Formal Linguistics and Language Education: New Empirical Perspectives (Educational Linguistics (43)) [1st ed. 2020] 3030392562, 9783030392567

This volume focuses on work that has its origin and motivation in formal linguistics and theory-driven research on the a

783 116 3MB

English Pages 229 [223] Year 2020

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Formal Linguistics and Language Education: New Empirical Perspectives (Educational Linguistics (43)) [1st ed. 2020]
 3030392562, 9783030392567

Table of contents :
Contents
Contributors
Formal Linguistics and Language Education: Bridging the Gap
1 Recent Attempts at Bridging the Gap
2 The Contributions
References
Part I Conceptual Foundations
Formal Linguistics and Language Education: A View from Bilingualism Research
1 Introduction
2 Why Not Before, Why Now?
3 Final Words
References
Grammatical Concepts for Pedagogical Grammar
1 Introduction
2 L2 Development and Grammatical Concepts
3 Rationale for Grammatical Concepts
3.1 Awareness, Language Learning and Language Education
3.2 Deriving Grammatical Concepts from Formal Grammar
4 Grammatical Concepts Illustrated
4.1 Pronouns
4.1.1 Person and Number Concepts
4.1.2 Gender Concept
4.1.3 Respect and Formality Concepts
5 Conclusions
References
Part II Native Language Settings
Teaching Word Order Variation with a Constraint-Based View on Grammar
1 Introduction
1.1 Word Order in German
1.2 Word Order in Textbooks
1.3 Outline
2 Word Order in Students' Text: A Quick Corpus Study
2.1 Procedure
2.2 Results
2.2.1 Word Orders Used by the Grade 5/6 Group
2.2.2 Word Orders Used by the Grade 11–13 Group
2.2.3 Comparison
3 The OT-View on Word Order Variation in German
3.1 Optimality Theory
3.2 Constraints on Word Order
3.3 Scenarios for Introducing Constraint Interaction
3.3.1 Who Can Go First in Traffic?
3.3.2 Where to Sit in the Class Room?
3.3.3 What to Do on Saturday Evening?
4 Word Order Variation in the Curriculum
4.1 Fifth Grade
4.2 Grade 11–13
5 Conclusion
References
Grammar Is Irrelevant – The Role of Epistemological Beliefs in Students' Learning Success
1 Introduction
2 A Domain-Specific Competence Model
2.1 Grammatical Content Knowledge
2.2 Epistemological Beliefs
2.2.1 A General Model of Epistemological Beliefs
2.2.2 Towards Modelling Epistemological Beliefs of German Grammar
3 Grammatical Content Knowledge, Beliefs, and Other Variables
3.1 Participants
3.2 Methods and Procedure
3.2.1 The Grammar Test
3.2.2 Level of Epistemological Beliefs
3.2.3 Motivation, Self-Concept & Learning Strategies
3.3 Results
4 Discussion & Conclusion
References
Shaking Students' Beliefs About Grammar: Some Thoughts on the Academic Education of Future Language Teachers
1 Introduction
2 What Do Students Today Think About Grammar?
2.1 Questionnaire 1
2.1.1 The Sample
2.1.2 Data Collection
2.1.3 Data Interpretation
2.2 Questionnaire 2
2.2.1 The Sample
2.2.2 Data Collection
2.2.3 Data Interpretation
2.3 Results and Summary of the Explorative Studies
3 How Can We Deal with Students' Beliefs?
3.1 Essential Areas of Knowledge About Grammar for Future Language Teachers
3.2 Shaking Students' Beliefs: How to Get the Essentials Across or Grammar for Prospective Teachers
3.2.1 Grammar Teaching for Future Teachers
3.2.2 Reflective Teaching: Taking Beliefs into Account
3.2.3 Alternative Ways of Teaching
3.3 Difficulties, Hitches
4 Summary and Outlook
References
Do Linguistic Landscapes Influence the Spelling Competence of Orthographic Beginners? Two Case Studies
1 Introduction
2 Linguistic Landscapes and Spelling Performance
3 How Linguistic Landscapes Affect Spelling Performance
3.1 Hypothesis
3.2 Orthographic Case Example: Linking Positions of N-N Compounds
3.3 Study 1: Frequency of Wrong Spelling in Linguistic Landscapes
3.3.1 Hypothesis
3.3.2 Method
3.3.3 Data Collection and Data Processing
3.3.4 Data Assessment
3.4 Study 2: Field Experiment on Perceiving Linguistic Landscapes
3.4.1 Hypothesis
3.4.2 Method
3.4.3 Test Persons, Data Recording and Data Processing
3.4.4 Data Assessment
4 Discussion and Summary
References
Part III Second/Third Language Settings
The Present Tense in English, Again
1 Introduction
2 Linguistic Property and Its Treatment
2.1 Introduction to Generative Terminology and Its Application to Tense
2.2 Generative Analysis of the Present Simple Form
2.3 Generative Analysis of the Present Progressive Form
2.4 Non-prototypical Interpretations of Present Tense Forms in English
3 Grammatical Aspect in French
3.1 French Aspectual Tenses
3.2 Narrative Commentaries in French
4 L2 Acquisition of Aspect in the English Present Tense
4.1 The Interpretability Hypothesis
4.2 Previous Research on Aspectual Tenses
4.3 Structural Priming Research
5 The Present Study
5.1 Research Questions and Methodology
5.2 Participants
5.2.1 Participants: Study One
5.2.2 Participants: Study Two
6 Results
7 Discussion
References
Post-instruction Processing of Generics in English by Japanese L2 Learners
1 Introduction
2 Explicit Learning and Underlying Linguistic Competence
3 Genericity and L2 Acquisition
3.1 Kind vs. General Statement
3.2 L2 Acquisition of Generics
4 On-Line Processing Studies of Articles
5 Previous Article Interventions
6 The Present Study
6.1 Participants
6.2 Research Questions
6.3 SPR Stimulus Set and Experimental Design
6.4 Procedure
6.5 Results
7 Discussion
8 Conclusion
References
Explicit and Implicit Knowledge of Article Semantics in Belarusian Learners of English: Implications for Teaching
1 Introduction
2 Theoretical Background
2.1 The MOGUL Framework
2.2 Definiteness
2.3 The Learning Task
2.4 Previous Research
3 Method
3.1 Participants
3.2 Tasks and Procedures
3.2.1 The Implicit Task
3.2.2 The Explicit Task
3.3 Analysis
3.3.1 Analysing the Implicit Task
3.3.2 Analysing the Explicit Task
3.3.3 Analysing the Relation Between Explicit and Implicit Knowledge
4 Results
4.1 Results for the Implicit Task
4.2 Results for the Explicit Task
4.3 The Relation Between Explicit and Implicit Knowledge
5 Discussion
5.1 Implications for Teaching
References
L2 Influence in L3 Acquisition: The Role of the L3
1 Introduction
2 Linguistic Description: V-to-T Movement
3 Background
3.1 Results of Our Previous Studies
3.2 Other Studies on L3 German
4 Design of This Study
4.1 Context of Study
4.2 Research Question and Prediction
4.3 Participants
4.4 Experiments
4.5 Procedure and Analysis
5 Results
6 Discussion
6.1 Interpretation of the Results
6.2 Implications for Third Language Teaching
7 Conclusion
References

Citation preview

Educational Linguistics

Andreas Trotzke Tanja Kupisch Editors

Formal Linguistics and Language Education New Empirical Perspectives

Educational Linguistics Volume 43

Series Editor Francis M. Hult, University of Maryland, Baltimore County, Baltimore, USA Editorial Board Marilda C. Cavalcanti, Universidade Estadual de Campinas, Campinas, Brazil Jasone Cenoz, University of the Basque Country, Leioa, Spain Angela Creese, University of Stirling, Stirling, United Kingdom Ingrid Gogolin, University of Hamburg, Hamburg, Germany Christine Hélot, Université de Strasbourg, Strasbourg, France Hilary Janks, University of Witwatersrand, Johannesburg, South Africa Claire Kramsch, University of California, Berkeley, USA Constant Leung, King’s College London, London, United Kingdom Angel Lin, Simon Fraser University, Burnaby, Canada Alastair Pennycook, University of Technology, Sydney, Australia

Educational Linguistics is dedicated to innovative studies of language use and language learning. The series is based on the idea that there is a need for studies that break barriers. Accordingly, it provides a space for research that crosses traditional disciplinary, theoretical, and/or methodological boundaries in ways that advance knowledge about language (in) education. The series focuses on critical and contextualized work that offers alternatives to current approaches as well as practical, substantive ways forward. Contributions explore the dynamic and multilayered nature of theory-practice relationships, creative applications of linguistic and symbolic resources, individual and societal considerations, and diverse social spaces related to language learning. The series publishes in-depth studies of educational innovation in contexts throughout the world: issues of linguistic equity and diversity; educational language policy; revalorization of indigenous languages; socially responsible (additional) language teaching; language assessment; first- and additional language literacy; language teacher education; language development and socialization in non-traditional settings; the integration of language across academic subjects; language and technology; and other relevant topics. The Educational Linguistics series invites authors to contact the general editor with suggestions and/or proposals for new monographs or edited volumes. For more information, please contact the Editor: Natalie Rieborn, Van Godewijckstraat 30, 3300 AA Dordrecht, The Netherlands.

More information about this series at http://www.springer.com/series/5894

Andreas Trotzke • Tanja Kupisch Editors

Formal Linguistics and Language Education New Empirical Perspectives

Editors Andreas Trotzke University of Konstanz, Konstanz, Germany Autonomous University of Barcelona Barcelona, Spain

Tanja Kupisch University of Konstanz Konstanz, Germany UiT The Arctic University of Norway (Tromsø) Tromsø, Norway

ISSN 1572-0292 ISSN 2215-1656 (electronic) Educational Linguistics ISBN 978-3-030-39256-7 ISBN 978-3-030-39257-4 (eBook) https://doi.org/10.1007/978-3-030-39257-4 © Springer Nature Switzerland AG 2020 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. 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. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Contents

Formal Linguistics and Language Education: Bridging the Gap. . . . . . . . . . . Andreas Trotzke and Tanja Kupisch

1

Part I Conceptual Foundations Formal Linguistics and Language Education: A View from Bilingualism Research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fatih Bayram and Jason Rothman Grammatical Concepts for Pedagogical Grammar . . . . . . . . . . . . . . . . . . . . . . . . . . Tom Rankin and Melinda Whong

11 21

Part II Native Language Settings Teaching Word Order Variation with a Constraint-Based View on Grammar. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Daniel Gutzmann and Katharina Turgay

45

Grammar Is Irrelevant – The Role of Epistemological Beliefs in Students’ Learning Success . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Daniela Elsner

69

Shaking Students’ Beliefs About Grammar: Some Thoughts on the Academic Education of Future Language Teachers. . . . . . . . . . . . . . . . . . . . . . . . . . Sandra Döring

91

Do Linguistic Landscapes Influence the Spelling Competence of Orthographic Beginners? Two Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Björn Rothstein Part III Second/Third Language Settings The Present Tense in English, Again . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 Amber Dudley and Roumyana Slabakova v

vi

Contents

Post-instruction Processing of Generics in English by Japanese L2 Learners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 Neal Snape Explicit and Implicit Knowledge of Article Semantics in Belarusian Learners of English: Implications for Teaching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 Anders Agebjörn L2 Influence in L3 Acquisition: The Role of the L3 . . . . . . . . . . . . . . . . . . . . . . . . . . 203 Rosalinde Stadt, Aafke Hulk, and Petra Sleeman

Contributors

Anders Agebjörn Department for Swedish, University of Gothenburg, Gothenburg, Sweden Fatih Bayram Department of Language and Culture, UiT The Arctic University of Norway, Tromsø, Norway Sandra Döring Institut für Germanistik, Universität Leipzig, Leipzig, Germany Amber Dudley Modern Languages and Linguistics, University of Southampton, Southampton, UK Daniela Elsner Institut für Sekundarbildung und Fachdidaktik, Pädagogische Hochschule Vorarlberg, Feldkirch, Austria Daniel Gutzmann Institut für deutsche Sprache und Literatur I, Universität zu Köln, Köln, Germany Aafke Hulk Department of Literary studies and Linguistics, ACLC (Amsterdam Centre of Language and Communication), Amsterdam, The Netherlands Tanja Kupisch Department of Linguistics, University of Konstanz, Konstanz, Germany Department of Language and Culture, UiT The Arctic University of Norway, Tromsø, Norway Tom Rankin School of Education, Johannes-Kepler University, Linz, Austria Jason Rothman Department of Language and Culture, UiT The Arctic University of Norway, Tromsø, Norway Universidad Nebrija, Madrid, Spain Björn Rothstein Germanistisches Institut, Ruhr-Universität Bochum, Bochum, Germany Roumyana Slabakova University of Southampton, Southampton, UK NTNU The Norwegian University of Science and Technology, Trondheim, Norway vii

viii

Contributors

Petra Sleeman Department of Literary studies and Linguistics, ACLC (Amsterdam Centre of Language and Communication), Amsterdam, The Netherlands Neal Snape Department of English Communication, Faculty of International Communication, Gunma Prefectural Women’s University, Tamamura-machi, Sawa-gun, Gunma Prefecture, Japan Rosalinde Stadt Department of Literary Studies and Linguistics, ACLC (Amsterdam Centre of Language and Communication), Amsterdam, The Netherlands Andreas Trotzke Department of Linguistics, University of Konstanz, Konstanz, Germany Center for Theoretical Linguistics (CLT), Autonomous University of Barcelona, Barcelona, Spain Katharina Turgay Institut für Germanistik, Universität Landau, Landau, Germany Melinda Whong Center for Language Education, The Hong Kong University of Science and Technology, Kowloon, Hong Kong SAR

