Mul Lex Rec in The Men Lex of Thi Lan Use

Secondebo
free oks ==> www.ebook777.com

Language Learning and Teaching
Weronika Szubko-Sitarek
Multilingual Lexical
Recognition in the
Mental Lexicon of
Third Language Users
WWW.EBOOK777.COM
free ebooks ==> www.ebook777.com
Second Language Learning and Teaching
Series editor
Mirosław Pawlak, Kalisz, Poland
WWW.EBOOK777.COM
About the Series
The series brings together volumes dealing with different aspects of learning and
teaching second and foreign languages. The titles included are both monographs and
edited collections focusing on a variety of topics ranging from the processes under-
lying second language acquisition, through various aspects of language learning in
instructed and non-instructed settings, to different facets of the teaching process,
including syllabus choice, materials design, classroom practices and evaluation.
The publications reflect state-of-the-art developments in those areas, they adopt a
wide range of theoretical perspectives and follow diverse research paradigms. The
intended audience are all those who are interested in naturalistic and classroom
second language acquisition, including researchers, methodologists, curriculum and
materials designers, teachers and undergraduate and graduate students undertaking
empirical investigations of how second languages are learnt and taught.
More information about this series at http://www.springer.com/series/10129
WWW.EBOOK777.COM
Multilingual Lexical
Recognition in the
Mental Lexicon of Third
Language Users
13
WWW.EBOOK777.COM
Department of English Language
and Applied Linguistics
Institute of English
University of Lodz
Lodz
Poland
ISSN 2193-7648 ISSN 2193-7656 (electronic)

ISBN 978-3-642-32193-1 ISBN 978-3-642-32194-8 (eBook)
DOI 10.1007/978-3-642-32194-8
Library of Congress Control Number: 2014946749
Springer Heidelberg New York Dordrecht London
© Springer-Verlag Berlin Heidelberg 2015

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or
information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed. Exempted from this legal reservation are brief excerpts
in connection with reviews or scholarly analysis or material supplied specifically for the purpose of
being entered and executed on a computer system, for exclusive use by the purchaser of the work.
Duplication of this publication or parts thereof is permitted only under the provisions of the Copyright
Law of the Publisher’s location, in its current version, and permission for use must always be obtained
from Springer. Permissions for use may be obtained through RightsLink at the Copyright Clearance
Center. Violations are liable to prosecution under the respective Copyright Law.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
While the advice and information in this book are believed to be true and accurate at the date of
publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for
any errors or omissions that may be made. The publisher makes no warranty, express or implied, with
respect to the material contained herein.
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)
WWW.EBOOK777.COM
Acknowledgments
Professor Jan Majer has been my mentor throughout my entire academic career.
His patience, understanding, knowledge and ability to provide constructive
criticism—at the same time enhancing motivation—are second to none and I am
sure this book would never have been published had it not been for his profes-
sional and friendly assistance. Thanks are also due to Dr. Anna Parr-Modrzejewska
for her help with the visual elements contained in this work, as well as for gen-
eral encouragement. I also wish to acknowledge the help provided by Professor
Mirosław Pawlak, in particular his kind attitude, extraordinary patience and gener-
osity in enabling me to publish this volume as part of the series; without his con-
stant encouragement my effort would no doubt have gone to waste.
WWW.EBOOK777.COM
Contents
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
1 Beyond Bilingualism: Issues in Multilingualism . . . . . . . . . . . . . . . . . . 1

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 The Multilingual World. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Multilingualism—Terminology and Definitions . . . . . . . . . . . . . . . . 5
1.3.1 Societal and Individual Multilingualism . . . . . . . . . . . . . . . . 5
1.3.2 Bilingualism Versus Multilingualism. . . . . . . . . . . . . . . . . . . 6
1.3.3 Multilingualism and the Levels and Breadth of Proficiency. . . 9
1.4 Towards a Holistic View of Multilingualism. . . . . . . . . . . . . . . . . . . 10
1.5 Characterizing Multilingual Learners . . . . . . . . . . . . . . . . . . . . . . . . 12
1.5.1 Contexts of Multilingual Acquisition
and Routes of Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.5.2 Cognitive Skills in Multilinguals. . . . . . . . . . . . . . . . . . . . . . 15
1.6 Research Themes in the Study of Multilingualism . . . . . . . . . . . . . . 18
1.6.1 Studies on Sociocultural Aspects of Multilingualism . . . . . . 20
1.6.2 Studies on Psycholinguistic Aspects of Multilingualism. . . . 22
1.6.3 Studies on Educational Aspects of Multilingualism . . . . . . . 24
1.6.4 Political Perspective of Multilingual Education in Europe. . . . 27
1.7 Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2 Modelling the Lexicon: Some General Considerations. . . . . . . . . . . . . 33

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.2 The Internal Composition of a Lexical Item . . . . . . . . . . . . . . . . . . . 34
2.3 Towards the Model of the Mental Lexicon . . . . . . . . . . . . . . . . . . . . 35
2.3.1 The Mental Lexicon Defined. . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3.2 The Internal Organization of the Lexicon . . . . . . . . . . . . . . . 40
2.3.3 The Internal Relations Within the Lexicon . . . . . . . . . . . . . . 41
2.3.4 Lexical Storage: The Full Listing Hypothesis
Versus the Decompositional Hypothesis . . . . . . . . . . . . . . . . 42
vii
WWW.EBOOK777.COM
viii Contents
2.4 Theories of Semantic Representation. . . . . . . . . . . . . . . . . . . . . . . . . 46

2.4.1 The Hierarchical Network Model. . . . . . . . . . . . . . . . . . . . . . 47
2.4.2 The Spreading Activation Model. . . . . . . . . . . . . . . . . . . . . . 48
2.4.3 The Componential Approach. . . . . . . . . . . . . . . . . . . . . . . . . 49
2.5 Models of Lexical Access in the Mental Lexicon . . . . . . . . . . . . . . . 52
2.5.1 The Serial Search Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
2.5.2 The Logogen Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2.5.3 The Cohort Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.5.4 Computational Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
2.6 Views on Language Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
2.6.1 The Modularity Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
2.6.2 Connectionism. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
2.7 Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3 Modelling Multilingual Representation and Processing. . . . . . . . . . . . 67

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.2 Mental Representation of Bilingualism. . . . . . . . . . . . . . . . . . . . . . . 68
3.2.1 The Separate Storage View. . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.2.2 The Shared Storage View. . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.2.3 The Mixed Storage View . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.3 Modelling Multilingual Representation. . . . . . . . . . . . . . . . . . . . . . . 80
3.4 Modelling Bilingual and Multilingual Processing. . . . . . . . . . . . . . . 84
3.4.1 Selective and Non-Selective Access. . . . . . . . . . . . . . . . . . . . 85
3.4.2 Task Dependent Access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
3.5 Models of Bilingual Visual Word Recognition . . . . . . . . . . . . . . . . . 88
3.5.1 The Bilingual Interactive Activation Model (BIA) . . . . . . . . 90
3.5.2 The Inhibitory Control Model . . . . . . . . . . . . . . . . . . . . . . . . 92
3.5.3 SOPHIA and the BIA+MODEL. . . . . . . . . . . . . . . . . . . . . . 93
3.6 Models of Multilingual Visual Word Recognition. . . . . . . . . . . . . . . 96
3.7 Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4 Visual Word Recognition in Multilinguals. . . . . . . . . . . . . . . . . . . . . . . 101

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.2 Word Stimuli Used in Visual Word Recognition Studies. . . . . . . . . . 102
4.2.1 Interlingual Homographs. . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.2.2 Interlingual Neighbours. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
4.2.3 Cognates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
4.3 Factors Affecting Visual Word Recognition in Bilinguals. . . . . . . . . 108
4.3.1 Word Superiority Effect. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
4.3.2 Word Length Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
4.3.3 Frequency Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
4.3.4 Neighbourhood Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
4.3.5 Cognate Facilitation Effect. . . . . . . . . . . . . . . . . . . . . . . . . . . 111
4.4 Factors Affecting Cognate Recognition. . . . . . . . . . . . . . . . . . . . . . . 115
4.4.1 The Effect of Cross-Linguistic Similarity on Cognate
Recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
WWW.EBOOK777.COM
Contents ix
4.4.2 The Cognate Facilitation Effect and the Word

Frequency Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
4.4.3 The Cognate Facilitation Effect and Task Demands . . . . . . . 119
4.5 Models of the Bilingual Lexicon and How They Account
for the Cognate Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
4.5.1 Shared Morphological Representation. . . . . . . . . . . . . . . . . . 120
4.5.2 Associatively Linked Orthographic Representations. . . . . . . 121
4.5.3 The Distributed Connectionist Model . . . . . . . . . . . . . . . . . . 122
4.5.4 The Localist Connectionist Model. . . . . . . . . . . . . . . . . . . . . 123
4.6 Bilingual Word Recognition in a Sentence Context. . . . . . . . . . . . . . 124
4.7 The Effect of the Cognate Status on Foreign Language
Vocabulary Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
4.8 Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5 Exploring the Cognate Facilitation Effect in Multilingual

Word Recognition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.2 Methods for Establishing Cognate Status in the
Multilingual Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
5.2.1 Similarity-Rating Study. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
5.2.2 Translation-Elicitation Task. . . . . . . . . . . . . . . . . . . . . . . . . . 134
5.2.3 The Comparison of Cognate Identification
Methodologies in the Multilingual Context. . . . . . . . . . . . . . 137
5.3 Investigating Cognate Effects in Trilingual
Visual Word Recognition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
5.4 Study 1: Exploring the Cognate Effect in Trilingual Visual
Word Recognition from the Weakest Language Perspective. . . . . . . 142
5.4.1 Research Hypotheses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
5.4.2 The Participants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
5.4.3 Procedure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
5.4.4 Materials. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
5.4.5 Results and Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
5.5 Study 2: The Role of Task Demands in the Trilingual
Processing of Cognates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
5.5.1 Research Hypotheses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
5.5.3 Procedure and Materials. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
5.6 Study 3: The Role of Relative Proficiency in the Third
Language in Trilingual Processing of Cognates. . . . . . . . . . . . . . . . . 155
5.6.1 Research Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
5.6.3 Materials and Procedure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
5.7 Study 4: The Role of Psychotypology in Trilingual
Processing of Cognates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
WWW.EBOOK777.COM
x Contents
5.7.1 Research Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

5.7.3 Materials and Procedure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
5.8 General Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
Conclusions and Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
Appendix A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Appendix B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Appendix C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Appendix D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Appendix E . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
WWW.EBOOK777.COM
Introduction
It is estimated that most language users in the world speak more than one
language, i.e. they are at least bilingual. In quantitative terms, then, monolingual-
ism may be regarded as the exception and multilingualism the norm (cf. Crystal
2003; Graddol 1997, 2006; Auer and Li Wei 2007; Aronin and Singleton 2012).
Nevertheless, it is only in the last two decades that issues of multilingualism started
to be examined closely and systematically. Indeed, it could no longer go unno-
ticed that a large part of the world’s population speaks several languages on a daily
basis and that a substantial proportion of language learners have some knowledge
of other languages beyond the second one. In fact, whether by living environment,
immigration or education, in many parts of the world the typical learner is no
longer a learner of a second language but rather of a third or additional language,
an L3 or an L4.
The wide spread of multilingualism, positively encouraged by people’s increas-
ing mobility across language borders, has led the ever-growing number of scholars
to investigate multilingual behaviour, as evidenced by the strong tradition of work
on sociolinguistic, educational and sociopolitical aspects of multilingualism (cf.
Cenoz and Genesee 1998; Cenoz et al. 2001b; Clyne et al. 2004; Edwards 1994,
2012; Kramsch 2006, 2014; Aronin and Singleton 2012). However, research on
the cognitive and psycholinguistic aspects of this young field of inquiry has been
much slower to appear. In fact, it was only recently that some researchers shifted
their interest to aspects relative to multilingual processing. Before describing the
structure of the present book, it is imperative to introduce some of the issues that
surround multilingualism. First, it will present some terminological inconsistencies
pertaining to the key notions in the field. Next, the existence of a mono- and bilin-
gual bias in multilingualism research will be discussed. An outline of the book
content concludes the present section, with a more precise introduction to the top-
ics discussed in each of the five chapters that follow.
As is usually the case with any other emerging field of inquiry, there is a
period of time in which the new area of research is perceived to be a mere exten-
sion of some related well-grounded fields. The most frequent tendency, then, is
to borrow the terminology already used in well-established fields such as SLA
xi
WWW.EBOOK777.COM
xii Introduction
(Second Language Acquisition) or bilingualism and adapt it to a new area, in the

case in question, multilingualism. These borrowing strategies are usually effec-
tive, but they can also be consequential if a term starts to be used with more than
one meaning, just as the terms “bilingual/bilingualism” and “multilingual/multi-
lingualism” are. When one reads definitions of bilingualism and multilingualism
in the literature, it becomes clear that the number of languages the individual is
familiar with is not central to the definition itself. The terms “bilingual/bilingual-
ism” and “multilingual/multilingualism” are often used as synonyms. For instance,
Myers-Scotton (2002) states that the term bilingual refers to “persons who
speak two or more languages” (Myers-Scotton 2002, p. 1, italics in the original),
whereas Grosjean (1992, p. 51) believes that “bilingualism is the regular use of
two (or more) languages, and bilinguals are those people who need and use two
(or more) languages in their everyday lives”. Indeed, as De Angelis (2007, p. 5,
italics in the original) observes, “in the literature on the topic the word bilingual
can refer to anything beyond the L1, when in actual fact the prefix ‘bi-’ means
‘two’ hence a bilingual can only be a speaker of two languages and not a speaker
of more than two languages by definition”. Analogously, the term multilingualism
tends to be used as a cover term for bi-, tri- and other lingualisms (cf. discussion in
Herdina and Jessner 2000; Cenoz et al. 2003).
Whatever the actual reasons for the use of the term multilingualism in place
of bilingualism, or of bilingualism in place of multilingualism, the use of these
terms as synonyms generates confusion in the field and makes it necessary to look
for additional information in the text itself in order to be able to identify whether
the author is talking about bilinguals or multilinguals. A similar confusion exists
around the terms used in reference to the language being learned. So far, research-
ers have referred to a second language as an L2, and to more than one non-native
language as L2s or Ln. A third or additional language has often been referred
to as an L3, regardless of whether it is a third, fourth or sixth language. On the
other hand, some scholars label languages according to the order of acquisition
(L3, L4 or L6) without taking into account issues of language proficiency. Last but
not least, a highly confusing term is the name of the field itself, which has been
labelled in at least four different ways: (a) Multiple Language Acquisition (MLA);
(b) Multilingual Acquisition; (c) Third Language Acquisition (TLA); and (d) Third
or Additional Language Acquisition. All of these terms are regularly found in the
literature but none of them has fully established itself to the present day, p robably
due to the weaknesses that each of them hold (cf. discussions in Herdina and
Jessner 2000; De Angelis 2007).
By and large, it seems that, in the literature on the topic, ambiguities and
uncertainties caused by terminological fuzziness can only be minimized by pro-
viding specific and detailed information on learners’ linguistic and educational
background. Hence in the practical part of the present volume a deliberate effort
will be made to specify the multilinguals using a set of parameters all of which
have already been shown to have some effect on multilinguals’ cognitive and
psycholinguistic processes (cf. Dörnyei 2005). The parameters used to specify
L3 learners invited to take part in the research reported in Chap. 5 included: the
WWW.EBOOK777.COM
Introduction xiii
number of languages known to the speaker; the sequence of acquisition of all

languages; the amount of formal instruction in each non-native language (years
and hours per week); the proficiency level in all non-native languages, and how the
proficiency level was measured, and the productive and receptive skills for each
language and how these were measured.
Another controversial issue within the field of multilingual studies concerns
some aspects of the monolingual and bilingual bias in third or additional language
research and relates to how a bilingual and by extension a multilingual person is
conceptualized. Notably, in the literature on the topic, bi- and multilinguals’ com-
petences are more prone to being viewed as separate and independent from one
another, and not as an intact whole. As Jessner points out, from a methodological
perspective, the fractional view of multilingualism is undoubtedly the most con-
venient, as it “allows one to easily separate source and target language information
in empirical research” (Jessner 2008, p. 30). Nevertheless, one cannot ignore the
fact that the mind of a multilingual contains information that belongs to several
languages and that the presence of linguistic information from various languages
is likely to lead to a state of integration of knowledge in the mind. Consequently,
both bi- and multilinguals must be looked at from a holistic perspective. As
Grosjean postulates, “a bilingual is NOT the sum of two complete or incomplete
monolinguals; rather, he or she has a unique and specific linguistic configura-
tion” (1992, p. 55, capitalized in the original). The contention that resonates in
a n umber of publications pertaining to different aspects of bilingual processing
(cf. De Angelis and Selinker 2001, Arabski 2002; Gabryś-Barker 2005, Waniek-
Klimczak 2007; Cenoz and Gorter 2013). By extension, it can be claimed that
a multilingual is neither the sum of three or more monolinguals, nor a bilingual
with an additional language. Rather, as De Angelis and Selinker (2001, p. 45)
propose, a multilingual is a speaker of three or more languages with unique lin-
guistic configurations, often depending on his/her individual history. It needs to
be remembered that existing theoretical perspectives increasingly view languages
as being interconnected with one another rather than being separate entities
(Jessner 2008; Cenoz and Gorter 2013; Komorowska 2013). The question today
is no longer whether linguistic information from different languages is integrated
or not, but to what extent it is integrated and how this integration may affect the
overall comprehension or production processes. Cook’s (1991, 1992, 1995) notion
of “multicompetence” (discussed in detail in Chap. 1) is perhaps one of the best
examples of a cognitivist theory which conceives knowledge as an integrated
whole in the mind. Moreover, it seems that even though the definition specifically
mentions the presence of two languages in the mind, the concept is sufficiently
neutral to be suitable to describe the integration of knowledge in the multilingual
mind as well (cf. Cook 2002, 2008).
One other form of bias which has come to light in the recent years is the
so-called bilingual bias which refers to the tendency to view multilinguals as
bilinguals with some additional languages rather than as speakers of several lan-
guages from the start. It also refers to the associated tendency of regarding bilin-
guals’ acquisition and production processes as default processes for multilinguals
WWW.EBOOK777.COM
xiv Introduction
(Jessner 2008a). This bias is so pervasive that it is virtually impossible to list

all the cases in which it can manifest itself. Still, the ever increasing number of
research studies on multilinguals point to the fact that L2 users’ behaviour can-
not adequately inform us about the phenomena related to multilingualism. This, in
turn, results in a significant rise of a general awareness among scholars that mul-
tilinguals are learners and speakers of their own right who should not be equated
with L2 learners without some careful vigilance. Today, most researchers would
agree that a general theory of non-native language acquisition cannot be based
on L2 learner behaviour alone. Instead, it must be capable of explaining how the
mind operates when two, as well as more than two, languages are involved, and
must be based on the knowledge and understanding of how the mind acquires,
stores, organizes and gains access to all the linguistic information that is available
to the learner, not just the information that belongs to the first or the second lan-
guage (cf. De Angelis 2007; Jessner 2008; Cenoz 2013).
It has to be borne in mind that due to the explicit nature of the phenomenon,
much work on multilingualism has been centered on the lexicon, this “central
module of a natural processing, its structure, and functioning” (de Bot 2004,
p. 17). Harely (2010, p. 145) claims that the first step in understanding both spo-
ken and written language processing is “doing something with words since words
are the building blocks of language”. In fact, understanding the processes under-
lying the visual recognition of isolated words remains a central endeavor in psy-
cholinguistics, cognitive psychology, and cognitive neuroscience. Additionally,
it is worth mentioning that over the past three decades, a prodigious amount of
work in visual word recognition has not only identified the many statistical prop-
erties associated with words (e.g., length, frequency of occurrence, concreteness,
cognate status, etc.) but also the effect of these properties on word recognition
performance (cf. Balota et al. 2006). A multitude of previous studies carried out
in different languages suggests that the distinction between cognate (words that
are similar in form and meaning) and noncognate (words only similar in mean-
ing) translations is consequential to the processing of this type of words (cf. the
cognate facilitation effect, Dijkstra 2005) and can be relevant in determining how
words are represented in the multilingual lexicon. Accordingly, one of the most
commonly researched lexical properties is cognate status.
Indeed, cognates have been very useful as tools to investigate the multilingual
mental lexicon and language (non)specificity of lexical access in both bilinguals
and multilinguals (cf. Friel and Kennison 2001). The general finding is that cog-
nates are produced, recognized and translated faster than noncognates (cf. Kroll
and Stewart 1994; Dijkstra et al. 1999; Costa et al. 2005). By way of example,
cognate more often than noncognate translation equivalents elicit associates
that are also translations of each other (cf. Taylor 1976), and that are translated
faster and more accurately from one language to the other (cf. De Groot 1992,
Sánchez-Casas et al. 1992). Particularly relevant for the present study is the find-
ing that in lexical decision tasks cognates tend to be responded to faster than non-
cognates (cf. Caramazza and Brones 1979; Cristoffanini et al.1986; De Groot and
Nas 1991; Dijkstra et al. 1998, 1999; Dufour and Kroll 1995; Kroll and Stewart
WWW.EBOOK777.COM
Introduction xv
1994; Lemhöfer et al. 2008; Sánchez-Casas et al. 1992; Schwartz and Kroll 2006;
Schwartz et al. 2007; Voga and Grainger 2007). The faster production, recognition
and translation of cognates are usually attributed to a common set of form-based
representations (e.g., orthographic, phonological, morphological) that are used to
process them in both (or more) languages. These processing differences between
cognates and other words in multilinguals are acknowledged to indicate differ-
ences in the representations depending on the word type (cognate/noncognate).
In the Polish context research on the multilingual mental lexicon and the role
cognates play in its organization and processing is still rather limited. Hence the
need of the present author to explore this dynamically growing field and to verify
the applicability and generalisability of the research findings to the Polish setting.
According to statistical data published by the Central Statistical Office of Poland
(Pl. GUS), English is the first and German the second foreign language in Polish
schools (cf. Dmochowska 2010). Accordingly, taking this language constellation
as the basis for empirical research guaranteed a wide number of reliable respond-
ents. The other reason for choosing to use this pair of foreign languages through-
out the experiments is related directly to the ready availability both of natural
cognates and noncognates between German and English.
The main aim of this book is to report on a set of experiments which employed
both offline as well as online methodologies to investigate the role of cog-
nates in the internal structure of the mental lexicon of Polish–English–German
trilinguals. Additionally, it is hoped that the effects uncovered by this kind of
empirical work can also be used in foreign language instruction, the contention to
be further addressed in the final section of the present work. The structure of this
volume is organized into five chapters, the first four of which contain the discus-
sion of the relevant theoretical concepts and the last one reports on the findings
of a series of experiments conducted among Polish–English–German learners at
the tertiary level. Chapter 1 provides a comprehensive and up-to-date review of
studies into multilingualism and third language learning and teaching, including
examples mainly from a European background. It describes sociolinguistic, psy-
cholinguistic and educational aspects of multilingualism and emphasizes current
research trends in a fairly young area of multilingual teaching. More importantly,
the chapter introduces some basic differences between Second and Third or
Additional Language Acquisition and overviews the “no-difference” assumption
that shapes so many of the SLA and TLA studies currently available. Chapter 2
introduces the discussion on the structure of the monolingual mental lexicon and
word recognition models. It tackles issues related to lexical storage and process-
ing in the monolingual context. First, it describes various propositions concern-
ing the internal structure of the mental lexicon of monolingual speakers. Next, it
provides a brief review of models of lexical access with a special focus on visual
perception. Thereby, a broader, diachronic perspective on the general role of the
mental lexicon in lexical processing is presented. The main concern of Chap. 3 is
the presentation and discussion of research to date on the lexical structure of the
bilingual and multilingual lexicon, as well as on the hypothesized changes in lexi-
cal organization over time and the role of language proficiency in bringing about
WWW.EBOOK777.COM
xvi Introduction
these changes. The chapter additionally examines the evidence of integration and
separation of knowledge in the mind, verifies the arguments for and against non-
selective access hypothesis and accounts for issues related to task dependence
of multilingual word recognition studies. Much of Chap. 4 offers the theoretical
background to the empirical research reported in the last chapter by reviewing evi-
dence for the special status of cognates. It needs to be noted here that the term
cognate is viewed from a psycholinguistic perspective in the light of which it is
not etymological relation but formal and semantic resemblance that is of impor-
tance. Finally, Chap. 5 reports on the results of two off-line studies (similarity-
rating and translation-elicitation tasks) analysing the perception of cognates from
the form similarity perspective. The studies were conducted to identify German–
English cognates and noncognates as viewed by Polish learners of English (L2)
and German (L3). The outcome of these experiments was a list of German–
English cognates as identified by the speakers of Polish used in a series of four
experiments on a series of visual perceptual experimental studies conducted within
the lexical decision task paradigm whose aim was to test the assumptions concern-
ing the special position of cognates (the cognate facilitation effect) within a multi-
lingual mind and to answer the question whether trilinguals rely upon their second
language lexical knowledge when recognizing L3 words. The results of the experi-
ments attest to simultaneous activation and parallel processing as well as interac-
tion among all the three languages in a trilingual mind. At the same time, they
point to the fact that cross-linguistic lexical access and the source and strength of
transfer may be constrained by variables such as L1 dominance, language profi-
ciency linguistic typology as well as task demands. A detailed description of all
four experiments is offered in the main body of the chapter, whereas its final part
discusses the results, provides a brief synthesis of the general conclusions and
offers some recommendations for further research.
Notably, discussions in each chapter reflect the fundamental belief that research
on multilingual behaviour can offer some valuable insights into the processes
connected with non-native language acquisition. Therefore, the analysis and dis-
cussion of the results accrued in the experiments will be followed by some sugges-
tions which when applied in the multilingual educational context could contribute
to the increased efficiency of the process of third-language education.
WWW.EBOOK777.COM
Chapter 1
Beyond Bilingualism:
Issues in Multilingualism
1.1 Introduction
This introductory chapter is meant to serve as a background for a more thorough

consideration of psycholinguistic issues related to multilingualism which will be
discussed in the subsequent chapters of the present work. In the following sub-
sections an attempt will be made to explore various aspects of the multifaceted
concept of multilingualism. The underlying objective of the chapter is to show that
multilingualism is a separate phenomenon from bilingualism since, as Herdina and
Jessner argue, it not only produces a quantitative shift but, above all, it leads to
a substantial change of quality in the speaker’s language system(s) (cf. Herdina
and Jessner 2000, 2002; Jessner 2013a). Multilinguals are learners who cannot be
simply equated with L2 learners, they are not extended bilinguals, but learners in
their own right, which remains not without influence on the process of L3 lan-
guage learning.
Following these introductory remarks included in the present section, the
next section discusses the contemporary multilingual situation in the world with
a special focus on diachronic perspective on the changing attitudes towards mul-
tilingualism. Some closer attention is given to the multilingual situation in Europe
and to the role of English in promoting multilingualism. Further, a discussion of
numerous interpretations of terminological inconsistencies related to multilin-
gualism offered in Sect. 1.3, will be supplemented with the consideration of a
number of definitional issues. Section 1.4 provides a comprehensive treatment of
the holistic view of multilingualism. Section 1.5 reviews selected classifications
of multilinguals based on the numbers of languages known by multilingual indi-
viduals, contexts of acquisition and routes of learning. The section subsequently
refers to cognitive abilities of multilingual users. Finally, Sect. 1.6 reviews recent
studies within the field of multilingualism as well as its quantitatively specified
version—trilingualism.

W. Szubko-Sitarek, Multilingual Lexical Recognition in the Mental Lexicon 1
of Third Language Users, Second Language Learning and Teaching,
DOI 10.1007/978-3-642-32194-8_1
WWW.EBOOK777.COM
2 1 Beyond Bilingualism: Issues in Multilingualism
1.2 The Multilingual World
The majority of people in the world speak several languages on a daily basis. It
can no longer go unnoticed that communicating in more than two languages has
become a dire necessity for the world today. Consequently, many language learn-
ers go on to study languages beyond the second one. As can be imagined, one
could name numerous reasons for the increasing interest in the phenomenon of
multilingualism, the most convincing of which postulates that humans are poten-
tially polyglot by nature since all of us “possess the capacity to learn several lan-
guages” (Mackey 1968, p. 555). Undoubtedly, there exists a multitude of more
pragmatic reasons accountable for that status quo, starting with the multilingual
situation of the majority of African countries and the rapid growth of international
cooperation and increasing contacts between the countries of the world; to, more
locally, the developing European integration. Another important factor is the mul-
tiplicity of a number of linguistic communities, especially in countries such as
Canada, Switzerland or Spain, where a vast number of research studies, especially
in the form of immersion programmes, have been conducted. Finally, the develop-
ment of English as world’s lingua franca (ELF) is yet another important variable
adding to the growth of studies into multilingualism. Interestingly, in contrast to
a widespread assumption that English in its role of a lingua franca constitutes a
serious threat to multilingualism, an ever-growing bulk of literature in this field
seems to suggest that ELF can be seen as one of the leading factors in the creation
of multilingualism today (cf. Hoffmann 2000; House 2003; Jessner 2006, 2013b;
Jenkins 2007; Seidlhofer 2004, 2011; Cogo 2008, 2012; Cenoz and Jessner 2009;
Archibald et al. 2011).
Nowadays it is universally recognized that to be multilingual is no longer an
aberration it was believed to be for years. Putting it bluntly, it is a dire necessity
for the majority of the world today (cf. Edwards 1994, p. 1). In fact, it has been
hypothesized that individual multilingualism is at least as frequent in the population
of the world as pure monolingualism, perhaps even more frequent (cf. Cook 1991;
Tucker 1998; Tokuhama-Espinosa 2003; Crystal 2003; Auer and Wei 2007; Aronin
and Singleton 2012). As Cook reports,
On some calculations there are more people in the world like the Cameroonian
(who speaks 4-5 languages) than like the Englishman; there are 3000-5000 languages in
the world but only about 150 countries to fit them all into (Cook 1991, p. 113).
Indeed, bearing in mind the statistical data concerning the distribution of lan-
guages within the countries of the world, gathered in Ethnologue, the most
comprehensive reference volume that catalogs all the known living languages
in the world today (cf. Lewis et al. 2014), the speculation that two-thirds of the
world’s children grow up in a multilingual environment seems quite plausi-
ble (Crystal 1997, 2003). And although this contention may be impossible to
document with full precision, many researchers refer to it as a powerful argu-
ment supporting the significance of multilingual research (cf. Cook 1992,
p. 578; Cenoz 2001a; Cenoz et al. 2003; Franceschini 2009). More importantly,
WWW.EBOOK777.COM
1.2 The Multilingual World 3
assumptions of that kind have resulted in a growing number of opinions

(cf. Cook 1993; Romaine 1995; Herdina and Jessner 2002; Auer and Li Wei 2007;
Kemp 2009) according to which linguistic and psycholinguistic research should no
longer be modelled on the monolingual speaker but should take multilingualism
as a point of departure, measuring other lingualisms against it. As Jessner com-
ments, “whereas a number of scholars in language acquisition research still base
their work on the monolingual native speaker norm, others have developed more
realistic viewpoints” (2008a, p. 15).
Undoubtedly, multilingualism is not a recent phenomenon. Going back in time
one could trace a number of examples of individual, societal or functional mul-
tilingualism in Europe in the centuries before 1800. In their book on the aspects
of multilingualism in European history Braunmüller and Gisella Ferraresi (2003)
show that the use of languages other than that of the main population was by no
means anything special for the European middle and upper classes of that time.
One reason for the existence of multilingualism was that “it was simply a neces-
sary precondition for mastering the various tasks in everyday life” (Braunmüller
and Ferraresi 2003, p. 3). It was only the political history of the 19th and the 20th
centuries and the ideology of “one state—one nation—one language” that gave
rise to the idea that “monolingualism has always been the default or normal case
in Europe” (Braunmüller and Ferraresi 2003, p. 2). As Auer and Li Wei point out:
it is a reasonable assumption that the marginal role research on multilingualism has
played within linguistics until some decades ago is a result of the monolingual bias of
(particularly) European thinking about language which came into being during a phase of
European history in which the nation states defined themselves not in the least by the one
(standard) language which was chosen to be the symbolic expression of their unity (Auer
and Li Wei 2007, p. 1).
And thus, for many decades it has been overlooked that in whatever form or con-
ditions the vast majority of the European population is multilingual. According to
Aronin’s (2005) diachronic approach to language use and behaviour, the attitude
towards individual multilingualism has since the 19th century gone through three
historical periods (or stages). They can be labeled monolingual (approximately
up to the late 1950s), bilingual (from the 1960s up to the late 1980s) and multi-
lingual (from the 1990s up to the present). The proposed diachronic division
implies that “in these periods particular language(s) arrangements were socially
significant and represented a dominant trend” (Aronin 2005, p. 10). Each of these
periods is also characterized by the prevailing societal attitude and interest
toward the phenomena and social response, like the world-wide spread of immer-
sion programs, or more recently, the rise of tertiary education1 and plurilingual
didactics (cf. Sect. 1.6.3).
1 The term “tertiary education” (also “tertiary languages”) has been adopted by some linguists
to refer to foreign languages learned after the first foreign language, i.e. as one’s second, third,
fourth, etc. foreign language (cf. Hufeisen 1991, 2004).
WWW.EBOOK777.COM
Initially, the great majority of researchers investigating psychological, cognitive

or linguistic development of children concentrated exclusively on monolingual
subjects, leaving the development of bilingual and multilingual children out of the
scope of their research. One could give a multitude of reasons accountable for that
status quo, one of them being the deep-seated belief in the ever-present prevalence
of monolingualism, another the absurd claim that bilingualism was destructive to
the overall development of children. Almost a hundred years ago Jespersen (1922)
considered bilingualism as a burden causing detrimental effects and responsible
for delayed development and reduced intelligence. The general assumption that
monolingual development should be perceived as a norm prevailed for many
years; however, for the last few decades, bilingual, and more recently, multilingual
subjects and communities have received ever-increasing attention among linguists,
applied linguists and psycholinguists all over the world.
The early interest in the phenomenon of multilingualism, or in fact bilingual-
ism, the former being a much more recent realm of studies, was triggered, or else
imposed, by a socio-political necessity to educate thousands of immigrants in the
United States. At first, it was believed that bilingual immigrant children needed
to be cured of their disorder, bilingualism, per se, being their disease. Thus, the
main target of a number of teachers and curriculum developers was to eliminate
the immigrant children’s native language. A few significant exceptions to be noted
are usually connected with linguists staying in close touch with bi- or multilingual
children (cf. Leopold 1939–1949) and with the multilingual children’s linguist
parents who having witnessed the real cognitive benefits coming from mul-
tilinguality did not approve of all the efforts aimed at eliminating the children’s
native languages. Changes started as late as in the 1970s. Firstly, due to the lack
of convincing evidence to support the irrational belief that bilingualism inhib-
ited the overall development of children. Secondly, due to increasing evidence
to the contrary coming from the ever extending group of bi- and multilinguals
(Cummins 1976; Peal and Lambert 1962).
Recent research on bi- and multilinguals proves that being exposed to
more than one language as a child does not have a negative effect on language
acquisition (cf. Genesee 2002; Tokuhama-Espinosa 2003; De Angelis 2007;
Paradis 2007; Randall 2007; Bialystok 2009; Kemp 2009; Luo et al. 2010;
Grosjean and Li 2013), nor is there any evidence that implies that being exposed
to two or more languages causes delays or disorders in language development
(cf. de Houwer 1999; Harley 2010). Furthermore, some linguists go so far as to
believe that children are born ready to be bilingual (cf. Baker 2006), while others
(cf. Albert and Obler 1978; Tokuhama-Espinosa 2003, 2008; Safont Jordà 2005)
postulate that bilinguals mature linguistically, and perhaps also cognitively (cf.
Kecskes and Albertazzi 2007), earlier than monolinguals, which gives them a defi-
nite advantage over their monolingual colleagues. Additionally, the learning of
more than one language is believed to develop linguistic awareness (Jessner 2006,
2013b), support transfer of strategies from L2 to L1 (Garfinkel and Tabor 1991)
and contribute to the development of multilingual competence (Gabryś-Barker
2005; Jessner 2006).
WWW.EBOOK777.COM
1.3 Multilingualism—Terminology and Definitions 5
1.3 Multilingualism—Terminology and Definitions
In line with the foregoing overview, contemporary researchers have, beyond

doubt, far more thorough knowledge than the scholars of the 1930s or even 1970s
about the cognitive and psychological processes occurring in the multilingual
brain. They are still, however, unable to provide conclusive answers to many mul-
tilingual issues such as explaining multilingual lexical storage, processing and
retrieval by children and adults or accounting for complex multilingual devel-
opment, nor are they able to agree on the precise definition of this multifaceted
phenomenon. Consequently, definitions of multilingualism are still “many and
wide-ranging” (Aronin and Singleton 2012, p. 1).
In fact, multilingualism being a relatively new field of research has until recently
been still perceived as a mere extension of bilingualism and second language acqui-
sition (SLA). In fact, many linguists still use the term bilingualism with regard to
all types of lingualisms. In the same way, the notion of second language acquisi-
tion has often been understood in its wider sense where second may refer to any
language that the learner has added after infancy (cf. Weinreich 1953; Baetens
Beardsmore 1986). Naturally enough, having a wealth of bilingual research at their
disposal, the substantial majority of linguists working with subjects fluent in three or
more languages have not only adopted the terminology generated and used within
bilingual studies, with regard to both the collection of data and its analysis, but have
also, to a large degree, based their own research on the theoretical framework of
bilingualism. It is imperative to realize, however, that although the terms bilingual-
ism and multilingualism tend to be used interchangeably, both phenomena are far
from being identical. Moreover, the fact that multilinguals are by no means easy sub-
jects to investigate does not help in developing one concise definition of the field,
either. Firstly, there is the problem of a variety of possible language sequences and
the multiplicity of sociolinguistic circumstances in which potential subjects acquire
their subsequent languages. Secondly, there is the age at which they start learn-
ing their languages. Finally, there are a number of other social, individual and atti-
tudinal factors, all creating such a variety of groups and subgroups that hardly any
study may pride itself on unbiased results. And thus, the present section endeav-
ors to define and classify some of the fundamental concepts referred to throughout
the book. By no means, though, does it purport to be an exhaustive overview of
the notions that form the conceptual foundation of this extremely complex field of
research. Instead, it makes an attempt to present multilingualism from different per-
spectives and by doing so to clarify multiple terminological inconsistencies.
1.3.1 Societal and Individual Multilingualism
To date, many scholars (especially in Francophone tradition) have endeavoured to

differentiate between multilingualism as a societal and individual phenomenon (cf.
Li Wei 2007). The first attempts derive from the works of Hamers and Blanc (1989),
WWW.EBOOK777.COM
who in their comprehensive account on bilingualism proposed two separate

definitions differentiating between bilinguality and bilingualism. The former refers
to the psychological state of an individual who has access to more than one lin-
guistic code as a means of social communication, whereas the latter comprises
bilinguality but also refers to the state of a linguistic community in which the bilin-
gual functions and where other individuals are bilingual. There does not seem to
be much of a risk to assume that the above-presented division can be extended
in relation to multilingual studies. That line of argument appears to be shared by
many researchers (cf. Herdina and Jessner 2002; Aronin and Ó Laoire 2004;
Hufeisen 2004; Aronin and Hufeisen 2009). The ardent proponents of the multilin-
gualism-multilinguality dichotomy are Larissa Aronin and Muris Ó Laoire. In their
influential article on cultural contexts of multilingualism the authors in question
specify and expand the notion of multilinguality using their ecological model as the
background (Aronin and Ó Laoire 2004). They posit that multilinguality is a per-
sonal characteristic and involves an individual’s store of all their interlanguages, as
well as metalinguistic awareness, language learning strategies and even emotional
factors such as opinions, preferences and knowledge about languages, their use and
the learning of them (ibid.). Thus, multilinguality, being “connected to personality
and intrapersonal aspects” (Aronin and Ó Laoire 2004, p. 18), can be defined as an
individual’s linguistic identity. By and large, while multilingualism, like bilingual-
ism, refers to a social situation when more than one language or several variations
of one language are used in a particular geographical area; multilinguality refers
more to “inner constructs of a single speaker” (Cenoz 2000, p. 39).
Importantly, in the literature there are two alternative terms used in reference
to multilinguality: individual multilingualism and plurilingualism. The first notion
(individual multilingualism) was coined by Cenoz and Genesee (1998, p. 17) and
focuses predominantly on language systems and language codes. The second term
(plurilingualism), typical of the EU context, was introduced by the Council of
Europe as a literal translation from French plurilinguisme denoting an individual’s
ability to use several languages in contrast to the multilingual nature of a given
society referred to as multilinguisme. Finally, it needs to be noted that there is still
a lot of confusion concerning the notions of multilingualism and plurilingualism
mainly owing to their inconsistent uses in documents and research papers. In the
present volume both multilingualism and plurilingualism will be used with refer-
ence to individuals rather than to societies. The terms bilingualism/bilingual and
trilingualism/trilingual, etc. will also be applied in the above meaning where the
reference is made to situations involving a precise number of languages.
1.3.2 Bilingualism Versus Multilingualism
An important issue of relevance in the context of clarifying terminology used with

reference to multilingualism is the notion of bilingualism. By definition a bilin-
gual person is familiar with two languages, whereas a multilingual one knows
WWW.EBOOK777.COM
three or more. In practice, definitions used in the literature on the topic do not
stick to the “number of languages” rule. The terms bilingualism/multilingualism
and bilinguals/multilinguals are often used as synonyms. Obviously, many pro-
posed definitions of bilingualism restrict themselves explicitly to two languages.
Nevertheless, the implicit suggestion that several lingualisms can be subsumed
under the concept of bilingualism was already illustrated in early publications in
the history of multilingual studies, e.g., in the works of Skuttnab-Kangas (1984)
who allowed for more than two languages to be present in the person she defined
as bilingual. According to Oksaar, bilingualism was “the ability of a person to use
here and now two or more languages as a means of communication in most situa-
tions and to switch from one language to the other if necessary” (Oksaar 1983, p. 19).
In a similar vein, Grosjean saw bilingualism as “the regular use of two (or more)
languages, and bilinguals (as) those people who need and use two (or more) lan-
guages in their everyday lives” (Grosjean 1992, p. 51), Myers-Scotton (2002, p. 1)
proposed that the term “bilingual refers to persons who speak two or more lan-
guages”, whereas Cenoz and Genesee (1998) defined the term as the final outcome
of the process of acquisition of several non-native languages. It can be concluded
that as for the number of languages involved, the term bilingualism “(…) has so
many contradictory definitions and associations in popular and academic usage
that it seems best to avoid it whenever possible” (Cook 2002, p. 4). For that rea-
son, Jessner (1996, 2008a, b) and Hoffmann (2000) look at bilingualism from an
entirely new, one could say “reversed”, perspective and refer to the phenomenon
as a quantitatively specified subtype of multilingualism “albeit a common one”
(Herdina and Jessner 2000, p. 84). Following this line of reasoning, additional
terms that started to appear recently in the literature on multilingualism have been
trilingualism and quadrilingualism. Naturally enough, they are used in reference
to a precise number of languages involved.
Equally disputable as the definitions of bilingualism and multilingualism are
the notions of a second (L2) and a third language (L3). Firstly, it should be noted
that in the technical sense L2, L3 or L4 are not necessarily equal to language num-
ber two, three or four in order of acquisition; instead, the numbers point to the
level of proficiency following the equation: the higher the number, the less profi-
cient the speaker (cf. Hufeisen 1991, 2004). Secondly, analysing the literature on
multilingualism it becomes obvious that many scholars still approve of the bilin-
gual bias and do not differentiate between L2, L3 and Ln learners claiming that
the processes behind non-native language acquisition are essentially the same (cf.
De Angelis 2007). Thus languages that are acquired after the first language (or
first languages, in the case of simultaneous bilingualism) are commonly termed
second languages, L2s or Ln, whereas all non-native language learners are often
labelled L2 learners. This extension of meaning proves that many scholars even
perceive the distinction between L2 and L3 or Ln acquisition as unnecessary. By
way of example, Haugen referred to multilingualism as “a kind of multiple bilin-
gualism” (1956, p. 9) suggesting no qualitative differences between the two phe-
nomena. Similarly, Sharwood Smith opined that “[s]econd language acquisition
(SLA) normally stands as a cover term to refer to any language other than the first
WWW.EBOOK777.COM
language learned by a given learner or group of learners” (Sharwood Smith 1994,

p. 7), whereas Baetens Beardsmore (1986) argued that,
[t]here is no evidence to suggest that the fundamental principles affecting language usage
are any different whether two, three or more languages are being used by one and the
same speaker, and the major question is whether they differ significantly from the cases
where only one language is being used (Baetens Beardsmore 1986, p. 3).
More recently, Li Wei (2007, p. 7) pointed out that although the word bilingual is
primarily used in reference to someone using two languages, it can also be taken
to include the many people in the world who have varying degrees of proficiency
in and interchangeably use three, four or even more languages. Also Mitchell and
Myles (2004, p. 6) promoted the view that foreign languages should be included
under one more general term second languages claiming that the underlying pro-
cesses are essentially the same for all non-native languages.
Of late, however, a rationale underlying the interchangeability of multilingualism
and bilingualism as supplied by the foregoing researchers has been called into ques-
tion. Undoubtedly, these two phenomena share many similarities both at the psycholin-
guistic as well as at the sociolinguistic level; still, there seem to be many good reasons
to recognize that multilingualism has “characteristics of its own” (Hoffmann 2001a,
p. 3). Proponents of distinguishing studies on L3/Ln acquisition from the research on
bilingualism (cf. Cenoz and Jessner 2009; Aronin and Singleton 2012) note that factors
affecting third or additional language acquisition are much more numerous and much
more complex than those involved in the process of L2 learning. The variables include
among others: age of acquisition, sequence of acquisition of all non-native languages,
proficiency level in all languages or context and manner of acquisition. Most impor-
tantly, L2 is taught to monolinguals, whereas L3 to bilinguals. Claiming that processes
involved in learning L2 and L3 are the same would by extension imply that monolin-
guals do not differ from bilinguals (cf. Sect. 1.4).
According to Herdina and Jessner, “learning a third language differs essentially
from learning a second—something third language learners themselves intuitively
perceive” (Herdina and Jessner 2002, p. 96). As a matter of fact, in the light of
their Dynamic Model of Multilingualism (2002) learning any subsequent language
results in a qualitative change in the speaker’s language system as the acquisition
of a new language leads to the development of new skills (i.e. language learning
skills, language management skills and language maintenance skills; cf. Herdina
and Jessner 2002, pp. 92–93). Consequently, the process of L3 acquisition should
be depicted on the multilingual continuum, ranging from monolingual acquisition
(the acquisition of a foreign language based on the command of one language)
through balanced bilingualism to the command of three (or more) languages. Such
a perception on L3 learning supports the view that bilinguals should not be equated
with trilinguals as they are not extended bilinguals, just like a bilingual is not the
extension of a monolingual. As Cenoz and Genesee (1998) point out,
[m]ultilingual acquisition and multilingualism are complex phenomena. They implicate
all the factors and processes associated with second language acquisition and bilingual-
ism as well as unique and potentially very complex factors and effects associated with the
interactions that are possible among the multiple languages being learnt and the processes
of learning them (Cenoz and Genesee 1998, p. 16).
WWW.EBOOK777.COM
L3 prototypically represents the concept of the acquisition of any foreign language

after the second language, regardless of whether it is a third, fourth or seventh lan-
guage. Assuming that bilingualism and multilingualism are two distinct phenomena,
another question arises concerning the position of trilingualism within the studies
on multilingualism. Is trilingualism only a quantitatively specified type of multi-
lingualism or does it differ qualitatively from other lingualisms (e.g. quadrilingual-
ism)? The answer to this question is strictly connected with routes of learning. By
way of example, in the formal instructed context where languages are taught con-
secutively, trilingualism cannot be perceived as just one type of multilingualism. It
definitely shares a lot with other multilingualisms just as it resembles bilingualism
in some respects. However, the fact that it is in the process of learning a third lan-
guage that learners can use their second language learning experience for the first
time makes trilingualism a distinct phenomenon both from bilingualism as well as
from other multilingualisms. For that reason proponents of distinguishing trilingual-
ism as a separate field of study maintain that it is a phenomenon in its own right. On
the one hand, it cannot be treated as a mere extension of bilingualism, on the other, it
also differs from learning a fourth or fifth language (cf. Hoffmann and Ytsma 2004).
Indeed, in recent years, there has been a growing number of linguists postulating
that due to the presence of a number of substantial qualitative discrepancies the
terms bilingualism, trilingualism and multilingualism should be used with due cau-
tion as the phenomena they refer to can by no means be equated (cf. Ringbom 1987;
Thomas 1988; Cenoz 2000, 2001; Herdina and Jessner 2002; Cenoz et al. 2003;
Jessner 2006, 2008a, b; De Angelis 2007; Kemp 2007; Aronin and Hufeisen 2009;
Cenoz and Gorter 2011a, b; Aronin and Singleton 2012; Singleton et al. 2013).
A final terminological refinement concerns the name of the field dealing with
acquisition of languages beyond L2. As De Angelis (2011) notices there are four
different labels regularly used in the literature in reference to the field itself:
Third Language Acquisition (TLA), Third or Additional Language Acquisition,
Multiple Language Acquisition (MLA) and Multilingual Acquisition. A compre-
hensive review of the weaknesses the terms suffer from has been proposed by De
Angelis 2007 (pp. 10–11) who herself favours the second proposition making it
even the title of the whole book. In the present volume the term Third Language
Acquisition will be applied to emphasize the situation in which a third language
is learned. Third or Additional Language Acquisition, in turn, will be used as an
umbrella term in reference to all languages learned beyond the L2 as it gives no
preference to any particular language.
1.3.3 Multilingualism and the Levels and Breadth

of Proficiency
A further terminological refinement that needs to be noted concerns the levels and
breadth of proficiency needed to apply the term to persons using two or more lan-
guages. As far as definitions of bilingualism based on competence are concerned,
they range from “native-like control of two languages” (Bloomfield 1933, p. 56)
WWW.EBOOK777.COM
and “active, completely equal mastery of two or more languages” (Braun 1937

p. 115) through “the practice of alternately using two languages” (Weinreich 1953,
p. 1) and “an ability to produce complete meaningful utterances in the other lan-
guage” (Haugen 1956, p. 7) to, more recently, “having a choice of two available
languages for conversation” (Votaw 1992, p. 299). Generally speaking, earlier
definitions were based on a monolingual bias and tended to restrict bilingual-
ism to equal mastery of two languages, while later ones have allowed much
greater variation in competence (cf. Cook 2000, Edwards 1994, 2012). Whereas
Bloomfield (1933) and Braun (1937) concentrated on the control an individual
has over his linguistic systems, Diebold (1964) left out the question of minimal
proficiency, and instead allowed for the fact that some people are bilingual to a
certain degree calling them “incipient bilinguals” (cf. semilinguals; Hockett 1958).
Macnamara (1967) followed this line of thinking and defined bilinguals as those
having some basic competence in one of the four skills (speaking, listening, read-
ing or writing), whereas Li Wei perceives a bilingual as someone who “can com-
municate in more than one language, be it active (through speaking and writing) or
passive (through listening and reading)” (2008, p. 4). Some linguists went even so
far as to endorse Edwards’s proposal declaring everyone familiar with even a few
expressions a bilingual:
Everyone is bilingual. That is, there is no one in the world (no adult, anyway) who does
not know at least a few words in languages other than the maternal variety. If, as an
English speaker, you can say c’est la vie or gracias or guten Tag or tovarisch – or even
if you only understand them – you clearly have some “command” of a foreign language
(Edwards 1994, p. 7).
The same competence controversy holds for multilingual users. The question
arises whether the label multilingual should be reserved only for those individuals
whose proficiency is native-like and balanced across all their languages and across
the range of language skills (cf. Aronin and Singleton 2012). Following Cook’s
studies on multicompetence (e.g. Cook 2007a, b, Sect. 1.4). It seems to be more
likely that the proficiency levels necessary for a language user to be classified as a
multilingual cannot be specified.
Finally, another question is connected to the proficiency threshold level in
non-native languages. Namely, when do individuals start being multilingual?
How many hours/years of instruction do they need to become bi- or multilin-
gual users? How proficient do they need to be in their L2, L3 or Ln? And how is
their language proficiency to be assessed? How much is enough? (see Aronin and
Singleton 2012, pp. 2–3 for a comprehensive discussion).
1.4 Towards a Holistic View of Multilingualism
Multilingual language users constitute a new testing ground in psycholinguistic

and educational research. As has been shown in the previous section, this interest
results, to a great extent, from the contention that a multilingual speaker exhibits
WWW.EBOOK777.COM
1.4 Towards a Holistic View of Multilingualism 11
linguistic features not necessarily compatible with those of bilingual speakers. The
assumption which is strictly related to the notion of multicompetence. The idea
of a ‘holistic’ interpretation of bilingualism as opposed to a monolingual ‘frac-
tional’ interpretation of bilingualism, initially put forward by Grosjean (1992,
2010), has over the last two decades attracted significant research attention in
the study of Third or Additional Language Acquisition. Similarly, the concept of
“holistic multicompetence” (cf. Cook 1992, 1995, 2008, 2012a, b) proposed by
Cook has been based on the essential tenet of holism which postulates that the
whole is more than the sum of its parts. By the same token, “a bilingual is NOT
the sum of two complete or incomplete monolinguals” (Grosjean 1992, p. 55, cap-
italized in the original), just like the trilingual is not the extension of the bilin-
gual speaker (Jessner 1996, 2006, 2008b; Herdina and Jessner 2002; Aronin and
Hufeisen 2009; Aronin and Singleton 2012). According to Cook’s pioneering defi-
nition relating to bilingualism, multicompetence is “the compound state of mind
with two grammars” (Cook 1992, pp. 557–558). Herdina and Jessner postulate
that multicompetence is an integrated entity “consisting of dynamically inter-
acting linguistic subsystems” (Herdina and Jessner 2002, p. 75), whereas Clyne
speaks about the multilingual’s “multilateral competence” (Clyne 1997) which
encompasses linguistic as well as procedural knowledge and finds evidence for
it in various aspects of multilingual, more precisely trilingual, behaviour. Finally,
taking the individual aspect of multilingual proficiency into account, Aronin and
Ó Laoire (2004) perceive the nature of multicompetence as an ecosystem or a
“bionic system” arguing at the same time that the study of multilingualism should
be based above all on the notion of identity. Unfortunately, space constraints pre-
vent this section from offering a detailed description of the multitude of works on
multicompetence. Therefore, in what follows, only a brief look at this fundamental
concept is presented.
First of all, Cook (1992) argues that what differs monocompetence from multi-
competence is not only the number of languages involved, but also the quality of the
linguistic knowledge. In his influential article presenting evidence for multicompe-
tence (1992), the author claims that since L2 users differ from L1 users in their L1
knowledge, metalinguistic awareness and cognitive processes, they are not simply
equivalent to two monolinguals but instead should be perceived as unique combina-
tions (Cook 1992). In the same vein, Cenoz and Genesee (1998, p. 19) indicate that
“multilinguals possess a configuration of linguistic competences that is distinct from
that of bilinguals and monolinguals competence”. Further, it is claimed that mul-
tilingual language competence contains the linguistic aspects, i.e. vocabulary and
grammar, from all the language systems of the learner, and also the pragmatic com-
ponent, consisting of sociolinguistic, discourse and strategic competences pertaining
to all the languages involved. Additionally, it includes the ability to function in mul-
tilingual contexts, which require decisions on code choice and code-switching. As
Grosjean postulates, bilinguals (and by extension multilinguals) have “a unique and
specific linguistic configuration” (Grosjean 1992, p. 55). And since for multilinguals
it is normal to move between different languages, switching, mixing and borrowing,
clearly, multilingual learner performance must exhibit linguistic features which are
WWW.EBOOK777.COM
exclusively multilingual. Therefore, paraphrasing Cook, it could be argued that the

language competence of the multilingual speaker must be “assessed in terms of his
or her total language repertoire, and not in relation to individual languages only”
(Cook 1991, p. 112). The contention that resonates with many scholars (cf. Herdina
and Jessner 2000, 2002; Aronin and Hufeisen 2009), who, like Cook, believe that
multicompetent speakers have a different command of their languages from that
possessed by monolinguals or even bilinguals.
Holistic approaches such as Cook’s notion of multicompetence as well as
Grosjean’s language mode conception (discussed in Chap. 3) which both describe
the multilingual as a multicompetent but specific speaker-hearer whose mind is not
comparable to the monolingual in either language, echo in one of the most influen-
tial models of multiple language acquisition, namely Herdina and Jessner’s (2002)
Dynamic Model of Multilingualism (DMM) inspired by and based on the dynamic
systems theory (DST). According to DMM, multilingual proficiency is defined as
the dynamic interplay between various psycholinguistic systems, cross-linguistic
interactions and the M(ultilingualism) factor which refers to all the effects in mul-
tilingual systems that distinguish multilinguals from monolinguals (Jessner 2008a,
p. 25). As Jessner summarizes, “according to DMM, the development of a multi-
lingual system changes over time, and is non-linear, reversible—resulting in lan-
guage attrition or loss—and complex” (ibid.).
Finally, the idea of multicompetence can also be traced in studies on the multilin-
gual mental lexicon advocating the view that the mental lexicon of a multilingual user
is fully integrated and “fundamentally nonselective”, no matter how many languages
are involved (e.g. Dijkstra 2003a, b). This contention rests on findings from a wide
range of experimental studies (cf. Chaps. 3 and 4) confirming that, when a particu-
lar word form is activated, formally similar words (interlingual homographs, interlin-
gual neighbours and psycholinguistic cognates) known to a given individual are also
activated, even if they have different language affiliation than the words in question.
Consequently, as Van Heuven et al. (2008, p. 2706) note, “the bilingual brain can-
not avoid language conflict, because words from the target and nontarget languages
become automatically activated”. Some more detailed discussion of the concept of
multicompetence in the multilingual brain will be continued in Chaps. 3 and 4.
1.5 Characterizing Multilingual Learners
While the previous sections endeavoured to shed some light on the terminological
problems related to specifying and differentiating the notions of bi- and multilin-
gualism by offering a concise review of definitions connected with these phenom-
ena, the subsequent paragraphs will focus on developmental patterns and cognitive
abilities of multilinguals. The following subsections will also seek to address
the pertinent issue of the qualitative differences that have been reported to occur
between multilingual and bilingual language users and which have been briefly
referred to in the foregoing discussion on multicompetence.
WWW.EBOOK777.COM
1.5 Characterizing Multilingual Learners 13
1.5.1 Contexts of Multilingual Acquisition

and Routes of Learning
For many years research has only applied the term multilinguality to cases in
which “learners are characterized as bilingual or multilingual as a result of natural
language acquisition processes and have a similar level of competence in all lan-
guages” (Cook 2000, p. 23). However, in recent years, the situation has changed
considerably. Consequently, with reference to the taxonomy of multilingual stud-
ies, many researchers distinguish, nowadays, between natural and formal multilin-
guality (cf. Fig. 1.1), at the same time emphasizing the complexity of both groups
and the variability among multilingual speakers themselves.
As shown in Fig. 1.1, natural multilinguality comprises citizens of bi- and
multilingual countries, children from mixed marriages and immigrants com-
ing to live in another country (both permanently or temporarily), whereas the
partially naturalistic setting of acquisition refers to natural bilinguals learning
a foreign language in the classroom context, e.g. a bilingual child brought up
with two school languages who learns another language in the formal setting. In
turn, formal multilinguality refers to monolingual or bilingual learners acquir-
ing their subsequent language(s) by means of formal instruction, especially in a
school context.
The complexity of the two former groups, is believed to be determined by
the social context, whose influence is difficult to measure since it is shaped by a
multitude of social, educational and political factors (cf. Barron-Hauwaert 2000;
Singleton et al. 2013). The formal type of multilinguality, on the other hand, the
one dependent on classroom instruction, is complex due to all possible language
configurations, typological and psychotypological proximity of the languages, the
learners’ motivation, the age factor, the previous linguistic knowledge and the prior
Fig. 1.1 Multilingual contexts of acquisition and subjects they concern (based on

Hoffmann 2001a; Gabryś-Barker 2005)
WWW.EBOOK777.COM
Table 1.1 Patterns of trilingual development (based on Cenoz 2000, p. 40)

The order of acquisition Description
(a) L1/L2/L3 All the three languages acquired simultaneously
(b) L1/L2 → L3 A bilingual person acquiring a foreign language
consecutively
(c) L1 → L2/L3 Both foreign languages acquired after L1
simultaneously
(d) L1 → L2 → L3 Languages acquired consecutively
language learning experience, the metalinguistic and strategic awareness and, last
but not least, the instruction itself. These points granted, there is no denying the
fact that “the very nature of the process of formal multilinguality may create pat-
terns very different from those of the natural multilingual context” (Gabryś-Barker
2005, p. 19).
In bilingual acquisition two basic orders of linguistic development have been
distinguished. These are: compound (L1 → L2) and coordinate (L1 + L2) bilin-
gualism. As suggested above, in the case of multilingualism the situation is far
more complex since the number of languages involved in multilingual acquisi-
tion multiplies their possible acquisition orders, which in turn further compli-
cates the study of the relationships among the languages involved. In her account
of the research on multilingual acquisition Jasone Cenoz distinguished four dif-
ferent paradigms to be followed by a trilingual subject; all of them amassed in
Table 1.1. The first paradigm (L1/L2/L3) illustrates a fully naturalistic situation
typical of multilingual African countries where three or even more languages
are used by individuals on a daily basis. Type b (L1/L2 → L3) is referred to as
partially naturalistic and includes, among others, children from mixed marriages
learning another language at school or bilingual immigrants acquiring the lan-
guage of the country they had come to live in. Finally, types c (L1 → L2/L3)
and d (L1 → L2 → L3), apart from being valid for immigrants, show fully for-
mal contexts of language acquisition by monolingual speakers learning their for-
eign languages in the classroom context either simultaneously or consecutively
(sequentially).
Due to such a big number of potential patterns of multilingual development,
and often contradictory opinions on the level of proficiency necessary to classify
someone as a bi- or multilingual, it seems almost impossible to design general
models and hypotheses concerning multilinguals. Therefore, every single study
needs to precisely specify what multilingual profile it pertains to. Otherwise, its
results may turn out not to be applicable to other multilinguals. In the research
reported in Chap. 5 special attention will be given to group d as it is the most
representative category of the Polish educational context (cf. Arabski 2002;
Widła 2005; Gabryś-Barker 2005; Siek-Piskozub et al. 2008) and the one ena-
bling to analyze the importance of prior L1 and L2 language knowledge on the
process of instructed third language acquisition (which is the underlying aim of
the present work).
WWW.EBOOK777.COM
1.5.2 Cognitive Skills in Multilinguals
The discussion of the influence multilinguality exerts on learners’ linguistic and

cognitive development constitutes an important area of multilingual studies.
Arguably, the evidence gathered by psycholinguists leaves little doubt that mul-
tilingual upbringing does influence the cognitive content, thought processes and
intelligence of the involved language users. The nature of this contention, how-
ever, has altered over the last hundred years; mainly due to the application of dif-
ferent and more precise methodologies, as well as more accurate interpretation of
the obtained data. Assuming that the difference between bilinguals and multilin-
guals is not only a quantitative but also a qualitative one, an intriguing question
arises of how different these two groups are when it comes to their cognitive skills.
Do multilinguals possess better cognitive abilities than bilinguals or, on the con-
trary, do they suffer from a linguistic loss since they speak many languages but
might not be fully competent in any of them? Finally, do prior language knowl-
edge of two (or more) languages and the learning experience gathered in the pro-
cess of learning these languages facilitate the process of subsequent language
acquisition in bilinguals?
In his most comprehensive review, Hakuta describes three stages of the history
of research on bilingualism and intelligence which he terms the periods of det-
rimental, neutral and additive effects, thus indicating a developmental pattern in
research towards a positive view of bilingualism (cf. Hakuta 1986). In the 19th
century and the first half of the 20th century, generally perceived as the detri-
mental period, many researchers considered bilinguals to be at a disadvantage,
stressing that by learning more than one language, they can suffer from “brain
overload”. As Cattell put it in 1887, bilinguals might pay a cognitive price for
being able to communicate in two languages. He claimed that “foreign languages
take up much time even after they have been learned, and may lead us once more
to weigh the gain and loss of a polyglot mental life” (Cattell 1887, p. 70; in de Bot
and Lowie 2010). Likewise, a few decades later Jespersen wrote that,
[i]t is, of course, an advantage for a child to be familiar with two languages: but without
doubt the advantage may be, and generally is, purchased too dear. First of all, the child in
question hardly learns either of the languages as perfectly as he would have done if he had
limited himself to one. It may seem on the surface, as if he talked just like a native, but he
does not really command the fine points of the language (…). Secondly, the brain effort
required to master the two languages instead of one certainly diminishes the child’s power
of learning other things which might and ought to be learnt (Jespersen 1922, p. 148).
A similar viewpoint was put forward by Weisgerber (1966; in Romaine 1995), who

argued that bilingualism could impair the intelligence not just of an individual but
that of the entire ethnic group. Even as late as in the 1950s it was still common-
place to blame multiple languages for problems such as stuttering and dyslexia (cf.
Tokuhama-Espinosa 2003). Such claims derived from various Intelligence Quotient
(IQ) tests which were conducted mainly among immigrants coming in search of a
better life to the United States. Most of the results pointed to the lower language
WWW.EBOOK777.COM
score and intelligence of the tested bilingual children, e.g., tests by Goddard (1917;
in Romaine 1995) or Brigham (1923; in Romaine 1995), who concluded that the
language handicap is responsible for mental retardation, in other words for lower
intelligence of the examined bilinguals. It was only much later that their results
were highly criticized, methodological shortcomings and biased interpretation of
the data being at the foot of this criticism.
The period from the late 1950s to the early 1960s is usually referred to as the
neutral period in the history of research on bilingualism and intelligence. During
this time methodological weaknesses of previous research were brought into light
and allowed for the development of the positive period. The development of cog-
nitivism and cognitive methodology in the late 1960s allowed researchers to look
at the processes involved in language acquisition and various aspects of facilitation
bilingual competences perform, thus initiating the period of additive effects of bilin-
gualism (cf. Hakuta 1986). Since then, bilinguals have been compared to monolin-
guals with respect to their cognitive abilities in a multitude of psycholinguistic and
neuropsychological studies. A contrasting attitude to those of Jespersen (1922) and
Weisgerber (1966) towards bilingualism and its influence on children’s development
came from what is now regarded as a landmark in the studies on bilingualism and
intelligence, namely Peal and Lambert’s article on “The Relation of Bilingualism to
Intelligence” published in 1962. The researchers found that bilingual children per-
formed better than monolingual children in both verbal and non-verbal intelligence
tests. They additionally noticed that bilinguals were significantly more successful at
certain types of tasks which required mental manipulation, all of which suggests the
positive influence of bilingualism on the children’s cognitive development. Having
studied all available evidence, Peal and Lambert argued that “intellectually [the
bilingual child’s] experience with two language systems seems to have left him with
a mental flexibility, a superiority in concept formation and a more diversified set of
mental abilities” (Peal and Lambert 1962 p. 20). It needs to be noted that theirs was
one of the earliest studies which accounted for variables like age, type of bilingual-
ity or social status in the development of individual bilinguals. Since then the view
that bilingualism promotes cognitive development and facilitates third language
acquisition has been reported in a number of studies (for a review see Cenoz 2003).
Bilinguals have been repeatedly reported to present higher scores in tests of diver-
gent thinking or creative thinking (cf. Baker 2006; Cenoz 2003), to obtain bet-
ter results in tasks demanding a high level of analysis (Bialystok 2001). They are
also more sensitive to the communicative needs of their interlocutors and use more
varied communication strategies (Thomas 1992). Carringer (1974) emphasizes the
greater cognitive flexibility of bilinguals which enables them to better separate form
and content because they have two terms for one referent. Of late, bilingualism has
been even reported to “(…) protect[s] against age-related cognitive decline” and
“postpone the onset of symptoms of dementia” (Craik et al. 2010).
Considering the above, the question arises whether knowledge of a second lan-
guage further enhances cognitive development and aids in the acquisition of lan-
guages beyond the L2. The majority of research studies in this area have confirmed
that prior language knowledge and prior learning experience facilitate the acquisition
WWW.EBOOK777.COM
of additional languages and exert a positive influence on cognitive development.

