge. It is possible that without this manipulation, a larger part of the phonological knowledge would have remained implicit. Our findings are consistent with the notion that morphological ‘‘rules’’ include distinct phonological and semantic representations, which may have different realizations in the brain (e.g., Joanisse & Seidenberg, 1999). Furthermore, the finding that all but one participant not only acquired the phonological transformation rapidly but did so prior to discovering the semantic aspect is compatible with the ease whereby phonological aspects of language are grasped and generalized in first and second language learning (e.g., Saffran, Johnson, Alsin, & Newport, 1999). There is evidence that the recognition of phonological structure may begin in utero and that infants can abstract and transfer ‘‘phonological constrains to new items long before they acquire their first meaningful word’’ (e.g., see review by Gomez & Gerken, 2000). Models of second language acquisition suggest that fluency in novel items can be explained as a bottom-up process wherein phonological habits (i.e., procedural knowledge) must be established first and the manipulation of the conceptual level must be delayed until all lower-level processes are well established. We would not, however, construe our findings as supporting the notion that in the learning of every linguistic task wherein both phonological and semantic features are combined, the phonological component is the first to be acquired. Rather, the acquisition of a non-salient phonological feature combined with an easy semantic concept might have resulted in the opposite effect. Moreover, the current results do not necessarily imply that the phonological and semantic aspects of the AMR were acquired successively and independently. It cannot be ruled out that the emergence of knowledge of each of these aspects was enhanced by the gains in the complementary aspect, through a dynamic interaction between the phonological and semantic processes employed to solve the linguistic problem (e.g., Bitan & Karni, 2003; Bitan, Manor, Morocz, & Karni, 2005). This view is consistent with several studies suggesting that in the course of learning a linguistic task (e.g., words, morphological rules), semantic and phonological constraints interact with each other (Haskell, McDonald, & Seidenberg, 2003).

210

S. Ferman and A. Karni

It has been argued that the learning of grammatical knowledge and its representation are mainly implicit (see review in Ullman, 2001, 2004). Nevertheless, explicit knowledge may be required for the learning of some features embedded in natural language. Learning the phonological transformation did not by itself constitute a sufficient condition enabling participants to go beyond chance level performance in the judgment and production of items (of the AMR) not previously encountered. Our results suggest that the discovery of the semantic (animate-inanimate) distinction and its requisite role in the AMR was crucial for accurate generalization to new items. Furthermore, our results indicate that the acquisition of the semantic aspect of the AMR was in the form of an explicit discovery – that is, the establishment of declarative knowledge. Errors such as using the incorrect consonant in the artificial suffix (v/r) which were often committed in our sample of adult learners, suggest that the artificial suffixes were not acquired as whole-unit forms, but rather that each phoneme of the suffix was acquired and represented separately in memory. Similarly, the phonological errors that were identified in the verb stems suggest that even the verb stems, previously known phonological patterns, underwent at least partial segmentation as a result of the required transformation. These results suggest that various phonological sub-components of the AMR transformed verbs were detected and represented in a modular manner at various stages of learning and support the notion that phonological knowledge may be organized as a sequence of phonemes (Saffran, Newport, & Alsin, 1996). We propose that this should be taken as an indication for the modularity of the phonological (and morpho-phonological) aspect of the AMR at least at some stages of the acquisition, rather than for gestalt-like acquisition of repeated exemplars. Taken together, our results provide support for the approach that the knowledge that constitutes a morphological ‘‘rule’’ is not ‘‘abstracted’’ into a single ‘‘rule’’ but rather is evolved as modular, multiple-component knowledge. Our results suggest that multiple levels of representation of a morphological rule (i.e., various components of the phonological transformation and the semantic aspect) can emerge during the course of learning. This finding lends support to ‘‘non-abstractionist’’, multiple probabilistic and connectionist models for linguistic learning and representation (e.g., Saffran, 2002). By analogy to non-linguistic skill learning, we propose that eventually, most participants evolved an effective task-specific routine for the application of the AMR to repeated and new items. This routine may be conceptualized as an alternative (i.e., available in parallel) to the multi-component process that constitutes the basis for task performance in the early stages of practice (along with specific item recollection). Thus, we propose that parallel to task performance based on increasingly more effective use of multiple pre-existing processing modules, a novel task-specific module for the artificial morphological transformation may have been established with extended practice. This finding is highly characteristic of skill learning. In the context of motor sequence learning, for example, the data suggest that the knowledge generated

10

Procedural and Declarative Memory in the Acquisition . . .

211

and retained in memory at different levels of practice on a given task may radically change in discrete steps and that novel movement modules can be established even in the adult brain (Korman et al., 2003; Sosnik, Hauptman, Karni, & Flash, 2004). Our results support the view that both memory systems are involved in the learning and use of particular aspects of language. Our findings however, constitute a marked departure from Ullman’s procedural/declarative model in three main points. First, Ullman’s model dichotomizes between grammatical and lexical learning by suggesting that all aspects of grammatical rules including phonology, non-lexical semantics, syntax, and morphology, are learned, retained and applied in actual performance by means of the procedural memory system and that all aspects of a word including meaning, phonology, and abstract representations such as word category are learned and applied by means of the declarative memory system (Ullman, 2001, 2004). Our findings, however, provide clear evidence that a morphological rule – a grammar rule in Ullman’s terms – is not acquired as a singular (whole) unit of knowledge, but rather that the phonological and the semantic aspects of the rule are learned separately (even successively), characterized by different time-courses, and are moreover supported by different contributions from the two memory systems. Our results can be interpreted to suggest that the phonological aspect of a morphological rule is learned implicitly and retained as procedural memory and that the acquisition of the semantic aspect of the rule requires an explicit learning stage making use of the declarative memory. This is consistent with the notion that the phonological aspect of a word (lexical item) is acquired implicitly and its meaning aspect is acquired explicitly (Ellis, 1994). Second, rather than a lexicon versus grammar dichotomy implemented in different memory systems, our results are in line with several studies suggesting that the knowledge underlying higher-level cognitive skills is subserved by complex interactions between the declarative and procedural memory systems (e.g., Bitan & Karni, 2003; Bitan et al., 2005) with perhaps both cooperative and competitive contributions to learning and performance (Poldrack & Packard, 2003). A more recent version of Ullman’s proposal raises some uncertainties about the demarcation of language elements that fall within the domains of either the procedural or the declarative memory system (Ullman, 2004). Third, our results go a step further in implying that the contribution of the procedural and declarative memory systems to the acquisition of each aspect of a linguistic rule is dynamic and dependent upon the amount of practice and upon the structure of the training experience. This notion of dynamics constitutes a marked departure from the view of a ‘‘static’’ procedural or declarative memory dichotomy for specific language functions and units. From this perspective, our findings suggest that the procedural memory is responsible for the acquisition and retention of the repeated items as well as the implicit learning of the phonological aspect of the AMR for the transformation of new items. However, later on in training, at least part of the implicit phonological knowledge may be transferred into explicit knowledge. That is, further training may

212

S. Ferman and A. Karni

result in developing declarative memory of the new phonological knowledge (i.e., the artificial suffixes) but apparently not of the preexisting Hebrew-valid phonological routines (i.e., vowel omission and stress shift). A similar dynamic process, but in the reverse direction, can be recognized in the acquisition of the semantic aspect of the morphological knowledge. Initially, the learning of the semantic aspect, at least in adults, appeared to depend upon the explicit discovery (declarative memory) of the animate-inanimate distinction. After additional practice, however, the semantic declarative knowledge of both the phonological and the semantic aspects of the AMR may undergo proceduralization, resulting in what we propose should be viewed as a set of multicomponent specific linguistic routines. There are, therefore, reasonable grounds to conjecture that the learning of a morphological ‘‘rule’’ is a multiple-phase process wherein both procedural and declarative memory systems contribute differentially, in different phases of experience, to the evolving linguistic competence. There should be no a-priory limit on language learning in adults, at least in laboratory settings. In the current study, young adults learned to judge and produce repeated and new items of an AMR whereby highly familiar, legitimate, native language (Hebrew) syntactical noun-verb forms were transformed. Such a paradigm can be considered a partial artificial language paradigm wherein the trained task was related to previously established skills (Hulstijn, 1997) and thus resembles many cases of laboratory non-linguistic skill learning paradigms in adults in which the novel skill builds upon existing skills (e.g., handwriting, Sosnik et al., 2004). Not only did the participants learn the AMR through their native language – more importantly, they made phonological, morphological, and semantic errors that clearly indicate the involvement of basic linguistic processes in learning the AMR. From this perspective, the current experimental paradigm can be considered as a laboratory paradigm for morphological acquisition in late stages of first language development (e.g., Berman, 2004, Nippold, 2007), as well as in second language acquisition in adulthood (DeKeyser, 1997). We propose, therefore, that the main findings presented here reflect basic learning phenomena in the domain of language competence. The drawing of theoretical and practical implications from the present laboratory artificial language research, however, should be cautious, and broader, ecologically valid data, including for children, are needed. Nevertheless, there are some indications that the ability of (young and healthy) adults to acquire skills including linguistic competence is not inferior top that of children before the onset of puberty. A study in which children (8- and 12-year-olds) were given an identical training protocol on the AMR has shown that both groups of children were slower and less affective in acquiring the linguistic skill compared to the adults (Ferman & Karni, 2010). It was proposed that this lack of childhood advantage reflected a developmental advantage of adults in both declarative and procedural memory systems in relation to acquiring and retaining language competence. A recent study has indicated that the acquisition, consolidation and retention of a sequence of

10

Procedural and Declarative Memory in the Acquisition . . .

213

finger movements (motor skill) are as effective in adults as in children; if at all there was some superiority in skill retention by the adults (Dorfberger et al., 2007). The results further suggested that the difference between sequence learning in children and adults may rest in the fact that the adult procedural memory consolidation process may be more selective than the one in children (before adolescence) and is thus more prone to interference by subsequent experience, including normal everyday experience. In the context of language learning, this result may be taken to indicate the possibility that second language acquisition in adults may be prone to interference by the use of the first language, but not so in children. Thus, we propose that while adults may have superior or at least as effective memory systems, declarative and procedural, as children, it may be the case that second language acquisition in children can be effectively consolidated in a mixed language environment, while for adults only complete immersion in the new language would lead to the effective expression of their language learning potential.

References Anderson, J. R., Bothell, D., Byrne, M. D., Douglass, S., Lebiere, C., & Qin, Y. (2004). An integrated theory of the mind. Psychological Review, 111, 1036–1060. Ari-Even Roth, D., Kishon-Rabin, L., Hildesheimer, M., & Karni, A. (2005). A latent consolidation phase in auditory identification learning: Time in the awake state is sufficient. Learning and Memory, 12, 159–164. Bates, E., Devescovi, A., & Wulfeck, B. (2001). Psycholinguistics: A cross-language perspective. Annual Review of Psychology, 52, 369–396. Bates, E., & Goodman, J. C. (1997). On the inseparability of grammar and the lexicon: Evidence from acquisition, aphasia, and real-time processing. Language and Cognitive Processes, 12, 507–584. Berman, R. A. (Ed.). (2004). Language development across childhood and adolescence: Crosslinguistic perspectives. TILAT (Trends in Language Acquisition Research) Series. Amsterdam: Benjamins. Bialystok, E. (1982). On the relationship between knowing and using linguistic forms. Applied Linguistics, 3, 181–206. Bitan, T., & Karni, A. (2003). Alphabetical knowledge from whole words training: Effects of explicit instruction and implicit experience on learning script segmentation. Cognitive Brain Research, 16, 323–337. Bitan, T., Manor, D., Morocz, A., & Karni, A. (2005). Effects of alphabetically, practice and type of instruction on reading an artificial script: An fMRI study. Cognitive Brain Research, 25, 90–106. Braine, M. D. S., Brody, R. E., & Brooks, P. J. (1990). Exploring language acquisition in children with miniature artificial language: Effects of item and pattern frequency, arbitrary subclasses, and correction. Journal of Memory and Language, 29, 591–610. Chomsky, N. (1995). The minimalist program. Cambridge, MA: MIT Press. Cleeremans, A., Destrebecqz, A., & Boyer, M. (1998). Implicit learning: News from the front. Trends in Cognitive Science, 2, 406–416. Cohen, N., & Poldrack, R. A. (1997). Memory for items and memory for relations in the procedural/declarative memory framework. Memory, 5, 131–178.

214

S. Ferman and A. Karni

Dahan, D., & Brent, M. (1999). On the discovery of novel wordlike units from utterances: An artificial-language study with implications for native-language acquisition. Journal of Experimental Psychology: General, 128, 165–185. DeKeyser, R. (1997). Beyond explicit rule learning: Automatizing second language morphosyntax. Studies in Second Language Acquisition, 19, 195–222. DeKeyser, R. M. (2001). Automaticity and automatization. In P. Robinson (Ed.), Cognition and second language instruction (pp. 125–151). New York: Cambridge University Press. DeKeyser, R. M. (2005). What makes learning second language grammar difficult? A review of issues. Language Learning, 55(Suppl 1), 1–25. Devescovi, A., Caselli, M. C., Marchione, D., Pasqualetti, P., Reilly, J., & Bates, E. (2005). A cross linguistic study of relationship between grammar and lexical development. Journal of Child Language, 32, 759–786. Dionne, G., Dale, P. S., Boivin, M., & Plomin, R. (2003). Genetic evidence for bidirectional effects of early lexical and grammatical development. Child Development, 74, 394–412. Dorfberger, S., Adi-Japha, E., & Karni, A. (2007). Reduced susceptibility to interference in the consolidation of motor memory before adolescence. PloS One, 2, e240. Ellis, N. (1994). Vocabulary acquisition: The implicit ins and outs of explicit cognitive mediation. In N. Ellis (Ed.) Implicit and explicit learning of languages (pp. 211–283). London: Academic. Ellis, N., & Schmidt, R. (1998). Rules or associations in the acquisition of morphology? The frequency by regularity interaction in human and PDP learning of morphosyntax. Language and Cognitive Processes, 13, 307–336. Elman, J., Bates, E., Johnson, M., Karmiloff-Smith, A., Parisi, D., & Plunkett, K. (1996). Rethinking innateness: A connectionist perspective on development. Cambridge, MA: MIT Press. Ferman, S., & Karni, A. (2010). No childhood advantage in the acquisition of skill in using an artificial language rule. PLoS One, 5(10), e13648. Retrieved October 27, 2010. Ferman, S., Olshtain, E., Schechtman, E., & Karni, A. (2009). The acquisition of a linguistic skill by adults: Procedural and declarative memory interact in the learning of an artificial morphological rule. Journal of Neurolinguistics, 22, 384–412. Gomez, R. L. (1997). Transfer and complexity in artificial grammar learning. Cognitive Psychology, 33, 154–206. Gomez, L. R., & Gerken, L. (2000). Infants artificial language learning and language acquisition. Trend in Cognitive Sciences, 4, 178–186. Haskell, T., McDonald, M., & Seidenberg, M. (2003). Language learning and innateness: Some implications of compound research. Cognitive Psychology, 47, 119–163. Hauptmann, B., & Karni, A. (2002). From primed to learn: The saturation of repetition priming and the induction of long-term memory. Brain Research. Cognitive Brain Research, 13, 313–322. Hauser, M. D., Chomsky, N., & Techman Fitch, W. (2002). The faculty of language: What is it, who has it, and how did it evolve? Science, 298, 1569–1579. Hulstijn, J. H. (1997). Second language acquisition research in the research lab: Possibilities and limitations. Studies in Second Language Acquisition, 19, 131–144. Joanisse, M. F., & Seidenberg, M. S. (1999). Impairments in verb morphology after brain injury: A connectionist model. Proceedings of the National Academy of Sciences, 96, 7592–7597. Johnson, K. (1996). Language teaching and skill learning. Oxford: Blackwell. Karni, A. (1996). The acquisition of perceptual and motor skills: A memory system in the adult human cortex. Brain Research. Cognitive Brain Research, 5, 39–48. Karni, A., & Sagi, D. (1993). The time course of learning a visual skill. Nature, 365, 250–252. Knowlton, B. J., & Squire, L. R. (1996). Artificial grammar learning depends on implicit acquisition of both abstract and exemplar-specific information. Journal of Experimental Psychology: Learning, Memory, and Cognition, 22, 169–181.

10

Procedural and Declarative Memory in the Acquisition . . .