Formal Linguistics and Language Education: Bridging the Gap Andreas Trotzke and Tanja Kupisch

Abstract In this chapter, we provide a brief introduction to recent work in linguistics that has its origin and motivation in formal linguistics and theoretical acquisition research, and on this basis indicates potential connections and contributions to language pedagogy, including students’ and teachers’ beliefs about what ‘grammar’ actually is. Keywords Applied linguistics · Educational linguistics · Language education · Language pedagogy · Second language acquisition · Second language classroom

1 Recent Attempts at Bridging the Gap Applying insights from formal linguistics to language teaching and teacher education is anything but new. However, recent years have seen a development of approaches within formal linguistics that aim to reach out and engage more actively with the field of language pedagogy (Whong et al. 2013; De Knop and Gilquin 2016; Marsden and Slabakova 2018; Gil and Rastelli 2018; Trotzke and Rankin 2020). Given the conceptual background of established language-teaching methodologies such as Communicative Language Teaching or Critical Pedagogy, modern teacher education has shifted from the rigorous study of language structure to a focus on communicative and sociolinguistic underpinnings of teaching languages in a classroom. As a result, Applied Linguistics has expanded to address a range

A. Trotzke () Department of Linguistics, University of Konstanz, Konstanz, Germany Center for Theoretical Linguistics (CLT), Autonomous University of Barcelona, Barcelona, Spain e-mail: [email protected] T. Kupisch Department of Linguistics, University of Konstanz, Konstanz, Germany Department of Language and Culture, UiT The Arctic University of Norway, Tromsø, Norway e-mail: [email protected] © Springer Nature Switzerland AG 2020 A. Trotzke, T. Kupisch (eds.), Formal Linguistics and Language Education, Educational Linguistics 43, https://doi.org/10.1007/978-3-030-39257-4_1

1

2

A. Trotzke and T. Kupisch

of general societal issues involving language and communication, while one of the original concerns of linguistics, namely its application to language pedagogy, has moved away from the focus of attention. To refer to the more specific goal of applying linguistics to language pedagogy, researchers have established the term Educational Linguistics (see Hult 2008 for the historical development of this field, as well as the book series the present volume appears in). More specifically, Hult (2008: 17–18) states that [t]he individual educational linguist, trained in any number of combinations of [ . . . ] relevant areas of study, might have her or his home in a variety of different departments, including anthropology, applied linguistics, area studies, education, English, foreign languages, linguistics, psychology, and sociology. Common to all educational linguists, though, is training in critical thinking of a transdisciplinary nature [ . . . ]

Given this definition and understanding of an educational linguist, it follows that more sociology-oriented research areas such as language policy and language planning are a crucial part of Educational Linguistics too (see Spolsky 2005; Hult 2018). Our present volume on Formal Linguistics and Language Education does not deal with those issues, but instead focuses on work that has its origin and motivation in formal linguistics and theory-driven research on the acquisition of grammar, and on this basis tries to establish links to language pedagogy, including students’ and teachers’ beliefs about what ‘grammar’ actually is. By ‘formal’ and ‘theoretical’ approaches, we do not mean to refer to traditional language pedagogies that consider teaching and analyzing formal grammar as a goal in itself (like in branches of historical philology). Rather, we understand ‘formal linguistics’ as an umbrella term that encompasses all approaches − in theory or empirical acquisition research − using modern tools to analyze linguistic items beyond the word level (such as immediate constituent analysis, structuralist accounts of the language system, and specific syntactic approaches such as X-bar theory). In the context of what we have said above about Applied Linguistics more generally and Educational Linguistics more specifically, we consider our volume a timely publication because insights from descriptive and theoretical linguistics, especially formal approaches to language and grammar, no longer have the natural channel to communicate issues and implications of language teacher training and Educational Linguistics that they once had. Furthermore, the work cited above, which has the explicit goal of linking formal linguistics to language pedagogy, often retains a strong theoretical allegiance, which makes it hard for outsiders to the theory to see the relevance to teaching. Also, in addition to being theoretically siloed, the contribution of work grounded in linguistic theory may face particular hurdles in addressing questions of educational relevance because often, as Widdowson (2003: 4) has put it, “the academic discipline of linguistics [ . . . ] is seen to be an abstruse field of enquiry at several removes from the reality of the language classroom” (see also Widdowson 2000). Recent studies have empirically shown that it is indeed the case that teachers refrain from consulting current linguistic research for reasons such as time constraints, insufficient access to online databases, insufficient knowledge of linguistic terminology (see Sato and Loewen 2019 and Marsden and Kasprowicz

Formal Linguistics and Language Education: Bridging the Gap

3

2017). Moreover, many sources that teachers actually do consult and that they think represent proper linguistic research might actually be of low quality, according to common standards and measures like the Social Sciences Citation Index, which help defining a field of inquiry such as linguistics and language acquisition research. Therefore, we, as linguists, should also take into account teachers’ understanding of what exactly counts as research (see, e.g., Borg 2010 on this point). This situation with its potential confusions and ambiguities has already led researchers to conclude that “the findings of academic research are bound to be no less misleading and unreliable than teachers’ experience and intuitions” (Medgyes 2017: 509), and that, consequently, there is nothing to gain from bridging the gap between linguistic research and language-pedagogy practitioners. At the same time, and on a more positive note, linguists have sought to bridge this gap, without too much theoretical terminology and commitment to a specific framework. We would like to especially highlight Dick Hudson’s work in this context. Hudson (2004, 2008, 2020) has repeatedly (and persistently) stressed that linguistics itself is fundamentally pedagogical, and it always has been. In particular, his work demonstrates that central linguistic concepts such as the first modern tree diagrams in syntax were developed and introduced in a pedagogical context: to improve the teaching of grammar in school. More generally, his historical observations indicate that our (read: the linguists’) concepts of language have always been heavily influenced by education because language users are impacted by the specific language pedagogy they have experienced in school, and we as linguists are concerned with the actual behavior of those users. Accordingly, Hudson (2004: 105–106) points out that linguistics, seen as a whole, has an important interface with education, and that research whose results cross this interface is just as important as that which feeds into, say, neuroscience or child development. [ . . . ] academic linguistics is weakened if we ignore the impact of education on language, so information must cross this interface in both directions.

Hudson thus argues for a two-way bridge between linguistics and education, encouraging both linguists and language-teaching practitioners to bridge the socalled research/practice divide. An important component of his work is that in this context he brings the formal properties of language to the fore. In line with this approach, the collection in our volume exemplifies that there is an increasing amount of (potential) research in formal linguistics more generally and in the linguistic approaches to language acquisition that might prove to be highly relevant and useful for teaching languages in the classroom. Specifically, the present volume aims at bridging the gap between the ‘social turn’ in Educational Linguistics indicated above and formal approaches to natural language to eventually explore new teaching methodologies. Our collection of recent empirical work in this domain focuses especially on current accounts that envisage a subsequent integration of issues and insights from modern formal linguistics into language-teaching practice. We thus challenge recent comments and polemics on the research/practice divide, both from those who encourage stronger connections between academic research and pedagogical practice, and from

4

A. Trotzke and T. Kupisch

those who see research as at best irrelevant to practice, or even as an unwelcome intervention in practice (see, e.g., Borg 2010; Marsden and Kasprowicz 2017; Medgyes 2017 for discussion). The contributions to this volume cover a wide range of empirical linguistic domains and concern aspects of morphosyntax, including word order, inflectional morphology, article systems, pronouns, compounding patterns, as well as orthography and students’ general beliefs about grammar. The first two articles in Part I of the volume (Conceptual Foundations) address the role of formal linguistic concepts for language teaching from a general perspective (Bayram and Rothman; Rankin and Whong), thereby introducing the broader picture and the conceptual starting point of many of the empirical papers in this volume. All other contributions report on empirical studies that cover settings relevant to the language classroom. In particular, while the first set of contributions (Part II: Native Language Settings) addresses the role of grammar instruction and language teaching in the native language, the second set (Part III: Second/Third Language Settings) focuses on empirical approaches to second and third language acquisition. In addition to discussing implications for language teaching, some of these studies explore experimentally how different teaching manipulations facilitate the learning of a second or third language. We consider this volume as one further step towards shaping the emerging field of educational applications of theoretical and descriptive linguistics, and we hope it is of interest to a broad audience interested in state-of-the-art approaches to questions of linguistic theory and language acquisition that find applications in pedagogy.

2 The Contributions Part I: Conceptual Foundations. In their short conceptual contribution, Fatih Bayram and Jason Rothman sketch recent attempts at building bridges between formal linguistics and language education from the view of current research on bilingualism. Specifically, they outline the fundamental distinction between language acquisition and language learning, which has played a crucial role over the last decades, especially in the generative literature on second language acquisition. Bayram and Rothman then provide a short overview of recent research using explicit interventions and instructions in bilingual acquisition contexts. The authors conclude that this work underlines that language teaching can indeed benefit from linguistic research and pedagogical interventions inspired by formal linguistics. Tom Rankin and Melinda Whong’s chapter “Grammatical concepts for pedagogical grammar” is rooted in generative approaches to second language acquisition. The authors first develop the notion of “grammatical concept,” suggesting that grammatical concepts can be derived from universal formal and semantic distinctions. Rankin and Whong then propose ways of using these concepts to

Formal Linguistics and Language Education: Bridging the Gap

5

illuminate learnability issues in second language acquisition. To illustrate their approach, they outline some specific principles for the pedagogic use of grammatical concepts and apply them to the phenomenon of pronominal paradigms and usage−a phenomenon that combines different formal, semantic, and pragmatic notions because pronoun systems across languages vary along these three dimensions. More generally, Rankin and Whong highlight the importance of a linguistically-informed comparative approach to language pedagogy and suggest how learners can make use of specific analytical ways familiar to linguists, but not to most teachers working in language pedagogy. Part II: Native Language Settings. Daniel Gutzmann and Katharina Turgay also focus on grammatical concepts from a formal approach to natural language. In “Teaching word order variation with a constraint-based view on grammar,” they argue that Optimality Theory can be used in the language classroom to teach word order patterns. Focusing on teaching German in German schools, they begin by illustrating the flexibility of German word order, explaining how it is determined by information-structural notions, such as topic and focus. At the same time, as the authors point out, German textbooks tend to represent German word order as if it was fairly rigid. Gutzmann and Turgay argue that students of German could benefit from being taught explicitly that word order is flexible and learn about the functions associated with the different word order patterns in order to structure their own texts coherently. They suggest that (a weighted version of) Optimality Theory can be useful for this purpose, illustrating two concrete scenarios that can be used to familiarize students with constraintbased thinking. The authors conclude by exploring how and where their approach can be implemented in the official (national and federal) standards of education in Germany. Related to the contribution by Gutzmann and Turgay, Daniela Elsner also explores grammar teaching in the German education system. Her article “Grammar is irrelevant: The role of epistemological beliefs in students’ learning success” deals with beliefs about grammar of students at German universities. Elsner draws on literature from education science that has shown a relation between epistemological beliefs and variables such as content knowledge, motivation, and self-concept. In accordance with this strand of research, Elsner postulates that epistemological beliefs function as a predictor for these other variables. The paper reports the results of a questionnaire study. While students’ epistemological beliefs about grammar do not correlate with any other tested variables (i.e., content knowledge, motivation, and self-concept), poor performance in grammar tests does correlate with a lower motivation of the students, a worse self-concept, and the use of repetitive learning strategies. Based on these results, Elsner highlights that motivation is a key factor triggering learning processes, pointing out that there is a great need for reviewing academic curricula and implementing new teaching methods in order to enhance students’ motivation. Sandra Döring is concerned with students’ beliefs about grammar too and investigates how these beliefs might be connected to common formal concepts of