Studies by Ringbom (cf. Ringbom 1987, 2001, 2007) imply that bilinguals may be
more successful learners of a foreign language than monolinguals, since they have
“a wider perspective on language” and “a greater awareness of language varia-
tion and the possibilities of expressing the same idea by different linguistic means”
(Ringbom 1987, p. 112). Also research by McLaughlin and Nation (1986) shows
that “knowing more than one language gave individuals an edge in learning addi-
tional languages because they could make certain educated guesses about language
structure” (McLaughlin and Nation 1986 in Tokuhama-Espinosa 2003). Other
researchers (cf. Bialystok 2001; Hufeisen 2004; Safont Jordà 2005; Jessner 2006;
Bialystok et al. 2010) list concept formation, higher sensitivity to communicative
needs of their interlocutors, and more advanced ability in specific uses of language
applied to certain types of tasks as abilities which bilinguals seem to be better at and
which may have a facilitative effect for third language acquisition.
Obviously, there is also the question of how proficient learners need to be to
cognitively benefit from their bilingualism in the process of third language acqui-
sition and whether the linguistic knowledge and experience can be transferred
into another language. To address this issue the reference needs to be made to
the Threshold Hypothesis and the Developmental Interdependence Hypothesis
proposed by Cummins (1976, 1979) in regard to cognitive advantages and disad-
vantages of bilingualism. On the one hand, Diaz (1985) claimed that bilingual-
ism promotes the development of cognitive flexibility from the early stages of
acquisition; on the other, basing his multilingual research on the Developmental
Interdependence Hypothesis and the Threshold Hypothesis, Lasagabaster (2001)
concluded that the more proficient the learner in the previous languages, the
stronger the likelihood that the languages will influence each other.
Yet another issue of relevance in the context of facilitative effects of bilingual-
ism on L3 acquisition is the question of how to account for bilinguals’ superiority
in language learning. This facilitative effect can be explained as related to learning
strategies, metalinguistic awareness and communicative ability, in particular if the
languages in contact are typologically close (cf. Sect. 5.7 in Chap. 5). As regards
metalinguistic awareness, it has been variably defined in the literature. For the pur-
pose of this discussion Malakoff’s definition will be used:
Metalinguistic awareness allows the individual to step back from the comprehension or
production of an utterance in order to consider the linguistic form and structure underly-
ing the meaning of the utterance. Thus a metalinguistic task is one which requires the
individual to think about the linguistic nature of the message: to attend to and reflect on
the structural features of language. To be metalinguistically aware, then, is to know how to
approach and solve certain types of problems which themselves demand certain cognitive
and linguistic skills (Malakoff 1992, p. 518)
Cenoz (2003) promotes the assumption that bilingualism has an indirect effect

on L3 acquisition through higher levels of metalinguistic awareness (the media-
tion hypothesis) which “facilitate the acquisition of language by exploiting the
cognitive mechanisms underlying these processes of transfer and enhancement”
(Cenoz 2003: 104). Admittedly, the cumulative facilitative effect of multilingualism
WWW.EBOOK777.COM
in the area of metalinguistic awareness has been reported in a number of studies

concerned with qualitative differences occurring between mono-, bi- and multilin-
guals (cf. Hufeisen 2004; Aronin and Hufeisen 2009; Pütz and Sicola 2010; Cenoz
and Genesee 1998; Jessner 1999, Thomas 1988, Cook 1991; Grosjean 1985;
Herdina and Jessner 2002; Romaine 1995; Lasagabaster 2000; Bialystok 2001,
2003; Hufeisen 2004; Jessner 2006). Research on the effects of bilingualism on
metalinguistic awareness has associated bilingualism with a higher ability to reflect
on language and to manipulate it (cf. Bialystok 1991, 2001; Cummins 1978, 1972;
Ricciardelli 1992). By way of example, Romaine defines metalinguistic skills of
bilinguals as the ability to use the language “to talk about or reflect on language”
(Romaine 1995, p. 114). She believes that a bilingual’s creativity and flexibility
may be positively influenced by the different semantic networks associated with
words in each language. Thomas (1988) concludes that apart from possessing
enhanced lexical knowledge, bilinguals develop more sensitivity to language as
a system, which helps them in the solution of certain linguistic tasks (especially
problem solving tasks). Similar findings come from the works of Bialystok (cf.
Bialystok 1991, 2001), who compared the ability of monolinguals and bilinguals
to solve problems in three different language domains. Analyzing the experiments
involving oral language use Bialystok found that bilingual children have height-
ened metalinguistic awareness because they are regularly forced to pay attention
to language forms (i.e. which language is spoken) in order to make decisions on
their own language choice, which places additional demands on their control abili-
ties and facilitates the process of additional language learning.
Much in the same vein, many researchers accentuate that the addition of sub-
sequent languages help “multilinguals develop new language-learning skills and
strategies” (Aronin 2005, p. 14), as well as greater sensitivity to social aspects of
communication simply by forcing them to navigate through a number of complex
pragmatic situations (cf. Baker 1993). Paraphrasing Jessner (2006), Aronin and
Singleton (2012) claim that even only receptive language knowledge of additional
languages “may enhance metalinguistic awareness and have a significant prepar-
atory value for further language acquisition” (Aronin and Singleton 2012, p. 3).
One of the most comprehensive studies on this topic has been provided by Jessner,
who conducted a detailed discussion on the use and usefulness of linguistic aware-
ness in multilingual language teaching concluding that additional knowledge in
the mind provides further metalinguistic knowledge learners can rely upon during
the learning process, particularly if the additional languages are studied in formal
learning contexts (Jessner 2006).
1.6 Research Themes in the Study of Multilingualism
Despite a growing number of thought-provoking papers focusing on issues pertain-

ing to multilingualism and its pragmatic implications, the literature on multilingual-
ism, as distinct from bilingualism, was, until recently, rather limited and deeply
WWW.EBOOK777.COM
1.6 Research Themes in the Study of Multilingualism 19
rooted in the bilingual background. In recent times, however, psycholinguists have

started to realize that multilingual studies based exclusively on research into bilin-
gualism and SLA are, undoubtedly, unable to deal with and explain all the phe-
nomena occurring in multilingualism (cf. Cenoz 2000; Cenoz et al. 2003; Herdina
and Jessner 2002; Hufeisen and Neuner 2004; Jessner 2008a, b; Aronin and
Singleton 2012). Many scholars subscribe to the view that not only are multilin-
gual acquisition and multilingualism more complex than their two-language coun-
terparts, but they also require different skills of the learner. Accordingly, a rising
number of multilingual research studies using alternative, more advanced method-
ologies have mushroomed in the last decades and the field once perceived as pioneer-
ing has grown enormously. As a reflection of the increased interest in multilingual
matters, over the past few years, a number of European universities have introduced
chairs in fields of study related to multilingualism and interculturalism. Research on
multilingualism has been reported in a number of publications including a special-
ized journal exclusively dedicated to multilingualism and TLA—The International
Journal of Multilingualism, and at several international conferences dedicated solely
to the themes of multilingualism and additional language acquisition. Considerable
increase in multilingual research has been reported even in typically monolingual
countries such as Poland (cf. Siek-Piskozub et al. 2008). In the present section a
brief and rather eclectic look will be taken at different areas of multilingual studies.
In their critical overview of multilingualism research published in one of
the first volumes of The International Journal of Multilingualism, Marx and
Hufeisen (2004) distinguish five major areas of interest in publications on mul-
tilingualism and third language learning (cf. Table 1.2). The suggested typology
clearly shows how complex and versatile a discipline multilingual research is.
Table 1.2 Types of studies into multilingualism and MLA (based on Marx and Hufeisen 2004,
pp. 146–150)
General category Areas of research
Models and state-of-the-art Research desiderata, models of multilingualism; projects
articles promoting multilingualism and TLA (e.g., EuroCom)
Empirical studies TLA/MLA acquisition by children—case studies; accounts of
the negative and positive effects of previous foreign language
learning experience; phenomena of transfer; studies on individ-
ual differences in multilingual contexts (e.g., learning strategies,
motivation, metalinguistic awareness); discourse analysis; stud-
ies on the multilingual mental lexicon; neurolinguistic studies
Sociocultural aspects of Students’ attitudes; multilingual societies; environmental
multilingualism influences
Educational aspects of Pedagogical implications of multilingual instruction; lifelong
multilingualism language learning; the teaching of tertiary languages; curricular
suggestions; didactical and methodical issues concerning L3
teaching; different multilingual classroom concepts, cross-
linguistic influence (CLIL)
Political aspects of language Language-political aspects; political dimension of
learning multilingualism
WWW.EBOOK777.COM
Further, it needs to be noted that research on multilingualism has had an

important boost recently and has been approached from a number of different per-
spectives. In her recent article, Cenoz (2013) summarizes the themes researched
in various areas of the study of multilingualism such as neurolinguistics, psy-
cholinguistics, linguistics, education, sociolinguistics, and language policy. Such
a proliferation of different perspectives which entail different theoretical frame-
works and a wide range of methodological approaches, may lead one to believe
that research on multilingualism is highly disorganized. However, it needs to be
noted that researchers in all these areas have different goals when they try to test
hypotheses or answer research questions. Thus, instead of calling the multilin-
gual research heterogenous, it seems justifiable to say that multilingualism is the
field of multiple facets. As Cenoz (2013, p. 7) reports at the individual level some
of the most relevant areas of multilingual studies are: the cognitive outcomes of
multilingualism, the relationship between language and thought in multilinguals,
multilingual language processing and the multilingual brain, and cross-linguistic
interaction. Discussing the societal level of the studies on multilingualism, Cenoz
enumerates multilingualism as a social construct, multilingual identities, multilin-
gual practices and multilingualism, multimodality, and new technologies. In what
follows a brief overview of some selected themes in the study of individual multi-
lingualism will be presented.
1.6.1 Studies on Sociocultural Aspects of Multilingualism
There is a considerably large number of studies centred around trilingual lan-

guage acquisition in children, the most numerous being case studies which focus
on developmental aspects of trilingual language behaviour. These are usually
reports on bilingual children (i.e. children who had two first languages acquired
from their parents), who have become trilingual because they grew up in an envi-
ronment where, between home and community, three languages were used. In all
cases different languages were “associated with different persons and/or situations
identified by the subjects using the correct language to address their interlocutors”
(Hoffmann 2001a, p. 13). As Hoffmann documents, these studies are principally
just short reports on how the children of a particular family became trilingual,
together with some comments about difficulties encountered maintaining the three
languages (cf. Table 1.3).
In brief, studies on early trilingualism report on the development of different
aspects of communicative competence in different languages and the interaction
between the languages. Nevertheless, despite the growing number of projects car-
ried out with trilingual children, still more systematic and rigorous research into
this field seems to be necessary. As Hoffmann observes, we still lack “long-term
accounts which follow children well into their teens when their languages have
become fully established but their use, preference and subsequent competence
may be influenced by affective variables” (Hoffmann 2001a, p. 13). Similarly,
WWW.EBOOK777.COM
Table 1.3 A selection of studies on trilingual acquisition in children

Name Year Research
Elwert 1973 Self-report on L3 upbringing
Oksaar 1977 Language transfer
Harding and Riley 1986 Transient trilingualism
Hoffmann 1985; Aspects of trilingual development
1991
Arnberg 1987 Transient trilingualism
Barron-Hauwaert 2000 Issues surrounding trilingual families
Tokuhama-Espinosa 2000 Raising multilingual children
Deweale 2000 Trilingual first language acquisition
Cenoz 2000; Trilingual acquisition
2001
De Houwer 2004 Trilingual input and children´s language use
Hoffmann and Ytsma 2004 Issues surrounding trilingual families
Hoffmann and Stavans 2007 Trilingual and multilingual acquisition
Stavans and Muchnik 2007 Language production in trilingual children
Montanari 2009 Pragmatic differentiation in early trilingual development
Aronin and Hufeisen 2009 Multiple language acquisition
Yip and Matthews 2010 Trilingual acquisition in bi- and multilingual contexts
Chevalier 2011 Caregiver responses to the language mixing of a young
trilingual
Quay 2013 The role of caregivers and peers in trilingual language
development
Unsworth 2013 Multilingual first language acquisition
more research in early quadrilingualism (cf. Cenoz 2000), or generally speaking

multilingualism, is needed for one to be able to identify the basic characteristics of
multilingual children’s behaviour.
Another influential group of studies analyzing sociocultural aspects of mul-
tilingualism comprises projects analyzing the linguistic behaviour of trilingual
immigrants. A representative example of this area seems to be Clyne’s report
about language processing in trilinguals in Australia or, as Clyne himself states,
about “some of the things trilinguals do” (Clyne 1997). Not only is this a long-
term and a large-scale study, but it also involves different combinations of lan-
guages. What all the participants have in common, however, is their bilingual
background. They all come either from bilingual families or from a bilingual
environment. Another common characteristic is the fact that they all acquired
English only after emigrating to Australia. The central question in Clyne’s
research addresses the issue of whether trilinguals process their languages dif-
ferently from bilinguals. To answer this perplexing question Clyne centres his
discussion on three linguistic phenomena: positive transfer, negative transfer
(interference) and code-switching, the analysis of which leads him to believe
that “bilinguals and trilinguals use the same kinds of mechanisms and processes,
WWW.EBOOK777.COM
but that these are more complex when three languages are involved” (Clyne 1997,
p. 112). More importantly, the results obtained from the research lead him to con-
clude that “trilinguals (are) trilingual in different ways” (Clyne 1997, p. 113).
Clyne speculates that some trilinguals are more like bilinguals who have two dis-
tinct standard languages and “a non-standard variety regarded as part of one of
them in some ways” (ibid.), whereas other trilinguals are more like double bilin-
guals, “with two pairs of languages where each pair has a special relationship
with their L3 (English) rather than with the other languages” (ibid.).
1.6.2 Studies on Psycholinguistic Aspects of Multilingualism
One of the most dynamic areas of bilingualism, and more recently, multilingual-
ism research involves the psycholinguistic study of both children and adults. It
needs to be noted that the majority of studies within this domain predominantly
focus on bilingual users, however, a steady growth of research dealing with more
than two languages cannot go unnoticed. Perhaps the most numerous group
of studies exploring the psycholinguistic aspects of multilingualism has been
conducted into the area of multilingual language processing. The mechanisms
involved in comprehension and production in two or more languages have been
examined with reference to phonetics, lexis, and grammar. The comprehensive
amount of studies concern spoken language processing including speech per-
ception and comprehension (Grosjean 2008, 2010) as well as speech production
(de Bot 1992, 2004; Costa et al. 2005). The central question posed here relates to
the dynamic character of multilingual speech production enabling bilinguals and
multilinguals “to generally produce language in the ‘intended’ language”. Another
group of studies deals with written language processing. Here special attention has
been given to visual word recognition (Dijkstra 2005, cf. Chap. 3), far less to the
processing features of bilingual text production (Manchón 2013). As regards mul-
tilingual recognition, the leading question posed by researchers has been whether
visual word recognition is restricted to the contextually appropriate language or
whether the other languages are also involved. Overall, the research results dem-
onstrate a considerable amount of interaction between the languages known by
a multilingual, which has led the majority of researchers to believe that lexical
access in multilinguals is basically nonselective with respect to language. Of late,
however, a growing amount of research has been conducted to verify whether lan-
guage nonselectivity is maintained or eliminated in context as the empirical data
gathered thus far show that when words are processed in sentence context, their
processing seems to be sensitive to the semantic and syntactic aspects of the sen-
tence (cf. Hartsuiker et al. 2004, 2008; in Riehl 2010). Research on visual word
recognition can pride itself on a relatively long tradition and throughout many
years of studies a number of factors affecting written lexical access have been
enumerated and explored (cf. Warren 2013). They constitute yet another area of
research into multilingual processing.
WWW.EBOOK777.COM
Within the psycholinguistic area much research has been dedicated to simultaneous
(Yip and Matthews 2007, 2010) and successive language acquisition (Bialystok 2001,
De Angelis 2007, Pavlenko 2009) in multilinguals. Here, a recurrent question
addressed by psycholinguists relates to cross-linguistic influences. In fact, the study of
different forms of cross-linguistic interaction, including code switching and code-mix-
ing, has had a long tradition in research on multilingualism. Interestingly, recent stud-
ies on third or additional language acquisition have confirmed the multidirectionality
of cross-linguistic interaction. Moreover, outcomes of numerous experiments indicate
that there could be closer links between users’ foreign languages than between these
languages and the mother tongue, which seems to be most probable in the case of
typologically related languages learned after L1 (see, e.g., De Angelis 2007). Cross-
linguistic interaction has also been explored in the context of the early acquisition of
two or more languages (Paradis 2007).
Yet another group of studies analyzing multilingualism from the psycholinguis-
tic perspective covers the bilingual (multilingual) brain and cognition. Here special
attention has been devoted to the bilingual lexicon. Different models of organiza-
tion of lexical word forms and their conceptual representations in the bilingual and
multilingual lexicon have been devised and tested in practice (cf. Chap. 3; De
Groot 2002, 2011, Pavlenko 2009). The central question posed by researchers deal-
ing with the multilingual brain concerns the multilingual lexical storage and retrieval
(Dijkstra 2003a, b) both in production and comprehension. It is worth mentioning that
new possibilities of examining the multilingual brain have opened along with the use
of neuroimaging techniques (MRI, fMRI, PET) and methods to analyze the electrical
activity of the brain (ERP, EEG). They offer the opportunity to relate language pro-
cessing to different parts of the brain and to explore some characteristics of bilingual
and, more recently, multilingual processing with more accuracy and from different
angles. Further, there are also studies that investigate the implications of being bilin-
gual on various aspects of non-linguistic cognition. Many researchers (Bialystok and
Craik 2009, Bialystok et al. 2009, 2010) work on verbal abilities of bilingual children
and adults in terms of vocabulary size, metalinguistic awareness and learning to read.
Finally, mention must be made of research projects on the relationship between
multilingualism and conceptualization (Pavlenko 2009, 2011). Whereas some
psycholinguists promote the view that the conceptual base of monolinguals and
multilinguals is identical (de Bot 2008), others support the position that there are
some significant qualitative differences between mono and multilingual concep-
tualisation (Kecskes 2010). Many of them explore the relationship between the
acquisition of additional languages and conceptual development and restructuring.
Interesting findings have been reported on areas such as visual perception, inner
speech, and gesturing (Cook and Bassetti 2011; Jarvis and Pavlenko 2008).
In the foregoing section a brief and rather eclectic look at the different areas of
the psycholinguistics of multilingualism has been offered. It needs to be noted that
studies on visual word recognition are of crucial importance for the experimental
part of this volume and as such they will be further elaborated on in the following
chapters entirely devoted to a more meticulous survey of experimental studies and
theoretical models of multilingual visual comprehension.
WWW.EBOOK777.COM
1.6.3 Studies on Educational Aspects of Multilingualism
Trilingualism through schooling or formal trilingualism has recently attracted

much interest since learning a third language in the school context has become a
common phenomenon all over the world. A full review of trilingual research per-
taining to the classroom context is not possible within the confines of this chapter;
therefore, only certain selected studies will be mentioned. The majority of stud-
ies concerning trilingualism through schooling deals with language processing in
various acquisitional contexts and language configurations. A representative selec-
tion of such contributions has been presented in Table 1.4. Notably, the bulk of
these projects focus on the influence of bilingualism on the acquisition of another
language and possible methodologies and strategies that might facilitate the acqui-
sition of a third language, as well as on differences occurring between third lan-
guage acquisition (TLA) and SLA.
In brief, three recurrent questions posed by linguists dealing with formal trilin-
guals concern: firstly, the very nature of the difference between second and third
language acquisition; secondly, the problem of whether bilingualism favours the
acquisition of a third or tertiary language(s)—in other words, to what extent an
individual’s previous foreign language knowledge and foreign language learning
skills facilitate L3 acquisition. Lastly, assuming that L2 does spur L3 acquisition,
a pertinent question arises of how to additionally support the process of L3 learn-
ing for faster progress to be made.
As regards the first question, in her 2008 study, Cenoz (2008, p. 3) emphasizes
that the major challenge is “the identification of the characteristics that distinguish
third language acquisition and multilingualism from second language acquisi-
tion and bilingualism” (Cenoz 2008, p. 2). She has also singled out three leading
research fields of TLA. The first field deals with the effect of bilingualism on third
language acquisition focusing on differences in bilingual and monolingual learn-
ers acquiring a target language. Another group of studies explores cross-linguis-
tic influence in third language acquisition and analyzes metalinguistic awareness.
Finally, some researchers address the issue of the age factor in third language
acquisition at school indicating quite surprisingly that instruction in foreign lan-
guages from an early age in the school context does not necessarily mean higher
achievement in all cases (cf. Cenoz 2008).
As for the issue concerning the facilitative role of bilingualism in TLA, the
findings are, for all language learners, more than encouraging. Bearing in mind
a number of multilingual studies, it can be concluded that bilinguals may indeed
have certain advantages with respect to general language proficiency and there-
fore they may be able to acquire a third language more easily than a monolin-
gual learns a second language. By way of illustration, the empirical findings of
the Basque studies (cf. Cenoz and Lindsay 1994; Lasagabaster 2000; Joaristi
et al. 2009) confirm that high levels of competence in L1 and L2 produce high
levels of competence in L3 since the Spanish-Basque bilingual pupils performed
significantly better than their monolingual peers when it came to learning an
additional language, the contention that has been further supported by Muñoz’s
WWW.EBOOK777.COM
Table 1.4 A selection of studies on trilingualism and multilingualism in the educational context

Name Year Research
Thomas 1988 Children from minorities (USA)
Bild and Swain 1989 Trilingualism in Canada
Valencia and Cenoz 1992 Trilingualism in the Basque country
Cenoz and Lindsay 1994 The age factor in L3 acquisition; the Basque country
context
Klein 1995 US context trilingualism and multilingualism
Genesee 1998 Trilingualism in Canada; focus on social motivations
Cenoz 1998 The Basque country L3 education
Muñoz 2000 Trilingualism through schooling in Catalonia
Lasagabaster 2000 Trilingual education in the Basque country
Björklund and Suni 2000 Trilingual education in Finland
Ytsma 2000 Description of trilingual primary education in Friesland
Iatcu 2000 Overview of english for Hungarian-Romanian bilinguals
Hinger 2001 Multilingual approach to language acquisition in school
context
Lasagabaster 2001 University students’ attitudes towards english as an L3
Spöttl 2001 Vocabulary learning by trilinguals
Arabski 2002 Learning strategies of L1, L2 and L3 lexis
Gabryś-Barker 2005 Multilingual storage, access and retrieval
Widła 2005 Acquisition of third languages
Lam 2007 Multilingual education in China
Joaristi et al. 2009 Trilingualism in the educational system of the Basque
country
Cenoz and Jessner 2009 Trilingualism in educational contexts
Cenoz 2009 towards multilingual education
Creese and Blackledge 2010 Translanguaging in a multilingual classroom
de Mejía 2012 Towards multilingual education
Hornberger and Link 2012 Translanguaging and transnational literacies in multilin-
gual classrooms
Ellis 2013 The ESL teacher as plurilingual
Wang and Kirkpatrick 2013 Trilingual education in Hong Kong—a case study
Hélot and Ó Laoire 2013 Language policy for the multilingual classroom
Le Pichon et al. 2013 Strategic competence in young plurilingual children
Catalan research (cf. Muñoz 2000). Additionally, the positive experimental find-

ings in southern Europe have been confirmed by the results in Finland, where
English as an L3 was examined in “Swedish minority immersion for Finnish-
speaking children” (cf. Björklund and Suni 2000; Ringbom 2007). Also the results
concerning general aspects of proficiency indicated that bilingualism has a positive
effect on third language acquisition when L3 acquisition takes place in additive
contexts and when bilinguals have acquired literacy skills in both their languages
(cf. Cenoz and Genesee 1998; Safont Jordà 2005).
WWW.EBOOK777.COM
Finally, there is the issue of using the advantage bilingualism offers in L3

instruction so that previous language knowledge and language learning experience
are no longer ignored or left unexploited. As Neuner (2004) observes from the
point of view of plurilingual didactics, the goal of language teaching has changed
fundamentally. The aim is no longer to ensure that learners accumulate as much
language knowledge and as many language skills as possible in the individual lan-
guages. In contrast, as Hufeisen indicates,
[i]t makes much more sense to structure the basic elements of foreign language learning
in such a way that profiles can be developed in the individual languages that correspond
to the communicative (…) needs of learners in using the language and that can be further
developed later on, if needed, after the completion of schooling (Hufeisen 2004, p. 8).
The claim that plurilingualism is innate in all individuals, because even in their
mother tongue they shift back and forth between a number of often clearly dif-
ferentiated language variants appeared already in late 1970s in the publications
by Mario Wandruszka who distinguished two types of this phenomenon. The
first type, referred to as “internal plurilingualism”, concerns the already men-
tioned language variants such as standard language, dialect, colloquial lan-
guage, specialist jargon or knowledge of earlier linguistic forms of their own
language that are nonetheless linked with each other. To this was added indi-
viduals’ ability to learn other languages in addition to their own in the course of
their intellectual development. The latter type was called “external plurilingual-
ism”. In fact, the term plurilingualism in its modern usage requires some clari-
fication since in the last few decades it has come to denote a range of different
meanings.
To date research has typically applied the term plurilingualism, as synony-
mous to individual multilingualism (cf. Sect. 1.3.2), to cases in which learners
are characterized as plurilingual as a result of natural language acquisition pro-
cesses and have a similar level of competence in all languages (cf. Hufeisen and
Neuner 2004). However, plurilingualism as it has emerged in the current discus-
sions of foreign language teaching and as it is also used in Council of Europe
documents is an utterly different concept. Using the metaphor of a “threshold
level” Christ (2001, p. 3) suggests that a person is plurilingual if, “with respect
to a number of languages, he/she has learned to cross the threshold into these
different language houses.” Indeed, as a result of the circumstances of their lives
and the demands of their jobs, plurilingual learners usually reach different lev-
els with respect to the four competences. More significantly, plurilingualism
is at present defined and distinguished from multilingualism as a plurilingual
competence that does not simply add on the skills in one language to those of
another, but rather combines and interrelates them in a variety of ways. This is a
plurilingualism that is inter alia characterized by interlinguistic and intercultural
awareness (cf. Abendroth-Timmer and Breidbach 2000; Jessner 2006; Aronin
and Hufeisen 2009). Beyond this, the plurilingual approach emphasizes the fact
that as an individual person’s experience of language in its cultural contexts
WWW.EBOOK777.COM
expands they do not keep their languages and cultures in strictly separated men-
tal compartments, but rather “builds up a communicative competence to which
all knowledge and experience of language contributes and in which languages
interrelate and interact” (Council of Europe 2001, 4; cf. Sect. 1.4). Basing on the
principles of Cook’s multicompetence (1992, 2007a, b) and Grosjean’s holistic
view of bilingualism, Neuner enumerates (2004, pp. 15–16) the following fea-
tures of plurilingual learners, emphasizing the specificity of the contemporary
interpretation of the term in question. Firstly, the level of competence and the
language profile in each language learned can be very different. Secondly, there
is no need to achieve the ideal of “near nativeness” in each of the languages to
be learned. Lastly, when several languages are learned, the learner does not begin
“at zero” in each case, but rather the existing language possession is continually
extended by each new language.
As regards types of plurilingual learners, the same author singles out three
types of the phenomenon in question: retrospective, retrospective-prospective
and prospective (cf. Neuner 2004, p. 15). Retrospective plurilingualism means
that learners bring their plurilingualism into the classroom. They are (to a
large extent) bilingual, with considerable knowledge of L2, the language being
taught, and thus have a definite advantage in terms of knowledge and skills in
this language over the other learners. Retrospective-prospective plurilingual-
ism refers to learners who bring their plurilingualism into the classroom and
therefore have a substantial lead in linguistic knowledge over the other learners,
but neither of these two languages is the subject being taught. Through teach-
ing in an L3 (or Ln) the learners are extending their plurilingualism. Finally,
prospective plurilingualism is used in relation to the learner who arrives in the
foreign language classroom as a monolingual and first begins to develop and
extend their plurilingualism in the teaching of the foreign language. This is the
situation that is assumed to be typical for foreign language teaching in the first
foreign language. The research presented in the final chapter of this volume
focuses on retrospective-prospective plurilinguals. It aims at analyzing their
ability to use the already acquired lexical knowledge in the process of visual
word recognition.
1.6.4 Political Perspective of Multilingual

Education in Europe
For many decades it has been overlooked that in whatever form or conditions the
vast majority of the European population is multilingual. In fact, it was not until
the development of the European integration that the monolingual bias began to
disappear. Nowadays, as some researchers believe, “the concept of the monolin-
gual native speaker has no place in the new Europe” (Cheshire 2002, p. 33). Today,
WWW.EBOOK777.COM
the European Union is founded on the “unity in diversity” principle—diversity of

cultures, customs and beliefs and of languages. Besides the 24 official languages of
the Union, there are 60 or so other indigenous languages and scores of non-indige-
nous languages spoken by migrant communities. As confirmed by a number of leg-
islative documents, respect for linguistic diversity is a core EU value. Article 21 of
the Charter of Fundamental Rights of the European Union forbids discrimination
on several grounds, specifically including language, while Article 22 guarantees
respect for cultural, religious and linguistic diversity. Clearly, Europe has come a
long way from marginalizing minority languages and establishing the dominance
of monolingual national standard ideologies to promoting multilingualism and
multilingual education, or, to put it in more European terms plurilingualism and
plurilingual education, as one of the leading aims of the EU language policy.
Since 2001 the European Commission’s survey and analysis service,
Eurobarometer, has carried out a series of research projects into the lan-
guage skills of European citizens and their attitudes towards languages
(cf. Eurobarometer 2001, 2006, 2011, 2012). The surveys show many intrigu-
ing results. With regard to the command of foreign languages, 53 % of the
respondents claimed that they could speak a foreign language in addition to
their mother tongue in the 2001 survey. Significantly, in 2006 this figure rose
to 56 %. According to the 2011 survey, the most multilingual EU citizens were
the Luxemburgers—96 % of the people know at least one other foreign lan-
guage, followed by Latvians (87 %) and Maltesians (86 %). It was also stated
that one in three Europeans interviewed was fluent in at least one other lan-
guage than the language in which the interview took place. Interestingly
enough, the surveys showed that smaller member states with several state
languages exhibited greater levels of multilingualism. This was also true for
countries with lesser-used native languages (Poland seems to be a typical rep-
resentative of this group) or “language exchange” with neighbouring coun-
tries. Most importantly, with regard to the attitude towards foreign language
learning, the survey confirmed that only a minority of Europeans considered
language learning unimportant—just 8 % in 2006. Recent data coming from
Eurobarometer 2012 show that 88 % Europeans consider language learning
useful and 98 % consider it useful for the future of their children (p. 7) which,
as Komorowska rightly observes, proves that “attitudes towards individual mul-
tilingualism have definitely changed from strongly negative to extremely posi-
tive” (2013, p. 466).
This change in attitude towards multilingualism is parallel to political actions
taken in the educational domain. Significantly enough, governments and many
educational institutions are in general becoming more supportive of multilingual-
ism, and the knowledge of foreign languages is increasingly being recognized as
an asset for individuals. In fact, the need to develop multilingual instruction has
already been noticed in the EU for almost two decades now. Consequently, numer-
ous educational programmes promoting life-long language learning and a working
knowledge of three languages (one to two languages from neighbouring territories
and a lingua franca) as a goal for all citizens have been implemented in the last
WWW.EBOOK777.COM
Fig. 1.2 Key areas for multilingual development (based on a new framework strategy for multi-
lingualism, 2005)
few years in many European countries.2 The European Commission’s long-term

objective3 is to increase individual multilingualism until every citizen has practical
skills in at least two languages in addition to his or her mother tongue. In
March 2002, the Heads of State or Government of the European Union meeting in
Barcelona identified the improvement of language skills as a priority and a group
of national experts on languages has produced recommendations to Member States
which have inspired some of the following proposals concerning key areas for
action, enumerated in Fig. 1.2.
2 Cf. The European Commission's White Paper Teaching and Learning: Towards the Learning
Society (1995) considers that European citizens should be proficient in three community lan-
guages and recommends foreign language teaching at pre-school level in order to allow for sec-
ond foreign languages in secondary school.
3 Cf. A New Framework Strategy for Multilingualism [Brussels, 22.11.2005 com (2005) 596
final].
WWW.EBOOK777.COM
When it comes to language teacher training, Cenoz and Gorter (2013) recently
proposed an innovative pulrilingual approach to the teaching of English, which
as they claim softens the boundaries between languages (Cenoz and Gorter 2013,
p. 591). The authors point out that learning English in Europe cannot be separated
from the use of other languages in education, and argue for a language policy that
moves from the traditional monolingual ideology towards adopting holistic pluri-
lingual approaches in the teaching of second and foreign languages. They believe
that a plurilingual approach highlights how learners relate the languages in their
repertoire to each other when learning English as an additional language and when
they use their languages in a social context (Cenoz and Gorter 2013, p. 597). They
conclude that,
it is time for TESOL professionals to use the opportunity to accelerate the learning pro-
cess by using plurilingualism as a resource and not as an obstacle by setting attainable
goals, using the learners’ plurilingual repertoire, integrating syllabi, and using learners’
linguistic creativity as a resource (Cenoz and Gorter 2013, p. 597).
As regards multilingual development programmes, one of the most prominent

example is European Intercomprehension Programme, more commonly known as
EuroCom. It is currently being run in the three main European language groups:
Romance, Slavic and Germanic: EuroComRom; EuroComSlav and EuroComGerm,
respectively (cf. www.eurocomcenter.eu; www.eurocom-frankfurt.de). The project
has been based on the idea of intercomprehension. It concentrates on the typologi-
cal similarities of some European tongues and employs them in order to facilitate
the simultaneous acquisition of a group of typologically-related languages. As Marx
and Hufeisen assert, the project “purports the idea that, within a specific language
family (…) speakers may converse in their L1s and be understood by their listeners,
who in turn will reply in their own mother tongue(s) and likewise be understood”
(Marx and Hufeisen 2004, p. 147). In other words, the EuroCom aims to provide
European citizens with knowledge that can help them develop sufficient compre-
hension in several languages that belong to the same language family. The basic
idea behind the project is that one can easily learn to understand a language close
to his or her own. The objective is to teach transfer-based deduction strategies to
enable easier multilingual receptive competence. Additionally, the project aims at
producing books and teaching materials that “make use of the facilitative potential
for reading comprehension inherent in speakers of languages that are related, but
are not so close as to be mutually comprehensible” (Ringbom 2007, p. 103).
1.7 Conclusion
In this chapter an attempt has been made to explore different aspects of the con-
cept of multilingualism. The presentation of numerous definitions of bilingualism
and multilingualism differing as regards content and scope proves how complex
and diversified the perception of these phenomena is. The picture that emerges
from the foregoing discussion is rather fragmentary and incomplete since attempts
WWW.EBOOK777.COM
1.7 Conclusion 31
to define a multilingual user are characterized by diversified, and quite often con-
tradictory, opinions and research findings. As Jessner (2006) puts it, multilingual-
ism and multilinguality are multifaceted constructs which can be viewed from a
variety of angles and whose study has only just begun.
As regards the multilingual research, it cannot go unnoticed that the body
of multilingual literature written in the last few decades has been growing fast.
However, there is no denying the fact that further extensive studies are still needed
before all the terminological inconsistencies are solved and concrete models are
forwarded. Marx and Hufeisen (2004) propose the following desiderata for the
study of multilingualism: more longitudinal studies; the use of more learner-ori-
ented data such as learner questionnaires, introspection and retrospection or think-
aloud protocols, as well as more in-depth research on the learner’s metalinguistic
awareness or research into positive results of multilingualism and TLA.
The above discussion has shed some light on terminological complexities and
definitional problems. Numerous theoretical issues related to the phenomenon of
multilingualism have been tackled including deliberation on such concepts as mul-
tilingual cognitive skills, multicompetence or metalinguistic awareness. Special
reference has been made to Cook’s (e.g. Cook 2007a, b) notion of holistic mul-
ticompetence which posits a very high degree of integration of language compe-
tence across languages. In the final part of the chapter the discussion has revolved
around selected research areas. Since the experimental part of the work concerns
research on the multilingual mental lexicon, special attention has been paid to the
presentation of studies dealing with the psycholinguistic aspect of multilingualism.
Now that the main issues concerning the concept of multilingualism have been
delineated, it is time to outline influential theoretical models and the underlying
research experiments concerning the structure of the mental lexicon of mono- and
multilingual users.
WWW.EBOOK777.COM
Chapter 2
Modelling the Lexicon: Some General
Considerations
2.1 Introduction
In order to understand current research on multilingual processing and to

appreciate the role early theories on the mental lexicon have had in shaping this
new area of inquiry, it is useful to go back in time and examine some of the most
influential works published over the years. Notably, the present chapter tackles
problems which do not have a specifically multilingual focus; however, in research
“relating to the multilingual mental lexicon the same kinds of organizational and
operational issues arise as in L1-focused research” (Singleton 1999, p. 83), the
difference being that in the case of L2 they are further complicated by questions
having to do with precisely the fact that more than one language comes into the
picture. Therefore, what is said in the present chapter with respect to L1 lexical
processing is also relevant to L2 or, by extension, to Ln.
In the following subsections, an attempt will be made to explore different
aspects of the multifaceted concept of the mental lexicon. The chapter begins
with the discussion of the hypotheses referring to the internal structure of the
lexical entries. Issues to be examined include, inter alia, the type of the stored
information as well as the way this information is organized within an entry.
Next, the chapter goes on to discuss the wider issue of the domain of the lexi-
con. It offers a brief presentation of different definitions of the mental lexicon.
Presently, it reviews the most influential monolingual models of lexical process-
ing to be discussed within two broader theoretical frames of reference—the mod-
ularity theory and connectionism. The discussion of the numerous lexical access
models will be supplemented with some research evidence that the models seek
to account for.

DOI 10.1007/978-3-642-32194-8_2
WWW.EBOOK777.COM
34 2 Modelling the Lexicon: Some General Considerations
2.2 The Internal Composition of a Lexical Item
The mental lexicon includes a large number of lexical entries containing all the
information on individual words. But what precisely are these individual words
and what do they consist of? Within the psycholinguistic tradition two proposi-
tions concerning the issue of the internal structure of the lexical entry merit atten-
tion; the first, whose most fervent supporter is Levelt (1989, 1993) and the other,
put forward by Bierwisch and Schreuder (1992).
Many linguists (cf. Aitchison 2003a, 2012; Levelt 1989, 1993) support the view
that all the information “behind” a word can be allocated to two separate compo-
nents: a semantic component called a lemma1 (including the information on the
word’s meaning, its connotations, style, and its syntactic pattern) and a formal one,
frequently referred to as a lexeme (including the word’s morphology, phonology
and orthography). According to Levelt, each lemma has a lexical pointer which
“indicates an address where the corresponding word-form/information is stored”
(Levelt 1989, p. 165). Levelt enumerates four main internal features of a lexical
item (cf. Fig. 2.1): meaning, defined as the semantic information which lists “a set
of conceptual conditions that must be fulfilled in the message for the item to
become selected” (Levelt 1989, p. 165), syntax (including syntactic arguments and
other properties), morphology and phonology. He also points to some stylistic,
pragmatic and affective attributes of a word.
Another influential perception of the internal structure of a lexical entry was
proposed by Bierwisch and Schreuder (1992). There are many similarities between
Levelt’s and Bierwisch and Schreuder’s models since both approaches argue in
favour of similar internal features of the lexical entry. Bierwisch and Schreuder
enumerate the phonetic form, the grammatical form, the argument structure and the
semantic form. There is, however, one fundamental difference between the models.
Fig. 2.1 The structure of
lexical representations in the
mental lexicon (adaptation
based on Levelt 1989)
1 The term lemma was first used by Kempen and Hujbers (1983) in reference to the part of the
lexical entry which relates to its meaning and syntax.
WWW.EBOOK777.COM
2.2 The Internal Composition of a Lexical Item 35
Namely, the manner in which the meaning is represented. The key question is
whether semantic representations of items are identical with general world knowl-
edge or whether it is possible to draw a direct line between word meanings and con-
cepts which represent encyclopedic information. Similarly to other psycholinguists
and unlike Levelt, who is a fervent proponent of the holistic approach to meaning
representation, Bierwisch and Schreuder (1992) believe that the internal structure
of the lexical meaning of the entry is a composition of more primitive units. In con-
trast, Levelt asserts that the meaning of lexical items is “represented as a whole
which cannot be decomposed into separate elements” (Levelt 1993, p. 28).
The findings so far, then, are that although there exists a general agreement as
to the constituents of a word, two conflicting approaches concerning the issue of
the representation of meaning are still under discussion. The first approach,
called the one-level model or the network model (cf. Levelt 1989), considers
semantic and conceptual knowledge identical. The other approach, known also as
the two-level model (cf. Bierwisch and Schreuder 1992), differentiates between a
word’s semantic meaning and the more general conceptual knowledge the item
refers to. It needs to be noted that this latter theory relates to the proposition
advocating the modularity of mind (cf. Fodor 1983, 1989). In the light of this
approach a human linguistic system forms a closed mental module which does
not depend on other mental faculties but is to some extent interconnected with
them. Hence a lexical item is connected with the more general conceptual
domain of world knowledge.2
All in all, the commonly adopted approach to the structure of a lexical entry
and to the representation of meaning is the compromise between the two presented
options. On the one hand, this approach draws a boundary between semantic and
conceptual knowledge and conceives of them as non-identical. On the other hand,
it admits that these two types of knowledge are strongly related (cf. Aitchison
2003a, 2012; Pickering and Garrod 2013; Randall 2007).
2.3 Towards the Model of the Mental Lexicon
The following sections pertain to a number of issues concerning the mental lexi-
con. Firstly, an attempt has been made to review various definitions of the phe-
nomenon in question from a diachronic perspective. Subsequent sections relate to
the internal organization of the mental lexicon; to be more precise to the actual
number of storage systems. In brief, questions arise how many lexicons are to be
found in the brain and how the semantic and formal (morpho-phonological and
orthographical) components of lexical entries are stored. Are they stored together
in a unitary modality-neutral lexicon or rather separately within two different
modality-specific lexicons, and, if so, are there any direct links between the two?
2 The modularity theory will be further discussed in Sect. 2.6.1.
WWW.EBOOK777.COM
2.3.1 The Mental Lexicon Defined
The term mental lexicon was introduced by Oldfield in 1966 (Oldfield 1966; in
Singleton 1999) and since then it has been the focus of attention of a number of
psycholinguists all over the world. It has been researched and re-defined from
various perspectives many times. One of the early definitions was proposed by
Fay and Cutler who attempted to describe the mental lexicon in terms of the lexi-
con metaphor as “the listing of words in the head” (1977, p. 509). The evidence
they cite to support their claim demonstrates that the majority of words, excluding
onomatopoeias, are characterized by arbitrary sound-meaning relations. Fay and
Cutler (1977, pp. 508–509) offer the following description of the mental lexicon:
What is this mental dictionary, or lexicon, like? We can conceive of it as similar to a
printed dictionary, that is, as consisting of pairings of meanings with sound representa-
tions. A printed dictionary has listed at each entry a pronunciation of the word and its
definition in terms of other words. In a similar fashion, the mental lexicon must represent
at least some aspects of the meaning of the word, although surely not in the same way as
does a printed dictionary; likewise, it must include information about the pronunciation of
the word although, again, probably not in the same form as an ordinary dictionary.
While some linguists compare the mental lexicon to a written dictionary, others
describe it as a network of interconnected nodes similar to bundles of neurons in
the brain. Aitchison (2003a, p. 248) rightly argues that “the mental lexicon is (…)
concerned above all with links, not locations” and observes that “the lexical con-
nections in the mind are far from what we normally imagine a dictionary or lexi-
con to be”. When a word is activated, other words of similar form (Stamer and
Vitevitch 2012), meaning (Mirman 2011), syntax (Kim and Lai 2012), orthogra-
phy (Carreiras et al. 2013) or emotional content (Bayer et al. 2012) are also acti-
vated, suggesting that the mental lexicon is complex and highly interconnected.
Emmorey and Fromkin (1988) propose to view the mental lexicon as
that component of grammar in which information about individual words and/or mor-
phemes is entered, i.e. what a speaker/hearer of a language knows about the form of the
entry (its phonology), its structured complexity (its morphology), its meaning (its seman-
tic representation), and its combinatorial properties (its syntactic, categorical proper-
ties) (…) also orthographical or spelling representation (Emmorey and Fromkin 1988; in
Gabryś-Barker 2005, p. 38).
According to Singleton (1999), the mental lexicon is a module in human long-

term memory which contains the speaker’s all knowledge concerning words in his
or her language(s). Marslen-Wilson rightly describes the mental lexicon as “the
central link in language processing” (1992, p. 9). In Levelt it is argued that the
speaker’s mental lexicon is a “repository of declarative knowledge about the words
of his language” (1989, p. 182). For the purpose of this work, however, a much
more recent definition by Roux (2013, p. 82) seems suitable, which sees the men-
tal lexicon as “the collective representation of words in the mind, which draws
together contextual, personal and interpersonal dimensions of meaning, and assists
most fundamentally in the acquisition, retention and expression of language.”
WWW.EBOOK777.COM
2.3 Towards the Model of the Mental Lexicon 37
Before anything more is said about the structure of the mental lexicon, it is
imperative to realize that finding common patterns in language errors is believed
to provide valuable information about the nature of the internal lexical storage sys-
tem. Thus, error analysis constitutes a basis and seems to be a perfect source of
data in research on language processing (cf. Fromkin 1973). Admittedly, errors in
any language system have an incalculable explanatory value. The evidence from
word searches and “slips of the tongue”, selection errors known as malapropisms,
but also psycholinguistic experiments and research with aphasic patients, show that
lexical items in the mental lexicon are interconnected in a wide variety of ways.
Fay and Cultler (1977) based their model of the mental lexicon on malaprop-
isms (cf. Vitevitch 1997; Goldrick et al. 2010). These are speech or writing errors
“in which a word similar in sound to the intended one is uttered as in The cold
is being exasperated by the wind instead of The cold is being exacerbated by the
wind” (Aitchison 2003b, p. 71). However, there are three basic conditions an erro-
neous word needs to meet in order to function as a malapropism. Firstly, the mean-
ing of the error and the target word needs to be unrelated. Secondly, erroneous
intrusion should sound similar to the intended word. Thus, using tattoo instead of
book cannot be classified as a malapropism; whereas substituting tattoo for taboo
would be. Lastly, the word becomes a malapropism providing it has the so-called
recognized meaning in the user’s language. Consequently, coining a non-existent
or ungrammatical word by adding some affixes does not make a word a malaprop-
ism. Moreover, Fay and Cutler claim that
(…) the malapropisms, have some interesting properties. First, the target and the error
are of the same grammatical category in 99 % of the cases. Second, the target and the
error frequently have the same number of syllables (87 % agreement in our list). Third,
they almost always have the same stress pattern (98 % agreement) (Fay and Cutler 1977,
pp. 507–508).
On the basis of their findings, Fay and Cutler proposed a model which assumes
that lexical storage is phonologically governed. The mental lexicon is conceived
of as a network which “lists entries that have similar phonological properties near
each other” (Fay and Cutler 1977, p. 512). Recapitulating, words beginning with
the same phoneme are listed together, whereas words sharing the same second
phoneme are grouped in a subcategory of that class and so on. It needs to be added
that Fay and Cutler do not exclude the possibility of arrangement by syntactic cat-
egory. Nevertheless, they do not provide any further details of such a concept.
Admittedly, there are considerable similarities between a traditional dictionary
and the human mental lexicon. They are both organized along some underlying
principles based on the characteristics that words share. Clearly, in the case of a
written dictionary, the basic criterion of organization is orthography. Words in a
book lexicon are always stored in alphabetical order. Consequently, if we want to
look up a word we need to identify its initial letter, find words beginning with that
letter and, finally, again in alphabetical order, exhaust the possibilities until the
right entry has been found. Locating the word enables us to gain access to all the
related data hidden “behind”—the semantic, phonetic, and pragmatic information.
WWW.EBOOK777.COM
Similar to a dictionary, the mental lexicon is comprised of a substantial number

of lexical entries with linguistic information “behind” them, the complexity of the
storage, however, being far more sophisticated.
In the first place, the lexical entries in a traditional dictionary are static, whereas
the mental dictionary is dynamic. Not only do languages evolve constantly, but the
individual linguistic knowledge of a language speaker also changes over time (cf.
Aitchison 2003a). Consequently, the mental representations change—new meanings
are added, while words which are rarely or never used become inaccessible. Another
critical difference between a tangible dictionary and the mental one is the accessibil-
ity of the information being stored. In a book dictionary we can easily get equal
access to any of the chosen entries. By contrast, words stored in our mind have dif-
ferent degrees of accessibility. It is argued that frequency of use, context and image-
ability3 are the most common factors influencing the accessibility of a given word.
A further, but concurrently the most radical, difference is the form of the stored
information. A written dictionary is simply an inventory of verbal information. The
dictionary in the human brain, on the other hand, includes both verbal—linguistic
and non-verbal—conceptual data. Schreuder and Flores d’Arcais (1989) describe
this characteristic feature of the human mental lexicon in the following way:
A word in the mental lexicon has, besides its lexical properties, nonverbal percepts, con-
ceptual representations and images that are derived from “real-life experience” and are
stored in episodic memory (Schreuder and Flores d’Arcais 1989, p. 422).
As Bakhtin (1981) formulates it, “every word smells of the context (…) in which
it has lived its intense social life” (Bakhtin 1981; in Gass and Selinker 1994,
p. 276). In communication, language users depend on the contexts in which words
appear to a significant degree, inferring word senses on the basis of linguistic as
well as non-linguistic data, the latter being frequently even more informative.
However influential the lexicon metaphor can be, many cognitive psychologists
and psycholinguists reject it claiming that the mental lexicon is much more than
just a repository of lexical items. The advocates of the cognitive approach posit
that the mental lexicon consists of concepts and their linguistic realizations, both
phonological and orthographic. They conceive of it is as a conceptual system. As
Gabryś-Barker puts it,
A mental lexicon should be seen more as a conceptual system than a pure inventory of
entries, a system which is composed of concepts and their linguistic realisations both pho-
nological and orthographic, and with strong emphasis put on lexical processing (…) that
is to say, access and retrieval as evidence of the working structure of the mental lexicon
(Gabryś-Barker 2005, p. 39).
Notably, the standard position in language processing is that the mental lexicon
is a largely fixed resource, acquired during early development. Although people
can of course add new lexical entries during their adult life, this is generally seen
3 According to Aitchison, imageability is “the extent to which something can be visualized”
(2003b, p. 57).
WWW.EBOOK777.COM
as a marginal activity. Studies of processing assume that people already know the
language that they use and that there is a clear demarcation between acquisition
and processing. (cf. Aitchison 2003a, 2012; Cutler 2005). In addition, the lexicon
is treated as a store that principally consists of small units (either words or mor-
phemes) and that knowledge of larger units is largely limited to idioms, which are
regarded as fairly peripheral to “core” language processing.
More recently, Pickering and Garrod (2013) proposed an alternative view
of the mental lexicon that is consistent with the Dynamic Systems Theory
(cf. Briggs and Peat 1989). They based their proposition on the evidence from
dialogues which shows that interlocutors make use of fixed or semi-fixed expres-
sions during a particular conversation with meanings that are established through
that conversation. They also argued that language users “routinize” (Pickering
and Garrod 2005, p. 87) these expressions by storing them in the mental lexicon,
normally for that conversation alone. This requires a conception of the lexicon
in which complex expressions (of all kinds, not just established idioms) can be
stored alongside more traditional lexical units. On this view, the lexicon can be
constantly and dynamically updated, and the strict division between acquisition
and adult usage is removed.
The final paragraphs of this section attempt to shed some light on the research
concerning the much debatable problem of the size of the mental lexicon. It is
generally believed that the mental lexicon is comprised of a huge number of lexi-
cal entries; however, its exact size remains undefined. In the early research con-
ducted by Seashore and Eckerson in 1940 (in Aitchison 2012) the number of
words stored in the mental lexicon of an educated adult was estimated at about
150 thousand receptive words with 90 % available for production. A similar study
carried out by Diller in 1978 resulted in an unpredictably high number of about
250 thousand words, whereas the more recent work by Levelt (1989) rated the pro-
ductive vocabulary of an educated adult at no more than 30 thousand word fami-
lies. According to Clark (1993), on the other hand, adult speakers of a language
have at their disposal between 20 and 50 thousand productive words, the amount
of receptive vocabulary being “considerably larger” (1993, p. 13). All things con-
sidered, average educated adult language users have at their disposal a production
vocabulary of between 20 and 50 thousand words and comprehension vocabulary
of between 150 and 250 thousand words.
Why are the research results so diverse? Many linguists postulate that such
sharp differences are connected with the failure to distinguish between produc-
tive and receptive vocabulary. Consequently, different experiments employ either
active vocabulary exclusively or involve both passive and active words. Some
researchers concentrate on active vocabulary (thus achieving lower numbers);
other experimenters use both passive and active words (those used exclusively
for comprehension and those for both comprehension and production). Another
typically cited explanation for such a discrepancy in the results are various, not
infrequently, incompatible methodologies. Nonetheless, whatever the answer to
the question about the amount of mental word-stores, the actual number seems to
exert little impact on the way the lexicon functions.
WWW.EBOOK777.COM
2.3.2 The Internal Organization of the Lexicon
Turning to the internal organization of the lexicon, the number of components of

the human word store is a complex issue which is far from being settled. There are
many models and the number of components of the mental lexicon they distinguish
markedly vary. Some scholars (cf. Carroll 1994) apply the term mental lexicon to
mean only the semantic sub-lexicon. Others (cf. Garman 1990) distinguish between
the semantic lexicon and the phonological one. Alternatively, there are models which
disregard the word’s orthographic representation and instead concentrate on two lev-
els called semantic and phonological sub-lexicons (cf. Levelt 1989; Aitchison 2003a,
2012). On the other hand, many psycholinguists perceive the orthographic repre-
sentation an inseparable component of a lexical item. Consequently, in their mod-
els of the mental word-store they describe two modality-specific phonological and
orthographic components within the formal layer of the lexicon (cf. Emmorey and
Fromkin 1988; Randall 2007; Fernández and Smith Cairns 2011). The validity of the
latter type of models has been proved by experiments involving priming effects of
different modalities on word production and recognition (cf. Harley 2004).
It is widely agreed that the semantic and formal components of a lexical item are
not stored together. Aitchison (1987, 2003a, 2012), Levelt (1989), Garman (1990),
or more recently Randall (2007) and Fernández and Smith Cairns (2011), all agree
that the semantic aspects of a word are located in one layer and the information on
the formal aspects is kept in a separate part of the word-store. The two levels, how-
ever, are assumed to be connected by a wide net of direct links. A common argu-
ment supporting this view refers to the tip-of-the-tongue phenomenon in which
the meaning of a word and a number of its syntactic properties are available for
the speaker but the word’s form cannot be retrieved (cf. Ecke 2009; Ecke and Hall
2013). By comparison, in his mental lexicon model Levelt (1989) adopted a twofold
lemma vs. lexeme distinction to the entire lexicon, thus creating two separate stores:
a lemma lexicon containing lemmas and a form lexicon comprised of morpho-pho-
nological forms. Clearly, this division has only a metaphorical function which is to
show that the internal organization of the mental lexicon is twofold, according to the
meaning of items, as well as according to their morpho-phonological features.
The still debatable problem of the number of lexicons coincides with the issue
of the modality of input and output. Are there two modality-specific lexicons or do
we use the same store both while reading and listening? Undoubtedly, the advan-
tage of the former assumption is the economy of storage, its drawback being the
expense of retrieval in contrast to the latter proposal characterized by the simpler
retrieval at the expense of complex storage. In short, the model which allowed for
the maximum storage capacity might, at the same time, invalidate the most effi-
cient retrieval. However, as Aitchison observes,
In dealing with words in the mind (…) we must treat storage and retrieval as interlinked
problems (…). Although common sense suggests that the human word-store is primarily
organized to ensure fast and accurate retrieval, we cannot assume that this is inevitable.
Humans might have adopted a compromise solution which is ideal neither for storage nor
for retrieval (Aitchison 2012, p. 10).
WWW.EBOOK777.COM
With regard to organization, then, Fay and Cutler (1977) believe that there is one
single mental lexicon for both production and comprehension instead of two sepa-
rate lexicons. This assumption has been based on the analysis of common speech
errors such as malapropisms or slips of the tongue. By contrast, Garman’s model
(1990) accounts for the existence of two separate specialized stores: one for gen-
erating and one for identifying words. Here the evidence supporting this view
comes, above all, from neuropsychological research which has proved a number of
discrepancies between comprehension of spoken and written input and production
of spoken and written output. According to Ellis and Young’s model (1988, 1996),
on the other hand, there is one semantic lexicon incorporating four-modality spe-
cific interconnected sub-lexicons.
2.3.3 The Internal Relations Within the Lexicon
The structure of the lexicon is not the only debatable issue concerning the human
word-store. Equally controversial is the matter of the relations within the mental
lexicon. A highly advanced classification of various internal connections occurring
in the mental lexicon was proposed by Levelt (1989), who distinguished between
intrinsic and associative links. Intrinsic relations occur when items are linked
through at least one component of the fourfold information on a word—meaning,
morphology, syntactic category or phonology. Associative relations, on the other
hand, hold between words which show no direct semantic, phonological or mor-
phological links, but which frequently co-occur in speech or writing.
Lexical items can be intrinsically related through their meaning. A word is
linked with its hyperonym (banana—fruit), co-hyponyms (banana—apple), near-
synonyms (wide—broad), antonyms (wide—narrow) etc. All these interrelated
links form a network called a semantic field. Another form of intrinsic links are
morphologically-determined relations between derivatives of one item, which
simultaneously share some semantic features (e.g., govern, government, govern-
mental, governor). Evidence supporting the existence of such types of relations
between individual lexical entries, again, comes from the analysis of speech
errors. Fay and Cutler (1977) and much later Fikkert (2005, 2007) point to yet
another type of intrinsic relation—the one based on phonological features which
may be responsible for substitution errors such as the already discussed mala-
propisms. The authors claim that “words with the same initial or final segments
seem to be connected as they cause errors in speech production such as week for
work” (Fay and Cutler 1977, p. 514). Finally, there is some evidence on syntacti-
cally conditioned connections between entries coming from research with aphasic
patients who have lost access to the entire class of words (cf. Haverkort 2005).
The second type of connections between entries in the mental lexicon are the
associative relations. This kind of link occurs between entries which do not share
any semantic, phonological or morphological features but which tend to co-occur
in language use. The existence of associative relations has been evidenced in a
WWW.EBOOK777.COM
variety of experiments using different methodologies, the most common of them

being priming tests (cf. Carr and Dagenbach 1990; Kroll and Sunderman 2003;
Dijkstra 2005; Dijkstra et al. 2010). It is argued that if a word is found to prime
another, then the words could be closely connected in the mind (cf. Aitchison
2012). Another group of experiments employed to support the existence of associ-
ative links in the mental lexicon encompasses association tests.4
2.3.4 Lexical Storage: The Full Listing Hypothesis Versus

the Decompositional Hypothesis
One of the most hotly disputed controversies connected with the mental lexicon
seems to be the issue of whether words are stored as whole units or as roots plus
affixes. The following paragraphs address two fundamental questions concerning
the lexical storage of polymorphemic words. Prior to the presentation of two influ-
ential hypotheses seeking to explain the storage of morphologically complex words,
a question addressing the problem of what precisely is stored will be tackled.
The issue of lexical storage is strongly related to the phenomenon of word
primitives which are commonly defined as the smallest meaningful elements
stored in the mental lexicon. For many decades linguists have tried to determine
how words consisting of more than one morpheme (e.g., government) are stored
within the lexicon. Are they stored as independent units, or, as many linguists have
suggested, are complex words decomposed into their constituent elements (e.g.,
govern and ment), which would support the morphemic organization of the lexi-
con? Depending on the perception of word primitives, linguists advocate in favour
of one of the following theories.
The Full Listing Hypothesis was first proposed by Butterworth in 1983 and
since then it has gained a number of supporters (cf. Henderson et al. 1994). In
the light of this theory, derivations are stored, similarly to a written dictionary, as
separate, independent entries (e.g., go and goer are stored as independent units).
Consequently, both for comprehension and production they are accessed sepa-
rately. In the light of more recent studies (cf. Vigliocco and Hartsuiker 2005), the
only advantage of this hypothesis seems to be the so-called access efficiency.
The alternative proposition, known also as the Decompositional Hypothesis,
has gained far more advocates (cf. Levelt 1989; Taft 2004; Frost and Ziegler 2007)
and for this reason the idea of the morphemically-governed organization of the
lexicon will be elaborated further in what constitutes the final paragraphs of the
present section.
In the Decompositional Theory, words are seen as bundles of morphemes,
and since morphemes are believed to be the smallest meaningful units of
4 For
a thorough discussion of this methodology see Gabryś-Barker (2005), Fitzpatrick (2007)
and Roux (2013).
WWW.EBOOK777.COM
language, consequently, the smallest element to be stored is no longer a word

but a morpheme. Morphemes are typically ascribed to one of two categories:
free morphemes (functioning as independent words) and bound morphemes
(all sorts of meaningful affixes which do not, however, function independently
and which require the accompaniment of a free morpheme, thus changing its
meaning and generating a new word). In the light of this hypothesis, to pro-
duce a morphologically complex word (also called a polymorphemic word)
separate morphemes need to be accessed and subsequently melded into one
unit (which may be very elaborate at times; e.g., anti-dis-establish-ment-arian-
ism constitutes six morphemes). Similarly, on encountering a polymorphemic
word our brain needs to decompose it into separate morphemes to be accessed
individually.
Critics of the Decompositional Theory (cf. Bozic et al. 2013) point, among
others, to the problem of the lengthening of the recognition time that the hypoth-
esis would need to endorse. Undoubtedly, due to the fact that in the case of com-
plex words many more units would have to be accessed, additional processing
would be inevitable. As a consequence, the amount of time necessary to access
a complex word would be much longer. On the other hand, scholars supporting
the Decompositional Hypothesis (cf. Levelt 1989; Taft 2004; Frost and Ziegler
2007) postulate that its obvious advantage seems to be the economy of storage.
Morphemic organization ensures that there is no redundancy in the representation
of related words created by using either derivational (e.g., trusty, distrust, untrust-
worthy) or inflectional (e.g., jumps, jumped, jumping) morphemes.
It needs to be stressed that there is a fair amount of experimental evidence
supporting the hypothesis under discussion. The literature on the topic abounds
with data coming from priming tasks, lexical decision tasks, spoken error anal-
ysis or experiments with brain-damaged subjects, in particular those suffering
from Broca’s aphasia. For instance, in priming tasks responses to a simple word
(hunt) are speeded by a prior presentation of a related word (hunter), suggesting
that these words share some entries in the mental lexicon (cf. Reichle and Perfetti
2003; Rossell et al. 2001; Dijkstra et al. 2005). Moreover, many experiments (cf.
Garrod 2006) have also confirmed that the priming effect accompanying the mor-
phologically related word pairs is stronger than that for word pairs overlapping in
exclusively orthographic (planet vs. plan) or semantic form (imitate vs. copy).
Another source of evidence in favour of the morphemically-governed organiza-
tion of the lexicon are the lexical decision tasks in which words are mixed with
nonwords (pseudowords). The oft-cited experiments measuring the reaction time
(RT) show that the longer the word (i.e. the more morphemes it has), the longer
the reaction time; in other words, the more time we need to decompose it to under-
stand the meaning of the constituent parts and to evaluate their validity (cf. Taft
1981; Reid and Marslen-Wilson 2003, 2007; Marslen-Wilson and Tyler 2007).
A further example utilizing the RT paradigm to support the Decompositional
Theory comes from the research conducted by MacKay concerning morphological
processing in language production (cf. MacKay 1978). When a group of partici-
pants were asked to derive nouns (such as government, existence, decision etc.) out
WWW.EBOOK777.COM
of the aurally presented verbs (govern, exist, decide), MacKay noticed that RTs
varied significantly depending on the “complexity” of the derivation. And thus,
government was identified as the fastest item (no phonological changes), exist-
ence was slower (resyllabification), while decision turned out to be the slowest
(two phonetic changes). The interpretation put forward by MacKay strongly sup-
ports the Decompositional Theory. He claims that the results confirm the assump-
tion that subjects are able to make those changes when producing morphologically
complex words, which means that such words are not stored simply as independ-
ent units. However, the popular criticism of the research was linked to the very
form of the experiment. Namely, the participants were explicitly instructed to
derive morphologically complex nouns from a presented list of verbs. The task
itself required derivational processing which may not occur normally. Thus, many
linguists find MacKay’s research unreliable.
A similar lexical decision task paradigm was adopted by Taft and Forster
(1975) and Taft (1981), who worked on the “prefix-stripping” strategy in word rec-
ognition. They concluded that in lexical decision tasks words with prefixes had
greater RTs than words without them. Thus, remind (prefix re-) was identified
faster than relish (“pseudoprefix”). The proposed interpretation of the obtained
results is as follows:
Morphological processor automatically “strips off” anything that looks like a prefix (e.g.,
“RE”), then searches for the base in the lexicon. With words like REMIND, it will find
MIND (real word), but with words like RELISH, will not find *LISH and will have to
restart the search for the whole string (Taft and Forster 1975, pp. 642–643).
It needs to be stressed, however, that the presented results and Taft and Forster’s
interpretation were also rejected by many who, as in the previous case, criti-
cized the methodology of the conducted experiment. The opponents claimed that
although the participants were not instructed to strip off prefixes, “maybe they were
implicitly told this by the kind of word list they got” (cf. Rubin et al. 1979, p. 760).
Aitchison, who is also an ardent advocate of the decompositional approach, to
support her theory uses error analysis of a spoken discourse—he example she gives
is She wash upped the dishes instead of She washed up the dishes (Aitchison 2003a,
p. 65). In her interpretation, the error may be suggestive of the organization of the
internal lexical storage. In other words, the error has been committed since the brain
has accessed the preposition up instead of the verb wash. The researcher believes that
such errors verify the Decompositional Hypothesis and prove the fact that words are
stored as morphemes. To generate this sentence our brain needs to access the verb
wash, the preposition up and the past tense morpheme -ed. If derivations were stored
as in the Full Listing Hypothesis, such an error would not occur—our brain would
store the word washed as a separate item. We would not have to go into the process
of building a word; instead, the already prepared items would wait to be accessed.
Yet another source of evidence supporting the hypothesis under discussion are the
results of experiments with patients suffering from Broca’s aphasia (cf. Tyler et al.
1995). On the basis of these and other findings, models of word recognition were
created which typically include a processing stage in which complex words are split
into their constituent morphemes before meaning-based representations are accessed.
WWW.EBOOK777.COM
It has been generally agreed upon that some morphologically complex

words share their lexical entries with the related forms. Nevertheless, the ques-
tion of precisely which complex words are stored as morphemic units remains
unanswered. It is worth mentioning that many researchers (e.g., Levelt 1989)
emphasize the difference between lexical entries and lexical items. Levelt pos-
tulates that not all lexical items constitute separate lexical entries. And thus,
inflections are items belonging to one single entry (e.g., going, goes, gone are
all to be included under go). Derivations, on the other hand, are to be treated
as separate entries (e.g., goer). The presented assumption has been confirmed by
some experiments showing that the decomposition into morphemes is typically
stronger for words composed of inflectional suffixes than for those formed with
derivational endings (cf. Stanners et al. 1979; Chialant and Caramazza 1995;
Blevins 2004).
All in all, hypotheses of morpheme processing are classified in relevance to
the type of explanation they offer for the identification of polymorphemic words.
Proponents of the Decompositional Theory claim that the meaning of a com-
plex word is composed of its constituent morphemes (cf. Taft and Forster 1975;
MacKay 1978). From this perspective, the meaning of schoolbooks would be
created by first identifying the word’s components (e.g., school+book+s) and
then accumulating its meaning from these components. Conversely, support-
ers of the Full Listing Theory argue that complex words are stored and repre-
sented as independent units (cf. Rubin et al. 1979; Butterworth 1983; Henderson
et al. 1994). In the light of the latter theory, the word schoolbooks is stored as
a single entity, with individual representations for their components: school and
books. Moreover, even the singular form of the word, schoolbook has its separate
representation.
To conclude, the understanding of how an adult native speaker/hearer pro-
cesses inflected word forms has increased considerably over the last decade.
Experimental studies using a range of different psycholinguistic methods and
techniques, e.g., lexical decision or priming, have led to a number of consistent
and replicable results, e.g., frequency effects for inflected word forms in lexi-
cal decision tasks or priming effects for inflected word forms in different kinds
of priming experiments. To account for the theoretical interpretation of these and
other results on morphological processing in adult native speakers, many hybrid
theories, including both separated and compositional representations, have been
proposed (cf. Marslen-Wilson and Tyler 1980; Caramazza et al. 1988; Taft 1994;
Clahsen et al. 2003, Marslen-Wislon 2007). In these theories, “complex words are
identified via a ‘race’ between compositional and whole-word lookup processes”
(Reichle and Perfetti 2003, p. 227). A good example of the hybrid theories are
dual-mechanism models which hold that morphologically complex word forms
can be processed both associatively, i.e. through stored full-form representations
and by rules that decompose or parse inflected word forms into morphological
constituents (Chialant and Caramazza 1995; Clahsen 2006; Bozic and Marslen-
Wilson 2010; Bozic et al. 2013). In brief, hybrid theories, highlight that some
words are more prone to decomposition than others.
WWW.EBOOK777.COM
2.4 Theories of Semantic Representation
In this section the problem of conceptual representation of meaning will be dis-

cussed in relation to the storage of conceptual features and their retrieval from
memory. The key question here is whether semantic representations of words are
identical with general world knowledge or whether it is possible to draw a line
between word meanings and concepts which represent encyclopedic information.
In brief, this section addresses the problem of the representation of meaning in our
mind.
Conceptual representations are assumed to build an independent network
that is frequently referred to as semantic or conceptual memory (Levelt 1993).
It needs to be noted that semantic memory is not the same as the mental lexicon,
which is often compared to a dictionary. Rather, it is a mental encyclopedia inde-
pendent of the formal linguistic representations of the lexical items (cf. Levelt
1993). Clark and Clark (1978) explain the distinction between the two by saying
that not all concepts stored in semantic memory have names in the mental lexi-
con. A typical way of presenting conceptual representations of lexical items is a
rich network of sense relations. Semantic information is given meaning only by
the way it relates to other information. Putting it bluntly, “words are organized
in an interconnected system linked by logical relationships” (Aitchison 2003a,
p. 103). And thus, a definition of a word (concept) is always created in relation to
other words (concepts).
Initially, it was believed that it is possible to measure the distance between
words in the network, thus defining their mutual relations. Some later studies,
however, postulated that the network is far more complicated and much less
stable than was once assumed. It is now generally agreed upon that concepts
are represented in a network of interconnected nodes and that the distance
between the nodes represents similarity between the items. And thus, a typi-
cal mode of describing conceptual representations is as an associative network.
Originally, associative links among lexical items were believed to be fixed and
stable and to reflect the internal organization of words in semantic memory. The
major research tool seeking to describe this static model were free association
tests (cf. Deese 1962, 1965). In this model the meaning of a word was believed
to be the sum of all its associations. Additionally, the model attempted to clas-
sify various relations among words such as syntagmatic and paradigmatic
relations.
As for the syntagmatic-paradigmatic shift, two prominent scholars dealing with
this intriguing phenomenon were Melčuk and Zholkovsky. They argued that, in
contrast to adults, children have words organized differently in their mind (Melčuk
and Zholkovsky 1988). They found that in word association tasks adults give asso-
ciations within the same category; i.e. the word sun typically evokes words such as
moon or star, whereas children tend to associate words paradigmatically; i.e. the
word sun triggers yellow, hot or shines. Altogether, it was concluded that the rela-
tions change with age from syntagmatic to paradigmatic.
WWW.EBOOK777.COM
2.4 Theories of Semantic Representation 47
Below basic models of semantic representations are presented; namely, the

hierarchical network model, the semantic feature model, and the spreading activa-
tion model.
2.4.1 The Hierarchical Network Model
As has been indicated earlier, the storage of conceptual representations can be

depicted as a system of interconnected elements. Hierarchical network models
posit that a word’s meaning depends on its relation (a network of relations) to
other words and that semantic information is arranged in a network. However, a
new notion introduced here is hierarchy. Collins and Quillian, the major propo-
nents of this model, argue that semantic representations of words belonging to
one category create a hierarchical system (cf. Collins and Quillian 1969, 1970).
And thus, as illustrated in Fig. 2.2, words with more general meanings are placed
higher in a network, whereas more specified words tend to be positioned lower in
the hierarchy; e.g., the word animal is located over fish, which in turn is superordi-
nate to salmon or shark.
Another significant assumption of the presented model is the cognitive econ-
omy according to which semantic information referring to more than one word is
stored at the highest possible node and is accessible to all the subordinate nodes
through the network of internal relations; e.g., the information that A salmon can
swim or that A salmon has fins is stored at the fish node which is superordinate to
the salmon node and is true of all fish. Essentially, word properties are stored at
the most general (i.e. the highest) level possible (cf. Fig. 2.2).
Fig. 2.2 A hierarchical network model of semantic information related to animals (adapted from
Collins and Quillian 1969)
WWW.EBOOK777.COM
Table 2.1 Predictions of the hierarchical network model theory that proved to be wrong