215

Korman, M., Raz, N., Flash, T., & Karni, A. (2003). From novelty to familiarity: Practice makes transfer imperfect. Proceeding of the National Academy of Sciences of the USA of America, 100, 12492–12497. Logan, G. (1988). Toward an instance theory of automatization. Psychological Review, 95, 492–527. MacKay, D. G. (1982). The problems of flexibility, fluency, and speed-accuracy trade off in skilled behavior. Psychological Review, 89, 483–506. MacWhinney, B. (1983). Miniature linguistic systems as tests of the use of universal operating principles in second language learning by children and adults. Journal of Psycholinguistic Research, 12, 467–478. Maquet, P., Laureys, S., Perrin, F., Ruby, P., Melchior, G., Boly, M., et al. (2003). Festina lente: Evidences for fast and slow learning processes and a role for sleep in human motor skill learning. Learning and Memory, 10, 237–239. Mathews, R. C., Buss, R. R., Staneley, W. B., Blanchard-Fields, F., Cho, J. R., & Druhan, B. (1989). Role of implicit and explicit processes in learning from examples: A synergistic effect. Journal of Experimental Psychology: Learning, Memory, and Cognition, 15, 1083–1100. McLeod, B., & McLaughlin, B. (1986). Restructuring or automaticity? Reading in a second language. Language Learning, 36, 109–123. Mishkin, M., Malamut, B., & Bachevalier, J. (1988). Memories and habits: Two neural systems. In G. Lynch, J. L. McGaugh, & N. W. Weinburger (Eds.), The neurobiology of learning and memory (pp. 64–88). New York: Guliford. Nippold, M. A. (2007). Later language development: School-age children, adolescents, and young adults (3rd ed.). Austin, TX: Pro-Ed. Pinker, S. (1994). The language instinct. New York: William Morrow. Pinker, S. (1999). Words and rules: The ingredients of language. New York: Basic Books. Poldrack, R., & Packard, M. G. (2003). Competition among multiple memory systems: Converging evidence from animal and human brain studies. Neuropsychologia, 41, 245–251. Pothos, E. M. (2007). Theories of artificial grammar learning. Psychological Bulletin, 133, 227–244. Reber, A. S. (1989). Implicit learning and tacit knowledge. Journal of Experimental Psychology: General, 118, 219–235. Reber, P. J., Knowlton, B., & Squire, L. (1996). Dissociable properties of memory systems: Differences in the flexibility of declarative and nondeclarative knowledge. Behavioral Neuroscience, 110, 861–871. Robertson, E. M., & Cohen, D. A. (2006). Understanding consolidation through the architecture of memories. Neuroscientist, 12, 261–271. Robinson, P. (1996). Learning simple and complex second language rules under implicit, incidental rule search and instructed conditions. Studies in Second Language Acquisition, 18, 27–67. Robinson, P., & Ha, M. A. (1993). Instance theory and second language rule learning under explicit conditions. Studies in Second Language Acquisition, 15, 413–438. Saffran, J. R. (2002). Constraints on statistical language learning. Journal of Memory and Language, 47, 172–196. Saffran, J. R., Johnson, E. K., Alsin, R. N., & Newport, E. L. (1999). Statistical learning of tone sequences by human infants and adults. Cognition, 70, 27–52. Saffran, J., Newport, E., & Alsin, R. (1996). Word segmentation: The role of distributional cues. Journal of Memory and Language, 35, 606–621. Schmidt, R. (1993). Psychological mechanisms underlying second language fluency. Studies in Second Language Acquisition, 14, 357–385. Segalowitz, N. (2003). Automaticity and second languages. In C. Doughty & M. Long (Eds.), Handbook of second language acquisition (pp. 382–340). Oxford: Blackwell.

216

S. Ferman and A. Karni

Sosnik, R., Hauptman, B., Karni, A., & Flash, T. (2004). When practice leads to coarticulation: The evolution of geometrically defined movement primitives. Experimental Brain Research, 156, 422–438. Squire, L. R., & Zola, S. M. (1996). Structure and function of declarative and nondeclarative memory systems. Proceeding of the National Academy of Sciences of the USA of America, 93, 13515–13522. Stadler, M. A. (1989). On learning complex procedural knowledge. Journal of Experimental Psychology: Learning, Memory and Cognition, 15, 1061–1069. Sun, R., & Zhang, X. (2004). Top-down versus bottom-up learning in cognitive skill acquisition. Cognitive Systems Research, 5, 63–89. Ullman, M. T. (2004). Contribution of memory circuits to language: The declarative/procedural model. Cognition, 92, 231–270. Ullman, M. T. (2001). A neurocognitive perspective on language: The declarative/procedural model. Nature Review Neuroscience, 2, 717–726. Walker, M. P., & Stickgold, R. (2006). Sleep, memory, and plasticity. Annual Review of Psychology, 57, 139–166.

Chapter 11

Reading in L1 and L2: Behavioral and Electrophysiological Evidence: A Comparison Between Regular and Dyslexic Readers Zvia Breznitz and Liat Fabian

11.1 Bilingualism: Different Definitions and Hypotheses What exactly is bilingualism and how can we define this phenomenon? Linguists have distinguished between various types of bilingualism and have proposed specific definitions of what it means to be a bilingual. It is important to understand the qualitative differences between the many definitions, which can influence both research and practice, including teaching, training and intervention. In general, a person who can understand, speak and read more than one language is considered to be bilingual (i.e., Breznitz, Oren, & Shaul, 2004; Fabbro, 1999). Neurolinguistic and psychological research on this subject has yielded several different ways to identify, classify and define different types of bilinguals. In his review, Fabbro (1999) defined the following types of bilinguals: a ‘‘coordinated bilingual’’ learns the second language after the first, before puberty, perhaps following relocation to a foreign country. For a ‘‘subordinate bilingual’’, one language serves as the mother tongue while the second serves as the mediator of the first. That is, the individual thinks of what he wishes to express in the first language and then translates it into the second language. ‘‘Early bilingualism’’ refers to the early acquisition of the two languages, while ‘‘late bilingualism’’ refers to the acquisition of the second language much later than the first. Other definitions attempt to describe the degree of fluency in the two languages. For instance, a ‘‘balanced bilingual’’, is an individual who has mastered both languages to the same extent, whereas a ‘‘dominant bilingual’’ is more fluent in one language than in the other (see Fabbro, 1999 for a comprehensive review).

Z. Breznitz (*) Edmond J. Safra Brain Research Center for the Study of Learning Disabilities, University of Haifa, Haifa 31905, Israel e-mail: [email protected] M. Leikin et al. (eds.), Current Issues in Bilingualism, Literacy Studies 5, DOI 10.1007/978-94-007-2327-6_11, Ó Springer ScienceþBusiness Media B.V. 2012

217

218

Z. Breznitz and L. Fabian

It is widely accepted among researchers that in order to attain a complete and accurate picture of bilingualism, in addition to examining the behavioral level, we must also look at the neurolinguistic level. This level defines bilinguals as individuals who have mastered, can understand, and speak either two languages, two dialects, or one of each and in addition can distinguish between two linguistic codes during verbal production without confusing the two (Fabbro, 1999). In order to be considered a bilingual, one does not have to attain the same level of competence and/or performance in both languages. It is interesting to note the common misconception that only a small percentage of the world’s population are bilingual, when in fact over 50% can be considered bilingual (Fabbro, 1999). There are even those that go so far as to claim that there is no such thing as a monolingual person. This claim is made on the basis of the assumption that once a person is able to understand and/or speak an additional language or jargon, beyond the native language spoken at home, he can be considered bilingual (Kroll & de Groot, 1997). In general, over the course of a lifetime people can develop different skills in accordance with the varying circumstances and needs that arise. In reference to the four basic skills of reading, writing, speaking and listening, a person’s ability may differ significantly between these language skills (i.e., Breznitz et al., 2004). In sum, due to the complex nature of the matter, it is difficult to precisely classify and distinguish between the many different types of bilinguals. Therefore, in order to adequately interpret research carried out on bilingualism, it is necessary to fully understand this complex phenomenon (Harris & McGhee Nelson, 1992). Over the years, several different hypotheses have been proposed in an attempt to better understand bilingualism in general, and second language acquisition in particular. Cummins’ (1979) Linguistic Interdependence Hypothesis posits that cognitive academic language proficiency is transferred from one language to another. This hypothesis claims that there is a significant relationship between one’s skills in two languages (L1 and L2), and it can therefore be assumed that low competence in L1 leads to low competence in L2 and high competence in L1 can predict a similarly high level of competence in L2 (Da Fontoura & Siegel, 1995). Many studies have supported this statement, indicating a high correlation between an individual’s reading skills in two languages (Da Fontoura & Siegel, 1995). In other words, a strong connection has been found between L1 and L2 language skills. As a result of Cummins’ hypothesis, the Central Deficit Hypothesis was proposed, claiming that individuals experiencing reading problems in their first language will also manifest similar difficulties in their second language, regardless of the language used (Abu-Rabia, 1997). There is a strong connection between L1 and L2 skills, however, having L1 difficulties is not considered to be the direct cause of L2 learning problems. Vellutino and Scanlon (1986) examined the language abilities of good and poor readers (including phonology, syntax, and semantics) and their research into native language reading disabilities is the basis for a hypothesis, similar to

11

Reading in L1 and L2: Behavioral and Electrophysiological Evidence . . .

219

Cummins’ Linguistic Interdependence Hypothesis, pertaining to first and second language acquisition and proficiency known as the Foreign Language Linguistic Coding Differences Hypothesis (LCDH) (Ganschow, Sparks, Javorsky, Pohlman, & Bishop-Marbury, 1991). This hypothesis aims to explain the different types of problems existing in foreign language acquisition, claiming that difficulties in the native language are the cause of foreign language difficulties and can predict these difficulties as well. The LCDH is based on research conducted among college students experiencing foreign language learning difficulties. In other words, according to the LCDH, the difficulties experienced in acquiring a foreign language result from deficiencies in at least one of the linguistic codes in the native language (Schwartz, 1997). Consequently, an individual will experience anywhere from mild to extreme difficulties with both the oral and written aspects of the second language, specifically from phonological processing deficits experienced in the native language (Schwartz, 1997; Sparks, Philips, Ganschow, & Javorsky, 1999). There are additional issues concerning second language acquisition, in particular the age of acquisition, proficiency level and amount of exposure to the second language. Over the years there has been consistent interest in the existence of a biologically ‘‘critical period’’ for language acquisition (Lenneberg, 1967), and specifically if this also pertains to second language acquisition (Johnson & Newport, 1989), such as the difference between early and late bilinguals. In addition to age of acquisition, proficiency level in L2 is also of interest. According to the Convergence Hypothesis (Green, 2003) different languages recruit a common system, indicating that L1 influences both acquisition and processing of L2 (Li & Green, 2007). As proficiency increases, qualitative differences between native first (L1) and second (L2) language speakers disappear, and L2 processing ‘‘converges’’ onto the neural representation of L1. In other words, this hypothesis states that L2 is acquired in the context of an existing language system (L1) and that the second language (L2) will receive convergent neural representation within the representations learned in L1 (Perani & Abutalebi, 2005). Neuroimaging studies have showed that when the degree of proficiency in L1 and L2 is comparable, there is common activation in the left hemisphere, both on word and sentence task, and across different languages (Chee, Tan, & Thiel, 1999; Klein, Milner, Zatorre, Zhao, & Nikelski, 1999). In addition, among bilinguals with low proficiency in L2, additional brain activity has been found, mostly in prefrontal areas (Briellman et al., 2004; De Bleser et al., 2003). These important issues will be discussed later in the chapter with relation to the different studies conducted. Another important issue that has interested researchers in the study of bilingualism has been the localization of languages in the bilingual brain, specifically whether the two different languages are localized in the same areas or in distinct areas of the brain (see Fabbro, 1999 for a comprehensive review). Several hypotheses have been proposed concerning this critical question. The first hypothesis suggests that all languages known by bilinguals are localized in the same cerebral areas. In contrast, a second hypothesis, states that each

220

Z. Breznitz and L. Fabian

language is localized in a separate area of the brain. The Differential Hypothesis (Ullman, 2001) states that adult L2 acquisition does not rely on the same brain mechanisms used in processing native L1 language. However, several studies conducted thus far have contradicted this hypothesis, showing that while performing grammatical tasks, the brain structures that are traditionally involved in grammatical processing, for example Broca’s region, were apparent in both L1 and L2 performance (Briellman et al., 2004; Sakai, Miura, Narafu, & Muraishi, 2004; Wartenburger et al., 2003). A third hypothesis states that the different languages in a bilingual brain are organized in the same cortical areas but in distinct neural circuits. This hypothesis is comprised of two important claims, the first being that the distinct cerebral cells subserve the different languages, whereas the second claim states that the same cells are active, yet in different combinations and interactions depending on the language involved (Fabbro, 1999). Finally, many researchers support the notion that language is partially organized in common areas and partially in specific and separate areas of the brain (Fabbro, 1999). In other words, this hypothesis is based on synthesis between three hypotheses namely, that brain activity during the bilingual process in different languages occurs in the same brain areas, in separate brain areas and within the same areas distinct neural circuits working independently for each language.

11.2 Bilingualism in Reading: Behavioral and Brain Research In the last decade, there has been a great deal of neuropsychological and neuroimaging research on the representation and processing of multiple languages in the brain (Perani & Abutalebi, 2005). These studies employed different methods, including event-related potentials (ERP), positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) on different bilingual populations (i.e., English, German, Italian, Japanese, Chinese) (Li & Green, 2007). In other words, neurophysiological research provides us with a more complete picture of how language is represented in the brain (Zhang & Wang, 2007). The neural basis of second language acquisition and processing has also been of interest, particularly with regard to whether both L1 and L2 are processed through the same neural mechanisms (for review see Perani & Abutalebi, 2005).

11.2.1 Behavioral Measures On the behavioral level, research has compared the processing times of bilinguals in both L1 and L2 while performing different tasks. Most often bilinguals are given reaction time tests, comparing either reading speed of words or texts in both languages, naming latencies of objects or colors, and word completion tasks (Cook, 1997). It is interesting to note that over a century ago, in 1887,

11

Reading in L1 and L2: Behavioral and Electrophysiological Evidence . . .

221

Cattell stated that it took more time to name an object in L2 than in L1, even when one attains a very high level of proficiency in L2 (in Keatley, 1992). Many years later, studies continue to address the difference between reading speed in L1 and L2. Favreau and Segalowitz (1983) suggested that the slower reading rates for L2 as compared to L1 were associated with a lower level of automaticity of word recognition in L2, which resulted from less time practicing language processing in the second language relative to the first. In other words, even among highly skilled bilinguals, their decreased automatization in L2 varied as a function of fluency and the age at which the second language was acquired. Segalowitz, Poulsen, and Komoda (1991) found that L2 reading rates were 30% slower than L1 reading rates among French/English bilinguals while reading equivalent L1 and L2 texts in both English and French. In a more recent study, Cook (1997) found that in comparison to L1, mental activities performed in L2 are usually fractionally less efficient in terms of both speed and accuracy. In sum, these studies provide behavioral evidence supporting the assumption that the differences in accuracy and fluency of word reading in L2 can be explained, at least in part, by the age of L2 acquisition and the amount of experience one has with the second language. In addition to comparing between L1 and L2 on behavioral measures of speed and accuracy, neuroscientific research on bilingualism has attempted to clarify whether the two languages are localized in shared or separate areas of the brain, how the brain processes the language in general and other influencing variables such as age of L2 acquisition, degree of proficiency and amount of exposure to language (Abutalebi, Cappa, & Perani, 2001; Perani & Abutalebi, 2005).

11.2.2 Imaging Studies To date, neuroimaging studies on bilingualism using positron emission tomography (PET) and functional magnetic resonance (fMRI) have demonstrated both identical and different representations of L1 and L2 in the brain. In particular, specific frontal, temporal, and parietal regions, along with subcortical structures, are differentially involved in linguistic processing of words and sentences (Friederici, 2002; Poeppel & Hickok, 2004). These studies have investigated the manner in which language is organized in the bilingual brain, using different types of linguistic stimuli: word, sentence and short story processing (Fabbro, 2001). 11.2.2.1 Word Processing Klein, Zatorre, Milner, and Meyer (1994) conducted a PET study with fluent English-French bilinguals, in which they were requested to repeat aloud words presented auditorially in both languages. Results showed that there was a similar cortical localization of activation for both languages. Klein, Milner,

222

Z. Breznitz and L. Fabian

Zatorre, Meyer, and Evans (1995) replicated the previous findings with fluent English-French bilinguals in an additional PET study. In this study, subjects were instructed to generate aloud synonyms of words presented to them auditorially in both languages. This study showed that during a word generation task in the less fluent language, there was greater activation of the left putamen in the subject’s brain. In an additional PET study conducted by Klein et al. (1999) on Chinese-English bilinguals, who had all acquired English (L2) in adolescence, subjects performed a verb-generation task and results showed that despite later acquisition of L2, both languages showed activation in the same cerebral structures. Furthermore, a study conducted by Chee et al. (1999) on Chinese-English bilinguals showed that the same brain areas were activated for both languages, regardless of age of acquisition, meaning that both early (L2 learned before age 6) and late (L2 learned after age 12) bilinguals showed indistinguishable activation patterns on a word-processing task in both their languages. Illes et al. (1999) used fMRI to determine whether L1 and L2 have a unitary or separate cortical system for specifically semantic performance. Fluent bilinguals in English and Spanish were tested on semantic and non-semantic judgments using words that were presented visually. Results showed greater activation for semantic relative to non-semantic decisions in left and right frontal regions. Furthermore, the locations of activations were similar for both English and Spanish languages. Based on the results, the researchers suggested that there is a common neural system in the frontal lobe for semantic processing of two languages (see also Klein et al., 1994, 1995, 1999). In sum, these studies show that there is a shared cortical system for semantic knowledge in the bilingual brain (Illes et al., 1999). In addition, level of proficiency may have greater influence on the cerebral representation of semantic processing than age of acquisition (Klein, Watkins, Zatorre, & Milner, 2006; Klein, Zatorre, et al., 2006; Perani et al., 1998; Wartenburger et al., 2003). More recently, Meschyan and Hernandez (2006) conducted an fMRI study in an attempt to examine the impact of language proficiency and orthographic transparency on word reading among Spanish-English bilinguals, who were more fluent in English (L2) than Spanish, which was their native language. Subjects were required to read words in both languages. Results showed that subjects read words in Spanish (L1) significantly slower than words read in English (L2). In addition, the fMRI revealed that there was greater activity in the articulatory motor system while reading words in the less proficient language. In sum, both the behavioral and neural data revealed that when required to read in the less practiced (and less proficient) language, greater articulatory motor effort is required. In addition, when reading Spanish, which is more orthographically transparent than English, greater activity was found in the superior temporal gyrus, which is an area activated in phonological processing, whereas while reading English, greater activity was shown in visual processing and word recoding regions (occipito-parietal border and inferior parietal lobe), indicating a lesser degree of influence resulting from exposure to language.

11

Reading in L1 and L2: Behavioral and Electrophysiological Evidence . . .