6

A. Trotzke and T. Kupisch

grammar. Her paper “Shaking students’ beliefs about grammar: Some thoughts on the academic education of future language teachers” starts with the observation that although grammar courses constitute a basic component of the first academic phase of language teacher education, German teacher trainees do not feel confident about their knowledge of grammar. Döring highlights that students enter university with certain beliefs about grammar, which work like filters and thus heavily impact the students’ learning success. Two large-scale questionnaire-based studies carried out at different German universities indicate that students see the relevance of grammar and grammar teaching, but nevertheless consider the topic too difficult and abstract. Based on these results, Döring first discusses formal areas of grammar knowledge that she considered essential for prospective teachers, and then goes on suggesting possibilities of how to teach these essentials whilst taking students’ beliefs about grammar into account. Her general point is that grammar teaching for prospective teachers has to be different from grammar teaching for future linguistics researchers. Döring concludes by proposing new teaching and learning methods, highlighting how these can form a new generation of teachers in the context of universities and schools. Björn Rothstein focuses on orthography – a topic of great importance for students at school. His article “Do linguistic landscapes influence the spelling competence of orthographic beginners? Two case studies” investigates to what extent wrong spellings in public spaces influence the spelling performances of beginning writers. In particular, Rothstein investigates the spelling of German NN compounds. Based on a corpus, he establishes the types and frequencies of misspellings. He then discusses a field experiment targeting the question whether wrong spellings in linguistic landscapes (here: ‘writing in public spaces’) has an impact on beginning writers. Although there were a lot of wrong spellings in the linguistic landscape investigated, the results of Rothstein’s study suggest that these wrong spellings do not influence orthographic performance. Rothstein points to methodological issues in his study, arguing that linguistic landscapes might affect the spelling performance of beginning writers after all. Part III: Second/Third Language Settings. While the papers in Part II are dedicated to language teaching in the first (native) language classroom, those in Part III are concerned with teaching a foreign (i.e., second or third) language. In “The present tense in English, again” Amber Dudley and Roumyana Slabakova investigate the linguistic forms that can express aspectual meanings when talking about present events in English. In their study, they tested the effect of structural priming on the use of aspectual tense morphosyntax (present progressive and the present simple) in the English present tense by English native speakers, French upper-intermediate and French advanced L2 learners of English. Their findings demonstrate a high level of individual variation within the group of native English speakers. Moreover, structural priming in task instructions modulated both native and non-native speakers’ choices of tense forms. In contrast to previous accounts, which interpreted optionality in aspectual choices as indicative of a representational deficit, Dudley and Slabakova argue that

Formal Linguistics and Language Education: Bridging the Gap

7

sensitivity to structural priming and the nature of the L2 input influence have a substantial impact on learner patterns. The implications are that teachers of English should take variation in the input into account, and researchers should be aware that structural priming in their instructions might affect study outcomes. The paper by Neal Snape, “Post-instruction processing of generics in English by Japanese L2 learners,” also deals with English as a second language, but from the perspective of native speakers of Japanese. The study focuses on article acquisition in generic contexts, testing in a self-paced reading task to what extent explicit instruction can affect real-time reading times. The general goal of this experiment is to find out whether explicit instruction can lead to faster reading times for the instruction group in the domain of definite kind sentences (e.g., The dodo is extinct), compared with a non-instruction group. The instructed group indeed displays faster reading times compared with the non-instruction group. Snape discusses some implications of his study for both theoretical concepts of second language acquisition and for teaching features of the English article system in the second language classroom. He concludes that even though explicit instruction might result in faster reading times and thus better performance, it is still unclear whether this improvement can be retained by learners over the long term. Anders Agebjörn is also concerned with article acquisition in a second language. In “Explicit and implicit knowledge of article semantics in Belarusian learners of English: Implications for teaching,” he investigates explicit and implicit knowledge of Russian and Belarusian learners of English, whose native languages lack articles. More specifically, Agebjörn asks whether the metalinguistic ability to explain what principles govern the use of definite and indefinite articles correlates with target-like use of articles in a communicative task that targets implicit knowledge. The study shows no correlation between explicit knowledge and the ability to use articles in the experimental setting. Agebjörn discusses his findings in the light of theoretical models in second language acquisition and suggests ways in which explicit instruction and knowledge can nevertheless be used to boost the development of implicit knowledge of English article semantics. The volume concludes with Rosalinde Stadt, Aafke Hulk and Petra Sleeman’s paper on “L2 influence in L3 acquisition: The role of the L3.” The paper is concerned with third language (L3) acquisition, comparing different third languages (French and German), while keeping the first two languages (L1 Dutch, L2 German) (and other variables) constant. While previous studies found a significant effect of English on French word order patterns, the question of the present study was whether a similar result can be observed when the L3 is German and not French. To this end, Stadt et al. investigated intermediate learners of German (in secondary school) with regard to word order patterns. The results show that English plays a significantly smaller role in intermediate learners of L3 German compared to L3 French learners. Stadt et al.’s findings underline the role of typological relatedness and structural similarity between the Germanic languages Dutch, English, and German. More generally, their contribution highlights the relevance of taking into account the diversity of language profiles in the language classroom because all previously acquired languages may affect learner outcomes.

8

A. Trotzke and T. Kupisch

References Borg, S. (2010). Language teacher research engagement. Language Teaching, 43, 391–429. De Knop, S., & Gilquin, G. (Eds.). (2016). Applied construction grammar. Berlin: Mouton de Gruyter. Gil, K. -H., & Rastelli, S. (Eds.). (2018). Second language teaching and generative linguistics. Instructed Second Language Acquisition, 2 (Special Issue). Hudson, R. (2004). Why education needs linguistics (and vice versa). Journal of Linguistics, 40, 105–130. Hudson, R. (2008). Linguistic theory. In B. Spolsky & F. Hult (Eds.), The handbook of educational linguistics (pp. 53–65). Oxford: Blackwell. Hudson, R. (2020). Towards a pedagogical linguistics. Pedagogical Linguistics, 1, 8–33. Hult, F. (2008). The history and development of educational linguistics. In B. Spolsky & F. Hult (Eds.), The handbook of educational linguistics (pp. 1–10). Oxford: Blackwell. Hult, F. (2018). Language policy and planning and linguistic landscapes. In J. W. Tollefson & M. Pérez-Milans (Eds.), Oxford handbook of language policy and planning (pp. 333–351). Oxford: Oxford University Press. Marsden, E., & Kasprowicz, R. (2017). Foreign language educators’ exposure to research: Reported experiences, exposure via citations, and a proposal for action. The Modern Language Journal, 101, 613–642. Marsden, H., & Slabakova, R. (2018). Grammatical meaning and the second language classroom: Introduction. Language Teaching Research, 23, 147–157. https://doi.org/10.1177/ 1362168817752718. Medgyes, P. (2017). The (ir)relevance of academic research for the language teacher. ELT Journal, 71, 491–498. Sato, M., & Loewen, S. (2019). Do teachers care about research? The research-pedagogy dialogue. ELT Journal, 73, 1–10. Spolsky, B. (2005). Is language policy applied linguistics? In P. Bruthiaux, D. Atkinson, W. G. Eggington, W. Grabe, & V. Ramanathan (Eds.), Directions in applied linguistics (pp. 26–36). Clevedon: Multilingual Matters. Trotzke, A., & Rankin, T. (2020). Editorial: Introduction to pedagogical linguistics. Pedagogical Linguistics, 1, 1–7. Whong, M., Gil, K.-H., & Marsden, H. (Eds.). (2013). Universal grammar and the second language classroom. Dordrecht: Springer. Widdowson, H. G. (2000). On the limitations of linguistics applied. Applied Linguistics, 21, 3–25. Widdowson, H. G. (2003). Defining issues in English language teaching. Oxford: Oxford University Press.

Part I

Conceptual Foundations

Formal Linguistics and Language Education: A View from Bilingualism Research Fatih Bayram and Jason Rothman

Abstract This chapter aims to underline the importance of bilingualism research from a formal linguistic perspective for second language pedagogy. In doing so, we highlight where the two fields of inquiry overlap with each other and offer insights into how language pedagogy can benefit from information gained by psycholinguistic studies on specific properties of grammar and its development. Keywords Bilingualism · Language pedagogy · Psycholinguistics

1 Introduction In recent years, some scholars who have traditionally worked on the implicit side of bilingual language acquisition have argued that language teaching could benefit from earnest bridges with generative theoretical linguistics, language acquisition, and processing research (e.g., Whong 2011; Long and Rothman 2013; VanPatten and Rothman 2015; Whong et al. 2014; Marsden and Slabakova 2019). Notable exceptions notwithstanding (e.g., White et al. 1991; VanPatten 1996), formal acquisition scholars have rarely endeavored to make such links prior to the last decade. There are likely several reasons for the apparent disconnect, including, but not limited to, differences in the goals and remit of these traditions. Formal linguistics (theory and acquisition) endeavors to describe and explain the psychological reality, development, and ultimate attainment of grammars in the mind, especially abstract hierarchical structure, acquired in a ‘naturalistic’ wild compatible

F. Bayram () Department of Language and Culture, UiT The Arctic University of Norway, Tromsø, Norway e-mail: [email protected] J. Rothman Department of Language and Culture, UiT The Arctic University of Norway, Tromsø, Norway Universidad Nebrija, Madrid, Spain e-mail: [email protected] © Springer Nature Switzerland AG 2020 A. Trotzke, T. Kupisch (eds.), Formal Linguistics and Language Education, Educational Linguistics 43, https://doi.org/10.1007/978-3-030-39257-4_2

11

12

F. Bayram and J. Rothman

with systematically collected data enabling theoretical justification and prediction. There are various formalisms for understanding grammar in a cognitive science sense that make use of distinct axioms and constructs to understanding language representations in the mind (see, e.g., Goldberg 2019; Jackendoff 2002). By formal linguistics, herein, we refer to generative linguistic theory and its language acquisition corollaries because this is the paradigm under which we work. In general terms, however, the leitmotif message of this paper is applicable regardless of the specific formalism to which one subscribes: language teaching can benefit from linguistic research. Language pedagogy theory and language teaching in the trenches are charged with a much more practical task and are, rightfully, beholden to a different set of questions and needs. Language teaching theory and practice necessarily engages with the dynamic reality of learning itself in the ‘non-naturalistic’ wild with the task of describing a set of grammar rules that are traditionally based on grammar books and/or the teacher’s personal knowledge/competence in the target language (but see, e.g., Rankin & Whong; Döring; and Elsner in this volume). Not surprisingly, then, audiences and stakeholders that are interested in and/or targeted by language pedagogy research and formal linguistic research are only partially overlapping. We see, however, that the overlapping audience is, at present, smaller than it should be. We will discuss our thoughts on the matter as well as research that supports the aforementioned view that insights from linguistic research can be useful for more effective language teaching in the remainder of this chapter. It is fitting that this piece should accompany the present edited collection of papers that, as a whole and individually, provides credence to the general idea that formal linguistics and language teaching are ripe for better communication and cross-fertilization for the mutual benefit of each.

2 Why Not Before, Why Now? Despite having language as their main object of study, formal linguistics and language pedagogy deal with different sets of constraints from their own realworld contexts, individually defined by distinctions in their approaches. In the case of language pedagogy, creating and understanding purposefully designed interventions, teacher training, understanding the learner needs and domain-general constraints on learning serves a dual purpose. The remit is to both understand how language learning unfolds in a classroom context in a general sense, but also to gain knowledge to maximize resources at various levels: what is reasonable to expect in the context of classroom learning and how to best attain learning goals towards a measurable target juxtaposed against procedural, financial, temporal, motivational and many other constraints (see Döring; Elsner in this volume). In this context, descriptions of the characteristics of a language system as described by formal theorists are only useful to language teachers to the extent that this set of information is transferrable to learners as a collection of grammar rules (see Rankin & Whong in this volume). Of course, generative approaches to bilingualism engages with

Formal Linguistics and Language Education: A View from Bilingualism Research

13

acquisition in non-naturalistic classroom environments as well. However, its focus is not on intervention or facilitation of learning in any sense. In fact, it is not seemingly interested in learning at all, a point to which we return in greater detail below. Instead, its remit is exclusively on the implicitly acquired system itself, its formal representations and how acquisition/processing of non-native language obtains. Generative approaches largely subscribe to a crucial distinction between acquisition and learning. The former relates to the passive act of forming linguistic representations on the basis of input/intake in conjunction with whatever domaingeneral and domain-specific mechanisms one has at her disposal. Alternatively, learning refers to the (typically conscious) act of obtaining a skill set from explicit effort and overt, purposeful teaching. While there might be specific talents for nonnative language learning in the form of language aptitude (see Wen et al. 2017 for review) and differences in language attitudes/motivation (see Dörnyei 1998, 2003; Dörnyei et al. 2016 for review), the general idea is that language learning in a classroom setting, as opposed to acquiring a language in a naturalistic environment, is not so different from the learning of other skill sets (DeKeyser 2007a, b, 2018). Theoretical paradigms differ considerably with respect to what cognitive/linguistic mechanisms they accept as contributors to the acquisition process (see Synder 2007; O’Grady 2005; Clark 2009; Ambridge and Lieven 2011; Guasti 2017). However, they all conclude that language acquisition in children is a virtually completely involuntary, unconscious activity. While the same paradigmatic splits exist for bilingualism and although there are active discussions even within a single paradigm regarding potential mechanistic differences between children and adult language learners, researchers focusing on adult bilingual language acquisition tend to agree that much of what adult bilinguals come to know is acquired implicitly (see VanPatten et al. 2020). Although passive acquisition can and does happen in parallel to learning in a second language classroom setting at all ages, the context of explicit instruction as the primary source of input in classroom language learners is quite distinct from the task of native language acquisition. Similarly, compared to the task of second language teaching and learning, native language teaching and its grammar instruction at school (e.g., German language classes for native speakers of German, English for native speakers of English) has its own constraints (see Turgay and Gutzmann; Döring; Elsner in this volume). Since language acquisition and processing theories are largely concerned with acquisition and language pedagogy with learning, this alone might explain the dearth of connections between the two fields. Within generative acquisition, it was (is) largely held true that there is no interface between explicitly learned and implicitly acquired knowledge. And so, rules that are explicitly learned would not (could not) become part of the underlying acquired grammar, but would remain as conscious metalinguistic knowledge, constituting a competitive system of its own (see Felix 1985; Rothman 2008; Long and Rothman 2013 for discussion). To the extent that explicit knowledge of rules—grammar rules taught in a classroom— coincides with an acquired/implicit mental representation of the same rules, it is difficult to tease the two apart (where they do not overlap is precisely where you can test and appreciate the distinction). Nevertheless, their qualitative natures are distinct. This distinction (implicit vs explicit knowledge) is reminiscent of Ullman’s