(adaptation based on Harley 2004)
Problem Sample sentences Model predicts Finding
Familiarity effect A. A bear is an animal B faster than A A faster than B
B. A bear is a mammal
Typicality effect C. A robin is a bird C = D C faster than D
D. An ostrich is a bird
Concept property E. An animal breathes E faster than F E = F
associations F. A bird breathes
To check their model Collins and Quillian employed sentence verification tasks
(cf. Collins and Quillian 1969, 1970). They assumed that it takes longer to verify a
sentence containing information from the most remote nodes in the hierarchy, e.g.,
A bear is a mammal, than a sentence using information from closer nodes, such as
A bear is an animal, since the lower levels inherit the information from the higher
levels. This kind of familiarity effect has not, however, been confirmed in empirical
research (cf. Table 2.1 below). The model has also been criticized for the invalid-
ity of accommodating the typicality effect which posits that all the words from the
same level of a given hierarchy, e.g., robin, ostrich, canary etc. are to be considered
equal. Hence the relation between robin and bird and ostrich and bird should be
perceived as equal. Nevertheless, the postulate has not been evidenced in sentence
verification tasks. Conversely, the research carried out proved the predictions of the
hierarchical framework inaccurate. Table 2.1 below compiles the basic problems
the model does not account for. For instance, in contrast to Collins and Quillian’s
assumptions rejecting the familiarity effect, familiar words are indeed recognized
faster than unfamiliar words irrespective of their position in the hierarchy.
2.4.2 The Spreading Activation Model
In view of the criticism validated by the results of the numerous experiments

(cf. Table 2.1) an improved version of the hierarchical network model was pre-
sented and until now seems to be the most satisfying model of the semantic mem-
ory. The basic change concerns the notion of hierarchy. Collins and Loftus, the
major advocates of the spreading activation theory, postulate that the meanings of
words form a network of semantic relations. The network, however, is not hier-
archical any more (cf. Collins and Loftus 1975). No longer are the links within
the network organized along the superordinate and subordinate principles. Instead,
it is argued that the relations between semantic representations are not of equal
importance. In brief, some nodes are more accessible than others and the degree
of accessibility depends on the frequency of use and the word’s typicality (Collins
and Loftus 1975). Additionally, the authors claim that the distance between nodes
is determined by structural characteristics, e.g., taxonomic relations (cf. Rosenman
and Sudweeks 1995) or the already-mentioned typicality (cf. Fig. 2.3).
WWW.EBOOK777.COM
Fig. 2.3 A diagrammatical representation of a possible semantic network for DOCTOR

(adapted from Collins and Loftus 1975)
More importantly, the model also seeks to account for the problem of semantic
economy. Whereas the hierarchical model assumed that the word’s semantic prop-
erties were stored, for the reason of economy, at the highest possible nodes thus
eliminating the redundancy, the revised theory has it that certain features which
are typically associated with a given word are stored with the semantic represen-
tation of that word, against cognitive economy, quite redundantly. Collins and
Loftus’ model also encompasses the typicality effect as developed by the prototype
theory (cf. Sect. 2.4.3). Hence the distance between two nodes is conditioned by
the typicality of these words and not by the hierarchy of organization; e.g., the
connection between bird and penguin is weaker than between bird and pigeon. To
test the efficacy of their model, Collins and Loftus employed the semantic priming
paradigm. The obtained results appeared to support the assumption of the auto-
matic spreading activation mechanism to be found in the processing of semantic
representations.
2.4.3 The Componential Approach
The semantic feature view stands in contrast to the hierarchical network model
(cf. Smith et al. 1974). This kind of approach, also termed the componential
approach, proposes that words can be decomposed into a bundle of primitive
semantic elements. As a result, words similar in meaning share some of their
semantic features known as defining features, but they also incorporate some char-
acteristic features specific only to them. This theory is connected with two influ-
ential categorisation theories to be discussed below.
WWW.EBOOK777.COM
Two contrasting standpoints concerning the phenomenon of the nonverbal,

conceptual representations are: the classical view and the prototype theory derived
from cognitivism. The classical theory is based in ancient Greece and it prevailed
in psychology, philosophy, and linguistics until the 1950s. It is based on objec-
tivism and essentialism which in turn constitute the very core of the Aristotelian
model of categorization. In the light of essentialism,
all reality is made up of objectively existing entities with properties and relations among
them. Some properties are essential and others are not. Classical categorization links cat-
egories to properties. Objectivist cognition assumes that people reason in terms of abstract
symbols and that those symbols get their meaning via a correspondence between those
symbols on the one hand and entities and categories in the world on the other. (Lakoff
1987, p.173)
Aristotle enumerates two aspects of a thing: essence described as “the parts which
are present in such things, limiting them and marking them as individuals, and by
whose destruction the whole is destroyed (…)” (Metaphysics 5.8.3.) and accidents
referred to as “that which attach to something and can be truly asserted, but neither
of necessity nor usually, e.g., if someone in digging a hole for a plant has found
treasure” (Metaphysics 5.30.1.). The above-mentioned aspects can be explained on
the example of the word flower. The essence of a flower is that it is a plant; its col-
our or smell is just accidental and does not influence the judgment of whether the
entity is a flower or not.
However influential the classical theory of categories may be, it has been “reap-
praised in all of the cognitive sciences” (Lakoff 1982, p. 3). In the 1970s a com-
peting theory of natural categorisation was proposed by Eleonor Rosch (1975).5
Since the theory centered on the so-called prototypical members of the group of
possible referents of a given word, it was labeled the prototype theory. It can be
briefly described as “a hypothesis that people understand the meaning of words by
reference to a highly typical example” (Aitchison 2003a, p. 94). In a short time the
theory gained a wide group of supporters, including Bolinger (1977), Lakoff
(1982, 1987), Wierzbicka (1985) or Langacker (1987) and more recently Smith
and Minda (2002) or Taylor (2003). Unlike the purely theoretical argumentation of
the objectivist metaphysics and psychology, the prototype theory is based on
empirical evidence, “experimental results and the interpretation of these results”
(Lakoff 1982, p. 8).
The subsequent paragraphs provide a short presentation of the number of dif-
ferences concerning these two highly influential theories. The first difference to
be discussed is the so-called componential analysis. In the classical view catego-
ries are defined in terms of a conjunction of necessary and sufficient conditions.
Entities can be described in terms of smaller parts—components or features of
binary structure (present [+] or absent [−]). All members of the category have to
share the same necessary and sufficient features (cf. Taylor 1990). It needs to be
noted that categories are homogenous; i.e. that all members have equal status and
5 The theory found its philosophical grounds in the works of Ludwig Wittgenstein (1953).
WWW.EBOOK777.COM
they need to share the same features. Thus, there are no worse or better examples.
In other words, no single cat is more cat-like than others (cf. Lakoff 1987). In the
prototype theory, on the other hand, entities belonging to one category do not have
to, and rarely do, possess the same inventory of features. Wittgenstein (1953) pre-
sents it as follows:
Consider for example the proceedings that we call: ‘games’. I mean board-games, card-
games, ball-games, Olympic games, and so on. What is common to them all? (…) For if
you look at them you will not see something that is common to all, but similarities, rela-
tionships, and a whole series of them at that (Wittgenstein 1953, p. 31).
Thus, it can be concluded that the underlying principle behind the categorisa-
tion is family resemblance (Wittgenstein 1953). However, not all members pos-
sess the same inventory of features still forming one common category. Rosch
and Mervis refer to a category as to “a set of items of the form AB, BC, CD,
DE. That is, each item has at least one, and probably several, elements in com-
mon with one or more other items, but no, or few, elements in common to all
items” (Rosch and Mervis 1975, p. 66). An item classified as belonging to a
given category shares features with a few others, but not necessarily with all,
members of the same category. Categories are not homogenous, which means
that e.g., some birds are more birdy, like the robin, while some are less birdy
like the penguin (cf. Rosch 1975). It could be concluded that the theory
accounts for worse and better members of one category. The most representa-
tive entities for the entire category are called prototypes. Prototypes have a
privileged place in memory as they occupy the central role in the category and,
consequently, are retrieved quicker.
Another discrepancy between the two theories concerns category boundaries.
In the classical view category boundaries are clear-cut and stable. And thus, the
decision whether an entity belongs to the category or not is based on objective
features. Moreover, no factors can influence those categories. As Lakoff points
out, “category boundaries do not vary. Human purposes, features of context, etc.
do not change the category boundaries” (Lakoff 1982, p. 15). Thus, they would
often demand redefinition or the creation of new categories. Internal definition is
the only factor affecting the category, the structure of a category is context inde-
pendent, no subjective factors can affect the category and thus, psychological
factors seem to be unimportant. No matter how humans perceive a given item, it
is categorized regardless of the subjective interpretation. In contrast, the proto-
type theory shows that there are no clear-cut boundaries. Instead, any boundaries
are described as flexible, susceptible to subjective factors such as human pur-
poses. Many experiments proved (cf. Black 1949; cited in Ungerer-Schmidt
1996; Labov 1973) that prototype-based categories merge into each other, and
their boundaries instead of being clear can be described as fuzzy. In his publica-
tion Labov (1973) elaborates on an experiment in which the subjects were asked
to name various containers (e.g., cup, bowl). The results showed that the labels
provided by the participants varied substantially. Furthermore, the same partici-
pants were not consistent in their responses. Labov later concluded that a word
WWW.EBOOK777.COM
has its core meaning which is most central and invariant as well as its peripheral
meanings.6 As a consequence, advocates of this theory emphasize that the mean-
ing of a word should be analyzed on a continuum.
To conclude, the objective of the present section was to discuss the two most
influential theories related to lexical meaning. As has been shown, the theories are
very different. Many scholars emphasize the fact that the prototype theory being
based on empirical evidence seems to be far more convincing than the classical
one based on non-empirical speculations (cf. Lakoff 1987). In the following this
section various models of lexical access will be examined.
2.5 Models of Lexical Access in the Mental Lexicon
Having discussed the issues concerning the structure of mental representation of

words and their meaning in the human mind, the chapter will now proceed to elab-
orate on the selection of the most influential models of lexical access and retrieval.
Obviously, it would be almost impossible, and for the sake of the present work
unnecessary, to discuss and compare all the models of lexical access that have
been proposed. Thus, this section has been limited exclusively to the most influen-
tial language processing models that can be found in psycholinguistics.
Lexical production and recognition are very quick processes. In his research
endeavouring to analyze word recognition patterns Marslen-Wilson (1989) found
that a word is recognized usually about 200 ms after its onset; this means even
before the speaker managed to finish uttering that word. Not only is the mecha-
nism of lexical access rapid, but it is also highly sophisticated and complex. Word
recognition involves receiving a perceptual signal, rendering it into the phonologi-
cal or orthographic representation and then accessing its meaning. The opposite
process of producing a word requires first choosing the meaning for the intended
concept, then recovering its phonological or orthographic representation, and
finally converting it into a series of motor actions.
To date, many methods have been used, many paradigms followed to analyze
lexical access in speech production and comprehension. A typical methodology
adopted to search for the key to the lexical access enigma has been the analysis of
malfunctions (e.g. different types of selection errors, slips of the tongue or the tip-
of-the-tongue phenomenon; see Aitchison 2012 for a detailed discussion). Other
methods used the already-mentioned picture naming, lexical decision tasks and
priming. Yet another source of research data derives from speech pathologies such
as aphasia. Aphasic patients who have lost parts or all of their linguistic abilities
have provided linguists with a substantial amount of data concerning the processes
of lexical access and retrieval (cf. Dell et al. 1997; Biran and Friedmann 2012).
6 Cf.Kellerman`s famous study of core and peripheral meanings of the German verb brechen
(break) (Kellerman 1978).
WWW.EBOOK777.COM
2.5 Models of Lexical Access in the Mental Lexicon 53
However complex and demanding the research on the mental lexicon might be,
psycholinguistic literature abounds with models of lexical access. There are many
properties according to which the models can be grouped. Firstly, some models
focus on word recognition, others on production. There are also models which try
to combine these two processes. Another distinctive property is the type of search
involved in lexical processing: here serial (indirect) or parallel (direct) models are
distinguished. The serial models have it that words are accessed individually, one
by one, at the phonological, orthographic and semantic levels. The parallel mod-
els, on the other hand, postulate that words are searched simultaneously. A further
property is interactivity—the question of whether lexical information can travel
backwards and forwards between different levels of lexical representation and
affect their retrieval.
What differentiates the models to be examined in the paragraphs to follow is
the sequence of interaction; a property which divides the models into direct and
indirect ones (cf. Garman 1990, p. 260). The first category of models distinguished
by Garman are indirect access models, which depict lexical processing as “look-
ing up a word in a dictionary” or “finding a word in a library” (Singleton 2000,
p. 170). The indirect access models, also known as multi-step models, are predi-
cated on two-stage access “via a search procedure and then a retrieval procedure”
(Singleton 1999, p. 84). A much-debated representative of the indirect kind of
model to be described below is Forster’s serial search model functioning within a
modular system paradigm.
When referring to direct access models Garman uses the metaphor of a “word-
processing package which allows items stored by name to be accessed simply by
the typing in of as many letters as are sufficient to distinguish the relevant name
from all other stored names” (Singleton 2000, p. 170). In other words, direct mod-
els view lexical processing as a one-stage phenomenon. Two oft-cited representa-
tives of the direct type of model to be discussed below are the logogen model and
the cohort model.
2.5.1 The Serial Search Model
The best known and one of the most influential indirect models is Forster’s auton-
omous search model (Forster 1976; Murray and Forster 2004). The processes
of access and retrieval described in this model resemble looking up a word in a
written dictionary or looking for a book in a library, the only difference being the
organizing principle, which in the case of a written dictionary is alphabetically-
governed, while in the mental lexicon it is claimed to be frequency-dependent.
This is how Garman summarizes Forster’s suggestion:
We enter, looking for a particular book; we do not go straight to the main shelves where
the books are located, since there are simply too many of them to permit efficient search
of them in this direct fashion. So we go instead to the catalogue. Searching through the
catalogue, we find something that matches what we are looking for; but we retrieve from
WWW.EBOOK777.COM
Fig. 2.4 Forster’s serial search model of lexical access (adaptation based on Foster 1976)
this stage of the process, not the book itself, but an abstract location marker, telling us
where to find the book on the shelves. Armed with this, we implement the second stage
of the process, by using the marker to guide us to the right book on the shelves (Garman
1990, pp. 266–267).
A similar metaphor has been adopted by Singleton, who compares the first stage to
finding the right page in a dictionary, the major difference between these two pro-
cesses being the aforementioned principle behind the organization of the entries.
The moment the “page” has been found (on the basis of the initial properties of
the signal), the search goes on through the “page” governed by word frequency.
Once the abstract location marker has been found, the second stage starts. Forster
assumes that lexical entries are searched sequentially until the appropriate word
is selected and believes that the mental lexicon consists of two levels: one con-
taining access files and the other the master file (the lexicon proper; cf. Forster
1976; Fig. 2.4). There are two stages of word processing. In the first stage, follow-
ing perceptual processing, the serial search of access files starts. The only infor-
mation on a word available in the access files is its address in the master file. To
put it differentially, access files comprise the stimulus features of a word, i.e. its
access code and the pointer to the matching entry in the master file. The master
file includes all the information on a given word—phonological, orthographic,
morphological, semantic and syntactic data. It needs to be noted that not only is
the master file a complete representation of each individual lexical item, but it also
includes cross-references between all the items stored in the master file, thus pro-
viding for the semantic priming effect. In order to cater for different modalities,
through which lexical items are perceived and generated, and two directions of
lexical access Forster proposed three separate access files which organize words
either by orthographic, phonological or syntactico-semantic properties and are
linked with the master file by pointers. These discrete operating subsystems pro-
cess lexical information independently of one another. A schematic visualization
of the model has been presented in Fig. 2.4.
WWW.EBOOK777.COM
Garman (1990) notes that Forster’s search model incorporates two key features
a good model of the mental lexicon should have. It is characterized by the diver-
sity of access and, at the same time, the unity of storage. Whatever the channel of
communication, it is always the same entry in the master file. Access to each word
depending on the channel leads always through a suitable access file:
If one is listening to speech, one processes each spoken word by going first to the pho-
nological access file; if one is reading written language one goes first to the orthographic
access file; and if one is producing language on the basis of particular meaning intentions,
one goes first to the syntactic/semantic access file. The access file in question then facili-
tates access to the master files (Singleton 2000, p. 174).
For convenience, the access files are divided into separate bins based on the initial
sound or letter. The words in a bin are arranged in a descending order of frequency
so that more frequent words can be searched faster and matched with the acoustic
string prior to low frequency words. In this way, Forster eventually managed to
accommodate the frequency effect within his model.7 The effect has been con-
firmed by a substantial body of evidence from lexical decision tasks where high
frequency words were identified faster than low frequency items. The model also
includes the lexicality effect assuming that there occurs exhaustive search for non-
words and terminating one for actual words. Before rejecting a nonword the brain
needs to search through the entire master file only to find an empty entry. This has
been confirmed by numerous lexical decision tasks which show that it takes about
150 ms longer to identify nonwords than actual words.8
Not only does the model account for the frequency effect and the lexicality effect,
but it is capable of accommodating the priming effect as well. Forster’s model is
not interactive in that it does not allow for the cross-referencing of access files and
the master file. Words in the master file are accessed only through one file at a time.
However, once an entry in the master file has been accessed, cross-references are
observed. Thus, the model is able to accommodate the effect of semantic priming.
If an individual sees the target word for doctor and then subsequently is shown the
word nurse, the response time for the latter word is rightly expected to decrease.
However detailed the model seems to be, there is still a substantial number of
controversies it has been unable to end. Firstly, the model faces the problem
of capacity limitations. The evidence from lexical decision tasks supports the
idea of empty entries for nonwords, which, if really there, would occupy a lot
of space redundantly. Secondly, speech seems to be much too rapid to accept the
idea that words are searched sequentially; the model allows for only one entry to
be searched and matched with the input at a time. Another repeatedly criticized
issue is the fact that the model does not allow for the influence of context on the
process of recognition. It also does not give an account of form-based priming
7 The original version of the model which presented it as a direct access model (cf. decision
trees; Forster 1976, p. 258) failed to incorporate the results supporting both the lexicality and the
frequency effect and proved quite to the contrary.
8 Again this effect was impossible to implement within the original version of the model.
WWW.EBOOK777.COM
effect and it cannot explain the role of similarity neighbourhood. Finally, the
model does not explain the influence of context on speech production (a phenom-
enon confirmed by the repetition priming effect). Due to the above limitations
the early version of the model underwent extensive changes (cf. Forster 1989).
For instance, in the revised version, Forster introduced a discrete comparator
for each lexical entry, thus solving the problem of limited capacity (Murray and
Foster 2004). He also proposed various models of activity among lexical entries.
However, it seems that the presented changes have actually transformed the
model in the direction of spreading activation models.
2.5.2 The Logogen Model
The logogen model, in contrast to its serial search equivalent, assumes one-stage
parallel processing. Initially, the model was proposed by the British psychologist
John Morton in 1969 to account for visual word recognition. Only later was it
developed and revised to incorporate both written word recognition and word
selection in speech production. The model comprises three elements: the logogen
system, the cognitive system and the response buffer. However, its key feature is
the logogen system which is defined to be a part of the nervous system responsible
for lexical processing (in the initial version of the model it was described as a neu-
ral unit, to be later changed to the more technical term logogen9). According to
Singleton, the logogen system is “a set of mechanisms (…) specialized for collect-
ing perceptual information and semantic evidence concerning the presence of
words to which the logogens correspond” (Singleton 1999, p. 86), whereas
Coltheart et al. (2001) describe it as an “information-gathering” device.
Initially, Morton postulated that there was a unitary logogen system, but due
to some empirical evidence he revised the idea and divided the system into three
parts. He distinguished two specialized input logogen systems: visual and audi-
tory and the output logogen system (cf. Morton and Paterson 1980). It has to be
stressed that none of these units includes semantic information about words. This
information is stored in the cognitive system, which includes “a collection of
semantic information of various kinds” (Singleton 1999, p. 86). The system may,
but does not have to, be incorporated in the lexicon itself. What merits special con-
sideration, however, is that meaning is not stored as a single unit for each word;
instead, it is computed when required.
In the logogen system every single item is represented by the corresponding
logogen which comprises the word’s features (phonological and orthographic
characteristics). The moment the acoustic or visual input reaches the logogen, it
is changed into appropriate phonological or orthographic representation which
9 logogen, from Greek logos, “word” and Greek and Latin gen—“birth”; “to bring to life” (after
Singleton 2000, p. 171).
WWW.EBOOK777.COM
launches the process of word finding. The next step is passing the information
to the cognitive system which specifies its semantic and conceptual aspects, and
finally to the logogen output system. It needs to be stressed that the links between
the elements of the model are bidirectional.
It has to be noted that one of the key features of the model is the so-called
threshold level. Each logogen has a “resting” threshold level. Once sufficient evi-
dence has been introduced into the system, the threshold level is reached and the
word is activated. This means that, e.g., in the case of a written word, even before
all the letters are identified, the word can already be recognized and its code sent
to the cognitive system. Clearly, threshold levels differ in value depending on the
words’ frequency. Thus a high frequency word has a much lower threshold than
a low frequency word and less activation will be needed to “fire” and thus access
this word (cf. Harley 2008). In this way the model accounts for the frequency
effects of words in a different way, by using the activation and raising of potentials
within different words.
Summing up, in this model word recognition is seen as a process of accumulat-
ing sufficient information to ultimately access a given word. Once enough infor-
mation has been gathered, the logogen’s threshold is exceeded, the code is passed
to the cognitive system and to the suitable output logogen. The key features of the
model are: directionality of access—every word has its logogen, interactivity—it
allows for the interaction of semantic and perceptual aspects, and parallel process-
ing—incoming information is checked against logogens when it reaches threshold.
The model is versatile in that it accounts for both visual and auditory processing.
Nevertheless, it is perceived as very complex and difficult to test experimentally. It
also leaves many empirical findings unaccounted for, one of them being the effect
of neighbourhood size.
2.5.3 The Cohort Model
The cohort model is another variant of the direct access model. It was first pro-
posed by the British psycholinguist William Marslen-Wilson (1973, 1975) and
later revised many times to incorporate new findings of psycholinguistic research
(Marslen-Wilson and Welsh 1978; Marslen-Wilson and Warren 1994; Marslen-
Wilson 2007). The model focuses on explaining the process of spoken word recog-
nition and does not account for other aspects of lexical access, namely visual word
recognition and word production. The model is based on the assumption that
words are recognized by their onsets in the left-to-right fashion of processing.
Once the initial segments of the word are uttered and received, all potential lexical
candidates commencing with that very sound(s) are activated and form an initial
cohort. This assumption is supported by the tip-of-the-tongue phenomenon (TOT),
according to which lexical access is possible when the beginning sounds of a word
become accessible (Biedermann et al. 2008). Spoken word recognition is assumed
to constitute three stages: access, selection and integration. In the access stage the
WWW.EBOOK777.COM
perceptual representation of the word triggers the activation of a set or, as

Marslen-Wilson (1992) suggests, a cohort of lexical items which share the same
acoustic features. On the basis of empirical findings10 Marslen-Wilson (1992) pos-
tulates that a cohort is activated even before the word has been pronounced to the
end. In fact, the very first sounds trigger the activation of a group of words begin-
ning with that particular sequence of phonemes referred to as the word-initial
cohort. As the subsequent sounds are pronounced, more information is given and
the cohort is narrowed down up to the word’s uniqueness point; the point at which
only one word candidate is left in the cohort. What proves the existence of the
uniqueness point is the finding that a word that has not been fully pronounced can
still be guessed. Moreover, the model also defines the point at which nonwords are
recognized; namely the point at which the sequence of pronounced sounds fails to
correspond to any word of the language. For instance, the nonword recognition
point for a potentially English word daffodip would be the very last sound /p/ since
“only this final sound rejects the possibility of a match” (Singleton 2000, p. 173).
In its early version the model was presented as fully interactive. Marslen-
Wilson postulated that a word can be recognized and selected even before it
reaches its uniqueness point due to contextual information. It was also believed
that a word can be eliminated from the cohort due to context. Indeed, many find-
ings obtained in the course of psycholinguistic research supported the assumption
that contextual information has a facillitatory impact on lexical processing. Much
evidence in favour of this proposal came from speech shadowing experiments (cf.
Marslen-Wilson and Welsh 1978) where the subjects were asked to retell a story
they heard. What merits special attention is the fact that words which in the story
were incorrect (e.g., mispronounced or misused) were successfully repaired in the
process of retelling the story. Moreover, retelling was not accompanied by any
hesitation pauses. Marslen-Wilson used this empirical data to support the assump-
tion of the importance of contextual information. He claimed that fluent restora-
tion proved the influence of context since the mispronounced words could have
only been corrected on contextual grounds. Another pool of evidence in favour of
the role of context came from word-monitoring (Marslen-Wilson and Welsh 1978)
and rhyme-monitoring studies (Marslen-Wilson and Tyler 1980).
Understandably, as every controversial assumption, the issue of context effects
had many opponents. Many critics emphasized the fact that context cannot cause
the elimination of words from a cohort. Ultimately, the mounting criticism and the
growing evidence against the validity of the context effect forced Marslen-Wilson
to reject the dominant role of context in auditory word recognition. As he pointed
out himself, the problem with the context-driven pre-selection lies mainly in the
fact that it does not account for the open-endedness and unpredictability of lan-
guage use (cf. Marslen-Wilson and Tyler 1980). It needs to be added, however,
that the more recent versions of the model assume that although “(…) contextual
10 To recognize a monosyllabic word it takes about 300 ms from the word onset and about
100 ms before its coda (cf. Marslen-Wilson and Welsh 1978).
WWW.EBOOK777.COM
information has no impact on the selection of the word-initial cohort, (…) once the
cohort has been established, word candidates which are inconsistent with the con-
text can begin to be deactivated” (Singleton 1999, p. 94).
What is interesting to note is that whereas the previously described logo-
gen model allowed for various levels of activation, the early version of the cohort
model, on the contrary, postulated the existence of the binary membership. It
asserted that an item is either active (on) or not (off). The first alternative refers
to the situation when a word still belongs to a cohort of word-candidates, the lat-
ter describing the situation when the word has already been eliminated from the
cohort. Ultimately, later versions of the model rejected the binary membership
and postulated the gradual membership. It was suggested that words which do not
receive further verification from the incoming acoustic representation have their
level of activation gradually lowered; they cannot, however, be eliminated from
the cohort. Conversely, they can be activated again if an appropriate signal occurs.
Consequently, in the more recent versions of the model (1990, 1993) the importance
of input goes beyond the uniqueness point and the deactivation of word candidates
is reversible; the alterations which have made the model maximally efficient.
Finally, it needs to be stressed that however controversial the model might be,
there is still a fair amount of experimental evidence to be found in favour of its
main assumptions. Singleton (2000), for instance, refers to the recognition of non-
words. The recognition of nonwords is shorter, he says, in those cases where rec-
ognition points come early in words. As Singleton reports, “the more contextually
predictable a word is, the shorter the sequence of sounds required to reduce the
cohort to a sole candidate” (Singleton 1999, p. 95). The recognition time is much
longer when recognition points appear later within a word. On the other hand, the
model is still criticized, the basic criticism being connected with the fact that the
model accounts for only one type of modality and fails to explain effects of fre-
quency or neighbourhood density. Additionally, some researchers hold it as highly
unlikely that we recognize words on the basis of “the noisy and ambiguous acoustic
signal which is speech” (Marcus and Frauenfelder 1985, p. 164; in Garman 1990).
2.5.4 Computational Models
The traditional “box-and-arrow” type models discussed so far were determined

by “the high-level theoretical principles themselves” (Norris 2013, p. 518) but
were unable to explain what processes were going on in the boxes. The situation
changed with the development of computational models of reading in the early
1980s. Recent computational models are able to handle realistic lexicons and sim-
ulate data from a range of different tasks (e.g. masked priming, lexical decision or
eye-movement control). Moreover, current models of word recognition can now
perform large-scale simulations using many thousands of words. Finally, they can
successfully simulate an interaction between the theoretical predictions and the
contents of the lexicon. They make clear assumptions about what is supposed to
WWW.EBOOK777.COM
Table 2.2 Major computational models of visual word recognition (adapted from Norris013)

Computational Author(s) The main phenomena the
models of visual word model simulates and the tasks
recognition used in simulation
The interactive McClelland and Rumelhart (1981) Word-superiority effect/percep-
activation Rumelhart and McClelland (1982) tual identification task
model (IA)
The spatial coding Davis (2010) Letter order/
model (SCM) lexical decision task; masked
priming task
The dual-route Coltheart et al. (2001) Reading aloud/lexical decision
cascaded model (DRC) task
The letters in time Adelman (2011) Letter order/
and retinotopic masked priming task; percep-
space (LTRS) tual identification task
The Bayesian Norris (2006) Word frequency, letter order,
reader (BR) Norris (2009) RT distribution/
Norris and Kinoshita (2012) lexical decision task; masked
priming task
Diffusion model Ratcliff (1978) Word frequency,
Ratcliff et al. (2004) letter order/
Gomez et al. (2013) lexical decision task
SERIOL Whitney (2008) Letter order/
Whitney (2011) lexical decision task; masked
Whitney and Cornelissen (2008) priming task
be going on in the boxes and later successfully work out differential predictions of
the models (cf. Norris 2005). For the reasons indicated above, there is a common
agreement among psycholinguists that computational models are to be preferred
over traditional “box-and-arrow” models. However, some obvious limitations all
the current models share, the main being their focus on a single domain of behav-
iour, should not be left unnoticed. Indeed, Norris raises a valid point when he
remarks that there is still a need for more integrated theories of word recognition
(2013, p. 523).
Table 2.2 presents a selection of the most influential computational models of
visual word recognition and points to the basic phenomena the models have been
developed to explicate. As for the modelling style of framework within which the
models have been created, the most influential style of computational models are
connectionist models, the earliest example of which is the Interactive Activation
model (IA),11 first proposed by McClelland and Rumelhart in 1981. This type of
modelling has for many years been favoured among researchers mainly due to the
fact that it is relatively “brain like” (Clark 1993) and relatively easy to understand.
An alternative style of modelling—mathematical or computational one—exploits
computational procedures or mathematical formulae. It needs to be noted,
11 McClelland and Rumelhart’s model will be discussed in details in Sect. 2.6.2 below.
WWW.EBOOK777.COM
however, that the interactive activation model (McClelland and Rumelhart 1981),
the Spatial Coding Model (Davis 2010) or the dual-route cascaded model
(Coltheart et al. 2001) which are typically visualized as connectionist models can
also be expressed mathematically (cf. Norris 2013).
2.6 Views on Language Processing
This section provides a contrastive description of the two hypotheses on the lin-
guistic storage (the modularity hypothesis and connectionism), with particular
emphasis on their strengths and weakness.
The modular view has it that the mind is “divided into separate compartments,
separate modules, each responsible for some aspect of mental life” (Cook and
Newson 1996, p. 31). The modularists claim that linguistic meaning is clearly
separated from other varieties of meaning and is represented and processed within
the language module (cf. Emmorey and Fromkin 1988). The proposed processing
is sequential (i.e. one thing at a time—an assumption which makes the process-
ing slow), symbolic (i.e. one token equals one concept) and procedural (linguistic
behaviour is governed by rules). The basic problem with this theory, however, is
its inflexibility.
Cognitive theories, which are commonly taken to be antipathetic to the modular
view, adopt the analogy of brain-style neuronal interactions and depict the mind
as a single system—an interactive network. They describe linguistic processing in
terms of connection strength rather than rules or patterns. It needs to be noted,
however, that in the recent decades the most current models have sought to com-
bine both the modular computational and the connectionist theory (cf. Dell 1988).
2.6.1 The Modularity Theory
Customarily, the origins of the modularity theory can be traced as early as in the
18th century, when a German anatomist Franz Josef Gall “developed the view that
each intellectual and behavioural attribute was controlled by a specific location
in the human brain” (Singleton 1999, p. 111). The current version of the hypoth-
esis became one of the most influential cognitive perspectives of the late 1960s.
The major proponents of this modular view of the mind are theoretical linguist
Noam Chomsky (1988) and psycholinguist Jerry Fodor (1983, 1989). Whereas
Chomsky’s interest in modularity is related exclusively to language acquisition
processes, Fodor’s work concentrates on aspects that are processing-oriented.
Since the focus of the present chapter is centered around issues relating to lan-
guage processing, only the Fodorian perspective is discussed.
The modularity hypothesis, according to Fodor (1983), postulates that “the
entire language faculty is a fully autonomous module [comprising] a number of
WWW.EBOOK777.COM
distinct, specialized, structurally idiosyncratic modules that communicate with

other cognitive structures in only a very limited way”12 (Singleton 2000, p. 176).
In the light of the Fodorian theory, modules are independently functioning cog-
nitive systems located within the language system. They can be defined in terms
of nine characteristic features. Five of the features refer to the way in which
modules process information, and as Fodor himself points out (Fodor 1989), are
also characteristic of acquired skills. These include: informational encapsulation
(i.e. the notion that it is impossible to interfere with the inner workings of a
module), unconsciousness (i.e. the assumption that it is difficult or impossible to
think about or reflect upon the operations of a module), speed (i.e. the idea that
modules are very fast), shallow outputs (i.e. the view that modules provide lim-
ited output, without information about the intervening steps that led to that out-
put), and obligatory firing (i.e. the claim that modules operate reflexively,
providing pre-determined outputs for pre-determined inputs regardless of the
context).13 Another three features, namely ontogenetic universals (i.e. the postu-
late that modules develop in a characteristic sequence), localization (i.e. the idea
that modules are mediated by dedicated neural systems), and pathological uni-
versals (i.e. the suggestion that modules break down in a characteristic fashion
following some damage to the system), characterize the biological status of
modules and play a crucial role in differentiating the behavioural systems from
learned habits.14 The final and concurrently the most controversial feature is
domain specificity, i.e. the assumption that modules deal exclusively with a sin-
gle information type.
It is beyond the scope of the present chapter to elaborate on all the aforemen-
tioned modular aspects, thus it will suffice to present Fodor’s two major and most
controversial postulates describing the processing of linguistic information as
domain-specific and informationally encapsulated.
Domain specificity asserts that each module is capable of processing only
certain linguistic information. Fodor emphasizes the fact that this feature of the
language module was confirmed in a number of experiments where both lin-
guistic and non-linguistic context of one and the same signal influenced the
way it was perceived by the subjects (cf. Liberman et al. 1967). Informational
encapsulation means that intramodular processing is unrelated to other oper-
ating systems and nonlinguistic cognitive processes and that modules do not
make use of other information available in the cognitive system as a whole.
In other words, Fodor postulates that language module is immune to non-
linguistic operations carried out outside the module such as general knowledge
12 This opinion does not equal with the claim that the language module has absolutely no con-
nection with other cognitive operations (cf. Aitchison 2003a).
13 The detailed description of the features has been based on Fodor (1983, 1985).
14 It is assumed that learned systems do not display these particular regularities (cf. Singleton
2000).
WWW.EBOOK777.COM
2.6 Views on Language Processing 63
or the influence of context (cf. Singleton 1999). He sees language processing

as a system limited to “a formal processor with no semantic role” (Fodor 1983,
p. 178). He also clearly distinguishes the linguistic processing from processing of
non-linguistic data.
The postulate that the language module is informationally encapsulated, and
thus context-independent, constitutes one of the most controversial and widely
disputed aspects of the Fodorian theory. Essentially, in the light of the abundant
evidence coming from psychological and psycholinguistic research this postu-
late is difficult to accept. There is a substantial body of research confirming the
facilitative role of general knowledge and context in language task perfor-
mance. Singleton (2000, p. 177) argues that cases have been reported when
multilingual speakers fail to understand or even recognize language which they
are fluent in if they do not expect to be exposed to that language. Another
source of counterevidence highlighting the importance of context in speech pro-
duction and comprehension derives from experiments carried out with hypno-
tized subjects who are able to interact. Furthermore, a pool of arguments
against encapsulation derives from empirical findings of experiments that
involve reduced-redundancy procedures such as cloze tests. In this type of lexi-
cal tasks participants are to fill in missing words removed from a cohesive text.
In order to do this they need to read the whole text. Results show that the more
predictable the target elements, due to some contextual clues, the more success-
ful the performance of the participants endeavouring to guess the missing words
(cf. Weir 1988). These findings lend support for the proposal that participants
actually utilize all aspects of contextual information (e.g., semantic or syntactic
clues) at the same time. Singleton claims that these results account for evidence
for effects of cognitive penetration (Singleton 1999, pp. 115–116) during pro-
cessing. Fodor, however, strongly refutes such an interpretation and suggests
that what might appear as contextual effect, might also be viewed as a matter of
interlexical excitation15 (Fodor 1983, p. 80). He presents his claim in the
following manner:
We can think of accessing the item in the lexicon as (…) exciting the corresponding node;
and we can assume that one of the consequences of accessing a node is that excitation
spreads along the pathways that lead from it. Assume, finally, that when excitation spreads
through a portion of the lexical network, response thresholds for the excited nodes are
correspondingly lowered (Fodor 1983, p. 80).
To conclude, from the presented evidence it can be inferred that in the Fodorian
model lexical knowledge is represented in the network of interconnected nodes.
It is perceived as a central part of a larger system which operates independently
from other systems. The underlying assumption posits that the mind is modular
and comprises special-purpose perceptual processors called modules.
15 It seems justified to emphasize that Fodor’s description of interlexical excitation shows direct
similarities with some claims of spreading activation theory (cf. Sect. 2.3.2).
WWW.EBOOK777.COM
2.6.2 Connectionism
The connectionist theory dates back to the works of McCulloch and Pitts, who in
the 1940s presented the first mathematical model describing the functioning of a
neuron (McCulloch and Pitts 1943; after Singleton 2000). However, the first influ-
ential models of the lexical processing within the connectionist paradigm were
proposed only after a long period of silence in the 1970s and 1980s. To explain
lexical processing, connectionism adopts the “brain metaphor” (Rumelhart and
McClelland 1986, p. 75), which is based on neurophysiological activity in the
brain. And thus, the basic feature of all connectionist models, known also as inter-
active network models, is an analogy between the human brain and the connection
of neurons. They all depict the mental lexicon as a network of nodes which have
various degrees of activation and perceive lexical processing as activation spread-
ing along a network of interconnected units. In fact, one of the major interests of
connectionism is to compute the algorithm that reflects how activation spreads
around the network and triggers individual nodes.
It is worth noting that the connectionist approach to lexical processing belongs
in a much broader parallel processing perspective which stands in stark contrast to
the aforementioned modular theory deriving from the serial processing tradition
(cf. Sect. 2.4.1). Firstly, whereas the parallel processing perspective advocates the
independence of processing operations (i.e. that many activated items can be han-
dled simultaneously), the serial perspective describes lexical processing as organ-
ized in stages (i.e. that activated items can be handled in the one-at-a-time order).
Secondly, the connectionist models call into question the Chomskyan/Fodorian
perception of language and the mind by rejecting the so-called symbolic para-
digm which posits that “mental operations involve the manipulation of symbols”
(Singleton 2000, p. 179). Instead, the connectionist paradigm seeks to describe
information processing in terms of the strength of connections between units in a
network rather than in terms of rules. As Singleton puts it, “it is not patterns that
are stored (…) but rather the connection strengths between elements at a much
lower level that allow these patterns to be recreated” (Singleton 2000, p. 180).
Proponents of parallel processing models stress that an obvious advantage of
these models over serial search models is that they can explain the enormous com-
plexity of the information processing in the brain. On the other hand, connection-
ist models are frequently criticized for their inability to account for syntactic and
semantic aspects of language processing. Indeed, the current versions concentrate
mainly on the lexical level.
In the subsequent paragraphs two approaches representative of the connection-
ist tradition will be briefly outlined: localist connectionism and distributed con-
nectionism, also referred to as parallel distributed processing (PDP).16 In the
16 Both localist connectionism and distributed connectionism approach will be further discussed
in the multilingual context in Chap. 4.
WWW.EBOOK777.COM
2.6 Views on Language Processing 65
localist models each item is represented by a single unit (node) which is symbolic
in nature and has a functional value (cf. McClelland and Rumelhart 1981;
Stemberger 1992; Dell 1988; Roelofs 1992, 1999). In contrast, distributed con-
nectionism models assume the existence of distributed representations which are
processed in parallel and where units do not bear any functional value. The most
representative example of the latter type is the parallel distributed processing
model designed by Seidenberg and McClelland (1989). The basic difference
between the localist and distributed connectionism models lies in the representa-
tion of words. In contrast to the localist connectionism models which assume
one-to-one correspondence of lexical units and their mental representations, in
the PDP models “knowledge of words is embedded in a set of weights on connec-
tions between processing units encoding orthographic, phonological, and seman-
tic properties of words, and the correlations between those properties”
(Seidenberg and McClelland 1989, p. 560). In these models there are no entries,
bins or logogens. The models do not accommodate the mental lexicon in the tra-
ditional sense, nor do they account for the traditional lexical access. As Singleton
observes, “different portions of information are simultaneously processed inde-
pendently of one another (‘in parallel’) on different levels (‘distributed’)”
(Singleton 2000, p. 179).
One of the first parallel processing models (pre-connectionist model) is
the Interactive Activation Model which was put forward by McClelland and
Rumelhart (1981). The model postulates that perceptual processing takes place
simultaneously at a more than one level (parallel processing). McClelland and
Rumelhart (1981) distinguished the feature level, the letter level, the word level
and higher levels responsible for top-down input to the word level. The model
is not only parallel but it also accounts for interactive processing, which means
that in the process of word comprehension there are two co-occurring factors,
namely lexical knowledge and the incoming information from the perceived
stimulus. Thus, the processing is both top-down (conceptually driven) and
bottom-up (data-driven) at the same time. As for the representation of words,
the model posits that lexical units have their corresponding nodes which are
stored in levels (localist tradition) and are linked with other nodes. It needs to
be stressed that nodes are connected bidirectionally with other nodes at different
levels of the network.
The model also accommodates the frequency effect. Nodes have their acti-
vation levels which are modified by the amount of activation they receive from
other nodes (neighbours). Nodes corresponding to frequently or recently used
lexical items have a lower level of activation and thus are selected faster than
nodes which represent words of lower frequency. Communication between
nodes is possible due to the spreading activation mechanism. McClelland and
Rumelhart (1981) posit that there are two types of connections within the sys-
tem of nodes: excitatory and inhibitory ones. The former is responsible for
increasing the activation level of connected nodes, the latter for decreasing
the level.
WWW.EBOOK777.COM
2.7 Conclusion
This chapter has been meant to serve as a background for a more thorough con-
sideration of various theoretical issues and empirical investigations concerning the
multilingual context which will be discussed in the subsequent chapter of the pre-
sent work. The underlying assumption has been that presenting and explicating the
most significant concepts pertinent to the modelling of the mental lexicon would
enable the reader to interpret and evaluate the research projects whose design and
findings are presented in Chaps. 4 and 5. In order to do so, an attempt has been
made to review the classical theories and models concerning the organization of
the monolingual mental lexicon. The chapter started with a short discussion on the
internal structure of a lexical entry. Subsequently, it provided a brief overview of
various definitions of the mental lexicon as an entity, from those depicting it as
a dictionary to the ones which view it as a network. Apart from tackling termi-
nological issues, the present chapter dealt with the phenomenon of the represen-
tation of meaning in the human mind. The discussion revolved around the most
influential models concerning the internal structure of the monolingual lexicon, as
well as the numerous models of lexical processing. With regard to lexical process-
ing, the chapter summarized and assessed the better-known psycholinguistic mod-
els concerning the organization and functioning of the mental lexicon (Forster’s
lexical search model, Morton’s logogen model, Marslen-Wilson’s cohort model),
and gave a brief account of most recent computational models of visual word
recognition. Finally, special consideration was given to modular and connection-
ist perspectives on lexical processing. It needs to be noted, however, that despite
its substantial size, the section on lexical processing has not exhausted the vast
scope of visual word recognition study. In particular, a lot more can be said about
the achievements of computational modelling (cf. Norris 2005, 2013). However,
as stated above, the chapter has been meant to serve as a background for a discus-
sion of the multilingual mental lexicon, which constitutes the main concern of the
present work.
Now that the main issues concerning the concept of the mental lexicon have
been delineated, it is time to outline the most prominent hypotheses and models
of language storage, processing and retrieval in relation to the mental lexicon of
multilingual speakers. More precisely, the following chapter will be devoted to the
presentation and discussion of the issues of single vs. multiple lexicons and lan-
guage selective vs. language nonselective lexical access.
WWW.EBOOK777.COM
Chapter 3
Modelling Multilingual Representation
and Processing
3.1 Introduction
In the case of multilingual speakers, who are in the focus of attention of the
present work, the complexity involved in L1 lexical storage and processing, dis-
cussed in the previous chapter, is further multiplied by the complications added by
other lexical systems, those of L2, L3, Ln. Probably, the most enduring questions
in psycholinguistic research on bilingual and, more recently, multilingual mental
lexicon that have been repeatedly posed by researchers since the 1960s (that is,
since the beginning of extensive psycholinguistic research on bilingual and, later,
multilingual mental representation and processing), concern separate/integrated
storage and selective/non-selective access and as such, are both linked to the dis-
cussion about dependence/interdependence dichotomy. As Kroll and Tokowicz
rightly observe, “life experience offers ample support for each alternative” (2005,
p. 531). On the one hand, multilinguals are able to function in each of their lan-
guages independently, on the other, instances of code-switching are not uncommon
in their discourse. And thus, two alternative theories concerning bilingual lexical
representation have been proposed: one is the common memory theory (commonly
referred to as one-store hypothesis), which postulates a single integrated memory
system for both languages. The other, known as the multiple-memory theory (two-
store hypothesis), claims that words from each language are represented separately.
Whereas the results of tests involving word association or comprehension of
code-switched sentences speak in favor of a two-store hypothesis, studies on
transfer and interference support the hypothesis of one store.1 These two theoreti-
cal models, however, appeared too simple to explain all the experimental findings
(e.g., the result that cognates were recognized with the same speed in both
1 A review of these early tests and their results is proposed by Kroll and De Groot (1997).

DOI 10.1007/978-3-642-32194-8_3
WWW.EBOOK777.COM
68 3 Modelling Multilingual Representation and Processing
languages). Additionally, it turned out that task requirements play an essential

role in language processing: while conceptually driven tasks (e.g., free recall)
measured semantic and conceptual representations and thus favoured a one mem-
ory system view, data-driven tasks (e.g., lexical decision) supported the two-store
hypothesis.
An obvious corollary of these conflicting research findings concerning these
two theories was the introduction of a three-store hypothesis claiming that both (or
more) languages are differentially connected to the same conceptual-experiential
information store (cf. Paradis 1981, 2004, 2007; Kroll and Stewart 1994; Kroll and
De Groot 1997). A leading question that has been raised within the framework of
this type of model is not whether there is one multilingual system or more, but
how and to what extent the words from the multilingual’s different languages are
interconnected at both the lexical and conceptual levels. According to Dijkstra
and Van Heuven (2002a), the answer to this question appears to depend on two
types of variables. One group of variables is related to the language user. These
are variables such as level of proficiency, experience, and learning environment of
the second language (cf. Kroll and Stewart 1994; Ó Laoire and Singleton 2009).
The other group, which will be enlarged upon in the course of the next chapter,
comprises word type variables such as concreteness, word frequency, and cognate
status (cf. De Groot and Nas 1991; De Groot 1993; Dijkstra et al. 1999; Smits
et al. 2006).
Overall, the following sections will address the question of how different lan-
guages are represented in our memory and how bi- and multilingual speakers access
these languages during language processing. The next two Sects. (3.2 and 3.3)
discuss diverse approaches to language storage by describing the state-of-the-art in
this field and presenting the most influential models of bi- and multilingual lexi-
cal representation. The subsequent sections are concerned with the language pro-
cessing with the special focus on visual word recognition. More specifically, Sect.
3.4 addresses the notion of selective/nonselective access to different languages and
includes the discussion on the role of task demands in visual word recognition;
whereas the two further sections present a range of bi- and multilingual models of
visual recognition. Throughout the chapter, it is also argued that the question of
mental representation of multilingualism cannot be answered for all speakers in the
same way as it is determined by a number of user- and language-related factors.
3.2 Mental Representation of Bilingualism
A key issue that arises with reference to bilingual and, by extension, multilingual
storage is whether language systems are stored together or separately. Weinreich’s
compound-coordinate distinction was the first statement of the shared and separate
store hypotheses of bilingual memory which provided the basis for the subsequent
research in this field. On the basis of the Saussurean (1959) distinction between the
WWW.EBOOK777.COM
3.2 Mental Representation of Bilingualism 69
Fig. 3.1 Weinreich’s
three types of bilingual
representation: a coordinate,
b compound, c subordinate
(adaptation based on
Weinreich 1953)
signifier and the signified, Weinreich proposed three types of bilingual lexical organi-
sation2 i.e., coordinate, compound and subordinate (cf. Weinreich 1953; cf. Fig. 3.1).
In 1963, Kolers related the models of the organisation of bilingual memory to
models of the nature of representation. He adapted Weinreich’s compound sys-
tem and labelled it the shared-store model (Kolers 1963). In this model, all lexical
knowledge is stored in one lexicon, concepts are represented in some sort of non-
linguistic, abstract form. Kolers also described what was essentially the coordinate
system and labelled this the separate-store model. Here, the bilingual’s representa-
tions of words encoded in a specific language are stored in a form that in some
way is specific to that language.
A few decades later, Obler and Gjerlow (1999, p. 12) proposed a range of dif-
ferent models representative of possible connections existing in the internal lexi-
cons of multilingual language users. They distinguished between separationist
(Fig. 3.2a), interdependent (Fig. 3.2b, c), or partial overlap views (Fig. 3.2d) of
the multilingual mental lexicon and their conceptual basis. The model depicted in
Fig. 3.2a assumes the independence of particular lexicons (cf. Weinreich’s coor-
dinate model in Fig. 3.1a), whereas Fig. 3.2b presumes their interdependence (cf.
Weinreich’s subordinate model in Fig. 3.1c). The latter model resembles a bilin-
gual dictionary in which the meanings of lexical items from L2 are direct trans-
lations of L1 words. Interestingly, the equivalence of the words from subsequent
2 A more detailed description of the types of lexical organisation described by Weinreich will be
presented in Sects. 3.2.1 and 3.2.2.
WWW.EBOOK777.COM
Fig. 3.2 (a–d) Models of possible types of multilingual representation (adaptation based on

Obler and Gjerlow 1999). a Separate lexicons, b interdependent lexicons, c integrated lexicons,
d partially overlapping lexicons
lexicon (in this case L3) seems to be received through the subordinate language,
most likely the one of higher proficiency.
Further, Fig. 3.2c presents a fully integrationist model (cf. the concept of multicom-
petence discussed in Cook 1993, 1996, 2007a, b) where there is one conceptual store
with different language forms. The model assumes the cross-consultation between the
languages (cf. Singleton 1999) and resembles Kroll and Stewart’s Revised Hierarchical
Model (1994), discussed in detail in Sect. 3.2.3.2. Finally, Fig. 3.3d shows partial over-
lap of particular lexicons in their relationship to a shared conceptual store, thus refer-
ring to Kroll and De Groot’s Distributed Feature Model (1997; cf. Sect. 3.2.3.1).
Clearly, the issue of bilingual and, more recently, multilingual storage is the
subject of a multitude of empirical studies on the basis of which many models and
hypotheses have been developed. Some of them argue for the separate and others
for the shared organisation of the multilingual mental lexicon. Successful, in part,
seems to be an attempt to resolve this conflict made by those psycholinguists who
opt for the mixed structure of the multilingual store. The two following sections
provide a critical overview of studies and arguments supporting either the separa-
tist or integrationist view of the mental representation of multilingualism, whereas
WWW.EBOOK777.COM
Fig. 3.3 The distributed feature model (adaptation based on Van Hell and De Groot 1998a)
Sect. 3.2.3 will be devoted to the more detailed presentation and discussion of the
hypotheses and models supporting the mixed-storage view.
3.2.1 The Separate Storage View
Weinreich’s coordinate type of bilingualism is characterized by two separate

conceptual systems. In other words, there is a signified for each signifier (cf.
Fig. 3.1a). The model assumes the independence of lexicons of the particular lan-
guages (cf. the Independence Hypothesis, Obler and Gjerlow 1999; Fig. 3.2a). In
more current psycholinguistic terminology this means that
a word in the bilingual’s one language (L1) (e.g., English book) and its translation in his
or her second language (L2) (the Russian kniga) are represented in two conceptual forms,
one for the word in each language (De Groot 1993, p. 28).
Grainger and Beauvillain (1987) provided experimental support for the model using
a lexical decision task3 where context effects were investigated. In their study sub-
jects were asked to decide if a given letter string is French or English. Lists of letter
strings were either mixed (included both English and French items) or blocked for
language (with words either exclusively English or French). In the study it took
longer to decide about language affiliation of the stimuli if they were mixed across
languages rather than when they were presented in monolingual lists, which has
been interpreted as a fact supporting the separatist view. More interestingly, reaction
times were the longest for those items which lacked language specific orthographic
3 Lexical decision tasks as well as other tasks typically used in the research on multilingual stor-
age and processing are further delineated in Chap. 4.
WWW.EBOOK777.COM
patterns and were presented immediately after items from the other language. By
way of example, it took longer for the participants to recognize the English word
time when it was preceded by the French lire than when it was preceded by the
English word life. This effect of the preceding word was absent if the target was
white, whose spelling is illegal in French (cf. Grainger and Beauvillain 1987).
Another pool of evidence for the separation of language-specific lexicons comes
from the modularity hypothesis (cf. Chomsky 1988 after Singleton 1999; Fodor
1983; Sect. 2.6.1); according to which “in the case of post-pubertal L2 learner, L1
and L2 lexical operations proceed in absolute isolation from each other” (Singleton
1999, p. 167). Also, neurological research studies applying imaging techniques
prove that different brain areas are activated when the informant performs in L1 and
when he or she switches to L2 (cf. Arabski 2004; Field 2004; Kovelman et al. 2008).
Substantial arguments supporting the separation theory have also been pro-
vided by research studies with aphasic patients (e.g., Albert and Obler 1978;
Arabski 2004; Cutler 2005; Ibrahim 2009; Paradis 2004, 2009). By way of exam-
ple, Edwards (1994) reports on bilingual patients who lose both their languages
and then recover them one by one, not necessarily in the order of their acquisition,
which leads him to a conclusion that recovery patterns for each of the languages
may not be identical. In a similar vein, Smith (1997) summarizes her findings col-
lected in language impairment studies concluding that
a.) For some bilinguals, there has been a reported loss of one language, with no impair-
ment in functioning in the second language; b.) different types of aphasia have been found
to occur in the two languages of a bilingual; and c.) the recovery patterns in aphasia often
differ for each of the several languages known to the patient (Smith 1997, p. 149).
It needs to be stressed, however, that Smith (1997) does not postulate a neuroana-
tomic separation of the systems but proposes a difference in the levels of activation
of the two lexicons.
3.2.2 The Shared Storage View
In contrast to the above-presented evidence supporting the separate-storage view,

studies into cross-linguistic interference and code-switches substantiate the claim
that the two languages do not operate in total separation; the organisation described
by Weinreich (1953) as compound bilingualism, and by Kolers (1963) as shared-
storage model. Admittedly, there must exist some kind of connection between the
two systems which could account for these phenomena. As Singleton puts it,
The fact that bilinguals switch between languages sometimes, apparently, without mean-
ing to - whether at one word level or over longer stretches, is highly relevant to the
question of separation/integration of lexicons […] and […] seems to rule out absolute sep-
aration (Singleton 1999, pp. 174-5).
In other words, transfer and code-switching, taken together, seem to eliminate

the possibility of total separation of languages in the mind. In his critical review
WWW.EBOOK777.COM
of studies on the multilingual lexicon, Singleton (1999) quotes the following

arguments in favour of the integrationist view. Firstly, he refers to cross-linguistic
influence (CLI) observed in multilingual production such as examples of trans-
fer errors or use of calques (cf. studies using translation tasks methodology;
Singleton 1999). Secondly, he describes multilingual behaviour with special focus
on the phenomenon of code-switching, defined as an unintentional insertion of
words/phrases from the other language. Finally, two more arguments quoted by
Singleton (1999), those of particular importance in the context of the research
presented in the practical part of the present work, come from studies on learning
strategies employed by multilinguals such as use of cognate words or interlingual
homographs as well as research projects pointing to quicker reaction times in the
retrieval of cognate words (cf. Chaps. 4 and 5).
Ample evidence for the presence of cross-linguistic influence in multilingual
performance, supporting the existence of the integrated lexical systems, emerges
also from research by Grauberg (1971, after Singleton 1999, p. 179)—the study
of errors; Chen and Ho (1986), Tzeglov et al. (1996), Van Heuven et al. (2011)
and Marian et al. (2013)—the cross-language version of Stroop task; Smith and
Kirsner (1982, after De Groot and Nas 1991) and MacLeod (1991)—the picture-
word interference paradigm; De Groot and Nas (1991), Jiang and Foster (2001),
Basnight-Brown and Altarriba (2007), Trofimovich and McDonough (2011), Hu
and Jiang (2011) and Guasch et al. (2011)—cross-language semantic priming
experiments; Clyne (2003), Toribio (2001) and Gullifer et al. (2013)—experi-
ments on code-switching and De Groot et al. (2002), Dijkstra (2005), Dijkstra
et al. (2010) and Fitzpatrick and Izura (2011)—reaction time studies.
An obvious corollary of the contradictory and inconclusive findings of numer-
ous experiments aimed at explaining the separate-shared dichotomy of the mul-
tilingual lexical storage that were presented in the preceding paragraphs, was the
introduction of models postulating two distinct levels of representation namely
the lexical, storing word forms, and the conceptual, storing meanings. In the light
of convincing arguments supporting both integrationist and separatist view, many
psycholinguists lent their support to models allowing for mixed multilingual stor-
age and shifted their attention to the degree of interconnectivity of the particular
lexicons at different levels of representation.
3.2.3 The Mixed Storage View
It seems appropriate to follow the discussion of the results of empirical investiga-

tions advocating in favour of either separate or integrated bilingual representation
with a presentation of research findings that provide evidence for the mixed stor-
age view. The research findings that these investigations led to have served as the
basis for the development of various theoretical models that have exerted a signifi-
cant influence on the studies of the representation and processing of information in
bilingual and multilingual memory.
WWW.EBOOK777.COM
An important example of the mixed storage view is Paradis’ Subset Hypothesis