223

11.2.2.2 Sentence Processing Kim, Relkin, Lee, and Hirsch (1997) conducted an fMRI study in an attempt to determine the developmental relationship between L1 and L2 among fluent English-French bilinguals. Half of the subjects had been exposed to the two languages during infancy (‘‘early bilinguals’’), and the other half had been exposed to their second language in early adulthood (‘‘late bilinguals’’). Subjects performed a silent sentence-generation task in both languages. Results indicated that among the ‘‘early’’ bilinguals, the cortical localization of activity for the two languages was similar both in Broca’s and Wernicke’s areas. Among the ‘‘late’’ bilinguals, however, no differences were observed in Wernicke’s area but spatially distinct activation sites were evident for L1 versus L2 in Broca’s area. The researchers concluded that age of language acquisition may be a significant factor in determining the functional organization of this area in the bilingual brain, with native and second languages less distinctly localized in early exposure to the second language. Both common and separate cortical activations for L1 and L2 were found in moderately fluent French-English bilinguals during processing of sentences (Dehaene et al., 1997). It is apparent that the fMRI studies have revealed both common and separate brain activity for L1 and L2. However, it is difficult to interpret the results as the studies used different tasks as stimuli. Illes et al. (1999) attempted to resolve the conflicting experimental data found in the previous fMRI studies (Dehaene et al., 1997; Kim et al., 1997) by stating that both had examined moderately fluent bilinguals instead of highly proficient bilinguals as did the remaining studies. As shown by Perani et al. (1996) and (1998), proficiency clearly affected the cortical localization and processing of the second language. Less proficient bilinguals in Perani et al.’s first study (1996) showed very different activation patterns in L1 and L2, whereas highly proficient late learners in the second study showed no differences between L1 and L2.

11.2.3 Electrophysiological Studies The reading process is based on the information processing system (Brandeis & Lehmann, 1994; Johnson, 1995) which is characterized by limitations of capacity and rapid decay (Baddeley, Logie, Bressi, Della-Sala, & Spinnler, 1986). Moreover, this process is also based on stages of activation from perception to processing and finally to output. In order to trace the activation of the brain online during reading, the time units of the brain sample must be smaller. A technology recently adopted in reading research capable of overcoming some of these time resolution limitations is electroencephalogram (EEG) utilizing the Evoked Related Potential (ERP) methodology. As visual and acoustic activity during decoding words lasts for mere tenths of milliseconds or less, detailed measurement of neural processing during the word

224

Z. Breznitz and L. Fabian

decoding process requires accurate time resolution. Compared with PET and fMRI measures that reflect hemodynamic and/or metabolic changes associated with neuronal activities with high-quality localization information but reduced temporal resolution in the order of seconds, ERPs have temporal resolution on the order of milliseconds with modest localization accuracy. EEGs record electrical signals during brain activity every 3.9 ms. This method permits direct observation of information processing at different levels of analysis, and can provide crucial information by means of real-time imaging of the neural system’s responses to sensory stimulation (Bentin, 1989; Zhang & Wang, 2007). ERPs are extracted from EEG data by means of averaging the brain responses to a number of equivalent trials in a given experiment. ERPs consist of various discrete components, or brain waves, that can be related to different stages of information processing in terms of amplitude and/or latency variations. The components are usually designated by their polarity (P, N) and by the latency of their maximal amplitudes in milliseconds. Areas of brain specialization can be identified by observing variations of amplitude and latency in ERP components across different scalp locations (see Halgren, 1990). ERP components reflect the time course of sensory and cognitive processes with millisecond resolution and the data is complementary to behavioral data in cognitive research. ERP components have been identified in different studies, which appear to be characteristic of certain types of brain activity during the reading process (see Shaul, 2007 for review). Each component is considered to be associated with a specific stage of activation in the reading process within an explicit time window and is complimentary to behavioral reaction time and accuracy that can only be seen at the output stage. The reading related components are P100-N100 (Johnstone, Barry, Anderson, & Coyle, 1996), P200-N200 (Luck & Hillyard, 1994; Picton, 1995), P300 (Palmer, Nasman, & Wilson, 1994), N400 (Kutas & Delong, 2007), P600 (Key, Dove, & Maguire, 2005) and MMN (Kujala, Belitz, Tervaniemi, & Na¨a¨ta¨nen, 2003). 11.2.3.1 ERP Research Evidence Weber-Fox and Neville (1996) used ERP methodology in a study carried out on adult Chinese-English bilinguals who had been exposed to their second language (English) at different points in their development. ERPs were obtained as subjects read sentences in English that included either semantic anomalies or syntactic violations. The event-related potentials showed that both early and late acquirers of L2 processed semantic anomalies similarly, although significant differences between the two groups were found while processing grammatical anomalies. The amplitudes and distributions of the N4 component obtained in ‘‘early’’ and ‘‘late’’ bilinguals were not significantly different from the responses of English monolinguals. That is, in English monolingual subjects, event-related brain potentials (ERPs) indicated greater activity in left posterior structures during the processing of the semantic components of

11

Reading in L1 and L2: Behavioral and Electrophysiological Evidence . . .

225

language (nouns, verbs, adjectives) and greater activity in left frontal lobe structures during the processing of the syntactic components of language (articles, conjunctions, prepositions). In particular, results showed that bilingual subjects displayed a similar pattern of localization for semantic anomalies, regardless of the age of second language acquisition. However, peak latencies elicited in bilinguals exposed to English late in life (between 11 and 16 years) occurred later, suggesting a slight slowing in processing. In contrast, during the syntactic processing, similar patterns of localization were found only in ‘‘early’’ bilinguals. Even with short delays in language exposure, altered ERPs were observed. With increased delays in language exposure, less activity in left frontal lobe structures was evident and greater changes in ERP asymmetries were found, with reduced specialization in left hemisphere language processing subsystems and increased right hemisphere involvement. The researchers concluded that these findings enabled them to distinguish between ‘‘balanced’’ bilingualism proper and the ability to speak a second language with quasinative fluency and accuracy. This study indicated that there might be differences in the cerebral cortical organization of languages according to age of acquisition and learning strategies (Fabbro, 2001). Hahne and Friederici (2001) also examined syntactic processing using auditory sentences that included different violations: syntactic, semantic or a combination of both. Subjects were native Japanese speakers who were late learners of German (L2) and they were compared to native German speakers. Results showed that when presented with syntactic violations, native German speakers showed ERP responses (biphasic LAN and P600) that were absent among L2 learners. Additional studies of bilinguals have linked ERP changes to reading in L1 and L2 languages. Ardal, Donald, Meuter, Muldrew, and Luce (1990) found that the N400 latency of bilinguals to semantic incongruencies was longer than that of monolinguals. In addition, the N400 latency in response to incongruency in L2 was longer than that for L1. First-language ERPs (French/ English) were almost identical in the groups of bilinguals. No age of second language acquisition effects on N400 latencies or amplitudes were found. Kluender and Kutas (1993) also reported an overall N400 reduction for the bilinguals’ processing of the less efficient language. Chen, Shu, Liu, Zhao, and Li (2007) conducted a more recent ERP study on L2 learning. Subjects were native Chinese speakers (L1) with English serving as their second language L2 (late learners) compared to native (L1) English speakers. All were presented with sentences in English which varied in subject-verb agreement. In general, results showed different ERP responses during specific syntactic processing of features not present in L1 (Chinese) among late English L2 learners as compared to native English speakers. Consistent with previous research on syntactic processing (e.g., Hahne & Friederici, 1999; Osterhout & Mobley, 1995; Weber-Fox & Neville, 1996), a biphasic left anterior negativity (LAN) P600 syntactic processing profile was found among the native English speakers. In comparison, the English L2 learners showed an N400 followed by a

226

Z. Breznitz and L. Fabian

P600 pattern, specifically while processing grammatically incongruent sentences. These ERP differences were present even when the behavioral data did not show a distinction between L2 and L1 English learners. These findings support the assumption that at a neural level, L1 language experience has an effect on L2 processing. In conclusion, these studies provide electrophysiological evidence supporting the assumption that there are differences between L1 and L2 that can be explained, at least in part, by the time of acquisition (age), amount of experience one has with the second language, degree of proficiency, cortical localization activities, and activation patterns.

11.3 Orthography Type of orthography is considered to be an additional aspect influencing reading in L2. In past studies, the role of orthography in bilingual reading proficiency was not thought to be an influencing factor leading to differences in proficiency between L1 and L2. It can be assumed that orthography may have been overlooked due to the fact that most of the comparisons made were between languages with similar orthographies, such as English and French. In particular, most of the studies were carried out on languages based on Roman scripts, which may influence the results. In addition, the bilinguals participating in these studies have many different levels of proficiency in their second language (Oren & Breznitz, 2005). The first to examine orthography as an influencing factor on both accuracy and rate among Hebrew L1 and English L2 speakers were Shimron and Sivan (1994). These researchers conducted two behavioral experiments in order to examine whether the orthography of readers’ first and second languages affects their reading time and comprehension in each language. They tested subjects who were considered both highly skilled and fluent bilinguals, with either Hebrew or English serving as their native language. In both experiments subjects read texts both in English and in Hebrew (with and without diacritic marks). Results showed that the native English speakers read the English texts significantly faster than the native Hebrew speakers. However, contrary to expectations, the native Hebrew speakers read the English (L2) texts faster than the Hebrew (L1) texts. The difference in reading rates and processing times between L1 and L2 was explained by the researchers as being attributed to the dissimilar orthographic features of each language. In relation to reading in general, this study implies that the mechanism underlying speed of word reading is influenced in part by the different orthographic features of languages, in addition to variance in individual ability (Breznitz, 2001; Breznitz, 2003 for review). In other words, in addition to what is known of the critical ages of language acquisition, implementing learning strategies, motivational factors, different types of scripts may also be related to differences in the rates of reading words.

11

Reading in L1 and L2: Behavioral and Electrophysiological Evidence . . .

227

The Script (Orthographic) Dependent Hypothesis (Frost, 1994) claims that the specific characteristics of a particular orthography lead to the use of different reading strategies. In languages with an irregular orthography, such as English, accurate word recognition skills develop more slowly in comparison to languages with more regular orthographies, such as vowelized Hebrew (Gholamain & Geva, 1999). With relation to reading difficulties, it is claimed that each language has special orthographic features, and therefore the connection between graphemes and phonemes in the different languages is the cause of different learning and reading problems (Abu-Rabia, 1997). In addition, AbuRabia (1997) states that some disabled learners who are able to implement high compensatory strategies in their first language are not recognized as having difficulties with L1, and often experience problems during second language acquisition and foreign language learning. With regard to the above, it is important to fully understand the different types of orthographies that exist. Different languages are generally categorized into two main groups based on their orthography. The first group is comprised of languages with phonologically deep orthographies, which means that there is great variance in the relationship between sounds and letters. In other words, a reader is not able to match the sound to the orthography in a simple manner. The second group consists of phonologically shallow orthographies, meaning that there is practically a one-to-one relationship between letters and sounds (phonemes), which minimizes the probability of phonological coding errors (Abu-Rabia, 1997; Grigorenko, 1999).

11.4 Reading Disabilities It is impossible to speak of second language acquisition (L2) among adult readers without addressing reading difficulties (disabilities) in first language (L1) and how the two are related. It is widely accepted that readers who experience difficulty in their native language (L1), are more likely to experience similar difficulties when acquiring an additional language (L2) (Breznitz et al., 2004). In particular, research on second language acquisition has shown that those with difficulties in their first language (L1) are prone to similar difficulties when learning a second language (L2). Reading disabilities in second language coexist with difficulties experienced in the native language (Ganschow et al., 1991). Most of the research on adults with learning disabilities has focused on dyslexics and has suggested that those diagnosed as dyslexic in childhood continue to suffer from dyslexia throughout their lifetime (Bruck, 1998; Lefly & Pennington, 1991). The current definition of dyslexia as suggested by the British Psychological Society (1999) states that ‘‘dyslexia is evident when accurate and fluent word reading and/or spelling develops very incompletely or with great difficulty’’ (p. 18). The adult dyslexic population can be divided into two types based on their reading proficiency. The first type consists of those who have been able to achieve a high level of reading proficiency, enabling them to continue on

228

Z. Breznitz and L. Fabian

to higher education, and are considered to be compensated dyslexics. The second type consists of those with reading skills that have remained at a low level of proficiency, meaning that they are unable to integrate into and adapt to formal learning environments (Miller-Shaul, 1999; Pennington et al., 1986). It is important to note that the majority of studies on adult dyslexics used compensated subjects belonging to the first type. This is the result of convenience as researchers have easier access to this population. Several studies have shown that approximately 25% of adult dyslexics are able to achieve reading levels similar to their peer reading levels (Felton, Naylor, & Wood, 1990; MillerShaul, 1999; Scarborough, 1984). However, even when these compensated adult dyslexics achieve acceptable or even high levels of reading performance, their cognitive profile remains stable showing evidence of the cognitive deficits characteristic of dyslexic individuals (Breznitz & Meyler, 2003). In light of this fact, it is important to fully understand the specific difficulties and obstacles that compensated adult dyslexics face when reading. Research has indicated that the most prominent and persistent difficulty that adult dyslexics continue to face are with the phonological processing skill (Bruck, 1992; Fawcett & Nicolson, 1991; Snowling, 1995) and that they continue to experience difficulty while performing tasks that rely a great deal on phonological skill, for instance tasks that require decoding of pseudowords, reading unfamiliar words, in addition to spelling tasks (which rely on orthographic ability as well). It is important to note that many adult dyslexics are able to reach high levels of reading comprehension, even though they experience difficulties with decoding. This may indicate reliance on contextual cues, which is considered to be a high-order skill, and may help compensate for problems with lower-level decoding skills (Stanovich, 1985). According to the British Psychological Society (1999), inaccuracy and the lack of fluency of word reading are currently considered to be the primary diagnostic features of a specific reading disability. Although research has suggested that the reading disabled can be categorized into different subtypes, it is commonly agreed that continuing deficits in decoding accuracy and speed of word recognition are the major characteristic of adult dyslexics (Breznitz, 2003; Lovett et al., 1994). It has been shown that dyslexic adults usually show improvement in word recognition accuracy, while their word reading rate continues to remain slower in comparison to regular non-dyslexic adult readers (Ben-Dror, Polatsek, & Scarpati, 1991). However, it seems that the word recognition difficulties encountered while reading are not the same across the different levels of reading. Specifically, the most prominent difficulties appear when performing a reading task that requires the decoding of pseudowords, and to a lesser degree when reading single words, and the least amount of word recognition difficulties occur while reading texts (Oren & Breznitz, 2005). A more recent theory regarding the factors that bring to dysfluent word decoding among dyslexic readers was suggested by Breznitz (2003 and 2006 for review). This theory, known as the Asynchrony Theory suggested that fluent and

11

Reading in L1 and L2: Behavioral and Electrophysiological Evidence . . .

229

accurate word decoding is derived from appropriate synchronization of the speed at which different brain systems are activated during reading. As such, the Asynchrony Theory puts forward the idea that speed of processing (SOP) and speed synchronization of the visual and auditory acoustic cerebral systems are crucial for effective word decoding to occur. A within modality (stage of activation) and between modality SOP gap leads to the ‘‘Asynchrony Phenomenon’’, which prevents the matching between graphemes and their phoneme representations and impairs the word decoding process (see Breznitz, 2006 for review). Based on an abundance of evidence supporting the phonological deficit hypothesis as an explanation for dyslexia, it has been suggested that the manifestation of reading difficulties differs across languages, depending on the depth of the phonological demands imposed by each linguistic system (Da Fontoura & Siegel, 1995; Geva & Wade-Woolley, 1998). However, based on the Asynchrony Theory, it can hypothesized that the orthographic structure of a language might also contribute to the speed at which the brain systems process reading and reading-related information that differs from language to language and between regular and dyslexic readers. These claims were the focus of studies conducted by Breznitz et al. (2004) and Oren and Breznitz (2005). In accordance with our understanding of adult dyslexia, as well as with research on second language, one can ask whether bilinguals can attain equivalent levels of reading fluency in both languages, or in our terms whether SOP synchronization within and between the relevant modalities in L1 is similar to that in L2. Breznitz et al. (2004) and Oren and Breznitz (2005) conducted studies in an attempt to verify whether Hebrew (L1) and English (L2) are processed in the same manner among bilingual regular readers as compared to bilingual dyslexic readers. In particular, the main focus of both studies was to verify whether the inaccuracy and slowness characterizing adult dyslexic readers in their first language (Hebrew) when reading words (Breznitz et al., 2004) and sentences (Oren & Breznitz, 2005) were also apparent in second language (English). Furthermore, Oren and Breznitz (2005) also sought to explore whether the intensity of brain activity in L1 (Hebrew) and L2 (English) during reading of unexpected sentence endings among bilingual regular and dyslexic readers is similar.

11.4.1 Hebrew and English It is important to understand the specific differences between both languages as Hebrew constitutes the official language of Israel while English serves as an official second language, and is mandatory for all Israeli pupils from the fourth grade. It is not surprising that some students experience difficulty acquiring English as a second language. Abu-Rabia (1997) claimed that the differences between the two orthographies, have served as the primary obstacle to learning English as a second language effectively.

230

Z. Breznitz and L. Fabian

It is important to compare these two languages because they are very different phonologically, orthographically and morphologically speaking (Ben-Dror, Bentin, & Frost, 1995; Ben-Dror, Frost, & Bentin, 1995; Bentin & Frost 1995; Deutsch & Rayner, 1999; Gronau & Frost 1997; Levin & Landsmann, 1989; Ravid, 1996; Shatil, 1996; Shimron, 1999). English orthography uses a Latin alphabet, whereas Hebrew uses a unique alphabet (Abu-Rabia, 1997; Shimron & Sivan, 1994). English orthography is phonologically deep and very irregular, therefore reading is more complex. When learning to read English, one must form a mental map of the different connections between phonemes (sounds) and graphemes (letters) (Grigorenko, 1999). In other words, each English grapheme represents a phoneme. In comparison, Hebrew is orthographically different, and is comprised of two writing systems that differ in their degree of print-to-sound correspondence (transparency). Namely, each Hebrew grapheme represents a syllable, including both vowel and consonant information. That being the case, there is vowelized (pointed) Hebrew and unvowelized (unpointed) Hebrew. The first is considered to have shallow and regular orthography, whereas the second is considered to have deep and often quite ambiguous orthography. Most often, vowelized Hebrew is found in children’s books and sacred scriptures and nonvowelized Hebrew is used in daily communication. Reading vowelized Hebrew requires the reader to employ bottom-up processing and to rely on phonological information, in comparison to reading unvowelized Hebrew, where top-down processing is needed along with strong use of lexical-semantic-contextual information (Frost, 1994; Shimron, 1999). In comparison to English, accuracy in decoding is less demanding in Hebrew, however, acquiring adequate reading rate is important in both orthographies (Shatil, 1996). With regard to reading, there are a few levels that can be examined: letters, words, and sentences. With this in mind, it is important to understand that Hebrew and English require the reader to employ different methods as there are a number of orthographical features in each language that influence reading proficiency (speed and decoding efficiency). Typographical letter structure is not the same in both languages, with Hebrew letters being more uniform and possessing more vertical and horizontal strokes, and less diagonals and curves, as compared to English. This uniformity in letter structure could negatively influence letter recognition, thus slowing reading time (Shimron & Sivan, 1994). At the word level, degree of word ambiguity also influences word reading, especially in nonvowelized Hebrew, where the reader must rely on the context in which the word is read, due to the fact that in the absence of diacritic marks one word may have multiple meanings. Affixation is another feature inherent in Hebrew. The use of affixes can assist the reader with the recognition of unvowelized single words; however, it also adds more information to the word. For example, Hebrew verbs are conjugated according to gender, number, tense and person; adjectives are conjugated according to gender and number; and nouns are conjugated according to relation, such as location and ownership. In other words, in Hebrew each word carries multiple sources of information, perhaps

11

Reading in L1 and L2: Behavioral and Electrophysiological Evidence . . .