14

F. Bayram and J. Rothman

(2001, 2016) distinction between declarative and procedural memory in language acquisition and the related question many have asked whether or not some linguistic structure that has been acquired via the declarative system can become procedural. Ullman has never denied that declarative knowledge can be automatized over time, making it more difficult to distinguish at the surface level between what is acquired via the procedural system and what is automatized via declarative knowledge. Automatized knowledge acquired through the declarative system cannot suddenly be considered procedural knowledge because it remains qualitatively different by virtue of the memory system through which it was originally acquired. If indeed there is no interface between acquisition and learning, then it stands to reason that theorists and practitioners alike would be less inclined to dialogue, much less see one another’s research as particularly relevant and useful. In 2019, however, this no interface position in generative approaches to non-native acquisition is not as steadfast as it was in prior decades, at least for some scholars. It might be the case that explicit instruction can have effects in mental representations. This can tell us many things about formal theory itself. In line with the above, some current generative L2 proposals that highlight and account for the difficulty adult learners have with inflectional (functional) morphology are particularly relevant. The Bottleneck Hypothesis (Slabakova 2008) and the Feature Reassembly Hypothesis (Lardiere 2009) each claim that morphology is particularly challenging, in fact, the very ‘bottleneck’ of acquisition for adult learners. The Feature Reassembly Hypothesis essentially claims that adults can acquire new L2 features without major problems. The challenge is the re-distribution of L1 features, particularly, how they bundle together, onto new morphological exponents. The Bottleneck Hypothesis further claims that structured input or, potentially even explicit knowledge, related to how features bundle together on target language morphological units could help learners to generate the parsing failures required to promote better, more efficient reassembling of features. The next step is to translate these linguistic formalisms and predictions for different L2 learner contexts, with their L1s in mind, into targeting the learning challenges by designing interventions that lead the learner’s mind to establish form-meaning mappings that assign newly acquired L2 features, and crucially, reassign the L1 features appropriately to relevant L2 morphology. This all points in the direction that structured input, as argued for decades in the Input Processing literature (see above) is a key factor for improving learning outcomes in the typical L2 classroom. Although some erroneously believe that generative linguistics de-values the role of input in the acquisition process, recent proposals highlight just how incorrect this is. Input is the core building unit of specific grammars (e.g., Yang 2016), especially in light of the concession that parametric settings themselves would not be part of the genetic endowment of Universal Grammar (Rothman and Chomsky 2018). It seems almost certain that imparting what decades of research on how nonnative grammars are acquired and represented, often unintuitively so, coupled with infusing linguistic descriptions to the baseline target grammars used in language teaching has the potential to improve the quality of language teaching. If the cognitive and linguistic study of language acquisition reveals seemingly hardwired

Formal Linguistics and Language Education: A View from Bilingualism Research

15

constraints or adaptive, dynamic ones to acquisition, it only stands to reason that pedagogical approaches working with and not, inadvertently, against such constraints will be more successful for more individuals. If a pedagogical grammar unwittingly overgeneralizes or underestimates some properties of language described, either knowingly for ease of teaching or simply because it is not as linguistically informed as it could be, there is a risk that such does more harm than good in the long run. Minimally, inaccurately learned knowledge will need to be unlearned down the road, fostering a context for pedagogically induced, lingering variation, as described by the Competing Systems Hypothesis (Rothman 2008) in the existing literature, as well as in Snape’s and Rankin & Whong’s chapters in the present volume, which look at learnability issues in second language acquisition, in Agebjörn’s contribution that focuses on the use of explicit knowledge in developing implicit knowledge of certain grammatical structures, and in Stadt, Hulk and Sleeman’s article that investigates the role previously learnt languages play in third language acquisition. Indeed, language teaching necessarily makes explicit that which is inherently implicit under child language acquisition scenarios. Language teaching must try to be better informed about language acquisition in the “wild” because it is itself a gap filler. Its utility is predicated on the fact that creating the same or comparable conditions to native, naturalistic language acquisition is virtually impossible on several planes. It makes sense that accuracy in linguistic description is paramount at the outset of curriculum materials and teaching methods development as well as teacher training. While we acknowledge that linguistic descriptions of grammatical structures as they are observed in the naturalistic “wild” and in the mental representation do not necessarily overlap with the grammar “rules” that are described and presented in language teaching materials, the existence of this mismatch underscores a missed opportunity where educational linguists should be working with formal linguists. In principle, theoretical linguists hold some useful keys for more efficient curriculum design; however, they lack the expertise to implement them, expertise which language pedagogy theorists have. Working together to translate such knowledge to two sets of audiences, teachers and students, in an appropriate way for them, the language teaching scholar, the formal acquisitionist, and the formal theorist can accomplish an important task together. Of course, work focusing on high-quality language teaching with the aim of building some bridges exists, especially by scholars such as Patsy Lightbown, Nina Spada, Sue Gass, Pilar Garcia Mayo and Bill VanPatten, to name a few. Although their work is linguistically informed and, thus, has a trickle-down effect towards what we are advocating, they have generally not attempted to explicitly link formal linguistics to teaching practice in the way we are discussing herein. A notable exception worthy of discussion is VanPatten’s Input Processing Theory and its corollary Processing Instruction approach to teaching that, in principle, incorporates insights from the former (VanPatten 1996, 2002a, b). Input Processing Theory highlights the fact that, irrespective of intension and often unintuitively, nonnative adult bilinguals process linguistic input via a series of hardwired constraints. Depending on the target language, these default processing strategies are, a priori,

16

F. Bayram and J. Rothman

more or less helpful. For example, one such constraint has been labeled the First Noun Principle which captures the default tendency for adult learners to process the first noun in a sentence as the subject, potentially irrespective of whether that is what is done in their native L1 or not. If one is learning English, this is a pretty good strategy given that English does not drop subjects like many languages and has a pretty robust, S(ubject), V(erb), O(bject) word order. However, if one is learning a language where subjects are often dropped, like Spanish, this is a less optimal strategy. Given that subjects are often dropped, the first noun encountered in Spanish is likely to be the object of the verb and not its subject. In parallel to Input Processing Theory and the underlying empirical research revealing the processing principles, VanPatten developed Processing Instruction, which asks and answers the key question “what do we do with insights from formal acquisition/processing theory?”. Details aside, Processing Instruction provides a rubric of intervention by which, via modified input and specifically designed tasks, the parser can be tricked into failures much sooner than would otherwise happen. The theory uncovers the bottlenecks in the acquisition process and the pedagogical intervention on the coattails of formal research that provides a way around them. VanPatten’s theories have spawned literally hundreds of studies. Although, unfortunately, some do not represent faithfully his original intension or are faithful to the bridging of theory to practice, those that do exemplify the possibility and the benefit of increasingly formally inspired pedagogical interventions. It is important to make it clear that we do not think that this is a direct link between teaching X and its potential outcome Y, but we do see how the outcome Y can be facilitated indirectly by insights from formal linguistic and processing research. Explicit grammar instruction, increasingly accurate descriptions of grammar and/or task development that works with our natural predispositions for language processing could help the learner get more input and make better use of what is available in the input so that the internal mechanisms get higher quantities and better qualities of proverbial gasoline to most efficiently perform the tasks of acquiring a new language system. Therefore, while we do not believe that conscious knowledge of morphological paradigms or drilling them is necessary or sufficient for language acquisition, such knowledge, especially in a classroom context, might facilitate processing and maximize the allocation of limited resources, optimizing opportunities for acquisition in the true sense of the word to take place. These are, of course, empirical questions that are being tested. Because of the promising results (see, e.g., Dudley and Slabakova, this volume) and the renewed enthusiasm within generative approaches to bilingual acquisition to make itself more meaningful to language teaching where possible, the timeliness of the present volume is a huge step in the right direction.

Formal Linguistics and Language Education: A View from Bilingualism Research

17

3 Final Words Before enduring and truly meaningful bridges can be formed, however, it is fair, if not prudent, to ask if bilingual language teaching can really benefit from formal linguistic theories and research. If so, then after revealing and defending why we must take steps to show how there is benefit, ideally mutual benefit, from building such bridges. We take it as a given that there should be an inherent connection between the scientific study of language—linguistics and psycholinguistics proper, irrespective of specific formalism—and pedagogical practice in the domain of teaching. While we strongly believe that language teaching should be informed by what we know regarding the mental constitution of language as well as constraints on how language is (likely) to be acquired, processed, stored and accessed in the mind/brain, bridges built are destined to be untraveled if we do not project beyond the obvious connections and benefits by making practical and usable insights. What we have provided herein is merely a taster of what has been done and what can be done. Recent venues such as the Educational Linguistics book series by Springer in which this volume will appear as well as the newly launched journal Pedagogical Linguistics published by John Benjamins are excellent venues attempting to fill the gap of much needed venues for achieving the goal of interdisciplinary work between formal linguistics and language pedagogy. The next years will prove the extent to which we can, and indeed desire, to take these insights and others various scholars have been discussing for some time now to the next levels. Information from formal linguistics and acquisition is necessary, but, we submit, not sufficient to accomplish the goals at hand. Knowing that, in principle, there can and should be more linguistically-informed pedagogical interventions designed to circumvent the pitfalls for adult language learners that are consistently observed and for which, at present (and any given time in the future), linguistic theory can provide some insights into is also not sufficient. Partnerships with pedagogical theorist to carve out practical ways to implement these insights is necessary. Combining the expertise on the ground that pedagogical theorist and educational linguistics have with the expertise of acquisitionist and other cognitive scientists can yield meaningful and enduring improvements to outcomes. As a final aside, it might be useful to consider some other related research on neurocognition as well as multilingualism in general. More than half of the world’s population is bilingual (Bialystok et al. 2012), with 54% of Europeans reported to be at least bilingual (19% bilingual, 25% trilingual and 10% four or more languages (Eurobarometer 2012) and 21.3% of the US population being bilingual (American Community Survey 2017). This means that an increasing number of children around the world grow up as bi/multilingual and enter the school system with linguistic knowledge/competences varying from one of their languages to the other(s). Work on additive multilingual acquisition and processing is very clear: all previous linguistic knowledge matters (see Rothman et al. 2019, for review). Successive L3 and L2 acquisition/processing is not the same (see Stadt, Hulk & Sleeman, in this volume). What does this mean for the theory and

18

F. Bayram and J. Rothman

practice of language teaching, which is traditionally fed by methods of teaching English as a second/foreign language to groups of (predominantly) monolingual speakers? Should this linguistically more diverse and cognitively more resourceful learner profile in the classroom play any role in the way teaching practitioners, policymakers, materials/curriculum developers and educationalists make decisions? How will future bridges to offer more linguistically-informed pedagogy deal with this reality? Bilingualism affects the mind/brain (Bialystok 2016; Bialystok et al. 2012) with the most contempory evidence suggesting that anatomical and cognitive adaptations in the bilingual brain are ameliorated by language experiences such as domains and frequency of use (DeLuca et al. 2019). What is the relevance of this literature for language teaching and how will this literature (or should it) also contribute bridges to language pedagogical theory?

References Ambridge, B., & Lieven, E. V. (2011). Child language acquisition: Contrasting theoretical approaches. Cambridge: Cambridge University Press. Bialystok, E. (2016). The signal and the noise. Linguistic Approaches to Bilingualism, 6(5), 517– 534. Bialystok, E., Craik, F. I., & Luk, G. (2012). Bilingualism: Consequences for mind and brain. Trends in Cognitive Sciences, 16(4), 240–250. Clark, E. V. (2009). First language acquisition. Cambridge: Cambridge University Press. DeKeyser, R. M. (Ed.). (2007a). Practice in a second language: Perspectives from applied linguistics and cognitive psychology. Cambridge: Cambridge University Press. DeKeyser, R. M. (2007b). Skills acquisition theory. In B. VanPatten & J. Williams (Eds.), Theories in second language acquisition (pp. 97–114). Mahwah: Erlbaum. DeKeyser, R. M. (2018). Age in learning and teaching grammar. In J. I. Lointas (Ed.), The TESOL encyclopedia of english language teaching (pp. 1–6). Hoboken: Wiley. DeLuca, V., Rothman, J., Bialystok, E., & Pliatsikas, C. (2019). Redefining bilingualism as a spectrum of experiences that differentially affects brain structure and function. Proceedings of the National Academy of Sciences, 116(15), 7565–7574. Dörnyei, Z. (1998). Motivation in second and foreign language learning. Language Teaching, 31(3), 117–135. Dörnyei, Z. (2003). Attitudes, orientations, and motivations in language learning: Advances in theory, research, and applications. Language Learning, 53(S1), 3–32. Dörnyei, Z., Henry, A., & Muir, C. (2016). Motivational currents in language learning: Frameworks for focused interventions. New York: Routledge. Eurobarometer. (2012, November). Public opinion in the European Union (First results from Eurobarometer 78). Brussels: European Commission. Felix, S. (1985). UG-generated knowledge in adult second language acquisition. In S. Flynn & W. O’Neill (Eds.), Linguistic theory in second language acquisition. Dordrecht: Reidel. Goldberg, A. E. (2019). Explain me this: Creativity, competition, and the partial productivity of constructions. Princeton: Princeton University Press. Guasti, M. T. (2017). Language acquisition: The growth of grammar. Cambridge, MA: The MIT press. Jackendoff, R. (2002). Foundations of language: Brain, meaning, grammar, evolution. Oxford: Oxford University Press.