(Paradis 1981, 2004, 2007) which connects the idea of shared and separate rep-
resentations claiming the existence of one common language system, in which
the bilingual’s two languages are represented as two subsystems developing in
language production and comprehension. Since words belonging to the same lan-
guage are normally used together, strong connections are formed between them.
Once one subsystem is activated, the other one is deactivated, where activation
follows language specification. Language specification, in turn, is done by lan-
guage cue, a preverbal message containing information about language choice.
Interestingly enough, although the very concept of subsystems may imply that lan-
guages are not fully integrated into a single mental structure, subsystems formed
in languages use do not have to be language specific and may contain items
belonging to both languages if these items are frequently used together (cf. Paradis
1997, 2004, 2007, 2009). To reiterate, in this model, a word can be conceived of
“as a multidimensional matrix of interconnected phonological, syntactic, morpho-
logical and semantic features” (Paradis 1997, p. 335).
This type of mixed organization is also reflected in the Revised Hierarchical
Model (cf. Sect. 3.2.3.2), as developed by Kroll and Stewart (1994). The model
postulates connections from the lexicons to the conceptual nodes that represent
word meaning. The more similar two concepts are, the more meaning features they
have in common. As concrete lexemes have a corresponding item in the visible
world, their conceptual features are supposed to show almost complete overlap,
whereas abstract words might only share some conceptual features (cf. Kroll and
Stewart 1994; Kroll and De Groot 1997; Kroll and Tokowicz 2005).
Another important issue that needs to be addressed in the discussion on the
mixed storage view is the fact that organisation of the multilingual mental lexicon
is described as mixed along two axes: vertical and horizontal. The former presup-
poses that the type of organisation differs across levels of the mental lexicon (the
lexical level is language-specific and the conceptual level is language-independent).
The latter, on the other hand, assumes that in the course of developing L2 pro-
ficiency, the organisation of the multilingual’s mental lexicon is a kind of con-
tinuum with separate and common types coexisting within a single multilingual
mind.
Viewed from the horizontal perspective, the problem of mental representa-
tion of multilingualism cannot be answered for all speakers in the same way. The
research to date has led to the assumption that the modelling of bilingual men-
tal representation is dependent on some external factors such as language profi-
ciency, age of acquisition, structural differences between languages or relative
language dominance. As Kroll and Stewart (1994) critically argued, most early
models of processing represented proficient bilinguals ignoring their learning his-
tory. Therefore, their Revised Hierarchical Model (cf. Sect. 3.2.1) was devised for
unbalanced bilinguals, in case of whom L1 words are assumed to be more strongly
connected to concepts than L2 words, which, on the other hand, are more strongly
related to their corresponding translation equivalents in L1. Notably, this type of
links seems to hold true for typical language-learning scenarios where languages
WWW.EBOOK777.COM
are learned via translation equivalents and the neural connections between these
equivalents are more strongly developed, whereas in uncontrolled acquisition situ-
ations the links between concept and lexical lemma might be stronger.
The significance of bilingual participants’ proficiency level was further recog-
nized by van Hell and Dijkstra who aptly state that the relative fluency in their two
languages
will affect the bilingual’s sensitivity to L1 interference when he/she is processing in L2,
and the sensitivity to L2 interference when processing in L1. The underlying rationale
here is that less activation is needed to recognize words that are used relatively frequently,
as are words in a language in which the speaker is relatively proficient (2002, p. 782).
Accordingly, it can be concluded that “bilingual memory representation is not
static but rather a dynamic system that can be influenced by language usage”
(Heredia 2008, p. 57).
Other studies that have indicated a proficiency effect on multilingual lexi-
cal organisation are reviewed by Christoffels et al. (2006) and include those by De
Groot and Comijs (1995), Kroll and Curley (1988) and Chen and Leung (1989).
They are all focused on adult bilinguals of different levels of proficiency and all
use picture-naming and word-translation methodologies. It needs to be noted, how-
ever, that they all adopt various proficiency measures, from the number of years of
instruction (Cieślicka 2000; van Hell and Dijkstra 2002) and self-assessment of the
language skills (Haigh and Jared 2007) to on-line vocabulary tests (Lemhöfer et al.
2004), and tend to allocate bilinguals into separate proficiency groups rather than to
investigate gradual changes that appear along with the increasing L2 (an important
exception being the research by Chambers and Cooke 2009). A further complicating
factor seems to be the fact that bilinguals from different proficiency groups may also
belong to different age groups and thus their bilingual representation may also be
dynamic. Here a vast number of current empirical studies pointing to the changing
patterns of bilingual representation of early and late bilinguals comes from the field
of neurolinguistics (cf. Wattendorf et al. 2001; Wartenburger et al. 2003; Ullman
2001; Kroll and Tokowicz 2005; Miozzo et al. 2010).
Another aspect concerning the organization of the multilingual lexicon, also
underpinned by neuroimages of the multilingual brain, is a great inter-subject vari-
ability between speakers. This is also reflected in the dynamic systems theory, as
suggested by de Bot et al. (2007). The proposed framework implies that all vari-
ables are interrelated so that changes in one variable will have an impact on all
other variables in the system. And thus, a major question raised by de Bot and his
colleagues (2007) is whether, as commonly assumed, individuals really have simi-
lar L1 systems. Further, in their view it is also possible that in multilingual repre-
sentation there might be different systems according to the individual’s linguistic
biography and acquisition contexts: different numbers of subsystems and different
modes of access (cf. de Bot et al. 2007).
Indeed, the question of mental representation of multilingualism cannot be
answered for all speakers in the same way. This contention has been confirmed
by Cieślicka (2000) who conducted the lexical decision task experiment on a
group of fluent and non-fluent Polish-English bilingual language users who were
WWW.EBOOK777.COM
also allocated to different groups basing on their preferred L2 vocabulary learning

strategies. Using the obtained research findings, she introduced the variable inter-
connection hypothesis according to which associative connections between differ-
ent lexical units in the bilingual’s mental dictionary vary in strength depending on
the type of a bilingual learner’ experience in his or her L2. The experiment showed
that both L2 proficiency and the preferred L2 vocabulary learning strategies exert
an important influence on the patterns of lexical representation and processing.
Consequently, she advocates the need for bilingual models to account for essential
factors influencing bilingual experience, such as “the context of L2 acquisition and
different learning strategies employed by bilingual learners to acquire L2 lexis”
(Cieślicka 2000, p. 52).
As transpires from this brief overview of the research studies presented in the
foregoing paragraphs, the traditional question of whether the languages are stored
together or separately was ill-formed. As Kroll and Tokowicz (2005) correctly
observe, the question requires that assumptions be made about a number of fea-
tures such as levels of representation, the storage changes at different stages of
learning or different learning contexts. And thus, the models presented in the fol-
lowing sections illustrate the ways in which assumptions have been made about
different levels of representation and the developmental issues.
3.2.3.1 The Distributed Feature Model
The Distributed Feature Model assumes language independent lexical and concep-
tual levels of representation with a language specific lemma level. It accounts for
the results obtained by a number of researchers in a variety of experiments such
as translation production, translation recognition, lexical decision, primed lexi-
cal decision, Stroop-type picture-word production and word association or (cf.
Grainger and Beauvillain 1987; Altarriba and Mathis 1997; Costa and Caramazza
1999; De Groot 1992; De Groot et al. 1994; Van Hell and De Groot 1998a, b).
Interestingly, all these experiments proved that it is not pure language mixing that
increases the processing time but the extent to which stimuli used in experiments
activate both languages, which points to the influence of task dependency on lexi-
cal processing.
Within the early version of the model, De Groot (1992, p. 1011) assumed
different storage patterns only at the semantic/conceptual level and proposed
that “words in the bilingual’s mental lexicon activate conceptual features which
are assumed to be distributed such that particular concepts correspond to sets of
activated features”. Van Hell and De Groot (1998a) later proposed to extend the
Distributed Feature Model and suggested the distribution of features not only at
the conceptual but also at the lexical level (cf. Fig. 3.3). It is imperative to indi-
cate, however, that the version of the model presented in Fig. 3.3 depicts only one
layer of lexical features. This simplification was done for the sake of clarity and
not to imply that multilayered representation of various aspects of lexical form is
excluded from the model.
WWW.EBOOK777.COM
It needs to be noted that the model attempts to accommodate major findings

within bilingual translation research, namely the common result that concrete
words and cognates are translated faster than abstract words and noncognates (cf.
De Groot and Nas 1991). The model distinguishes between word types in terms of
patterns in which L1 and L2 words are stored. The claim is that some words, nota-
bly concrete nouns and cognates, are more likely to map onto virtually the same
pool of semantic features across languages than abstract nouns and noncognates.
The more overlap between semantic features, the more quickly the translation will
be retrieved and the more likely bilinguals will be to consistently produce the
same response. As far as concreteness effect is concerned, concrete words, being
typically names for objects shared across languages and having close translation
equivalents, are thought to activate very similar, or the same, sets of features for
both languages known to an individual. In contrast, since abstract words are more
culturally determined and interpretation-sensitive, they are less likely to be shared
across languages and to have close translation equivalents. Consequently, they are
likely to share fewer features than concrete words. By analogy, as regards cognate
effect, cognates have been reported to be translated more rapidly than noncognates
because the former share more features than the latter.4
Recapitulating, the model assumes language independent lexical and concep-
tual levels of representation with a language specific lemma level. Van Hell and De
Groot (1997) distinguish the level of lemmas as a form of representation of acti-
vation patterns between lexical forms and meanings. When contextual factors are
available, lemmas additionally reflect syntactic processing which selectively weigh
the activation of features, otherwise the lemma level only reflects word to meaning
mappings. Admittedly, the fact that the lemma level is language specific does not
exclude the possibility of cross-linguistic influence nor does it deny access to the
shared inventory of lexical and semantic features but it enables an individual to
function in a monolingual mode if only one of the languages is active.
3.2.3.2 The Revised Hierarchical Model
The Revised Hierarchical Model was proposed by Kroll and Stewart (1994) to
“characterize the consequences of differential expertise in the two languages for
the connections between words and concepts” (Kroll and Stewart 1994, p. 157).
The model includes independent lexical representations for each language, with
L1 assumed to be larger than L2, and a shared conceptual representation. Of inter-
est is that it focuses on the connections between words and concepts, not on the
structure. The model is based on the classic study of Potter et al. (1984) who refer-
ring to the empirical findings supporting the assumption that words may be stored
separately from concepts (cf. Potter 1979), proposed two separate models of
4 Empiricalevidence for concept mediated translation of cognates was found also by Kroll and
Stewart (1994) and will be referred to in Chap. 4.
WWW.EBOOK777.COM
Fig. 3.4 a The word

association model (adaptation
based on Potter et al. 1984).
b The concept mediation
model (adaptation based on
Potter et al. 1984)
bilingual memory representation—the Word Association and Concept Mediation

Models (cf. Fig. 3.4a, b).
The Word Association Model (cf. Fig. 3.4a) assumes that second language
words are associated to first language words and that only through first language
mediation can second language words gain access to concepts (cf. Weinreich’s
subordinate model). In contrast, the Concept Mediation Model (Fig. 3.4b) predicts
that second language words directly access concepts. In other words, within the
Concept Mediation Model two lexicons are connected only through the common
conceptual system (cf. Weinreich’s compound model), whereas within the Word
Association Model there are direct links between lexical representations of transla-
tion equivalents of the two languages.
Evidence for both models was found by Kroll and Curley (1988) in the experi-
ments that involved picture naming in participants’ L2 and translation of words
from L1 to L2.5 Kroll and Curley (1988) found that non-proficient bilinguals rely
on word-to-word mappings, as in the Word Association Model, whereas proficient
bilinguals access concepts in L2 performance, which is in line with the Concept
Mediation Model. In order to account for this shift from word association to con-
cept mediation with the increase of L2 proficiency, Kroll and Stewart (1994)
designed the Revised Hierarchical Model (RHM).
In the RHM, the L1 lexicon is significantly larger than the L2 lexicon (cf.
Fig. 3.5). The model assumes both direct lexical connections of the Word
Association Model and conceptual links of the Concept Mediation Model. Lexical
links from L2 to L1 are stronger (depicted with a continuous line) than those
5 For detailed description of empirical data used in the original research by Potter et al. (1984)
see the review by Kroll and Tokowicz (2005).
WWW.EBOOK777.COM
Fig. 3.5 The revised
hierarchical model of lexical
and conceptual representation
in bilingual memory
(adaptation based on Kroll
and Stewart 1994)
from L1 to L2 (depicted with a dotted line). By and large, this asymmetry in the
strength of connections between words in the two languages and meaning is a
characteristic feature of this model.
The model represents the consequences of the learning history of the late sec-
ond language learner for whom lexical and conceptual representations are already
in place for L1 when L2 learning begins. Early in the acquisition, L2 words are
strongly linked with the L1 items and only weak links connect L1 words with
those of L2. Kroll (1993) emphasizes the fact that
the strength of the connections between the L1 and L2 lexicons and between each lexicon
and conceptual memory is hypothesized to vary as a function of relative fluency in L2 and
language dominance (Kroll 1993, p. 70).
In other words, with the increasing proficiency, stronger and stronger semantic
links between concepts and L2 items are developed.
What is interesting to note is that the empirical observation that led to the RHM
was initially the finding that translation from L1 to L2 is typically slower and
more error-prone than translation from L2 to L1. It took less time for proficient
participants to translate words from L2 to L1 (backward translation) than from L1
to L2 (forward translation). To explain these findings Kroll and Stewart claimed
that backward translation operates within the lexical level (hence faster), unlike
forward translation which is conceptually mediated (slower). By analogy, Kroll
and Dussias assume that
[i]n the L2 to L1 direction, the strongly associated translation equivalents will be assessed
directly. In the L1 to L2 direction, the bias to activate the meaning of the L1 word will
encourage reliance on a translation route that engages semantics. The latter process will
require additional processing and also the potential negotiation of lexical competition
prior to selecting an L2 response. The L1 to L2 direction is hypothesized to be particularly
difficult for less proficient bilinguals for whom the concept to L2 links are relatively weak
(Kroll and Dussias 2004, p. 176).
Despite convincing evidence supporting both the existence of a larger translation

asymmetry at lower levels of L2 proficiency and the fact of L1 to L2 translation
changing most dramatically with increasing L2 skill (cf. Kroll and Dijkstra 2002),
there are still certain doubts concerning the fact that the RHM model is not com-
pliant with the recent research on lexical representations.6 For instance, the results
6 For a critical review and assessment of the model see Kroll et al. (2010).
WWW.EBOOK777.COM
Fig. 3.6 The revised
hierarchical model (RHM)
with integrated lexicons
and distributed semantic
representations; Dutch-
English stimulus example
(adaptation based on Desmet
and Duyck 2007, p. 177)
of Duyck and Brysbaert’s (2004) study contrast with the RHM’s developmental
hypothesis, which states that L2 lexico–semantic mappings may only develop in
very high proficiency levels.
A critical review of the model was also proposed by Desmet and Duyck (2007)
who noted its two important problems. Firstly, as the researchers rightly note, the
model still contains two separate lexicons, which is clearly at odds with the large
body of evidence against lexical autonomy. Secondly, Desmet and Duyck suggest
that the RHM should be more specific about the nature of semantic representa-
tions, assuming gradual semantic feature overlap depending on word variables
(similar to the way semantics are conceived in the Distributed Feature Model; cf.
Sect. 3.2.3.1). According to Desmet and Duyck (2007), an updated version of the
RHM with the modifications concerning the above mentioned arguments would
probably look like the model presented in Fig. 3.6.
The model aptly illustrates the current mixed storage theories where words are
stored in a language specific manner and concepts, which underlie the words, form
a single language-independent store.
3.3 Modelling Multilingual Representation
The preceding sections focused on the lexical storage in the bilingual mental lexi-
con. The vexing question that arises in the context of the present work, however,
is what happens to the lexical storage system when more than two languages are
involved. In the following paragraphs an attempt will be made to delineate one of
the most prominent issues in the multilingual lexicon studies, namely, the degree
to which lexical operations are separate or integrated within a multilingual mind.
In turn, the problem of multilingual processing will be tackled in Sect. 3.7.
As already mentioned in Sect. 3.3, early studies (e.g., Meara 1982) assumed that
the L1 mental lexicon was qualitatively distinct from the mental lexicons of other
WWW.EBOOK777.COM
3.3 Modelling Multilingual Representation 81
Fig. 3.7 Model of
multilingual memory
representation (adaptation
based on Gabryś-Barker
2005, p. 64)
languages, the difference in quality implying the separate storage. Another typical
argument against integration of lexical systems of many languages in one mind was
put forward by the proponents of the modularity hypothesis (cf. Sect. 2.6.1). Strong
evidence supporting the separatist view came also from studies of aphasia and lan-
guage loss in multilinguals (cf. Sect. 3.2.1). Interestingly enough, some early research-
ers referred to the psychotypology effect arguing that it may support selectivity which,
by implication, may suggest separate storage (cf. Singleton 2003 for review).
In contrast, more recent studies promote the idea of the high degree of intercon-
nectivity within a multilingual mind. One of the most fervent supporters of this
view is Vivian Cook (e.g. 1996, 2007a, b) who has for many years been advocating
for the complete integration of language competence across languages—the holis-
tic multicompetence hypothesis (cf. Chap. 1). Further, extensive evidence in favour
of integration comes from Franceschini et al.’s research findings, comprehensively
reviewed in the authors’ contribution on a range of brain-imaging studies published
in what seems one of the first volumes devoted entirely to the multlingual acquisi-
tion, storage and processing (Franceschini et al. 2003; in Cenoz et al. 2003). The
results of the conducted studies, all lead to the conclusion that lexical-semantic
aspects of the processing of all languages known to an individual are subserved by
essentially the same areas of the cortex (cf. Franceschini et al. 2003). Importantly,
the authors of the study account for the role of the level of proficiency and the age
of acquisition in the organisation of the lexicon. Their research findings imply that a
high level of proficiency acquired in a late learned L2 may mask the differences con-
nected with the onset time of acquisition (Franceschini et al. 2003, p. 163). The con-
tention which, if confirmed, seems to be of importance for the pedagogical context,
which for many years has been strongly promoting early language education.
The role of language proficiency for the multilingual lexicon structure was also
confirmed in the research by Gabryś-Barker (2005). Basing on the data gathered in
a series of association tests conducted on Polish-English-German and Portuguese-
English-German trilinguals, Gabryś-Barker put forward a multilingual adaptation
of Kroll and Stewart’s Revised Hierarchical Model. In this model, depicted in
Fig. 3.7, items in the multilingual memory are interconnected by way of two types
of links, lexical and conceptual. The lexical links exist “within the same language
referring to the form as a factor and across languages using translation equivalents”
WWW.EBOOK777.COM
(2005, p. 64), whereas the conceptual links exist “within the same language and
across languages manifested as a semantic field search, e.g., coordination or super-
ordination” (2005, p. 64). It is believed that, in the multilingual memory words are
accessed either via lexical links or conceptual links depending on a set of factors,
such as language dominance in the multilingual competence and performance of a
learner, language proficiency in all the languages, the form of a linguistic task, and
the type of a linguistic stimulus (Gabryś-Barker 2005, p. 64).
Gabryś-Barker concludes that the gathered research findings confirmed “a con-
ceptually based L1 mental lexicon, separation of the lexicons with growing lexi-
cal competence (L2) and strong connections between high proficiency L1 and L2,
and L3 (the lowest level proficiency) lexicons, in other words, integration between
these lexicons” (2005, p. 73).
Important evidence for the full integration of language systems within the mul-
tilingual mind can be found in the research conducted by Dijkstra (2003a, b) and
Dijkstra and Van Heuven (e.g., 1998, 2002a) who have devoted much time to the
investigations of the multilingual processing. Since the present section is pre-
dominantly focused on mental representation, both bilingual as well as multilin-
gual models proposed by Dijkstra will be delineated in Sect. 3.7, which is entirely
devoted to the problem of multilingual lexical retrieval. The present discussion
will instead offer evidence of a high degree of cross-lexical connectivity (also
referred to in the practical chapters of this work), which is what most researchers
on the multilingual lexicon mean by integration.
Psychologists, psycholinguists as well as neurolinguists opine that the most
important principle for the organization of the mental representation is the ability
to categorize new information basing on similarity to existing knowledge repre-
sentations (cf. Rosch 1975; Fay and Cutler 1977—discussed in Sect. 2.4.3 of the
previous chapter). Consequently, many theories aimed at designing a multilin-
gual mental representation model refer to the exploitation of similarity between
new lexical input and previous lexical knowledge. An important example is the
Parasitic Model originally devised to explain aspects of L2 vocabulary acquisi-
tion and later extended to incorporate the L3. The model makes strong hypotheses
about the initial stages of the cognitive processes involved in additional language
vocabulary development and concentrates on the automatic, unconscious detec-
tion and adaptation of similarity between novel lexical input and the information
already stored in the mental lexicon (cf. parasitic learning strategy; Hall 2002; Hall
et al. 2009). According to the Parasitic Model, “new lexical representations will be
integrated where possible, into the rest of the network via connections with preex-
isting representations (…), at points of similarity or overlap between them” (Hall
and Ecke 2003, p. 72)—the phenomenon documented in a number of studies on the
role cognates in CLI (e.g., Ringbom 1987, 2001, 2007; Cenoz et al. 2001a). The
model predicts that new words are integrated into the existing lexical network with
the least possible redundancy and as rapidly as possible. As Hall and Ecke (2003,
p. 71) declare, the ability to integrate new information within the existing knowl-
edge based on the criterion of similarity is “essential for the development of con-
ceptual relations and networks as well as for the acquisition and organization of the
WWW.EBOOK777.COM
3.3 Modelling Multilingual Representation 83
mental lexicon”. By way of illustration, when a new L3 verb has a cognate in one
of the previous languages learned, but not in the other, learners will tend to assume
that the new verb is used in the frame of its cognate in the L1 or L2 (cf. Hall and
Ecke 2003). For example, the L3 German verb brechen, at the early stage of lexical
acquisition, is connected parasitically to the L2 English form break, rather than the
L1 Spanish form romper, because of the cognate relationship between the L2 and
L3 forms. Thus, the learner will be led to assume that the German form brechen is
used nonreflexively, like its English L2 cognate counterpart break, rather than with
a reflexive pronoun, as in the case of the Spanish L1 translation romper (cf. Hall
et al. 2009, p. 163). Overall, the model adopts the view of the mental lexicon as “an
interconnected network of processing units through which the activation spreads”
(Hall and Ecke 2003, p. 7). Items in the lexicon are connected on the basis of
similarity—both within a single language and between languages. They are repre-
sented in the form of a lexical triad (cf. Triad Model; Hall 1992 in Hall et al. 2009)
which comprises three levels: form (or lexeme) which is a representation of the
phonological and/or orthographic features of a word, syntactic frame (associated
with Levelt et al.’s, (1999) notion of lemma) which includes a specification of a
syntactic category and any other feature regarding sentential deployment of a word,
and concept (representation of meaning).
The role of typological or rather psychotypological similarity for the integra-
tion of linguistic systems in the multilingual mind resonates also in the model
proposed by Herwig. In her view, the mental lexicon comprises “dynamically
interacting subsystems of a common linguistic system, subject to individual vari-
ation and change over time” (Herwig 2001, p. 115). To shed some light on the
multilingual mental lexicon, Herwig adopts Paradis’ Subset Hypothesis for TLA,
advocating for three languages to form a single system with three identical, lan-
guage-specific networks of connections. In the case of SLA, in the initial stages
of bilingual development, there is an extended system where the second language
is strongly related to the L1 items. With progress in L2 proficiency, the L2 net-
work becomes less L1-dependent to the extent of becoming a fairly independent
subsystem. According to Herwig, the same pattern of development is repeated in
TLA. The point is that it is not necessarily L1 that undergoes the initial extension.
Instead, it seems that the language that is typologically closer to L3 will undergo
the initial extension of the network. As Herwig puts it, “the structure of the lexi-
con, the connectivity of its subsystems and their interrelation depend on a number
of factors, such as perceived linguistic distance, proficiency of the user, or method
of acquisition” (Herwig 2001, p. 115).
The effect of psychotypology was also exploited by Singleton, who in 2002
proposed a model of a trilingual speaker based on Weinreich’s assumptions.
Adopting Weinreich’s models for trilinguals, Singleton concentrated predomi-
nantly on the subordinate organisation of languages. The key question he asked
was to which of the previously acquired languages the L3 would become sub-
ordinate. On analysing empirical data he concluded that L3 will be subordinate
towards the language which is perceived as typologically closer (in his example
Italian to Spanish). He also hypothesized that with progress in the acquisition of
WWW.EBOOK777.COM
L3 the psychotypologically close languages will develop a compound structure

which finally may turn into a coordinate one. Singleton (2002) illustrates his idea
with an example of English-Spanish bilingual’s acquisition of Italian in three
stages of development. In the very early stages of TLA, Italian forms are linked to
Spanish forms, Spanish being typologically closer to Italian than English. Later as
progress occurs, Italian and Spanish words will share only concepts. According to
Singleton (2002), in the highly advanced stages of TLA, all three languages may
develop a coordinate structure. It means that each language will have separate rep-
resentations at the lexical and conceptual levels. It needs to be noted that Singleton
emphasized a high degree of connectivity between the structures, which enables
transfer of items in all the systems. At the same time the separation of the lan-
guages allows for the choice of a prime source of cross-linguistic influence.
The data reported above demonstrate clearly that the architecture of the multilin-
gual mental lexicon admits CLI from all possible source languages and at all repre-
sentational levels. In the light of the discussed research findings, it seems legitimate
to claim that CLI is the major determinant of the acquisitional and organisational
patterns of the multilingual mental lexicon. Still, the scope of possibilities for the
organisational arrangements for multilingual users is very large and highly com-
plex but only due to a number of possible patterns of multilingual development
(cf. Chap. 1). Considering only a trilingual learner there is a number of possibili-
ties. For instance, following Singleton’s example (2003) in the case of a coordinate
bilingual learning an L3, the new linguistic system needs to become subordinate to
either of the languages. The vexing question arises, to which one? The native one or
the psychotypologically similar one? Or maybe the new language item associates
itself with both languages but to a different degree? If so, is the degree of interactiv-
ity dependent on the level of proficiency or maybe on the learning history, as sug-
gested by Cieślicka’s variable interconnection hypothesis (cf. Sect. 3.3). How will
the connections change if we consider subordinate bilinguals learning L3? How
do they incorporate a new language item? And further, what about subsequent lan-
guages L4, L5, Ln; how are they linked to the existing system of previously learned
languages, most likely of varying proficiency? There is no one valid answer to
all these questions but only due the fact that accounting for so many possibilities
within one model becomes a very challenging task. Consequently, research results
vary depending on the adopted methodology and the analyzed target group. The
recurring finding, however, is the dominant role of psychotypology in the organi-
sational patterns of the multilingual lexicon, the issue which will also constitute a
leading motif of the empirical studies presented in Chaps. 4 and 5.
3.4 Modelling Bilingual and Multilingual Processing
The previous sections focused on the issue of bi- and multilingual representation and
showed that the majority of current research arguments in favour of the mixed stor-
age and high interconnectivity of language systems within one mind. To get a better
WWW.EBOOK777.COM
3.4 Modelling Bilingual and Multilingual Processing 85
insight into how languages are interconnected, the present sections takes a closer
look at the mechanisms of how the languages are accessed. However, before some
information on lexical access is provided, it is worth mentioning that research stud-
ies that address the question of whether lexical candidates from different languages
are activated during bilingual word recognition basically exploit three main types
of stimulus materials: interlingual homographs, also referred to as homographic
noncognates or more typically false friends (words with the same form but differ-
ent meaning across languages such as fin meaning end in French), homographic
cognates (words sharing both form and meaning across languages such as wolf in
English and German) and interlingual neighbours which resemble interlingual hom-
ographs except that their forms do not overlap completely between the languages
(such as Dutch mand and English sand; example after De Groot 2013a, p. 76). These
three types of stimuli are typically used since it is believed that if reaction time dif-
ferences between these words and matched control words arise, this will be probably
due to their existence in two or more languages (cf. Dijkstra 2005; De Groot 2013
for a thorough discussion of all the types of stimuli).
3.4.1 Selective and Non-Selective Access
The debate over the selective/non-selective access dichotomy has not ceased to
inspire theoretical considerations and empirical investigations. In a number of early
studies on lexical access, language selective position was favoured. On the basis of
their research conducted with a group of Spanish-English bilinguals, Gerard and
Scarborough (1989) concluded that the participants were accessing each of their two
lexicons selectively when they performed a monolingual task. Similarly, in a pleth-
ora of early research studies (cf. Caramazza and Brones 1979; in Singleton 1999;
Soares and Grosjean 1984; in Dijkstra 2005) no clear reaction time (RT) differences
were observed between test items and controls. Also some later studies replicated the
RT null results under similar experimental circumstances (cf. De Groot et al. 2000;
a lexical decision experiment with interlingual homographs and cognates). These
research findings were interpreted as constituting support for language specific
access as the participants of the studies seemed to have been operating in a language
specific manner without any influence from their L2. It has to be stressed, however,
that the early understanding of language selective access implied partition of the lex-
icon by language. Clearly enough, this assumption has evolved in the course of psy-
cholinguistic research and it is now generally agreed that language selective access
does not have to imply the existence of separate lexicons for each of the languages,
but it may be the result of different activation patterns for L1, L2 and Ln (cf. Costa
and Caramazza 1999; Paradis 2004, 2009; De Groot 2011).
In recent years, more and more studies reported evidence in support of lan-
guage nonselective access with respect to form (orthographic and phonological),
as well as semantic representations. Obviously enough, this does not mean that
words from the two or more languages cannot be distinguished anymore; rather,
WWW.EBOOK777.COM
language information is thought to be available at a later point in time than the

word activation itself, but it cannot prevent an initial activation of word candidates
from the non-target language.
According to de Bot (2004), in the last two decades psycholinguistic research has
accumulated a substantial body of evidence to support the nonselective view, a selec-
tion of which will be presented in the following paragraphs. Firstly, an important
source of information supporting the nonselective access hypothesis is data from
eye-tracking studies and brain-imaging studies (cf. Marian and Spivey 2003;
Wartenburger et al. 2003; Bartolotti and Marian 2012). The research findings docu-
mented in these neurolinguistic studies are commonly interpreted as evidence for
simultaneous activation of both languages in the early phonetic stages of perception.
Much evidence in favour of nonselective access has also been obtained in experi-
ments that used interlingual (orthographic) neighbours as stimulus materials.
According to Dijkstra (2005, p. 187), n interlingual neighbour is “any word differ-
ing by a single letter from the target word with respect to length and letter position”.
It has been confirmed that in monolingual word recognition many possible words
initially become active on the presentation of a letter string, and the reader is usually
not aware of them; only the word that is eventually recognized becomes available to
awareness.7 Similarly, empirical studies show that neighbours from both the same
and the other language (interlingual orthographic neighbours) are activated during
the presentation of a target word. This provides evidence that, with respect to ortho-
graphic codes, the lexicon of bilinguals is integrated and nonselective in nature.
Notably, Jared and Kroll (2001) in their word naming study showed that the same
conclusions hold for the phonological part of the bilingual lexicon.
Further support for the existence of nonselective access in bilingual visual word
recognition comes from a plethora of studies which observed RT differences for
interlingual homographs and cognates under different experimental conditions. In
fact, evidence from studies using cognates and interlingual homographs and hom-
ophones follows a similar logic to the neighbourhood studies. If access is language
selective, the fact that words are cognates or have many neighbours in another lan-
guage should have no effect on reaction times. If access is nonselective, candidates
from both languages will present themselves and this competition will again lead
to longer reaction times. A large number of studies have been designed to investi-
gate this assumption and the overwhelming evidence in favour of the nonselective
access hypothesis cannot go unnoticed.
One of the studies supporting the nonselective access that exploited cognates
and interlingual homographs is an experiment conducted by Dijkstra et al. (1999)
in which Dutch-English bilinguals performed an English lexical decision task with
English words varying in their degree of orthographic (O), phonological (P), and
semantic (S) overlap with Dutch words. On the basis of the collected data, Dijkstra
et al. (1999) concluded that languages are accessed in a nonselective manner and that
7 The notion of the neighbourhood was introduced in Sect. 2.5.3 of the previous chapter, where
competition between the activated words constituting the fundamental tenet of the cohort model
was discussed.
WWW.EBOOK777.COM
3.4 Modelling Bilingual and Multilingual Processing 87
orthographically identical cognates and interlingual homographs can be identified faster

than matched controls because they share lexical and sublexical orthographic representa-
tions across languages. This sharing leads to stronger activation of the orthographic rep-
resentations during recognition and therefore to faster RTs. (Dijkstra et al. 1999, p. 511).
In yet another study, Lemhöfer and Dijkstra (2004) found that cognates were rec-
ognized faster than the matched English and Dutch controls. Because at the same
time the homographs (having an identical orthographic form across languages)
showed no effects (relative to Dutch controls), the effect for cognates seemed to
be determined at least by the fact that their meanings overlap across languages.
The obvious interpretation was that the cognates’ semantics in both languages
must have been co-activated. Also, the data coming from cross-linguistic priming
and repetition effect tasks (cf. Woutersen 1997; Jiang and Forster 2001; Basnight-
Brown and Altarriba 2007) clearly support the nonselective access view. Primed
lexical decision tasks revealed that semantically related words prime each other,
even when prime and target appear in different languages (cf. Duyck 2005).
Lexical decisions were facilitated by cross-linguistic orthographic and semantic
similarity relative to control words that belonged only to English.
Overall, the research results in multilingual word recognition demonstrate a
considerable amount of interaction between the languages known by a multilin-
gual, which has led the majority of researchers to believe that lexical access in
multilinguals is basically nonselective with respect to language (Caramazza and
Brones 1979; De Groot et al. 2000; Kroll and Dijkstra 2002; De Groot 2013b),
automatic (i.e., not under control of the reader/listener), and, although task
dependent, its first processing stages might remain unaffected by nonlinguistic
factors (cf. Dijkstra 2003b). In this view, lexical word representations from both
(or more) languages are activated even in situations where only one language
is relevant. Consequently, the group of word candidates that compete for selec-
tion within the word recognition or production process is not restricted to one
language.
Importantly, it has to be noted that all the above mentioned experiments con-
cern out-of-context word recognition performance. The empirical data reviewed
in the previous paragraphs indicate that language nonselectivity is a compelling
feature of this type of recognition. This means that word candidates form differ-
ent languages initially become active on the presentation of a letter string. This
nonselectivity seems to hold for all representations that characterize words (e.g.,
orthographic, phonological, and semantic codes). Further, multilingual word
recognition also seems to be automatic in the sense that the process takes place
relatively unaffected by nonlinguistic contextual factors. This applies not just to
words from the native language (L1), but also to words from the L2. At the same
time, De Groot (2013) raises a valid point when she remarks that some research
is still needed to verify whether language nonselectivity is maintained or elimi-
nated in context since the empirical data gathered thus far show that when words
are processed in sentence context, their processing seems to be sensitive to the
semantic and syntactic aspects of the sentence (cf. Hartsuiker et al. 2004, 2008; in
Riehl 2010).
WWW.EBOOK777.COM
3.4.2 Task Dependent Access
As has been pointed out, there is a growing amount of experimental studies

supporting the view that lexical access is nonselective under many circum-
stances. However, some researchers point to the possibility that even if access
to the identification system is basically nonselective in nature, particular cir-
cumstances might allow it to operate in a language selective way (cf. Dijkstra
2005). In fact, there is solid evidence in the literature that task demands can
influence multilingual performance to a significant extent. By way of example,
in the experiments carried out by Dijkstra and Van Heuven (2002a) the inform-
ants appeared to be extremely sensitive to small variations in task demands and
the composition of the word lists. Many researchers even claim that it may be
inappropriate to talk about multilingual word recognition in general without
specifying the precise task and experimental circumstances under which it takes
place because performance is both task and context dependent (cf. De Groot
et al. 2000; Dijkstra et al. 2000; Dijkstra et al. 1998; De Groot et al. 2002).
A highly influential theoretical approach that assumes that the relative acti-
vation state of words and languages is context sensitive is Grosjean’s Language
Mode Hypothesis. The language mode framework proposed by Grosjean (1982,
2001, 2008, 2010), presupposes integration of languages into a single system and,
in the same time, predicts that language processing mechanisms and languages as
a whole can be active to different extents. According to Grosjean, the activation of
languages in the mind may be influenced by the context of interaction (topic, level
of formality, communicators). Grosjean (2001, p. 3) states that,
at any given point in time and based on numerous psychological and linguistic factors, the
bilingual has to decide, usually quite unconsciously, which language to use and how much
of the other language is needed – from not at all to a lot.
This means that the degree of cross-linguistic influence is determined by the

monolingual/bilingual mode of the speaker. The mode in which the speaker
operates is, in turn, influenced by the context of interaction in its broad mean-
ing. Hence the language mode hypothesis is capable of accommodating the
effect which the context has on cross-linguistic influence. However, it seems in
place here to mention that Dijkstra and Van Hell (2003) tested the language mode
hypothesis with trilinguals and found strong evidence against it, concluding that
it is impossible for multilinguals to operate in an exclusively monolingual mode
(Dijkstra and Van Hell 2003).
3.5 Models of Bilingual Visual Word Recognition
The previous chapter proposed a comprehensible review of monolingual word

recognition models. One of them viewed word recognition as a list of words that
are searched in serial order to identify the candidate most consistent with the
WWW.EBOOK777.COM
3.5 Models of Bilingual Visual Word Recognition 89
perceptual input (cf. Forster 1976, 1989; Sect. 2.6.1). Another advocated the view
of a set of word “detectors” that compete to collect evidence that their word was
present in the input (cf. Morton 1969; in Singleton 1999; Sect. 2.5.2). Both theo-
ries have been criticized for being little more than verbal descriptions and more
recently they have been replaced by more realistic computational models which
“force clarity on theories because they require previously vague descriptive
notions to be specified sufficiently for implementation to be possible” (Thomas
and Van Heuven 2005; cf. Sect. 2.5.4).
Within the bilingual domain, a number of computational models have been
used. Notably, many computational models of bilingual word comprehension
have worked within the connectionist tradition. In fact, bilingual researchers have
appealed to two different types of connectionist models in studying processes
of comprehension: localist and distributed. Localist models are seen as direct
descendants of the original word detector models proposed in the 1970s. These
models do not incorporate change according to experience, their focus within mul-
tilingual research has been to investigate the static structure of the word recog-
nition system in the adult multilingual. One good example of a localist model is
the Bilingual Interactive Activation [BIA] model and its revised version known
as BIA+. Another important representative of this group of models is Semantic,
Orthographic, and Phonological Interactive Activation [SOPHIA] model (Dijkstra
and van Heuven 2002). The Bilingual Interactive Model of Lexical Access
[BIMOLA] (Léwy and Grosjean 1997, 2008) also belongs to the group of localist
models but focuses on spoken word recognition.
Distributed models, on the other hand, tend to represent individual entities (like
words) as patterns of activity spread over sets of units. The entity represented by a
network cannot, therefore, be identified by looking at a single unit, but only as a
code over several units. Consequently, they are able to account for both language
independence and language interaction within a single network. More interest-
ingly, distributed models tend to focus on experience-driven change, specifically
on learning to map between codes for different types of information (such as a
word’s spoken form and its meaning). Since these models incorporate changes
according to experience they are often applied to issues of language acquisition
and change in language dominance over time. Some prominent examples of the
distributed developmental models are: the Bilingual Single Network [BSN]
(Thomas 1997); the Bilingual Simple Recurrent Network [BSRN] (French 1998)
or the Self-Organizing Model of Bilingual Processing [SOMBIP] (Li and Farkaš
2002), and its later version known as the Developmental Lexicon Model [DevLex]
(Li et al. 2007).8
Importantly enough, the distinction between the localist and distributed mod-
els is not a dichotomous one, but rather a continuum. According to Van Heuven
and Dijkstra (2010, p. 105), although localist and distributed models have different
advantages for studying various phenomena of bilingual language processing, the
8 For a thorough discussion of the computational bilingual models (see Warren 2013,
pp. 231–235).
WWW.EBOOK777.COM
characteristics of these models must eventually be combined. A final model must

reflect both how the bilingual system is acquired as well as details of its process-
ing dynamics in the adult state.
In the following paragraphs, bilingual localist-connectionist models that focus
on different aspects of lexical recognition will be reviewed along with the support-
ing evidence. One—the BIA model—explores the way in which the orthography
of the written language may be shared (Dijkstra and Van Heuven 1998) whereas
its successor—the SOPHIA model—contains also phonological and semantic rep-
resentations (Van Heuven and Dijkstra 2001). In turn, the BIA+ (Dijkstra and Van
Heuven 2002a, b), which in fact constitutes a revised version of the BIA model,
and the Inhibitory Control Model (Green 1986, 1998) consider the importance of
the demands posed by different tasks and include a language-external control sys-
tem. It needs to be stressed that, since the focus of the empirical part of the pre-
sent work is placed on visual word recognition, the models described below were
specifically selected to characterize only this aspect of processing. Consequently,
models referring to speech production and recognition will not be accounted for.
3.5.1 The Bilingual Interactive Activation Model (BIA)
The Bilingual Interactive Activation Model (BIA) by Dijkstra and Van Heuven
(1998) is an implemented localist-connectionsit model (processing units are separate
and unitary nodes in a network) that extends the McClelland and Rumelhart (1981)
Interactive Activation Model (IA; cf. Chap. 2) to the bilingual case. Similarly to
monolingual word recognition, processing is assumed to be initiated bottom-up upon
presentation of visual input and nonselectively so that all information resembling the
input is activated. The BIA differentiates between four hierarchically organized levels
of different linguistic representations: letter features, letters, words, and language tags
(or language nodes). The model hypothesizes that all lexical representations from the
same language are linked bi-directionally to a single language node at a superior rep-
resentational level. This means that activation from letters spreads to words and then
to language nodes and vice versa. Consequently, the activation levels of the language
node can influence representations at the word level. As depicted in Fig. 3.8, when a
string of letters is presented to the model, it activates features at each letter position,
which subsequently leads to the activation of letters possessing those features, and,
simultaneously, to the inhibition of other letters. Thus activated letters excite ortho-
graphic forms of both languages which contain those letters in appropriate positions,
which in turn leads to competition among words sharing overlapping or identical
orthography in the bilingual’s two languages (e.g., cognates). In Fig. 3.8 such excita-
tory connections are indicated by arrows, with arrowheads pointing in the direction
of activation spread. To allow for language selection, the model introduces a sepa-
rate level of representation within the lexicon—the level of language nodes. It is at
this level that words from the non-selected language are inhibited in such a way that
the inhibition increases the likelihood of selection of the intended language words,
WWW.EBOOK777.COM
Fig. 3.8 The bilingual
interactive activation model
(adaptation based on Dijkstra
and Van Heuven 1998,
p. 207)
which, however, does not affect the initial activation of words in both languages.
In Fig. 3.8 inhibitory connections are indicated by lines with closed circles.
Extensive evidence for the BIA model can be found in studies in which aspects
of word type have been experimentally manipulated, like in language-specific and
language-general decision tasks. A growing number of recent studies using this
methodology have provided support for the claim that lexical access is nonselec-
tive and that bilinguals cannot help but respond as if information in both languages
was active. Kroll and Dussias (2004) offer a comprehensive review of these stud-
ies which include the use of interlingual homographs (cf. Dijkstra and Van Heuven
1998; De Groot et al. 2000, Smits et al. 2006), cross-language neighbours (cf.
Jared and Kroll 2001; Kerkhofs et al. 2006; Dijkstra et al. 2010) and, most impor-
tantly for the present work, cognates (cf. Dijkstra and Van Heuven 1998; Van Hell
and Dijkstra 2002; Kroll and Sunderman 2003; Hall et al. 2009).
Despite substantial and convincing evidence supporting the validity of the model,
certain doubts concerning some aspects of bilingual word recognition that are not
fully accounted for by Dijkstra and Van Heuven’s proposition have been raised by
the authors themselves. Notably absent in the BIA model is the representation of
semantics. The assumption at this level of analysis is that lexical form properties
of words in the bilingual’s languages are activated in a bottom-up fashion and only
WWW.EBOOK777.COM
later in processing does the output of the lexical identification system interact with
semantics and higher-level context. Another line of criticism concerns the lack of
phonological representation, which explains why the model fails to cater for results
of between-language phonological priming found e.g., by Brysbaert et al. (1999)
or Brysbaert and Dijkstra (2006). Most importantly, the model did not account for
task and context effects, although there is clear evidence in the literature that task
demands can influence bilingual performance to a considerable extent (cf. De Groot
et al. 2000). The more recent variations of the BIA model: SOPHIA and later BIA+
seek to fill in some of these theoretical gaps (cf. Sect. 3.5.3).
3.5.2 The Inhibitory Control Model
As indicated above, the BIA model failed to incorporate the distinction between the
actual word identification system and a task/decision system. It the light of numer-
ous studies confirming that bilingual word processing often depends on the task that
is to be performed and the context in which it is performed, this is a rather obvious
shortcoming. Indeed, it seems impossible to talk about multilingual word recogni-
tion in general without describing the conditions under which it takes place and the
goal that needs to be achieved. The Inhibitory Control Model (IC) of bilingual word
recognition offered by Green (1986, 1998) and presented in Fig. 3.9 successfully
meets these ends. It accounts for the fact that bilingual language processing always
takes place within a particular task context and with certain goals in mind. As the IC
model has been more often applied to bilingual language production and translation
tasks than to visual word recognition processes, the following discussion will be lim-
ited to those aspects of the model which are of importance for the experimental part
of present text, namely task dependency.
Fig. 3.9 The inhibitory
control model (adaptation
based on Green 1998)
WWW.EBOOK777.COM
The topmost part of the model and its key element is the “conceptualizer”,
which is language independent and builds conceptual representations on the basis
of information stored in the long-term memory. These representations are driven
by the “goal”, i.e. an intention to produce a word in a particular language. The
supervisory attentional system (SAS), which is the main element of the model,
along with the lexico-semantic system and a set of language task schemas, medi-
ate the planning. For Green (1998, p. 72) the term schema refers to “mental
devices or networks that individuals may construct or adapt on the spot in order
to achieve a specific task and not simply to structures in long-term memory”. In
various language tasks different schemas (e.g. translation schemas or word pro-
duction schemas) compete for output. Once the schema is specified, it is retrieved
from memory and modified if the task demands it. A given language task schema
regulates the outputs from the lexico-semantic system whose representations fol-
low Levelt’s (1989) and Levelt et al.’s (1999) model where each lexical concept
is associated with a lemma containing syntactic specification of the concept.
Importantly, each lemma is tagged for language and the language tag is also a part
of the conceptual representation. In order to achieve a linguistic goal, a particular
task schema is activated under the control of SAS. Next, the task schema regulates
activation and suppression of tags at the lemma level. Activated lemmas send acti-
vation to the associated word forms in a given language and, finally, output in the
appropriate language is produced.
Despite its numerous merits, Green’s model is not free of certain shortcomings,
one of them being the fact that it does not fully account for the way the recogni-
tion of items within the lexico-semantic system takes place. It is worth mention-
ing that this criticism points to an aspect that the BIA has also failed to address,
namely a more precise description of the interactions at the sub-lexical level,
including some phonological representations.
3.5.3 SOPHIA and the BIA+ MODEL
Due to its similarity, the common basis and the same developers, the last two
models will be discussed together. As already stated in Sect. 3.5.1, an important
problematic issue criticized in the BIA model was the lack of semantic and pho-
nological levels of representation. These gaps have been addressed by Van Heuven
and Dijkstra (2001) in the semantic, orthographic and phonological interactive
activation model (SOPHIA). Accordingly, the architectural structure of SOPHIA
depicted in Fig. 3.10 incorporates the BIA system but also includes levels for
semantics and phonology. More specifically, nodes at one particular level can
activate and inhibit units in neighbouring levels via excitatory and inhibitory con-
nections, respectively. Moreover, representations at a particular orthographic or
phonological level mutually inhibit each other via lateral inhibition. However, the
connections between an orthographic level and its analogous phonological level
are of an excitatory nature only (i.e. they do not inhibit each other). For instance,
WWW.EBOOK777.COM
Fig. 3.10 The semantic,
orthographic, and
phonological interactive
activation (SOPHIA)
model of word recognition
(adaptation based on Van
Heuven and Dijkstra 2001)
the presentation of the visual stimulus ‘cat’ will activate its orthographic word
node for ‘cat’ and the corresponding phonological node [k t].
Another alteration introduced to SOPHIA is the inclusion of a semantic system
which can directly interact with remaining structural levels in the model. Finally,
and most importantly, another modification to the SOPHIA model concerns the
language nodes. In the BIA word nodes were connected to their corresponding
language node via excitatory pathways. In turn a language node exerted inhibitory
influence on all the word nodes from the opposing language. The latter inhibitory
connections have been removed from SOPHIA.
SOPHIA did not, however, resolve all the problems identified in the BIA,
in particular it did not cater for lack of context and task demands system.
Consequently, within the new version of the BIA model—BIA+, Dijkstra and
WWW.EBOOK777.COM
Fig. 3.11 The BIA+ model

and Van Heuven 2002a)
Van Heuven (2002a) proposed two clearly separated systems—a word identi-

fication system, which incorporates the BIA system with additional levels for
semantics and phonology, and a task schema system. Importantly, the informa-
tion proceeds from the word identification system toward a task/decision system
without any influence of this task/decision system on the activation state of words,
as shown in Fig. 3.11. In other words, unlike in the BIA model, where the identi-
fication system does not allow language nodes to feed activation back to the word
level, which excludes the influence of task demands or context on activation lev-
els (cf. Fig. 3.8), the only function of language nodes in the BIA+ is to specify
which language lexical items belong to. Importantly, the effect of context, task
demands, and other nonlinguitsic factors is dealt with within the task schema part
of the model. Dijkstra and Van Heuven (2002a, b) added that part to the model as
they state that nonlinguistic context effects arising from instruction or participant
expectancies can affect the way information from the word identification system
is used, but not the activation state of word candidates.
Of importance for the practical chapters of the present work seems to be the
fact that the BIA+ postulates that language information becomes available rather
late and provides only limited feedback from language nodes to the lexical level;
hence language membership can only have limited influence on word recognition
(Dijkstra and Van Heuven 2002a, p. 186).
WWW.EBOOK777.COM
3.6 Models of Multilingual Visual Word Recognition
Until recently few researchers in psycholinguistics made the step from

bilingualism to multilingualism and the majority of models limited themselves to
the processing of two languages. However, of late there has been a growing inter-
est related to the simultaneous processing of more than one foreign language.
Notably, although multilinguals are generally not considered to be extended bilin-
guals, models of multilingual lexical processing are typically extensions of respec-
tive monolingual and bilingual models and hypotheses. Multilingual adaptations
of the Levelt’s Blueprint for the Speaker: de Bot’s Multilingual Processing Model
(2004), Müller-Lancé’s Connective Model (2003) and Li Wei’s model of multilin-
gual lemma activation (2003, 2006) are the most prominent examples of such a
situation. However, since all the above mentioned models are above all production
models and the main focus of the present work rests in visual word recognition,
they will not be referred to in the present section. Instead, in what follows, a more
detailed description of Dijkstra’s Trilingual Interactive Activation (TIA) model
will be offered (Dijkstra 2003a).
The TIA model (cf. Fig. 3.12) seems particularly interesting for the present dis-
cussion since it is solely devoted to the processing of written language executed
by multilinguals. Additionally, it assumes a single integrated lexicon “that consists
of a mix of words from the [three] languages” (Dijkstra 2003a, p. 17) and accom-
modates both the influence of language typology and the recency effect. Similarly
to the Bilingual Interactive Activation Model (BIA), its trilingual version consists
of four levels of language representations, namely: feature level, letter level, word
level and language level. The mechanics of the model are similar to those already
described within the Monolingual Interactive Activation Model (cf. Chap. 2), and
its bilingual counterpart, the BIA model (Sect. 3.5.1). The process of word rec-
ognition starts when the feature nodes detect a configuration of visual features
of the input letter string and excite those letters whose features match the input
and simultaneously inhibit the unmatching ones. In Fig. 3.12 excitation is marked
by arrows and inhibition by dotted lines. Similar procedure is repeated between
the letter level and the word level, i.e., letters excite words in which these letters
are present and inhibit those words from which the letters are absent. It is impor-
tant to note that high frequency words and the ones most recently used are recog-
nized faster than the lower frequency ones and the ones which have not been used
recently, since their representations have a higher resting level of activation (the
phenomena traditionally termed as the frequency effect and the recency effect, for
thorough review see Warren 2013).
A crucial notion which is strongly related to the process of word recognition in
the model is the already mentioned neighbourhood effect. This stands for the fact
that, initially, many words are activated; e.g., on the presentation of the English
word wind, other words such as bind, kind, wild, wink, would receive some acti-
vation (cf. Dijkstra 2005). With the inclusion of words from three languages, the
density of words increases and the competition of words for recognition becomes
WWW.EBOOK777.COM
3.6 Models of Multilingual Visual Word Recognition 97
Fig. 3.12 Trilingual
interactive activation model
2003a)
stronger. Consequently, the moment when a letter string is recognized as a particular

word is delayed. However, simulations reported by Dijkstra (2003a) and Dijkstra
et al. (2010) suggest that adding another lexicon to a bilingual lexical system results
in a relatively small change in the processing time (recognition of L1 items is only
30 ms slower in the reported simulation). The possible reason might be that not all
words are affected by the fact that additional items are included in the lexical net-
work. Obviously enough, because foreign language words are probably lower in
subjective frequency, their effect on the native language will be smaller than that
of within-language neighbours. However, competition with L1 words will increase
with the increasing proficiency in the foreign languages.
Another factor which influences the speed of word recognition is the language
distance. Dijkstra (2003a, p. 20) suggests that the mental lexicon, though inte-
grated across languages, is not homogenous as far as lexical density is concerned,
which has an influence on word processing. He states that
[l]earning other language words leads to an increase in the number of neighbours of L1
target words, but this increase is less than in the case of new L1 words, because the other
language words are positioned somewhat more distant in lexical space. As a consequence,
words from more distant languages having the same script will interfere less than words
from closely related languages (Dijkstra 2003a, p. 20).
WWW.EBOOK777.COM
Apart from bottom-up factors that influence the speed of processing (such as the
language-specific cues9), there are top-down factors which affect activation of
items belonging to the languages known to the multilingual speaker. The top-down
factors are twofold, either non-linguistic, such as context, task demands, and stim-
ulus list compositions, or linguistic, such as morphological representations, sen-
tence level information, or language membership information. Strong evidence for
the influence of the linguistic, sentential context on lexical processing has been
found in the monolingual domain but, again, it is not certain if this could be zto
multilingual processing. As far as language membership is concerned, Dijkstra
et al. (1999) found evidence against its facilitative effect. They tested bilingual
participants on a set of Dutch-English homographs embedded in mixed language
lists. It was found that participants were unable to exclude effects from the non-target
language on interlingual homographs identification.
Summing up, it can be seen that word recognition is aided by bottom-up and
top-down factors. Neighbourhood density, storage patterns within the mental lexi-
con, as well as linguistic and non-linguistic factors appear to speed up the pro-
cess of word recognition, which helps multilinguals to cope with the processing
load related to the greater number of items that they have in their mental word
stores. The model is of special importance as the research findings will be ana-
lyzed within its framework.
3.7 Conclusion
Whereas the vast majority of accounts on the mental lexicon of foreign language
users have been concerned mainly with the processing of two languages, the pre-
sent chapter has gone beyond this, paying special attention to the interaction of more
than two languages and thus offering theoretical basis to the multilingual analysis
described in Chaps. 4 and 5. As has been demonstrated throughout the chapter, the
issue of multilingual storage is the subject of a growing range of empirical studies
on the basis of which many models and hypotheses have been developed. Some of
them argue for the separate and others for the shared organisation of the multilingual
mental lexicon. Successful, in part, seems to be an attempt to resolve this conflict
made by those psycholinguists who opt for the mixed structure of the multilingual
store. It has also been shown that lexical storage and access depend on a number of
different variables both user- and word type-determined.
The present chapter was divided into two main parts. The first of these con-
sidered multilingual mental representation theories with the special focus on the
separate/shared dichotomy. The second part was devoted to the presentation and
9 Itis suggested that the recognition might be facilitated by language specific cues such as dia-
critical markers in French or Polish words or onset capitals in German nouns. However, Kroll
and Dijkstra (2002) admit that this aspect of processing awaits further research as evidence for
the aforementioned facilitation is still limited.
WWW.EBOOK777.COM
3.7 Conclusion 99
discussion of multilingual processing with a special focus on visual recognition. The

chapter provided an overview of models of bilingual and multilingual representation
and processing. It needs to be noted that this review was necessarily selective mainly
due to the length limitations of the present text and its primary focus on the visual
recognition which is why production models have not been included.
The research reported in the next chapter focuses on the role of one word type
variable, namely cognate status. Since the language affiliation of cognates is
ambiguous, they are of special interest for designing a model of a multilingual lex-
icon. Indeed, cognates have been the most important stimulus material in studies
attempting to unravel the processing of bilingual and multilingual word recogni-
tion. Therefore, Chap. 4 reviews the evidence for the special status of cognates,
focusing on the cognate facilitation effect. It also reports on two cognate identifi-
cation studies, the results of which has been deployed in the studies discussed in
the final chapter of the present work.
WWW.EBOOK777.COM
Chapter 4
Visual Word Recognition in Multilinguals
4.1 Introduction
Now that the main issues concerning the multilingual mental lexicon storage and
retrieval have been delineated, it is time to outline the most essential aspects con-
cerning visual word recognition in multilinguals. Even a brief inspection of articles
on visual word recognition reveals that three groups of words typically exploited
in this kind of research are interlingual homographs, interlingual neighbours and
cognates (cf. Sect. 3.4). Through such words, a wealth of studies in the last decade
have confirmed that during the initial stages of word identification by multilinguals,
word candidates from several languages are often co-activated. In accordance with
these results, several word recognition models propose that multilingual word rec-
ognition involves an initial language nonselective access into an integrated lexicon.
The present chapter focuses on the last category of words which have also been
used in the empirical experiments described in the final chapter of the book. Cross-
language cognates are of special interest for designing a model of the multilingual
lexicon, because there is a possibility that at least part of their representations is
shared between two languages. The cross-linguistic form overlap of cognates has
also been used by researchers to explore whether words from different languages
are co-activated during the reading, listening, and speaking of multilinguals. If
responses to such “special” items differ from those to language-specific control
items, this can be seen as evidence that both readings of the cognates have become
active and affect each other. A number of experimental bilingual, and more
recently multilingual, studies have demonstrated a substantial difference in the
processing of cognates and monolingual control words (cf. Caramazza and Brones
1979; Dijkstra et al. 1998; Dijkstra et al. 1999; De Groot et al. 2002; Lemhöfer
and Dijkstra 2004; Lemhöfer et al. 2004, 2008; Van Heuven and Dijkstra 2010;
Van Heuven et al. 2011). In all these studies, lexical decisions on cognates were
faster and/or more accurate than those on control words. Moreover, cognates have

DOI 10.1007/978-3-642-32194-8_4
WWW.EBOOK777.COM
102 4 Visual Word Recognition in Multilinguals
been found to be translated faster (cf. Sanchez Casas et al. 1992; De Groot et al.
1994), to be more effective as masked or unmasked primes in visual lexical deci-
sion than noncognate translations (cf. Cristoffanini et al. 1986; Gollan et al. 1997),
to be easier to learn (cf. De Groot and Keijzer 2000), and to be named faster in
word naming (cf. De Groot et al. 2002) or in picture naming (cf. Costa et al. 2000;
Cieślicka and Kujałowicz 2006; Hoshino and Kroll 2008). Interestingly, the cog-
nate effect has also been observed in bilinguals using different scripts (cf. Gollan
et al. 1997; Kim and Davis 2003; Hoshino and Kroll 2008). Studies using electro-
physiological and neuroimaging methods (cf. De Bleser et al. 2003; de Bruin et al.
2013) also indicate that cognates are processed differently from controls.
Much of the present chapter offers theoretical background to the empirical
research documented in Chap. 5 by reviewing evidence for the special status of
cognates. First, the chapter will present word stimuli used in visual word recog-
nition studies with a special focus on the terminology and typology used in con-
nection to cognates viewed from the psycholinguistic perspective. Second, special
attention will be given to factors affecting visual word recognition in bilinguals
elaborating on their mutual interaction and its role in the visual recognition. Here,
much space will be devoted to the cognate facilitation effect, the essence of which
is that bilinguals produce and recognize cognates faster than noncognates. Further,
factors affecting cognate recognition will be delineated together with models of
the bilingual lexicon accounting for the cognate facilitation effect. Next, a brief
presentation of the more recent line of studies devoted to bilingual word recog-
nition in a sentence context will follow. In the final subchapter the effect of the
cognate status on the foreign language vocabulary learning will also be explicated.
4.2 Word Stimuli Used in Visual Word Recognition Studies
A vast number of research studies addressing the problem of language selective

or language nonselective visual word recognition has been based on the belief
that the similarity of form and meaning affects word recognition and will help
to explain patterns of mutilingual storage and retrieval. Consequently, the three
types of word stimuli that have been typically used by researchers are: interlingual
homographs, interlingual neighbours and cognates. The following subsections will
be devoted to the presentation and discussion of these three groups of words. It
needs to be noted that both homographs and neighbours were initially exploited in
the multitude of studies within the monolingual domain and only later have been
adopted by researchers working in the bi- and more recently multilingual contexts.
4.2.1 Interlingual Homographs
Interlingual homographs are words that exist in two languages but their meanings
are different in these languages. In other words, they share the written or spoken
WWW.EBOOK777.COM
4.2 Word Stimuli Used in Visual Word Recognition Studies 103
form (homographs and homophones, respectively) but not the meaning. Many
studies have employed interlingual homographs to confirm the nonselectivity of
the multilingual lexical retrieval processes. It has been believed that if the inter-
lingual homographs and their matched controls are recognized in the same time
it means that the other meaning of the homograph is not activated which in turn
means that the lexical access is selective. If, however, the interlingual homograph
is recognized faster than its matched control it means that the access is nonselec-
tive. Such a difference in decision times and error rates between the homographs
and control words is commonly referred to as “the homograph effect”.
One of the first works in which interlingual homographs were used was
Beauvillain and Grainger’s primed lexical decision task study. The study exploited
French words to prime English targets. The results showed that a priming effect
occurred in the case of homographic primes which proves that both meanings of
a homograph are activated nonselectively. However, considering the outcomes of
the study in the light of Grosjean’s language mode hypothesis (e.g. 1982, 2008), it
must be noted that Beauvillain and Grainger’s research did not give clear evidence
about the nonselective access. According to Grosjean’s theory, both language sub-
systems were activated since the stimulus material included items from both lan-
guages. Consequently, later studies were unilingual, which means that the research
material came from only one of the bilingual’s languages. The basic question the
studies tried to address and ultimately managed to confirm was whether both lan-
guages can be activated if the experimental material from only one language is
used (e.g. Dijkstra et al. 1998).
As far as the factors affecting the homograph effect are concerned, the most
influential ones include task dependency, level of proficiency and word frequency.
Indeed, it has been confirmed that the type of task influences the direction of the
homograph effect. In a language neutral decision task it is facilitatory as responses
to homographs are faster, whereas in a language specific task the effect is inhibi-
tory, which means that responses to homographs take longer to occur (cf. De
Groot et al. 2000). Further, many studies using interlingual homographs in the
lexical decision task (Jared and Szucs 2002) confirm that in unbalanced bilinguals
the stronger language can be immune to an influence of the weaker language. Also
the frequency of the selected interlingual homographs seems to be an important
factor as reviewed in De Groot et al. (2000). All these examples clearly show
that the occurrence of the homograph effect is not that easy to predict and if the
answer to the question whether language access is selective or not was only based
on the research exploiting interlingual homographs, it would not be unequivo-
cal. Instead it depends on a number of factors such as the experimental material
(Dijkstra et al. 1998), prior activation of the non-target language (Jared and Szucs
2002), the type of task (Dijkstra et al. 1998), the proficiency level (Jared and
Szucs 2002) or, even the linguistic context (Elston-Güttler et al. 2005). Finally, it
should be noted that the effects obtained with interlingual homographs and homo-
phones have been far less consistent than those observed for cognates (cf. sect.
4.2.3). As De Groot (2013, p. 82) observes they do not rule out either one of the
two theoretical positions regarding bilingual word recognition (selective vs. non-
selective access).
WWW.EBOOK777.COM
4.2.2 Interlingual Neighbours
Another group of words typically used in the visual recognition studies are inter-
lingual neighbours. In the research on monolingual visual recognition it was
proved that orthographically similar words are already activated on the presenta-
tion of only one of those words (Andrews 1989). Importantly, the time necessary
to recognize a given word depends on the number and frequency of its ortho-
graphic neighbours—those with more neighbors are recognized slower, those
which do not have many neighbours are, in turn, recognized faster.
In 1992 Grainger and Dijkstra conducted one of the first bilingual studies using
the interlingual neighbours and proved that the neighbourhood effect also works
beyond languages. The study by Van Heuven et al. (1998) confirmed that the
recognition of target words in L1 and L2 is affected by similar word forms both
within the target language and within the other language. The authors not only
noticed that the identification latency for L2 words depended on the number of L1
neighbours, but also that the identification latency for L1 words was influenced by
the number of L2 neighbours, which was interpreted as suggesting that the influ-
ences of L2 on L1 visual word recognition work at proficiency levels lower than
generally assumed. It needs to be noted that, so far, the interlingual neighbourhood
effect was confirmed for studies using many different types of tasks and paradigms
(e.g., lexical decision or priming tasks).
4.2.3 Cognates
The third group which will be discussed in more detail are cognates. Within the
three types of word stimuli used in the visual word recognition studies, cognates
constitute a very special group. The traditional definition of cognates has its roots
in historical linguistics which sees them as lexical units (words) in two or more
languages that are “part of a relation defined in terms of direct descent from a
common word (or morpheme) belonging to a given shared ancestral language”
(Carroll 1992, p. 100). In other words, in the traditional linguistics approach,
cognate pairs are to be found in etymologically related languages. They are lexi-
cal items of similar form (which can be explained in terms of regular phonologi-
cal change in each language) and having the same (or similar) meaning (e.g.,
German Hand and English hand, German trinken and English drink, German rot
and English red, etc.). Due to their common etymological origins, cognate pairs
are very important in historical linguistics. By comparing the pronunciation of
etymologically related words, linguists can postulate tendencies in the change
of the phonological properties of words from the common ancestor language to
the present-day languages of the language family. In view of the above, cognates
can only occur naturally in languages with some shared roots. Interestingly, in
unrelated languages, some “cognates” may also exist (cf. Kim and Davis 2003;
WWW.EBOOK777.COM
cf. Sect. 4.4.1) but strictly speaking they are borrowings: words taken from a for-
eign language, usually to fill a newly-formed linguistic need as a result of cultural
contact or, increasingly, technological development, and often phonologically
or orthographically adapted into the language, such as tişört (Turkish, T-shirt)
and futball (Hungarian, football). Such words are definitely quite different from
English wolf and German Wolf, where both the language-specific forms have
evolved from a common origin and long been part of the language.
However, considered from the point of view of lexical storage and processing
as well as language acquisition, the linguistic definition of cognates based on ety-
mological bonds is more than unsatisfactory. Perceived from the psycholinguis-
tic perspective, that is, how and why language users or learners identify words
as being the “same” in two languages, etymological relatedness is “neither a suf-
ficient, nor a necessary condition” (Carroll 1992, p. 98). Indeed, it seems highly
unlikely for this information to be encoded in the lexical representation of a word
in our mind. Instead, metalinguistic information, like the linguistic history of a
word, is postulated to be represented in memory separate from lexical knowledge
and is not supposed to be “part of the automatic processing of structural informa-
tion in word activation” (Carroll 1992, p. 103). Thus, in psycholinguistic literature,
cognates are any two words with shared aspects of spelling, sound, and meaning
across two or more languages, whereas cognate-pairing is usually described as
a form of automatic cross-linguistic activation of lexical addresses that have the
same or similar formal properties in two languages (ibid.). To recapitulate, in lin-
guistics, cognates are often defined as words that share a common etymological
origin. Within psycholinguistic research, however, focus is given not to etymo-
logical relations but to language processing, since only aspects of the underlying
mental representation (orthography, phonology, and semantics) are hypothesized
to affect processing. Historical linguistic roots of a word are not represented in
the mind of the language user, and thus the linguistic definition of cognates based
on etymology does not capture the form-meaning overlap that psycholinguists as
well as language practitioners are interested in (cf. Carroll 1992; Ringbom 2007).
Accordingly, in the present publication, which seeks to explore issues related to
multilingual storage and processing—both independent of metalinguistic informa-
tion—the psycholinguistic definition will be employed.
As described in previous sections, during lexical processing several units can
be activated simultaneously in the mental lexicon. In multilinguals, formally
similar lexical items in two or more languages can also receive parallel activa-
tion (cf. Grainger and Dijkstra 1992). Apparently, in the same way as formally
similar words in one language (e.g., stay, steak, stage) can prime each other (that
is, they make each other more easily accessible), interlingual cognate pairs (e.g.,
G. singen—E. sing; G. blau—E. blue) seem to prime each other between languages
(cf. Costa et al. 2000; Costa et al. 2005). Moreover, lexical items in two languages
do not need to be completely identical in form for neighbourhood activation. On
this account, research most often reports on cross-language cognates, meaning both
formally identical or almost identical words in two languages (cf. Beauvillain and
Grainger 1987; Gerard and Scarborough 1989; Sect. 4.4.1). Consequently, many
WWW.EBOOK777.COM
linguists postulate that cognates can be considered as simply cross-linguistic ortho-

graphic and/or phonological neighbours (cf. Dijkstra 2005; Ringbom 2007).
Another term used in reference to cognates is that of interlingual homophones/
homographs. Some psycholinguists distinguish between cross-linguistic homo-
phones/homographs with complete, partial or no semantic overlap. Cross-
linguistic homographs are cognate pairs that have identical or almost identical
orthographic properties in two languages (e.g., German Lust and English lust;
German Land and English land; German Milch and English milk; German Sturm
and English storm), whereas cross-linguistic homophones are cognate pairs that
have almost identical phonological properties (e.g., German Igel meaning “hedge-
hog” and English eagle1; both pronounced in a very similar way). It seems crucial
to realize, however, that cross-linguistic homographs/homophones tend to be used
to describe words of similar form but different meanings and are also known as
homographic/homophonic noncognates.
Another issue to be discussed in this section concerns terminology used in rela-
tion to formally similar words with various degrees of semantic overlap. In the
light of the semantic resemblance criterion, words of similar form can be divided
into true and false cognates. It needs to be noted that a wealth of research, espe-
cially in the fields of lexicology and lexicography as well as in foreign language
teaching, have focused on the latter group of formally similar but semantically dif-
ferent words from two or more languages. There is a plentitude of labels used in
relation to such words. In French writings the familiar term faux amis (Koessler
and Derocquigny 1946; Koessler 1975) is used, whereas deceptive cognates
(Lado 1957), deceptive words (Wełna 1977), misleading cognates (Taylor 1976)
or treacherous twins (Kirk-Greene 1981) are typically found in English, though
the French faux amis is often readily borrowed or translated into false friends (cf.
Kirk-Greene 1981). Indeed as for the terminology applied in relation to cognates,
there is no denying the fact that in the literature on lexical units of different lan-
guages which have identical or similar form, a variety of terms are in use.
As regards the general typology of cognates based on the semantic overlap,
Table 4.1 presents the categories referred to by Carroll (1992). A short description
of each type is offered in the middle column of Table 4.1.
A large number of language teaching publications distinguish between cross-
lexical homophones/homographs that share most of their semantic components
and can be attributed to the same concept (true cognates), and cross-lexical homo-
phones/homographs that share no or only a few semantic components and thus are
not attributable to the same concept in most contexts (false cognates). Similarly
to the typology described by Carroll, false cognates can be divided into two fur-
ther groups in terms of the degree of semantic overlap. Cognate pairs that share
some meaning properties and can, in certain contexts, be considered as translation
equivalents, but not in their prototypical meaning instance (e.g., German Land and
1 The examples taken from Sherkina (2003).
WWW.EBOOK777.COM
Table 4.1 Types of cognates (based on Carroll 1992)