231

impeding cognitive processing in reading and slowing it down (Shimron & Sivan, 1994). Finally, the direction of reading may also affect reading rate. A body of evidence shows that when words are presented in the right visual field, they are usually better identified as compared to left visual field presentation. Hebrew is read from right to left, which means that the graphemes are presented first in the left visual field, whereas English is read from left to right, meaning that the words appears first in the right visual field. This may influence word recognition and may explain to some extent why English is read faster than Hebrew (Shimron & Sivan, 1994; Silverberg, Bentin, Gaziel, Obler, & Albert, 1979). Breznitz et al. (2004) and Oren and Breznitz (2005) utilized behavioral, electrophysiological (ERP) and low resolution electromagnetic tomography (LORETA) (Pascual-Marqui, Michel, & Lehmann, 1994) methodologies in an attempt to examine the differences between bilingual regular and dyslexic readers when reading sentences and single words, respectively. The subjects were 50 university students (mean age 25.6 years, SD ¼ 2.14) divided into two groups: 25 dyslexic and 25 chronologically age-matched regular readers. All subjects were ‘‘subordinate bilinguals’’ who acquired Hebrew (L1) prior to English (L2). In addition, all subjects were from middle-class backgrounds, right-handed and displayed regular or corrected-to-regular vision in both eyes. None of the participants had a history of neurological or emotional disorders and all were paid volunteers. The dyslexic readers were recruited through the Haifa University Student Support Service in the north of Israel that aids students with learning disabilities; they had all been diagnosed as dyslexic in childhood and were classified as impaired readers by the Student Support Service. To control for gender differences, each group contained only males. The subjects were tested using different Hebrew and English tests as well as lexical decision paradigms in Hebrew (Breznitz, 1997) and English (Shatil, 1997a, 1997b, 1997c) using electrophysiological parameters and ERP and LORETA solution methodologies. At the word (Breznitz et al., 2004) and sentence (Oren & Breznitz, 2005) levels, results indicated discrepancies in the processing profiles of the dyslexic and regular bilingual readers, in both first and second languages. In general, the amplitudes of an early N1 component, a P2 component, and a later P3 component were higher and latencies were longer among dyslexic readers during processing of information in first and second languages (L1 and L2), but the differences between the two groups were more pronounced in English (L2). In addition, the results indicated that the localization of brain activation exhibited by current density in Hebrew and English was similar for regular but different for dyslexic readers. The results of both studies were in line with the ‘‘Linguistic Interdependence Hypothesis’’ (Cummins, 1979) suggesting that competence in L1 is related to competence in L2 among regular readers only. The dyslexic readers were significantly slower, less accurate and displayed later ERP components as compared to regular readers, and they displayed a significant disadvantage when reading English as compared to Hebrew. Furthermore, their patterns of brain activation in the two languages were less related and revealed less cross-language transfer of reading related skills from L1 to L2.

232

Z. Breznitz and L. Fabian

In sum, these studies revealed the differential effects of orthography on the cognitive processes underlying reading ability among skilled and impaired readers and demonstrated that the characteristics of the deficits in L1 and L2 among dyslexic readers are similar. It would seem that skilled readers should be able to master reading in English as L2 with the same level as in Hebrew L1. Moreover, among impaired readers, reduced automaticity coupled with higher phonological demands of the irregular English orthography may lead to additional reading difficulties than in otherwise shallow and consistent orthographies.

11.4.2 The Current Study Previous results indicated the ability of the N1 component to differentiate between dyslexic and regular readers at the word level in both Hebrew (L1) and English (L2), with more pronounced differences in English. Based on these findings, the current study examined whether the ERP component can also differentiate between the two groups at the letter level, when the only requirement is symbol perception and there is no semantic processing. As letters in Hebrew and English sound similar, the ability of dyslexics as compared to regular readers to distinguish between letters in the two languages was examined in the visual domain only. An oddball paradigm was employed and behavioral (reaction time and accuracy) and electrophysiological (ERP) measures were obtained. Participants were the ‘‘subordinate bilingual’’ subjects from Breznitz et al. (2004) and Oren and Breznitz (2005), who had acquired Hebrew (L1) prior to English (L2). Two experiments for letter discrimination and perception were administered. One contained 2 English letters and the other 2 Hebrew letters. Stimuli consisted of 120 items in each experiment: 40 target and 80 no-target letters presented to the subjects in the center of the computer screen. Each letter was 1/4 mm in height and the target letter in each language was ‘‘bet’’ in Hebrew and /b/ in English while the non-target was ‘‘dalet’’ in Hebrew and /d/ in English. In each language, stimulus duration was 170 ms and the ISI was 1,000 ms. Subjects were asked to press a button in response to a target stimulus and to ignore the non-target stimuli. In the two experiments, two pronounced ERP components were identified for both groups of subjects. An early N100 component evoked at approximately 100 ms post stimuli onset (Hyde, 1997) and a later P300 component evoked at approximately 300 ms post stimuli onset (Erez & Pratt, 1992; Palmer et al., 1994). Group (dyslexics controls) language (English Hebrew) repeated measures MANOVAs were carried out in order to evaluate differences between groups and languages in accuracy, reaction time, amplitudes and peak latencies of each component for each of the experimental tasks. The analysis for the N100 was carried out on the T5 and T6 electrodes (Maurer, Feldmann, Bromme, & Kalenka, 2005) and for the P300 on the Cz and Pz electrodes (Palmer et al., 1994). Table 11.1 present the means and standard deviations for the ERP amplitudes and latencies and the behavioral reaction time and accuracy among dyslexics and controls in each experiment for each component.

416.69 (73.64) 99.07 (1.14) 144.01 (29.45) 136.28 (36.21) 395.84 (67.16) 400.84 (62.12) –2.01 (5.42) –1.07 (2.84) 8.08 (4.65) 9.09 (7.37)

Dyslexic

444.25 (74.90) 97.38 (8.97) 166.98 (33.42) 152.09 (36.76) 483.73 (58.86) 476.49 (67.16) –4.12 (3.68) –1.13 (3.32) 7.45 (4.90) 10.00 (4.90)

* p < 0.5; ** p < 0.01; *** p < 0.001

Reaction time Accuracy N100 latency T5 N100 Latency T6 P300 latency Cz P300 Latency Pz N100 amplitude T5 N100 amplitude T6 P300 amplitude Cz P300 amplitude Pz

Language

English

359.29 (39.90) 99.79 (1.05) 99.11 (27.09) 106.12 (30.11) 364.35 (42.87) 362.40 (38.42) –4.47 (8.33) –1.33 (8.75) 10.30 (11.62) 13.13 (8.63)

Regular 366.71 (60.76) 99.54 (0.85) 101.09 (26.45) 108.34 (23.78) 376.08 (42.73) 372.08 (42.73) –3.61 (5.11) –1.13 (5.02) 8.63 (9.58) 12.19 (8.13)

18.46*** N.S 4.42* 15.15*** N.S 3.50*

13.84** N.S N.S

Main effect of language

4.25* N.S 14.21**

Main effect of group

Table 11.1 Mean and SD of English and Hebrew visual letters processing Hebrew English Hebrew

N.S

N.S

5.92*

4.15* N.S 11.98**

Group language interaction

11 Reading in L1 and L2: Behavioral and Electrophysiological Evidence . . . 233

234

Z. Breznitz and L. Fabian

As it can be seen from Table 11.1, when processing letters in both languages, the dyslexic readers were found to be significantly slower than regular readers despite a near perfect level of accuracy. Later N100 and P300 latencies as well as later reaction times were found among the dyslexic readers as compared to the control group. Moreover, when comparing the processing time between Hebrew and English, data revealed significant differences only within the dyslexics group. The dyslexics processed the English letters slower than the Hebrew letters. This slowness begins at the written symbol perception stage (N100), continues on to the processing stage (P300) and then on to the reaction time. Even though in the current study the dyslexic group was classified as bilingual, it seems that even simple processing of an alphabetic symbol in the second language (English) was slower as compared to the first language (Hebrew). Based on our current and previous results (Breznitz et al., 2004; Oren & Breznitz, 2005), it is clear that subordinate bilinguals, with a high level of proficiency in both their native language, Hebrew (L1) and their second language, English (L2), manage to achieve automaticity when decoding the alphabetic code at the letter level in both languages. However, when semantic processes are introduced into reading at higher levels (words and sentences), reading in L2 was found to be slower in comparison to reading in L1. Nevertheless, based on the high level of decoding accuracy of the regular readers in L2, it can be argued that this slowness is within the limitations of the information processing system and does not affect their L2 reading. At the same time, the dyslexic subordinate bilinguals with Hebrew (L1) as their native language and English (L2) as their second language were less accurate and slower as compared to the regular readers in both languages and at all reading levels. Moreover, for the dyslexics, the between-language reading rate gap from letters to words to sentences was wider as a result of slow L2 reading. This slowness affects their level of accuracy. These findings provide initial support for the Asynchrony Theory (see Breznitz, 2006) regarding the slowness of dyslexic readers when reading L2 materials. In conclusion, this chapter provides a comprehensive review of the theories regarding reading abilities among bilinguals, both regular and dyslexic readers. Different behavioral and brain imaging studies were presented throughout this chapter, providing support for these theories.

References Abu-Rabia, S. (1997). Verbal and working-memory skills of bilingual Hebrew-English speaking children. International Journal of Psycholinguistics, 13(1), 25–40. Abutalebi, J., Cappa, S. F., & Perani, D. (2001). The bilingual brain as revealed by functional neuroimaging. Bilingualism: Language and Cognition, 4, 179–190. Ardal, S., Donald, M. W., Meuter, R., Muldrew, S., & Luce, M. (1990). Brain responses to semantic incongruity in bilinguals. Brain and Language, 39(2), 187–205.

11

Reading in L1 and L2: Behavioral and Electrophysiological Evidence . . .

235

Baddeley, A. D., Logie, R., Bressi, S., Della-Sala, S., & Spinnler, H. (1986). Dementia and working memory. Quarterly Journal of Experimental Psychology: Human Experimental Psychology, 38(4-A), 603–618. Ben-Dror, I., Bentin, S., & Frost, R. (1995). Semantic, phonologic, and morphologic skills in reading disabled and regular children: Evidence from perception and production of spoken Hebrew. Reading Research Quarterly, 30(4), 876–893. Ben-Dror, I., Frost, R., & Bentin S. (1995). Orthographic representation and phonemic segmentation in skilled readers: A cross-language comparison. Psychological Science, 6(3), 176–181. Ben-Dror, I., Polatsek, A., & Scarpati, S. (1991). Word identification in isolation and in context by college students. Brain and Language, 31, 308–327. Bentin, S. (1989). Electrophysiological studies of visual word perception, lexical organization, and semantic processing: A tutorial review. Language and Speech, 32, 205–220. Bentin, S., & Frost, R. (1995). Morphological factors in visual word identification in Hebrew. In L. B. Feldman (Ed.), Morphological aspects of language processing (pp. 271–292). Hillsdale, NJ: Lawrence Erlbaum. Brandeis, D., & Lehmann, D. (1994). ERP mapping: A tool for assessing language disorders. In H. J. Heinze, T. F. Mante, & G. R. Mangun (Eds.), Cognitive electrophysiology (pp. 242–250). Boston: Birkhauser. Breznitz, Z. (1997). Parsing tests. Unpublished tests, University of Haifa, Mt. Carmel, Israel. Breznitz, Z. (2001). The determinants of reading fluency: A comparison of dyslexic and average readers. In M. Wolf (Ed.), Dyslexia, fluency and the brain (pp. 245–276). Cambridge, MA: York Press. Breznitz, Z. (2003). Speed of phonological and orthographic processing as factors in dyslexia: Electrophysiological evidence. Genetic, Social and General Psychology Monographs, 129(2), 183–206. Breznitz, Z. (2006). Fluency in reading: Synchronization of processes. Mahwah, NJ: Lawrence Erlbaum and Associates. Breznitz, Z., & Meyler, A. (2003). Speed of lower-level auditory and visual processing as a basic factor in dyslexia: Electrophysiological evidence. Brain & Language, 85(1), 166–184. Breznitz, Z., Oren, R., & Shaul, S. (2004). Brain activity of regular and dyslexic readers while reading Hebrew as compared to English sentences. Reading and Writing, 17(7–8), 707–737. Briellman, R. S., Saling, M. M., Connell, A. B., Waites, A. B., Abbott, D. F., & Jackson, G. D. (2004). A high-field functional MRI study of quadrilingual subjects. Brain and Language, 89, 531–542. British Psychological Society. (1999). Dyslexia, literacy and psychological assessment. Leicester: British Psychological Society. Bruck, M. (1992). Persistence of dyslexics’ phonological awareness deficits. Developmental Psychology, 28, 874–886. Bruck, M. (1998). Outcomes of adults with childhood histories of dyslexia. In C. Hulme & J. R. Malatesha (Eds.), Reading and spelling: Development and disorders (pp. 179–200). Mahwah, NJ: Lawrence Erlbaum. Chee, M. W. L., Tan, E. W. L., & Thiel, T. (1999). Mandarin and English single word processing studied with functional magnetic resonance imaging. The Journal of Neuroscience, 19(8), 3050–3056. Chen, L., Shu, H., Liu, Y., Zhao, J., & Li, P. (2007). ERP signatures of subject-verb agreement. Bilingualism: Language and Cognition, 10(2), 161–174. Cook, V. (1997). The consequences of bilingualism for cognitive processing. In A. de Groot & J. F. Kroll (Eds.), Tutorials in bilingualism: Psycholinguistic perspectives (pp. 279–300). Mahwah, NJ: Lawrence Erlbaum. Cummins, J. (1979). Linguistic interdependence and the educational development of bilingual children. Review of Educational Research, 49(2), 222–251.

236

Z. Breznitz and L. Fabian

Da Fontoura, H. A., & Siegel, L. S. (1995). Reading, syntactic, and working memory skills of bilingual Portuguese-English Canadian children. Reading and Writing, 7(1), 139–153. De Bleser, R., Dupont, P., Postler, J., Bormans, G., Speelman, D., Mortelmans, L., et al. (2003). The organization of the bilingual lexicon: A PET study. Journal of Neurolinguistics, 16, 439–456. Dehaene, S., Dupoux, E., Mehler, J., Cohen, L., Paulesu, E., Perani, D., et al. (1997). Anatomical variability in the cortical representation of first and second language. Neuroreport, 8(17), 3809–3815. Deutsch, A., & Rayner, K. (1999). Initial fixation location effects in reading Hebrew words. Language and Cognitive Processes, 14(4), 393–421. Erez, A., & Pratt, H. (1992). Auditory event-related potentials among dyslexic and normalreading children: 3CLT and midline comparisons. International Journal of Neuroscience, 63, 247–264. Fabbro, F. (1999). The neurolinguistics of bilingualism. UK: Psychology Press. Fabbro, F. (2001). The bilingual brain: Cerebral representation of languages. Brain and Language, 79, 211–222. Favreau, M., & Segalowitz, N. S. (1983). Automatic and controlled processes in the first- and second-language reading of fluent bilinguals. Memory and Cognition, 11(6), 565–574. Fawcett, A. J., & Nicolson, R. I. (1991). Vocabulary training for children with dyslexia. Journal of Learning Disabilities, 24(6), 379–383. Felton, R. H., Naylor, C. E., & Wood, R. B. (1990). Neuropsychological profile of adult dyslexics. Brain and Language, 39, 485–497. Friederici, A. D. (2002). Towards a neural basis of auditory sentence processing. Trends in Cognitive Science, 6, 78–84. Frost, R. (1994). Prelexical and postlexical strategies in reading evidence from a deep and shallow orthography. Journal of Experimental Psychology: Learning, Memory and Cognition, 20(1), 116–129. Ganschow, L., Sparks, R. L., Javorsky, J., Pohlman, J., & Bishop-Marbury, A. (1991). Identifying native language difficulties among foreign language learners in college: A ‘‘foreign’’ language learning disability? Journal of Learning Disabilities, 24(9), 530–541. Geva, E., & Wade-Woolley, L. (1998). Component processes in becoming English-Hebrew biliterate. In A. Y. Durgunoglu & L. Verhoeven (Eds.), Literacy development in a multilingual context: Cross-cultural perspectives (pp. 85–110). Mahwah, NJ: Lawrence Erlbaum Associates. Gholamain, M., & Geva, E. (1999). Orthographic and cognitive factors in the concurrent development of basic reading skills in English and Persian. Language and Learning, 49(2), 183–217. Green, D. W. (2003). The neural basis of the lexicon and the grammar in L2 acquisition. In R. van Hout, A. Hulk, F. Kuiken, & R. Towell (Eds.), The interface between syntax and the lexicon in second language acquisition (pp. 197–218). Amsterdam: John Benjamins. Grigorenko, E. L. (1999). Current trends in education in Russia: Preliminary outcomes indicative of students’ cognitive development. International Journal of Psychology, 34(3), 175–184. Gronau, N., & Frost, R. (1997). Prelexical phonologic computation in a deep orthography: Evidence from backward masking in Hebrew. Psychonomic Bulletin and Review, 4(1), 107–112. Hahne, A., & Friederici, A. D. (1999). Electrophysiological evidence for two steps in syntactic analysis: Early automatic and late controlled processes. Journal of Cognitive Neuroscience, 11, 193–204. Hahne, A., & Friederici, A. D. (2001). Processing a second language: Late learners’ comprehension mechanisms as revealed by event-related brain potentials. Bilingualism: Language and Cognition, 4, 123–141.