Formal Linguistics and Language Education: A View from Bilingualism Research

19

Lardiere, D. (2009). Some thoughts on the contrastive analysis of features in second language acquisition. Second Language Research, 25(2), 173–227. Long, D., & Rothman, J. (2013). Generative approaches and the competing systems hypothesis. In J. W. Schwieter (Ed.), Innovative research and practices in second language acquisition and bilingualism (pp. 63–82). Amsterdam: John Benjamins Publishing. Marsden, H., & Slabakova, R. (2019). Grammatical meaning and the second language classroom: Introduction. Language Teaching Research, 23(2), 147–157. O’Grady, W. (2005). Syntactic carpentry: An emergentist approach to syntax. Mahwah: Erlbaum Print. Rothman, J. (2008). Aspect selection in adult L2 Spanish and the competing systems hypothesis: When pedagogical and linguistic rules conflict. Languages in Contrast, 8(1), 74–106. Rothman, J., & Chomsky, N. (2018). Towards eliminating arbitrary stipulations related to parameters: Linguistic innateness and the variational model. Linguistic Approaches to Bilingualism, 8(6), 764–769. Rothman, J., González Alonso, J., & Puig-Mayenco, E. (2019). Third language acquisition and linguistic transfer. Cambridge: Cambridge University Press. Slabakova, R. (2008). Meaning in the second language. Berlin: Walter de Gruyter. Synder, W. (2007). Child language: The parametric approach. New York: Oxford University Press. Print. U.S. Census Bureau. (2017). American community survey. Washington, DC: U.S. Census Bureau. 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(2), 105–122. Ullman, M. T. (2016). The declarative/procedural model: A neurobiological model of language learning, knowledge, and use. In G. Hickok & S. L. Small (Eds.), Neurobiology of language (pp. 953–968). San Diego: Elsevier. VanPatten, B. (1996). Input processing and grammar instruction in second language acquisition. Norwood: Ablex. VanPatten, B. (2002a). Processing instruction: An update. Language Learning, 52(4), 755–803. VanPatten, B. (2002b). Communicative classrooms, processing instruction, and pedagogical norms. In M. S. Gass, K. Bardovi-Harlig, S. Magnan & J. Walz (Eds.), Pedagogical norms for second and foreign language learning and teaching (pp. 105–118). Amsterdam: John Benjamins. VanPatten, B., & Rothman, J. (2015). What does current generative theory have to say about the explicit-implicit debate. In P. Rebuschat (Ed.), Explicit and implicit learning of languages (pp. 91–116). Amsterdam: John Benjamins. VanPatten, B., Keating, G. D., & Wulff, S. (Eds.). (2020). Theories in second language acquisition: An introduction. Routledge. Wen, Z. E., Biedro´n, A., & Skehan, P. (2017). Foreign language aptitude theory: Yesterday, today and tomorrow. Language Teaching, 50(1), 1–31. White, L., Spada, N., Lightbown, P. M., & Ranta, L. (1991). Input enhancement and L2 question formation. Applied Linguistics, 12(4), 416–432. Whong, M. (2011). Language teaching: Linguistic theory in practice. Edinburgh: Edinburgh University Press. Whong, M., Gil, K. H., & Marsden, H. (2014). Beyond paradigm: The ‘what’ and the ‘how’ of classroom research. Second Language Research, 30(4), 551–568. Yang, C. (2016). The price of linguistic productivity: How children learn to break the rules of language. Cambridge, MA: The MIT Press.

Grammatical Concepts for Pedagogical Grammar Tom Rankin and Melinda Whong

Abstract This paper develops an approach to pedagogical grammar based on the notion of grammatical concepts. Grammatical concepts are based on the sort of properties that are used to explain grammatical patterns and the acquisition of grammar in formal linguistics. It is proposed that these properties can be exploited for language pedagogy as they provide teachers with a deeper understanding of issues of learnability with respect to grammar. The concepts are illustrated on the basis of the pronoun system of English, demonstrating how a range of different underlying grammatical concepts can coincide to regulate syntactic and semantic patterns in a particular morphosyntactic paradigm. Keywords Pedagogical grammar · Formal linguistics · Grammar teaching · Language awareness · Knowledge about language

1 Introduction With the existence of ‘grammar books’ and the teaching of ‘grammar rules’, there is a notion of grammar as something that is fixed. Linguists of all theoretical persuasions will know that it is not, of course. One useful way to influence language pedagogy to more contemporary understanding of grammar is through the consideration of grammatical concepts. The notion of grammatical concepts can be understood with respect to the ability to analyse language and correctly to apply grammatical terminology in that analysis. On this understanding, knowing about grammatical concepts means one can differentiate between categories (noun phrases

T. Rankin () School of Education, Johannes-Kepler University, Linz, Austria e-mail: [email protected] M. Whong Center for Language Education, The Hong Kong University of Science and Technology, Kowloon, Hong Kong SAR e-mail: [email protected] © Springer Nature Switzerland AG 2020 A. Trotzke, T. Kupisch (eds.), Formal Linguistics and Language Education, Educational Linguistics 43, https://doi.org/10.1007/978-3-030-39257-4_3

21

22

T. Rankin and M. Whong

vs. verb phrases, subjects, predicates, etc.), and one can identify and describe elements of morphosyntactic paradigms (tense forms, case-marking, etc.). This conceptualisation also challenges the “static nature of grammar pedagogy” and relies mainly on traditional descriptive notions of grammar which trace their origins to the late nineteenth century (see Van Rijt and Coppen 2017: 364). If pedagogical grammar has remained relatively staid and not developed in line with advances in understanding of grammar in linguistics, then linguists have to bear responsibility for this state of affairs. With some notable exceptions (e.g. Hudson 2004, 2008), theoretical, formally-oriented linguistics and language acquisition research has typically shied away from formulating ideas for education. In addition, terminological complexities and theory-internal assumptions act as barriers to easy translation from academic findings to educational practice. Recently, however, there has been increasing interest in potential applications of findings from formal approaches to second language acquisition research (SLA) to questions of classroom practice (e.g. Whong et al. 2013; Marsden and Slabakova 2018). In this contribution, we draw on ideas from generative grammar and findings from generative approaches to SLA to re-examine and rebrand the notion of a grammatical concept as a constructive way to apply findings from formal linguistics to language teaching. We will propose grammatical concepts based on universal formal and semantic distinctions in language which can potentially be used to illuminate learnability issues in any specific target language, as well as contributing to the broader educational goals of language learning. The aim is to mediate theoretical notions that underpin findings in SLA in such a way as to make them potentially pedagogically useful. As such, we also explore the process of applied linguistic mediation itself. We proceed as follows. First, we provide a very brief summary of some fundamental aspects of generative SLA and in doing so develop the idea of what a grammatical concept is. Then we explore how linguistics interacts (or not) with questions of pedagogical applications and conclude that the notion of grammatical concepts is a potentially useful contribution. Finally, we outline some principles for the pedagogic use of grammatical concepts, before concluding with a more detailed illustration of the ideas on the basis of pronominal paradigms and usage.

2 L2 Development and Grammatical Concepts The approach to L2 development which inspires this grammatical concepts approach is based on Universal Grammar (UG) (see White 2006 for introduction to UG-based SLA). The technical formulation of the syntactic analyses in the UG framework continually develop, and on-going research refines and rethinks the precise linguistic properties which may be part of UG or may be emergent based on language input and processing constraints. It would therefore be unhelpful to base any formulations for pedagogical grammar too closely on the specific details of current formulations of any syntactic theory. Instead, the process of mediation

Grammatical Concepts for Pedagogical Grammar

23

requires consideration of the generally agreed upon basic elements of the theoretical framework so that these can be distilled in ways that are pedagogically relevant. The resulting presentation of pedagogical features is then not beholden to specific instantiations of theory, which are liable to change. Models of SLA within a broad UG framework offer a range of proposals which may make different assumptions about the nature and role of particular variables in the L2 acquisition process. For example, major debates have focussed on the extent to which grammatical representations at the initial state of L2 acquisition are transferred from the native language, and whether properties of UG remain available in late L2 acquisition after a critical period (see White 2006: Ch. 3). Details notwithstanding, the consensus is that L1 grammatical representations play some major role in the acquisition of a second language and that UG continues to constrain L2 grammar, at least to the extent that learners are never found to develop ‘wild grammars’ which fall outside the bounds of what UG licenses. During the course of development, L2 performance may resemble neither features of the L1 nor the L2, but it will be a result of some combination of UG-licensed features. Readers familiar with generative SLA will recognise this characterisation as more or less Schwartz and Sprouse’s (1996) Full Transfer/Full Access model. But the overall thrust of the ideas is more important than the way they are expressed in particular theoretical models. The major points are the implication of L1 properties and the potential creative restructuring of those L1 properties in response to L2 input. A crux of the difficulty in the restructuring of L2 grammar is the complexity involved in the interaction of existing linguistic knowledge, properties of the input and the expression of underlying morphosyntactic and semantic features in the surface morphological and lexical forms of the target language. Given the influence of L1 grammar, this mapping problem between the underlying features and their overt expression is especially intricate where the native and target languages have different ways of realising mappings between semantic or morphosyntactic features and their overt expression. Readers familiar with generative SLA will also recognise this depiction as deriving from Lardiere’s (2008, 2009) Feature Reassembly Hypothesis. Again, the details of a model formulated to address particular empirical and theoretical questions in SLA need not be discussed extensively. Only some fundamental assumptions need to be explicated to motivate the nature of grammatical concepts as they will be formulated later. The surface morphological and lexical forms of languages are built from formal, semantic and phonological features. The formal features dictate the morphosyntactic distribution of lexical items; semantic features mark interpretable properties with some semantic import; the phonological features constrain their phonological realisation. L2 learners must “reconfigure or remap features from the way these are represented in the L1 into new formal configurations on possibly quite different types of lexical items in the L2. This is a formidable learning task that goes far beyond the simple ‘switch-setting’ or ‘selecting’ metaphors that have sometimes been used to characterize the acquisition of a L2 grammar” (Lardiere 2009: 175). Even though semantically-relevant distinctions might be universal across all languages, their instantiation, i.e. the way that features are assembled, in the

24

T. Rankin and M. Whong

functional and lexical categories of individual languages can differ widely, resulting in learnability issues due to L1 influence in addition to inherent complexity in semantic-syntactic interfaces. An example from Choi and Lardiere (2006) serves to illustrate the issue on the basis of the L2 acquisition of Korean by English speakers. Korean has lexical items which seem on the surface comparable to English wh-words: they may be offered as translations of wh-words and are also implicated in questions. However, there are different meaning associations and syntactic distributions. Without going into major technical detail, wh-words in English are assembled out of a wh-operator and an interrogative feature [+Q] so that they refer to or question the existence of some particular referent, as in the translation in (1). Korean exploits the same features but distributes them differently. The interrogative feature is expressed as a morphological ending on verbs so that a declarative morpheme includes [−Q] and an interrogative morpheme is [+Q]. The examples in (1) and (2) illustrate. (1) John-un Mary-ka mues-ul sassnun-ci an-ta. John.TOP Mary.NOM thing.ACC bought.Q know.DECL ‘John knows what Mary bought.’ (2) John-un Mary-ka mues-ul sass-ta-ko an-ta. John.TOP Mary.NOM thing.ACC bought.DECL.C know. DECL ‘John knows that Mary bought something.’ This gives rise to a learnability issue. An English-speaking learner of Korean whose grammar represents Korean mues like English what will not detect a change in meaning between these Korean sentences as, for the learner, the [+Q] feature is an intrinsic part of the lexical item. Reassembling the features so that interrogative force is represented by a verb ending rather than via lexical and word order changes is difficult. In experiments, this is precisely what Choi and Lardiere (2006) found: intermediate proficiency L1 English-L2 Korean learners did not reliably detect the meaning difference in these types of Korean sentence and treated them both as meaning ‘ . . . know what Mary bought’. What sort of grammatical concepts could be leveraged to develop a pedagogical grammar treatment of such issues? One could of course concentrate on traditional paradigms by exploring Korean verb endings and situating the interrogative/declarative meaning distinction in the wider paradigm of verbal morphology and its various functions. One could concentrate on the paradigm of wh-words in Korean and the meanings and functions associated with them. Taking a grammatical concept view, by contrast, would involve the use of something conceptual as the starting point and situating a range of potential surface forms with respect to the underlying conceptual-semantic notion, rather than taking the surface patterns as the organising principle for pedagogical grammar. This idea will be fleshed out below. As a foretaste, in the context of the English-Korean example we have been looking at, the distinctions between the sorts of sentences illustrated in (1) and (2) are one manifestation of broader conceptual-semantic distinctions between existential interpretation and other meaning-interpretations associated with different classes of words or