Type of cognates Definition Examples
True cognates Words which are etymologically German Buch and English book;
related and whose semantic German braun and English brown
properties completely or almost
completely overlap
Deceptive cognates Words which are etymologically German Land and English land
related and whose semantic or German Frieden and English
properties partially overlap. These freedom
cognates are either no longer
translation equivalents, or only in
certain contexts, but they still share
some features of meaning
False cognates Words which are etymologically German pregnant and English preg-
related and whose semantic nant; German sensibel and English
features no longer overlap sensible
Accidental Words which have no obvious German Rock meaning skirt and
cognates etymological relationship and do English rock
not share any features of meaning,
but which nevertheless bear much
formal resemblance
The examples that are presented above were not taken from Carroll (1992). The original exam-
ples were replaced by German-English cognate pairs
English land or German Frieden and English freedom), are called deceptive false
cognates, whereas word-pairs that share no semantic features and cannot be trans-
lation equivalents in any context are typically referred to as accidental (or formally
determined) false cognates (e.g. German Rock meaning skirt and English rock).
In the above categorization etymological relationship, or the lack of it, is still a
fundamental aspect in the definition of cognate types. In cognitive psycholinguis-
tics, however, the genetic relatedness of words is not taken into account and the
type of the cognate-pairing is determined by the differences in the mental repre-
sentations of formally similar words and the nature and degree of formal resem-
blance. Indeed, the speed and accuracy of cognate recognition depends, above all,
on their cross-linguistic form similarity. Hence in terms of the nature of the simi-
larity of form, psycholinguistic typology concentrates on orthographic and pho-
nological cognates. There are many studies that have examined the influence of
phonological and orthographic overlap of cognates in comprehension tasks. As
Browne (1982) rightly observes, visual accessing of the lexicon can give rise to
different cognate pairs from those arising from auditory processing. Moreover,
there is now quite some evidence that visually presented words are not recognized
on a visual basis alone. Readers rely on the spoken—phonological—informa-
tion enclosed in the orthographic stimulus when they identify written words, and
researchers no longer quarrel about whether phonology is involved in visual word
recognition but to what extent it is involved (cf. De Groot 2013a).
Another distinction important for lexical recognition research is the one based
on the degree of similarity (identical vs. non-identical cognates). For instance,
WWW.EBOOK777.COM
Dijkstra et al.’s (2010) study, more fully described in Sect. 4.4.1, analyzed how
cross-linguistic similarity affects Dutch-English identical and non-identical cog-
nate recognition. Results showed that the emergence and the directionality of
cross-language form similarity effects can depend not only on task demands but
also on cognate type (identical vs. non-identical).
4.3 Factors Affecting Visual Word Recognition

in Bilinguals
Research on visual word recognition can pride itself on a relatively long tradi-
tion and throughout many years of studies a number of factors affecting written
lexical access have been enumerated and explored (cf. Warren 2013). Before the
most important, and most commonly analysed factors affecting visual word rec-
ognition are delineated, it seems justified to note that the influence of many of
those effects was primarily investigated within the monolingual domain and only
more recently has it been explored in the multilingual context. It also needs to be
observed that the multitude effects influencing lexical recognition do not exist in
separation. Instead, they interact making visual word recognition a highly complex
phenomenon. In the multilingual context, the complexity caused by interaction of
the various effects is further complicated by the occurrence of additional language
or languages. Accordingly, the following paragraphs elaborate on the most typical
factors affecting visual word recognition trying to contextualize them within the
multilingual domain.
4.3.1 Word Superiority Effect
The word superiority effect was described as one of the first factors influencing
visual word recognition. In 1887 Cattell confirmed that identification of words
is easier and faster than the identification of the so-called nonwords—words that
from the phonological perspective can occur in a language but do not exist in a
given language. Cattell proved that the identification of letters within a string of
letters creating a meaningful word is easier than identifying words within a mean-
ingless stream of words. The results of Cattell’s and a number of later studies
prove that word recognition is not based on a letter-by-letter sequence. Instead,
the effect accounts for the existence of some top-down processing from words to
letters. More recent studies distinguish between legal and illegal nonwords (word
and wlod, respectively; cf. Warren 2013). Interestingly enough, the word superi-
ority effect concerns both types of words form this category. In the experiments
presented in the practical chapter, nonwords are used along with bi- and trilingual
cognates and control words.
WWW.EBOOK777.COM
4.3 Factors Affecting Visual Word Recognition in Bilinguals 109
4.3.2 Word Length Effect
Another important factor influencing our visual lexical recognition that has been
recurrently explored in the monolingual domain is quite physical and refers to
the very length of the word. Word length can be based on orthographic measures
(number of letters) or phonological measures (number of phonemes and sylla-
bles). Importantly, as New et al. (2006, p. 45) report in their review article on the
word length effect in written lexical access, these different measures are gener-
ally highly interconnected, and they also correlate with other variables (such as the
number of orthographic neighbors and the printed frequency) that influence word
recognition. Although it seems reasonable that our recognition should slow down
together with the word length, a variety of tasks (such as perceptual identifica-
tion, lexical decision, naming or eye tracking) used to examine the effects of word
length on visual word recognition offers rather inconsistent results ranging from
inhibitory (longer words are more difficult) to null effects. Frederiksen and Kroll
(1976), for instance, obtained reliable inhibitory length effects in naming but not
in lexical decision, while Hudson and Bergman (1985) found length effects in both
types of tasks. It needs to be noted that the factor under discussion is of crucial
importance also for the bilingual visual recognition studies where the effects of
word length on the recognition of foreign language words have been investigated
in a number of studies (New et al. 2006; Lemhöfer et al. 2008). Such a criterion
was also used in the experiments presented in the final chapter of the book.
4.3.3 Frequency Effect
Results of lexical decision tasks show that high frequency words are recognized
faster than low-frequency words, the relation known as the frequency effect (e.g.
Howes and Solomon 1951; Schilling et al. 1998). Participants respond faster to
high-frequency words than to low-frequency words in almost any lexical process-
ing task, including lexical decision, reading aloud, semantic categorization, and
picture naming. Not only response times are quicker but also error rates are lower
in the case of words that are used more frequently. In the bilingual domain, some
evidence suggests that the frequency effect might even be larger in the second as
compared with the first language (van Wijnendaele and Brysbaert 2002). In a lexi-
cal decision experiment conducted with Dutch-English bilinguals, Duyck et al.
(2008) compared the effect of word frequency on visual word recognition in the
first language with that in the second language and found that, even though corpus
frequency was matched across languages, bilinguals showed a considerably larger
frequency effect in their second language.
As the authors of the research suggest, the findings support models of lexical
access that incorporate the frequency effect as the result of an asymptotic learn-
ing process (Duyck et al. 2008, p. 852). However, a recent study by Peeters et al.
WWW.EBOOK777.COM
(2013) conducted on late French-English bilinguals confirmed facilitatory effects of

word frequency for both languages. The analysis on the identical cognates revealed
effects of both English and French frequency. Interestingly, the cognate facilita-
tion effect was larger for cognates with a low English frequency than for cognates
with a high English frequency. The error analyses showed fewer errors for cog-
nates compared to control words and fewer errors for words with a high English
frequency compared to words with a low English frequency. The authors note that
the results support a claim made by Gollan, et al. (2008) that an increase in the use
of a word leads to a smaller frequency effect. The study was the first to show, both
behaviourally and electrophysiologically, that not only the frequency of the target
reading of the cognate (in this case English) but also the frequency of the non-
target reading (in this case French) is of importance for processing (Peeters et al.
2013, p. 328). Lastly, it needs to be noted that a number of visual word recognition
studies confirmed that length is strongly connected with frequency. For instance,
it would be difficult to state which of these two factors is responsible for shorter
response times in the case of short and frequent words. Frequency and length are
two important areas to be controlled in the process of research stimulus selection.
4.3.4 Neighbourhood Effect
Lexical neighbours, already mentioned in Sect. 4.2.2, are words with similar spell-
ing patterns. The role of a target’s orthographic neighbourhood in visual word rec-
ognition has been investigated in a number of studies both in the monolingual as
well as in the bilingual context. However, the role that activated neighbours play
in target recognition has so far not been determined unequivocally (cf. Pugh et al.
1994). As Warren points out, “a number of properties of the neighbourhood are
important for visual word recognition” (2013, p. 143). Above all, the neighbour-
hood size or density is strongly connected to the lexical frequency characteristics
of the neighbourhood and of the target itself. Generally, responses to low-fre-
quency words are affected by the neighbourhood effect, whereas in the case of
high-frequency words the effect is not likely to occur. To complicate things even
further, the nature of this effect depends on task demands. Depending on task type,
the neighbourhood density or word frequency, the effect yields facilitatory, inhibi-
tory or null results (cf. Pugh et al. 1994).
The neighbourhood effect has also been analysed within the bilingual context.
As already mentioned, in 1998 Van Heuven et al. (1998) conducted a series of pro-
gressive demasking and lexical decision experiments investigating how the rec-
ognition of target words exclusively belonging to one language is affected by the
existence of orthographic neighbours from the same or the other language of bilin-
guals. The results showed that increasing the number of orthographic neighbours
in Dutch systematically slowed response times to English target words in Dutch/
English bilinguals, whereas an increase in target language neighbours consistently
produced inhibitory effects for Dutch and facilitatory effects for English target
WWW.EBOOK777.COM
words (Van Heuven 1998, p. 458). The experiments were interpreted as providing
evidence for parallel activation of words in an integrated Dutch/English lexicon.
4.3.5 Cognate Facilitation Effect
As stated above, cognates have been one of the most important sources of stimulus
materials in studies attempting to unravel the process of bilingual word recogni-
tion. The cognate facilitation effect (also CFE), according to which cognates are
processed faster than noncognates, has often been taken as evidence for a multi-
lingual lexicon that stores words of two or more languages in an integrated fash-
ion and/or for a lexical access procedure that activates word candidates in several
languages in parallel (the nonselective access hypothesis). Clearly, cognates have
been very useful as tools to investigate the language (non)specificity of lexical
access in bilinguals (cf. Friel and Kennison 2001 for an overview). A number of
experimental studies have demonstrated processing differences between cognates
and other words in bilinguals, thus indicating differences in the representations
depending on the word type (cognate/noncognate). In many tasks that involve
speech production or recognition, cognates have a processing advantage over
non-cognate translation equivalents and interlingual homophones or homographs.
Worth mentioning here is the fact that the cognate facilitation effect has also been
reported in recent ERP studies. Midgley et al. (2011) recorded the electroencepha-
logram (EEG) of English-French bilinguals while cognates and matched control
words were presented in the participants’ L1 (English) and L2 (French) language
blocks. The findings led the authors to the conclusion that the mapping from form
to meaning is facilitated for cognates, and that access to the bilingual lexicon is
language non-selective. To provide background for the experiments described in
Chap. 5, the subsequent paragraphs will expand on issues pertaining to different
aspects tackled in the cognate studies outlining both the research methods they
used as well as the obtained empirical findings.
4.3.5.1 Association Tasks
In his article published in 1976, Taylor tried to answer the question whether the
similarity between French and English words is a factor to be considered in bilin-
gual language behaviour. For the purpose of the research, conducted within a con-
tinued free-association task paradigm, he selected a number of French and English
translation equivalents which were also formally similar (i.e. in sound and mean-
ing). He noticed that
(…) in continued-word association to the two types of key words, French – English bilin-
guals produced different patterns of responses. More response words to the similar than to
the dissimilar key words tended to be translation equivalents between the two languages
(Taylor 1976, p. 85).
WWW.EBOOK777.COM
English and French cognate translations (e.g., carrot and carotte) elicited the same
primary associates (vegetable and légume, respectively) more often than non-
cognates did. Consequently, the response list for cognates had far more common
associates than the list of noncognates. Van Hell and De Groot (1998a) in their
study on Dutch-English bilinguals used a similar methodology and received simi-
lar results. Again, the participants produced associations to cognates faster than
to noncognates. Associations were also faster to concrete than to abstract words,
and faster to nouns than to verbs. The other finding was that within- and between-
language associations were more similar for cognates, concrete words and nouns.
Essentially, in both experiments (Taylor 1976 and Van Hell and De Groot 1998a),
cognates yielded far more associates than noncognates. For cognates, associ-
ates were more often translation equivalents than they were for noncognates.
Additionally, associates for cognates were far easier to generate than those for
noncognates. The presented findings were taken as evidence that the concep-
tual representation in bilingual memory depends, at least to some extent, on the
similarity of form. On the basis of their results, the authors suggested also that
those wordtypes for which association was faster (concrete words, nouns and cog-
nates) share more conceptual features. Nevertheless, Sherkina (2003, 2004) doubts
whether this has to be always true for cognates. She supports her claim with three
significant postulates. Firstly, there are cognates which share some of their mean-
ings between two languages, but not all. Secondly, there are cognates which have
more conceptual features in one language than in the other; that is, they have a
narrower meaning in one language than in the other. Moreover, there are non-
cognate translation equivalents which share most or even all conceptual features,
such as words for certain animals and plants, some scientific terms, and calques.
Accounting for the above, Sherkina posits that “the difference between cognates
and noncognates in the word association task is more likely to occur in the recog-
nition component than in the association part” (Sherkina, 003, p. 137).
4.3.5.2 Categorisation and Translation Tasks
Cognates have been reported to have an advantage over noncognates in a num-

ber of studies using categorization tasks. Dufour and Kroll (1995) who used this
methodology to analyze a group of more fluent and less fluent English-French
bilinguals concluded that “(…) subjects were faster at categorizing words in both
languages when the language of the category name matched the language of the
target word” (Dufour and Kroll 1995, p. 166). Similarly, experiments exploiting
translation as the basic research tool confirmed the privileged position of cognates
which were reported to be translated more quickly than noncognate words (cf.
De Groot 1992; Sánchez-Casas et al. 1992). Both backward and forward transla-
tion tasks conducted by Kroll and Stewart (1994) and De Groot and Poot (1997)
showed that when participants (bilinguals) were asked to translate from their L1
to their L2, they tended to translate cognates faster, give more responses for cog-
nates and provide correct responses to cognates more often than to noncognates.
WWW.EBOOK777.COM
By the same token, some previous experiments proved that translation times for
cognates may be equal regardless of translation direction, whereas for noncognates
the results are different. In cued translation tasks, for instance, L2 to L1 transla-
tion is faster than L1 to L2 for noncognates (cf. De Groot et al. 1994; Costa et al.
2000). Also in translation recognition tasks (in which pairs of words from two lan-
guages are presented and participants need to decide whether the words form a
translation pair) bilinguals are found to make faster and more correct decisions for
cognate pairs than for noncognate pairs (cf. De Groot and Comijis 1995). More
recent translation experiments also lend support to the special status of cognates.
For example, Yudes et al., in their 2010 study on cognate effects in bilingual lan-
guage comprehension, found that in the translation decision task, faster and more
accurate responses were associated with cognates.
4.3.5.3 Priming Tasks
Cognate translations typically produce stronger priming than do noncognate trans-

lations in a variety of priming paradigms. It needs to be noted that the priming
effect occurs when the presentation of a word (the prime) in one language facili-
tates recognition of its translation (the target) when presented later relative to its
presentation after an unrelated word (cf. Voga and Grainger 2007, p. 938). For
example, in long-lag priming studies, where the prime and the target are separated
by a large number of intervening trials, only cognates have been shown to pro-
duce facilitation (cf. Cristoffanini et al. 1986; Lalor and Kirsner 2000). In turn,
De Groot and Nas (1991) found a strong cross-language semantic priming effect
(e.g., Dutch rivier priming English water) for cognates. The effects were as pro-
found as those for within-language semantic priming, though Friel and Kennison
(2001) cite Grainger and Frenck-Mestre (1998) and point out that similar results
have also been obtained for noncognates. On the whole, an increasing number of
studies support the assumption that strong translation priming effects occur for
cognates (cf. Gerard and Scarborough 1989, Costa, et al. 2005; Davis et al. 2010),
weaker (some claim that non-existent; cf. Duyck 2005; Dimitropoulou et al. 2010)
for noncognates.
4.3.5.4 Lexical Decision Tasks
In many reaction time (RT) studies, which measure the CFE by comparing RTs for
cognates and matched language-specific control words, cognates were responded
to faster than control words that exist in only one language. This has since long
been established by studies on bilingual word recognition in the visual modal-
ity (cf. Caramazza and Brones 1979; Cristoffanini et al. 1986; De Groot and Nas
1991; Sánchez-Casas et al. 1992; Kroll and Stewart 1994; Dufour and Kroll 1995;
Dijkstra et al. 1998; Dijkstra et al. 1999; Schwartz and Kroll 2006; Schwartz et al.
2007; Voga and Grainger 2007; Lemhöfer et al. 2008). However, more recently,
WWW.EBOOK777.COM
the effect has also been observed in the auditory modality (cf. Marian and Spivey
2003) and in word production (cf. Costa et al. 2000; Costa et al. 2005).
Apparently, RT studies involve a variety of experimental paradigms. One of
the most typical is a lexical decision task during which participants are asked to
decide whether a letter string forms a word in either of the languages involved
in the experiment. A substantial body of evidence with monolingual subjects has
revealed that the time needed to recognize a word is influenced by the character-
istics of words that are formally similar to the stimulus (cf. Grainger and Dijkstra
1992; Rastle and Davis 2008). Similarly, in bilingual systems, both lexicons might
be initially activated while an extended cross-lexical search appears to be con-
ducted on the basis of the formal features of the input (Smith 1997). Since the
lexical decision task methodology has been used in a number of visual lexical rec-
ognition studies, it seems justified to take a closer look at some selected experi-
ments conducted within this research paradigm. Thus, the subsequent paragraphs
will centre around visual lexical decision experiments devised to corroborate the
facilitatory cognate effect.
Gerard and Scarborough (1989) conducted a visual lexical decision task with
proficient English-Spanish bilinguals whose dominant language was English. The
input set consisted of Spanish or English lexical items that were noncognates in
the two languages (e.g., dog—perro), or cognates (e.g., actual—actual, having
the same meaning), or homographic noncognates (e.g., red—red, meaning net in
English). The input stimuli (single words) were presented in two different condi-
tions; (1) as Spanish items, (2) as English items. The results of the test showed
that bilinguals generally processed L2 (Spanish) information more slowly than L1
(English) lexical items, but otherw there were no effects of formal similarity; the
bilinguals appeared to be functioning like monolinguals. However, a repetition of
the same orthographic pattern across languages produced facilitation for cognates
as well as for homographic noncognates. Gerard and Scarborough explained these
results claiming that “the cross-language repetition effect appears to be a general
effect of encoding the same orthographic pattern” (Gerard and Scarborough 1989,
p. 312).
A similar cognate advantage was identified in lexical decision tasks with bilin-
guals by Cristoffanini et al. (1986). They found a 100-ms advantage for cognates
in English-target lexical decisions with Spanish-dominant bilinguals. They argued
that a possible explanation of the cognate advantage is that both the English and
Spanish lexical entries of a cognate are activated. Given that lexical access is
faster in the bilingual’s dominant language, the lexical entry in that lexicon will be
accessed first and will prime the corresponding lexical entry in the other language
lexicon (cf. Solomyak and Marantz 2009).
A comparable facilitation effect was obtained in Dijkstra et al.’s (1999) visual
word recognition study. Their stimuli included cognates and interlingual homo-
phones/homographs with different degrees of phonological and orthographic
similarity. Therefore, there were 6 conditions (“S” stands for semantically simi-
lar, “O”—orthographically similar, and “P”—phonologically similar): SOP, SO,
SP (cognates); OP, O, and P (homophones or homographs). Each test word was
WWW.EBOOK777.COM
matched in target language frequency and in word length with a control word
which was completely dissimilar to its translation equivalent. In the first experi-
ment, the task was progressive demasking (subjects had to identify a word which
alternated with a mask); in the second, the task was lexical decision. Cognates
were recognized faster than noncognate control words. Reaction times for SOP,
SO and O were faster than for their control words, for SP—not different from, and
for P and OP—slower than control words. The results showed that semantic and
orthographic overlap led to faster RT, but phonological overlap slows them down.
Interestingly enough, a positive correlation between phonological similarity and
RT was found. This was taken to explain why, in a number of recognition studies,
cognates have a processing advantage, but interlingual homophones do not. Access
to the former is facilitated by at least semantic overlap, as well as, in the case of
homographic cognates, by orthographic overlap; the effect of phonological overlap
is not strong enough to neutralize the facilitation from two sources. Interlingual
homophones/homographs do not have semantic overlap, and, as Dijkstra et al.
(1999, p. 510) put it, “facilitation from orthography and inhibition from phonology
cancel each other”.
4.4 Factors Affecting Cognate Recognition
In the previous subchapter an eclectic look was taken at factors affecting visual
word recognition in bilinguals with the increased focus on the special status of
cognates resulting in their facilitatory effect. In the subsequent sections, it seems
reasonable to take a closer look at factors affecting cognate recognition which
include cross-language similarity, task-dependency, and word frequency.
4.4.1 The Effect of Cross-Linguistic Similarity on Cognate

Recognition
Many linguists (cf. Lemhöfer and Dijkstra 2004; Lemhöfer et al. 2004; Solomyak
and Marantz 2009) point to the fact that cognates overlapping in form trigger
faster recognition. A question arises, however: how does the cognate facilita-
tion effect depend on cross-linguistic similarity? More precisely, what happens
if cognates are presented that are nonidentical in the language pair the bilingual
knows? For instance, are German-English bilinguals affected in their recognition
of the German word Nacht by its similarity (but nonidentity) to the English word
night? If form-identity is not required, how much form-overlap between the two
readings of a cognate will still induce a facilitation effect? The answer to these
questions can be found in the experiment conducted by Van Hell and Dijkstra
(2002) who had trilinguals with Dutch as their native language, English as their
L2 and French as their L3 perform a word association task or a lexical decision
WWW.EBOOK777.COM
task in their L1. Stimulus words were (mostly) nonidentical cognates such as bak-
ker (English baker, French boulanger) and droom (English dream, French rêve).
Shorter association and lexical decision times were observed for Dutch-English
cognates than for noncognates. For trilinguals with a higher proficiency in French,
faster responses in lexical decision were found for both Dutch-English and Dutch-
French cognates. The conclusion was that even when their orthographic and pho-
nological overlap across languages is incomplete, cognates may be recognized
faster than noncognates.
Similar results have been obtained by Dijkstra et al. (2010) who examined the
influence of phonological and orthographic overlap of cognates in languages that
share the same alphabet. The results of the study were interpreted as suggesting
that the occurrence and the directionality of cross-language form similarity effects
may depend on task demands as well as cognate type. The study analyzed how
cross-linguistic similarity affects Dutch-English identical and nonidentical cognate
visual recognition using different types of lexical decision tasks. Performing one
of three tasks, Dutch-English bilinguals processed cognates with varying degrees
of form overlap between their English and Dutch counterparts (e.g., lamp-lamp
vs. flood-vloed vs. song-lied). In lexical decision, reaction times decreased going
from translation equivalents without any cross-linguistic orthographic overlap to
very similar but non-identical cognates. In turn, identical cognates showed a big
discontinuous processing advantage and were additionally subject to facilitation
from phonological similarity. In language decision, the effect of orthographic sim-
ilarity reversed: a cognate inhibition effect arose, the size of which increased with
orthographic similarity. Here identical cognates were markedly slower than other
cognates. In progressive demasking, no orthographic similarity effect was found
for non-identical cognates, but a semantic similarity effect arose. In addition, there
was a facilitation effect for identical cognates of low English frequency. More
importantly, for identical cognates, and only for identical cognates, a marked facil-
itatory effect of phonological similarity was present. This finding is in line with
the previous conclusions, according to which cognates with high orthographic
similarity across languages were named faster when they had a highly similar pho-
nological code than when they were more distinct. This contention was further
confirmed in yet another lexical decision task study. Schwartz et al. (2007) asked
English-Spanish bilinguals to name English and Spanish noncognate words as
well as cognate words that varied in their degree of phonological and orthographic
overlap. The results showed that cognates with high orthographic and phono-
logical overlap (O+P+) were named faster than cognates with high orthographic
overlap but low phonological similarity (O+P−). However, when the orthographic
form of cognates was different (O−P+ and O−P−), the effects of phonology
were not statistically reliable. The authors interpreted the results as evidence that
cross-language activation forwards from orthography to phonology. Similarly to
Dijkstra et al. (1999), Schwartz et al. showed that the facilitatory effects typically
related to lexical overlap across languages can be ‘reduced or turned into inhibi-
tion when there is not a consistent mapping across all codes’ (2007, p. 122).
WWW.EBOOK777.COM
4.4 Factors Affecting Cognate Recognition 117
Both Schwartz et al.’s (2007) as well as Dijkstra et al.’s (2010) studies provided
evidence that is consistent with a localist connectionist view of bilingual mem-
ory. In view of this proposal, the cross-linguistic similarity of cognates leads to
a greater semantic activation, since the associated meaning receives activation
from two lexical representations rather than one, as it occurs with noncognates.
However, the degree of semantic activation depends on the orthographic and pho-
nological similarity of cognates due to the existence of inhibitory connections
between lexical representations. Consequently, the speed and accuracy of cognate
reading depends on their cross-linguistic similarity (identical and non-identical
cognates). Accordingly, it seems reasonable to believe that greater overlap causes
more facilitation. A number of experiments show that the degree of semantic and
phonological overlap positively correlates with the magnitude of effects caused by
sharing (cf. Dijkstra and Van Heuven 2002a, b; Lemhöfer et al. 2008; Solomyak
and Marantz 2009; Dijkstra et al. 2010).
Summing up, many language recognition studies confirm the assumption that
even not completely identical but only orthographically similar cognates yield
comparable effects. Based on the results of the studies presented above, it can be
concluded that only partial similarity is sufficient and that orthographic identity or
maximal phonological overlap are apparently not required for the cognate facili-
tation effect to arise. Importantly, recent work has shown that even for bilinguals
whose languages do not share script (e.g., Korean-English, Greek-French, Hebrew-
English, Japanese-English), cognate facilitation effects can be observed (cf. Gollan
et al. 1997; Voga and Grainger 2007; Kim and Davis 2003; Hoshino and Kroll
2008). The very recent research by Allen and Conklin (2013) more fully explores
what underpins the cognate ‘advantage’ in different script bilinguals (Japanese-
English). Instead of using the more traditional binary cognate/noncognate distinc-
tion, the authors used continuous measures of phonological and semantic overlap,
L2 proficiency and lexical variables (e.g., frequency). The results of the lexical
decision task demonstrated that increased phonological similarity (e.g., bus/basu/
vs. radio/rajio/) leads to faster response times.
4.4.2 The Cognate Facilitation Effect and the Word

Frequency Effect
On the surface, the word frequency effect (also WFE) resembles the cognate
facilitation effect to a significant extent. Both yield shorter naming latencies (for
high-frequency words and cognates), and, unlike in priming, this is not caused
by previous access to these or related words since both effects appear even after
repeated access to experimental items (cf. Jescheniak and Levelt 1994; De Groot
and Keijzer 2000; Duyck et al. 2008 on the WFE, Costa et al. 2000; Van Assche
et al. 2009; Dijkstra et al. 2010 on the CFE). Naturally, the question arises whether
these two effects might be related. Or, as Sherkina puts it: “might they be manifes-
tations of the same effect, within and across languages?” (2003, p. 142).
WWW.EBOOK777.COM
A wide range of studies with unbalanced bilinguals lends support to the

assumption that words from a dominant language are processed faster than their
equivalents from a non-dominant language (cf. Kroll and Stewart 1994). The ques-
tion remains, however, why the activation of dominant language words is stronger.
It seems reasonable to claim that naming latencies are in general faster in a domi-
nant language because the dominant language is used more often. As a conse-
quence, the words of a dominant language have higher frequency for a bilingual
than the words of a non-dominant language. Accordingly, if a cognate is shared
between two languages, and both language-specific frequencies contribute to its
overall frequency, then its frequency in the non-selected language is what seems
to cause the cognate facilitation effect. From the above discussion, it follows that
the amount of facilitation should positively correlate with the frequency in the
non-selected language. And this is exactly what is observable in various nam-
ing experiments: during naming in L1, the magnitude of the CFE depends on L2
frequency, but during naming in L2, on L1 frequency (cf. Sherkina 2003; Duyck
et al. 2008). Judging from the above-listed experiments, it seems plausible that for
multilinguals cognates must have higher frequency, since the overlapping parts in
their orthographic and/or phonological representations occur in both languages.
Sherkina (2003) has demonstrated that bilinguals do perceive cognates as having
higher frequency. She conducted a frequency-rating task, in which bilinguals rated
cognate frequency higher than monolinguals did.
A further extension of the same asymmetry was found for trilinguals by Van
Hell and Dijkstra (2002). Relying on bilingual experiments, they assumed that just
as with L1 and L2, lexical items belonging to L3 would have lower frequency for
a trilingual than lexical items belonging to L2. In other words, the weaker the lan-
guage is, the lower mean frequency its items have, and the less they contribute to
the CFE. The authors’ 2001 experiment confirmed this assumption: in L1 nam-
ing, facilitation was greater if a word had a cognate in L2 than if it had a cog-
nate in L3 (in the latter case, the CFE was present numerically, but did not reach
significance). In that experiment, L1 was the native language of the subjects, and
their strongest language, L2 was weaker, and L3 was their weakest language.
Notably, to verify their findings, for another experiment Van Hell and Dijkstra
(2002) selected trilinguals whose fluency levels in L2 and L3 were comparable.
As the authors predicted, more fluent L3 knowledge resulted in a significant CFE
for words with cognates in L3 as well as for words with cognates in L2. Within
the frequency account, higher fluency in L3 indicates higher mean frequency of
its lexical items, and the change for the words with cognates in L3 is exactly what
the frequency account predicts. If this is the case, then the cognate facilitation
effect can indeed be interpreted as the cross-language manifestation of the word
frequency effect.
More recently, however, Peeters et al. (2013) conducted a study in which the
word frequency effect was tested on identical cognates. It is one of the first studies
which manipulates the frequency of both readings of cognates. Firstly, the study
showed both behaviourally and electrophysiologically that both readings of cog-
nates are important in cognate processing. Secondly, the results clearly confirmed
WWW.EBOOK777.COM
4.4 Factors Affecting Cognate Recognition 119
that cognate facilitation is larger for cognates with a low compared to cognates
with a high frequency in the target-language. To explain the differential effects of
both cognate readings Peeters et al. (2013) proposed that identical cognates would
be represented twice in the bilingual brain. These two representations can be char-
acterized by their own language-specific frequency and “can be linked to their own
language-specific plural markers, gender, and syntactic category (Peeters et al.
2013, p. 329).
4.4.3 The Cognate Facilitation Effect and Task Demands
The issue of task dependency has already been referred to in Chap. 3.

Nevertheless, it seems reasonable to briefly address it once again—this time in
the context of the cognate facilitation effect. As stated before, in psycholinguis-
tic experiments, task demands have been confirmed to strongly affect language
activation. As for the effect of task demands on cognate processing, it has been
repeatedly proved in numerous studies on multilingual visual word recognition
(e.g., Dijkstra et al. 1998 2010). A series of three lexical decision experiments
conducted by Dijkstra et al. (1998) showed that interlingual homographs may
be recognized faster than, slower than, or as fast as monolingual control words
depending on task requirements and language intermixing. The results were con-
firmed in the much later study conducted by Dijkstra et al. in 2010. Again, the
researchers found that although the three conducted experiments included nearly
the same test stimulus materials, lexical decision led to different result patterns
than language decision and progressive demasking. The described results highlight
the adaptability of the bilingual lexico-semantic system to varying task demands.
More importantly, the task-dependent results seem to provide evidence in favour
of a localist connectionist account, since only this approach, as implemented
in the BIA/BIA+ models, has proposed both a processing and a task account
(cf. Dijkstra and Van Heuven 2002; Sects. 3.6.1 and 3.6.3).
4.5 Models of the Bilingual Lexicon and How They

Account for the Cognate Effect
Most researchers are now convinced that there is a common conceptual store for
all languages of a bilingual (or a multilingual), the contention which has been
demonstrated in many experimental studies reviewed in Chap. 3. However, as
Dijkstra et al. (2010) rightly observe, the biggest challenge for a model of the
multilingual lexicon in general is that it should advance a rational explanation for
both cross-language effects and the ability of multilinguals to use one language at
a time. As has been reviewed in the previous sections, cognates more often give
different results in a number of experiments than noncognates. Therefore, many
WWW.EBOOK777.COM
psycholinguists propose that lexical representations for cognates are stored differ-
ently than those for other words. In order to be able to account for cognate effects,
a multilingual mental lexicon model should allow sharing both conceptual and
lexical elements between two lexicons at some level. Theoretically, non-sharing
alternatives are possible, but in this case there should be some special process-
ing mechanisms, and, as Sherkina (2003) points out, a major challenge for them
would be in accounting for the frequency effects in the processing of cognates.
Over the years, several theoretical accounts attempting to answer the question of
how models of bilingual lexicon accommodates the cognate effect have been pro-
posed. In the subsequent sections, four theoretical positions discerned by Dijkstra
et al. (2010) will be presented with a special focus on the aspects of both cross-
linguistic similarity and word frequency.
4.5.1 Shared Morphological Representation
The first proposition discussed by Dijkstra et al.(2010), initially advocated by

Sánchez-Casas and colleagues (cf. Sánchez-Casas et al. 1992; Sánchez-Casas and
García-Albea 2005; Davis et al. 2010) and Kirsner and colleagues (cf. Kirsner, et al.
1993; Lalor and Kirsner 2000), assumes that a cognate relation between two words
constitutes a special case of a morphological relation that may exist between words
within the same language and that is reflected in the joint storage of morphologi-
cally related words in memory. Of late, Davis et al. (2010) put forward this proposal
noting that the lexicon could be structured on the basis of morphological properties,
such that words that have the same etymological root would share a representation
at the morphological level. Such morphological representations would be independ-
ent of language. By way of illustration, a Polish-English-German multilingual would
have one memory representation containing the Polish words student, studentka, stu-
diować, the English words student, study and studied and the German words Student,
Studentin, studieren. To check the validity of this model, Dijkstra and Van Heuven
(2002a) and Dijkstra et al. (2010) conducted a series of experiments aimed at ver-
ifying the claim according to which shared morphological representations are cre-
ated when bilinguals perceive that two word translations have a minimally similar
orthographic core. They hypothesized that if such representations existed the cog-
nate effect for identical and nearly-identical cognates would be of a similar size. In
other words, the assumption was that if cognates are represented by an all-or-none
morphological representation, the level of cross-linguistic similarity might not be
such an important characteristic of cognates since as long as some minimal degree
of orthographic and semantic overlap is present, the shared morphological represen-
tation would be established. Consequently, if the hypotheses were to be confirmed,
it would mean that a determining factor of the cognate facilitation effect is not the
level of cross-linguistic similarity but the word frequency of the cognates in either or
both languages.
WWW.EBOOK777.COM
4.5 Models of the Bilingual Lexicon… 121
Notably, the results of the experiments conducted by Dijkstra and Van

Heuven (2002a) and Dijkstra et al. (2010) offered only little support for the
above-presented views. The assumption that “there should be no effect of ortho-
graphic similarity across a whole range of orthographical similarity of cognates”
(Dijkstra et al. 2010, p. 11) was not confirmed. Instead, data gathered in this
research indicated that RTs for cognates became shorter if the level of ortho-
graphic similarity increased. Further evidence against shared morphological
representation of cognates was collected by Voga and Grainger (2007), who
tested Greek-French bilinguals performing a lexical decision task on French
target words preceded by masked Greek prime words. The authors claimed that
if the cognate effect was indeed based on a shared morphological representa-
tion, priming with cognate translations should evoke a similar RT pattern as
cross-language morphological priming (cf. Voga and Grainger 2007). However,
the patterns for the two conditions were quite different (cf. Voga and Grainger
2007). Cognate primes produced faster RTs than morphological primes at both a
50 ms and a 66 ms prime duration, while a significant morphological facilitation
effect (relative to a control condition) appeared only at the 66 ms prime duration
(Voga and Grainger 2007, p. 947).
4.5.2 Associatively Linked Orthographic Representations
A second theoretical position attempting to offer a convincing model accounting

for the representation of cognates assumes that translation equivalents, both cog-
nates and noncognates, are characterized by shared semantic representations and
associatively linked word form (orthographic) representations. By way of exam-
ple, the German-English cognate pair Lampe-lamp might have a common meaning
representation and two linked lexically-orthographic representations. Such a posi-
tion was strongly advocated by De Groot and Nas (1991) who, in the summary of
their article reporting on four experiments in which the representation of words in
a Dutch-English bilingual lexicon was examined, stated that
[our] data are most consistent with the view that (1) the representations of both cog-
nate and noncognate translations at the lexical level of representation are connected, (2)
cognate translations share a representation at the conceptual level and these shared rep-
resentations are connected to those of associatively related words at the same level; (3)
noncognate translations are represented in separate concept nodes and these nodes only
have connections to those of associatively related words of the same language (De Groot
and Nas 1991, p. 117).
This view was taken to predict frequency-dependent cognate facilitation effects

in various tasks. If the additional assumption is made that multiple orthographi-
cally similar words become activated on the basis of the input, the effects of ortho-
graphic similarity across languages could also be accounted for (cf. Dijkstra and
Van Heuven 2002b, p. 222).
WWW.EBOOK777.COM
However, more difficult to explain by this account are the cognate facilitation
effects observed by Voga and Grainger (2007). In their Greek-French masked
priming task, overlap-dependent facilitation effects arose for Greek-French cog-
nate primes relative to unrelated control primes (cf. Voga and Grainger 2007;
Experiments 2 and 3). It is hard to see how L1-L2 word associations could be
directly affected by (phonological) form overlap when the two languages involved
are characterized by different scripts. In addition, as Dijkstra et al. (2010) rightly
argue, a task account is badly needed to explain why lexical decision, language
decision, and progressive demasking lead to different result patterns for associa-
tively linked cognate representations. In contrast, somewhat distinct predictions
are proposed with respect to the role of cross-linguistic similarity and word fre-
quency by distributed connectionist models such as proposed by Thomas and
Van Heuven (2005); French and Jacquet (2004), and Li and Farkaš (2000), and
by localist connectionist models such as BIA (cf. Dijkstra, van Jaarsveld and
ten Brinke 1998; Voga and Grainger 2007) and BIA+ (Dijkstra and Van Heuven
2002a, b). Consequently, in the light of the research, the model promoting the
associatively linked orthographic representations of cognates seems to be highly
unlikely.
4.5.3 The Distributed Connectionist Model
A third theoretical proportion trying to account for the cognate facilitation effect
is that of distributed connectionist model (cf. De Groot 1992; Thomas 1997;
French and Jacquet 2004; Thomas and Van Heuven 2005). Distributed mod-
els hold that the two readings for identical and nearly identical cognates share
most of the connections. Therefore, representations of these cognates in two
languages “lie closely together in multidimensional space” (Thomas and Van
Heuven 2005). Consequently, responses to such cognates should be faster than
for control words, because, “being in the same region, the attractors of the two
representations of a cognate exert a joint force” (Dijkstra et al. 2010, p. 11).
Responses to cognates with decreasing levels of similarity should gradually
become slower because the joint force decreases. Additionally, a sharp rise in
RTs would not be expected when one compares form-identical to slightly non-
identical cognates.
According to Thomas and Van Heuven (2005), in distributed connectionist
models, word attractors lying in the same area of a multidimensional space set up
by stimulus dimensions exert a joint “pull” during the word recognition process.
Co-activated cognate readings lead therefore to faster RTs than language-specific
control words resulting in the cognate facilitation effect. Furthermore, because the
joint attractor force is larger, the closer the two representations are, the faster RTs
for these cognates. Finally, as observed by Thomas and Van Heuven (2005) and
WWW.EBOOK777.COM
4.5 Models of the Bilingual Lexicon… 123
Dijkstra et al. (2010), a higher frequency of use might strengthen the connections
between representations and allow faster movement through multidimensional
space. As confirmed by numerous experiments (cf. Thomas 2002; Thomas and Van
Heuven 2005; Dijkstra and Van Heuven 2002a, b and Dijkstra et al. 2010), the dis-
tributed connectionist model, as presented above, offers a convincing explanation
to the cognate facilitation effect. Nevertheless, it is only the localist connectionist
model (BIA and later BIA+) that managed to account for the cognate effect occur-
ring under different task demands.
4.5.4 The Localist Connectionist Model
A fourth theoretical model proposed by Dijkstra et al. (2010) is based on local-

ist connectionist framework. In this model, in contrast to noncognate translations,
cognate translation equivalents share more orthographic, semantic, and/or pho-
nological features across languages (cf. Dijkstra, et al. 1999; Voga and Grainger
2007). This account proposes to see the cognate facilitation effect in reading as an
orthographic-semantic priming effect. As Dijkstra and Van Heuven (2002b) point
out, overlapping orthographic and semantic representations of both languages
become active upon the presentation of one of the readings of the cognate, lead-
ing to a facilitated recognition of cognates relative to noncognates. In this view,
the CFE depends on both cross-linguistic similarity (e.g., number of shared let-
ters in comparable word positions) and word frequency. Both a larger similarity
of the two readings of a cognate and a higher frequency result in a more strongly
activated shared semantic representation. Notably, this viewpoint also accounts
for a cognate facilitation effect for cognates that partly share their phonological
representations.
This final model accommodated the processing consequence of an increase
in cross-linguistic similarity of cognates in view of which its counterpart in the
other language becomes more activated. By way of example, the English word
lamp activates German Lampe to a larger extent than English school activates
German Schule. Additionally, the obvious consequence of the fact that both
reading of the cognate converge at the semantic level is the increase in the
semantic activation, with the largest coactivation appearing for identical cog-
nates. Interestingly, a number of experiments (Sherkina-Lieber 2004, Dijkstra
and Van Heuven 2002b, and Dijkstra et al. 2010) show that even a slight change
in similarity from identical to nearly-identical cognates results in either discon-
tinuous or markedly non-linear change in cognate recognition speed (cf. Dijkstra
and Van Heuven 2002a, b). This is because the introduction of even a small mis-
match implies that, instead of one representation (for form-identical cognates),
participants activate two representations that inhibit each other via lateral inhibi-
tion (cf. Dijkstra et al. 2010).
WWW.EBOOK777.COM
Worth mentioning here is the fact that while early research on the represen-
tations of cognates (cf. Dijkstra et al. 2010) allowed for the possibility of the
existence of a single symbolic representation for identical cognates with two rep-
resentations for nonidentical cognates, in their recent study, Peeters et al. (2013)
basing on the behaviourally and electrophysiologically gathered data, developed
the model in which identical cognates have, at some level (e.g., morphology), two
form-related representations and a (partially) shared semantic representation. In
their proposal identical cognates have two representations at the morphological
level, in between a shared orthographic form and a shared semantic representation
concluding that,
Although prima facie it might not seem optimally efficient to have two form-related
representations for the same word, when we keep in mind that identical cognates can
have different plural markers, gender and/or relative frequencies across languages and
are generally learnt in a different context (L1 at home, L2 at school/abroad), this might
nonetheless be the way such identical words are represented. The language-specific plural
markers, gender, and relative frequencies could then be related to a language-specific
morpheme for both readings of the identical cognates (Peeters et al. 2013, p. 331).
Recapitulating, the above models presented four different viewpoints concern-

ing the representation of cognates in a bilingual mind. As the results of experi-
ments conducted by Dijkstra et al. (2010) confirm, both cross-linguistic similarity
as well as word frequency are significant determinants of cognate recognition
time, which in turn is a marker of crucial importance for creating a reliable model
of bilingual lexical storage and retrieval. Finally, in the light of the discussion on
multilingual lexicon presented in Chap. 3, the question arizes whether similar
effects can be achieved for trilingual cognates and whether the localist connection-
ist model can be used to explain them. Accordingly, one of the aims of the experi-
ments presented in the following Chapter will be to verify the BIA+ approach to
cognate representations in a trilingual context.
4.6 Bilingual Word Recognition in a Sentence Context
As has been shown above, the majority of studies using interlingual homo-
graphs, interlingual neighbours or cognates in relation to visual lexical recogni-
tion focused on isolated words. They all show that lexical representations from
both languages are activated when reading in one language (language non-selec-
tive lexical access). However, more recently the three types of words have also
been used to analyse lexical recognition in sentential context (cf. Elston-Güttler
et al 2005; Van Assche 2012). In other words, a question has been posed of
whether language non-selective access generalizes to word recognition in sen-
tence contexts which provides not only a language cue but also semantic con-
straint for upcoming words.
WWW.EBOOK777.COM
4.6 Bilingual Word Recognition in a Sentence Context 125
One of the early hypotheses assumes that the presentation of words in a sen-
tence context restricts lexical activation to words of the target language only. As
Van Assche et al. (2013) aptly state, this would actually be “quite an efficient strat-
egy to speed up word recognition, because it reduces the number of lexical can-
didates” (2013, p. 3). Moreover, evidence coming from the monolingual context
shows that semantic and syntactic restrictions imposed by a sentence are used to
speed up recognition of upcoming words. For instance, in their study with ambigu-
ous words Binder and Rayner (1998) proved that monolingual sentential context
affects lexical access of such words. Still, the results of many visual recognition
studies conducted within the bilingual domain suggest that the language of the
preceding words is insufficient to restrict lexical access to words of the target lan-
guage, even when reading in the native language (cf. Schwartz and Kroll 2006;
Duyck et al. 2008; van Hell and De Groot 2008; Titone et al. 2011; Van Assche
et al. 2009, 2011; Bultena et al. 2013). As Van Assche et al. (2013) conclude in
their comprehensive review of the studies on bilingual word recognition in a
sentence context,
mixed results have been obtained for semantically constraining sentences, but recent studies
using time-sensitive eye movement recordings suggest that even a strong semantic context
does not necessarily eliminate cross-lingual activation effects, at least for early interaction
effects reflected in early reading time measures (Van Assche et al. 2013, pp. 3–4).
4.7 The Effect of the Cognate Status on Foreign Language

Vocabulary Learning
Beyond doubt, one of the aims of second and subsequent language acquisition
studies should be exploring processes responsible for mastering the foreign
language lexical system with a view to identifying effective solutions which
are transferable to the language classroom. Thus, an important research area
that awaits a more thorough exploration is the role of the facilitatory cognate
effect in the process of foreign language acquisition. So far, findings from many
studies in line with evidence from psycholinguistic research show a cognate
facilitation effect in L2 word learning that makes cognates easier to learn and
remember than non-cognate translations. It has been confirmed that the formal
similarity of cross-language cognates provides means for language learners to
connect L2 and L1 lexical forms and thus to process the higher level represen-
tations such as concepts or mental images through the L1 word form and not
directly from the L2 lexical representation in word recognition. Analogously, it
has been proved that language learners can activate L2 lexical forms through its
L1 similarity neighbour in word production. Results of numerous studies indi-
cate that cognates provide learners with a number of significant advantages. By
way of illustration, Lotto and De Groot (1998) found that cognates and high-fre-
quency words were easier to learn than noncognates and low-frequency words,
whereas De Groot and Keijzer (2000) proved that cognates and concrete words
WWW.EBOOK777.COM
were easier to learn and less susceptible to forgetting than noncognates and
abstract words. Also, Ellis and Beaton (1993) corroborated the assumption that
“native-to-foreign learning is shown to be easier the more the FL words conform
to the phonological and orthographic patterns of the native language” (1993, p.
559). Friel and Kennison in their state-of-the-art article on cognates in foreign
language learning offer a comprehensive summary of cognate-induced advan-
tages claiming that,
participants with no prior knowledge of the to-be-learned language recall a higher per-
centage of cognates than noncognates, achieve a high level of recall in fewer learning
sessions for cognates than noncognates, and yield faster response latencies in translating
cognates than they do for noncognates (Friel and Kennison 2001, p. 249).
In this connection, Lotto and De Groot (1998) and De Groot and Keijzer (2000)
suggested a possible source for the superior foreign language vocabulary learn-
ing performance for cognates, considering both the learning stage (storage) and
the testing stage (retrieval) as possible loci of the effect. Their explanation extends
the view of bilingual memory representation that assumes shared representations
for cognates, but language-specific representations for noncognates (cf. Sánchez-
Casas et al. 1992; cf.). If true, the learning of a FL word that shares a noncognate
relation with the corresponding L1 word involves creating a new entry in memory,
whereas learning a cognate word may only involve adding new information to, or
adapting, a representation already stored there prior to the learning episode. The
latter process may be less demanding than the former, causing the learning advan-
tage of cognates over noncognates. As is apparent from the above review, findings
of many studies actually lead to the conclusion that cognates can be a significant
source of positive transfer and can facilitate vocabulary acquisition in L2 as they
offer “an encouraging springboard” (Aronin and Hufeizen 2009, p. 9) into learn-
ing of a new language (cf. Nation 1990; Odlin 1989; Ringbom 2001, 2007; Aronin
and Hufeizen 2009).
4.8 Conclusion
The present chapter has cautioned against characterizing cognate-pairing in terms

of the etymological origins of word-pairs or their meaning. Instead it has pro-
moted the view according to which in language comprehension cognate-pairing
should be based on processing terms as a form of automatic activation of lexical
addresses that have the same or similar shape (cf. Carroll 1992). As transpires
from the foregoing overview of visual word recognition in bilinguals, the special
status of cognates is of crucial importance for modelling multilingual processing.
Although the cognate facilitation effect is reported in many studies and taken to be
evidence in favor of language non-selective lexical access, there is no consensus
about cognate representation. The localist connectionist position, however, seems
to be the best accounted for.
WWW.EBOOK777.COM
4.8 Conclusion 127
The above discussion of the special status of cognates both in modelling visual
recognition and in foreign language learning and teaching has not included the
multilingual context. The following chapter will be devoted solely to the descrip-
tion of empirical experiments conducted in the multilingual, or more specifically
trilingual, domain. The chapter will address the question of whether the cognate
facilitation effect extends to three languages and whether cognates in three lan-
guages lead to stronger facilitation effects (cf. Lemhöfer et al. 2004; Dijkstra et al.
2010; Szubko-Sitarek 2012).
WWW.EBOOK777.COM
Chapter 5
Exploring the Cognate Facilitation
Effect in Multilingual Word Recognition
5.1 Introduction
The foregoing chapters have been devoted to the discussion of various issues related
to multilingual processing. Theoretical in nature, they have concentrated on three
main areas: the presentation of various aspects of the multifaceted concept of multi-
lingualism, the discussion of various theoretical models developed to operationalize
the mechanisms responsible for mono- and multilingual lexical processing and the
delineation of different factors influencing the multilingual lexical access. Special
attention has been given to the account of the cognate facilitation effect occurring
in the bilingual mental lexicon. The present chapter, in turn, makes an attempt to
verify the assumptions concerning the storage and processing of cognates in the
context of the trilingual lexicon. More specifically, it will be devoted to the presen-
tation of the findings of a series of experiments aimed at verifying the existence of
the cognate facilitation effect in the trilingual lexical processing.
As transpires from the foregoing overview of different lexical units used in the
research on the multilingual storage, access and retrieval, cross-language cognates
are of special interest for designing a model of the multilingual lexicon since, as
many psycholinguists claim, there is a possibility that at least part of their repre-
sentations is shared between languages. Hence the present chapter will be entirely
devoted to the analysis of the special position of cognates in the lexical retrieval in
the multilingual mind. The first part of the chapter reports on the outcomes of the
off-line studies whose aim was the analysis of the perception of cognates from the
form similarity perspective. The outcome of the study is a list of German-English
cognates as identified by the speakers of Polish (cf. Appendix A and B). A selection
of items from the list created in these two tasks was used as stimulus material in a
set of four experiments conducted with a view to exploring the language-specific or
language-integrated nature of trilingual lexical processing by examining the issues
connected with the lexical organization of cognates. More specifically, the research

DOI 10.1007/978-3-642-32194-8_5
WWW.EBOOK777.COM
130 5 Exploring the Cognate Facilitation Effect …
studies were conducted to investigate how cross-linguistic overlap in orthography

and semantics affects multilingual word recognition in different variants of the lexi-
cal decision task. The experiments will be reported in detail in the second part of the
chapter. The first experiment tested whether language nonselective access hypoth-
esis extends to trilinguals. In other words, it was set up to investigate whether tri-
linguals co-activate all three languages in a monolingual task. Experiments 2, 3 and
4 were devised to analyze the influence of different language- and user-determined
factors (such as task dependency, language proficiency and psychotypology, respec-
tively) on the storage and retrieval of cognates in a trilingual mind. Additionally,
their aim was to examine the influence of second foreign language knowledge on
native language performance in an exclusively native language context. It needs to
be noted that, since the research stimuli used in the experiments comprised only
cognate nouns, all results and their implications for connections between languages
in the multilingual mental lexicon refer to this particular group of words. Last but
not least, the conducted experiments were to verify the BIA+ approach to cognate
representations. Below, when the experiments of the present study are described, the
predictions of the BIA+ and its trilingual version (the TIA) with respect to repre-
sentational issues and task demands will be considered in more detail.
5.2 Methods for Establishing Cognate Status

in the Multilingual Context
The first step in the study, whose experiments are to be described in the present
chapter, was to create a reliable corpus of German-English cognate vocabulary. To
achieve this goal an empirical means of identifying cognate words from the psycho-
linguistic perspective had to be established. In one early research study of cognate
effects (Taylor 1976), the experimenter herself decided which words were “similar”,
“dissimilar”, and “misleading” cognates, and only used items that, in her judgment,
fell unambiguously into one of the first two categories. However, this subjective
procedure made the results of her experiments unreliable. Almost two decades later,
two more objective techniques for identifying cognates were tested. First, De Groot
and Nas (1991) proposed a similarity rating task, in which lists of translation pairs
were given to Dutch-English bilinguals, who rated the similarity of the word-pairs
on a 7-point scale. Mean ratings across a pool of participants were used to confirm
the experimenter’s own classification. A totally different procedure, widely known
as a translation-elicitation task, was introduced by Kroll and Stewart (1994) and
Dufour and Kroll (1995). The researchers gave English speakers a list of written
words in the foreign language (Dutch) asking them to guess each word’s English
translation. The idea was that a cognate word in a foreign language must formally
overlap with its native-language translation equivalent to trigger the meaning.
A contextual cue was provided, with items grouped together under their category
headings (e.g., vegetables, clothing). Those words which were correctly translated
by more than 50 % of the participants were designated as cognates.
WWW.EBOOK777.COM
5.2 Methods for Establishing Cognate Status … 131
To develop a database of cognates and noncognates for use in the psycholin-

guistic research described in the experiments reported in this chapter, as well as
to evaluate the above-mentioned techniques for rating similarity in the trilingual
context, both methods have been tested. The participants selected for the research
were divided into two groups. One group rated the similarity of translation-equiv-
alent word-pairs in German and English. The other group completed the transla-
tion-elicitation task. Additionally, mention must be made of the fact that only L2
(English) and L3 (German) words were used in both tasks. Naturally, several words
selected for the tasks bore some formal resemblance to their Polish counterparts.
Accordingly, this aspect seemed to have influenced the participants’ decisions to a
significant extent. Details concerning the L1 influence visible in the participants’
choices from both tasks will be discussed in Sect. 5.8.
5.2.1 Similarity-Rating Study
The first study was modelled on the similarity-rating technique, initially created by
De Groot and Nas (1991) and later developed by Dijkstra et al. (2010). The aim
of the study was to construct a list of words on a continuum ranging from items
identical in form and meaning, via translation equivalents that moderately resem-
ble each other with respect to orthography and phonology, to words that have a
completely different form in two languages. The participants were instructed to
base their ratings on overlapping sound and appearance between translation pairs.
The outcomes of the experiment were juxtaposed with the list of L2–L3 cognates
pre-selected by the author herself. Mean ratings across a pool of participants were
used to confirm the experimenter’s own classification. Additionally, the study was
meant to check whether the correlation between word orthographic (O) and pho-
nological (P) aspects, confirmed for L1–L2 lexical relation was also valid for
L2–L3 connection. The research was carried out in two stages. The first stage
involved the selection of the German-English word-pairs and the creation of a ran-
domised list of word-pairs. The following stage involved the collection of data.
5.2.1.1 The Participants
Ninety-eight students took part in the first stage of selection. They all filled biodata
charts, where they were asked to declare what languages they had learnt. For the
next stage only those were selected who declared at least B1 proficiency in German.
Sixty two trilingual subjects (mean age 23.4 years, SD 2.1, 48 women 14 men) took
part in the second stage of the experiment. They were all 2nd and 3rd year students
of English Department at the University of Łódź and the Higher Vocational State
School in Włocławek. First, the participants completed a language history question-
naire in which they were asked to self-assess their proficiency in the foreign lan-
guages they knew. The results of the questionnaire are listed in Table 5.1. All the
WWW.EBOOK777.COM
Table 5.1 The language experience questionnaire of trilingual participants in the similarity-

rating task
English German
Mean SD Mean SD
Number of years of learning a language 8.75 2.97 4.23 2.43
Self-rated language proficiency level in that language (1–5) 3.97 1.4 1.78 1.2
participants had Polish as their native language; they had had experience with the
English language for at least 7 years (mean 8.75, SD 2.97) and with German for at
least 3 years (mean 4.23, SD 2.43). They were not paid for participating, nor did
they receive extra course credits.
5.2.1.2 The Design of the Study
To obtain a list of test word-pairs with a variable degree of cross-linguistic ortho-

graphic and phonological overlap (from identity to no overlap whatsoever), 62 partici-
pants rated two groups of English-German word-pairs that were translations of each
other. The translation pairs to be evaluated were drawn from the CELEX database.1
They were all nouns. Only those translation pairs were selected for which the English
words consisted of 3–6 letters and for which the German counterparts had some
orthographic or phonological overlap. This selection resulted in a list of 50 candidates.
Next, another group of 50 control words which were completely dissimilar to their
translation equivalents were selected. They were all matched in target language fre-
quency and in word length with candidate words. In brief, the list consisted of two
parts of 50 words each. Within each part there were 25 translation equivalents with a
varying level of formal overlap and 25 control words. By selecting these two groups
of words, it was ensured that cognate-like items did not dominate the ultimate list of
100 words. The order of items was then changed in such a way that 2 highly similar
English-German nouns, and 2 pairs of little similarity, as judged by the experimenter,
began each list. This was done to enable the participants to begin the similarity-rating
task with confident responses. Clearly, no reference was made to the presence of cog-
nate and noncognate conditions. Finally, it has to be pointed out that some German
words selected for the study contained characters not found in the English alphabet:
three umlauted vowels: ä, ö, ü. This Spelling factor allowed the researcher to compare
similarity ratings for pairs in which the German word contained an umlauted vowel
with pairs without such characters. The reason for the inclusion of those words was
the prediction that the presence of typically German characters would strongly influ-
ence the participants’ perception. The testees were expected to rate word-pairs con-
taining characters not found in English to be less similar than those without these
characters. All the more so as such a situation had already been described for the
1 cf., http://celex.mpi.nl/scripts/entry.pl and Baayen et al. (1995).
WWW.EBOOK777.COM
bilingual context (Friel and Kennison 2001) where the subjects rated word-pairs con-
taining characters not found in their L1 to be less similar than those without these
characters.
5.2.1.3 Procedure
Sixty-two participants rated a list of German-English item pairs for cross-linguistic

similarity by choosing a similarity-rating on a 5-point rating scale. Each item in this
list was followed by 5 little squares. The participants were instructed to tick the left-
most square (1) if they thought there was no similarity at all between the words in a
pair, and the rightmost square (5) if they thought there was perfect similarity between
the words in a pair. Additionally, they were asked to use the entire range of num-
bers from 1 to 5 and, at the same time, and not to be concerned about how often
they use a particular number. To verify the correlation between the role of orthogra-
phy and phonology in the word recognition process, the same list of items had to be
rated for orthographic (O) and phonological (P) similarity. In order to lessen the par-
ticipants’ workload, a single semi-randomized list of the 100 English-German trans-
lation equivalents was divided into two parts, each containing 50 nouns. Thirty one
participants were asked to rate List 1 words for O and List 2 words for P, whereas
the other group was instructed the other way round. Because every participant only
rated one list for O and the other for P, every word was rated for O by only half of
the participants (31 subjects). The same was true for P. The participants had no more
than 15 min for the completion of each list. Therefore, they were asked to work fairly
quickly but not to be careless in their ratings.
5.2.1.4 Results and Discussion
Lists from 5 subjects had to be rejected due to too high a number of not completed
answers. To achieve the same number of answers on each aspect (O and P), the data
for one more participant were not included in the statistical analysis. Additionally,
items that were inadvertently skipped by the participants, as well as their matched
partners, were eliminated from the analysis resulting in a loss of less than 5 % of the
data. Mean similarity ratings for O and P were calculated for each word pair showing a
similar rating pattern as reflected in the high correlation (r = 0.82, p < 0.001) between
them. These results seem to suggest that from a language learner perspective, an L2–
L3 word pair with a highly similar orthography is also generally perceived as the one
having highly similar phonology and vice versa. Interestingly enough, the obtained
pattern of results was similar to the one for L1–L2 pairs. Both orthographic and pho-
nological aspects turned out to be important in the process of rating similarity between
foreign language items. It needs to be noted, however, that the similarity rating for P
was calculated on the basis of the word’s written form, which might have influenced
the judgment. Clearly, the spoken version of the similarity-rating study is needed to
establish the factual role of phonology in the process of identifying cognate pairs.
WWW.EBOOK777.COM
Table 5.2 Mean similarity-rating for the spelling factor

Without-German specific With-German specific
characters characters
Mean similarity-rating 3.30 (0.43) 1.80 (0.53)
As indicated by the data accrued in the course of this experiment, there was no
significant correlation between test items O- or P-rating scores and their English
frequency, as confirmed by correlations of 0.1 and −0.04, respectively. This sug-
gests that the participants were not influenced by the English frequency of the lexi-
cal items while they were judging the orthographic or phonological similarities of
the word pairs which seems to be justifiable considering the fact that English was
the participants’ second language and the proficiency level clearly did not allow for
this factor to influence the ratings as it happened to be the case in the research (cf.
De Groot and Nas 1991; Friel and Kennison 2001) where participants judged the
similarity of word pairs where one item came from their native language. Both the
stimuli and ratings for the list of words are tabulated in Appendix A.
As noted before, an additional aim of the presented experiment was to determine
whether words containing German-specific characters (i.e. umlauted vowels) were
responded to differently than other words. Table 5.2 displays mean similarity ratings
for the two conditions of this factor, hereafter referred to as the spelling factor.
The main effect of German-specific characters reached significance as con-
firmed by the analysis of variance (ANOVA). Similarity ratings for translation pairs
in which the German word contained a German-specific character were significantly
lower than for those in which the German word did not contain such characters
F(1, 98) = 908.11, p < 0.001. However, the spelling effect should be interpreted with
caution, as there were many more words without German-specific characters than
words with German-specific characters. Furthermore, the difference observed can
be due to some inherent distinctions between the two types of words. Namely, the
German words containing German-specific characters were lower in mean word fre-
quency than the words without such characters (39 vs. 96). It needs to be noted that
the words containing German-specific characters were not included in the following
multilingual visual word recognition experiments. However, a study employing such
words is needed to establish the role of diacritical marks both in rating the form simi-
larity and, more generally, in cognate processing.
5.2.2 Translation-Elicitation Task
To verify the validity of the similarity-rating task, another experiment was con-
ducted. As mentioned before, it employed the translation-elicitation paradigm, ini-
tially used by Kroll and Stewart (1994) and Dufour and Kroll (1995), described in
the introduction to this section. The difference again was the language combina-
tion. Unlike the original study where the participants basing on the formal resem-
blance translated foreign language words form a language they did not know into
WWW.EBOOK777.COM
their native tongue, in the experiment reported below the participants were asked
to translate unknown foreign language words into their L2 which they declared to
know only at the B1/B2 level. Consequently, to do the task they had to activate their
second language lexical store as the majority of German items did not bear any for-
mal resemblance to their Polish counterparts.
5.2.2.1 The Participants
A total of 30 English language students from the Foreign Languages Teacher

Training College and the English Institute of the University of Łódź, all of whom
had no knowledge of German (as declared in the language background question-
naires distributed before the experiment), volunteered for the study. Although 21 of
the participants spoke other languages, none had ever learnt German or any other
Germanic languages, apart from English. Their mean age was 22.3 years (SD 2.45);
there were 25 females and 5 males. Their self-rated level of proficiency in English
measured within the 5 point scale was 3.50 (SD 1.9), which was comparable to the
respondents from the similarity rating task.
5.2.2.2 The Design of the Study
The list of 100 German nouns, previously used in the similarity rating task,
was collated and grouped by category wherever possible (e.g., food; parts of
body). The items for which no clear category label could be given were sim-
ply grouped under the heading “Other Items”. All the words were presented in
the written form. The order of categories was random, as was the order of indi-
vidual words within each category. Additionally, mention must be made of the
fact that some of the selected German-English pairs bear some orthographic as
well as phonological similarity to their L1-Polish semantic counterparts. This
factor must have facilitated the participants’ choices, the outcomes of which will
be discussed in subsequent paragraphs. The subjects were asked to go through
the list of the German words one by one, and for each word, to write down what
they thought its English translation might be. They were encouraged to offer a
response for every word, even if they felt it was a complete guess, in preference
to leaving a blank. No guide to German spelling-sound relationships was offered.
Classifications were, therefore, based on orthographic rather than phonological
similarity. For practical reasons, the study was conducted in the paper-and-pencil
format. Omitted responses were scored as incorrect answers. Spelling errors
and plurals did not invalidate otherwise correct translations. The mean propor-
tion of correct responses was tallied for each item, and the results are tabulated in
Appendix B.
WWW.EBOOK777.COM
5.2.2.3 Results and Discussion
Admittedly, cognate status is a continuum, but the experimental manipulations in the

studies to be reported in this chapter required stimuli to be classified in a binary fash-
ion. Consequently, apart from some exceptions to be described below, words with cor-
rect translation rates of over 60 % were classified as cognate, while words that were
correctly translated on fewer than 40 % of trials were classified as noncognate. Words
whose translation rate fell between 40 and 60 % were initially considered indetermi-
nate. However, if they clearly had little orthographic overlap, they were subsequently
designated noncognate. Interestingly enough, several items which lack orthographic
overlap (cf. Table 5.3) were correctly identified by many participants. They were
not included in the list of cognates since the requirement taken in this study was
that true cognates must exhibit form similarity confirmed in a similarity-rating task.
Nevertheless, it is worth analysing those items more closely and pointing to possible
factors that might have helped the participants identify the translation equivalents of
the, otherwise, unknown German words.
Undoubtedly, at least some of the meanings of the German words listed in
Table 5.3 (Kartoffel, Zitrone) must have been derived through their formally similar
Polish counterparts. Clearly, the testees must have made more effort to generalise
and to draw connections between related concepts in order to generate more prob-
able sounding responses. And thus the influence of the native language pointing to
the integrated lexicon and nonselective access was unexpectedly documented. As
regards the rest of the words, it seems that a number of German items (e.g., Zimmer
or Flasche) are comprehensible even among Polish students who have never for-
mally studied the language, partly due to their being borrowings or loanwords in
the Polish language and partly because of some historical and socio-cultural bonds
between the two neighbouring countries.
By and large, since the items presented in Table 5.3 lack the form overlap which
is, by definition, a requisite for cognates, they were excluded even when correctly
identified by a substantial number of respondents. It needs to be noted, however,
that Kroll and Stewart (1994) and later Dufour and Kroll (1995) in their exhaustive
studies on the methods of identifying cognates, false cognates and noncognates did
not impose the form overlap condition. As a consequence, they ended up designat-
ing several word-pairs as cognate despite their lack of any form similarity what-
soever (e.g., citron—lemon; crayon—pencil; from the two studies respectively). In
Kroll and Stewart’s (1994) procedure, all items translated correctly by more than
50 % of English speakers were considered cognates.
Table 5.3 Words identified German word English equivalent Translation rate

by the participants as (%)
cognates probably due to
Zitrone Lemon 60
their L1 overlap
Kartoffel Potato 59
Flasche Bottle 58
Zimmer Room 57
WWW.EBOOK777.COM
Table 5.4 Frequencies and length of cognate and noncognate items