11

Reading in L1 and L2: Behavioral and Electrophysiological Evidence . . .

237

Halgren, E. (1990). Insights from evoked potentials into the neuropsychological mechanisms of reading. In A. B. Scheibel & A. F. Wechsler (Eds.), Neurobiology of higher cognitive function (pp. 103–150). New York: Guilford Press. Halsband, U. (2006). Bilingual and multilingual language processing. Journal of PhysiologyParis, 99, 355–369. Harris, R. J., & McGhee Nelson, E. M. (1992). Bilingualism: Not the exception any more. In R. J. Harris & E. M. McGhee Nelson (Eds.), Cognitive processing in bilinguals (pp. 3–14). North Holland: Elsevier. Hyde, M. (1997). The N1 response and its applications. Audiology and Neuro-Otology, 2, 281–307. Illes, J., Francis, W. S., Desmond, J. E., Gabrieli, J. D. E., Glover, G. H., Pildrack, R., et al. (1999). Convergent cortical representation of semantic processing in bilinguals. Brain and Language, 70(3), 347–363. Johnson, R. (1995). Effects of color on children’s naming of pictures. Perceptual and Motor Skills, 80, 1091–1101. Johnson, J. S., & Newport, E. L. (1989). Critical period effects in second language learning: The influence of maturational state on the acquisition of English as a second language. Cognitive Psychology, 21, 60–99. Johnstone, S. J., Barry, R. J., Anderson, J. W., & Coyle, S. F. (1996). Age-related changes in child and adolescent event-related potential component morphology, amplitude and latency to standard and target stimuli in an auditory oddball task. International Journal of Psychophysiology, 24, 223–238. Keatley, C. W. (1992). History of bilingualism research in cognitive psychology. In R. J. Harris (Ed.), Cognitive processing in bilinguals (pp. 15–49). Oxford, England: North-Holland. Key, A. P. F., Dove, G. O., & Maguire, M. J. (2005). Linking brainwaves to the brain: An ERP primer. Developmental Neuropsychology, 27, 183–215. Kim, H. S., Relkin, N. R., Lee, K. M., & Hirsch, J. (1997). Distinct cortical areas associated with native and second languages. Nature, 388(6638), 171–174. Klein, D., Milner, B., Zatorre, R. J., Meyer, E., & Evans, A. C. (1995). The neural substrates underlying word generation: A bilingual functional-imaging study. Proceedings of the National Academy of Sciences USA, 92(7), 2899–2903. Klein, D., Milner, B., Zatorre, R. J., Zhao, V., & Nikelski, J. (1999). Cerebral organization in bilinguals: A PET study of Chinese–English verb generation. Neuroreport, 10, 2841–2846. Klein, D., Watkins, K. E., Zatorre, R. J., & Milner, B. (2006). Word and nonword repetition in bilingual subjects: A PET study. Human Brain Mapping, 27, 153–161. Klein, D., Zatorre, R. J., Chen, J. K., Milner, B., Joelle, C., Belin, P., et al. (2006). Bilingual brain organization: A functional magnetic resonance adaptation study. Neuroimage, 31, 366–375. Klein, D., Zatorre, R. J., Milner, B., & Meyer, E. (1994). Left putaminal activation when speaking a second language: Evidence from PET. Neuroreport, 5(17), 2295–2297. Kluender, R., & Kutas, M. (1993). Bridging the gap: Evidence of ERPs on the processing of unbound dependencies. Journal of Cognitive Neuroscience, 5(2), 196–214. Kroll, J. F., & de Groot, A. M. B. (1997). Lexical and conceptual memory in the bilingual: Mapping form to meaning in two languages. In A. M. B. de Groot & J. F. Kroll (Eds.), Tutorials in bilingualism: Psycholinguistic perspectives (pp. 169–199). Mahwah, NJ: Lawrence Erlbaum Associates. Kujala, T., Belitz, S., Tervaniemi, M., & Na¨a¨ta¨nen, R. (2003). Auditory sensory memory disorder in dyslexic adults as indexed by the mismatch negativity. European Journal of Neuroscience, 17, 1323–1327. Kutas, M., & Delong, K. A. (2007). A sampler of event-related brain potential (ERP) analyses of language processing. In Z. Breznitz (Ed.), Brain research in language (pp. 153–186). New York: Springer.

238

Z. Breznitz and L. Fabian

Lefly, D. L., & Pennington, B. F. (1991). Spelling errors and reading fluency in compensated adult dyslexics. Annals of Dyslexia, 41, 143–162. Lenneberg, E. H. (1967). Biological foundations of language. New York: Wiley. Levin, I., & Landsmann, L. T. (1989). Becoming literate: Referential and phonetic strategies in early reading and writing. International Journal of Behavioral Development, 12(3), 369–384. Li, P., & Green, D. W. (2007). Neurocognitive approaches to bilingualism: Asian languages. Bilingualism: Language and Cognition, 10(2), 117–119. Lovett, M. W., Borden, S. L., DeLuca, T., Lacerenza, L., Benson, N. J., & Brackstone, D. (1994). Treating the core deficits of developmental dyslexia: Evidence of transfer of learning after phonologically- and strategy-based reading training programs. Developmental Psychology, 30(6), 805–822. Luck, S. J., & Hillyard, S. A. (1994). Electrophysiological correlates of feature analysis during visual search. Psychophysiology, 31(3), 291–308. Maurer, M. H., Feldmann, R. E., Bromme, J. O., & Kalenka, A. (2005). Comparison of statistical approaches for the analysis of proteome expression data of differentiating neural stem cells. Journal Proteome Research, 4, 96–100. Meschyan, G., & Hernandez, A. E. (2006). Impact of language proficiency and orthographic transparency on bilingual word reading: An fMRI investigation. NeuroImage, 29, 1135–1140. Miller-Shaul, S. (1999). Speed of visual-orthographic and auditory phonological processing as a factor in dyslexia: A comparison between young and adult dyslexics by means of electrophysiological measures. Unpublished MA thesis, University of Haifa, Israel. Oren, R., & Breznitz, Z. (2005). Task detection difficulty: Effects on ERPs in a same-different letter classification task. Journal of Neurolinguistics, 18(2), 127–151. Osterhout, L., & Mobley, L. A. (1995). Event-related potentials elicited by failure to agree. Journal of Memory and language, 34, 739–773. Palmer, B., Nasman, V. T., & Wilson, G. F. (1994). Reading processes in L1 and L2 among dyslexic as compared to regular bilingual readers: Behavioral and electrophysiological evidence. Biological Psychology, 38(2–3), 199–214. Pascual-Marqui, R. D., Michel C. M., & Lehmann, D. (1994). Low resolution electromagnetic tomography: A new method for localizing electrical activity in the brain. International Journal of Psychophysiology, 18(1), 49–65. Pennington, B. F., McCabe, L. L., Smith, S. D., Lefly, D. L., Bookman, M. O., Kimberling, W. J., et al. (1986). Spelling errors in adults with a form of familial dyslexia. Child Development, 57(4), 1001–1013. Perani, D., & Abutalebi, J. (2005). The neural basis of first and second language processing. Current Opinion in Neurobiology, 15, 202–206. Perani, D., Dehaene, S., Grassi, F., Cohen, L., Cappa, S. F., Dupoux., et al. (1996). Brain processing of native and foreign languages. NeuroReport, 7(15–17), 2439–2444. Perani, D., Paulesu, E., Galles, N. S., Dupoux, E., Dehaene, S., Bettinardi, V., et al. (1998). The bilingual brain: Proficiency and age of acquisition of the second language. Brain, 121(10), 1841–1852. Picton, T. W. (1995). The neurophysiological calculation of auditory discrimination. Ear and Hearing, 16, 1–5. Poeppel, D., & Hickok, G. (2004). Towards a new functional anatomy of language. Cognition, 92, 1–12. Ravid, D. (1996). Accessing the mental lexicon: Evidence from incompatibility between representation of spoken and written morphology. Linguistics, 34(6), 1219–1246. Sakai, K. L., Miura, K., Narafu, N., & Muraishi, Y. (2004). Correlated functional changes of the prefrontal cortex in twins induced by classroom education of second language. Cerebral Cortex, 14, 1233–1239. Scarborough, H. S. (1984). Continuity between childhood dyslexia and adult reading. British Journal of Psychology, 75(3), 329–348.

11

Reading in L1 and L2: Behavioral and Electrophysiological Evidence . . .

239

Schwartz, R. L. (1997). Learning disabilities and foreign language learning: A painful collision. Retrieved April 2000, from http://ldonline.com/ld_indepth/foreign_lang/painful_ collision.html Segalowitz, N., Poulsen, C., & Komoda, M. (1991). Lower level components of reading skill in higher level bilinguals: Implications for reading instruction. AILA Review, 8, 51–58. Shatil, E. (1996). Predicting reading ability: Evidence for cognitive modularity. Unpublished doctoral dissertation, University of Haifa, Mt. Carmel, Israel. Shatil, E. (1997a). One minute tests in Hebrew and English. Unpublished tests, University of Haifa, Mt. Carmel, Israel. Shatil, E. (1997b). Oral language test for English as a second language. Unpublished test, University of Haifa, Mt. Carmel, Israel. Shatil, E. (1997c). Basic screening tests. Unpublished battery, University of Haifa, Mt. Carmel, Israel. Shaul, S. (2007). Event related potentials (ERPs) in the study of dyslexia. In Z. Breznitz (Ed.), Brain research in language (pp. 51–92). New York: Springer. Shimron, J. (1999). The role of vowel signs in Hebrew: Beyond word recognition. Reading and Writing, 11(4), 301–319. Shimron, J., & Sivan, T. (1994). Reading proficiency and orthography: Evidence from Hebrew and English. Language Learning, 44(1), 5–27. Silverberg, R., Bentin, S., Gaziel, T., Obler, L. K., & Albert, M. L. (1979). Shift of visual field preference for English words in native Hebrew speakers. Brain and Language, 8(2), 184–190. Snowling, M. J. (1995). Phonological processing and developmental dyslexia. Journal of Research in Reading, 18(2), 132–138. Sparks, R. L., Ganschow, L., Patton, J., Artzer, M., Siebenhar, D., & Plageman, M. (1997). Prediction of foreign language proficiency. Journal of Educational Psychology, 89(3), 549–561. Sparks, R. L., Philips, L., Ganschow, L., & Javorsky, J. (1999). Comparison of students classified as LD who petitioned for or fulfilled the college foreign language. Journal of Learning Disabilities, 32(6), 553–565. Stanovich, K. E. (1985). Explaining the variance in reading ability in terms of psychological processes: What have we learned? Annals of Dyslexia, 35, 67–96. Ullman, M. T. (2001). A neurocognitive perspective on language: The declarative/procedural model. National Review of Neuroscience, 2, 717–726. Vellutino, F., & Scanlon, D. (1986). Linguistic coding and metalinguistic awareness: Their relationship to verbal and code acquisition in poor and regular readers. In D. Yaden & S. Templeton (Eds.), Metalinguistic awareness and beginning literacy (pp. 115–141). Portsmouth, NH: Heinemann. Wartenburger, I., Heekeren, H. R., Abutalebi, J., Cappa, S F., Villringer, A., & Perani, D. (2003). Early setting of grammatical processing in the bilingual brain. Neuron, 37, 159–170. Weber-Fox, C. M., & Neville, H. J. (1996). Maturational constraints on functional specializations for language processing: ERP and behavioral evidence in bilingual speakers. Journal of Cognitive Neuroscience, 8(3), 231–256. Zhang, Y., & Wang, Y. (2007). Neural plasticity in speech acquisition and learning. Bilingualism: Language and Cognition, 10(2), 147–160.

Chapter 12

Identification of Grammatical Functions in Two Languages Mark Leikin and Elina Ritvas

12.1 Introduction Reading processing in bilingual populations has been studied intensively in the past few decades (e.g., Bialystok, 2007; Perfetti et al., 2007). These studies, however, dealt mostly with relatively young readers and focused more on phonological/ orthographic and lexical-semantic properties than on morphological and syntactic aspects of second language (L2) reading. The morphosyntactic characteristics of sentence processing in later bilinguals have been studied far less, especially from the neurophysiological point of view (Clahsen & Felser, 2006a; Hahne & Friederici, 2001; Suh et al., 2007). However, many behavioral studies indicate that, in contrast to lexical-semantic aspects, syntactic and phonological characteristics of L2 are particularly difficult to master during late acquisition (Ellis, 2002; Hohenstein, Eisenberg, & Naigles, 2006). Neurocognitive studies, though less numerous, provide additional evidence for the view that one’s age at the time of L2 acquisition is critical for the mastery of that language and for the functional specialization of the language in the brain (for reviews, see Fabbro, 2001). Recent fMRI studies have shown that syntactic processing is supported by a frontotemporal network, with specific increased involvement of anterior portions of Superior Temporal Gyrus (STG) (Humpries, Kimberly, Buchsbaum, & Hickok, 2001; Friederici, Meyer, & von Cramon, 2000; Ru¨schemeyer, Fiebach, Kempe, & Friederici, 2005). Semantic processing, however, selectively recruits portions of the Inferior Frontal Gyrus (IFG) bilaterally, but with clear dominance in the left hemisphere (Bookheimer, 2002; Ru¨schemeyer et al., 2005). In this context, it was hypothesized (Liu & Perfetti, 2003) that other brain mechanisms may be involved in syntactic processing by late language learners, different from those involved in such processing by native speakers. Recent brain imaging studies investigating sentence processing seem to confirm that this is indeed the case (Friederici, M. Leikin (*) Laboratory for Neurocognitive Research of Giftedness, Department of Learning Disabilities, Faculty of Education, University of Haifa, Haifa, Israel e-mail: [email protected] M. Leikin et al. (eds.), Current Issues in Bilingualism, Literacy Studies 5, DOI 10.1007/978-94-007-2327-6_12, Ó Springer ScienceþBusiness Media B.V. 2012

241

242

M. Leikin and E. Ritvas

Hahne, & Saddy, 2002; Suh et al., 2007). In particular, it was reported that brain activation for L1 and L2 use shows little, if any, overlap in late bilinguals (e.g., Hahne & Friederici, 2001; Perani et al., 1998; Suh et al., 2007). Moreover, syntactic abilities in L2 seem to depend not only on the age of acquisition but also on the characteristics of bilinguals’ native language (e.g., Liu & Perfetti, 2003). Several previous (mostly neuropsychological) studies have suggested distinct non-overlapping cortical representations of the two languages in bilinguals (e.g., Paradis, 2004; Paradis & Goldblum, 1989; Suh et al., 2007), and different ERP patterns have been observed in first language processing of bilinguals and monolinguals (Donald, Meuter, & Ardal, 1986). However, several other studies found evidence of overlapping cortical representations in bilinguals (e.g., Sarfarazi & Sedgwick, 1996). Empirical findings of locations of activation for the two languages using fMRI have also been contradictory (e.g., Illes et al., 1999). The same results were obtained in behavioral studies of bilingual language processing: some have suggested separate activation of L1 and L2 (e.g., Gerard & Scarborough, 1989) and others (e.g., Dijkstra, 2001; Kroll & Tokowicz, 2001) simultaneous activation of the two languages. These contradictions have been explained by differences in methodologies and subject populations (e.g., Billingsley-Marshall, Sakari, Pataraia, & Papanicolau, 2006). It has also been suggested that it is necessary to distinguish between different bilingual processing styles for phonetic and syntactic information versus lexical-semantic information (e.g., Marian, Spivey, & Hirsch, 2003). Two additional factors may account for these contradictions: differences between L1 and L2 and differences in the type of bilingualism (e.g., early vs. later bilingualism). The present study proposed to investigate the processes of identification of a word’s grammatical functions by two comparable groups of participants.