Grammatical Concepts for Pedagogical Grammar

25

grammatical paradigms (related concepts would be interrogative, specific, universal interpretations etc.). With respect to Korean and English, we see that the existential concept underlies a range of linguistic patterns that are variably realised in different paradigms. Existential force may be carried by pronouns and quantifiers or other grammatical constructions in English, for example the distinction between realisations of some and any or the use of existential there. The point is that there is nothing inherently incomprehensible about existential meaning in itself. It is just that Korean makes multi-functional use of wh-expressions such that they may optionally express existential force depending on grammatical context, as in (3) to (5) (from Gil and Marsden 2013: 124). As indicated by the translations, the expression of existential force by the choice of some vs. any is also intricately grammatically constrained in English (see Gil and Marsden 2013). (3) Nwu-ka cha-lul masiko iss-nayo? who.NOM tea.ACC drink PROG.Q ‘Is anyone/someone drinking tea?’/‘Who is drinking tea?’ (4) Nwu-ka sen-ul nemti-myen, kispal-ul tul-era. who.NOM line.ACC cross.COND flag.ACC raise.IMPER ‘If anyone/someone crosses the line, raise the flag.’ (5) Nwu-ka cha-lul masiko isseyo. who.NOM tea.ACC drink PROG ‘Someone (*anyone) is drinking tea.’ Taking a shared underlying conceptual foundation as the starting point for a pedagogical description of grammar would allow us to consider a range of grammatical and lexical properties which are linguistically related, but which have traditionally been presented more disjointedly in pedagogically oriented reference grammars because of a lack of recognition of the grammatically relevant concepts like existential in traditional grammars. In this way, linguistic commonality can be emphasised. In this example, learners can be guided to the existence of existential in language, including their own; the foreignness of a foreign language is then relativized by making existing knowledge from the L1 explicit and linking this to new instantiations of the same concept in the L2. The logic of basing pedagogical grammar on grammatical concepts is that they will help learners when they are confronted with new, apparently idiosyncratic grammatical phenomena in the L2, drawing more heavily on meaning instead of the more common approach of paradigms, form and pattern matching. At this stage, there are a number of open questions. Perhaps most pressing is the issue of whether this sort of pedagogical grammar would actually promote learning. The idea is derived from what we know about the nature of learning difficulties from generative L2 acquisition studies, but there is of course the difficult issue of the (non-)interface between implicit and explicit knowledge. Just because we raise learners’ awareness for a learnability issue does not mean that such explicit knowledge will (or can) be transformed into implicit grammatical competence in the target language. Consideration of issues connected with implicit versus explicit knowledge and related questions is far beyond the scope of the present contribution.

26

T. Rankin and M. Whong

In lieu of this, in the following, we offer a rationale for grammatical concepts as a way of addressing the pedagogical motivation behind the idea. We then discuss practical issues related to the definition of grammatical concepts before turning to a fuller illustration.

3 Rationale for Grammatical Concepts It is worth emphasising at this point that we are not claiming that grammatical concepts represent a framework for doing something with grammar in the sense of grammar tasks, nor a method for delivering instruction. The extent to which this might be possible and/or useful is left for colleagues with expertise in teaching. The focus is instead on the content of pedagogical grammar materials and the sort of information that could be presented there, with the aim that this content can potentially contribute to a useful form of linguistic awareness. The assumption is that language acquisition researchers and linguists can provide useful grammatical content. It is for teachers to choose how and when to use such content in the delivery of instruction. In addition to recognising the need for expertise in teaching methodology, such questions are circumscribed by a wide range of additional factors such as curriculum constraints and educational culture.

3.1 Awareness, Language Learning and Language Education It is possible to build a clear educational rationale for developing grammatical concepts as a way to raise language awareness. One aspect of this rests on a distinction between language learning and language education (Kramsch 2008: 5–7). Language learning involves proficiency development and training for transactional or professional contexts while language education is “inscribed within the schooling career of adolescents and young adults” (Kramsch 2008: 6). Rather than solely facilitating communicative competence or foreign language proficiency, language education seeks to provide “an aesthetic education that links taste and moral value and expands learners’ notion of the good and the beautiful” (Kramsch 2008: 7). Seen from this perspective, one can side-step the issue of implicit linguistic knowledge being the sole aim of teaching. Foreign language instruction is not designed solely with the aim of producing proficient users of the target languages. As a subject on school and university curricula, languages are also a vehicle for addressing wider educational and inter-cultural concerns. This is apparent in language education policy, which may increasingly mandate ideas of plurilingualism or translanguaging, and similar concepts, which do not see as their goal mastery of successive non-native languages. This is reflected, for example, in calls to “decompartmentalise language education” (Beacco and

Grammatical Concepts for Pedagogical Grammar

27

Byram 2007: 86).1 Such an approach might involve, among other things, including aspects of language awareness in the teaching of all languages “to show what is common to the functioning of all natural languages” (Beacco and Byram 2007: 86). A grammatical concepts approach represents a way of organising the content of pedagogical grammar that addresses this educational concern. Content which emphasises the underlying formal and semantic properties which are universal to any languages readily facilitates connections between different languages and highlights aspects of awareness which would be transferable to the learning of any language. The overarching theme emerging from this discussion is one of language awareness, which highlights explicit knowledge about what language is.2 Again, we are confronted with the implicit/explicit tension. In educational terms, this can be seen in the trend away from formal understandings of grammar in language teaching in the modern era, to prioritise communicative competence (Hymes 1972; Canale and Swain 1980) and social turn agendas (Block 2003). This development is an understandable reaction against a perceived excessive earlier pedagogical concentration on explicit focus on forms and the learning of grammatical rules for their own sake, at the expense of communicative ability. However, it would be wrong to assume that these notions are diametrically opposed. Within a policy context which seeks to promote language learning and language education, and in which multilingualism is a key factor, using formal grammatical ideas can address the educational imperative of raising conscious awareness of linguistic properties. However, it can also be leveraged for learning of linguistic features as we know that it is the mapping between surface forms and the underlying formal and semantic notions they encode which poses acquisition issues for learners. We can see this combined thinking embracing both ‘grammar’ and ‘communication’ in previous language pedagogy work. As expressed in one well known language teaching methodology text: “Grammar exists to ‘mean’ and without grammar it is impossible to communicate beyond a very rudimentary level” (Nunan 1995: 153). Despite the fact that language education policy and conceptual frameworks like Language Awareness would seem to make input from formal linguistics and

1 We

draw on European Union (EU) language education policies as an instance of a welldocumented, readily accessible approach. Policies and recommendations in other jurisdictions will obviously vary. It is hoped that readers can relate the European instantiation to local concerns, which are likely to be variations on similar societal themes connected to multilingualism, linguistic diversity, (super)diversity, etc. (see Blommaert and Rampton 2011). 2 Language awareness is a rather nebulous concept that is used in different ways by different scholars. It may encompass approaches which are diametrically opposed to one another, ideologically and methodologically (Van Lier 1996: 81). Our working definition comes from the Association of Language Awareness (2018) (emphasis in the original): “We define Language Awareness as explicit knowledge about language, and conscious perception and sensitivity in language learning, language teaching and language use. It covers a wide spectrum of fields. For example, Language Awareness issues include exploring the benefits that can be derived from developing a good knowledge about language, a conscious understanding of how languages work, of how people learn them and use them.”

28

T. Rankin and M. Whong

language acquisition a welcome contribution, as we noted in the introduction, the relationship between linguistics and language acquisition on the one hand and language pedagogy research and applied linguistics on the other has not been an easy one. This is not helped by fundamental differences between different schools of linguistics. Language can legitimately be viewed as a tool for communication, with grammatical properties emerging from communicative constraints. It can be seen as an abstract formal system. It may be viewed primarily as a cognitive phenomenon represented in the minds of its users, or a social phenomenon whose defining characteristic is variable use for interaction in different contexts. These are genuine points of difference in academic explanation of linguistic phenomena, and such fundamental differences complicate the linguist’s task in articulating educationally relevant insights. In seeking to develop ideas for pedagogical grammar founded in formal linguistics, the aim is not to privilege this as the only legitimate approach in opposition to other linguistic theories. If an educational aim is to foster knowledge of the nature of language, then this must necessarily encompass its multi-faceted nature as a formal system and a functional tool, a cognitive system and a medium for social interaction, and much more besides. Brumfit (2001: 17) observes: Language is the cheapest scientific data available to schools, and pupils [ . . . ] invariably enjoy thinking about it, as a socially significant system and as an abstract system alike, if the teacher is committed, knowledgeable and enthusiastic.

The important practical point is that each is a valid point of view on language and there is value in being able to think about language from each perspective, especially in the context of the language classroom, where one perspective will sometimes be more useful than another for particular pedagogical and educational goals. To be clear, we are not advocating a return to grammar-translation as a didactic tool for the teaching for foreign languages. Nor are we advocating Focus on Form (Long 1991) necessarily, but rather an informed consideration of formal linguistic properties as an integral element of language education in a way that integrates developments in linguistics. Our observation is that the substantial body of literature on Focus on Form, for example, assumes that the ‘forms’ in question are fixed. In the same way that an earlier excessive focus on formal properties of language led to an imbalanced and skewed language education, insufficient attention to formal linguistics does a disservice to teachers and learners alike. In short, we argue for including the formal nature of language as an element of language education and teacher training, but suggest this needs to be done via an approach which is accessible and useful for language teachers.

3.2 Deriving Grammatical Concepts from Formal Grammar As the preceding discussion has highlighted, there is no direct path from any linguistic or language acquisition theory to applications in pedagogy. Widdowson (2000a, b, 2003) has emphasised the need for mediation between linguistics

Grammatical Concepts for Pedagogical Grammar

29

and pedagogy in order to construct applied linguistics models. He summarises (Widdowson 2003: 10–11): it is not within the brief of linguists to make useful theories. On the contrary, as soon as they start doing that, they lose their scholarly independence and with it their value to the non-scholarly world. This value depends not on making useful theories, but making theories useful.

Knowledge and concepts that form the basis of formal language acquisition theories therefore have to be adapted and mediated to make them amenable to different uses by a different audience. However, Widdowson (2003: 11) goes on to say that linguists themselves cannot judge what use might be made of linguistic theory and that this is for others to decide. We respectfully disagree. It is surely valuable for linguists and acquisition researchers to attempt to make their theories useful. It seems strange to presuppose that it is impossible for theorists or specialists (in linguistics or any other field) to step outside their own framework, reflect on it and attempt to explore its pedagogical or societal relevance. Perhaps ironically given this discussion, one source of inspiration for the notion of grammatical concepts comes from Widdowson’s (2016) own discussion of “virtual language”, which he defines as follows: “It is not a system of actual encodings: it is a generative encoding potential whose properties can only be inferred from its variable use” (p. 33, our emphasis). His argument is that knowing something about the encoding potential would be helpful for learners in getting to grips with the learnability issues involved in actual variable occurrence of the overt forms in use. The affinity to the language acquisition models previously discussed are not difficult to discern: UG makes all potential grammatical properties available and the same universal meaning distinctions are available to learners (encoding potential), but establishing how these abstract features are bundled in morphological and lexical surface forms poses learnability issues (actual encodings). The practical task remains to define those grammatical concepts (or virtual encoding potential in Widdowson’s terms) which underpin grammar, and which can help learners to make sense of grammatical complexity in a way that is both useful, and in line with contemporary understanding of how language works. An exhaustive explanation is not realistic here. The following will be a programmatic sketch. A grammatical concept represents an underlying general linguistic property which cuts across a range of cross-linguistic grammatical differences. As such, grammatical concepts are not tied to specific target languages, or specific language combinations in L2 teaching contexts. The idea is that because the concepts are applicable cross-linguistically, they should prove potentially useful to demonstrate and explain points of usage in any target-language, or to explicate learnability difficulties in any language pair. The notion would therefore be amenable to both language-neutral treatments and language-specific instantiations. For example, a language-neutral treatment might be aimed at language teachers generally with exemplification drawn from a wide range of languages in order to foster a way of thinking about grammar. A language-specific treatment could be a conceptual grammar of Japanese or German or any individual language, which uses grammatical

30

T. Rankin and M. Whong

concepts as an organising principle to present the grammar of a particular language in a fresh way for teachers and/or learners. Actually providing grammatical concepts necessitates some indication of possible concepts around which pedagogical grammars could be built. No definitive, exhaustive list is possible; rather, as with any pedagogical or reference grammar, there is potential variation with respect to the level of granularity of presentation/organisation. For example, in a traditional grammar of English, one might choose to take Articles as a unit or chapter, or one might chose the broader category Nouns, and subsume the properties of articles within that, etc. Any presentation of grammatical concepts would be subject to the same variation. We propose two principles on which grammatical concepts should be based. Firstly, they should draw on what we know from linguistic research to be relevant underlying formal or semantic principles which constrain acquisition. Secondly, concepts should be pitched at a level that facilitates maximum cross-linguistic comparison. This immediately rules out anything too language-specific. That is, a familiar topic like ‘the present progressive’ would not make a good grammatical concept. Instead, the occurrence of progressive forms (as well as other tenses and aspects from other languages) would exemplify a broader aspectual concept such as Perfectivity, which would include consideration of predicate types, the contribution of types of object, etc. (see Slabakova 2008: 143ff. for a detailed consideration of L2 acquisition challenges involving aspectual distinctions).The overall aim of a pedagogical treatment of aspect would be to heighten awareness of form-meaning mismatches, and the relevance of the underlying concepts even in the absence of an overt grammatical reflex such as a perfective morpheme. Other appropriate candidates for concepts which are implicated in formal and semantic properties cross-linguistically and relate to different overt forms would be areas like: unaccusativity, resultativity, genericity, aspect, etc. Some of these are familiar from existing pedagogical or descriptive grammars (aspect), while some remain confined in the main to linguistic or theoretical grammar (unaccusativity). Some would lend themselves to different levels of granularity, as in perfectivity as one sub-type of aspect among many. And, there would be cross-reference and cross-fertilization across concepts, as in the connection between generic meaning and habitual aspect. As a reminder, this is not a ‘how to’ proposal but a content proposal. Educators would use grammatical concept information as they see fit and choose the extent of information to employ accordingly. For example, in a highly multilingual classroom setting, a teacher might choose to refer to realisations in many different languages in order to address different potential learnability issues. Teachers working within a functional-notional curriculum might choose particular functional features which fit with aspects of the curriculum. The point is that the content would be attuned to specific requirements. In the next section, we develop our approach by exploring one specific area of grammar, in order to exemplify the potential of a grammatical concept approach.