Frequency Word length
English German English- English German English-
mean mean German mean mean German
correlation correlation
Cognates 1.35 (0.77) 1.83 (0.71) 0.82 4.75 (1.93) 5.30 (1.98) 0.71
Noncognates 1.37 (0.50) 1.61 (0.62) 0.76 5.20 (2.02) 6.10 (2.51) 0.48
Another relevant factor was the choice of translation words. Admittedly, there is
more than one way to translate almost every word, especially when no context is
provided. In the present study, the responses were coded by the experimenter as cor-
rect or incorrect translations, based on whether the suggested translation would be
the first definition of a word in the English-German Duden-Oxford Dictionary.
Essentially, 47 word-pairs were deemed to be cognate, and 53 word-pairs were
identified as noncognate. As Table 5.4 shows, the cognate and noncognate sets of
words were comparable in their mean logarithmic frequencies2 and length in both
languages. As in the studies conducted for L1–L2 language combination, also in the
present experiment the cognate word-pairs were more similar in length across lan-
guages than the noncognate word-pairs which may be interpreted as reflecting the
importance of form similarity for the perception of cognates.
Basing on their research, Van Hell and De Groot’s (1998) indicated that cognate
word-pairs are closer than noncognate translation equivalents in meaning and there-
fore also in their context and frequency of use in the two languages. Interestingly,
this suggestion was also confirmed in the present study where the correlation
between word frequencies across languages was also somewhat higher for cognate
word-pairs. This finding, however, needs to be treated with due caution as it would
be difficult to prove whether the correlation was only dependent on L2 frequency.
Considering the selected items, all of them coming from basic vocabulary list, it
seems highly probable that L1 word frequency also played its role.
5.2.3 The Comparison of Cognate Identification

Methodologies in the Multilingual Context
One of the most important stages in psycholinguistic research on word process-

ing is the construction of items. To be able to draw valid and general conclusions
on the basis of an experiment’s outcome, the selection of words has to be per-
formed with the utmost care. The crucial manipulation of the stimulus words in the
experiments to be presented below concerned their cognate status. To help in the
2 Sincethe scales of measurement differ for the German and English frequency norms, values
should not be compared directly. Therefore, logarithmic frequencies were introduced. All the
data concerning the logarithmic frequencies were taken from CELEX database.
WWW.EBOOK777.COM
selection process, two independent tests, briefly presented in the foregoing sections,
were conducted. It has to be noted that following this approach, it was not neces-
sary a priori to decide if a particular item was to be considered as a cognate or a
noncognate, thus avoiding artifacts in allocating particular items to one group or
the other. The results from the two experiments correlated highly (r = 0.85), with
cognates scoring a mean similarity rating of 4.27 on a 5-point scale, while noncog-
nates scored a mean of 1.62, and led to the compilation of a reliable list of German-
English cognates as perceived from the psycholinguistic perspective.
Last but not least, it seems reasonable to refer the present results to a project by
Friel and Kennison (2001) who conducted an extensive research study aimed at ver-
ifying the effectiveness of the two methods of identifying cognates, false cognates
and non-cognates. It needs to be stressed that as their project focused on L1–L2
cognate relations the level of similarity was rated exclusively from the native lan-
guage perspective. In turn, in the tests described above, respondents were asked to
rate and translate pairs of words deriving from foreign languages which primarily
necessitated the activation of the second language lexicon and interestingly did not
exclude the activation of the first language store. In the following paragraphs the
results of the present tests will be compared to the outcomes of Friel and Kennison
study.
Of the 35 stimuli common to the present sets and those used by Friel and
Kennison (2001), classifications agreed for 13 cognates and 22 noncognates. The
four discrepant items were: milk, frog, mouse and stone. Some points should be
made as to the source of the discrepancies. All of the items were clearly cog-
nate in the present study (with translation rates of at least 70 %), but noncognate
in Friel and Kennison’s study. All these words fall within the cognate range in
Friel and Kennison’s similarity ratings, confirming that they are quite similar in
for. Also, the correct translation was given for those items more often than any
other translation, but the rate simply failed to exceed 50 % which was the border
line. The possible reason may be connected with the fact that when no context
was provided, some participants were reminded of other orthographically similar
words instead. Providing a context—the category groupings in this study—was
apparently sufficient to favour the correct translation and consequently it facili-
tated the task. On the other hand, Friel and Kennison’s (2001) primary reason for
randomising, rather than categorising, their lists of stimuli was that many of their
items were abstract words which could not clearly be categorised. Undoubtedly,
the categorised items may in some sense have had an unfair advantage over the
“other” items in the presented study, not so much by a process of elimination as
Friel and Kennison suggest, because the lists were by no means exhaustive of cat-
egory items, but simply by favouring the potential translation which falls within
the category. Another difference between the studies is that all the items which
Friel and Kennison (2001) found to be cognates (i.e., items identified correctly by
more than 50 % of participants) did overlap in form. That is, there were no items
for which correct translations were often produced despite the lack of form simi-
larity, such as the Zitrone-lemon pair. The possible explanation for this discrep-
ancy was already discussed in the previous paragraphs. In sum, two previously
WWW.EBOOK777.COM
described methods for identifying cognates were tested in a new trilingual con-
text. They both were shown to yield similar results. The resulting database of
German-English cognates was further used in a set of experiments devised to ver-
ify the nonselective access hypothesis in the trilingual context. The results of the
experiments are reported in the following sections.
5.3 Investigating Cognate Effects in Trilingual

Visual Word Recognition
The previous chapters discussed problems related to models and hypotheses which
propose how lexical items are stored and processed by monolinguals, bilinguals
and, most importantly, multilinguals. As has been shown, the question of whether
words from different languages are stored in one or more containers has recently
been replaced by the issue of how words from different languages in a multilin-
gual person’s lexicon are accessed. The question of whether there is one single
integrated lexicon (cf. Cook 1996), two or more separate lexicons (cf. Smith
1997; Singleton 1999, 2002) or two or more subsystems under one big system
(cf. Paradis 2004) ceased to be in the centre of researchers’ attention. Instead, the
central question most of the multilingual research is concerned with nowadays is
language-selective versus language-nonselective access (cf. de Bot 2004, p. 17).
As has been documented in the theoretical chapters, the majority of empirical evi-
dence gathered in psycholinguistic and neurolinguistic studies seem to support
the contention that during bilingual lexical access, even if the two languages are
indeed represented differently, both are activated, although perhaps to different
degrees (cf. De Groot and Nas 1991; Dijkstra et al. 1999; De Groot et al. 2000;
Jared and Kroll 2001; De Groot 2002; de Bot 2004; Dijkstra 2007). According to
the BIA+ model (Dijkstra and Van Heuven 2002a), which constitutes the theo-
retical background for a number of lexical processing experiments conducted
within the bilingual domain, the visual presentation of a word to a bilingual per-
son leads to parallel activation of orthographic input representations both in the
native language (L1) and in the second language (L2). These representations then
activate associated semantic and phonological representations, leading to a com-
plex interaction (or resonance process; Dijkstra and Van Heuven 2002a, p. 183)
between codes from which the lexical candidate corresponding to the input word
emerges and is recognized. Analogously, in the case of trilingual lexical process-
ing research, many linguists (cf. Dijkstra 2003a, b; Li Wei 2003, 2006; Hall and
Ecke 2003; Dijkstra 2007) argue for a unified lexicon with language nonselective
access. Some of them, like Dijkstra (2003a, b, Experiment 1 in the present chapter)
focus on visual perception, whereas others, e.g., Li Wei (2003, 2006), deal with
production. Furthering the question of the nonselectivity of language access, it
needs to be stressed that it can be influenced and, as some linguists claim, lim-
ited by a number of factors. Indeed, previous investigations identified several vari-
ables influencing the lexical organization of non-native languages and the manner
WWW.EBOOK777.COM
in which words in non-native languages (L2, L3, etc.) are linked to one another
and to their translation-equivalents in the native language (L1) (cf. De Groot and
Hoeks 1995; Grosjean 1997; Dijkstra et al. 1999; Grosjean 2008). In studying the
interaction of multilinguals’ languages at least three factors seem to influence the
strength of connections between orthographically and semantically overlapping
words. First, task demands and experimental stimulus materials. Second, the mul-
tilingual’s relative language proficiency. Third, the perceived similarity between
languages (psychotypological effect, cf. Kellermann 1983).
The question whether the three languages known to the learner simultane-
ously influence visual word recognition performance has been addressed in
Experiment 1. To test the influence of the above mentioned factors on the mul-
tilingual processing, three separate experiments (Experiments 2, 3 and 4) have
been conducted. As regards the first factor pertaining to the task demands and
the language of the stimulus material, there is clear evidence in the literature
that task demands can affect multilingual performance to a considerable extent.
Many researchers even claim that it may be inappropriate to talk about multilin-
gual word recognition in general without specifying the precise task and experi-
mental circumstances under which it takes place because performance is both task
and context dependent (cf. De Groot et al. 2000; Dijkstra et al. 2000; De Groot
et al. 2002; Dijkstra 2007; Lemhöfer and Radach 2009). Similarly, according to
Dijkstra’s BIA+ model, the result patterns for cognates will differ depending
on task demands. To verify this assumption for a trilingual lexicon, the results
of Experiment 2 will be contrasted with the data from Experiment 1 and will
undergo further analysis. Clearly, to achieve reliable results, in both experiments
the same group of stimulus words will be tested. The second factor in the dis-
cussion of whether knowledge of one language affects performance in the other
language concerns the relative language proficiency. According to Van Hell and
Dijkstra (2002, p. 783), “relative language proficiency will affect the bilingual’s
sensitivity to L1 interference when he/she is processing in L2, and the sensitiv-
ity to L2 interference when processing in L1”. The underlying rationale here
is that less activation is needed to recognize words that are used relatively fre-
quently, as are words in a language in which the speaker is relatively proficient.
Experiment 3 analyzes the role of the second foreign language proficiency on
the organization of cognates, which is studied in greater detail by manipulating
the subjects’ proficiency in the nontarget languages. In Experiments 1 and 2, tri-
linguals were most fluent in their L1 (Polish), less fluent in their L2 (English),
and least fluent in their L3 (German). In Experiment 3 trilinguals with a higher
level of proficiency in German were tested. These trilinguals’ proficiency levels
in L2 and L3, as declared by the participants themselves, were comparable. It was
expected that the level of proficiency may exert some influence on the strength
of interlingual connections. Additionally, numerous previous studies have shown
that processing in the weaker language, L2, can be influenced by knowledge of
the stronger language, L1 (e.g., De Groot et al. 2000; Dijkstra et al. 1998, 1999;
Van Heuven et al. 1998; Grosjean 2008). In a divergence from these studies,
Experiments 2 and 3 test the effects of L2 and L3 on L1 in an exclusively native
WWW.EBOOK777.COM
5.3 Investigating Cognate Effects in Trilingual Visual Word … 141
language context. The rationale behind these experiments is that a minimal level
of nontarget language proficiency seems to be required before any weaker lan-
guage effects become noticeable in L1 processing. Finally, the third factor that
influences the interaction of the multilingual’s languages to be focused upon in
this chapter, is the perceived similarity of the languages involved. Hall and Ecke
(2003) explained transfer-related mistakes by way of links which are formed
across languages at the word or concept level. They proposed that the links
are formed on the basis of similarity between items already well known to the
speaker (belonging to L1 and/or L2) and the newcomers from the L3. They also
suggested that more transfer between a speaker’s L3 and L2 than between his/her
L3 and L1 might be a reflection of a link between foreign languages in the mind.
Moreover, according to Singleton (2002, 2003), it seems plausible that in the case
when L3 is typologically closer to L2 than to L1, in the early stages of the pro-
cess of acquiring L3, L3 forms will be linked to L2 ones rather than to those of
L1. Singleton (2003) implies also that with growing L3 proficiency, strong lexical
connections develop between L2 and L3 items. The connections get even stronger
under the influence of psychotypology (cf. Kellermann 1983). Experiment 4 aims
to verify the role of perceived similarity supported by the L2 status factor in the
organization of the trilingual lexicon.
As regards the methodology, what is worth mentioning is the fact that the bulk
of the studies investigating transfer, which have been reported on in the TLA lit-
erature and used as a basis for models of multilingual mental lexicons, employed
off-line methodologies—pen-and-paper formula, such as e.g., questionnaires.
Such methodologies offer more time and thus allow for extensive use of strate-
gies by participants. In contrast, in the experiments reported in the present dis-
sertation, an on-line methodology was adopted since it enables the researcher to
investigate processing in real time. Unlike off-line studies, on-line experiments
do not allow for the same amount of conscious control and thus might be more
potent in terms of shedding light on the composition of languages in the mul-
tilingual mind. In this study, an experimental setup has been chosen for which
the most reliable and frequently replicated bilingual cognate effects have so far
been obtained, namely a lexical decision task. The logic in using this type of task
is that it requires individuals to search their lexicons for a lexical representation
that matches the letter string presented. The representation of a lexical item con-
tains information regarding the word’s orthography, phonology, and semantics,
and these aspects of the word are retrieved during the task. Finally, it needs to be
pointed out that Experiments 1–4 were conducted in the order of presentation and
each next experiment was designed on the basis of the conclusions drawn from
the previous one. In the subsequent paragraphs, a more detailed outline of the
rationale behind the design of the experiments as well as details concerning the
participants, materials, procedures and results will be provided, Next, a general
discussion will follow.
WWW.EBOOK777.COM
5.4 Study 1: Exploring the Cognate Effect

in Trilingual Visual Word Recognition
from the Weakest Language Perspective
Research on bilingual word recognition suggests that lexical access is nonselective

with respect to language, i.e., that word representations of both languages become
active during recognition (cf. Chap. 3). However, until now, it is still uncertain
whether the large degrees of interlingual interaction found for bilinguals extend
to multilinguals. One piece of evidence supporting nonselective access is that
bilinguals recognize cognates faster than noncognates (cf. Chap. 4). In fact, any
difference between how cognates and “monolingual” words are processed by mul-
tilinguals would indicate that the other, currently irrelevant, language must have
played a role as well, at least as long as the two groups of words are compara-
ble with respect to all dimensions other than language membership. To investigate
whether the nonselective access hypothesis holds also for trilinguals and three lan-
guages, Polish-English-German trilinguals were asked to carry out a lexical deci-
sion task in their third (weakest) language—German (cf. De Groot et al. 2002;
Dijkstra et al. 1998, 1999; Lemhöfer and Radach 2009). In the present experiment,
the word materials included purely German control words, “double” cognates that
overlapped in Polish and German, but not in English, and “triple” cognates with
the same form and meaning in Polish, German, and English (cf. Lemhöfer et al.
2004). While the triple cognates came from the studies reported in Sect. 5.2, the
list of double Polish-German cognates was taken from the former unpublished
studies of the present author, which applied the above mentioned methodologies
to identify the level of orthographical overlap between Polish and German lexical
items.
5.4.1 Research Hypotheses
First, it was expected that the participants would react faster and more accurately
to German-Polish cognates than to German control words. Such a finding would
replicate the standard cognate effect (cf. Djikstra and Van Heuven 2002a, b) for a
new language combination and provide additional evidence in support of language
nonselective access. Second, an even stronger view of language nonselective access
would be supported if three languages at a time can influence word recognition. In
that case, the cognate status of the stimuli with respect to English should have an
additional effect on top of the standard cognate effect. In other words, the recog-
nition performance for German-Polish-English cognates should be even faster and
more accurate than that for German-Polish cognates. If that turned out to be true,
the experiment would replicate the Lemhöfer et al. (2004) study conducted for a
different combination of languages and add further evidence to the discussion
on the nonselective visual word recognition in the multilingual mental lexicon.
WWW.EBOOK777.COM
5.4 Study 1: Exploring the Cognate Effect in Trilingual … 143
Specifically, the study reported in the present section sought to address the follow-
ing hypotheses:
H1. If there is language nonspecific selection in lexical access, faster RTs will be
found for Polish-German cognates than for German control words.
H2. If language nonspecific selection in lexical access extends to three languages,
RTs found for “triple” Polish-English-German cognates will be even shorter
than those obtained for “double” Polish-German cognates.
5.4.2 The Participants
The experiment involved 27 trilinguals with Polish as their mother tongue (L1),
and English and German as their second (L2) and third language (L3), respec-
tively. All the participants were “unbalanced” trilinguals; i.e., they were not as
proficient in their second and third languages—English and German—as in their
mother tongue—Polish. All of the testees were students (second- and third-years)
in the Institute of English Studies at the University of Łódź. All of them had learned
English and German as a foreign language at school. English from about the age
of 9, German from about the age of 14. The participants’ L2 competence was not
tested before the study. It was assumed that their status of students of English
Philology ensured advanced L2 proficiency. Before the experiment, however, a lan-
guage background questionnaire was administered, in which, 31 prospective partici-
pants were asked to report on their competencies in all of their foreign languages.
4 students had to be excluded from the experiment due to their relatively advanced
knowledge of Spanish that could affect the results. Additionally, a few respond-
ents declared some competence in French, Italian or Russian, but those selected
for the experiment described their competence as very low. To ensure that the tes-
tees constituted a homogenous sample as far as their L3 (German) competence is
concerned, only those students were selected who described their L3 level as pre-
intermediate (A2) or intermediate (B1), as specified by the Common European
Framework for Languages descriptors (henceforth CEFR). A summary of the par-
ticipants’ learning history is given in Table 5.5.
The data from 3 participants had to be excluded due to their faulty performance,
which sets the mortality rate of the experiment at 11 %. The remaining 24 partici-
pants were between 20 and 24 years old with the mean of 21.66. 17 were female,
7 were male. All the participants were right-handed. They were not paid for their
participation, nor given any course credits.
Table 5.5 The language experience questionnaire of the trilingual participants in Experiment 1

English German
Mean SD Mean SD
WWW.EBOOK777.COM
5.4.3 Procedure
The experimental session took about 30 min during which the participants carried
out a German lexical decision task involving triple Polish-English-German cognates
(hereafter referred to as PGE cognates), double Polish-German cognates (PG cog-
nates) and German control words. Testing took place individually on a PC; a modi-
fied program based on the Reaction Time Instrument Builder3 was used. The
subjects were seated at the 17-in. computer screen, where stimuli were presented in
black 14 point uppercase TIMES NEW ROMAN on a white background. One but-
ton on the side of the dominant hand of the participants was assigned to the “yes”
response, the other button to the “no” reaction. At the beginning of each trial, a fixa-
tion point appeared in the middle of the screen for 1 s, then a blank interval of
500 ms followed. Next, the test word appeared. The item stayed in view until a
response had been provided or until a time-out of 5 s had passed. The next trial was
started 500 ms after the response was given. All items were presented in uppercase
letters, because in German the case of the first letter can be a cue for the syntactic
class of a word (nouns are written with a capital). The experiment consisted of two
blocks of 48 items each. The first two items of each block were warm-up items (a
nonword and a German filler word) which were not included in the analyzes. The
participants took a short break (5 min) between the blocks. The order of items
within the lists was pseudo-randomised with no more than four words or nonwords
in a row. In order to provide an exclusively L3 context (cf. Grosjean’s language
mode theory; Grosjean 1989), before the test all the participants were asked to read
a short German text. Additionally, the task instruction explaining that in this task the
testees were to decide as quickly and as accurately as possible whether a letter
string appearing on the screen was an existing German word or not was also written
in German. A 5-min-long training session preceded the experimental part.
5.4.4 Materials
As noted before, one of the most important stages in psycholinguistic research on

word processing is the construction of items. To be able to draw valid and general
conclusions on the basis of an experiment’s outcome, the selection of words has to
be performed with utmost care. The crucial manipulation of the stimulus words in
the present experiment concerned their cognate status. Three critical groups of L3
words (48 words in total) were compared with respect to the latencies and accuracy
of their recognition: one quarter of the L3 words were cognates with their transla-
tions in (L1) Polish (e.g., DACH, meaning roof in both Polish and German), one
quarter of the L3 words were cognates with their translations in both (L1) Polish
3 www.er.uqam.ca/nobel/r21270/react_time; downloaded on the 2nd March, 2007; modified with

the help of an external experimenter in March 2007.
WWW.EBOOK777.COM
Table 5.6 Characteristics of the stimulus items used in Experiment 1

Item type Mean Mean Mean written
number of number German English Polish
letters of syllables frequency frequency frequency
per million per million per million
PGE 4.66 1.41 39.72 41.33 37.17
cognates
PG cognates 4.88 1.52 42.09 – 40.03
German 4.58 1.50 41.79 – –
controls
German 4.58 1.33 43.48 – –
fillers
and (L2) English (e.g., PLAN) and the remaining words were noncognates—
German control words that were different from both their Polish and English trans-
lations (e.g., KOPF, meaning Eng.: head, Pl.: głowa). In addition to the test stimuli,
2 German words as well as 2 PGE cognates, 2 PG cognates and 2 nonwords, all
different from any of the test stimuli, were selected as practice items. In the word
materials only nouns were used because they are the only content words that pos-
sess the same lemma form in all three languages. Verbs and many adjectives are
morphologically marked by suffixes in both German and Polish (e.g., sing-en or
śpiew-ać, meaning to sing), while they are not marked in English, which usually
results in different lemma forms of these words in the three languages. All stimu-
lus items are listed in Appendix C and The word characteristics are presented in
Table 5.6. The two groups of stimuli: cognates and control words will be described
separately.
5.4.4.1 Cognates
12 Polish-German-English cognates were selected from CELEX database (cf. Baayen

et al. 1995). They possessed the same spelling and meaning in all three languages, like
the word PLAN. All of them were singular forms of nouns with a length of between 4
and 6 letters and no more than two syllables. It was attempted to choose only cognates
with both English and Polish frequencies high enough to potentially affect the responses.
While in German and Polish, the chosen items are used as nouns only, it was inevitable
that in English, some of these nouns are also used as verbs. However, in all cases the verb
meaning was closely related to the noun meaning (e.g., PLAN, ECHO), so that semantic
competition can be largely ruled out. For the group of Polish-German cognates 12 nouns
with orthographic and semantic overlap in Polish and German (e.g., DACH, PECH) were
selected that matched the Polish-English-German cognates with respect to length, num-
ber of syllables and frequency. Matching took place on an item-by-item basis with sup-
port of the Statistica 9 software. All cognate translations had a nonidentical phonology,
whereas 41 % of the PG cognates had a nonidentical orthography. Additionally, only such
cognates were selected that in CEFR lists are assigned to no higher than B1 level.
WWW.EBOOK777.COM
5.4.4.2 Nonwords
A number of nonwords that was equal to the number of words (48) was generated
by changing one or more letters in an existing German noun of 4–6 letters. All non-
words were orthographically legal in German and they did not exist as words in any
of the three languages. They were all created with the help of WordGen (cf. Duyck
et al. 2004), an online program that uses the CELEX and Lexique lexical data-
bases for word selection and nonword generation in Dutch, English, German, and
French. 14 items were adopted from the studies by Lemhöfer and Dijkstra (2004)
and Lemhöfer et al. (2004). The nonwords were matched to the word items in terms
of their mean length and number of syllables—the mean number of letters 4.68 and
the mean number of syllables 1.54.
5.4.5 Results and Discussion
For the RT analysis, only correct responses were considered. The overall error rate
amounted to 9.4 %, including 8.1 % on test words. Furthermore, reaction times
exceeding two times the standard deviation from the item mean counted as outli-
ers and were excluded from the set of valid responses. Outliers accounted for 1.7 %
of all the responses and they were approximately equally distributed across condi-
tions. Erroneous responses on test words (8.1 %) were excluded from the analysis,
so that 9.8 % of the data were discarded in total. The data on the erroneous response
matched item partners were not excluded since such a procedure could have led
to too high a percentage of the excluded data points. Moreover, in many studies
employing similar methodology (e.g., Lemhöfer et al. 2004; Van Hell and De Groot
2008) results obtained with and without the exclusion of the data on the outliers and
erroneous response matched item partners yielded comparable results.
Similarly to the Lemhöfer et al. (2008) study, error rates and RTs were analyzed
over participants only, because the selected cognates and controls were matched item-
by-item and can be seen as an almost exhaustive set of items with the given restric-
tions. The data gathered in the experiment were entered into the Statistica 9 program
for statistical analysis. The relevant data are summarised in Table 5.7.
Analysis of the obtained RTs indicates that all three languages became activated
and influenced the subject’s responses to the following targets. As can be seen from
the data illustrated in Figure 5.1, the participants responded much faster (743.17 ms)
Table 5.7 The mean RT SD ER SD
RTs and error rates for
PGE cognates 743.17 25.74 5.2 3.7
participants in Experiment 1
PG cognates 765.92 32.74 7.8 4.1
German controls 819.58 66.48 11.4 6.2
German fillers 823.17 63.15 15.0 8.1
nonwords 897.73 91.16 7.7 4.9
WWW.EBOOK777.COM
Fig. 5.1 Mean RTs obtained 819

for PGE cognates, PG 820
cognates and control words in
Experiment 1 800
780
765 3*(FRJQDWHV
ms
760 3*FRJQDWHV
743 FRQWUROV
740
720
700
to PGE cognates than to PG cognates (765.92 ms) and to control words (819.58 ms).

These data indicate that both English and Polish were activated in the course of the
experiment and that the word status influenced lexical access.
To investigate the influence of item type on RT, paired samples t-test was con-
ducted whose results are presented in Table 5.8. In brief, the obtained data point
to the statistically significant activation of PGE and PG cognates. Planned com-
parison showed that PGE cognates were recognized 76.42 ms faster than mono-
lingual German control words. This difference turned out to be statistically
significant (t(11) = 3.94; p < 0.005). Similarly, PG cognates were recognized
53.67 ms faster than control words. This difference also reached statistical signifi-
cance (t(11) = 2.92; p < 0.05). Finally, cognate effect obtained for PGE cognates
was bigger (22.75 ms) than that obtained for PG cognates, which was significant at
t(11) = 2.36; p < 0.05.
Also the analysis of RTs with ANOVA yielded similar results. There was a sig-
nificant main effect of item type, F(2, 33) = 9.004, p < 0.01. Moreover, planned
comparison showed that PGE cognates were recognized faster than PG cog-
nates, F(1, 46) = 7.162, p < 0.01. Finally, PG cognates were responded to sig-
nificantly faster than control words, F(1, 46) = 12.584, p < 0.01. Similarly, the
analysis of error rates revealed a comparable pattern. As can be seen in Fig. 5.2,
item type significantly influenced error rates; F(2, 33) = 5.062, p < 0.05. Also
planned comparisons indicated that there were significantly more errors on PG cog-
nates than on PGE cognates, and more on German controls than on PG cognates.
Both differences turned out to be significant: F(1, 46) = 17.697, p < 0.001 and
F(1, 46) = 5.319, p < 0.05, respectively.
Table 5.8 Paired samples t-test for the three types of stimuli words (PGE cognates, PG cognates
and German controls) in a German lexical decision task (Experiment 1)
Paired samples Mean SD SME t df Sig.
(2-tailed)
PGE versus control −76.42 67.26 19.42 −3.94 11 0.002
PG versus control −53.67 63.67 18.38 −2.92 11 0.014
PGE versus PG −22.75 33.45 9.65 −2.36 11 0.038
WWW.EBOOK777.COM
Fig. 5.2 Mean ERs obtained 20

for PGE cognates, PG
cognates and control words in 15
Experiment 1 11.4
ER (%) 10 PGE cognates
7.8 PG cognates
5.2 controls
5
Statistically significant cognate effects obtained for PGE and PG cognates con-
firmed by RTs and ERs analyzes conducted both with paired-samples t-test as well
as one-way ANOVAs point to language nonspecific selection which extends to three
languages. Thus, both hypotheses have been confirmed. First, the “standard” cog-
nate effect in lexical decision for a new language combination (Polish and German)
was replicated. Polish-German cognates were responded to faster than exclusively
German control words. Second, an additional cognate effect on top of the standard
cognate effect could be demonstrated for the trilingual population: words that had
the same form and meaning in all three languages (Polish, German, and English)
were recognized even faster and more accurately than the matched Polish-German
cognates with a dissimilar English translation. This indicates that during the recog-
nition of words in a given foreign language, not only the mother tongue, but also
another non-native language (English) exerts influence on recognition performance.
Admittedly, the obtained effects cannot be explained without the involvement of
all three languages: If the participants had selectively activated their German lex-
icon, there should have been no cognate effect whatsoever; if they had only acti-
vated the relevant lexicon (German) and their native language (Polish), there would
not have been any RT difference between PG and PGE cognates. Not only do the
present data reveal that the cognate effect can accumulate over languages but they
also show that co-activation of three lexicons occurs even within the same words.
It seems that even with a constraining experimental setting (German task, German
instruction) the trilingual participants did not manage to completely deactivate their
L1 and/or L2. Both of these languages became activated when the participants saw
the triple cognate word and both of these languages significantly affected lexical
decision times in reaction to the presented cognates.
Clearly, the presented interpretation of the obtained results could be challenged.
Firstly, it could be claimed that the difference between PGE and PG group might have
been caused by the fact that not all words in the PG group were identical cognates. It
needs to be noted here that non-identical cognates are reported to be recognized more
slowly (cf. Lemhöfer and Dijkstra 2004; Dijkstra 2007). However, an additional anal-
ysis conducted exclusively on the pairs of identical cognates repeated the results as
confirmed by one-way ANOVA F(1, 12) = 4.986, p < 0.05. Secondly, it could be
argued that the triple cognate effect might have been obtained due to the higher level
of proficiency in L2—English. Similarly, the lower level of L3 might have led to
stronger L1–L3 connections triggering faster RTs for PG cognates. Doubtless, a
WWW.EBOOK777.COM
similar experiment with participants declaring comparable, advanced levels of L2 and

L3 proficiency would additionally help to verify the hypotheses. Finally, although the
results do show that it is possible to have activation of three languages simultaneously,
the present data alone leave the possibility that they may not all be activated in less
favourable circumstances, for instance if the target4 language is L1 or if the partici-
pants’ proficiency in their foreign languages changes. The influence of both these fac-
tors is to be verified in the subsequent experiments.
5.5 Study 2: The Role of Task Demands in the Trilingual

Processing of Cognates
If the mental lexicon of a multilingual is organized on the basis of item characteristics

as the outcomes of Study 1 seem to indicate, words from all the known languages
might be activated in response to incoming information. Interestingly enough, a num-
ber of previous studies, all of whose results were interpreted in terms of the language
nonselective view, differed in the languages that were relevant for task performance
and/or the languages of the stimuli with which multilinguals (or, in fact, mainly bilin-
guals) were presented. In some studies, bilinguals were shown stimuli in both lan-
guages, and they had to respond to items from both languages (e.g., Dijkstra et al.
1998; Lemhöfer and Dijkstra 2004) or only to items from one language (e.g., Van
Heuven et al. 1998; Dijkstra 2003a). In contrast, in some other studies the stimulus
list and the task demands employed words from one target language only. Definitely,
the majority of studies were conducted in the first foreign language setting (L2; De
Groot et al. 2000; Dijkstra et al. 1998; Van Heuven et al. 1998). However, some
studies employing the typically less dominant second foreign language have also
been reported of late (L3; Lemhöfer et al. 2004; Lemhöfer and Radach 2009; Study
1 above). Importantly, Van Hell and De Groot (2008), propose that the influence of
one language on the other would be best analysed in an experimental setting in which
multilinguals perform a lexical decision task exclusively in their dominant language
(L1), and in which they are presented with L1 words. Admittedly, taking the nonselec-
tive access view to an extreme would imply that “words from both languages are acti-
vated even when the bilinguals are performing in their native and dominant language
and in a purely native language context” (Van Hell and Dijkstra 2002, p. 782). It needs
to be emphasised, however, that for lexical decision tasks performed in L1, the find-
ings are far from clear. Caramazza and Brones (1979) failed to find a cognate effect
in the dominant-language task, so did Gerard and Scarborough (1989). Other authors,
on the other hand, reported some effects of L2 knowledge on L1 word performance
under similar circumstances (cf. De Groot et al. 2000; Van Hell and Dijkstra 2002;
De Groot and Van Hell 2005; Lemhöfer et al. 2008). More importantly, Van Hell and
Dijkstra (2002) documented a response time (RT) advantage for lexical decisions on
native language words that were cognates with respect to a second language and even
4 The word target is used to denote the language of the task.
WWW.EBOOK777.COM
for decisions on words that were cognates with respect to a third language. On the
whole, however, the typically stronger cognate effects from L1 on L2 than from L2
on L1 prove that L2 representations are generally activated less strongly or less rap-
idly than L1 representations, implying that they have less chance to affect the response
when L1 is the target. In other words, when cognates are processed in a second or
third language context, the first-language reading not only becomes active but it facil-
itates recognition as well (cf. Dijkstra 2007; Lemhöfer et al 2008; Study 1 above).
Consequently, it seems reasonable to claim that, in the previous experiment, before
the L3 (German) target reading of a cognate became active, the L1 (Polish) reading
had already affected target processing. In contrast, in the Polish language-specific lex-
ical decision task, multilinguals can be assumed to respond to the first reading of the
cognate they identify, which will often be their L1 reading. Thus, the cross-linguistic
effect measured relative to Polish controls is expected to be considerably smaller. In
fact, the question arises whether cognate effects are still present in the recognition of
L3 cognates in a first-language context and if that be the case, will the cumulative
influence of the stronger and the weaker foreign language (L2 and L3) lead to even
faster L1 recognition, thus adding evidence to the nonselective access extending to
three languages. Taken together, the issue under investigation is whether the cognate
effects found for PGE and PG cognates in the German task relative to German (L3)
control words (cf. Study 1) can also be demonstrated in comparison with Polish (L1)
control words in a purely L1 (Polish) task. If so, semantic and orthographic overlap of
cognates should cause facilitatory effects on word recognition. Alternatively, effects
caused by form similarity may be task dependent and nonsignificant relative to the
Polish control words. As Lemhöfer et al. (2008) rightly observe, this option would
indicate that whereas the simultaneous activation of an L1 code affects the recognition
of words in the second and/or third language, the reverse is not the case.
5.5.1 Research Hypotheses
The present study set out to investigate the generality of the language-nonselective
access hypothesis. More precisely, by changing the language of the task from L3 to
L1 the present study, in contrast to the experiment reported in the previous section,
tested the dependence of lexical retrieval on task demands, as well as the influence
of the foreign languages (L2 and L3) on the native language visual word recogni-
tion. Specifically, the following hypotheses were addressed:
H1. If the nonselective access hypothesis holds true in the exclusively native lan-
guage context and the weaker foreign language is potent enough to influence
the dominant language processing, performance will lead to different result pat-
terns for cognates than for noncognates. Faster RTs will be found for Polish-
German cognates than for Polish control words.
H2. If language nonspecific selection in lexical access in the exclusively
native language context extends to three languages and foreign languages
are potent enough to influence the dominant language processing, triple
WWW.EBOOK777.COM
5.5 Study 2: The Role of Task Demands in the Trilingual Processing … 151
Polish-English-German cognates will be processed even faster than double

Polish-German cognates.
H3. If nonselective access is task dependent, different RTs will be found for the
same group of double and triple cognates depending on the language of the
task. PGE and PG cognates will be processed faster in comparison to their
matched controls in the German decision task than in the Polish context.
Nineteen participants (14 women and 5 men), drawn from the group involved in
Experiment 1, took part in the present study. Five students from the previous experi-
ment did not participate in the present task. The two sessions were conducted within
the interval of 6 weeks.
5.5.3 Procedure and Materials
The apparatus and procedure of the lexical decision task in the present experi-
ment were similar to those used in Experiment 1. The stimulus materials consisted
of triple and double cognates from the previous study. However, since the lan-
guage of the task was L1, German control words were replaced with Polish con-
trols. Additionally, orthographically non-identical PG cognates were transformed
into their Polish equivalents. As in Experiment 1, the three groups of words (i.e.,
cognates with English and German, cognates with German, and control words)
were matched item-by-item for Polish frequency, length and number of syllables.
A group of Polish fillers (12) and a group of nonwords (48) were also included.
Again, in the word materials only nouns were used. All stimulus items are listed in
Appendix D and word characteristics are presented in Table 5.9.
5.5.3.1 Control Words and Fillers
For the present Polish language specific lexical decision experiment, only Polish
control words that were purely Polish nouns were included. They were noncog-
nates, and resembled neither their English, nor their German translation in either
orthography or phonology; examples are Polish NOGA (Ger.: Bein, Eng.: leg) or
PIES (Ger.: Hund; Eng.: dog). In order to keep the proportion of cognates in the
experiment at no more than 50 % of the words 12 additional pure Polish fillers were
included with similar characteristics as the Polish control words.
WWW.EBOOK777.COM
Table 5.9 Characteristics of the words used in Experiment 2

number of number of German English Polish
letters syllables frequency frequency frequency
PGE 4.66 1.41 39.72 41.33 37.17
cognates
PG cognates 4.88 1.52 42.09 – 40.03
Polish 4.4 1.75 – – 40.95
controls
Polish fillers 4.58 1.75 – – 43.12
5.5.3.2 Nonwords
A number of nonwords that was equal to the number of words (48) was constructed
by changing one or more letters in an existing Polish noun of 4–6 letters. All non-
words were orthographically legal in Polish. They did not exist as words in any of
the three languages. The nonwords were matched to the word items in terms of their
mean length and number of syllables—the mean number of letters 4.53 and the
mean number of syllables 1.47.
For the analysis of RTs, only correct reactions were considered. The overall error
rate was 7.4 %. Furthermore, RTs that lay more than two standard deviations away
from the item mean were considered outliers. The percentage of outliers among the
correct trials was 2.3 %. Errors on test words accounted for 5.9 %. In total 8.2 %
of the data were excluded. The mean RTs, standard deviations, and error rates are
listed in Table 5.10.
As can be inferred from the data illustrated in Fig. 5.3, the participants
responded faster (728 ms) to PGE cognates than to PG cognates (746 ms) and con-
trol words (757 ms).
To investigate the influence of item type on RT, paired samples t-test was con-
ducted whose results are presented in Table 5.11. Planned comparison showed that
RTs and error rates for the
PGE cognates 728.58 21.53 4.2 3.1
PG cognates 746.33 43.22 5.3 5.3
Polish controls 757.17 36.24 8.4 6.2
Polish fillers 760.03 39.33 9.1 7.7
Nonwords 792.15 47.15 10.2 11.1
WWW.EBOOK777.COM
Fig. 5.3 Mean RTs obtained for PGE cognates, PG cognates and control words
PGE cognates were recognized 29 ms faster than monolingual Polish control words.
This difference turned out to be statistically significant (t(11) = 4.11; p < 0.005;
F(1, 36) = 8.740, p < 0.005). PG cognates were recognized only 11 ms faster than
control words. This difference failed to reach statistical significance (t(11) = 0.79;
p = 0.44). Similarly, the difference between PGE cognates and PG cognates, turned
out to be not statistically significant (t(11) = 1.77; p = 0.10). Similarly, the analysis
of error rates revealed the same pattern. The fewest errors were made on PGE cog-
nates, more errors were made on PG cognates, and the most errors were made on
Polish control words. Nevertheless, only pairwise comparisons for PGE cognates
and Polish control words gave statistically significant results (F(1, 22) = 4.405;
p < 0.05).
Interestingly enough, only the group of PGE cognates compared with their
matched controls yielded statistically significant results both in RTs and ERs analy-
sis. Since the difference in processing PG cognates and controls did not reach the
significance level, nor did the difference between PGE and PG cognates, it may be
argued that the weaker foreign language does not affect visual word recognition in
the exclusively native language context. Consequently, Hypotheses 1 and 2 could
not be corroborated. No significant influence of the weaker foreign language on the
dominant language processing has been reported. Nor was any data found that could
Table 5.11 Paired samples t-test for the three types of stimuli words (PGE cognates, PG cog-
nates and Polish controls) in a Polish lexical decision task (Experiment 2)
(2-tailed)
PGE versus control 28.58 24.03 6.94 4.11 11 0.001
PG versus control 10.83 47.36 13.67 0.79 11 0.44
PGE versus PG 17.75 34.62 9.99 1.77 11 0.10
WWW.EBOOK777.COM
support Hypothesis 2, according to which nonselective access in an exclusively

native language context extends to three languages. To be sure, statistically signifi-
cant results for PGE cognates in comparison to their matched controls are likely to
be obtained due to the presence of the stronger foreign language—English (L2), not
the cumulative effect of two foreign languages. Thus, it can be argued that the non-
selective access hypothesis tested in the native language context has been found to
be valid only as far as the stronger foreign language is concerned. Notably, another
possible reason for the lack of L3 influence might be the fact that the trilingulas’ L3
proficiency, in comparison to their target language proficiency, was too low to affect
target language processing in any noticeable way. In Study 3 this factor will be fur-
ther manipulated.
As regards Hypothesis 3, to analyze task dependency, the RT results, achieved
for PGE cognates and for PG cognates under two different task conditions, have
been correlated (cf. Table 5.12). By using the same stimulus materials as well as
the same group of Polish-German-English trilinguals, in both a German language-
specific and a Polish language-specific lexical decision task, the effects of cross-
linguistic overlap were compared for exclusively German and exclusively Polish
lexical decision variants. In line with the assumption formulated in Hypothesis 3,
the paired difference for both PGE and PG cognates in comparison to their matched
controls in a German lexical decision task turned out to be bigger than in the case
of the Polish version of this task. Planned comparisons showed that both the dif-
ference in processing PGE cognates as well as PG cognates in comparison to their
matched controls in two different language settings reached the significance level as
confirmed by one-way ANOVAs: F(1, 41) = 8.695, p < 0.005 (p = 0.005) and F(1,
41) = 5.972, p < 0.05, respectively.
Following many other researchers (cf. Lemhöfer et al. 2008; Dijkstra et al.
2010), the obtained data can be interpreted with the support of the BIA+ model,
according to which the activation of different lexical representations is permanently
monitored by a decision system that subserves task performance and the process
of decision making. As has been shown in Chap. 3, Dijkstra’s BIA+ model pre-
dicts that even for the same stimulus materials, different tasks will lead to sys-
tematically different response patterns, because responding can occur at different
moments in time and can be based on different information sources. Overall, the
assumption formulated in Hypothesis 3 that the result patterns for cognates will dif-
fer depending on task demands has been confirmed. However, as has already been
stated, the achieved results might have been additionally affected by the difference
Table 5.12 Paired differences in RTs achieved for PGE and PG cognates in comparison to their
matched controls in German and Polish lexical decision tasks (Experiments 1 and 2)
German controls Polish controls
Mean RT SD Mean RT SD
PGE cognates versus controls 76.42 67.26 28.58 24.03
PG cognates versus controls 53.67 63.67 10.83 47.36
WWW.EBOOK777.COM
in the subjects’ L2 and L3 proficiency. In view of the above, some further research
using participants declaring comparable levels of L2 and L3 proficiency seemed
advisable.
5.6 Study 3: The Role of Relative Proficiency in the Third

Language in Trilingual Processing of Cognates
In Experiments 1 and 2, trilinguals with different levels of proficiency in the non-

target languages L2 (English) and L3 (German) were tested. In Experiment 1, it
was observed that the weakest language processing was influenced by both L1 and
L2 language knowledge. Experiment 2, in turn, proved that the strongest language
processing was influenced by L2, but that L3 did not play a significant role in L1
recognition. The analysis of the results of Experiments 1 and 2 for PGE and PG
cognates proved that task dependency does play a role in word recognition—the
same group of PGE and PG cognates reported different RTs in comparison to their
matched controls depending on the language of the task. It needs to be stressed,
however, that PGE cognates in Study 2 were significantly faster to recognize than
Polish controls, which implies that despite the context being changed, there was still
some cross-linguistic mediation, but only for the stronger nontarget language.
In the present experiment, which aimed at evoking the cross-linguistic influence
for the weaker foreign language, the L3 proficiency was manipulated. It was assumed
that the lack of L3 influence on L1 processing, documented in Experiment 2, could
have been caused by the fact that the participants’ proficiency in L3 was too weak
to exert a significant effect on L1 processing. Consequently, following Dijkstra and
Van Heuven (2002a), it was expected that activation of weaker language knowledge is
only noticeable when the speaker is relatively fluent in the nonnative language. Thus,
in Experiment 3 the level of proficiency in L3 was increased by testing trilinguals that
were equally fluent in L2 and L3. The question was whether processing in L1 can be
influenced by L2 and L3 nontarget language knowledge providing relatively high pro-
ficiency in both of the nontarget languages.
5.6.1 Research Hypothesis
If, following the previous research in this area, the manipulation of the trilinguals’
proficiency in their third language may qualify the emergence of effects of a weaker
foreign language knowledge on native language processing, performance should
lead to result patterns different for cognates than for noncognates. The hypothesis
for the present experiment is as follows:
H: If the nonselective access hypothesis in the exclusively native language con-
text depends on foreign language proficiency, the participants with higher L3 level
will recognize PG cognates faster than those whose L3 proficiency is lower.
WWW.EBOOK777.COM
In Experiments 1 and 2, the participants were recruited from a group of trilingual

students of English Philology who in a language questionnaire declared knowl-
edge of German at, at least, A2 level. In contrast, the aim of the present experi-
ment was to test trilinguals who were comparable to those of Experiments 1 and
2, but with a higher level of proficiency in German. Hence students from the
first year (second term) of the German Language Department at the University
of Łódź were recruited. A group of 17 people completed a language background
questionnaire before the experiment. For the present task only those were selected
(13 students) who rated their L3 and L2 competence at B1 level at the least. On a
scale form 1 (very little experience) to 5 (very much experience), the mean self-
rating of their language proficiency was in 3.87 German (SD 2.1) and in 3.76
English (SD 1.9). The participants were between 20 and 26 years old (mean 23.4).
A summary of the participants’ learning history is presented in Table 5.13.
It needs to be stressed that this was a small scale study whose major aim was to
replicate Experiment 2 for a different group of participants. The small scale of the
study in terms of the number of testees was determined by low availability of sub-
jects with high German and English proficiency.
5.6.3 Materials and Procedure
The materials were identical to those used in Experiment 2. The procedure fol-
lowed in Experiment 2 was adopted in Experiment 3 to calculate mean RTs for
the cognates with English and German, for the cognates with German, and for the
noncognates.
The data of 2 participants with exceptionally high error rates (above 15 %) were
excluded from the analysis. For the remaining 11 participants, the overall error
rate amounted to 7.6 % (5.8 % on test words). In addition, RTs that lay more than
two standard deviations away from the mean (2.8 % of all data) were classified as

English German
Mean SD Mean SD
WWW.EBOOK777.COM
5.6 Study 3: The Role of Relative Proficiency in the Third Language … 157
RTs and error rates for the
PGE cognates 730.92 23.58 4.6 5.6
PG cognates 745.00 43.91 5.1 6.2
Polish controls 758.75 53.58 7.9 7.9
Polish fillers 762.12 41.27 8.7 9.1
Nonwords 803.21 46.15 11.5 10.3
outliers and omitted from the analysis. In total 8.6 % of the data were excluded.
Mean lexical decision times and errors are presented in Table 5.14.
In Experiment 3, the same stimulus materials as in Experiment 2 were tested.
Now, a cognate advantage arose in words that were cognates with their L3 trans-
lations but still, by conventional criteria, this difference is considered to be not
quite statistically significant (t(11) = 2.03, p = 0.06). Again, the results for PGE
cognates reached significance (t(11) = 2.34, p < 0.5), probably due to the influ-
ence from L2 English. This assumed role of English appears to be interesting
especially if considered in relation to the subjects’ linguistic profile. Table 5.15
shows the results of the paired samples t-test for the three types of stimuli words.
With respect to the hypothesis set in the present experiment, to test the gener-
ality of the language-nonselective access and the dependence of lexical retrieval
on language proficiency, cross-task comparisons were conducted. The calculated
RT difference for PG cognates as compared with their matched controls from
Experiment 2 was juxtaposed with the parallel data obtained in Experiment 3 (cf.
Table 5.16). The level of L3 proficiency appeared to exert no effect on the visual
recognition of cognates. PG cognates were not faster to recognize when presented
to more fluent L3 users. Clearly, the fact that the comparison did not yield statisti-
cally significant results (F(1, 30) = 0.037, p = 0.84.) falsifies the hypothesis. On
the other hand, it can still be argued that the participants’ L3 level was still too
low to influence native language performance in an exclusively native language
context.
To sum up, the findings resulting from the nontarget, and weaker, language pro-
ficiency manipulation support the theoretical position that multilinguals’ process-
ing system is profoundly nonselective with respect to language only as far as the
stronger foreign language is concerned. Hence it seems legitimate to say that a cer-
tain level of weaker foreign language proficiency is required before any L3 effects
Table 5.15 Paired samples t-test for the three types of stimuli words (PGE cognates, PG cog-
nates and Polish controls) in a Polish lexical decision task in Experiment 3
(2-tailed)
PGE versus control 27.83 23.17 11.85 2.34 11 0.038
PGE versus PG 14.08 31.25 8.49 1.65 11 0.125
WWW.EBOOK777.COM
Table 5.16 Paired differences in RTs achieved for PG cognates in comparison to their matched

controls in two Polish lexical decision tasks conducted by the participants with varying L3 profi-
ciency level (Experiments 2 and 3)
Lower L3 proficiency— Higher L3 proficiency—
Experiment 2 Experiment 2
Mean RT SD Mean RT SD
PG cognates versus controls 10.83 47.36 13.75 29.45
become noticeable in L1 processing. To prove that assumption, however, some addi-

tional tests with more advanced L3 learners seem to be indispensable.
5.7 Study 4: The Role of Psychotypology in Trilingual

Processing of Cognates
As has been mentioned in the theoretical chapters, several studies on L3 acquisi-

tion have convincingly shown a qualitative difference between the acquisition of
an L2 and the subsequent acquisition of an L3. Some studies even indicate that
L2 takes on a stronger role than L1 in the initial state of L3 learning (cf. Williams
and Hammarberg 1998; Bardel and Falk 2007), which is typically related to the
L2 status factor. According to Bardel and Falk (2007), the L2 status factor might
be an outcome of the higher degree of similarity between L2 and L3 than between
L1 and L3, regarding age of onset, outcome, learning situation, metalinguistic
knowledge, learning strategies and degree of awareness in the language learning
process. Additionally, it has been widely acknowledged that while all the previ-
ously acquired languages potentially affect the development of the L3 system, it is
linguistic typology and psychotypology that play a crucial role in determining the
“privileged” source of transfer. Accounting for the above assumptions, the final
experiment described in the ongoing chapter aims to verify Singleton’s (2003)
prediction concerning the cumulative influence of the L2 status and psychotypol-
ogy on the organization of the multilingual lexicon. The rationale was that the
effect of psychotypology supported by the L2 status factor induces stronger con-
nections between L2 and L3 than between L1 and L3 lexical items in the mental
lexicon of the speaker whose foreign languages are typologically related to one
another and not closely related with the native one.
5.7.1 Research Hypothesis
The manifestation of strong connections between L2 and L3 lexical representations

and not between L1 and L3 ones in the trilingual mental lexicon will be mirrored in
WWW.EBOOK777.COM
5.7 Study 4: The Role of Psychotypology in Trilingual Processing … 159
reaction times to cognate words presented in the lexical decision task. The hypoth-
esis for the present experiment is as follows:
H: If stronger interaction between typologically related foreign languages (L2
and L3) known to the speaker than between his/her L1 and L3 is the manifestation
of strong connections between L2 and L3 representations and not between L1 and
L3 ones in the multilingual mental lexicon, the cumulative effect of psychotypology
and the L2 status will lead to faster RTs found for English-German cognates than
for Polish-German cognates.
A group of 17 trilinguals with Polish as their mother tongue (L1), and English and
German as their second (L2) and third language (L3), respectively, participated in
the experiment. All of the participants were students (III year) of English Philology
at the Academy of International Relations in Łodź. All of them had learned English
and German as a foreign language at school. English from about the age of 13,
German from about the age of 16. The participants’ L2 competence was not tested
before the study. It was assumed that their status of students of English Philology
ensured advanced L2 proficiency. Additionally, only those students were selected
for whom the pass mark at the English practical exam was at least 4 (good). Before
the experiment a language background questionnaire was administered, in which 19
participants were asked to report on their competencies in all of their foreign lan-
guages and the frequency of use of these languages. A few respondents declared
some competence in foreign languages other than English and German, but those
selected for the experiment (17) described their competence as very low. The
reported languages included Spanish and Russian. Again, all the participants were
“unbalanced” trilinguals, i.e., they were not as proficient in their second and third
languages—English and German—as in their mother tongue Polish (cf. Table 5.17).
The data from 3 participants had to be excluded from analysis which sets the
mortality rate at the level of 17.6 %. The remaining 14 participants were between
21 and 36 years old with the mean of 28.5. 10 were female, 4 were male. The
participants were not paid for their participation, nor given any course credits.
To ensure that the testees constituted a homogenous sample as far as their
L3 German competence is concerned only those students were selected who
described their L3 level as pre-intermediate or intermediate (A2 and B1, respec-
tively). A summary of the participants’ learning history is given in Table 5.17.

English German
Mean SD Mean SD
WWW.EBOOK777.COM
The small scale of the study in terms of the number of the participants was
determined by low availability of subjects with German as their L3 declaring at
least A2 level proficiency.
5.7.3 Materials and Procedure
Three critical groups of words were compared with respect to the latencies and
accuracy of their recognition: German control words that were different from both
their English and Polish translations (e.g., GELD; Eng.: money, Pl.: pienia˛dze),
Polish-German cognates that were not cognates to English (e.g., DACH, meaning
roof in both Polish and German), and English-German cognates that overlapped in
orthography and meaning in L2 and L3 but were different in L1 (e.g., FINGER;
Pl.: palec). The group of Polish-German cognates from Experiment 1 was used.
Additionally 12 English-German (EG) cognates were selected from the CELEX
database. They had identical or similar spelling and meaning in both languages,
like the word FINGER. They matched the Polish-German cognates with respect to
the length, number of syllables and frequency. All of them were singular forms of
nouns with a length of between 3 and 6 letters and no more than two syllables. Both
German fillers as well as nonwords from Experiment 1 were used in the present
experiment. All stimulus items are listed in Appendix E and word characteristics are
presented in Table 5.18.
As for the procedure, the language-specific lexical decision task for the weak-
est language was used. The procedure employed in the previous experiments was
adopted to calculate mean RTs for the cognates with English, for the cognates with
Polish, and for the noncognates.
Table 5.18 Characteristics of the stimulus items used in Experiment 4

number of number of German English Polish
letters syllables frequency frequency frequency
EG cognates 4.66 1.33 40.72 41.06 –
PG cognates 4.88 1.52 42.09 – 40.03
German 4.58 1.50 41.79 – –
controls
German 4.58 1.33 43.48 – –
fillers
WWW.EBOOK777.COM
5.7 Study 4: The Role of Psychotypology in Trilingual Processing … 161
For the RT analysis, only correct responses were considered. Reaction times
exceeding two times the standard deviation from the subject means counted as outli-
ers and were excluded from the set of valid responses. Outliers accounted for 1.7 %
of all the responses and they were approximately equally distributed across con-
ditions. Erroneous responses (7.8 % for test words) were also excluded from the
analysis. The remaining data were entered into Statistica 9 program for statistical
analysis. Mean RTs were calculated for the cognates with English, for the cognates
with Polish, and for the noncognates; all of them are presented in Table 5.19.
In order to analyse the recorded RTs, a t-test for paired-samples whose results are
listed in Table 5.20 was conducted. The effect of word type turned out to be not sig-
nificant (F(2, 33) = 1.23; p = 0.34). Mean lexical decision times to the L3 words that
were cognates with English were only slightly shorter than those to the noncognates
and failed to reach significance. Interestingly enough, however, the difference between
the L3 words that were cognates with Polish and the noncognates turned out to be
significant: t(11) = 3.06; p < 0.05. So did the difference between L1–L3 cognates and
their L2–L3 matched counterparts: t(11) = 2.75; p < 0.05.
Despite the fact that cross-linguistic influence between typologically related for-
eign languages has been observed in numerous off-line TLA studies (cf. Hall and
Ecke 2003; Ringbom 2007), the present study, which investigated on-line trilingual
processing, did not prove stronger connections between L2 and L3 lexical items
than between L1 and L3 ones. The data obtained in the present experiment provide
no statistically significant evidence for the effect of psychotypology which would be
powerful enough to foster strong connections between typologically related foreign
languages. Hence the hypothesis set in the present experiment has to be rejected.
Contrary to the initial assumptions, faster RTs documented for PG cognates than for
Table 5.19 The mean RTs and ERs for the participants in Experiment 4

RT SD ER SD
EG cognates 826.50 30.74 7.8 4.9
PG cognates 811.42 31.46 6.4 3.1
German controls 830.25 30.98 9.3 5.4
German fillers 843.78 29.67 9.8 5.2
Nonwords 889.17 41.18 11.4 7.9
Table 5.20 Paired samples t-test for the three types of stimuli words (EG cognates, PG cognates
and German controls) in a German lexical decision task in Experiment 4
(2-tailed)
EG versus control 3.75 21.04 2.66 1.40 11 0.186
EG versus PG 15.08 21.89 5.46 2.75 11 0.018
WWW.EBOOK777.COM
835
830
830
826
825
820 PG cognates
ms
EG cognates
815 German controls
811
810
805
800
Fig. 5.4 Mean RTs obtained for PG cognates, EG cognates and German control words
their EG matched counterparts (cf. Fig. 5.4) prove that there are strong connections
between L3 and L1, even though L1 (Polish) is typologically more distant from L3
(German) than L2 (English), and it lacks the special status of the second language.
To be sure, lexical links between L1 and L3 might have developed as a result
of learning habits. As appears from the post hoc questionnaire, the participants of
Experiment 4 learnt their L3 in an L1-based environment. Their teacher of German
was a Polish native speaker, who did not resort to English during their German
instruction. They also declared wide access to Polish-German and German-Polish
dictionaries in the course of their L3 learning process. Taking all these data into
consideration, it seems reasonable to claim that the structure of their mental lexi-
cons may have been affected by the nature of their learning experience. It seems
that this factor may be powerful enough to foster strong connections between lan-
guages extensively and concurrently used in education.
5.8 General Discussion
Recapitulating, the four experiments reported in the present chapter were performed
with two main goals in mind. Firstly, the aim was to take a deeper look at the nature
of the representation of cognates in the trilingual mental lexicon by including them in
a language specific lexical decision task in which all three languages were involved.
Secondly, the subsequent experiments were carried out in order to investigate any pos-
sible influence of a set of factors (such as task dependency, language proficiency and
psychotypology), identified in the course of research within TLA, on the representa-
tion of cognates in the multilingual mental lexicon. More precisely, in Experiment 2,
the BIA+ model’s predictions about task demands and multilingual lexicon were veri-
fied by changing the task (the weakest vs. the strongest language context) but keeping
WWW.EBOOK777.COM
5.8 General Discussion 163
the stimulus materials constant. In Experiment 3, relative language proficiency and its
impact on lexical recognition was tested. The final experiment aimed at checking the
cumulative effect of the L2 status factor and psychotypology on the multilingual stor-
age and retrieval of cognates. The experiments have led to several conclusions about
cognate processing and representation in trilinguals, to be discussed in the paragraphs
to come.
Consistent with previous findings in the literature, and as expected, the reported
experiments proved that trilinguals processed cognates more quickly and accurately
than they processed noncognate control words. Obviously, the overlap of form and
meaning across languages facilitated lexical access. Before some more detailed
analysis of the results of the experiments is offered, it seems justifiable to reiterate
the most plausible accounts of the cognate facilitation effect on the basis of the con-
ducted experiments. In the literature on the topic, mainly concentrated on bilinguals,
different accounts of the observed bilingual cognate effects are proposed, depend-
ing on the way cognates are thought to be represented within the bilingual language
system. First, orthographically identical cognates may share a single orthographic
representation across the two languages, as some researchers suggest (cf. De Groot
and Nas 1991; Sánchez Casas et al. 1992). In this view, the cognate effect is sim-
ply a consequence of the cumulative frequency of a cognate across all relevant lan-
guages. Because a multilingual encounters a cognate more frequently than a word
that exists in only one language, it is the standard word frequency effect that causes
the RT advantage of cognates. The results obtained for the group of PGE cognates
in Experiment 1 appear to be in line with this explanation. Interestingly enough, the
single orthographic representation account is related to a learning-based explanation
of the cognate effect, according to which language learners can make use of pre-
existing L1 memory representations during the acquisition of cognates (cf. De Groot
and Keijzer 2000; De Groot and Van Hell 2005). However, there is a growing body
of evidence that the obtained result patterns for cognates are not just a consequence
of cumulative frequency effects (cf. De Groot et al. 2000; Dijkstra et al. 1999;
Dijkstra 2007). In fact, many linguists point out that this account is not easily rec-
oncilable with the finding of facilitatory cognate effects for cognates that are ortho-
graphically similar, but not identical, such as the Polish-German pair cel versus Ziel,
which was used in the described experiments. Given that many cognates in the pre-
sented experiments had a different orthographic representation in the two languages
(almost 46 % in PG group), it seems unlikely that the facilitatory cognate effects can
be explained by assuming that a cognate’s effective frequency is simply that of the
sum of the frequencies of two words across languages. Indeed, the participants in all
the tasks were faster at recognizing identical cognates to similar ones. Nevertheless,
the latter group still triggered the facilitatory cognate effect as confirmed by planned
comparisons conducted on the group of non-identical cognates and their matched
controls. Therefore, a second possibility that may account for both orthographically
identical and nonidentical cognate effects involves semantic feedback. Even if cog-
nates do not share the same orthographic representations for all languages to which
they belong, it can be assumed that they activate the same conceptual representa-
tion, because the overlap of their meanings in the different languages is usually large
WWW.EBOOK777.COM
(cf. De Groot and Nas 1991; Van Hell and De Groot 1998a; De Groot and Van Hell
2005). Cognates would then be recognized faster because their semantic representa-
tions (receiving input from two rather than one orthographic representation) are acti-
vated to a larger degree (cf. BIA+ model). What is significant is that this account of
the cognate effect is in line with recent studies on monolingual word recognition that
also argue in favor of semantic-to-orthographic feedback (cf. Pecher 2001; Reimer
et al. 2001). Also the data gathered in the described experiments can be interpreted
within this localist connectionist framework, in the light of which cognate translation
equivalents are “only ‘special’ relative to noncognate translations in that they share
more orthographic, semantic, and/or phonological features across languages” (Voga
and Grainger 2007, p. 943). According to this account, the cognate facilitation effect
in reading might in fact be an orthographic-semantic priming effect: Overlapping
orthographic and semantic representations of both languages become active upon
the presentation of one of the readings of the cognate, leading to a facilitated rec-
ognition of cognates relative to noncognates. In this model, the cognate facilitation
effect depends on both cross-linguistic similarity (the factor that in all the experi-
ments could additionally trigger faster recognition of PGE words, all of which were
identical cognates) and word frequency. It has been argued that both a larger similar-
ity of the two readings of a cognate and a higher frequency result in a more strongly
activated shared semantic representation. Since the factor of cross-linguistic similar-
ity might have influenced the reported results, its role will be further analyzed in the
section devoted to final conclusions where some future experiments aimed at verify-
ing the role of this very factor are to be discussed.
More importantly, it appears that all of the enumerated approaches are extend-
able to three languages and reconcilable with the current findings. Just as the bilin-
gual cognate effect has been interpreted in terms of coactivation of two languages
(cf. De Groot et al. 2002; Dijkstra et al. 1999; Dijkstra 2007), it seems reasonable
to believe that data gathered in Experiment 1 provide evidence for the simultane-
ous involvement of all three languages during the word recognition process in L3.
As evidenced in Experiment 1, the obtained results reveal that the cognate effect
can accumulate over languages: while cognate status in one language caused shorter
word recognition latencies, the additional cognate status in one additional language
speeded up responses even more. Thus, it can be claimed that the notion of nonse-
lective lexical access that has recently received growing support within the bilingual
domain, seems to generalise to trilinguals and three languages. Within experiments
2–4 each experiment focused on a different set of factors which were meant to influ-
ence lexical connections between the languages known to the participants. It was
predicted prior to experiments 2–4 that, if the influence of factors in question is
potent enough to foster stronger connections with only one of the languages, the
facilitation effect will be observed only for this pair of languages. The absence of
the facilitation effect is not to mean that there are no links with the other language
(EG relation in Experiment 4), but that the connections are too weak to induce this
effect.
As mentioned before, Experiment 1 proved that processing the weakest language
(L3) word entails automatic, parallel activation of candidate words in the dominant,
WWW.EBOOK777.COM
stronger languages; not only L1 but also L2. It was concluded that in accordance
with the nonselective access hypothesis the presentation of a word in one language
automatically activates words from both the target and the nontarget languages in
parallel. From this it logically follows that cross-linguistic effects may arise in both
directions, manifesting themselves not only in a nondominant but also in a domi-
nant target language. This assumption gave rise to the hypotheses set in the sec-
ond experiment, whose aim was twofold. First, the influence of foreign language
knowledge on native language performance in an exclusively native language con-
text was studied. The objective was to verify the assumption in the light of which
weaker language knowledge may influence performance in the dominant language.
Needless to say, such a finding would additionally support nonselective access in
multilingual lexicon. Second, using the same group of PG cognates in Experiments
1 and 2 allowed the researcher to verify the influence of task demands on the rec-
ognition of words and the hypothesis that cognate effects might be task dependent.
As regards the nonselective position, it has been both strengthened and extended.
Experiment 2 showed that processing words in the strongest language can be influ-
enced by weaker language knowledge. In other words, the finding that foreign
language knowledge (in this case L2) affects L1 target word processing in an exclu-
sively native language context provided additional support for the theoretical posi-
tion that the language processing system of multilinguals is profoundly nonselective
with respect to language. On the other hand, the nonnative language influence on
the mother tongue was clear only for L2 since only PGE cognates were recognized
faster than their matched controls; L1–L3 cognates did not generate statistically sig-
nificant results. Admittedly, it was L2 that supported the recognition. The compari-
son of the results for PG cognates from Experiment 2 with those from Experiment 1
showed that the same list of stimulus words, recognized by the same group of sub-
jects, generated different RTs. Clearly, language nonselective access in the case of
three languages seems to be task dependent. And although more research is needed
to find out the exact nature of the demands posed by various tasks, Experiment 2
showed that the different results they produce may be informative with respect to
the underlying representation of cognates. Summing up, Experiment 2 showed that
processing words in the strongest language can be influenced by weaker language
knowledge—in this case, L2 knowledge. However, since L3 did not affect L1, it
was hypothesized that a certain level of weaker language proficiency is required
before any weaker language effects become noticeable in L1 processing.
In order to strengthen the interpretation of the post hoc analysis in Experiment 2
and the conclusion that L1 performance can be influenced by weaker language
knowledge if the multilingual has reached a certain level of weaker language pro-
ficiency, another experiment was performed (cf. Experiment 3). This time, how-
ever, trilinguals with a higher level of proficiency in L3 were used. Now again, as in
Experiment 2, lexical decision times to the triple cognates were shorter than those
to the noncognates. Similarly, responses on the L1 words that were cognates with
German were faster, though not statistically significant, than on the noncognates.
Accordingly, it can be concluded that, the manipulation of the trilinguals’ profi-
ciency in L2 and L3 revealed that their proficiency in L3 was too weak to exert any
WWW.EBOOK777.COM
effect on L1 processing. Nevertheless, although no statistically significant results

were reported, still it seems likely that non-native language proficiency may play
a role in cross-language lexical activation. In other words, an implication of the
preceding observation may be that weaker language knowledge may influence per-
formance in the dominant language as multilinguals gain more proficiency in the
weaker language.
The final experiment aimed at verifying the influence of psychotypology
and the L2 status on the representation of cognates in the multilingual lexicon.
Contrary to expectations, no results confirming the role of psychotypology have
been obtained. Indeed, the results of Experiment 4 can be taken to mean that the
cumulative influence of these factors was too weak to foster stronger connections
between the participants’ typologically related languages. It has to be stressed that
these findings do not follow the general pattern of the results obtained in TLA
studies where psychotypology and the L2 status have been found to have a robust
effect on transfer (cf. Odlin 1989; Williams and Hammarberg 1998; Singleton
2002; Hall and Ecke 2003; Ringbom 2001, 2007; Forte 2009). Clearly, the hypoth-
esis put forward by Singleton (2002), that if an L3 is more closely related to L2
than to L1, its representations at the lemma level will be linked to L2 rather than
L1, does not stand up to scrutiny. The discrepancy between the present results and
those obtained from research on transfer may be ascribed to a number of factors,
such as a type of methodology used in the experiment, strategies employed by tri-
lingual speakers, learning experience of the speakers or the low level of L3 pro-
ficiency. Firstly, the choice of methodology might have decided about the results
obtained in the experiment in question. In contrast to many experiments into cross-
linguistic influence that have been conducted in paper-and-pencil methodology
and thus relied heavily on off-line processing, Experiment 4 employed on-line
processing, which does not allow conscious control and thus seems to be more
potent in terms of shedding light on the composition of languages in the multi-
lingual mind. Secondly, the fundamental difference in methodologies is directly
related to another aspect which needs to be taken into account, namely strategiz-
ing. As has been already mentioned, studies investigating cross-linguistic influence
which rely on off-line processing allow for strategies being consciously employed
by the speaker to enhance his/her L3 performance. Given more time to think what
language item to produce, speakers may consciously resort to other languages they
know and use them as props in L3 production. Therefore, it is possible that psy-
chotypology and the L2 status play an important role as far as strategies are con-
cerned, as the case of an off-line study reported by Singleton (2002) shows very
well. However, as soon as strategizing is limited by task demands, the influence of
these factors cannot be evidenced. Thirdly, stronger lexical links between L1 and
L3 might have developed as a result of learning habits or the so-called learning
experience factor. Since L3 was the weakest language which was still being learnt,
the role of learning experience was still stronger than in the case of the more
advanced L2. Definitely, the influence of learning experience needs to be left for
future analysis. It seems that this factor may be powerful enough to foster strong
connections between languages extensively and concurrently used in education.
WWW.EBOOK777.COM
Finally, it may also be the case that the languages involved in the experiment were
still too similar and thus the psychotypology effect, typical for languages coming
from different language families, could not have been activated.
All in all, the findings triggered many additional questions that could constitute a
starting point for another set of experiments and that will be partly addressed in the
final part of this volume, together with the implications of the outcomes discussed
in this section for the structure of the multilingual lexicon and the conceptualiza-
tion of a multilingual speaker. A few tentative recommendations stemming from the
present observations concerning the ways in which the conclusions based on the
outcomes of the described experiments could be adopted in the process of foreign
language learning and teaching will be offered in the subsequent section, which is
also a conclusion to the present volume.
WWW.EBOOK777.COM
Conclusions and Implications
This final section provides a critical synthesis of the major findings presented
throughout this volume. It aims to show how research on multilingual processing
can add to the current body of knowledge on third or additional language acquisi-
tion. First, it once again outlines the goals and the structure of the work, offering a
brief summary of the general findings presented in the context of the dispute over
the conceptualization of a multilingual learner. Next, it refers to limitations of the
conducted experiments and enumerates some suggestions for future research. Last
but not least, the final paragraphs address the role of the obtained results for multi-
lingual didactics.
The primary goal of this volume was to offer an insight into the structure of the
multilingual mental lexicon which is a product of the interplay of a whole array of
cross-linguistic factors in the minds of multilingual speakers. More specifically,
the book focused on the empirical investigations of the storage and retrieval of
double and triple cognates in a multilingual mind. All the aspects, analyzed in the
experiments reported in the practical chapter, are part of a broader question of how
multilinguals make their lexical decisions and how they recognize words from dif-
ferent languages. In order to place the investigations in a larger context, the text
was divided into five chapters. The preliminary theoretical considerations referring
to a plethora of multilingual issues included in Chap. 1 provided a relevant back-
ground against which most important problems concerning the mental lexicon of
mono- and multilingual learners were presented in Chaps. 2 and 3, respectively.
In Chap. 4 the focus of attention was shifted to the special position of cognates
in visual word recognition in the multilingual mind. Chap. 5 offered a detailed
account of a series of experiments conducted in order to verify the assumptions
concerning the storage and processing of cognates in the context of the Polish-
English-German lexicon.
The underlying aim of the present work was to explore whether multilinguals
should be considered as learners and speakers in their own right as opposed to L2
learners and if so, what consequences this would bring to third or additional lan-
guage teaching. As mentioned in Chap. 1, on the one hand, there are scholars (cf.
Baetens Beardsmore 1986; Li Wei 2007; Mitchell and Myles 2004) who assume