12.2 Identification of the Grammatical Functions of Words: A Comparison Between Hebrew and Russian The act of sentence understanding requires, among other things, the processing of words in sentences so that words are assigned an immediate syntactic categorization, allowing attribution of their grammatical functions (e.g., subject and predicate). This, in turn, is used to construct a single preliminary phrase structure (Mitchel, 1987). The grammatical functions and syntactic structure of the sentence as a whole carry meanings of their own, which must be extracted by the reader when comprehending the sentence (Kako, 1999). Recent literature suggests that identification of the grammatical functions of words may be possible by reference to varied sources of information, such as word order, inflectional morphology, and the lexical-morphological properties of individual

12

Identification of Grammatical Functions in Two Languages

243

words (Clahsen & Felser, 2006a; Ferstl & Flores d’Arcais, 1999; Hahne, Mueller, & Clahsen, 2006). But various sources of information appear to contribute differently to this process in different languages. It has been suggested (Kempe & MacWhinney, 1999) that the greater the availability of a cue, the larger the processing benefits associated with the presence of such a cue and the smaller the effect of other converging information. In English, for example, the syntactic order of sentence components is usually fixed, so word order is highly important for sentence processing (Bates, Devescovi, & D’Amico, 1999). The situation is different, however, when the syntactic structures are not necessarily in a fixed order, as is the case in Hebrew and Russian (Berman, 1985; Wade, 1994). In Hebrew, as in other Semitic languages, derivational morphology is the most characteristic feature of the language (Deutsch, Frost, Pollatsek, & Rayner, 2005), so that recognition of the lexical-morphological characteristics of words makes available important syntactic information (Leikin, 2002; Shimron, & Sivan, 1994). Parsers may use ‘‘lemma information,’’ detailed lexical-syntactic information, to achieve the syntactic categorization of words (Mitchel, 1987). For example, they may use information about argument structures in which different verbs can take part in the thematic roles of syntactic arguments. The parser may attribute grammatical functions to words in sentences using the lexical entry of a word together with other sources of information (Leikin, 2002). Several studies investigating the role of morphological units in the Hebrew mental lexicon have suggested that verbs and nouns are organized differently (e.g., Frost & Granier, 2000; Deutsch et al., 2005). Particularly, it was concluded that word patterns in the nominal system do not govern the process of the lexical system, whereas verbal-pattern morphemes have a role in lexical organization within the verbal system. The Russian language belongs to the East Slav group of languages, itself part of the Slavonic branch of the Indo-European family. The stock of words in the Russian language has increased over time by borrowing, affixation, and composition procedures. Knowledge of the main principles of Russian affixation helps extend the vocabulary because it enables understanding the precise meaning of a word and recognizing the word’s relationship with other words derived from the same root. The reader must be able to identify the basic component of a Russian verb, noun, or adjective. Moreover, Russian is a highly inflected language. Meaning is much more dependent on the ending of words and less on word order than it is in English. Word order is much more flexible in Russian than it is in English, because in Russian it is primarily inflection that establishes the relationship between the words (Offord, 1998). A certain degree of similarity between Russian and Hebrew may be hypothesized because Russian is also characterized by rich derivational morphology and a mostly pragmatic syntactic order (Akhmanova, 1971; Berman, 1985; Wade, 1994).

244

M. Leikin and E. Ritvas

Recent studies by Breznitz and Leikin (Breznitz & Leikin, 2000, 2001; Leikin, 2002, 2008; Leikin & Breznitz, 1999, 2001) used the ERP technique to examine the contribution of the grammatical functions of words to sentence processing and to the identification of grammatical function among average Hebrew-speaking adult readers. Brain activity was examined during the processing of various parts of a sentence. The obtained effect concerned the amplitudes and latencies of the N100/P200, P300, and P600 ERP components. This effect was manifested in the form of distinct differences between three central parts of the sentence (subject, predicate, and object), with the largest differences observed for the predicate. The findings suggest that changes in ERP amplitudes and latencies did not relate to ordinal word position and word-class differences between target words, but were associated with the words’ grammatical functions. The effect was understood as an indication of the processes involved in the identification of the words’ grammatical functions (Leikin, 2002). In this case, Hebrew-speaking readers tended to utilize the predicate-oriented morphologically-based strategy for processing the grammatical functions of words. This strategy was reflected by the descending order of the grammatical functions of words according to the level of their activation: predicate > subject > direct object modifier. This strategy, however, seems to be language-related and complicated to a degree even for native speakers of Hebrew (Leikin & Breznitz, 2001; Leikin, 2008). Moreover, this strategy seems to develop at relatively later stages of language/reading acquisition (Sokolov, 1984), and Hebrew-speaking readers use several other procedures for processing the grammatical functions of words, including the word-order strategy (Leikin, 2002). Selection of a particular strategy may be influenced by different factors, including the lexical-morphological characteristics of the stimuli. Note, however, that the results obtained in Hebrew cannot be generalized or even adequately interpreted in the absence of data from other languages. Therefore, examination of the process of identification of the words’ grammatical functions in other languages becomes important for singling out the universal and specific features of sentence processing. Another aspect of the problem has to do with cognitive and neurophysiologic characteristics of bilingual language processing. Several studies on language processing by bilinguals have used ERP measures. Investigation of semantic aspects of language processing found the latency of the N400 component to semantic anomalies to be delayed in fluent bilinguals compared with monolinguals and in L2 compared with L1 within the bilingual group (Alvarez, Holcomb, & Grainger, 2003; Hahne et al., 2006; Weber-Fox & Neville, 1996). For phrase structure violations, monolinguals showed differential effects for an early left anterior negativity (N125) and a second left lateralized negativity (N300500), which were followed by a P600 effect (Hahne et al., 2006; Weber-Fox & Neville, 1996). Late L2-learners displayed a more bilateral ERP pattern for syntax-related negativity (N300500) and an absence of modulation of P600, which has been interpreted to reflect a process of syntactic

12

Identification of Grammatical Functions in Two Languages

245

reanalysis and repair (Friederici, 1999). Other studies, however, detected P600 effects in L2 learners in the same conditions as in native speakers (e.g., Hahne & Friederici, 2001). Considered together, these results suggest that morphosyntactic processing of L2 learners depends, among other things, on their language proficiency (Clahsen & Felser, 2006a). The present study aimed to examine brain activity through ERP amplitudes and latencies and by areas of brain activation in adult native (Russian and Hebrew) and bilingual (Russian/Hebrew) readers when processing the grammatical functions of words during the reading of sentences in different languages. Primarily, we proposed to investigate differences between processing strategies and patterns of brain activation in Hebrew and Russian as first languages (L1). We also intended to trace differences in ERP measures between processing patterns in first and second languages. We hypothesized that there are substantial differences between sentence processing in Hebrew and Russian and between processing in L1 and L2. Specifically, we suggested that readers identify the grammatical functions of words using morphologically based verb-oriented strategies in Hebrew and morphologically based but not verb-oriented strategies in Russian. We expected also that all characteristics of native and second language processing outlined above have clear neurophysiological correlates and are manifested by changes in the amplitudes and latencies of N100/P200, P300, and P600 ERP components, with variation in the localization of brain activation for L1 and L2/L3 sentence processing.

12.3 Method 12.3.1 Participants Thirty-six normally reading, male university students from middle-class backgrounds participated in this study, consisting of 16 native speakers of Hebrew (who had also participated in Leikin’s study, 2008), and 20 native speakers of Russian. The Russian group consisted of bilingual individuals who had immigrated to Israel at least 5 years earlier, have since acquired a thorough command of Hebrew (after puberty), and still read fluently in their native language. Adequate levels of reading proficiency were assessed for both groups in both languages using standard reading tests for adults (National Institute for Testing and Evaluation, 1996; Test for Russian as a First Language, 2006). Participants ranged in age from 18 to 29 years (see Table 12.1). Each participant was right-handed and displayed normal or corrected to normal binocular vision. Participants did not report any history of reading disabilities or hard neurological signs. All participants were paid volunteers from the University of Haifa.

246

M. Leikin and E. Ritvas

Table 12.1 Mean results of native Russian and Hebrew speakers in behavioral and readinga baseline tasks Russian Hebrew speakers speakers Measure M S.D. M S.D. t Age Raven matrices (raw scores) Digit span standard score Decoding, accuracy – words Decoding, speed – words (sec per item) Decoding, percent accuracy – pseudowords Decoding, speed – pseudowords (sec per item) Decoding, percent accuracy – connected text Decoding, output – connected text (words per minute) Reading comprehension (correct items)

25.9 58.6 11.5 98.9 0.6 98.8 0.7 99.5 270.2

1.82 15.33 3.70 2.03 0.11 1.87 0.16 0.46 30.69

25.9 70.0 12.7 99.4 0.5 95.5 1.2 99.6 141.7

2.06 19.86 2.18 0.59 0.08 4.47 0.23 0.47 23.86

0.14 1.96 1.13 1.10 3.74** 2.79* 7.50*** 0.17 13.85***

4.8

0.40

3.2

1.28

4.90***

* p < 0.05, ** p < 0.01, *** p < 0.001 a All the reading tasks were performed in the native language: Russian for native Russian speakers and Hebrew for native Hebrew speakers

12.3.2 Behavioral Baseline Measures 12.3.2.1 General Ability I. General nonverbal ability was assessed by the Raven Standard Progressive Matrices (Raven & Court, 1976), a nonverbal test for measuring IQ. II. Verbal Short Time Memory/Working Memory was assessed by the Digit Span Test (WAIS, Wechsler, 1994), in which the examinee is asked to repeat a number sequence identical to that presented by the examiner (in the same order for Digits Forward and in reverse order for Digits Backward). The longest correct sequence memorized was recorded and a standard score derived. 12.3.2.2 Reading Ability Several tests were used to obtain estimates of reading accuracy, time, and comprehension in Hebrew and Russian. The first set of tests provided measures of decoding accuracy for real words and pseudowords. The second set assessed reading time in context and evaluated reading comprehension. 12.3.2.3 Decoding Skills One Minute Tests (Shatil, 1997) for Hebrew language. This battery included two subtests in which participants were asked to read lists as quickly and accurately as possible within 1 min. The first list contained 100 real words

12

Identification of Grammatical Functions in Two Languages

247

arranged in order of increasing length (1 to 5 syllables) and decreasing frequency. The second list was comprised of 100 pseudowords arranged in order of increasing length (1 to 5 syllables). The decoding skills of the Russian (L1) group in Russian language were assessed by reading 40 pseudowords and 40 words arranged in order of increasing length (1 to 5 syllables) and decreasing frequency (Test for Russian as a First Language, 2006). Scores were based on the number of words/pseudowords read correctly. To obtain a comprehensive decoding score, Z-scores were first calculated for each of the tests (i.e., words and pseudowords for each language). 12.3.2.4 Reading Comprehension, Accuracy, and Speed in Context Text-reading performance was measured using 3 texts from the Reading Test section of the Israeli Psychometric SAT test (National Institute for Testing and Evaluation, 1996) containing one text in Hebrew and one in Russian (for the group of native Russian speakers). Each text contained a short story comprised of 25 sentences (about 300 words). The story was read silently in each language. Subsequently, 5 multiple-choice questions were presented. Comprehension scores were based on the total number of correct answers for each text.

12.3.3 Electrophysiological Baseline Measures 12.3.3.1 Target Detection Tasks These tasks were administered in order to habituate the participants to the experimental situation and to verify correct response based on brain activity (P300 component). The tasks were administered in auditory and visual modalities. For the auditory modalities, stimuli were 1000 and 3000 Hz tones presented consecutively through a PC speaker. For the visual modalities, stimuli were two Hebrew block letters, 0.64 cm (1/4 in) high, presented successively at the center of a computer screen. Stimulus presentation time depended upon the latency of the participants’ responses. Targets occurred 20% (n ¼ 20) of the time and non-targets 80% of the time (n ¼ 80). Stimuli were presented for duration of 250 ms at an ISI of 700 ms. Participants were asked to count the target stimuli and ignore the non-targets. 12.3.3.2 Experimental Stimuli Twenty five groups of sentences (75 sentences) in each of the two languages (Hebrew and Russian) were presented. Each group contained three sentences of 5–8 words each. Two multiple-choice questions were asked on each item. In each language the items were regular declarative sentences in SVO (subjectverb-object) syntactic order and included the three grammatical functions under study: subject, predicate, and direct object. Stimuli included the same words with interchangeable grammatical functions and positions in the sentences. In

248

M. Leikin and E. Ritvas

each group of three sentences the same word (noun) appeared in three different grammatical roles. Thus, the target word appeared at the beginning of the sentence (subject), in the intermediate or the terminal position (predicate; predicate nominative partly in Russian), and in the intermediate or the terminal position (object). For example, in Hebrew: ‫( את המוצרים השונים החבילה הכילה‬The package contained the various products), ‫( הקופסה הגדולה הייתה חבילה‬The big box was a package); ‫( אנחנו שולחים את החבילה לחברים‬We are sending the package to friends); and in Russian: Нефть разлилась по всему пляжу (The oil washed over the beach), Геологи нашли нефть в пустыне (Geologists found oil in the desert), Черная жидкость – нефть (The black liquid is oil). The length of the sentences in Hebrew and Russian were 5.7 and 5.9 words, respectively, with an average word length of 4.5 letters in Hebrew and 4.6 in Russian. We were only able to match word frequencies in Hebrew and Russian approximately, because there are no reliable data on word frequencies in Hebrew [except for the relatively old and small Balgur’s dictionary for school children (Balgur, 1968) (Zasorina, 1977)]. 12.3.3.3 Procedure Each testing session lasted approximately 60 min. Participants were seated in a quiet room, 1.5 m from a PC (Windows) computer screen. Experimental task presentation was counterbalanced. Participants were connected to an Electrocap and were instructed to remain quiet and refrain from moving during the testing session. They were also told that it was important to avoid excessive eye movements and blinking as much as possible. The items appeared one at a time on the computer screen. Participants were instructed to begin reading the item the moment it appeared on the screen. After reading the item, they pressed a button on the keyboard, whereupon the text was automatically erased and the multiple-choice questions were displayed. Participants indicated their answer by pressing a number on the keyboard (1–4) that corresponded to the answer chosen. Measures of comprehension and reaction time were determined for each reading item in the two presentation conditions. The two forms were counterbalanced across one experimental condition. 12.3.3.4 Window Presentation (With Electrophysiological Measures) The stimulus sentences appeared word by word. This manner of presentation was adopted in order to reduce eye movements by focusing the participant’s gaze on the center of the computer screen. As word presentation rates were calculated for each individual reader, presentation rates differed across participants. Thirty-two channels of electroencephalogram (EEG) were recorded using a Bio-Logic Brain Atlas III computer system with brain mapping capabilities. The system used a bandpass of 0.1–70 Hz interfaced with a 20-channel, 12-bit A/D converter. The EEGs were sampled at a rate of 250 Hz (dwell time ¼ 4.0 ms)

12

Identification of Grammatical Functions in Two Languages

249

beginning 100 ms before stimulus onset. A full array of electrodes was placed according to the International 10/20 system (Jasper, 1958) by means of an Electro-cap (a nylon cap fitted over the head with 9 mm tin electrodes sewn into it). Nineteen scalp electrodes were used according to standard 10/20 system locations: PF1, PF2, F7, F3, FZ, F4, F8, T3, C3, CZ, C4, T4, T5, P300, PZ, P4, T6, O1, O2, all referenced to an electrode on CVII (the seventh vertebra) and grounded to Fpz. In addition, one electrode was applied diagonally below the left eye to monitor eye movement. During data collection electrode impedances were kept below 5 K by first treating scalp areas with a mildly abrasive cleanser (Omni-Prep) and using an electrolyte gel (Electro-gel). Trial onset was marked on the Oz channel of EEG by a positive polarity 5 mV pulse delivered from an IBM-PC 486 computer. The pulse was delivered at the beginning of each word in each item. Signal averaging of the raw EEG data was performed off-line. EEG data were separated into discrete trials. After the eye movement correction, we determined the averages of the individual trials according to the experimental data set. There were three average trials, one for each of the three sentence elements across items (separately for each language). The averages were combined to form one set per participant, resulting in one data set for each participant and reflecting averaged EEG activity of the words representing subject, predicate, and object in two languages and for first and second language conditions. Evoked potentials were measured for each participant, for each word, in every item. Only single trials free from eye movements and associated with correct responses were averaged to obtain the event-related potentials. Grand averages over conditions and subjects were then performed for each experiment for each of the 19 scalp electrodes. ERP peaks were first identified and then validated by a machine-scoring algorithm. Latencies were measured from stimulus onset. Amplitudes were measured relative to the mean voltage of each channel during the pre-stimulus baseline.

12.4 Results 12.4.1 Background Behavioral Measures The performance of the Russian and Hebrew groups on behavioral cognitive and reading (in native language) baseline tasks is shown in Table 12.1. No differences were evident between the two groups on the general ability and working memory capacity tests, but significant differences were found between the two groups on most of the reading measures in their native languages (Russian and Hebrew, respectively). Russian readers exhibited significantly shorter reading times for most of the reading tasks than Hebrew readers did, and performed more accurately on these and on the reading comprehension test. Thus, the Russian readers were found to read more fluently and more accurately in their native language than Hebrew readers did in theirs. Russian speakers performed more poorly on Hebrew reading measures than in

250

M. Leikin and E. Ritvas

Table 12.2 Comparison of Russian speakers performing reading-related tasks in Russian and in Hebrew Russian Hebrew Measure M S.D. M S.D. t Decoding, percent accuracy – words Decoding, speed – words (sec per item) Decoding, percent accuracy – pseudowords Decoding, speed – pseudowords (sec per item) Reading comprehension (correct items)

98.9 0.7 98.8 0.7 4.8

2.03 0.11 1.87 0.16 0.40

96.0 1.3 91.5 1.7 2.8

2.60 0.40 4.00 0.44 1.06

4.08* 6.01** 9.13** 12.61** 7.44**

* p < 0.01, ** p < 0.001

their native language. They were significantly less accurate in reading Hebrew words and pseudowords and on the reading comprehension test. Moreover, the Russian speakers read significantly slower in Hebrew than in Russian on timed tests for words and pseudowords (Table 12.2).

12.4.2 Experimental Tasks In both groups of participants, P100, P200, P300, N400, and P600 ERP waves were identified for subject, predicate, and object in each sentence, in all reading items, and in both languages (see Figs. 12.1 and 12.2 for examples). Baseline adjustment was performed according to the average values, which were measured in the interval between 200 ms and stimulus appearance separately for each grammatical function. The most prominent ERP appeared in the Cz electrode. Mixed factorial ANOVAs were performed to compare brain activity between both groups of participants when reading sentences in their native languages. At the first stage, two ANOVAs were performed, separately for amplitudes and latencies. ERP component, grammatical role, and scalp region served as withinsubject factors, while group served as the between-subject factor. Additionally, mixed factorial ANOVAs were performed to compare brain activity among Russian speakers when reading sentences in Hebrew and Russian. Furthermore, mixed factorial ANOVAs were performed to compare brain activity among Russian speakers and Hebrew speakers when reading sentences in Hebrew. ERP component, language, grammatical role, and scalp region served as withinsubject factors, while group served as the between-subject factor.