Grammatical Concepts for Pedagogical Grammar

31

4 Grammatical Concepts Illustrated The discussion here centres on pronouns. The choice of pronouns is a practical one to facilitate a brief but informative presentation. Pronoun paradigms form a closed class of items that can be relatively straightforwardly translated between languages. However, pronouns are interesting in that different languages encode different formal, semantic and pragmatic concepts in pronouns, thus illustrating underlying general commonalities in meaning potential as well as differences in grammatical manifestation. The point of reference and comparison will be English. Again, this reflects practical concerns and should not be read as any ideological commitment to the status of individual languages. For better or worse, English is the most taught second/foreign language, with second language speakers of English far out numbering mother tongue speakers (see Seidlhofer 2011). It is therefore most likely to be implicated in language education internationally in one form or another. However, quite apart from reinforcing any cultural or educational ideology with respect to English as a foreign or second language, one insight that can be taken away from our approach is that English represents just one possible encoding of grammatical concepts, with no inherently special status in linguistic terms. Indeed, one could easily envisage how this sort of foundation could be built upon to discuss social, cultural and educational issues with respect to views of linguistic imperialism, language endangerment and similar concerns. But we return to our area of expertise to clarify the approach.

4.1 Pronouns A variety of formal and conceptual distinctions may be encoded in pronominal forms. The specific patterns that are grammatically realised in any one language must be acquired based on the contrasts marked in the input to learners (see Harley and Ritter 2002). An exploration of the possible types of distinctions and how they are realised serves to heighten conscious awareness of non-obvious properties of the native language, making implicit knowledge explicit in a cross-linguistic context. The paradigm of personal pronouns in English is a closed class of functional words as in other more or less closely related languages, illustrated in Table 1.3 A native speaker of English, or a learner of English whose L1 shares similar properties, will likely assume that this paradigm is the logical way to pronominally encode the world. There are, however, a range of grammatical concepts and distinctions encoded in pronouns in different languages. The particular concepts explored below are Person and Number, Gender, and Respect and Formality, which

3 We concentrate on personal pronouns for purposes of a manageable presentation. There are clearly

many other interesting facets that could be explored (distribution of reflexive pronouns, use of demonstrative pronouns, strong vs. weak paradigms, etc.)

32

T. Rankin and M. Whong

Table 1 English personal subject pronouns

Person 1st 2nd 3rd

Number Singular I You He/she/it

Plural We You They

combine in different ways to result in different arrays of usage patterns depending on the realisation of grammatical concepts in particular languages.

4.1.1

Person and Number Concepts

Personal pronouns typically distinguish the person and number of their referents. English makes three person distinctions (1st, 2nd, 3rd person) and two number distinctions (singular, plural). However, this is just one potential encoding of the underlying concepts of person and number. For speakers of English and similar languages, a straightforward singular (=one), plural (=more than one) grammatical encoding of entities is the clear binary distinction that must be realised in overt forms. However, number may be instantiated very differently in other languages. Indeed, Corbett (2000: 2) points out that English and familiar Indo-European languages have in fact “quite unusual number systems”. As applied to personal pronouns, there are languages which distinguish between dual (=two entities), trial (=three entities) and paucal (=a few entities, but fewer than plural) in addition to singular and plural. For instance, Upper Sorbian, a Slavic language spoken in Germany, differentiates between singular, dual and plural as in (6) (from Corbett 2000: 20). Larike, spoken in Indonesia, uses specific forms for exactly three, as in (7) (Corbett 2000: 21). (6) (7)

Ja ‘I’ mój ‘we two’ Duma hima aridu naPa. House that.1PERS.TRIAL.EXCL own.it ‘We three own that house.’

my ‘we’

This is valuable knowledge for language education from the perspective of general language awareness. It relativizes a narrow view of what language can do based on a small sample of Indo-European languages, and does so in an interesting way by illustrating that it is possible that speech communities which are less technologically advanced may have more sophisticated grammatical machinery than European languages (in the sense that they encode finer conceptual distinctions). Similarly, knowledge of the differential realisations of grammatical concepts provides a way of reflecting on variation and so-called non-standard or dialectal usages. Many speakers of English for example will encode plural consistently across the pronoun paradigm and so use forms such as youse, yiz, y’all, etc. depending on regional variety (see Hickey 2014: 253). Viewed from the perspective

Grammatical Concepts for Pedagogical Grammar

33

of the concept of number, this can be seen to address a gap in the standard English paradigm and thus encode number in a more consistent way. As we see in Table 1, the second person pronoun you does not change its form in singular versus plural usage, unlike the other personal pronouns, and unlike any number of languages with distinct you.sing./you.pl. forms. Rather than indicating a lack of literacy or linguistic sophistication, innovations such as youse are logical encodings of underlying grammatical concepts, which also index a sense of belonging to a particular speech community (see humorous discussion in McNally 2013). Moreover, there is a concept connected with person that is not systematically realised at all in English and closely related languages: Associativity. In languages that express associativity grammatically, first person plural pronouns can distinguish between inclusive (you and I) and exclusive (I, some other(s) but not you) usage. The South American language Jarawara, for example, uses inclusive ee and exclusive otaa as in (8) and (9) (from Dixon 2016: 39–40). (8) (9)

otaa jete na-bone ee jete na bone

‘We (not including you) are going out hunting’. ‘We (including you) are going out hunting’.

Given that this distinction is conceptual, it is obviously generally comprehensible regardless of one’s native language, but the grammatical encoding of such differences is problematic for learners. Even if one is not learning a language which has this distinction, awareness of it provides a way to view issues in familiar languages. Languages which encode the inclusive/exclusive distinction do not have ambiguity in cases where English and similar languages do. Zimmer (2010) says that this ambiguity has engendered its own humorous tradition of jokes, as in (10). (10) Surrounded by wild, screaming Indians, the Lone Ranger desperately asks Tonto, ‘What will we do?’ Tonto replies, ‘What do you mean “we”, paleface? Even if English does not have a formally recognised grammatical reflex of the concept, it is at stake in usage questions such as the ‘royal we’, the ‘editorial we’ or “nurse we” (on nurse we, see Collins and Postal 2012: 217, after Joseph 1979). Consider first “nurse we” as illustrated in (11) from Collins and Postal (2012: 217) said by a nurse addressing a patient.4 (11) Are we taking good care of ourselves today? Even though English does not consistently formally encode the inclusive/exclusive distinction directly on pronouns, traces of this can be found in variation in this structure. For example, some English speakers allow two forms of the first person plural reflexive (ourselves and ourself ), and this seems to pattern with exclusive usage of we. In (12), the use of ourself would be possible for many speakers and it indicates that we does not include the speaker, in effect it means singular you, resulting in the singular form of the associated reflexive. 4 Of course, much the same applies to any similar communicative situation, the rather unfair naming

convention is derived from similar examples discussed in the literature involving this particular sort of interaction.

34

T. Rankin and M. Whong

(12) Are we taking good care of ourself today? The same analysis can be applied to editorial or royal we, with a complementary pattern of meaning, instead of we meaning in effect singular you, editors and royals can use we to mean in effect I. Seen from this grammatical perspective, this is simply English making use of the conceptual distinctions of associativity that language permits. However, it provides a way of reflecting critically on “the presumptuous use of we that inspires so much outrage” (Zimmer 2010) and thus helps learners to consider their own usage and its effects from a linguistically informed perspective. Investigating the conceptual basis of constructions such as these may be helpful for learners and educators to navigate purist or aesthetic reactions, as well as issues of variability.

4.1.2

Gender Concept

In English, animacy and biological sex are marked on pronouns, leading to a three-way distinction, but only in third person singular (he/she/it). Given that the grammatical concept of gender permits the encoding of natural and/or grammatical gender, it is unsurprising that other languages make more or less extensive use of similar distinctions. French distinguishes masculine and feminine forms in the third person plural, as in (13). Arabic requires gender to be marked on second and third person, singular and plural; second person examples are illustrated in (14) and (15). (13) ils (masculine they); elles (feminine they) (14) anta (you singular masculine); anti (you singular feminine) (15) antum (you plural masculine); antunna (you plural feminine) Difficulties of acquiring and using grammatical gender are particularly well attested in the acquisition and processing literature (e.g. Hopp 2013; McCarthy 2008; Sabourin and Stowe 2008, among others). For learners whose L1 lacks a grammatical gender system, or for learners whose L1 instantiates a different type of grammatical gender system, learning and using grammatical gender in a foreign language represents an extreme learnability issue. This is unsurprising given the intricacies of grammatical gender, the assignment of which can be affected by lexical, phonological and morphological properties. There is thus no easy recommendation of something to teach to address the issues, which will vary depending on the nature of the target language and the particular gender system it instantiates. However, by exploring something of the implications of the grammatical concept, it is possible to raise awareness for the variety of properties that may be potentially marked in any one language. In this way, increased sensitivity for the types of distinctions that languages mark can be fostered, priming learners to look for such distinctions in their input, and hopefully leading to gains in learning.

Grammatical Concepts for Pedagogical Grammar

35

In addition to learning about the grammatical properties, the realisation of the concept of gender also leads to tricky sociolinguistic usage issues. In France, for example, there has been charged debate about the correct use of titles as illustrated in (16). The issue being that ministre (government minister) is a “masculine” noun supposedly grammatically requiring the masculine article le, but if a woman is a government minister, should she be addressed with the feminine la? Is grammatical masculinity of ministre just a hangover from outmoded gender roles where ministers were always men? (16) Madame la (fem) minstre/Madame le (masc) ministre. Linguistic controversy also follows the use of gender-neutral epicene pronouns in English, for example singular they can evoke prescriptivist and socio-politically charged discussion (see Newman 1992, 1996). A grammar-concept treatment will of course never resolve such issues but it can put them into some linguistic perspective. For example, as outlined in the discussion of person and number, the connection between grammatical form and the type of referent is not necessarily grammatically marked in English as opposed to other languages, such that we may refer in fact to the second person singular while you serves a double function as both singular and plural. In this context, any objections to the use of a ‘plural’ pronoun to refer to a singular referent cannot be linguistically justified. Moreover, the choice of gender pronouns can obey its own grammatical and conceptual rules quite apart from the social or prescriptive notions of connections to biological gender. Variation in dialectal forms can illustrate. For example, Wagner (2004) points out that dialects of Southwest England and Newfoundland express a mass-count distinction in the use of gendered pronouns, so that individual count nouns are referred to with he/him while abstract or mass nouns are referred to as it (17). While typologically rare, similar distinctions can be found in other languages, illustrating that it is a natural potential encoding of conceptual distinctions (Wagner 2004: 282). (17) they drained ’im into that brook. (Cf. they drained it (i.e. a river) into that brook.) This perspective helps to neutralise the charged socio-political aspects of linguistic controversies, or at least furnishes linguistically informed ways of thinking which put the educational and social aspects of language use in a different perspective. In this case, while grammatical gender often relates systematically to biological gender as expressed in pronouns, it can also be seen as an overt way to mark certain conceptual distinctions, which can be instantiated very differently in systems for classifiers in different languages. Many of these issues are specific to individual languages. But the contention is that a general understanding on the conceptual level may prove useful for teachers even if they do not have to teach a language which consistently encodes gender. English lacks a robust grammatical gender system, but English teachers who understand something of the concept could use it to shed light on variable usage and prescriptive questions in English.