DOI 10.1007/978-3-642-32194-8
WWW.EBOOK777.COM
170 Conclusions and Implications
that there is no meaningful difference between bilinguals’ and multilinguals’ pro-

cesses and accordingly classify all learners of one or more non-native languages
as L2 learners, especially when proficiency in the previously learned non-native
languages is low. On the other hand, there are researchers (cf. Cenoz and Jessner
2009; Aronin and Singleton 2012) who argue that this position is not accept-
able, as meaningful differences between these learners’ processes exist and must
be accounted for. The end balance of this discussion seems to be in favour of the
view that a difference between the two types of speakers exists, as prior linguis-
tic knowledge has been repeatedly shown to affect multilinguals’ processes in a
number of ways. Clearly, multilinguals have more knowledge that can be used and
drawn upon during the acquisition process which bilinguals do not have at their
disposal, and this additional knowledge seems to play more of a central role than
was previously assumed. A number of studies have been carried out to explore
the effects of bilingualism on TLA, to mention the early research into TLA by
Ringbom (1987) and Thomas (1988) and later studies by Jessner (2006), Ringbom
(2007), Hufeisen and Marx (2007) or Neuner et al. (2009). Typically, life with two
or more languages has been confirmed to be advantageous, not only with respect
to language knowledge but also cognitive and sociopragmatic development.
Research indicates that bilingual children outperform monolinguals in the acqui-
sition of third or additional languages (cf. Hufeisen and Jessner 2009; Neuner
et al. 2009). Accordingly, the role of the L2 in a trilingual system merits a careful
scrutiny and thus, in what follows, the results reached in the experiments will be
analyzed in reference to this discussion.
One common argument in favour of the view that a difference between the two
types of users exists is the contention that multilingual learners are influenced
both by their L1 and the non-native languages they know. In a multilingual sys-
tem, cross-linguistic influence (CLI) takes place not only between the L1 and the
L2 but also between the L2 and the L3, and the L1 and the L3, as well as in the
reverse direction. The findings of Experiments 1–4 confirmed this type of multidi-
rectional transfer. Consequently, it seems legitimate to say that the native language
does not always have a privileged status and must be looked at together with other
possible sources of transfer. Since multilinguals have knowledge of more than
two languages by definition, the possible sources of lexical transfer automati-
cally increase with the number of languages the individual is familiar with. More
extremely, it has been proved that multilingual processing can be influenced by
more than one language at the same time—a phenomenon referred to as combined
cross-linguistic influence (cf. Hoffmann 2001b; Singleton 2002; Ringbom 2007;
De Angelis 2007). Although it still remains little explored to date, the cumula-
tive cognate facilitation effect shows that two or more languages can interact with
one another and concur in influencing the target language (cf. Experiments 1–4
in Chap. 5). In other words, when more sources are available, they have to be
accounted for, whether used or unused by the learner.
Obviously, this multisource influence depends on a variety of factors. One vari-
able repeatedly referred to in the discussion on the multilingual user concerns lan-
guage proficiency (cf. Experiments 2 and 3). In a host of previous studies with
WWW.EBOOK777.COM
Conclusions and Implications 171
respect to the relationship between proficiency in the non-native languages and

CLI (cf. De Groot and Poot 1997; Christoffels et al. 2006), the influence was
equally found to occur from non-native languages speakers know well and from
non-native languages speakers do not know well. This opinion, however, has not
been confirmed by the experiments reported in Chap. 5 of the present book (cf.
Experiments 2 and 3). The influence of the weaker foreign language (L3) did
not yield the same results as the impact exerted by L2 which is in line with the
tenets of Cummins’ Threshold Hypothesis and Developmental Interdependence
Hypothesis (1976, 1979), both promoting the view that the higher the competence
attained in the previous languages, the stronger the likelihood that some influence
will occur. Consequently, there are two main issues which remain to be tackled
in future research on multilingual processing. The first one relates to the type of
relationship that non-native languages establish with one another at a single point
in time and over time, especially in view of the rapid changes in proficiency level
non-native languages are subject to. The other is the question of how proficient
learners need to be before their prior knowledge begins to affect the production
and development of a target language to a significant extent and before they can
begin to be classified as L3 learners in empirical investigation.
Yet another issue affecting multilingual processing raised in this volume con-
cerns language distance, the factor which has been repeatedly discussed as one of
the crucial triggers of cross-linguistic influence in relation to L2 as well as bilin-
gual learners. With multilinguals, however, the role of language distance is not as
straightforward as it may be with L2 learners due to the possible combination of
close and distant languages in the mind. According to Singleton (2003), when mul-
tilinguals have knowledge of related and unrelated languages, evidence suggests
that they are overwhelmingly influenced by the language that is the closest to the
target language, regardless of whether this is the native or a non-native language.
The same author claims that the native language does not seem to have a privileged
status. Experiment 4 did not support this contention, possibly due to the fact that the
testees’ L3 proficiency was too low in comparison to their L2. Additionally, stronger
lexical links between typologically more distant L1 and L3 might have developed
as a result of learning habits or preferred vocabulary learning strategies. Cieślicka
(2000) proved that the strategy of L2 learning significantly influences the strength
of connections obtaining in bilingual lexical networks. Admittedly, some further
research verifying this assumption for multilingual lexicon seems to be required.
As regards the limitations of the study, it has to be admitted that it was difficult
to fully control many individual as well as extraneous variables. Consequently,
there occurred some shortcomings that need to be acknowledged and addressed.
One of the extraneous variables that might have affected the validity of the experi-
ments relates to population characteristics. According to Seliger and Shohamy
(1989, p.107), this variable is linked to the degree to which “the sample population
in the study has the same characteristics as the population to which the research
findings are to be applied”. The fact that the participants of the experiments were
students at a philological department definitely limits the generalisability of the
findings. Another participant-related variable influencing the results is connected
WWW.EBOOK777.COM
with the number of respondents. Due to considerable difficulties in finding sub-

jects with advanced levels of both English and German the research groups were
not too numerous (ca. 20 people). Luckily, the participant mortality rate was not
statistically significant and, as the dropout was random, it did not affect the group
homogeneity. Yet another limitation concerns the type of methodology itself.
Since the lexical decision task entails single word recognition, it seems reason-
able to believe that some research is still needed to verify whether the outcomes
of the discussed experiments are maintained or eliminated in a sentence context.
The more so, as there is a growing amount of empirical data showing that when
words are processed in sentence context, their processing seems to be sensitive
to the semantic and syntactic aspects of the sentence (cf. Hartsuiker et al. 2004;
Hartsuiker and Pickering 2008; in Riehl 2010). All in all, it seems legitimate to
predict that context effects could influence the findings to a large extent.
Another important issue that might have influenced the results and that merits
further investigation in the context of the multilingual processing is recency of use.
It is believed that learners are more likely to be influenced by those languages that
are used often, and in the recent past (cf. Grosjean 2001, 2008). Recency of use
has been found to variably affect the amount of influence from the source to the
target language. This hypothesis, however, has not been widely tested with mul-
tilinguals and its validity remains to be empirically assessed. Still, the results of
Experiments 1 and 2 proved that the recently used language influences the recog-
nition time of both double and triple cognates and thus confirmed the task depen-
dency of the multilingual lexical processing. Indisputably, some further research
is needed to check if the degree to which the languages of a multilingual interact
depends on whether the individual is in a mono- or multilingual language mode.
The above paragraphs pointed to some limitations of the study and enumerated
possible areas for future research. Undoubtedly, more research is needed to exam-
ine the relationship between the strength of association and factors such as typol-
ogy and proficiency level; particularly so as in the Polish context few research
studies on these factors are currently available, works by Danuta Gabryś-Barker
and Anna Cieślicka being a prominent exception.
Undoubtedly, it would be imprudent to claim that the outcomes of the presented
experiments constitute sufficient grounds for the formulation of far-fetched ped-
agogical recommendations, and there surely exists the need to explore the issue
much further; however, some tentative suggestions seem admissible. Nowadays,
in our fast developing, frontier-free world, an increasing proportion of language
learners have some command of other languages beyond their native one. In fact,
whether by education, immigration or living environment, in many parts of the
world the typical learner is no longer a learner of a second language but rather
a learner of a third or additional language, for instance an L4 or an L5. In fact,
the rapidly growing number of multilinguals is a reality which more and more
language teachers and researchers are being confronted with in their daily work.
Consequently, one important research area that awaits a more thorough explora-
tion is the role of previous language knowledge and previous language learning
strategies in the processes of subsequent language acquisition.
WWW.EBOOK777.COM
Over the decades, three major language acquisition research lines were
developed, each of which was expressed in various models and hypotheses (cf.
Hufeisen 2004; Riemer 2002): the contrastive hypothesis (which focused on com-
paring and contrasting languages, and to which we owe the delineation of interfer-
ence); the nativistic hypothesis (which postulated an inherent language acquisition
sequence within each individual and identified development-related errors); and
the interlanguage hypothesis (which described language acquisition as a dynamic
and systematic process focused on the new language to be learned, and in par-
ticular described transfer processes). It needs stressing, however, that all of these
hypotheses assumed as self-evident the involvement of only two languages and
“implied, at best in passing, that the acquisition or learning of other (foreign) lan-
guages did not differ from the learning of the first foreign language” (Hufeisen
2004, p. 7). Needless to say, the situation has changed radically within the last two
decades. The attention directed towards the investigation of multilingual learners
has resulted in the intense research activity in the field of multilingualism which
in turn has helped raise a general awareness both among scholars and practitio-
ners about the distinctiveness of multilinguals` language learning processes and
how these may differ from those of monolingual learners of a second language.
The integrating of the multilingual research findings into the classroom has led
to the development of tertiary language didactics which is now seeking to apply
this qualitative difference to advantage, including in L3 teaching learners’ previ-
ous cognitive and emotional experiences. What is also worth noting is the fact that
if the languages to be learned are very similar, this recourse can also be closely
related to linguistic aspects; if they are rather different, then it is more the case
that learning strategies and cognitive factors come into play. Interestingly enough,
it is admitted that particularly in the latter case, it is not absolutely necessary for
teachers to be perfect speakers of L2 (cf. Jessner 2006, 2008a, b; Aronin and
Hufeisen 2009) since it is not primarily the languages that they activate, but rather
the learning potential established during the learning of preceding languages. And
thus, what tertiary language teaching in school can achieve is, above all, the devel-
opment of language learning awareness that would help to learn subsequent lan-
guages efficiently.
As evidenced in the conducted experiments, in third or additional language
acquisition, an important aspect is the role of prior linguistic knowledge, since if
the learners perceive enough similarity between the two (or more) languages, they
rely strongly on their mother tongue and/or non-native languages. Therefore, the
facilitative role of cognates in TLA seems to be of no minor importance. By way
of illustration, it has been proved that fewer learning sessions are needed to help
trilinguals recall cognates than noncognates (cf. Carroll 1992). Similarly, response
latencies in translating cognates have been reported to be faster than for noncog-
nates (cf. Ellis and Beaton 1995, De Groot and Keijzer 2000). In fact, in SLA,
vocabulary acquisition based on detection and exploitation of similarities between
novel lexical input and prior lexical knowledge is a well-documented concept,
known as the parasitic strategy (cf. Hall 2002; Hall and Ecke 2003). Undeniably,
this approach can be equally applicable to the field of TLA. Hall et al. (2009) go
WWW.EBOOK777.COM
even as far as to define parasitism as a default mechanism in L3 vocabulary acqui-

sition. More precisely, the parasitic model they propose presupposes that new
words are integrated into existing lexical network with least possible redundancy
and as rapidly as possible in order to become accessible for communication. The
authors further claim that the multilingual lexicon admits cross-linguistic transfer
from all possible source languages and at all representational levels. The special
status of cognates has also been described in research by Müller-Lancé (2003)
whose large-scale study on the use of Romance languages in Germany (Müller-
Lancé 2003) led to distinguishing between productive (or retrieval) and receptive
(or inferencing) strategies; which strategy is utilized depends on formerly acquired
lexical competences in other foreign languages.
Clearly, as has been shown, languages cannot be studied in isolation since
acquiring or using one language is related to the other languages known.
Unfortunately, since there seems to be very little common ground between lan-
guage teachers and researchers, rarely are the theoretical advancements translated
into practical applications. As Pawlak (2007) observes, the notorious gap between
theory and practice that has plagued SLA since time immemorial, is largely due to
the discrepancy between research tools and classroom reality:
[…] even though many of the studies conducted in recent years claim to be directed at
improving foreign language pedagogy and their outcomes serve as a basis for mak-
ing more or less definitive recommendations for effective instruction, such guidelines
appear to find little reflection in classroom practice. Such a situation is due to the fact that
research findings are seldom disseminated among teachers and even if this happens, they
are often inaccessible to practitioners due to specialist terminology employed, the use of
advanced statistics and the preoccupation with reliability and validity rather than practical
applications (Pawlak 2007, p. 57).
The same is essentially true of TLA. Thus, without further investigations, the
provision of recommendations and the development of materials for class-
room use, it is difficult to imagine that third or additional language teach-
ing could find its way into the language classroom. Thus, is hoped that this
volume can help teachers, testers as well as curriculum and material designers
to have a better understanding of multilingual processing. Only by being famil-
iar with the mental processes involved in multilingual acquisition can teach-
ers understand the problems their learners have to face when learning a third
or additional foreign language, while course book writers can produce more
efficient teaching materials. All in all, languages being taught in the class-
room need to be linked in order to profit from the synergies and to exploit the
resources that many of the pupils already have available through their prior lan-
guage knowledge (cf. Cenoz and Gorter 2013). Therefore, one of the most
difficult aims of future work on language teacher education will be to make
sure that all language teachers are experts on multilingualism, even if they
teach only one language. This issue is addressed by Krumm (2005, p. 35),
who emphasizes that the most difficult correction of the educational system will
be not to train teachers in English or French but experts in multilingualism, who
teach a certain language but also accept at the same time the multilingualism of
WWW.EBOOK777.COM
their learners (Cenoz and Gorter 2013; Szubko-Sitarek 2014; Komorowska 2013).
In fact, an approach to softening the boundaries between languages has been put
forward by Cenoz and Gorter (2013). This approach promotes using the resources
multilingual learners have at their disposal in order to improve efficiency in lan-
guage teaching. It explores the possibility of establishing bridges between second
and foreign language teaching. As Hufeisen and Marx (2007) rightly claim, aware-
ness raising and interacting with samples of cross-linguistic similarities indeed
result in creating a mental model for intercomprehension.
The current multilingual projects (cf. EuroCom) focus on the question of how
the teaching and learning of tertiary languages can be structured in such a way
as to consciously incorporate the learner's existing language knowledge and lan-
guage learning experience (mother tongue, first foreign language). More precisely,
the purpose of many researchers and foreign language teachers should be to teach
subsequent foreign languages in a manner different from that used for the first for-
eign language in order to tap the potential already developed through utilizing the
knowledge of the mother tongue and the first foreign language.
WWW.EBOOK777.COM
Appendix A
A.1 Similarity-Rating Study: Stimuli and Ratings
Ratings were collected for English and German translation pairs that were test or
control items
Test Test O-rating P-rating Mean Con Con O-rating P-rating Mean
English German OP English German con con OP
rating rating
con
Bank Bank 5.00 5.00 5 Road Weg 1.00 1.00 1.00
Film Film 5.00 5.00 5 Tree Baum 1.00 1.00 1.00
Finger Finger 5.00 5.00 5 Proof Beweis 1.15 1.12 1.13
Gold Gold 5.00 5.00 5 Bell Glocke 1.32 1.25 1.28
Lift Lift 5.00 5.00 5 Knife Messer 1.00 1.00 1.00
Moment Moment 5.00 5.00 5 Woman Frau 1.15 1.05 1.1
Park Park 5.00 5.00 5 Sign Zeichen 1.45 1.15 1.3
Plan Plan 5.00 5.00 5 Chair Stühl 1.00 1.00 1.00
Radio Radio 5.00 5.00 5 River Fluss 1.00 1.00 1.00
Ring Ring 5.00 5.00 5 Duty Pflicht 1.00 1.00 1.00
Ski Ski 5.00 5.00 5 Fork Gabel 1.00 1.00 1.00
Sport Sport 5.00 5.00 5 Autumn Herbst 1.00 1.00 1.00
Talent Talent 5.00 5.00 5 Song Lied 1.00 1.00 1.00
Test Test 5.00 5.00 5 Aim Ziel 1.13 1.25 1.19
Trend Trend 5.00 5.00 5 Potato Kartoffel 2.30 2.50 2.4
Winter Winter 5.00 5.00 5 Horse Pferd 1.23 1.15 1.19
Wolf Wolf 5.00 5.00 5 Pillow Kissen 1.15 1.20 1.17
Echo Echo 5.00 4.87 4.93 Lemon Zitrone 1.21 1.15 1.18
Hotel Hotel 5.00 4.80 4.9 Leg Bein 1.25 1.00 1.13

DOI 10.1007/978-3-642-32194-8
WWW.EBOOK777.COM
178 Appendix A
Test Test O-rating P-rating Mean Con Con O-rating P-rating Mean
English German OP English German con con OP
rating rating
con
Grass Gras 4.50 5.00 4.75 Church Kirche 1.35 1.27 1.31
Garden Garten 4.60 4.87 4.73 Window Fenster 1.15 1.21 1.18
Hand Hand 5.00 4.30 4.65 Head Kopf 1.00 1.00 1.00
Fish Fisch 4.00 5.00 4.5 Mirror Spiegel 1.00 1.00 1.00
Doctor Doktor 4.50 4.28 4.39 Fear Angst 1.00 1.00 1.00
Lamp Lampe 4.32 3.98 4.15 Skirt Rock 1.00 1.00 1.00
Kiss Kuss 3.89 3.98 3.93 Cheese Käse 1.35 1.06 1.20
Hunger Hunger 4.20 3.57 3.88 Juice Saft 1.00 1.00 1.00
Summer Sommer 4.23 3.54 3.88 Animal Tier 1.00 1.00 1.00
Nose Nase 3.70 3.90 3.8 Advice Ratschlag 1.00 1.00 1.00
Milk Milch 4.28 3.30 3.79 Watch Uhr 1.25 1.13 1.19
Mouse Maus 2.50 5.00 3.75 Cherry Kirsche 1.65 1.35 1.5
Baker Bäcker 4.10 3.30 3.7 Soap Seife 1.25 1.15 1.2
Angel Engel 4.12 3.17 3.64 Spoon Löffel 1.00 1.25 1.12
Chance Chance 5.00 2.21 3.6 Size Grösse 1.15 1.13 1.14
Thing Ding 3.10 3.90 3.5 Room Zimmer 1.50 1.45 1.47
Tomato Tomate 3.92 3,03 3.47 Pear Birne 1.24 1.30 1.27
Book Buch 3.54 3.25 3.39 View Aussicht 1.13 1.15 1.14
School Schule 3.78 2.86 3.32 Power Macht 1.00 1.15 1.07
Boat Boot 3.45 3.12 3.28 Bag Tasche 1.00 1.00 1.00
Frog Frosch 2.30 1.23 3.18 Corner Ecke 1.00 1.00 1.00
Soup Suppe 3.12 2.85 2.98 Shirt Hemd 1.13 1.25 1.19
Carrot Karotte 3.15 2.75 2.95 Monkey Affe 1.13 1.00 1.06
Water Wasser 2.70 3.10 2.9 Fruit Obst 1.00 1.00 1.00
Sugar Zucker 2.98 2.76 2.87 Skin Haut 1.00 1.00 1.00
Coffee Kaffe 2.96 2.76 2.86 Bottle Flasche 1.00 1.00 1.00
Salt Salz 2.30 2.90 2.6 Bird Vogel 1.00 1.00 1.00
Stone Stein 2.50 1.90 2.2 Shop Geschäft 1.25 1.14 1.19
Lion Löwe 1.90 1.70 1.8 Tooth Zahn 1.63 1.50 1.56
Rain Regen 1.78 1.45 1.61 Train Zug 1.00 1.00 1.00
King König 1.56 2.35 1.50 Forest Wald 1.00 1.00 1.00
Abbreviations for test words: Test English English orthographic word form, Test German German
orthographic word form, O-rating mean orthographic rating, P-rating mean phonological rating,
Mean OP rating mean of O and P ratings per stimulus word. The same organization holds for the
matched control items, referred to as “con”
WWW.EBOOK777.COM
Appendix B
B.1 Translation-Elicitation Task and Similarity-Rating

Task
List of items, rate of correct translations and the status
German English Translation Translation- Similarity- Similarity-
word word rate (%) elicitation rating rating status
status
Bank Bank 100 Cognate 5 Cognate
Chance Chance 100 Cognate 3.6 Cognate
Doktor Doctor 100 Cognate 4.39 Cognate
Echo Echo 100 Cognate 4.93 Cognate
Engel Angel 100 Cognate 3.64 Cognate
Film Film 100 Cognate 5 Cognate
Finger Finger 100 Cognate 5 Cognate
Fisch Fish 100 Cognate 4.5 Cognate
Garten Garden 100 Cognate 4.73 Cognate
Gold Gold 100 Cognate 5 Cognate
Gras Grass 100 Cognate 4.75 Cognate
Hand Hand 100 Cognate 4.65 Cognate
Hotel Hotel 100 Cognate 4.9 Cognate
Hunger Hunger 100 Cognate 3.88 Cognate
Kaffe Coffee 100 Cognate 2.86 Cognate
Karotte Carrot 100 Cognate 2.95 Cognate
Lampe Lamp 100 Cognate 4.15 Cognate
Lift Lift 100 Cognate 5 Cognate
Moment Moment 100 Cognate 5 Cognate

DOI 10.1007/978-3-642-32194-8
WWW.EBOOK777.COM
180 Appendix B

status
Nase Nose 100 Cognate 3.8 Cognate
Park Park 100 Cognate 5 Cognate
Plan Plan 100 Cognate 5 Cognate
Radio Radio 100 Cognate 5 Cognate
Ring Ring 100 Cognate 5 Cognate
Schule School 100 Cognate 3.32 Cognate
Sommer Summer 100 Cognate 3.88 Cognate
Sport Sport 100 Cognate 5 Cognate
Talent Talent 100 Cognate 5 Cognate
Test Test 100 Cognate 5 Cognate
Tomate Tomato 100 Cognate 3.47 Cognate
Trend Trend 100 Cognate 5 Cognate
Winter Winter 100 Cognate 5 Cognate
Wolf Wolf 100 Cognate 5 Cognate
Bäcker Baker 96 Cognate 3.7 Cognate
Kuss Kiss 96 Cognate 3.93 Cognate
Maus Mouse 96 Cognate 3.75 Cognate
Milch Milk 96 Cognate 3.79 Cognate
Salz Salt 96 Cognate 2.6 Cognate
Regen Rain 91 Cognate 1.61 Cognate
Wasser Water 91 Cognate 2.9 Cognate
Zucker Sugar 91 Cognate 2.87 Cognate
Buch Book 86 Cognate 3.39 Cognate
Suppe Soup 86 Cognate 2.98 Cognate
Boot Boat 78 Cognate 3.28 Cognate
Stein Stone 78 Cognate 2.2 Cognate
Frosch Frog 71 Cognate 3.18 Cognate
Zitrone Lemon 60 Noncognate due to 1.18 Noncognate
lack of form overlap
Kartoffel Potato 59 Noncognate due to 2.4 Noncognate
Schi Ski 59 Cognate due to 4.4 Cognate
form overlap
Ding Thing 58 Cognate due to 3.5 Cognate
form overlap
WWW.EBOOK777.COM
Appendix B 181

status
Flasche Bottle 58 Noncognate due to 1.00 Noncognate
Zimmer Room 57 Noncognate due to 1.47 Noncognate
Ziel Aim 55 Noncognate due to 1.19 Noncognate
Geschäft Shop 55 Noncognate due to 1.19 Noncognate
Löwe Lion 54 Noncognate 1.8 Noncognate
König King 52 Noncognate 1.50 Noncognate
Macht Power 41 Noncognate 1.07 Noncognate
Frau Woman 38 Noncognate 1.1 Noncognate
Kirche Church 31 Noncognate 1.31 Noncognate
Spiegel Mirror 30 Noncognate 1.00 Noncognate
Kopf Head 23 Noncognate 1.00 Noncognate
Kirsche Cherry 22 Noncognate 1.5 Noncognate
Angst Fear 18 Noncognate 1.00 Noncognate
Affe Monkey 17 Noncognate 1.06 Noncognate
Käse Cheese 17 Noncognate 1.20 Noncognate
Tasche Bag 17 Noncognate 1.00 Noncognate
Weg Road 17 Noncognate 1.00 Noncognate
Seife Soap 15 Noncognate 1.2 Noncognate
Fenster Window 14 Noncognate 1.18 Noncognate
Fluss River 13 Noncognate 1.00 Noncognate
Zahn Tooth 13 Noncognate 1.56 Noncognate
Saft Juice 12 Noncognate 1.00 Noncognate
Zeichen Sign 9 Noncognate 1.3 Noncognate
Grösse Size 9 Noncognate 1.14 Noncognate
Glocke Bell 7 Noncognate 1.28 Noncognate
Baum Tree 5 Noncognate 1.00 Noncognate
Birne Pear 4 Noncognate 1.27 Noncognate
Haut Skin 4 Noncognate 1.00 Noncognate
Wald Forest 4 Noncognate 1.00 Noncognate
Zug Train 4 Noncognate 1.00 Noncognate
WWW.EBOOK777.COM
182 Appendix B

status
Tier Animal 4 Noncognate 1.00 Noncognate
Uhr Watch 3 Noncognate 1.19 Noncognate
Aussicht View 0 Noncognate 1.14 Noncognate
Bein Leg 0 Noncognate 1.13 Noncognate
Beweis Proof 0 Noncognate 1.13 Noncognate
Ecke Corner 0 Noncognate 1.00 Noncognate
Gabel Fork 0 Noncognate 1.00 Noncognate
Hemd Shirt 0 Noncognate 1.19 Noncognate
Herbst Autumn 0 Noncognate 1.00 Noncognate
Kissen Pillow 0 Noncognate 1.17 Noncognate
Lied Song 0 Noncognate 1.00 Noncognate
Löffel Spoon 0 Noncognate 1.12 Noncognate
Messer Knife 0 Noncognate 1.00 Noncognate
Obst Fruit 0 Noncognate 1.00 Noncognate
Pferd Horse 0 Noncognate 1.19 Noncognate
Pflicht Duty 0 Noncognate 1.00 Noncognate
Ratschlag Advice 0 Noncognate 1.00 Noncognate
Rock Skirt 0 Noncognate 1.00 Noncognate
Stühl Chair 0 Noncognate 1.00 Noncognate
Vogel Bird 0 Noncognate 1.00 Noncognate
WWW.EBOOK777.COM
Appendix C
Stimulus materials used in the German lexical decision task—Experiment 1

PGE PG English Control word English Polish
cognates cognates translation German translation translation
Film Konto account Flug flight lot
Test Kino cinema Kirche church kościół
Trend Balkon balcony Bild picture obraz
Park Pech bad luck Tasse cup filiżanka
Hotel Handel commerce Geld money pienia˛dze
Sport Grippe flu Kopf head głowa
Talent Punkt point Heirat marriage małżeństwo
Echo Dach roof Baum tree drzewo
Plan Druck print Tante aunt ciocia
Bank Ziel aim Sache thing rzecz
Radio Urlaub holiday Auge eye oko
Moment Ferien holidays Obst fruit owoc
German Fillers:
SCHULD, GEDULD, ENGEL, KUNST, ZELT, WALD, DING, DUFT, ECKE,

MEER, EILE, STUHL.
Nonwords:
AKTIK, BLANG, BOCH, BRAFT, DASE, DELM, DIEGE, DINNE,

DROTTE, ESSEL, FALM, FANK, FAUS, FERST, FILD, FLON, FUTOR,

DOI 10.1007/978-3-642-32194-8
WWW.EBOOK777.COM
184 Appendix C
GARGE, HADEL, HAMPE, HARSE, HONNE, HORFT, KANKO, KERKEL,

KLONDE, KRECHT, LASE, LEIL, LOGA, LORF, MARM, MOTAR, NEIS,
NOCK, NUTER, PEIM, PROMA, PUNST, RACHT, RUBE, SONIE, STEIE,
TEIER, TEIN, TINDE, TORM, ZINGE.
WWW.EBOOK777.COM
Appendix D
Stimulus materials used in the Polish Lexical decision task—Experiments 2 and 3

PGE cognates PG cognates Control word Polish
Film Konto Ptak
Test Kino Noga
Trend Balkon Kwiat
Park Pech Okno
Hotel Handel Zegar
Sport Grypa Osoba
Talent Punkt Brat
Echo Dach Owoc
Plan Druk Jajko
Bank Cel Ryba
Radio Urlop Pies
Moment Ferie Rower
Polish Fillers:
BUTY; KREW; LUSTRO; MISKA; OGON; OWAD; PASEK; PLAMA;

SMOK; STAW; STOPA; WINDA.
Nonwords:
PIARES; PLAZAT; OKAN; KLOZA; FAZEL; MAGA; MESTA; HAMET;

SKLOT; BUKO; MIASEK; BUSTO; TKABO; KATAN; REKTON; KROT;
MEKA; RAGA; ATOR; MAWA; ROFA; BULA; BALAR; TULA; KROLA;

DOI 10.1007/978-3-642-32194-8
WWW.EBOOK777.COM
186 Appendix D
MAZA; HATAK; RUKAW; PIETA; PORTA; MAGA; KRATAT; PAWEC;

MIOSKA; ABATUR; PAKUR; MARES; NUTER; SPUNK; TORMA; MANT;
BREZ; LAMKO; WODEG; PTUG; DRUF; SMET; SOBA.
WWW.EBOOK777.COM
Appendix E
Stimulus materials used in the German Lexical decision task—Experiment 4

PG cognates EG cognates Control word
German
Konto Wolf Kirche
Kino Ski Tasse
Balkon Grass Bild
Pech Gold Flug
Handel Finger Heirat
Grippe Winter Kopf
Punkt Lift Geld
Dach Ring Baum
Druck Hunger Auge
Ziel Garten Sache
Urlaub Hand Tante
Ferien Fisch Obst
German Fillers:
SCHULD, GEDULD, ENGEL, KUNST, ZELT, WALD, DING, DUFT, ECKE,

MEER, EILE, STUHL.
Nonwords:
AKTIK, BLANG, BOCH, BRAFT, DASE, DELM, DIEGE, DINNE,

DROTTE, ESSEL, FALM, FANK, FAUS, FERST, FILD, FLON, FUTOR,

DOI 10.1007/978-3-642-32194-8
WWW.EBOOK777.COM
188 Appendix E
GARGE, HADEL, HAMPE, HARSE, HONNE, HORFT, KANKO, KERKEL,

KLONDE, KRECHT, LASE, LEIL, LOGA, LORF, MARM, MOTAR, NEIS,
NOCK, NUTER, PEIM, PROMA, PUNST, RACHT, RUBE, SONIE, STEIE,
TEIER, TEIN, TINDE, TORM, ZINGE.
WWW.EBOOK777.COM
References
Abendroth–Timmer, D. and S. Breidbach (eds.) 2000. Handlungsorientierung und

Mehrsprachigkeit. Fremd– und mehrsprachiges Handeln in interkulturellen Kontexten.
Frankfurt am Main: Peter Lang.
Adelman, J.S. 2011. “Letters in time and retinotopic space”, Psychological Review 118, 570–582.
Aitchison, J. 2003a. Words in the Mind: an Introduction to the Mental Lexicon, 3rd edn. Oxford:
Blackwell Publishing.
Aitchison, J. 2003b. A Glossary of Language and Mind. Edinburgh: Edinburgh University Press.
Aitchison, J. 2012. Words in the Mind: an Introduction to the Mental Lexicon, 4th edn.
Wiley–Blackwell.
Albert, M. and L. Obler. 1978. The Bilingual Brain. New York: Academic Press.
Allen, D. and K. Conklin. 2013. Cross-linguistic similarity and task demands for Japanese–
English bilingual processing. PLOS ONE 8/8: e72631.
Altarriba, J. and K.M. Mathis. 1997. “Conceptual and lexical development in second language
acquisition”, Journal of Memory and Language 36: 550–568.
Altarriba, J. and R.R. Heredia. (eds.). 2002. An Introduction to Bilingualism: Principles and
Processes. New York, London: Lawrence Erlbaum Associates.
Andrews, S. 1989. “Frequency and neighborhood effects on lexical access: Activation or
search?”, Journal of Experimental Psychology: Learning, Memory and Cognition 15:
802–814.
Arabski J. (ed.). 2002. Time for words: Studies in foreign language vocabulary acquisition.
Frankfurt am Main: Peter Lang.
Arabski, J. 2002. “Learning strategies of L1, L2 and L3 lexis”, in: J. Arabski (ed.), 209–215.
Arabski, J. 2004. Pragmatics and Language Learning. Kraków: Universitas.
Arabski, J. (ed.). 2007. On foreign language acquisition and effective learning. Katowice:
Wydawnictwo Uniwersytetu Śla˛skiego.
Archibald, A., A. Cogo and J. Jenkins (eds.). 2011. Latest Trends in ELF Research. Newcastle
upon Tyne: Cambridge Scholars Publishing.
Aristotle. Aristotle in 23 Volumes translated by Hugh Tredennick. Cambridge, MA, Harvard
University Press; London, William Heinemann Ltd. 1933, 1989.
Arnberg, L. 1987. Raising Children Bilingually: The Preschool Years. Clevedon: Multilingual
Matters.
Aronin, L. 2005. “Theoretical perspectives of trilingual education”, The International Journal of
the Sociology of Language 171: 7–22.
Aronin, L. and B. Hufeisen. (eds.). 2009. The Exploration of Multilingualism: Development of
Research on L3, Multilingualism and Multiple Language Acquisition. Amsterdam: John
Benjamins.

DOI 10.1007/978-3-642-32194-8
WWW.EBOOK777.COM
190 References
Aronin, L. and B. Hufeisen. (eds.). 2009. The exploration of multilingualism. Development of

research on L3, multilingualism and multiple language acquisition, Amsterdam: Benjamins.
Aronin, L. and D. Singleton. 2012. Multilingualism. Amsterdam: John Benjamins.
Aronin, L. and M. O´Laoire. 2004. “Exploring multilingualism in cultural contexts: towards
a notion of multilinguality”, in: C. Hoffmann and J. Ytsma (eds.), 11–29. Clevedon:
Multilingual Matters.
Auer, P. and Li Wei. (eds.) 2007. Handbook of Multilingualism and Multilingual Communication.
Berlin: Mouton de Gruyter.
Baayen, H. and R. Schreuder (eds.). 2003. Morphological structure in language processing.
Berlin: Mouton de Gruyter.
Baayen, R.H., R. Piepenbrock and L. Gulikers. 1995. The CELEX Lexical Database (Release 2)
[CD–ROM]. Philadelphia, PA: Linguistic Data Consortium, University of Pennsylvania.
Baetens Beardsmore, H. 1986. Bilingualism: Basic Principles. Clevedon: Multilingual Matters.
Bailey, C.N. and R.W. Shuy (eds.). 1973. New Ways of Analyzing Variation in English.
Washington: Georgetown University Press.
Baker, C. 1993. Foundations of Bilingual Education and Bilingualism. Clevendon: Multilingual
Matters.
Baker, C. 2006. Foundations of Bilingual Education and Bilingualism. 4th Edition. Clevedon:
Bakhtin, M.M. 1981. “Discourse in the novel” (C. Emerson & M Holquist, Trans.), in: M.
Holquist (ed.), 259–422.
Balota, D.A., M. Yap and M.J. Cortese. 2006. “Visual word recognition: the journey from fea-
tures to meaning (a travel update)”, in: M.J. Traxler and M.A. Gernsbacher (eds.), 285–377.
Bardel, C. and Y. Falk. 2007. “The role of the second language in third language acquisition: the
case of Germanic syntax”, Second Language Research 23/4: 459–484.
Barron–Hauwaert, S. 2000. Issues surrounding trilingual families: children with
simultaneous exposure to three languages. Retrieved: 14th May 2008, from:
www.spz.tu-darmstadt.de/projekt_ejournal
Bartolotti, J. and V. Marian. 2012. “Language Learning and Control in Monolinguals and
Bilinguals“, Cognitive Science 36/6: 1129–1147.
Basnight–Brown, D.M. and J. Altarriba. 2007. “Differences in semantic and translation prim-
ing across languages: The role of language direction and language dominance”, Memory and
Cognition 35: 953–965.
Bassetti, B. and V. Cook. (eds.). 2011. Language and Bilingual Cognition. Oxford, UK:
Psychology Press.
Bassetti, B. and V. Cook. 2011. “Language and cognition: The second language user”, in: B.
Bassetti and V.J. Cook (eds.), 143–190.
Bayer, M., W. Sommer and A. Schacht. 2012. “P1 and beyond: Functional separation of multiple
emotion effects in word recognition”, Psychophysiology, 49, 959–969.
Beauvillain, C. and J. Grainger. 1987. “Accessing interlexical homographs: Some limitations of a
language–selective access”, Journalof Memory and Language 26: 658–672.
Bhatia, T. and W. Ritchie (eds.). 2004. Handbook of Bilingualism. Cambridge, MA: Blackwell
Publishers.
Bialystok E. and F.I.M. Craik. 2009. “Cognitive and linguistic processing in the bilingual
mind”, Current Directions in Psychological Science 19 (1): 19–23.
Bialystok E., F.I.M. Craik, D.W. Green and T.H. Gollan. 2009. “Bilingual minds”, Psychological
Science in the Public Interest 10, 3: 89–129.
Bialystok E., G. Luk, K.F. Peets and S. Yang. 2010. “Receptive vocabulary differences in mono-
lingual and bilingual children”, Bilingualism: Language and Cognition 13, 4: 525–531.
Bialystok, E. (ed.). 1991. Language processing in bilingual children. Cambridge: Cambridge
University Press.
Bialystok, E. 1991. “Metalinguistic dimensions of language proficiency”, in: E. Bialystok (ed.),
113–140.
WWW.EBOOK777.COM
References 191
Bialystok, E. 2001. “Metalinguistic aspects of bilingual processing”, Annual Review of Applied

Linguistics 21: 169–181.
Biedermann, B., N. Ruh, L. Nickels and M. Coltheart. 2008. “Information retrieval in tip of the
tongue states: New data and methodological advances”, Journal of Psycholinguistic Research
37: 171–198.
Bierwisch, M. and R. Schreuder. 1992. “From concepts to lexical items”, Cognition 42: 23–60.
Bild, E.R. and M. Swain. 1989. “Minority language students in a French immersion programme:
their French proficiency”, Journal of Multilingual and Multicultural Development 10/3:
255–274.
Binder, K.S. and K. Rayner. 1998. “Contextual strength does not modulate the subordinate
bias effect: Evidence from eye fixations and self-paced reading“, Psychonomic Bulletin and
Review5/2: 271–276.
Biran, M. and N. Friedmann. 2012. “The representation of lexical-syntactic informa-
tion: Evidence from syntactic and lexical retrieval impairments in aphasia”, Cortex
48/9:1103–1127.
Björklund, S. and I. Suni. 2000. “The Role of English as L3 in Swedish Immersion Programme
in Finland: Impacts on language teaching and language relations”, in: J. Cenoz and U.
Jessner. (eds.), 198–22.
Black, M. 1949. Language and Philosophy: Studies in Methods. Ithaca, New York: Cornell
University Press.
Blevins, J. 2004. Evolutionary Phonology: The Emergence of Sound Patterns. Cambridge:
Cambridge University Press.
Bloomfield, L. 1933. Language. New York: Holt, Rinehart and Winston.
Bolinger, D. 1977. Meaning and Form. London and New York, Longman.
Bouma, H. and D. Bouwhuis (eds.). 1994. Attention and Performance X: Control of Language
Processes. Hillsdale, NJ: Lawrence Erlbaum.
Bozic, M. and W.D. Marslen-Wilson. 2010. “Neurocognitive contexts for morphological
complexity: Dissociating inflection and derivation”, Language and Linguistics Compass
4/11:1063–1073.
Bozic, M., L.K. Tyler, L. Su, C. Wingfield and W.D. Marslen-Wilson. 2013. “Neurobiological
systems for lexical representation and analysis in English”, Journal of Cognitive
Neuroscience 25: 1678–1691.
Braun, M. 1937. “Beobachtungen zur Frage der Mehrsprachigkeit”, Göttingische Gelehrte
Anzeigen: 115–130.
Braunmüller, K. and F. Ferraresi. 2003. Aspects of Multilingualism in European Language
History. Amsterdam: John Benjamins.
Briggs, J. and F.D. Peat. 1989. Turbulent Mirror. New York: Harper & Row.
Brigham, C.C. 1923. A Study of American Intelligence. Princeton: Princeton University Press.
Brown, G., K. Malmkjaer and J. Williams. (eds.). 1996. Performance and Competence in Second
Language Acquisition. Cambridge: Cambridge University Press.
Browne, R. 1982. Aural and visual recognition of cognates and their implications for the
teaching of cognate languages. Ph.D. dissertation, Harvard University, distributed by UMI
Dissertation Information Service.
Brysbaert, M. and T. Dijkstra. 2006. “Changing views on word recognition in bilinguals”, in: J.
Morais and G. d’Ydewalle. (eds), 25–37.
Brysbaert, M., G. Van Dyck and M. Van de Poel. 1999. “Visual word recognition in bilinguals:
Evidence from masked phonological priming”, Journal of Experimental Psychology: Human
Perception and Performance 25: 137–148.
Bullinaria, J., D. Glasspool and G. Houghton (eds.). 1997. Proceedings of the 4th Annual Neural
Computation and Psychology Workshop.
Bultena, S., T. Dijkstra and J.G. van Hell. 2013. “Cognate effects in sentence context depend on
word class, L2 proficiency, and task”, The Quarterly Journal of Experimental Psychology,
DOI: 10.1080/17470218.2013.853090
WWW.EBOOK777.COM
192 References
Butterworth, B. (ed.). 1983. Language Production Volume 2: Development, Writing and Other
Language Processes. London: Academic Press.
Butterworth, B. 1983. “Lexical representation”, in: B. Butterworth (ed.), 257–294.
Caramazza, A., A. Laudanna and C. Romani. 1988. “Lexical access and inflectional morphol-
ogy”, Cognition 28: 297–332.
Caramazza, A., and I. Brones. 1979. “Lexical access in bilinguals”, Bulletin of the Psychonomic
Society 13: 212–214.
Carr, T. H. and D. Dagenbach. 1990. Semantic priming and repetition priming from masked
words: Evidence for a center- surround attentional mechanism in perceptual recognition.
Journal of Experimental Psychology: Learning, Memory, and Cognition 16: 341–350.
Carreiras, M., M. Perea, C. Gil-López, R. Abu Mallouh and E. Salillas. 2013. “Neural correlates
of visual versus abstract letter processing in Roman and Arabic scripts“, Journal of Cognitive
Neuroscience 25/11: 1975–1985.
Carringer, D. 1974. “Creative thinking abilities of a Mexican youth. The relationship of bilin-
gualism”, Journal of Cross–cultural Psychology 5: 492–504.
Carroll, S. 1992. “On cognates”, Second Language Research 8: 93–119.
Cattell, J.M. 1887. “Experiments on the association of ideas”, Mind 12: 68–74.
Cenoz J. 2001. “The effect of linguistic distance, L2 status, and age on cross–linguistic influence
in third language acquisition”, in J. Cenoz, B. Hufeisen and U. Jessner (eds.) 2001a, 8–20.
Cenoz, J. and F. Genesee (eds.). 1998. Beyond Bilingualism. Multilingualism and Multilingual
Education. Clevedon: Multilingual Matters.
Cenoz, J. 2000. “Research on multilingual acquisition”, in: J. Cenoz and U. Jessner (eds.),
39–53.
Cenoz, J. 2003. “The Role of Typology in the Organization of the Multilingual Lexicon”, in: J.
Cenoz, B. Hufeisen and U. Jessner (eds.), 103–116.
Cenoz, J. 2008. “Teaching through Basque: achievements and challenges”, Language, Culture
and Curriculum 21: 1–4.
Cenoz, J. 2013. “Discussion: towards an educational perspective in CLIL language policy and
pedagogical practice”, International Journal of Bilingual Education and Bilingualism 16:
389–394.
Cenoz, J. and D. Gorter. 2011a. “Focus on Multilingualism: a Study of Trilingual Writing”,
TheModern Language Journal 95: 356–369.
Cenoz, J. and D. Gorter. 2011b. “A Holistic Approach to Multilingual Education: Introduction”,
The Modern Language Journal 95: 339–343.
Cenoz, J. and D. Gorter. 2013. “Towards a Plurilingual Approach in English Language Teaching:
Softening the Boundaries Between Languages”, TESOL Quarterly 47: 591–599.
Cenoz, J. and D. Lindsay. 1994. “Teaching English in a primary school: a project to introduce a
third language to eight year olds”, Language and Education 8: 201–210.
Cenoz, J. and F. Genesee. (eds.). 1998. Beyond Bilingualism. Multilingualism and Multilingual
Education. Clevedon: Multilingual Matters.
Cenoz, J. and F. Genesee. 1998. “Psycholinguistic perspectives on multilingualism and multilin-
gual education”, in: J. Cenoz and F. Genesee (eds.), 16–34.
Cenoz, J. and U. Jessner. (eds.). 2000. English in Europe. The Acquisition of a Third Language.
Clevedon: Multilingual Matters.
Cenoz, J. and U. Jessner. 2009. “The study of multilingualism in educational contexts”, in: L.
Aronin and B. Hufeisen. (eds.), 121–138.
Cenoz, J., B. Hufeisen and U. Jessner. (eds.). 2001a. Cross–linguistic Influence in Third
Language Acquisition. Clevedon: Multilingual Matters.
Cenoz, J., B. Hufeisen and U. Jessner. (eds.). 2001b. Looking Beyond Second Language
Acquisition. Tübingen: Stauffenburg Verlag.
Cenoz, J., B. Hufeisen and U. Jessner. (eds.). 2003. Multilingual Lexicon. Kluwer, Dordrecht.
Cenoz, J., B. Hufeisen and U. Jessner. 2003. “Why investigate the multilingual lexicon”, in: J.
Cenoz, B. Hufeisen and U. Jessner (eds.), 1–11.
WWW.EBOOK777.COM
References 193
Chambers, C. G. and H. Cooke. 2009. “Lexical competition during second-language listen-

ing: Sentence context, but not proficiency, constrains interference from the native lexicon”,
Journal of Experimental Psychology: Learning, Memory, and Cognition 35/4: 1029–1040.
Chapelle, C.A. (ed.). 2012. The Encyclopedia of Applied Linguistics. Oxford: Wiley-Blackwell.
Chen, H.C. and C. Ho. 1986. “Development of Stroop interference in Chinese–English bilin-
guals”, Journal of Experimental Psychology; Learning, Memory and Cognition 12, 397–401.
Chen, H.C. and Y.S. Leung, 1989. “Patterns of lexical processing in a non–native language”,
Cheshire, J. 2002 .“Who we are and where we’re going”, in: P. Gubbins and M. Holt (eds.),
19–34.
Chialant, D. and A. Caramazza. 1995. “Where is morphology and how is it processed? The case
of written word recognition” in: L. Feldman (ed.), 55–76.
Chomsky, N. 1988. Language and Problems of Knowledge. Cambridge MA: MIT Press.
Christ, H. 2001. “Wie das Postulat der Erziehung zur Mehrsprachigkeit denFremdsprachenu-
nterricht insgesamt verändert”, Fachverband Moderne Fremdsprachen, Landesverband
Niedersachsen, Mitteilungsblatt 2, 2–9.
Christoffels, I.K., A.M.B. De Groot and J.F. Kroll. 2006. “Memory and language skills in simul-
taneous interpreters: The role of expertise and language proficiency”, Journal of Memory and
Language 54: 324–345.
Cieślicka A. and A. Kujałowicz. 2006. “Cross–linguistic interactions in the trilingual lexicon:
Language selective or non–selective access of interlingual homographs in L3 visual word rec-
ognition?”, in: K. Dziubalska–Kołaczyk (ed.), 131–151.
Cieślicka, A, 2000. “The effect of language proficiency and L2 vocabulary learning strategies
on patterns of bilingual lexical processing”, Poznan Studies in Contemporary Linguistics 36:
27 – 53.
Clahsen, H., I. Sonnenstuhl and J.P. Blevins. 2003. “Derivational morphology in the German
mental lexicon: a dual mechanism account”, in: H. Baayen and R. Schreuder (eds.), 125–155.
Clark, E.V. 1993. The Lexicon in Acquisition. Cambridge: Cambridge University Press.
Clark, H.H. and E.V. Clark. 1978. “Universals, relativity and language processing”, in: J.H.
Greenberg (ed.), 225–278.
Clyne, M. 1997. “Some of the things trilinguals do”, The International Journal of Bilingualism
1/2: 95–116.
Clyne, M. 2003. Dynamics of Language Contact. English and Immigrant Languages.
Cambridge: Cambridge University Press.
Clyne, M., C. Rossi Hunt and T. Isaakidis. 2004. “Learning a community language as a third lan-
guage”, The International Journal of Multilingualism: 1/1: 33–52.
Cogo, A. 2008. “English as a lingua franca: form follows function”, English Today 24/3: 58–61.
Cogo, A. 2012. “ELF and super-diversity: a case study of ELF multilingual practices from a busi-
ness context”, Journal of English as a Lingua Franca 1/2: 287–313
Collins, A.M. and E.F. Loftus. 1975. “A spreading activation theory of semantic processing”,
Psychological Review 82: 407–428.
Collins, A.M. and M.R. Quillian. 1969. “Retrieval time from semantic memory”, Journalof
Verbal Learning and Verbal Behavior 8: 240–248.
Collins, A.M. and M.R. Quillian. 1970. “Does category size affect categorisation time?”, Journal
of Verbal Learning and Verbal Behaviour 9: 432–438.
Coltheart, M., K. Rastle, C. Perry, R. Langdon and J. Ziegler. 2001. “DRC: A dual route cas-
caded model of visual word recognition and reading aloud”, Psychological Review 108:
204–256.
Cook V. 2012a “‘Multi-competence’ and ‘Nativeness and language pedagogy’”, in: C. Chapelle,
(ed.), 3768–3774; 4173–4176.
Cook V. 2012b. “‘The native speaker’ and ‘Multicompetence’”, in: P. Robinson, (ed.), 454–455;
447–450.
Cook V.J. 2007a. “Multi–competence: black hole or wormhole for Second Language Acquisition
research?”, in: Z–H Han (ed.), 16–26.
WWW.EBOOK777.COM
194 References
Cook V.J. 2007b. “The goals of ELT: reproducing native–speakers or promoting multi–compe-
tence among second language users?”, in: J. Cummins and C. Davison (eds.), 237–248.
Cook, V.J. (ed.). 2002. Portraits of the L2 User. Clevedon, Multilingual Matters.
Cook, V.J. (ed.). 2003. Effects of the Second Language on the First. Clevedon: Multilingual
Matters.
Cook, V.J. 1991. “The poverty–of–the–stimulus argument and multi–competence”, Second
Language Research 7/2: 103–117.
Cook, V.J. 1992. “Evidence for multi–competence”, Language Learning 42/4: 557–591.
Cook, V.J. 1993. “Holistic multi–competence: jeu d’esprit or paradigm shift?”, in: B. Kettemann
and W.W. Wieden (eds.), 4–8.
Cook, V.J. 1995. “Multi–competence and the learning of many languages”, Language, Culture
and Curriculum 8/2: 93–98.
Cook, V.J. 1996. “Competence and multi–competence”, in: G. Brown, K. Malmkjaer and J.
Williams. (eds.), 57–69.
Cook, V.J. 2000. “Is transfer the right word? Paper presented at the 7th International
Pragmatics Conference, July, 2000. Budapest.” Retrieved: 12th January 2010 from
http://homepage.ntlworld.com/vivian.c/Writings/Papers/Transfer2000.htm
Cook, V.J. 2002. “Background to the L2 user”, in: V.J. Cook (ed.), 1–28.
Cook, V.J. 2008. Second Language Learning and Language Teaching. 4th ed. London: Hodder
Education.
Cook, V.J. and M. Newson. 1996. Chomsky’s Universal Grammar: An Introduction. 2nd ed.
Oxford: Blackwell.
Copeland, J.E. and P.W. Davis. (eds.). 1981. The Seventh LACUS Forum. Columbia, SC:
Hornbeam Press.
Costa, A., A. Caramazza, A. And N. Sebastian–Galles. 2000. “The cognate facilitation effect:
Implications for models of lexical access”, Journal of Experimental Psychology: Learning,
Memory, and Cognition 26: 1283–1296.
Costa, A., and A. Caramazza. 1999. “Is lexical selection in bilingual speech production lan-
guage–specific? Further evidence from Spanish–English and English–Spanish bilinguals”,
Bilingualism 2: 231–244.
Costa, A., M. Santesteban and A. Caño. 2005. “On the facilitatory effects of cognates in bilingual
speech production”, Brain and Language 94: 94–103.
Craik F.I., E. Bialystok and M. Freedman. 2010. “Delaying the onset of Alzheimer disease: bilin-
gualism as a form of cognitive reserve”, Neurology, 75 (19): 1726–1729.
Cristoffanini, P., K. Kirsner and D. Milech. 1986. “Bilingual lexical representation: The status of
Spanish–English cognates”, Quarterly Journal of Experimental Psychology 38A: 367–393.
Crystal, D. 1997. English as a Global Language. Cambridge: Cambridge University Press.
Crystal, D. 2003. English as a Global Language. Second Edition. Cambridge: Cambridge
University Press.
Cummins, J. 1978. “Bilingualism and the development of metalinguistic awareness”, Journal of
Cross-Cultural Psychology 9: 131–149.
Cummins, J. 1979. “Linguistic interdependence and the educational development of bilingual
children”, Review of Educational Research 49: 222–251.
Cummins, J. and C. Davison (eds.). 2007. Handbook on English Language Teaching. London;
New York: Springer.
Cummins, J. 1976. “The influence of bilingualism on cognitive growth: A synthesis of research
findings and explanatory hypotheses”, Working Papers on Bilingualism 9: 1–43.
Cutler, A. (ed.). 2005. Twenty-first Century Psycholinguistics: Four Cornerstones. London:
Erlbaum.
Davis, C. J. 2010. “The spatial coding model of visual word identification”, Psychological
Review 117: 713–758.
Davis, Ch., R. Sánchez–Casas, J.E.García–Albea, M. Guasch, M. Molero and P. Ferré. 2010.
“Masked translation priming: Varying language experience and word type with Spanish–
English bilinguals”, Bilingualism: Language and Cognition 13/2: 137–155.
WWW.EBOOK777.COM
References 195
De Angelis, G. 2007. Third or Additional Language Acquisition. Clevedon, UK: Multilingual