12.4.3 Comparison of Brain Activity Between Two Groups of Participants When Reading Sentences in Their Mother Tongue After performing two ANOVAs, one on all ERP component amplitudes and one on all ERP component latencies, two significant effects were found. First, a main effect of group on latency (F1,23 ¼ 10.66, p < 0.01) occurred significantly

12

Identification of Grammatical Functions in Two Languages Hebrew Speakers

Cz

Russian Speakers

251

6 4

Subject Predicate Object

µV

Pz

2 0 −2 −4

−200 −100 0 100 200 300 400 500 600 700 ms

Fig. 12.1 ERP wave forms identified for subject, predicate, and object in Hebrew

later among Hebrew speakers than among Russian speakers. Second, a main effect of region on amplitude (F4.20,96.58 ¼ 38.29, p < 0.001) was found, with amplitudes significantly higher at anterior regions than at posterior ones. Additionally, a region group interaction effect on amplitude (F4.20,96.58 ¼ 7.45, p < 0.001) and latency (F3.07,70.57 ¼ 4.61, p < 0.01) was found. In the majority of cases both groups showed a gradual incline on amplitude from left-anterior to right-posterior regions and an inverse relation in latency magnitude. A grammatical role x group interaction effect on latency (F1.72,39.48 ¼ 5.10, p < 0.05) was also found. The results showed an inverse relation between the Hebrew and Russian groups in latency magnitude for grammatical role. Finally, a region grammatical role group interaction effect was found on amplitude (F4.18,96.18 ¼ 3.15, p < 0.05). In the majority of cases, Russian readers exhibited significantly higher amplitudes than Hebrew readers did. Following this initial analysis, separate ANOVAs were performed on the amplitudes and latencies of P100, P200, P300, N400, and P600 (a total of 12 ANOVAs). Grammatical role and scalp region served as within-subject factors and group as the between-subject factor. In order to control the overall probability of type I errors, only effects that had previously been found to

252

M. Leikin and E. Ritvas

Cz

Russian Speakers

6

Subject Predicate Object

4

Pz

2 0 −2 −4 −200

−100

0

100

200

300

400

500

600

700

ms

Fig. 12.2 ERP wave forms identified for subject, predicate, and object in Russian

be significant across ERP components were examined. In all ANOVAs Greenhouse-Geisser correction was applied for sphericity values lower than 0.75, and Huynh-Feldt correction for sphericity values greater than 0.75 (Field, 2005). In this case, the following significant effects were found: 1. Main effect of group on latencies N100 (F1,26 ¼ 8.80, p < 0.01), P100 (F1,26 ¼ 14.97, p < 0.01), P200 (F1,29 ¼ 6.44, p < 0.05), N400 (F1,27 ¼ 13.55, p < 0.01), and P300 (F1,29 ¼ 5.74, p < 0.05). In all cases latencies among native Hebrew speakers were longer than among Russian speakers. A similar effect was found in Leikin’s (2008) study, which investigated syntactic processing in Hebrew and English by native and bilingual adult speakers, with the Hebrew L1 group showing longer latencies than the English L1 group reading in their native languages. 2. Main effect of region on P100 (F2.08,54.19 ¼ 19.14, p < 0.001), P200 (F2.75,79.70 ¼ 58.99, p < 0.001), N400 (F2.92,78.91 ¼ 32.39, p < 0.001), and P300 (F3.00,86.98 ¼ 48.42, p < 0.001) amplitudes. In all cases higher amplitudes were found in posterior than in central regions and in central regions than in anterior ones. Lower amplitudes were found in the left hemisphere, particularly in central and posterior regions. 3. Region group interactions effects on N100 (F1.48,38.41 ¼ 3.97, p < 0.05), P200 (F2.75,79.70 ¼ 16.49, p < 0.001), N400 (F2.92,78.91 ¼ 3.13, p < 0.05), P300

12

Identification of Grammatical Functions in Two Languages

253

Fig. 12.3 Region group interaction effect on N400 amplitude

(F3.00,86.98 ¼ 6.46, p < 0.01) and P600 (F2.50,72.38 ¼ 6.70, p < 0.01) amplitudes. Both groups showed higher amplitudes in right-posterior than in central regions and in central regions than in left-anterior ones (Fig. 12.3). In addition, a marginally significant effect on N400 latency (F2.50,67.38 ¼ 2.53, p ¼ 0.075) was found. 4. Grammatical role group interaction effects on N400 (F2,54 ¼ 5.57, p < 0.01) and P300 (F2,58 ¼ 10.54, p < 0.001) latencies. The results showed an inverse relation between the Hebrew and Russian groups in latency magnitude for grammatical role. In the Hebrew L1 speaking group the longest latencies were elicited by the object and the lowest ones by the subject, whereas Russian speakers showed the opposite latency activation order: subject > object > predicate (see Leikin, 2008 for details). 5. Region grammatical role x group interaction effects on P600 amplitude (F4.66,135.20 ¼ 3.90, p < 0.01), and a marginally significant interaction on P200 amplitude (F3.19,92.47 ¼ 2.51, p ¼ 0.06). In the majority of cases, Russian readers exhibited significantly higher amplitudes than Hebrew readers did. In the majority of cases both groups showed a gradual incline on amplitude from left-anterior to right-posterior regions.

12.4.4 Comparison of Brain Activity Among Russian Speakers When Reading Sentences in Russian and Hebrew Two ANOVAs performed separately for amplitudes and latencies revealed two significant effects. First, a marginally significant effect of language on latency

254

M. Leikin and E. Ritvas

(F2,26 ¼ 2.99, p ¼ 0.068) was shown. The latencies were significantly longer while reading sentences in Hebrew than in Russian. Second, a main effect of region on amplitude (F2.54,33.02 ¼ 5.27, p < 0.01) was found. In the majority of cases, amplitudes in the posterior regions were greater than those in central regions, and these were greater than amplitudes in anterior regions. A marginally significant effect on latency was also found (F3.15,40.89 ¼ 2.33, p ¼ 0.086). Latencies in the posterior regions were longer than those in central regions, which were longer than those in anterior regions. Following this initial analysis, separate ANOVAs were performed on the amplitudes and latencies of P100, P200, P300, N400, and P600 (a total of 12 ANOVAs). Language, grammatical role, and scalp region served as withinsubject factors. The following significant effects were found: 1. A main effect of language on N400 latency (F1.53,19.94 ¼ 6.15, p < 0.05) was found. Latencies in the processing of Russian sentences were shorter than those in the processing of Hebrew sentences. 2. A main effect of region on P100 (F2.69,24.22 ¼ 6.37, p < 0.01), P200 (F1.62,19.45 ¼ 3.84, p < 0.05), N400 (F2.41,31.36 ¼ 4.02, p < 0.05), P300 (F2.57,33.38 ¼ 9.24, p < 0.001), and P600 (F1.60,20.73 ¼ 9.61, p < 0.01) amplitudes was found. P100, P200, N400, and P300 amplitudes in posterior regions were greater than those in central regions, and these were greater than those in anterior regions (Fig. 12.4), whereas P600 amplitudes were

Fig. 12.4 Main effect of region on P100 amplitude

12

Identification of Grammatical Functions in Two Languages

255

greater in anterior than in central regions and in central regions than in posterior ones. Additionally, P200 amplitudes in left central and posterior regions were smaller than those in medial regions. A main effect of region on P200 latency (F2.30,27.61 ¼ 5.00, p < 0.05) was also found. Latencies in posterior regions were greater than those in central regions, which were greater than those in anterior regions. Additionally, latency in the right posterior region was significantly greater than in the left posterior region.

12.4.5 Analysis of the Sources of Activity To describe the experimental conditions and timings in which there were differences in the sources of brain activity, comparisons of scalp maps were performed (TANOVA – Strik, Fallgatter, Brandeis, & Pascual-Marqui, 1998) using the LORETA-Key computer program. Separate comparisons were carried out between Russian and Hebrew speakers for subject, predicate, and object. Moreover, within-subject comparisons were carried out for reading sentences in Hebrew and Russian. Finally, comparisons of brain activity during the reading of words with different grammatical functions (e.g., subject vs. predicate) were performed. Altogether, 21 comparisons were performed. To control the general probability of type I error, significance of greater than 0.0024 was needed. Significant differences (or differences with borderline significance) were found in 8 comparisons with 18 timings. Maps of current density produced with the LORETA-Key computer program were compared with timings in which differences in scalp maps were found (Strik et al., 1998). An analysis revealed several significant differences. First, comparing subject and predicate in native Hebrew speakers reading Hebrew sentences, a stronger activation was found when reading a predicate than a subject in two timings (Fig. 12.5). There were differences in the medial frontal gyrus (168 ms after stimulus onset) and in the left inferior frontal gyrus (285 ms).

(a)

(b)

Fig. 12.5 Comparison of brain activity in native Hebrew speakers processing subjects and predicates in Hebrew. (a) 168 ms after stimulus onset. (b) 285 ms after stimulus onset

256

M. Leikin and E. Ritvas

Fig. 12.6 Comparison of brain activity in native Russian speakers processing subjects and predicates in Russian

Second, comparing subject and predicate in native Russian speakers reading Russian sentences, a stronger activation was found when reading a subject than a predicate in the left superior frontal gyrus, between 270 and 277 ms after stimulus onset (Fig. 12.6). On the one hand, the timing shown in two above-noted cases (between 168 and 285 ms) seems to resemble time intervals that are known as P200 and P300 ERP components. Although the P200 component is generally thought to reflect late sensory and early perceptual processes, it is also sensitive to different grammatical classes (Pulvermuller, Preissl, Lutzenberger, & Birbaumer, 1996) and lexical-morphological distinctions (Breznitz, Oren, & Shaul, 2004; Leikin, 2002). In turn, P300 appears to be associated with stimulus classification and updating in short-term memory (Donchin, 1981). However, it can also be elicited by morphosyntactic violations (Coulson, King, & Kutas, 1998) and processing of grammatical functions (Leikin, 2002). On the other hand, the regions found to be sources of differences are known to be associated with syntactic processing (e.g., Brauer & Friederici, 2007; Dapretto & Bookheimer, 1999; Friederici et al., 2000) and even with semantic processing (Friederici et al., 2000; Seghier et al., 2004; Vitacco, Brandeis, Pascual-Marqui, & Martin, 2002). Third, comparing subject in Russian speakers reading Russian and Hebrew sentences, a stronger activation was found reading a subject in Russian in the time window that started at 188 ms after stimulus onset and continued up to 434 ms. Within this time slice, there were several small windows in which significant differences were found: between 188 and 285 ms it was generally shown in the right frontal medial and pere-sylvian areas and in the left supramarginal gyrus, angular gyrus, and post-central medial areas. The same sources of differences were found between 313 and 434 ms, with the differences being more prominent in posterior areas (SMG, AG), in the medial postcentral areas, and in the right posterior middle temporal gyrus (Fig. 12.7). It should be noted

12

Identification of Grammatical Functions in Two Languages 188−285 ms

313−434 ms

188−285 ms

313−434 ms

257

Fig. 12.7 Comparison of brain activity in native Russian speakers processing subjects in Russian and in Hebrew

that not all phenomena found in our study may be easily explained. For example, it is right in regard to the lack of late time interval (something like P600) and to the contradictory existence of significant involvement of Right Hemisphere areas in the processing, though in the literature there are a few references concerning the contribution of right STG, temporal median gyrus, and posterior cingular gyrus to sentence processing (Lauro, Tettamanti, Cappa, & Papagno, 2008; Mercure, Ska, & Yves, 2006). Fourth, comparing predicates in native Russian speakers reading Russian and Hebrew sentences, a stronger activation was found when reading in Russian in the anterior cingulate (231–242 ms after stimulus onset) and in the cingulate gyrus (273–270 ms). Fifth, comparing subject and object in Russian speakers reading in Russian, a stronger activation was found when reading a subject than an object in the left supramarginal gyrus, in the angular gyrus, and in the anterior cingulate (285–297 ms after stimulus onset).

258

M. Leikin and E. Ritvas

12.5 Discussion The present study examined brain activity by means of ERP amplitudes and latencies, and areas of brain activation in adult monolingual and bilingual readers processing the grammatical functions of words read in sentences in two languages: Hebrew and Russian. Behavioral data indicated that Russian L1 readers were more accurate and faster on most reading related tasks in their native language than were Hebrew L1 speakers reading in Hebrew. This finding confirms that the Russian speakers were proficient L1 readers despite having immigrated to Israel at least 5 years earlier. The finding is also consistent with previous studies that found Russian to be more complicated phonologically than Hebrew (Akhmanova, 1971; Shimron, 1993). This phonological complexity has a positive influence on the development of phonological awareness at the phoneme level in Russianspeaking children (Zaretsky, 2002). Therefore, it may be suggested that enhanced phonological awareness helped the Russian speaking group read more fluently and more accurately than the Hebrew L1 group. But Russian L1 speakers performed more poorly on most of the reading and reading-related tasks in Hebrew than they did in their native language. These results are also consistent with previous literature on bilingualism (e.g., Fabbro, 2001) that supports the hypothesis that one language, usually the native one, is more effective and fluent than the other. The most important outcome of the study was the revelation of significant differences in ERP measures between Russian and Hebrew readers. All our hypotheses were supported by the research findings. The results showed that different languages elicit different brain responses in bilingual (Russian/ Hebrew) and monolingual (Hebrew) speakers. There were also differences in ERP measures and brain localization (see the results of the LORETA-Key analysis) as well as in cognitive strategies employed in the processes of identifying the grammatical functions of words in the first and second language. Furthermore, the results confirmed previous findings suggesting that grammatical processing in L2 is fundamentally different from grammatical processing in one’s L1 (Clahsen & Felser, 2006a, 2006b). Comparison between the two groups reading in their native languages already revealed significant differences. Hebrew speakers exhibited consistently higher amplitudes and longer latencies of P300, N400, and P600 in all Hebrew sentence elements than did the Russian speakers. As mentioned earlier, a similar effect was found in Leikin’s (2008) study, which compared syntactic processing in Hebrew and English by native and bilingual readers. It has been shown that the Hebrew L1 group demonstrated higher amplitudes and longer latencies than the English L1 group reading in their native languages (Leikin, 2008). Higher amplitudes are thought (Taylor & Keenan, 1990) to represent greater brain effort during information processing. In turn, longer latencies are believed to reflect slower processing speed (Breznitz, 2006). Thus, the results

12

Identification of Grammatical Functions in Two Languages

259

seem to demonstrate that in Hebrew processing of words according to their grammatical functions is slower even among Hebrew L1 speakers and requires greater effort. Thus our study produced surprising results. It seems reasonable to expect more effective performance requiring less effort in one’s native language than in a foreign one. But these unexpected results (slow processing in the Hebrew language) can be explained by the fact that Hebrew has a rich derivational morphology. In general, Hebrew words can be decomposed into two morphemes, the root and the word pattern, but these morphemes are not appended to one another linearly; rather the phonemes of each morpheme are intertwined. Roots and word patterns are abstract structures, and only their joint combination forms specific words. Although these morphemes carry some semantic and syntactic information, their meaning is often obscure and changes for each root-pattern combination (Berman, 1985). The slower cerebral processing manifested in ERP latency in Hebrew than in Russian was observed at the initial phase of stimulus processing (P200), that is, at the stimulus classification stage, and continued into the updating in working memory stage (P300). Moreover, the higher N400 amplitude and longer N400 latency among Hebrew L1 speakers in reading Hebrew than in reading English sentences (Leikin, 2008) suggest considerable lexical-semantic processing (e.g., Breznitz & Leikin, 2000; Leikin, 2008; Osterhout, 1997). Our findings are consistent with the results obtained by Deutsch and Bentin (2001), who examined the on-line processing of semantic and syntactic information during agreement analysis using converging evidence from eye movements and ERP. Deutsch and Bentin argued that, consistent with findings obtained in other inflected languages (such as Italian, Spanish, French, and Dutch), and unlike English, the process of computing syntactic agreement in highly inflected languages is affected by semantic or conceptual factors, or both. The researchers concluded that the linguistic properties of a language, rather than differences between production and comprehension, account for discrepancies between various studies on the interrelation between semantic and syntactic processing of agreement. Thus, morpho-syntactic and lexical-semantic information in Hebrew plays an important role in processing a sentence and understanding its meaning. It may be suggested, therefore, that these multiple lexical factors decelerate cerebral processing in Hebrew. The study found a main effect of region on P100, P200, P300, and N400 amplitudes. A comparison made between groups and within each group revealed higher amplitudes in posterior than in central and in central than in anterior regions. If we accept the hypothesis that amplitude shows evidence of cerebral effort, it may be suggested that processing words with different grammatical functions requires greater effort in posterior regions. Various researchers have suggested that semantic and morphological features are processed in temporal middle and inferior regions and in left inferior frontal areas (see Heim, 2005 for a review). But the results for the processing of morpho-syntactic information are not entirely conclusive. Recent clinical findings suggest involvement of the cerebellum in syntactic and other higher cortical processing

260

M. Leikin and E. Ritvas

(e.g., Dogil et al., 2002). Deutsch and Bentin (2001), studying the ability to distinctively measure the manifestation of semantic and syntactic processes in ERPs, found a pattern of interactions between markedness, animacy, and syntactic congruity effects that were relatively obscure in eye movements. Whereas in eye movements both animacy and markedness influenced the effect of syntactic incongruity on first-pass duration, each of these factors affected a different ERP component: the congruity effect interacted only with markedness for the syntactically associated P600 and only with animacy for N400. Consequently, the ERP data disentangled the semantic and syntactic processes that interacted during the initial reading of the word in context, as reflected by firstpass duration. The absence of P600 sensitivity to animacy corroborated the view that the linguistic process indexed by this component is not sensitive to semantic information. Note that the data we obtained from P600 amplitudes revealed that higher amplitudes were evident in anterior regions than in posterior ones. A biphasic ERP pattern with an early anterior negativity followed by a P600 has been reported in several sentence processing studies (see Friederici et al., 2002, for review). P600 has been interpreted to reflect controlled rather than automatized processing (Friederici et al., 2002; Hahne & Friederici, 2001), specifically effortful syntactic integration processes at the sentence level. The P600 we obtained suggests that L1 and L2 speakers also use later sentence-based processes. Therefore, the P600 may also be responsive to processes related to word recognition that are required for its immediate integration into the local syntactic structure, such as morphological analysis. We suggest that our participants employed early processes of word-internal semantically based automatized processing that appeared in posterior brain regions, and at a later stage employed at least partially controlled syntactic processing that took place in anterior regions. The second research hypothesis postulated differences in ERP measures and in cognitive strategies between processes of identification of the grammatical functions of words in L1 and L2. A comparison between Russian- and Hebrewspeaking groups reading in their native languages registered the highest activation by N400 amplitude and the longest N400 latency, revealing a different processing profile. Moreover, the present study found different activation patterns in L1 Hebrew readers than did previous research. In previous experiments different grammatical functions influenced sentence processing differently, as reflected in ERP amplitudes and latencies (Breznitz & Leikin, 2000; Leikin, 2002). This effect was explained to indicate processes of identification of the grammatical functions of words. The patterns of brain activation among the Hebrew L1 group obtained in the present study were neither predicate-oriented nor entirely word-order types. In the Hebrew group this related to significantly longer N400 and P300 latencies for the object (object > predicate > subject). In the Russian group the N400 and P300 latencies reflected another pattern (subject > object > predicate). Thus, the patterns of brain activation were partly different in Hebrew and Russian speakers. In turn, these data confirm