Matters.
De Angelis, G. 2011. “Teachers’ beliefs about the role of prior language knowledge in language
learning”, International Journal of Multilingualism 8/3: 216–234.
De Bleser, R., P. Dupont, J. Postler, G. Bormans, D. Speelman, L. Mortelmans, M. Debrock.
2003. “The organisation of the bilingual lexicon: a PET study”, Journal of Neurolinguistics
16: 439–456.
de Bot, K. 2004. “The multilingual lexicon: Modelling selection and control”, International
Journal of Multilingualism 1: 17–32.
de Bot, K. and L. Lowie 2010. “7. On the stability of representations in the multilingual lexicon”,
in: M. Pütz and L. Sicola (eds.), 117–134.
de Bot, K., W. Lowie, M. Verspoor. 2007. “A dynamic system theory approach to second lan-
guage acquisition”, Bilingualism: Language and Cognition 10: 7–21.
de Bruin, A., A. Roelofs, T. Dijkstra and I. FitzPatrick. 2014. “Domain-general inhibition areas
of the brain are involved in language switching: FMRI evidence from trilingual speakers”,
NeuroImage 90: 348–359.
De Groot A.M.B and J.F. Kroll (eds.). 1997. Tutorials in Bilingualism. Psycholinguistic
Perspectives. Mahwah, NJ: Lawrence Erlbaum.
De Groot, A.M.B. 1992. “Determinants of word translation”, Journal of Experimental
Psychology: Learning, Memory, and Cognition 18: 1001–1018.
De Groot, A.M.B. 1993. “Word–type effects in bilingual processing tasks: support for a mixed–
representational system”, in: R. Schreuder and B. Weltens (eds.), 27–52.
De Groot, A.M.B. 2002. “Lexical representation and lexical processing in the L2 user”, in: V.
Cook (ed.), 29–63.
De Groot, A.M.B. 2011. Language and cognition in bilinguals and multilinguals: An introduc-
tion. New York-Hove: Psychology Press
De Groot, A.M.B. 2013a. “Reading”, in F. Grosjean and P. Li (eds.), 73–99.
De Groot, A.M.B. 2013b. “Bilingual memory”, in F. Grosjean and P. Li (eds.), 171–191.
De Groot, A.M.B. and C.J. Hoeks. 1995. “The development of bilingual memory: Evidence from
word translation by trilinguals”, Language Learning 45: 683–724.
De Groot, A.M.B. and G.L.J Nas. 1991. “Lexical representation of cognates and noncognates in
compound bilinguals”, Journal of Memory and Language 30: 90–123.
De Groot, A.M.B. and H. Comijs. 1995. “Translation recognition and translation production:
Comparing a new and an old tool in the study of bilingualism”, Language Learning 45:
467–509.
De Groot, A.M.B. and J.F. Kroll (eds.). 1997. Tutorials in Bilingualism. Psycholinguistic
Perspectives. Hillsdale, NJ: Lawrence Erlbaum.
De Groot, A.M.B. and J.G. Van Hell. 2005. “The learning of foreign language vocabulary”, in:
J.F. Kroll and A.M.B. De Groot (eds.), 9–29.
De Groot, A.M.B. and R. Keijzer. 2000. “What is hard to learn is easy to forget: The roles of
word concreteness, cognate status, and word frequency in foreign–language vocabulary learn-
ing and forgetting”, Language Learning 50/1: 1–56.
De Groot, A.M.B. and R. Poot. 1997. “Word translation at three levels of proficiency in a sec-
ond language: The ubiquitous involvement of conceptual memory”, Language Learning 47:
215–264.
De Groot, A.M.B., L. Dannenburg and J.G.Van Hell. 1994. “Forward and backward word transla-
tion by bilinguals”, Journal of Memory and Language 33: 600–629.
De Groot, A.M.B., P. Delmaar and S.J. Lupker. 2000. “The processing of interlexical
homographs in translation recognition and lexical decision: Support for non–selective
access to bilingual memory”, Quarterly Journal of Experimental Psychology 53A/2:
397–428.
De Groot, A.M.B., S. Borgwaldt, M. Bos and E. van den Eijnden. 2002. “Lexical decision and
word naming in bilinguals: Language effects and task effects”, Journal of Memory and
Language 47: 91–124.
WWW.EBOOK777.COM
196 References
De Houwer, A. 1999. “Two or more languages in early childhood: Some general points and prac-
tical recommendations”, Retrieved: 4th March 2005, from www.cal.org/resources/digest/earl
ychild.html
De Houwer, A. 2004. “Trilingual input and children’s language use in trilingual families in
Flanders”, in: C.Hoffmann and J. Ytsma. (eds.), 118–139.
de Lacy, P. (ed.). 2007. Handbook of phonological theory. Cambridge, MA: Cambridge
University Press.
De Saussure, F. 1959. Course in General Linguistics. Translated and annotated by Roy Harris.
London: Duckworth.
Deese, J. 1962. “The structure of associative meaning”, Psychological Review 69: 161–175.
Deese, J. 1965. The Structure of Association in Language and Thought. Baltimore: Johns
Hopkins University Press.
Dell, G.S. 1988. “The retrieval of phonological forms in production: Tests of predictions from a
connectionist model”, Journal of Memory and Language 27: 124–142.
Dell, G.S., F. Chang and Z.M. Griffin. 1999. “Connectionist Models of Language Production:
Lexical Access and Grammatical Encoding”, Cognitive Science 23: 517–542.
Desmet, T. and W. Duyck. 2007. “Bilingual Language Processing”, Language and Linguistics
Compass 1/3: 168–194.
Diaz, R.M. 1985b. “The intellectual power of bilingualism” Quarterly Newsletter of the
Laboratory of Cornporative Human Cognition 7:16–22.
Diebold, A.R. 1964. “Incipient bilingualism”, in: D. Hymes (ed.), 495–511.
Dijkstra, A. 2003a. “Lexical processing in bilinguals and multilinguals: The word selection prob-
lem”, in: J. Cenoz, B. Hufeisen and U. Jessner (eds.), 11–26.
Dijkstra, A. 2003b. Lexical storage and retrieval in bilinguals, in: R. van Hout, A. Hulk, F.
Kuiken and R. Towell (eds.), 129–150.
Dijkstra, A. 2005. Bilingual visual word recognition and lexical access, in: J.F. Kroll and A. De
Groot (eds.), 178–201.
Dijkstra, A. 2007. The multilingual lexicon, in: M.G. Gaskell. (ed.), 251–265.
Dijkstra, A. and W. van Heuven. 1998. “The BIA–model and bilingual word recognition”, in: J.
Grainger and A. Jacobs (eds.), 189–225.
Dijkstra, A. and W.J.B. van Heuven. 2002a. “The architecture of the bilingual word recognition
system: From identification to decision”, Bilingualism: Language and Cognition 5: 175–197.
Dijkstra, A. and W.J.B. van Heuven. 2002b. “Modeling bilingual word recognition: Past, present
and future. Authors’ response”, Bilingualism: Language and Cognition 5: 219–224.
Dijkstra, A., H. Van Jaarsveld and S. Ten Brinke. 1998. “Interlingual homograph recognition:
Effects of task demands and language intermixing”, Bilingualism: Language and Cognition
1: 51–66.
Dijkstra, A., J. Grainger and W.J.B. van Heuven. 1999. “Recognition of cognates and interlingual
homographs: The neglected role of phonology”, Journal of Memory and Language 41/ 4: 496
–518.
Dijkstra, A., M. Timmermans and H. Schriefers. 2000. “On being blinded by your other lan-
guage: Effects of task demands on interlingual homograph recognition”, Journal of Memory
and Language 42: 445–464.
Dijkstra, T. and J.G. van Hell. 2003. Testing the Language Mode hypothesis using trilinguals.
InternationalJournal of Bilingual Education and Bilingualism, 6: 2–16.
Dijkstra, T., K. Miwa, B. Brummelhuis, M. Sappelli and H. Baayen. 2010. “How cross–language
similarity and task demands affect cognate recognition”, Journal of Memory and Language
62: 284–301.
Diller, K. C. 1978. The Language Teaching Controversy. Rowley, MA: Newbury House.
Dimitropoulou, M., J.A. Duñabeitia and M. Carreiras. 2010. “Masked translation priming effects
with low proficient bilinguals”,Experimental Psychology 57 98–107.
Dmochowska, H. (ed.) 2010. Mały Rocznik Statystyczny – Polska 2010. GUS. Zakład
Wydawnictw Statystycznych: Warszawa.
Dörnyei, Z. 2005. The psychology of the language learner: Individual differences in second lan-
guage acquisition. Mahwah, NJ: Lawrence Erlbaum.
WWW.EBOOK777.COM
References 197
Doughty, C. and M. Long (eds.). 2003. Handbook of Second Language Acquisition. Cambridge,
MA: Blackwell Publishers.
Dufour, R. and J.F. Kroll. 1995. “Matching words to concepts in two languages: A test of the
concept mediation model of bilingual representation”, Memory and Cognition 23: 166–180.
Duyck, W. 2005. “Translation and associative priming with cross–lingual pseudohomophones:
evidence for nonselective phonological activation in bilinguals”, Journal of Experimental
Psychology: Learning Memory and Cognition 31: 1340–59.
Duyck, W. and M. Brysbaert. 2004. “Forward and backward number translation requires con-
ceptual mediation in both balanced and unbalanced bilinguals”, Journal of Experimental
Psychology: Human Perception and Performance 30: 889–906.
Duyck, W., D. Vanderelst, T. Desmet and R.J. Hartsuiker. 2008. “The frequency effect in second–
language visual word recognition”, Psychonomic Bulletin and Review 15/4: 850–855.
Dziubalska–Kołaczyk, K. (ed.). 2006. IFAtuation: A Life in IFA. A Festschrift for Professor Jacek
Fisiak on the Occasion of His 70th Birthday. Poznań: Wydawnictwo Naukowe UAM.
Ecke, P. 2009. “The tip–of–the–tongue phenomenon as a window on (bilingual) lexical retrieval”,
in: A. Pavlenko (ed.) 185–209.
Ecke, P., C.J. Hall. 2013. “Tracking tip-of-the-tongue states in a multilingual speaker: Evidence
of attrition or instability in lexical systems?”, International Journal of Bilingualism
17/6: 734–751.
Edwards, J. 1994. Multilingualism. London: Routledge.
Edwards, J.R. 2012. Multilingualism: Understanding Linguistic Diversity. Continuum.
Ellis, A.W. and A.W. Young. 1996. Human cognitive neuropsychology. Hove, UK: Erlbaum
Ellis, N. and A. Beaton. 1993. “Factors affecting foreign language vocabulary: imagery key-
word mediators and phonological short-term memory”, Quarterly Journal of Experimental
Psychology 46A: 533–558.
Ellis, N.C. and A. Beaton. 1995. “Psycholinguistic determinants of foreign language vocabulary
learning”, in: B. Harley (ed.), 107–165.
Elston-Güttler, K.E., T.C. Gunter and S.A. Kotz. 2005. “Zooming into L2: Global language con-
text and adjustment affect processing of interlin-gual homographs in sentences”, Cognitive
Brain Research 25/1: 57–70.
Elwert, W. 1973. “Das zweisprachige Individuum: ein Selbstzeugnis“, in: Studien zu
romanischen Sprachen. Bd. 4 1–81. Wisebaden: Franz Steiner Verlag.
Emmorey, K. and V. Fromkin. 1988. The mental lexicon, in: F. Newmeyer (ed.), 124–149.
Evens M. (ed.). 1988. Relational Models of the Lexicon. Cambridge: Cambridge University
Press.
Fay, D. and A. Cutler. 1977. Malapropisms and the structure of the mental lexicon. Linguistic
Inquiry 8: 505–520.
Feldman L. (ed.). 1995. Morphological Aspects of Language Processing: Cross-Linguistic
Perspectives. Hillsdale, NJ: Lawrence Erlbaum Associates. Clahsen, H. 2006. “Linguistic
perspectives on morphological processing”, in: D. Wunderlich (ed.), 355–388.
Ferguson C.A., L. Menn and C. Stoel–Gammon (eds.). 1992. Phonological Development:
Models, Research, Implication. Timonium, MD: York Press.
Fernández, E.M. and H. Smith Cairns. 2011. Fundamentals of Psycholinguistics. Hoboken, NJ:
Wiley–Blackwell.
Field, J. 2004. Psycholinguistics: The Key Concepts. London: Routledge.
Fikkert, P. 2005. “Getting sounds structures in mind. Acquisition bridging linguistics and psy-
chology?” in: A.E. Cutler (ed.), 43–56.
Fikkert, P. 2007. “Acquiring phonology”, in: P. de Lacy (ed.), 537–554.
Fishman, J.A. (ed.). 1968. Readings in the Sociology of Language. The Hague: Mouton.
Fitzpatrick, T. 2007. “Word association patterns: unpacking the assumptions”, International
Journal of Applied Linguistics 17/3: 319–331.
Fitzpatrick, T. and C. Izura. 2011. “Word association in L1 and L2”, Studies in Second Language
Acquisition 33: 373–398.
Fodor, J. 1983. The Modularity of Mind: an Essay on Faculty Psychology. Cambridge, MA: MIT
Press.
WWW.EBOOK777.COM
198 References
Fodor, J. 1989. “Why should the mind be modular?”, in: A. George (ed.), 1–22.
Forster, K.I. 1976. “Accessing the mental lexicon”, in: R. J. Wales and E.W. Walker (eds.),
257–287.
Forster, K.I. 1989. Basic issues in lexical processing, in: W.D. Marslen–Wilson (ed.), 75–107.
Forte, R. 2009. “Transfer in L3 Acquisition: The Role of Typology”, in: Y.I. Leong (ed.), 89–15.
Franceschini R. 2009. “The genesis and development of research in multilingualism: perspectives
for future research”, in: L. Aronin and B. Hufeisen (eds.), 27–61.
Franceschini, R., D. Zappatore and C. Nitsch. 2003. “Lexicon in the Brain: What Neurobiology
has to Say about Languages”, in: J. Cenoz, B. Hufeisen and U. Jessner. (eds.), 153–166.
Frederiksen, J.R. and J.F. Kroll. 1976. Spelling and sound: Approaches to the internal lexicon.
Journal of Experimental Psychology: Human Perception and Performance 2/3: 361–379.
French, R.M. 1998. “A simple recurrent network model of bilingual memory”, in: M.A.
Gernsbacher, M.A. and S.J. Derry. (eds.), 368–373.
French, R.M. and M. Jacquet. 2004. “Understanding bilingual memory: models and data”,
Trends in Cognitive Sciences 8: 87–93.
Friel, B.M. and S.M. Kennison. 2001. “Identifying German–English cognates, false cognates,
and non–cognates: methodological issues and descriptive norms”, Bilingualism: Language
and Cognition 4: 249–274.
Fromkin, V.A. (ed.). 1973. Speech Errors as Linguistic Evidence. The Hague: Mouton.
Frost, R. and J. Ziegler. 2007. “Speech and Spelling Interaction: The Interdependence of Visual
and Auditory Word Recogntion”, in: G. Gaskell (ed.), 107–118
Gabryś–Barker, D. 2005. Aspects of Multilingual Storage, Processing and Retrieval. Katowice:
Wydawnictwo Uniwersytetu Śla˛skiego.
Garfinkel, A. and K.E.Tabor. 1991. “Elementary School Foreign Languages and English Reading
Achievement: A New View of the Relationship”. Foreign Language Annals 24: 375–382.
Garman, M. 1990. Psycholinguistics. Cambridge: Cambridge University Press.
Garrod, S. 2006. “Psycholinguistic Research Methods”, in Encyclopedia of Language &
Linguistics (2nd Edition), Vol. 10. S. 251–257.
Gaskell, G. (ed.). 2007. Oxford Handbook of Psycholinguistics. Oxford: Oxford University Press.
Gass, S, and L. Selinker (eds.). 1983. Language Transfer in Language Learning. Rowley, MA:
Newbury House.
Gass, S. M. and L. Selinker. 1994. Second Language Acquisition. Hillsdale, NJ: Lawrence
Erlbaum.
Genesee, F. 2002. “Portrait of the bilingual child”, in: V.J. Cook (ed.), 167–197.
George, A. (ed.), 1989. Reflections on Chomsky. Oxford: Blackwell.
Gerard, L.D. and D.L. Scarborough, 1989. “Language–specific access of homographs by bilin-
guals”, Journal of Experimental Psychology: Learning, Memory, and Cognition 15: 305–315.
Gernsbacher, M.A. and S.J. Derry. (eds.), Proceedings of the Twentieth Annual Cognitive Science
Society Conference. NJ: LEA.
Goddard, H.H. 1917. “Mental tests and the immigrant”, Journal of Delinquency 2: 243–277.
Goldrick, M., J. Folk and B. Rapp. 2010. “Mrs. Malaprop's neighborhood: Using word errors to
reveal neighborhood structure”, Journal of Memory and Language 62: 113–134.
Gollan, T.H., K.I. Forster, and R. Frost. 1997. “Translation priming with different scripts:
Masked priming with cognates and noncognates in Hebrew–English bilinguals”, Journal of
Experimental Psychology: Learning, Memory, and Cognition 23: 1222–1139.
Gomez, P. et al. 2013. “A diffusion model account of masked versus unmasked priming: are
they qualitatively different?”, Journal of Experimental Psychology: Human Perception and
Performace, 39/6: 1731–1740.
Graddol, D. 1997. “The Future of English? A guide to forecasting the popularity of the English
language in the 21st century”, The British Council. Retrieved: 21st November 2009 from
www.britishcouncil.org/learning–elt–future.pdf.
Graddol, D. 2006. “English Next”, The British Council. Retrieved: 4th January 2011 from
www.britishcouncil.org/learning–research–english–next.pdf.
Grainger, J. and A. Jacobs (eds.), Localist Connectionist Approaches to Human Cognition.
Hillsdale, NJ: Lawrence Erlbaum.
WWW.EBOOK777.COM
References 199
Grainger, J. and C. Beauvillain. 1987. “Language blocking and lexical access in bilinguals”,
Quarterly Journal of Experimental Psychology 39A: 295–319.
Grainger, J. and Ch. Frenck-Mestre. 1998. “Masked Priming by Translation Equivalents in
Proficient Bilinguals”, Language and Cognitive Processes 13/ 6: 601–623.
Grainger, J. and T. Dijkstra. 1992. “On the representation and use of language information in
bilinguals”, in: R.J. Harris (ed.), 207–220.
Grauberg, W. 1971. “An error analysis in German of first–year university students”, in: G. Perren
and J. Trim (eds.), 11–26.
Green, D.W. 1986. “Control, activation and resource”, Brain and Language27: 210–223.
Green, D.W. 1998. “Mental Control of the Bilingual Lexico–Semantic System. Bilingualism”,
Language and Cognition 1: 67–81.
Greenberg, J.H. (ed.). 1978. Universals of Language: Method and Theory, vol. l., Stanford, CA:
Stanford University Press.
Grosjean F. and P. Li (eds.). 2013. The Psycholinguistics of bilingualism. Malden, MA & Oxford:
Wiley-Blackwell.
Grosjean, F. (ed.). 2008. Studying bilinguals. Oxford: Oxford University Press.
Grosjean, F. 1982. Life with Two Languages: An Introduction to Bilingualism. Cambridge, MA:
Harvard University Press.
Grosjean, F. 1985. “The bilingual as a competent but specific speaker–hearer”, Journal of
Multilingual and Multicultural Development 6/6: 467–477.
Grosjean, F. 1989. “Neurolinguists, beware! The bilingual is not two monolinguals in one per-
son”, Brain and Language 36: 3–15.
Grosjean, F. 1992. Another view of bilingualism, in: R.J. Harris (ed.), 51–62.
Grosjean, F. 1997. “Processing mixed languages: Issues, findings and models” in: A.M.B. De
Groot and J.F. Kroll (eds.), 225–255.
Grosjean, F. 2001. “The bilingual’s language modes”, in: J.L. Nicol (ed.), 1–22.
Grosjean, F. 2008. Studying Bilinguals. Oxford–NewYork: Oxford University Press.
Grosjean, F. 2010. Bilingual: Life and Reality. Cambridge, MA: Harvard University Press.
Gruneberg, M., P. Morris and R. Sykes (eds.). 1988. Practical Aspects of Memory, Vol. 2.
London: John Wiley and Sons.
Guasch, M., R. Sánchez-Casas, P. Ferré and J.E. García-Albea. 2011. “Effects of the degree
of meaning similarity on cross-language semantic priming in highly proficient bilin-
guals”, Journal of Cognitive Psychology 23/8: 942–961.
Gubbins, P. and M. Holt (eds.). 2002. Beyond Boundaries: Language and Identity in
Contemporary Europe. Clevedon: Multilingual Matters.
Gullifer, J.W., J.F. Kroll and P.E. Dussias. 2013. “When Language Switching has No Apparent
Cost: Lexical Access in Sentence Context“, Frontiers in psychology 4: 1–13.
Haigh, C. A. and D. Jared. 2007. “The activation of phonological representations by bilinguals
while reading silently: Evidence from interlingual homophones”, Journal of Experimental
Psychology: Learning, Memory & Cognition 33:623–644.
Hakuta, K. 1986. Mirror of Language. The Debate on Multilingualism. New York: Basic Books.
Hall, C.J. 1992. “Making the right connections: Vocabulary learning and the mental lexi-
con”, Unpublished manuscript, Universidad de las Américas, Puebla: ERIC Document
Reproduction Service No. ED 363 128.
Hall, C.J. 2002. “The automatic cognate form assumption: Evidence for the Parasitic Model of
vocabulary development”, International Review of Applied Linguistics 40: 69–87.
Hall, C.J. and P. Ecke. 2003. “Parasitism as a default mechanism in L3 vocabulary acquisition”,
in: J. Cenoz, B. Hufeisen and U. Jessner (eds.), 71–85.
Hall, C.J., D. Newbrand, P. Ecke, U. Sperr, V. Marchand and L. Hayes. 2009. “Learners’ implicit
assumptions about syntactic frames in new L3 words: The role of cognates, typological prox-
imity and L2 status”, Language Learning 59/1: 153–202.
Hamers, J. and M. Blanc. 1989. Bilinguality and Bilingualism. Cambridge: Cambridge
University Press.
Han, Z.–H. (ed.), 2007. Understanding Second Language Process. Clevedon: Multilingual
Matters.
WWW.EBOOK777.COM
200 References
Harding, E. and P. Riley. 1986. The Bilingual Family: a Handbook for Parents. Cambridge:
Cambridge University Press.
Harley, B. (ed.). 1995. Lexical Issues in Language Learning. Amsterdam: John Benjamins.
Harley, T.A. 2004. “Does cognitive neuropsychology have a future?”, Cognitive
Neuropsychology 21: 3–16.
Harley, T.A. 2010. Talking the Talk: Language, Psychology and Science. Hove: Psychology
Press.
Harley, T.A. 2008. The psychology of language: from data theory. London: Psychology Press.
Harris, R.J. (ed.). 1992. Cognitive Processing in Bilinguals. Amsterdam: Elsevier Science
Publishers.
Hartsuiker, R.J., M.J. Pickering and E. Veltkamp. 2004. “Is syntax separate or shared between
languages?”, Psychological Science 15: 409–414.
Hartsuiker, R.J., M.J. Pickering. 2008. “Language integration in bilingual sentence production”,
Acta Psychologica 128: 479–489.
Haugen, E. 1956. Bilingualism in the Americas. Alabama: American Dialect Society.
Haverkort, M. 2005. “Linguistic representation and language use in aphasia”, in: A. Cutler. (ed.),
57–68.
Henderson, L., J. Wallis and K. Knight. 1994. “Morphemic structure and lexical access”, in: H.
Bouma and D. Bouwhuis (eds.), 211–226.
Herdina, P. and U. Jessner. 2000. The dynamics of third language acquisition, in: J. Cenoz and U.
Jessner (eds.), 84–98.
Herdina, P. and U. Jessner. 2002. A Dynamic Model of Multilingualism. Clevedon: Multilingual
Matters.
Heredia, R. and J. Altarriba (eds.). 2002. Bilingual Sentence Processing. Amsterdam: Elsevier.
Heredia, R.R. 2008. “Mental models of bilingual memory”, in: J. Altarriba and R.R. Heredia.
(eds.), 39–67.
Herwig, A. 2001. “Plurilingual lexical organisation: Evidence from lexical processing in L1–L2–
L3–L4 translation”, in: J. Cenoz, B. Hufeisen and U. Jessner (eds.). 2001a, 115–137.
Hinger, B. 2001. “An intensive multilingual approach to language acquisition in school context”,
in: J. Cenoz, B. Hufeisen and U. Jessner (eds). 2001b, 121–136.
Hockett, C.F. 1958. A Course in Modern Linguistics. New York: Macmillan.
Hoffmann, C. and A. Stavans. 2007. “The evolution of trilingual code–switching from infancy
to school age: the shaping of trilingual competence through dynamic language dominance”,
InternationalJournal of Bilingualism 11/1: 55–72.
Hoffmann, C. and J. Ytsma. (eds.). 2004. Trilingualism in Family, School and Community.
Hoffmann, C. 2000. “The spread of English and the growth of multilingualism with English in
Europe”, in: J. Cenoz and U. Jessner (eds.), 1–21.
Hoffmann, C. 2001a. “Towards a description of trilingual competence”, InternationalJournal of
Bilingualism 5/1: 1–17.
Hoffmann, C. 2001b. “The status of trilingualism in bilingualism studies”, in: J. Cenoz, B.
Hufeisen and U. Jessner (eds.), 2001a, 13–18.
Hoffmann, C. and J. Ytsma. (eds.) 2004. Trilingualism in Family, School and Community.
Holquist, M. (ed.). 1981. The dialogic imagination: Four essays by M. M. Bakhtin. Austin, TX:
University of Texas Press.
Hoshino, N. and J.F. Kroll. 2008. “Cognate effects in picture naming: Does cross–language acti-
vation survive a change of script?”, Cognition 106: 501–511.
House, J. 2003. “English as a lingua franca: A threat to multilingualism?”, Journal of
Sociolinguistics 7/4: 556–578.
Howes, D.H. and R.L. Solomon. 1951. Visual duration threshold as a function of word probabil-
ity”, Journal of Erperimental Psychology 41: 401–410.
Hu, G. and N. Jiang. 2011. “Semantic integration in listening comprehension in a second language
Evidence from cross-modal priming”, in: P. Trofimovich and K. McDonough (eds.), 199–218.
WWW.EBOOK777.COM
References 201
Hudson, P.T.W. and M.W. Bergman. 1985. “Lexical knowledge in word recognition: Word length
in naming and lexical decision tasks”, Journal of Memory & Language 24: 46–58.
Hufeisen, B. 1991. Englisch als erste und Deutsch als zweite Fremdsprache. Frankfurt am Main:
Peter Lang Verlag.
Hufeisen, B. 2004. “A brief introduction to the linguistic foundations”, in: B. Hufeisen and G.
Neuner (eds.), 7–13.
Hufeisen, B. and Aronin, L. 2009. “On the genesis and development of L3 research, multilingual-
ism and multiple language acquisition”, in: L. Aronin and B. Hufeisen (eds.), 1–11.
Hufeisen, B., and G. Neuner, (eds.). 2004. The Plurilingualism Project: Tertiary language learn-
ing – German after English. Graz: European Centre for Modern Languages.
Hufeisen, B. and N. Marx. 2007. “How can DaFnE and EuroComGerm contribute to the concept
of receptive multilingualism?”, in: J. ten Thije and L. Zeevaert (eds.), 307–321.
Husén, T. and S. Opper (eds.). 1983. Multicultural and Multilingual Education in Immigrant
Countries. Wenner Gren Symposium Series (38), Oxford: Pergamon Press.
Hymes, D. 1964. Language in Culture and Society. New York: Harper and Row.
Iatcu, T. 2000. “Teaching English as a third language to Hungarian–Romanian bilinguals”, in: J.
Cenoz and U. Jessner (eds.), 157–178.
Ibrahim, R. 2009. “The cognitive basis of diglossia in Arabic: Evidence from a repetition priming
study within and between languages”, Psychology research and behavior management 2: 93.
Jared, D. and C. Szucs. 2002. “Phonological activation in bilinguals: Evidence from interlingual
homograph naming”, Bilingualism: Language and Cognition 5/3: 225–239.
Jared, D. and J.F. Kroll. 2001. “Do bilinguals activate phonological representations in one or
both of their languages when naming words?”, Journal of Memory and Language 44: 2–31.
Jarvis, S. and A. Pavlenko. 2008. Cross-linguistic influence in language and
cognition. Routledge.
Jenkins, J. 2012. “English as a Lingua Franca from the classroom to the classroom” ELT Journal
66, (4): 486–494.
Jescheniak, J.D. and W.J.M. Levelt. 1994. “Word frequency effects in speech production:
retrieval of syntactic information and of phonological form”, Journal of Experimental
Psychology: Learning, Memory, and Cognition 20: 824–843.
Jespersen, O. 1922. Language. London: Allen and Unwin.
Jessner, U. 1996. “Towards a dynamic view of multilingualism”, in: M. Pütz. (ed.), 17–30.
Jessner, U. 2006. Linguistic Awareness in Multilinguals: English as a Third Language.
Edinburgh: Edinburgh University Press.
Jessner, U. 2008a. “Teaching third languages: Findings, trends and challenges”, Language
Teaching 41: 15–56.
Jessner, U. 2008b. “A DST model of multilingualism and the role of metalinguistic awareness”,
The Modern Language Journalm92/2: 270–283.
Jessner, U. 2013a. “Dynamics of multilingualism”, in: C.A. Chapelle (ed.), 1798–1805.
Jessner, U. 2013b. “Teaching a Third Language”, in: C.A. Chapelle (ed.), 5662–5665.
Jiang N. and K.I. Forster. 2001. “Cross–language priming asymmetries in lexical decision and
episodic recognition”, Journal of Memory and Language 44: 32–51.
Joaristi, L., L. Lizasoain, J.F. Lukas, and K. Santiago. 2009. “Trilingualism (Spanish, English
and Basque) in the educational system of the Basque country”, InternationalJournal of
Multilingualism 6/1:105–126.
Kaplan, R. (ed.). 2002. Handbook of Applied Linguistics. Oxford: Oxford University Press.
Karpińska–Szaj, K. (ed.), Nauka je˛zyków obcych w dobie integracji europejskiej. Łask: Oficyna
Wydawnicza LEKSEM.
Kecskes, I. 2010. “Situation-bound utterances as pragmatic acts”, Journal of Pragmatics 42, 11:
2889–2897.
Kecskes, I. and L. Albertazzi (eds.). 2007. Cognitive Aspects of Bilingualism. New York:
Springer.
Kellerman, E. 1983. “Now you see it, now you don’t”, in: S. Gass and L. Selinker (eds.),
112–134.
WWW.EBOOK777.COM
202 References
Kellerman, E. 1978. “Giving Learners a Break: Native Language Intuitions as a Source of
Predictions about Transferability”, Working Papers on Bilingualism 15: 59–92.
Kemp, Ch. 2009. “Defining multilingualism”, in: L. Aronin and B. Hufeisen. (eds.), 11–27.
Kempen, G. and P. Hujbers. 1983. “The lexicalization process in sentence production and nam-
ing: Indirect election of words”, Cognition 14: 185–209.
Kerkhofs, R., A. Dijkstra, D. Chwilla and E. De Bruijn. 2006. “Testing a model for bilingual
semantic priming with interlingual homographs: RT and ERP effects”, Brain Research 1068:
170–183.
Kettemann, B. and W.W. Wieden (eds.). 1993. Current Issues in European Second Language
Research. Tübingen: Gunter Narr.
Kim, A. and V.T. Lai. 2012. “Rapid interactions between lexical semantic and word form
analysis during word recognition in context: Evidence from ERPs”, Journal of Cognitive
Neuroscience 24: 1104–1112.
Kim, J. and C. Davis. 2003. Task effects in masked cross–script translation and phonological
priming”, Journal of Memory and Language 49: 484–499.
Kirk–Greene, C.W.E. 1981. French False Friends. London: Routledge and Kegan Paul.
Kirsner, K., E. Lalor, and K. Hird. 1993. “The bilingual lexicon: Exercise, meaning and morphol-
ogy”, in: R. Schreuder and B. Weltens (eds.), 215–248.
Klein, E.C. 1995. “Second versus third language acquisition”, Language Learning 45: 419–465.
Koessler, M. 1975. Les faux amis des vocabulaires anglais et americain. Paris: Libraire Vuibert.
Koessler, M. and J. Derocquigny. 1946. Les faux amis. Paris: Librairie Vuibert.
Kolers, P.A. 1963. “Interlingual word associations”, Journal of Verbal Learning and Verbal
Behaviour 2: 291–300.
Komorowska, H. 2013. “Multilingualism: Its open and hidden agendas” SSLLT 3, 4: 463–482.
Kovelman, I., S.A. Baker and L.A. Petitto. 2008. “Bilingual and monolingual brains compared:
a functional magnetic resonance imaging investigation of syntactic processing and a possible
“neural signature” of bilingualism”, Journal of cognitive neuroscience 20/1: 153–169.
Kramsch, C. 2006. “The multilingual subject”, InternationalJournal of Applied Linguistics 16/1:
97–110.
Kramsch, C. 2014. “Teaching Foreign Languages in an Era of Globalization: Introduction”, The
Modern Language Journal 98/ 1: 296–311.
Kroll, J.F. 1993. “Assessing conceptual representations for words in a second language”, in: R.
Schreuder and B. Weltens (eds.), 53–81.
Kroll, J.F. and A. Dijkstra. 2002. The bilingual lexicon, in: R. Kaplan (ed.), 301–321.
Kroll, J.F. and A.M.B. De Groot (eds.). 2005. Handbook of Bilingualism: Psycholinguistic
Approaches. Oxford: Oxford University Press.
Kroll, J.F. and A.M.B. De Groot. 1997. “Lexical and conceptual memory in the bilingual:
Mapping form to meaning in two languages”, in: A.M.B. De Groot and J.F. Kroll (eds.),
169–199.
Kroll, J.F. and E. Stewart. 1994. “Category interference in translation and picture naming:
Evidence for asymmetric connections between bilingual memory representations”, Journal of
Memory and Language 33: 149–174.
Kroll, J.F. and G. Sunderman. 2003. “Cognitive processes in second language acquisition: The
development of lexical and conceptual representations”, in: C. Doughty and M. Long (eds.),
104–129.
Kroll, J.F. and J. Curley. 1988. “Lexical memory in novice bilinguals: The role of concepts
in retrieving second language words”, in: M. Gruneberg, P. Morris and R. Sykes (eds.),
389–395.
Kroll, J.F. and N. Tokowicz. 2005. Models of bilingual representation and processing, in: J.F.
Kroll and A.M.B. De Groot (eds.), 531–553.
Kroll, J.F. and P. Dussias. 2004. “The comprehension of words and sentences in two languages”,
in: T. Bhatia and W. Ritchie. (eds.), 169–200.
Labov, W. 1973. The boundaries of words and their meanings, in: C. N. Bailey and R. W. Shuy
(eds.), 340–373.
Lado, R. 1957. Linguistics Across Cultures. Ann Arbor: University of Michigan Press.
WWW.EBOOK777.COM
References 203
Lakoff, G. 1982. “Categories”, in: Linguistic Society of Korea, eds., 1–88. Linguistics in the
Morning Calm. Seoul: Hanshin.
Lakoff, G. 1987. Women, Fire and Dangerous Things: What Categories Reveal about the Mind.
Chicago: University of Chicago Press.
Lalor, E. and K. Kirsner. 2000. “Cross–lingual transfer effects between English and Italian cog-
nates and noncognates”, InternationalJournal of Bilingualism 4: 385–398.
Langacker R.W. 1987. Foundations of Cognitive Grammar, Volume I, Theoretical
Prerequisites. Stanford, California: Stanford University Press.
Lasagabaster, D. 2000. “Three languages and three linguistic models in the Basque educational
system”, in: J. Cenoz and U. Jessner (eds.), 157–178.
Lasagabaster, D. 2001. “University students` attitudes towards English as an L3”, in: J. Cenoz, B.
Hufeisen and U. Jessner. (eds.). 2001a, 43–52.
Lemhöfer, K. and A. Dijkstra. 2004. “Recognizing cognates and interlingual homographs:
Effects of code similarity in generalized lexical decision”, Memory and Cognition 32: 4,
533–550.
Lemhöfer, K. and R. Radach. 2009. “Task context effects in bilingual nonword processing.
Experimental Psychology”, 561: 41–47.
Lemhöfer, K., A. Dijkstra and M. Michel. 2004. “Three languages, one ECHO: Cognate effects
in trilingual word recognition”, Language and Cognitive Processes 19: 585–611.
Lemhöfer, K., A. Dijkstra, H. Schriefers, R.H. Baayen, J. Grainger and P. Zwitserlood. 2008.
“Native language influences on word recognition in a second language: A mega–study”,
Leong,Y.I. (ed.). 2009. Third Language Acquisition and Universal Grammar. Clevedon:
Leopold, W. (Volume 1–1939, Volume 2– 1947, Volume 3 and 4– 1949. Speech Development of a
Bilingual Child. A Linguist`s Record. Evanston. Ill.: Northwestern University Press.
Levelt, W.J.M. 1989. Speaking: from Intention to Articulation. Cambridge, MA: Bradford
Books/MIT Press.
Levelt, W.J.M. 1993. Lexical Access in Speech Production. Oxford: Blackwell.
Levelt, W.J.M., A. Roelofs and A.S. Meyer. 1999. “A theory of lexical access in speech produc-
tion”, Behavioral and Brain Sciences 22: 1–38.
Lewis, M.P., G.F. Simons and C.D. Fennig (eds.). 2014. Ethnologue: Languages of the World,
Seventeenth edition. Dallas, Texas: SIL International. Online version: http://www.ethnologu
e.com.
Léwy, N. and F. Grosjean. 2008. The Léwy and Grosjean BIMOLA model, in: F. Grosjean. (ed.),
201–210.
Li Wei. 2003. “Activation of lemmas in the multilingual mental lexicon and transfer in third lan-
guage learning”, in: J. Cenoz, B. Hufeisen and U. Jessner (eds.), 57–70.
Li Wei. 2006. “The multilingual mental lexicon and lemma transfer in third language learning”,
International Journal of Multilingualism 32: 88–104.
Li Wei (ed.). 2007. The Bilingualism Reader. London: Routledge.
Li Wei. 2008. “Research perspectives on bilingualism and multilingualism”, in: L. Wei and M.G.
Moyer (eds.), 1–17.
Li Wei and M.G. Moyer (eds.). 2008. The Blackwell Guide To Research Methods In Bilingualism
And Multilingualism. Oxford: Blackwell.
Li, P. and I. Farkaš. 2002. “A self–organizing connectionist model of bilingual processing”, in: R.
Heredia and J. Altarriba (eds.), 59–85.
Li, P., X. Zhao and B. MacWhinney. 2007. “Dynamic self-organization and early lexical develop-
ment in children”, Cognitive Science 31: 581–612.
Liberman, A.M., F.S. Cooper, D.P. Shankweiler and M. Studdert–Kennedy. 1967. “Perception of
the speech code”, Psychological Review 74: 431–461.
Lotto, L. and A.M.B. De Groot. 1998. “Effects of learning method and word type on acquiring
vocabulary in a new language”, Language Learning 48: 31–69.
Luo L., G. Luk Bial and E. Bialystok. 2010. “Effect of Language Proficiency and Executive
Control on Verbal Fluency Performance in Bilinguals”, Cognition 114: 29–41.
WWW.EBOOK777.COM
204 References
MacKay, D.G. 1978. “Derivational rules and the internal lexicon”, Journal of Verbal Learning
and Verbal Behaviour 17: 61–71.
Mackey, W.F. 1968. “The description of bilingualism”, in: J.A. Fishman (ed.). 554–584.
MacLeod, C.M. 1991. “Half a century of research on the Stroop effect: An integrative review”,
Psychological Bulletin 109: 163–203.
Macnamara, J. 1967. “The bilinguals linguistic performance”, Journal of Social Issues 232:
58–77.
Malakoff, M.E. 1992. “Translation ability: A natural bilingual and metalinguistic
skill”, Advances in psychology, 83: 515–529.
Manchón, R,M. 2013. “Writing“, in: F. Grosjean and P.Li. (eds.), 100–116.
Marcus, S. and U. Frauenfelder. 1985. “Word recognition – uniqueness or deviation? A theoreti-
cal note”, Language and Cognitive Processes 1: 163–169.
Marian, V. and M. Spivey. 2003. “Competing activation in bilingual language processing:
Within– and between–language competition”, Bilingualism: Language and Cognition 6:
97–115.
Marian, V., H.K. Blumenfeld, E. Mizrahi, U. Kania, and A-K Cordes. 2013.“Multilingual Stroop
performance: Effects of trilingualism and proficiency on inhibitory control”, International
Journal of Multilingualism 10/1: 82–104.
Marslen–Wilson, W. and P. Warren. 1994. “Levels of perceptual representation and processes in
lexical access: Words, phonemes, and features”, Psychological Review 101: 653–675.
Marslen–Wilson, W. and A. Welsh, A. 1978. “Processing interactions and lexical access during
word recognition in continuous speech”, Cognitive Psychology 10: 29–63.
Marslen–Wilson, W. and L.K. Tyler. 1980. “The temporal structure of spoken language under-
standing”, Cognition 8: 1–71.
Marslen–Wilson, W.D. (ed.). 1989. Lexical Representation and Process. Cambridge, MA: MIT
Press.
Marslen-Wilson, W.D. 2007. “Morphological processes in language comprehension”, in G,
Gaskell. (ed.), 175–193.
Marslen-Wilson, W.D. and L.K. Tyler. 2007. “Morphology, language and the brain: the
decompositional substrate for language comprehension”, Philos Trans R Soc Lond B Biol
Sci 362(1481): 823–36.
Marx, N. and B. Hufeisen. 2004. “Critical overview of research on third language acquisition and
multilingualism published in the German language”, InternationalJournal of Multilingualism
1/ 2: 141–154.
McClelland, J.L. and D.E. Rumelhart. 1981. “An interactive activation model of context effect in
letter perception: Part 1. An account of basic findings”, Psychological Review 88: 375–407.
McCulloch, W.S. and W. Pitts. 1943. “A logical calculus of ideas immanent in nervous activity”,
Bulletin of Mathematical Biophysics 5: 115–133.
McLaughlin, B. and R. Nation. 1986. “Experts and novices: An information–processing
approach to the ‘good language learner’ problem”, Applied Psycholinguistics 7: 41–56.
Meara, P. 1982. “Word associations in a foreign language:, Nottingham Linguistics Circular 11:
29–38.
Melčuk, I. and A. Zholkovsky. 1988. “The Explanatory Combinatorial Dictionary”, in: M. Evens
(ed.), 41–74.
Midgley, K., P.J. Holcomb and J. Grainger. 2011. “Effects of cognate status on word comprehen-
sion in second language learners: an ERP investigation”, Journal of Cognitive Neuroscience
23: 1634–1647.
Miozzo, M., A, Costa, M, Hernández and B. Rapp. 2010. “Lexical processing in the bilingual
brain: Evidence from grammatical/morphological deficits”, Aphasiology 24: 262–287.
Mirman, D. 2011. “Effects of near and distant semantic neighbors on word produc-
tion”, Cognitive, Affective, and Behavioral Neuroscience 11/1: 32–43.
Mitchell, R. and F. Myles. 2004. Second language learning theories. Second edition. London:
Hodder Arnold.
WWW.EBOOK777.COM
References 205
Montanari, S. 2009. “Pragmatic differentiation in early trilingual development”, Journal of Child

Language 36: 597–627.
Morais, J. and G. d’Ydewalle. (eds.). 2006. Bilingualism and Second Language Acquisition.
Brussels: KVAB.
Morton, J. 1969. “Interaction of information in word recognition”, Psychological Review 76:
165–178.
Müller–Lancé, J. 2003. “A strategy model of multilingual learning”, in: J. Canoz, B. Hufeisen
and U. Jessner (eds.), 117–132.
Muñoz, C. 2000. Bilingualism and trilingualism in school students in Catalonia, in: J. Cenoz and
U. Jessner (eds.), 157–178.
Murray, W.S. and K.I. Forster. 2004. “Serial mechanisms in lexical access: The rank hypothesis”,
Psychological Review, 111, 721–756.
Myers-Scotton, C. 2002. Contact Linguistics: Bilingual Encounters and Grammatical Outcomes.
Oxford University Press.
Nation, I.S.P. 1990. Teaching and Learning Vocabulary. New York: Newbury House.
Neuner, G. 2004. “The concept of plurilingualism and tertiary language didctics”, in: Hufeisen,
B. and G. Neuner. (eds.), 13–34.
New, B., L. Ferrand, C. Pallier, M. Brysbaert. 2006. “Reexamining the word length effect in
visual word recognition: New evidence from the English Lexicon Project”, Psychonomic
Bulletin and Review 13/1: 45–52.
Newmeyer, F. (ed.), 1988. Linguistics: The Cambridge Survey, vol 3. Cambridge: Cambridge
University Press.
Nicol, J.L. (ed.). 2001. One Mind, Two Languages: Bilingual Language Processing. Oxford:
Blackwell.
Norris D. 2005. “How do computational models help us develop better theories?”, in: A. Cutler.
(ed.), 331–346.
Norris, D. 2006. “The Bayesian reader: explaining word recognition as an optimal Bayesian
decision proces”, Psychological Review 113: 327–357.
Norris, D. 2009. “Putting it all together: a unified account of word recognition and reaction-time
distributions”, Psychological Review 116: 207–219.
Norris, D. 2013. “Models of visual word recognition”, Trends in Cognitive Sciences, 17/10:
517–524.
Norris, D. and S. Kinoshita. 2012. “Reading through a noisy channel: why there’s nothing special
about the perception of orthography”, Psychological Review 119: 517–545.
Ó Laoire, M. and D. Singleton. 2009. “The role of prior knowledge in L3 learning and use:
Further evidence of psychotypological dimensions”, in: L. Aronin and B. Hufeisen (eds.),
79–103.
Obler, L.K. and K. Gjerlow. 1999. Language and the Brain. Cambridge: Cambridge University
Press.
Odlin, T. 1989. Language Transfer. Cambridge: Cambridge University Press.
Oksaar, E. 1983. “Multilingualism and multiculturalism from the linguist’s point of view”, in: T.
Husén and S. Opper (eds.), 17–36.
Oldfield, R. C. 1966. “Things, words and the brain objects”, Quarterly Journal of Experimental
Psychology 18: 340–353.
Paradis, M. 1981. “Neurolinguistic organization of a bilingual’s two languages”, in: J.E.
Copeland and P.W. Davis, (eds.), 486–494.
Paradis, M. 1997. “The cognitive neuropsychology of bilingualism” in: A.M.B. De Groot and J.F.
Kroll (eds.), 331–354.
Paradis, M. 2004. A Neurolinguistic Theory of Bilingualism. Amsterdam: John Benjamins.
Paradis, M. 2007. “The neurofunctional components of the bilingual cognitive system” in: I.
Kecskes and L. Albertazzi (eds.), 3–28.
Paradis, M. 2009. Declarative and Procedural Determinants of Second Languages. Amsterdam:
John Benjamins.
WWW.EBOOK777.COM
206 References
Pavlenko, A. (ed.). 2009. The Bilingual Mental Lexicon: Interdisciplinary Approaches. Clevedon:
Pavlenko, A. 2011. Thinking and speaking in two languages. Clevedon, UK: Multilingual
Matters.
Pawlak, M. 2007. “Questions from the language classroom revisited: What does SLA
research have to tell us about effective foreign language pedagogy?” in: J. Arabski (ed.),
57–72.
Peal, E. and W. Lambert. 1962. “The relation of bilingualism to intelligence”, Psychological
Monographs 76: 1–23.
Pecher, D. 2001. “Perception is a two–way junction: Feedback semantics in word recognition”,
Psychonomic Bulletin and Review 8: 545–551.
Peeters, D., T. Dijkstra and J. Grainger. 2013. “The representation and processing of identical
cognates by late bilinguals: RT and ERP effects”, Journal of Memory and Language 68:
315–332.
Perren G. and J. Trim (eds.). 1971. Applications in Linguistics. London: Cambridge University
Press.
Pickering, M.J. and S. Garrod. 2005. “Establishing and using routines during dialogue:
Implications for psychology and linguistics”, in: A. Cutler (ed.), 85–101.
Pickering, M.J. and S. Garrod. 2013. “An integrated theory of language production and compre-
hension”, Behavioral and Brain Sciences 36/4: 329–347.
Potter, M.C. 1979. “Mundane symbolism: The relations among objects, names, and ideas”, in:
N.R. Smith and M.B. Franklin (eds.), 281–294.
Potter, M.C., K.F. So, B. Von Eckardt and L. Feldman. 1984. “Lexical and conceptual representa-
tion in beginning and proficient bilinguals”, Journal of Verbal Learning and Verbal Behavior
23/ 1: 23–38.
Pugh, K.R., K. Rexer and L. Katz. 1994. “Evidence of flexible coding in visual word recog-
nition”, Journal of Experimental Psychology: Human Perception and Performance 20:
807–825.
Pütz M. and L. Sicola (eds.). 2010. Cognitive Processing in Second Language Acquisition: Inside
the learner's mind. Amsterdam / Philadelphia: John Benjamins.
Pütz, M. (ed.). 1996. Language Choices: Conditions, Constraints and Consequences.
Amsterdam: John Benjamins.
Quetz, J. and G. von der Handt (eds.). 2002. Neue Sprachen lehren und lernen.
Fremdsprachenunterricht in der Weiterbildung. Bielefeld: Bertelsmann.
Randall, M. 2007. Memory, psychology and second language learning. Amsterdam: John
Benjamins.
Rastle, K. and M.H. Davis. 2008. “Morphological decomposition based on the analysis of
orthography”, Language and Cognitive Processes 23: 942–971.
Ratcliff, R. 1978. “A theory of memory retrieval”, Psychological Review 85: 59–109.
Ratcliff, R., A. Thapar, P. Gomez and G. McKoon. 2004. “A diffusion model analysis of the
effects of aging in the lexical-decision task”, Psychology and Aging 19: 278–289.
Reichle, E.D. and C.A. Perfetti. 2003. “Morphology in word identification: A word–experience
model that accounts for morpheme frequency effects”, Scientific Studies of Reading, 71:
219–238.
Reid, A.A. and W.D. Marslen-Wilson. 2003. “Lexical representation of morphologically complex
words: Evidence from Polish”, in: R.H. Baayen and R. Schreuder.(eds.), 287–336.
Reimer, J.F., J.S. Brown and T.C. Lorsbach. 2001. “Orthographically mediated inhibition effects:
Evidence of activational feedback during visual word recognition”, Psychonomic Bulletin and
Review 8: 102–110.
Ricciardelli L.A. 1992. “Bilingualism and cognitive development in relation to threshold theory”,
Journal of Psycholinguistic Research 21, 4: 301–316.
Riehl, C.M. 2010. “The mental representation of bilingualism”, Wiley Interdisciplinary Reviews:
Cognitive Science 1/5: 750–758.
Riemer, C. 2002. “Wie lernt man Sprachen?”, in: J. Quetz and G. von der Handt. (eds.), 49–82.
WWW.EBOOK777.COM
References 207
Ringbom, H. 1987. The Role of First Language in Foreign Language Learning. Clevedon:
Ringbom, H. 2001. “Lexical transfer in L3 production”, in: J. Cenoz, B. Hufeisen and U. Jessner
(eds.), 59–68.
Ringbom, H. 2007. Crosslinguistic Similarity in Foreign Language Learning. Clevedon:
Robinson, P. (ed.). 2012. Routledge Encyclopedia of Second Language Acquisition. London:
Routledge.
Roelofs, A. 1992. “A spreading activation theory of lemma retrieval in speaking”, Cognition 42:
107–142.
Roelofs, A. 1999. “Phonological segments and features as planning units in speech production”,
Language and cognitive processes 14: 173–200.
Romaine, S. 1995. Bilingualism. Second Edition. Oxford: Blackwell.
Rosch, E. 1975. “Cognitive representations of semantic categories”, Journal of Experimental
Psychology: General 104: 192–233.
Rosch, E. and C.B. Mervis. 1975. “Family resemblances: Studies in the internal structure of cat-
egories”, Cognitive Psychology 7: 442–460.
Rosenman, M. A. and F. Sudweeks. 1995. “Categorization and prototypes in design”, in: P.
Slezak, T. Caelli and R. Clark. (eds.), 189–212.
Rossell S.L., C.J. Price and A.C. Nobre. 2003.“The anatomy and time course of semantic prim-
ing investigated by fMRI and ERPs”, Neuropsychologia 41: 550–564.
Roux, P.W. 2013. “Words in the Mind: Exploring the relationship between word association and
lexical development”, Polyglossia 24: 80–91.
Rubin, G. S., C.A. Becker and R.H. Freeman. 1979. “Morphological structure and its effect on
word recognition”, Journal of Verbal Learning and Verbal Behaviour 18: 757–767.
Rumelhart, D. and McClelland, J. 1986. Parallel Distributed Processing. Cambridge, MA: MIT
Press.
Rumelhart, D.E. and J.L. McClelland. 1982. “An interactive activation model of context effects
in letter perception: II. The contextual enhancement effect and some tests and extensions of
the model”, Psychological Review 89: 60–94.
Safont Jordà, M.P. 2005. Third Language Learners: Pragmatic Production and Awareness.
Sánchez Casas, R.M., C.W. Davis and J.E. García Albea. 1992. “Bilingual lexical processing:
Exploring the cognate/noncognate distinction”, European Journal of Cognitive Psychology 4:
293–310.
Sánchez–Casas, R. and J.E. García–Albea. 2005. “The representation of cognate and noncognate
words in bilingual memory: Can cognate status be characterized as a special kind of morpho-
logical relation?”, in: J.F. Kroll and A.M.B. De Groot (eds.), 226–250.
Sandra, D. and M. Taft (eds.). 1994. Morphological Structure, Lexical Representation and
Lexical Access. Hove: Lawrence Erlbaum.
Schilling H.H., K. Rayner and J.I. Chumbley. 1998. “Comparing naming, lexical decision, and
eye fixation times: word frequency effects and individual differen-ces”, Mem Cognition
26/6:1270–81.
Schreuder, R. and B. Weltens (eds.). 1993. The Bilingual Lexicon. Amsterdam/Philadelphia: John
Benjamins.
Schreuder, R. and G.B. Flores d’Arcais. 1989. “Psycholinguistic issues in the lexical representa-
tion of meaning”, in: W.D. Marslen–Wilson (ed.), 409–436.
Schwartz, A.I. and J.F. Kroll. 2006. “Bilingual lexical activation in sentence context”, Journal of
Schwartz, A.I., J.F. Kroll and M. Diaz. 2007. “Reading words in Spanish and English: Mapping
orthography to phonology in two languages”, Language and Cognitive Processes 22:
106–129.
Seashore, R. H. and L.D. Eckerson. 1940. “The measurement of individual differences in general
English vocabularies”, Journal of Educational Psychology 31: 14–38.
WWW.EBOOK777.COM
208 References
Seidenberg, M. and J.L. McClelland. 1989. “A distributed developmental model of word recogni-
tion and naming”, Psychological Review 96: 523–568.
Seidlhofer, B. 2004. “Research Perspectives on Teaching English as a Lingua Franca”, Annual
Review of Applied Linguistics 24: 209–239.
Seidlhofer, B. 2011. Understanding English as a Lingua Franca. Oxford:Oxford University
Press.
Sharwood Smith, M.A. 1994. Second Language Acquisition: Theoretical Foundations. London:
Longman.
Sherkina, M. 2003. “The cognate facilitation effect in bilingual speech processing”, Toronto
Working Papers in Linguistics 21: 135–151.
Sherkina–Lieber, M. 2004. “The cognate facilitation effect in bilingual speech processing: The
case of Russian–English bilingualism”, Cahiers linguistiques d'Ottawa 32: 108–121.
Siek–Piskozub T., A. Wach and A. Raulinajtys. 2008. “Research on foreign language teaching in
Poland 2000–2006”, Language Teaching 41/1: 51–86.
Singleton, D. 1999. Exploring the Second Language Mental Lexicon. Cambridge: Cambridge
University Press.
Singleton, D. 2000. Language and the Lexicon. London: Edward Arnold.
Singleton, D. 2002. “Cross-linguistic interactions in the multilingual mental lexicon”, in
Proceeding of the Second International Conference on Third Language Acquisition
(CD-ROM) ed. Hooghiemstra,M., Leeuwarden, Fryske Akademy,
Singleton, D. 2003. “Perspectives on the multilingual lexicon: a critical synthesis”, in: J. Cenoz,
B. Hufeisen and U. Jessner (eds.), 167–176.
Singleton, D., J. Fishman, L. Aronin, L. and M. O Laoire. 2013. Current multilingualism: a new
linguistic dispensation. Berlin, Mouton de Gruyter.
Skutnabb–Kangas, T. 1984. Bilingualism or Not. Clevedon: Multilingual Matters.
Slezak, P., T. Caelli and R. Clark. (eds.). 1995. Perspectives on cognitive science: Theories,
experiments and foundations. Norwood, NJ: Ablex.
Smith, E.E., E.J. Shoben and L.J. Rips. 1974. “Structure and process in semantic memory: A fea-
tural model for semantic decisions”, Psychological Review 81: 214–241.
Smith, J.D. and J.P. Minda, 2002. “Distinguishing prototype-based and exemplar-based pro-
cesses in`` category learning”, Journal of Experimental Psychology: Learning, Memory, &
Smith, M.C. 1997. “How do bilinguals access lexical information?”, in: A.M.B. De Groot and
J.F. Kroll (eds.), 145–168.
Smith, M.C. and K. Kirsner. 1982. “Language and orthography as irrelevant features in colour–
word and picture–word Stoop interference”, Quarterly Journal of Experimental Psychology
34: 153–170.
Smith, N.R. and M.B. Franklin (eds.). 1979. Symbolic functioning in childhood. Hillsdale: N.J.:
Erlbaum.
Smits, E., H. Martensen, A. Dijkstra and D. Sandra. 2006. “Naming interlingual homographs:
Variable competition and the role of the decision system”, Bilingualism: Language and
Soares, C. and F. Grosjean. 1984. “Bilinguals in a monolingual and a bilingual speech mode: The
effect on lexical access”, Memory and Cognition 12/4: 380–386.
Solomyak, O. and A. Marantz. 2009. “Lexical access in early stages of visual word processing: A
single–trial correlational MEG study of heteronym recognition”, Brain and Language 108/3:
191–196.
Spöttl, C. 2001. “Expanding students` mental lexicons. A multilingual student perspective”, in: J.
Cenoz, B. Hufeisen and U. Jessner. (eds.). 2001b, 161–178.
Stanners, R.F., J.J. Neiser, W.P. Hernon and R. Hall. 1979. “Memory representation for morpho-
logically related words”, Journal of Verbal Learning and Verbal Behavior 18: 399–412.
Stavans, A. and M. Muchnik. 2007. “Language production in trilingual children: insights on code
switching and code mixing”, Sociolinguistic Studies 1/3: 483–511.
WWW.EBOOK777.COM
References 209
Stemberger, J.P. 1992. “A connectionist view of child phonology: Phonological processing

without phonological processes,” in: C.A. Ferguson, L. Menn and C. Stoel–Gammon (eds.),
165–189.
Szubko-Sitarek, W. 2012. “Cognate facilitation effects in trilingual word recognition”, Studies in
Second Language Learning and Teaching 1/ 2: 189–208.
Szubko-Sitarek, W. 2014. “On the Role of Pre-service Language Teachers’ Beliefs on
Multilingual Education in Developing Teacher Training Programs” in: W. Szubko-Sitarek, Ł.
Salski and P. Stalmaszczyk (eds.)., 107–116.
Szubko-Sitarek W., Salski Ł. and Stalmaszczyk P. (eds). 2014. Language Learning, Discourse
and Communication. Springer International Publishing Switzerland.
Taft, M. 1981. “Prefix stripping revisited”, Journal of Verbal Learning and Verbal Behaviour 2:
289–297.
Taft, M. 1994. “Interactive–activation as a framework for understanding morphological process-
ing”, in: D. Sandra and M. Taft (eds.), 271–294.
Taft, M. 2004. “Morphological decomposition and the reverse base frequency effect”, Quarterly
Journal of Experimental Psychology 57A: 745–765
Taft, M. and K.I. Forster. 1975. “Lexical storage and retrieval of prefixed words”, Journal of
Verbal Learning and Verbal Behaviour 14: 638–647.
Taylor I. 1976. “Similarity between French and English words: A factor to be considered in bilin-
gual language behavior?”, Journal of Psycholinguistic Research 51:, 85–94.
Taylor, I. 1990. Psycholinguistics: Learning and Using Language. Upper Saddle River, NJ:
Prentice Hall.
Taylor, J.R. 2003. Linguistic categorization. Oxford University Press.
Thije ten, J. and L. Zeevaert (eds.). 2007. Receptive Multilingualism. Linguistic Analyses,
Language Policies and Didactic Concepts. Amsterdam: John Benjamins.
Thomas, J. 1988. “The role played by metalinguistic awareness in second and third language
learning”, Journal of Multilingual and Multicultural Development 9/3: 235–246.
Thomas, J. 1992. “Metalinguistic awareness in second-and third-language learning”, Advances in
psychology 83: 531–545.
Thomas, M.S.C. 1997. “Distributed representations and the bilingual lexicon: One store or
two?”, in: J. Bullinaria, D. Glasspool and G. Houghton (eds.), 240–253.
Thomas, M.S.C. 2002. “Theories that develop”, Bilingualism: Language and Cognition 5/3:
216–217.
Thomas, M.S.C. and W. Van Heuven. 2005. “Computational models of bilingual comprehen-
sion”, in: J.F. Kroll and A.M.B. De Groot (eds.), 202–225.
Titone, D., M. Libben, J. Mercier, V. Whitford and I. Pivneva. 2011. „Bilingual lexical access
during L1 sentence reading: The effects of L2 knowledge, semantic constraint, and L1–L2
intermixing“, Journal of Experimental Psychology: Learning, Memory, and Cognition 37/6:
1412–31.
Tokuhama–Espinosa, T. (ed.). 2003. The Multilingual Mind: Issues Discussed By, For, and About
People Living with Many Languages. Westport: Praeger.
Tokuhama-Espinosa, T. 2008. Living languages: Multilingualism across the lifespan. Westport,
CT: Greenwood.
Toribio, A.J. 2001. “On the emergence of bilingual code–switching competence”, Bilingualism:
Language and Cognition 4/3: 203–231.
Traxler M.J. and M.A. Gernsbacher (eds.). 2006. Handbook of Psycholinguistics 2nd edition.
Amsterdam: Elsevier.
Trofimovich, P. and K. McDonough (eds.). 2011. Applying priming methods to L2 learning,
teaching and research: Insights from Psycholinguistics. Amsterdam: John Benjamins.
Tucker, G.R. 1998. “A global perspective on multilingualism and multilingual education”, in: J.
Cenoz and F. Genesee. (eds), 3–16.
Tyler L.K., H.E. Moss, M. Cooke and R. Ostrin. 1995. “Automatic access of lexical information
in Broca`s aphasia: Against the automaticity hypothesis”, Brain and Language 48: 131–162.
WWW.EBOOK777.COM
210 References
Tzeglov, J., R. Henik, R. Sneg and O. Baruch. 1996. “Unintentional word reading via the pho-
nological route: The Stroop effect with cross–script homophones”, Journal of Experimental
Psychology; Learning, Memory, and Cognition 22: 336–349.
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: 105–122.
Ungerer, F. and H. Schmid. 1996. An Introduction to Cognitive Linguistics. London: Longman.
Valencia, J.F. and J. Cenoz. 1992. “The role of bilingualism in foreign language acquisition”,
Journal of Multilingual and Multicultural Development 13/5: 433–449.
Van Assche E, W. Duyck and R.J. Hartsuiker. 2012. “Bilingual word recognition in a sentence
context”, Front. Psychology 3:174. doi: 10.3389/fpsyg.2012.00174.
Van Assche, E., D. Drieghe, W. Duyck, M. Welvaert and R.J. Hartsuiker. 2011. “The influence
of semantic constraints on bilingual word recognition during sentence reading”, Journal of
Van Assche, E., W. Duyck, R.J. Hartsuiker and K. Diependaele. 2009. “Does bilingualism change
native-language reading? Cognate effects in sentence context”, Psychological Science 20:
923–927.
Van Assche, E., W. Duyck, W. and R.J. Hartsuiker. 2013. Phonological Recoding in Error detec-
tion: A Cross-sectional Study in Beginning Readers of Dutch. PLoS ONE 8/12: 1–11.
Van Hell, J.G. and A.M.B. De Groot. 1998a. “Conceptual representation in bilingual memory:
Effects of concreteness and cognate status in word association”, Bilingualism: Language and
Van Hell, J.G. and A.M.B. De Groot. 1998b. “Disentangling context availability and concreteness
in lexical decision and word translation”, Quarterly Journal of Experimental Psychology,
Section A: Human Experimental Psychology 51: 41–63.
Van Hell, J.G. and A.M.B. De Groot. 2008. “Sentence context modulates visual word recognition
and translation in bilinguals”, Acta Psychologica 128: 431–451.
Van Hell, J.G. and T. Dijkstra. 2002. “Foreign language knowledge can influence native language
performance in exclusively native contexts”, Psychonomic Bulletin and Review 9: 780–789.
Van Heuven, W.J.B. and A. Dijkstra. 2001. The Semantic, Orthographic, and Phonological
Interactive Activation Model. Poster presented at the 12th Conference of the European
Society for Cognitive Psychology, Edinburgh, Scotland.
Van Heuven, W.J.B. and T. Dijkstra. 2010. “Language processing in the bilingual brain: fMRI
and ERP support for the BIA + model”, Brain Research Reviews 64: 104–122.
Van Heuven, W.J.B., A. Dijkstra and J. Grainger. 1998. “Orthographic neighborhood effects in
bilingual word recognition”, Journal of Memory and Language 39: 458–483.
Van Heuven, W.J.B., H. Schriefers, T. Dijkstra and P. Hagoort. 2008. “Language Conflict in the
Bilingual Brain” Cereb. Cortex 18, 11: 2706–2716.
Van Heuven, W.J.B., K. Conklin, E.L. Coderre, T. Guo and T. Dijkstra. 2011. “The influence
of cross-language similarity on within- and between-language Stroop effects in trilinguals”,
Frontiers in Psychology 2: 1–15.
Van Hout, R., A. Hulk, F. Kuiken and R. Towell (eds.). 2003. The Interface between Syntax and
the Lexicon in Second Language Acquisition. Amsterdam: John Benjamins.
van Wijnendaele, I. and M. Brysbaert. 2002. “Visual word recognition in bilingu-als: phono-
logical priming from the second to the first language”, Journal of Experimental Psychology:
Human Perception and Performance 28/3: 616–27.
Vigliocco and Hartsuiker. 2005.“Maximal input and feedback in production and comprehension”,
in: A. Cutler. (ed.), 209–228.
Vitevitch M.S. 1997. “The neighborhood characteristics of malapropisms”, Lang Speech 40:
211–28.
Voga, M. and J. Grainger. 2007. “Cognate status and cross–script translation priming”, Memory
and Cognition,35: 938–952.
Voorwinde, S. 1981. “A lexical and grammatical study in Dutch–English–German trilingualism”,
ITL Review of Applied Linguistics 52: 3–30.
WWW.EBOOK777.COM
References 211
Votaw, M.C. 1992. “A functional view of bilingual lexico-semantic organization”, in: R.J. Harris.
(ed.), . 299–321.
Wales, R.J. and E.W. Walker (eds.). 1976. New Approaches to Language Mechanisms.
Amsterdam: North–Holland.
Wandruszka, M. 1979. Die Mehrsprachigkeit des Menschen. München, Zürich: Piper.
Warren, P. 2013. Introducing Psycholinguistics. Cambridge: Cambridge University Press.
Wartenburger, I., H.R. Heekeren, J. Abutalebi, S.F. Cappa, A. Villringer. 2003. “Early setting of
grammatical processing in the bilingual brain”, Neuron 37: 159–170.
Wattendorf, E., B. Westermann, D. Zappatore, R. Franceschini, G. Lüdi, E-W Radü and C.
Nitsch. 2001. “Different languages activate different subfields in Broca's area”, Neuroimage
13: 624.
Weinreich, U. 1953. Language in Contact: Findings and Problems. New York: Linguistic Circle
of New York.
Weir, C. 1988. Communicative Language Testing with Special Reference to English as a Foreign
Language. Exeter: University of Exeter.
Weisgerber, L. 1966. “Vorurteile und Gefahren der Zweisprachigkeit”, Wirkendes Wort 16:
73–89.
Wełna, J. 1977. “Deceptive words: A study in the contrastive lexicon of Polish and English”,
Papers and studies in contrastive linguistics 7: 73–84.
Whitney, C. 2008. “Comparison of the SERIOL and SOLAR theories of letter-position encod-
ing”, Brain and Language 107: 170–178.
Whitney, C. 2011. “Location, location, location: how it affects the neighborhood (effect)”, Brain
and Language 118: 90–104.
Whitney, C. and P. Cornelissen. 2008. “SERIOL Reading”, Language and Cognitive Processes
23: 143–164.
Widła, H. 2005. “Sposób akwizycji kolejnych je˛zyków: Problemy metodologiczne i implikacje
dydaktyczne“, in: K. Karpińska–Szaj (ed.), 37–48.
Wierzbicka, A. 1985. Lexicography and Conceptual Analysis. Ann Arbor: Karoma.
Williams, S. and B. Hammarberg. 1998. “Language switches in L3 production: Implications for a
polyglot speaking model”, Applied Linguistics 19/3: 295–333.
Wittgenstein, L. 1953. Philosophical Investigations. Translated by G. E. M. Anscombe, Oxford:
Basil Blackwell.
Woutersen, M. 1997. The Organisation of the Bilectal Lexicon. Nijmegen: University of
Nijmegen.
Wunderlich D. (ed.). 2006. Advances in the theory of the lexicon. Mouton de Gruyter: Berlin.
Yip, V. and S. Matthews. 2007. “Relative Clauses in Cantonese-English Bilingual Children:
Typological Challenges and Processing Motivations”, Studies in Second Language
Acquisition 29: 277–300.
Yip, V. and S. Matthews. 2010. “The acquisition of Chinese in bilingual and multilingual con-
texts”, International Journal of Bilingualism 14, 1: 127–146.
Ytsma, J. 2000. “Trilingual primary education in Friesland”, in: J. Cenoz and U. Jessner (eds.),
222–236.
Yudes, C., Macizo, P. and Bajo, T. 2010. “Cognate effects in bilingual language comprehension
tasks”, Cognitive Neuroscience and Neuropsychology 21/7: 507–512.
WWW.EBOOK777.COM

Mul Lex Rec in The Men Lex of Thi Lan Use

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mul Lex Rec in The Men Lex of Thi Lan Use

Uploaded by

Copyright:

Available Formats

Secondebo

free oks ==> www.ebook777.com

Second Language Learning and Teaching

About the Series

More information about this series at http://www.springer.com/series/10129

ISSN 2193-7648 ISSN 2193-7656 (electronic)

Library of Congress Control Number: 2014946749

Springer Heidelberg New York Dordrecht London

© Springer-Verlag Berlin Heidelberg 2015

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

1 Beyond Bilingualism: Issues in Multilingualism . . . . . . . . . . . . . . . . . . 1

2 Modelling the Lexicon: Some General Considerations. . . . . . . . . . . . . 33

2.4 Theories of Semantic Representation. . . . . . . . . . . . . . . . . . . . . . . . . 46

3 Modelling Multilingual Representation and Processing. . . . . . . . . . . . 67

4 Visual Word Recognition in Multilinguals. . . . . . . . . . . . . . . . . . . . . . . 101

4.4.2 The Cognate Facilitation Effect and the Word

5 Exploring the Cognate Facilitation Effect in Multilingual

5.7.1 Research Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

Conclusions and Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

(Second Language Acquisition) or bilingualism and adapt it to a new area, in the

number of languages known to the speaker; the sequence of acquisition of all

(Jessner 2008a). This bias is so pervasive that it is virtually impossible to list

This introductory chapter is meant to serve as a background for a more thorough

© Springer-Verlag Berlin Heidelberg 2015

1.2 The Multilingual World

assumptions of that kind have resulted in a growing number of opinions

Initially, the great majority of researchers investigating psychological, cognitive

1.3 Multilingualism—Terminology and Definitions

In line with the foregoing overview, contemporary researchers have, beyond

1.3.1 Societal and Individual Multilingualism

To date, many scholars (especially in Francophone tradition) have endeavoured to

who in their comprehensive account on bilingualism proposed two separate

1.3.2 Bilingualism Versus Multilingualism

An important issue of relevance in the context of clarifying terminology used with

language learned by a given learner or group of learners” (Sharwood Smith 1994,

L3 prototypically represents the concept of the acquisition of any foreign language

1.3.3 Multilingualism and the Levels and Breadth

and “active, completely equal mastery of two or more languages” (Braun 1937

1.4 Towards a Holistic View of Multilingualism

Multilingual language users constitute a new testing ground in psycholinguistic

exclusively multilingual. Therefore, paraphrasing Cook, it could be argued that the

1.5 Characterizing Multilingual Learners

1.5.1 Contexts of Multilingual Acquisition

Fig. 1.1 Multilingual contexts of acquisition and subjects they concern (based on

Table 1.1 Patterns of trilingual development (based on Cenoz 2000, p. 40)

1.5.2 Cognitive Skills in Multilinguals

The discussion of the influence multilinguality exerts on learners’ linguistic and

A similar viewpoint was put forward by Weisgerber (1966; in Romaine 1995), who

of additional languages and exert a positive influence on cognitive development.

Cenoz (2003) promotes the assumption that bilingualism has an indirect effect

in the area of metalinguistic awareness has been reported in a number of studies

1.6 Research Themes in the Study of Multilingualism

Despite a growing number of thought-provoking papers focusing on issues pertain-

rooted in the bilingual background. In recent times, however, psycholinguists have

Further, it needs to be noted that research on multilingualism has had an

1.6.1 Studies on Sociocultural Aspects of Multilingualism

There is a considerably large number of studies centred around trilingual lan-

Table 1.3 A selection of studies on trilingual acquisition in children

more research in early quadrilingualism (cf. Cenoz 2000), or generally speaking

1.6.2 Studies on Psycholinguistic Aspects of Multilingualism

1.6.3 Studies on Educational Aspects of Multilingualism

Trilingualism through schooling or formal trilingualism has recently attracted

1.2 The Multilingual World

1.3 Multilingualism—Terminology and Definitions

1.3.1 Societal and Individual Multilingualism

1.3.2 Bilingualism Versus Multilingualism

1.3.3 Multilingualism and the Levels and Breadth

1.4 Towards a Holistic View of Multilingualism

1.5 Characterizing Multilingual Learners

1.5.1 Contexts of Multilingual Acquisition

1.5.2 Cognitive Skills in Multilinguals

1.6 Research Themes in the Study of Multilingualism

1.6.1 Studies on Sociocultural Aspects of Multilingualism

1.6.2 Studies on Psycholinguistic Aspects of Multilingualism

1.6.3 Studies on Educational Aspects of Multilingualism

1.6.4 Political Perspective of Multilingual

2.2 The Internal Composition of a Lexical Item

2.3 Towards the Model of the Mental Lexicon

2.3.1 The Mental Lexicon Defined

2.3.2 The Internal Organization of the Lexicon

2.3.3 The Internal Relations Within the Lexicon

2.3.4 Lexical Storage: The Full Listing Hypothesis Versus

2.4 Theories of Semantic Representation

2.4.1 The Hierarchical Network Model

2.4.2 The Spreading Activation Model

2.4.3 The Componential Approach

2.5 Models of Lexical Access in the Mental Lexicon

2.5.1 The Serial Search Model

2.5.2 The Logogen Model

2.5.3 The Cohort Model

2.6 Views on Language Processing

2.6.1 The Modularity Theory

3.2 Mental Representation of Bilingualism

3.2.1 The Separate Storage View

3.2.2 The Shared Storage View

3.2.3 The Mixed Storage View

3.2.3.1 The Distributed Feature Model

3.2.3.2 The Revised Hierarchical Model