12

Identification of Grammatical Functions in Two Languages

261

the assumption (Breznitz & Leikin, 2000) that normal readers of Hebrew identify the grammatical functions of words at least partly by means of the words’ lexical-morphological properties, while they are simultaneously affected by word order as well. In the Russian speaking group the results showed morphologically based, noun-oriented strategies, as we had hypothesized. The modification of activation patterns was yet another interesting result of the study. We speculate that this result is due to the character of the stimulus sentences in the new experiment (see also Leikin, 2008). Previous results (Leikin & Breznitz, 1999) were obtained by means of stimulus sentences that included verbs functioning as predicates as well as various prepositions. In the present study, as in Leikin (2002, 2008), one-third of the sentences was verbless. Verbless sentences are not common in Russian and are not accepted in Hebrew. Moreover, the lexical-morphological characteristics of the words that played different grammatical roles were equalized to a certain degree. In addition, in the Hebrew sentences only the preposition ett appeared as the object marker. Thus, the obtained activation pattern may be explained by the reduction of lexical-morphological differences between the three sentence elements. This seems especially critical for verbless sentences. Absence of the verb can cause a relative decrease in brain activation in response to the predicate, and equalization of intensity and timing measures for the predicate and the object. This appears to confirm the significance of the verb for processing Hebrew sentences (Leikin & Breznitz, 1999). Furthermore, in the present study, unlike in Leikin (2002), the paradigm included one-third more of the experimental stimuli, and the stimuli (the sentences) were easier and clearer from a grammatical point of view. Region x grammatical role x group interactions revealed higher 600 amplitudes in the Russian L1 group than in the Hebrew L1 group. This may suggest that syntactic processing for the Russian L1 group required greater effort for the same task in their native language than it did for Hebrew L1 speakers. Less effortful syntactic processing may be an indication of more natural and more language-directed processing. We suggest therefore that changes in ERP amplitudes and latencies in Hebrew did not relate to ordinal word position and wordclass differences between target words but were associated with the grammatical functions of the words, whereas in Russian the grammatical functions of the words played a less prominent role. Finally, note that the between-group comparison of reading in Hebrew revealed considerable differences in processing measures and brain localization. Recent studies, which investigated morphological and syntactic processing in adult native and non-native speakers who acquired their L2 after childhood, showed that grammatical processing in L2 was fundamentally different from that in one’s native language (see Clahsen & Felser, 2006b for a review). A recent study found that, although participants showed bidirectional lexical transfer, they displayed only L1-to-L2 grammatical transfer (Hohenstein et al., 2006). To account for the observed L1/L2 differences in processing, Clahsen and Felser (2006b) proposed the shallow structure hypothesis (SSH), according to which the representations that adult L2 learners produce during

262

M. Leikin and E. Ritvas

processing contain less syntactic detail than those of native speakers. In turn, adult L2 learners were seen to rely more on lexical–semantic cues for interpretation during sentence processing. As the N400 is believed to index lexical– semantic processing, the fact that L2 (Russian/Hebrew) speakers in our study showed N400 and P600 responses is fully consistent with SSH. The differences found between native Hebrew speakers and native Russian speakers reading in Hebrew are also with SSH. Acknowledgement This work has been supported by the Israel Scientific Foundation (ISF), grant (9812).

References Akhmanova, O. S. (1971). Phonology, morphonology, morphology. The Hague: Mouton. Alvarez, R. P., Holcomb, P. J., & Grainger, J. (2003). Accessing word meaning in two languages: An event-related brain potential study of beginning bilinguals. Brain and Language, 87, 290–304. Balgur, R. (1968). Hebrew language: Word frequency [in Hebrew]. Tel Aviv: Otzar Hamoreh Publishers. Bates, E., Devescovi, A., & D’Amico, S. (1999). Processing complex sentences: A crosslinguistic study. Language and Cognitive Processes, 14, 69–123. Berman, R. (1985). The acquisition of Hebrew. In D. Slobin (Ed.), The crosslinguistic study of language acquisition (Vol. 1, pp. 255–371). Hillsdale, NJ: Erlbaum. Bialystok, E. (2007). Acquisition of literacy in bilingual children: A framework for research. Language Learning, 57(1), 45–77. Billingsley-Marshall, R. L., Sakari, S., Pataraia, E., & Papanicolau, A. C. (2006). Brain mechanisms supporting distinct languages. Learning Disabilities Research & Practice, 20(1), 31–38. Bookheimer, S. (2002). Functional MRI of language: New approaches to understanding the cortical organization of semantic processing. Annual Review of Neuroscience, 25, 151–188. Brauer, J., & Friederici, A. D. (2007). Functional neural networks of semantic and syntactic processes in the developing brain. Journal of Cognitive Neuroscience, 19(10), 1609–1623. Breznitz, Z. (2006). Fluency in reading: Synchronization of processes. Mahwah, NJ: Lawrence Erlbaum Associates Publishers. Breznitz, Z., & Leikin, M. (2000). Syntactic processing of Hebrew sentences in normal and Russian readers: Electrophysiological evidence. The Journal of Genetic Psychology, 161(3), 359–380. Breznitz, Z., & Leikin, M. (2001). Effects of accelerated reading rate on processing words’ syntactic functions by normal and Russian readers: Event-related potentials evidence. The Journal of Genetic Psychology, 162(3), 276–296. Breznitz, Z., Oren, R., & Shaul, S. (2004). Brain activity of regular and dyslexic readers while reading Hebrew as compared to English sentences. Reading and Writing: An Interdisciplinary Journal, 17, 707–737. Clahsen, H., & Felser, C. (2006a). Grammatical processing in language learners. Applied Psycholinguistics, 27, 3–42. Clahsen, H., & Felser, C. (2006b). Continuity and shallow structures in language processing. Applied Psycholinguistics, 27, 107–126. Coulson, S., King, J. W., & Kutas, M. (1998). Expect the unexpected: Event-related brain response to morphosyntactic violations. Language and Cognitive Processes, 13, 21–58. Dapretto, M., & Bookheimer, S. Y. (1999). Form and content: Dissociating syntax and semantics in sentence comprehension. Neuron, 24, 427–432.

12

Identification of Grammatical Functions in Two Languages

263

Deutsch, A., & Bentin, S. (2001). Syntactic and semantic factors in processing gender agreement in Hebrew: Evidence from ERPs and eye movements. Journal of Memory and Language, 45, 200–224. Deutsch, A., Frost, R., Pollatsek, A., & Rayner, K. (2005). Morphological preview benefit effects in reading: Evidence from Hebrew. Language and Cognitive Processes, 20(1–2), 341–371. Dijkstra, T. (2001). What we know about bilingual word recognition: A review of studies and models. Plenary Address at the International Symposium on Bilingualism, Bristol. Dogil, G., Ackermann, H., Grodd, W., Haider, H., Kamp, H., Mayer, J., et al. (2002). The speaking brain: A tutorial introduction to fMRI experiments in the production of speech, prosody and syntax. Journal of Neurolinguistics, 15, 59–90. Donald, M. W., Meuter, R., & Ardal, S. (1986). Event-related brain potentials in the first and second languages of bilinguals. Society of Neuroscience Abstracts, 12(1), 721. Donchin, E. (1981). Surprise! . . . Surprise? Psychophysiology, 18, 493–513. Ellis, N. C. (2002). Frequency effects in language processing: A review with implications for theories of implicit and explicit language acquisition. Studies in Second Language Acquisition, 24(2), 143–188. Fabbro, F. (2001). The bilingual brain: Cerebral representation of languages. Brain and Language, 79, 211–222. Ferstl, E. C., & Flores d’Arcais, G. (1999). The reading of words and sentences. In A. D. Friederici (Ed.), Language comprehension: A biological perspective (pp. 175–210). Berlin: Springer. Field, A. (2005). Discovering statistics using SPSS (2nd ed.). London: Sage. Friederici, A. D. (1999). The neurobiology of language processing. In A. D. Friederici (Ed.), Language comprehension: A biological perspective (pp. 265–304). Berlin: Springer. Friederici, A. D., Hahne, A., & Saddy, D. (2002). Distinct neurophysiological patterns reflecting aspects of syntactic complexity and syntactic repair. Journal of Psycholinguistic Research, 31(1), 45–63. Friederici, A. D., Meyer, M., & von Cramon, D. Y. (2000). Auditory language comprehension: An event-related fMRI study on the processing of syntactic and lexical information. Brain and Language, 74, 289–300. Frost, R., & Granier, J. (2000). Cross-linguistic perspectives on morphological processing: An introduction. Language and Cognitive Processes, 15(4–5), 321–328. Gerard, L. D., & Scarborough, D. L. (1989). Language-specific lexical access of homographs by bilinguals. Journal of Experimental Psychology: Learning, Memory, and Cognition, 15(2), 305–315. Hahne, A., & Friederici, A. D. (2001). Processing a second language: Late learners’ comprehension mechanisms as revealed by event-related brain potentials. Bilingualism: Language and Cognition, 4(2), 123–141. Hahne, A., Mueller, J., & Clahsen, H. (2006). Morphological processing in a second language: Behavioral and Event-related brain potential evidence for storage and decomposition. Journal of Cognitive Neuroscience, 18(1), 121–134. Heim, S. (2005). The structure and dynamics of normal language processing: Insights from neuroimaging. Acta Neurobiologiae Experimentalis, 65, 95–116. Hohenstein, J., Eisenberg, A., & Naigles, L. (2006). Is he floating across or crossing afloat? Cross-influence of L1 and L2 in Spanish-English bilingual adults. Bilingualism: Language and Cognition, 9(3), 249–261. Humpries, C., Kimberly, T., Buchsbaum, B., & Hickok, G. (2001). Role of anterior temporal cortex in auditory sentence comprehension: An fMRI study. Neuroreport, 12, 1749–1752. Illes, J., Francis, W. S., Desmond, J. E., Gabrieli, J. D. E., Glover, G. H., Poldrack, R., et al. (1999). Convergent cortical representation of semantic processing in bilinguals. Brain and Language, 70(3), 347–363.

264

M. Leikin and E. Ritvas

Jasper, H. H. (1958). The ten-twenty electrode system of the International Federation. Electroencephalography and Clinical Neurophysiology, 10, 371–375. Kako, E. T. (1999). The event semantics of syntactic structure (meanings of words, grammatical roles). Dissertation Abstracts International: Section B. The Sciences and Engineering, 59, 6085. Kempe, V., & MacWhinney, B. (1999). Processing of morphological and semantic cues in Russian and German. Language and Cognitive Processes, 14, 129–171. Kroll, J. F., & Tokowicz, N. (2001), The development of conceptual representation for words in a second language. In J. Nicol (Ed.), One mind, two languages: Bilingual language processing (pp. 49–71). Malden, MA: Blackwell. Lauro, R., Tettamanti, M., Cappa, S. F., & Papagno, C. (2008). Idiom comprehension: A prefrontal task? Cerebral Cortex, 18, 162–170. Leikin, M. (2002). Processing words’ syntactic functions in normal and Russian readers. Journal of Psycholinguistic Research, 31(2), 145–163. Leikin, M. (2008). Syntactic processing in two languages by native and bilingual adult readers: An electrophysiological study. Journal of Neurolinguistics, 21, 349–273. Leikin, M., & Breznitz, Z. (1999). Syntactic processing of Hebrew sentences: ERP measures. Genetic, Social, and General Psychology Monographs, 2, 173–191. Leikin, M., & Breznitz, Z. (2001). Effects of accelerated reading rate on syntactic processing of Hebrew sentences: Electrophysiological evidence. Genetic, Social, and General Psychology Monographs, 127(2), 193–209. Liu, Y., & Perfetti, C. A. (2003). The time course of brain activity in reading English and Chinese: An ERP study of Chinese bilinguals. Human Brain Mapping, 18(3), 167–175. Marian, V., Spivey, M., & Hirsch, J. (2003). Shared and separate systems in bilingual language processing: Converging evidence from eyetracking and brain imaging. Brain and Language, 86(1), 70–82. Mercure, E., Ska, B., & Yves, J. (2006). Metaphoric and nonmetaphoric alternative meanings of ambiguous words: Differential processing?(case study). Journal of Medical Speech – Language Pathology, 14(3), 157–159. Mitchel, D. C. (1987). Reading and syntactic analysis. In J. Beech & A. Colley (Eds.), Cognitive approaches to reading (pp. 87–112). Chichester: Wiley. National Institute for Testing and Evaluation. (1996). The Israeli standard achievement test for higher education. Jerusalem: Author. Offord, D. (1998). Using Russian: A guide to contemporary usage. Cambridge: Cambridge University Press. Osterhout, L. (1997). On the brain response to syntactic anomalies: Manipulations of word position and word class reveal individual differences. Brain and Language, 59, 494–522. Paradis, M. (2004). A neurolinguistic theory of bilingualism. Amsterdam/Philadelphia: John Benjamins Publishing Company. Paradis, M., & Goldblum, M. C. (1989). Selective crossed aphasia in a trilingual aphasic patient followed by reciprocal antagonism. Brain and Language, 36(1), 62–75. Perani, D., Paulesu, E., Galles, N. S., Dupoux, E., Dehaene, S., Bettinardi, V., et al. (1998). The bilingual brain: Proficiency and age of acquisition of the second language. Brain, 121(10), 1841–1852. Perfetti, C. A., Liu, Y., Fiez, J., Nelson, J., Bolger, D. J., & Tan, L. H. (2007). Reading in two writing systems: Accommodation and assimilation of the brain’s reading network. Bilingualism: Language and Cognition, 10(2), 131–146. Pulvermuller, F., Preissl, H., Lutzenberger, W., & Birbaumer, N. (1996). Brain rhythms of language: Nouns versus verbs. European Journal of Neuroscience, 8, 937–941. Raven, J., & Court, J. (1976). Raven’s progressive matrices. Oxford: J.C. Raven Ltd., Oxford Psychologists Press.

12

Identification of Grammatical Functions in Two Languages

265

Ru¨schemeyer, S. A., Fiebach, C. J., Kempe, V., & Friederici, A. (2005). Processing lexical semantic and syntactic information in first and second language: fMRI evidence from German and Russian. Human Brain Mapping, 25, 266–286. Sarfarazi, M., & Sedgwick, E. M. (1996). Event related potentials (ERP) as a function of language processing in bilinguals. Electroencephalography and Clinical Neurophysiology, 99(4), 347. Seghier, M. L., Lazeyras, F., Pegna, A. J., Annoni, J. M., Zimine, I., Mayer, E., et al. (2004). Variability of fMRI activation during a phonological and semantic language task in healthy subjects. Human Brain Mapping, 23(3), 140–155. Shatil, E. (1997). One minute tests. The Department of Learning Disabilities at the University of Haifa, Israel. Shimron, J. (1993). The role of vowels in reading: A review of studies of English and Hebrew. Psychological Bulletin, 114(1), 52–67. Shimron, J., & Sivan, T. (1994). Reading proficiency and orthography: Evidence from Hebrew and English. Language Learning, 44, 5–27. Sokolov, J. L. (1984). Development in the functionality of grammatical cues in Hebrew sentence comprehension. Papers and Reports on Child Language Development, 23, 115–124. Strik, W. K., Fallgatter, A. J., Brandeis, D., & Pascual-Marqui, R. D. (1998). Threedimensional tomography of event-related potentials during response inhibition: Evidence for phasic frontal lobe activation. Evoked Potentials-Electroencephalography and Clinical Neurophysiology, 108, 406–413. Suh, S., Yoon, H. W., Lee, S., Chung, J. Y., Cho, Z. H., & Park, H. W. (2007). Effects of syntactic complexity in L1 and L2; An fMRI study of Korean-English bilinguals. Brain Research, 1136, 178–189. Taylor, M. J., & Keenan, N. K. (1990). Event-related potentials to visual and language stimuli in normal and dyslexic children. Psychophysiology, 27, 318–327. Test for Russian as a First Language. (2006). Ministry of Education, Israel. Vitacco, D., Brandeis, D., Pascual-Marqui, R., & Martin, M. (2002). Correspondence of event-related potential tomography and functional magnetic resonance imaging during language processing. Human Brain Mapping, 17(1), 4–12. Wade, T. L. B. (1994). A comprehensive Russian grammar. Oxford: Blackwell. Weber-Fox, C., & Neville, N. J. (1996). Maturational constraints on functional specialization for language processing: ERP and behavioral evidence in bilingual speakers. Journal of Cognitive Neuroscience, 8, 231–256. Wechsler, D. (1994). Wechsler intelligence scale for adults (WAIS-3). Winsor: NFER. Zaretsky, E. (2002). Effects of oral language on sound segmentation skills: Crosslinguistic evidence. In F. Windsor & L. M. Kelly (Eds.), Investigations in clinical phonetics and linguistics (pp. 201–212). Mahwah, NJ: Lawrence Erlbaum Associates. Zasorina L. E. (Red.). (1977). Frequency dictionary of Russian language. Moskva: Russkii Jazyk (in Russian).