36

4.1.3

T. Rankin and M. Whong

Respect and Formality Concepts

The encoding of levels of formality or respect in pronouns is a further grammar concept distinction which is widespread but which English does not make major use of. In this respect, English is fairly typical as most languages do not overtly encode politeness or honorifics on pronouns (Helmbrecht 2013). Nevertheless, a binary polite versus formal distinction is instantiated relatively frequently crosslinguistically, often interacting with number so that plural pronouns represent the polite or formal form, and the singular the familiar form. This is familiar from the politeness distinctions found in European languages, as in French tu (you singular/familiar) and vous (you plural/you formal). Exploration of this type of feature provides a window into the nature of pragmatics and cultural conventions of a target language speech community. By doing so from a grammatical perspective of considering the general distinctions that are available in language, one can gain greater insight into how social relations and interaction might be instantiated differently, even in socio-culturally similar speech communities. For example, the German du/Sie distinction and the Russian ty/vy distinction reflect the same familiar/polite usage illustrated for French. However, usage may differ significantly. Once a form of address is established between interlocutors in German, it does not vary. If one is on familiar terms with an interlocutor, the familiar form of address will be used consistently. If one is on unfamiliar terms, the polite form will be used consistently. One may progress from using the polite to the familiar form as the nature of an interpersonal relationship changes. By contrast, in Russian, it is possible to switch between the familiar and polite forms with the same interlocutor within the same interaction, depending on the topic under discussion and the discourse context (Helmbrecht 2013). It can be seen, then, that even a binary distinction has the potential to encode a range of complex levels of formality when viewed cross-linguistically. This is, however, only one way of expressing such distinctions. The concept of formality may find overt expression in more intricate distinctions. Helmbrecht (2013) points out that Marathi, spoken in India, distinguishes between the second person form t¯u used for family members and intimate persons, two separate honorific forms te and he for people with higher social status, and a polite high honorific form a¯ pan. for priests and teachers and in very formal contexts. It may be that the distinction is not realised on pronouns, but elsewhere in forms of address, or register choices, illustrating a full range of levels of politeness and differential forms of grammatical encoding of conceptual distinctions. For example, languages like Korean and Japanese involve an extensive system of honorific marking. In Korean, every sentence and utterance which includes reference to someone for whom deference is expected will employ a separate verb ending, or honorific marker, si. The example in (18) illustrates, where reference to father as subject involves addition of si to the verb, as compared to (19).

Grammatical Concepts for Pedagogical Grammar

(18)

(19)

apeci-ka o-si-nta. Father.SUBJ come.HON.PRES.PROG ‘Father is coming.’ tongsayng-i onta sibling.SUBJ come.PRES.PROG. ‘Brother (or Sister) is coming.’

37

(Lee and Ramsey 2000: 239–40)

There are formally six ‘speech levels’, with every verb in Korean required to reflect one of the six levels, though the ‘middle’ levels, blunt and familiar are no longer routinely used and becoming obsolete (Sohn 1999: 413). The verb endings for each of the levels (in declarative sentences) are as follows: (20)

Plain level Intimate level Familiar level Blunt level Polite level Deferential level

-ta -a/-e -ne-y -(s)ol-(s)wu -(a/e)yo -(su)pnita

This demonstrates the diversity of overt encodings of shared grammatical concepts. The same sort of distinction can be more or less intricately expressed, and can be expressed in different ways, depending on the grammatical machinery the language makes available, for example by changing the form of address, or the morphological agreement patterns required. The common thread running through these discussions of features associated with pronouns (or similar features expressed elsewhere in the linguistic system) is that a more conceptual, linguistically-informed comparative approach highlights the nature of the grammatical tools that languages have at their disposal to express a variety of conceptual distinctions. This aids learning by sensitising learners to the nature of grammar generally in order to provide analytical ways of approaching the learning of any specific languages in a way that is grammatically faithful, but which invites a more conceptually engaging approach to grammar. Moreover, it provides ways of reflecting on issues of language in use, variation and prescriptivism which may be hotly debated in one’s native language(s) as well as a target L2.

5 Conclusions Linguistics has often been seen as an esoteric and abstract subject far removed from the concerns of teachers tasked with teaching either specific foreign languages, or the language arts in learners’ native language. However, given broader educational policy goals of cross-linguistic awareness in addition to just language proficiency, we have claimed that consideration of abstract, universal underlying properties of language in fact represents a valuable take on pedagogical grammar. We formulated

38

T. Rankin and M. Whong

this programmatically in terms of the notion of grammatical concepts. Making these ideas available for education would represent an important step in rethinking the nature of pedagogical grammar. In terms of models in language pedagogy, the ideas are perfectly amenable to approaches already outlined in the literature. There have been calls for foreign language pedagogy to involve “teaching for transfer” (Cummins 2008), and contemporary ideas of multilingual pedagogies invoking multilingualism, plurilingualism or translanguaging continue variations on similar themes. A grammatical concepts approach provides a way of thinking about how to consider the grammatical properties of any language and allows for comparisons across multiple languages, and could thus serve as a vehicle to translanguaging practices. The grammatical concepts notion also shares similar conceptual foundations to Butzkamm’s (2003; see also Butzkamm and Caldwell 2009) idea that “we only learn language once”. Rather than assuming that learning of a new language needs to start from scratch, much of the grammatical (and lexical, phonological, etc.) knowledge from the native language can be applied to the L2 input. This notion of “bilingualization” sees existing L1 linguistic resources as a valuable source of knowledge and skills for L2 learning, rather than as a hindrance in the form of cross-linguistic influence. Grammatical concepts can be formulated so as to make explicit any relevant L1 knowledge and skills that could be positively employed to aid learning of L2 grammar. There has of course been much debate about the use of the L1 in foreign language teaching (e.g. Turnbull and Dailey-O’Cain 2009), but it is surely a sound pedagogical principle, no matter what subject is being taught, to build on existing knowledge. To assume that foreign language learners are linguistic blank slates ignores the rich implicit and explicit knowledge they already have that can be brought to bear on the task of learning the L2. Moreover, given the nature of our modern, globalised world, chances are that most classrooms will include learners who, in addition to competence in a standard national variety, will have knowledge of variants of a heritage language, or (at least one) Lingua Franca variety, whether that’s English, Chinese, Spanish, Swahili or Arabic, to name the largest. Unfortunately, especially in standard literacy education in national varieties, such competence may in fact be seen as a hindrance or a problem to be overcome. We argue that this large body of implicit knowledge about a range of subtle and complex features of our native language(s) is a valuable resource that can be harnessed in education in order to explicate any number of grammatical issues in ‘standard’, ‘dialectal’, ‘minority’ or ‘foreign’ languages. The ability to harness these resources, however, depends on awareness on the part of teachers, and teacher trainers of the sort of use that can be made of formal properties of language to illustrate abstract points and draw comparisons between different languages and varieties to illuminate questions of form and usage. This depends on some training in (formal) linguistics, which, given the development of the field of applied linguistics, cannot be taken for granted. Increasing engagement from formal linguists and language acquisition researchers can help to make relevant findings from the field available and applicable to educators, as long as we are aware

Grammatical Concepts for Pedagogical Grammar

39

of the imperative to mediate knowledge rather than apply or impose a linguistic perspective. Being able to draw knowledgably and strategically on a wide range of grammatical properties in order to illustrate the nature of language would aid in the attainment of education goals for language learning, as set out poetically by Hawkins (1984: 6): We are seeking to light fires of curiosity about the central human characteristic of language which will blaze throughout our pupils’ lives. While combating linguistic complacency, we are seeking to arm our pupils against fear of the unknown, which breeds prejudice and antagonism. Above all we want to make our pupils’ contacts with languages, both their own and that of their neighbours, richer, more interesting, simply more fun.

References Association of Language Awareness. (2018). About. http://www.languageawareness.org/ ?page_id=48. Accessed on 14 Aug 2018. Beacco, J. -C., & Byram, M. (2007). Guide for the development of language education policies in Europe. From linguistic diversity to plurilingual education. Language Policy Division, Council of Europe. https://www.coe.int/t/dg4/linguistic/Source/FullGuide_EN.pdf Block, D. (2003). The social turn in second language acquisition. Edinburgh: Edinburgh University Press. Blommaert, J., & Rampton, B. (2011). Language and superdiversity. Diversities, 13(2), 1–21. Brumfit, C. (2001). Individual freedom in language teaching: Helping learners to develop a dialect of their own. Oxford: Oxford University Press. Butzkamm, W. (2003). We only learn language once. The role of the mother tongue in FL classrooms: Death of a dogma. The Language Learning Journal, 28(1), 29–39. Butzkamm, W., & Caldwell, J. (2009). The bilingual reform: A paradigm shift in foreign language teaching. Tübingen: Narr-Studienbücher. Canale, M., & Swain, M. (1980). Theoretical bases of communicative approaches to second language teaching and testing. Applied Linguistics, 1, 1–47. Choi, M.-H., & Lardiere, D. (2006). The interpretation of wh-in-situ in Korean second language acquisition. In A. Belletti, E. Bennati, C. Chesi, E. DiDomenico, & I. Ferrari (Eds.), Language acquisition and development: Proceedings of GALA 2005 (pp. 125–135). Cambridge: Cambridge Scholars Press. Collins, C., & Postal, P. (2012). Imposters: A study of pronominal agreement. Cambridge, MA: The MIT Press. Corbett, G. (2000). Number. Cambridge: Cambridge University Press. Cummins, J. (2008). Teaching for Transfer: Challenging the Two Solitudes Assumption in Bilingual Education. In J. Cummins & N. H. Hornberger (Eds.), Encyclopedia of language and education, Volume 5: Bilingual education (pp. 65–75). Berlin: Springer. Dixon, R. M. W. (2016). Are some languages better than others? Oxford: Oxford University Press. Gil, K.-H., & Marsden, H. (2013). Existential quantifiers in second language acquisition: A feature reassembly account. Linguistic Approaches to Bilingualism, 3(2), 117–149. Harley, H., & Ritter, E. (2002). Person and number in pronouns: A feature-geometric analysis. Language, 78(3), 482–526. Hawkins, E. W. (1984). Awareness of language: An introduction. Cambridge: Cambridge University Press. Helmbrecht, J. (2013). Politeness distinctions in pronouns. In M. Dryer & M. Haspelmath (Eds.), The World Atlas of language structures online. Leipzig: Max Planck Institute for Evolutionary Anthropology. (Available online at http://wals.info/chapter/45. Accessed on 2018-08-20).

40

T. Rankin and M. Whong

Hickey, R. (2014). A dictionary of varieties of English. Malden: Wiley-Blackwell. Hopp, H. (2013). Grammatical gender in adult L2 acquisition: Relations between lexical and syntactic variability. Second Language Research, 29(1), 33–56. Hudson, R. (2004). Why education needs linguistics (and vice versa). Journal of Linguistics, 40(1), 105–130. Hudson, R. (2008). Linguistic theory. In B. Spolsky & F. Hult (Eds.), The handbook of educational linguistics (pp. 53–65). Oxford: Blackwell. Hymes, D. (1972). On communicative competence. In J. B. Pride & J. Holmes (Eds.), Sociolinguistics: Selected readings (pp. 269–293). Middlesex: Penguin Books. Joseph, B. (1979). On the agreement of reflexive forms in English. Linguistics, 17, 519–523. Kramsch, C. (2008). Applied linguistic theory and second/foreign language education. In N. Van Deusen-Scholl & N. H. Hornberger (Eds.), Encyclopedia of Language and Education, Volume 4: Second and Foreign Language Education (pp. 3–15). Berlin: Springer. Lardiere, D. (2008). Feature-assembly in second language acquisition. In J. Liceras, H. Zobl, & H. Goodluck (Eds.), The role of formal features in second language acquisition (pp. 106–140). New York: Lawrence Erlbaum Associates. Lardiere, D. (2009). Some thoughts on a contrastive analysis of features in second language acquisition. Second Language Research, 25, 173–227. Lee, I., & Ramsey, R. S. (2000). The Korean language. New York: State University of New York Press. Long, M. (1991). Focus on form: A design feature in language teaching methodology. In K. De Bot, R. Ginsberg, & C. Kramsch (Eds.), Foreign language research in cross-cultural perspective (pp. 39–52). Amsterdam: John Benjamins. Marsden, H., & Slabakova, R. (2018). Grammatical meaning and the second language classroom: Introduction. Language Teaching Research. https://doi.org/10.1177/1362168817752718. Electronic publication ahead of print version. McCarthy, C. (2008). Morphological variability in the comprehension of agreement: An argument for representation over computation. Second Language Research, 24(4), 459–486. McNally, F. (2013). The correct Youse of English. The Irish Times, December 11. Newman, M. (1992). Pronominal disagreements. Language in Society, 21, 447–475. Newman, M. (1996). Epicene pronouns: The linguistics of a prescriptive problem. New York: Garland. Nunan, D. (1995). Language teaching methodology. A textbook for teachers. New York: Phoenix ELT. Sabourin, L., & Stowe, L. A. (2008). Second language processing: When are first and second languages processed similarly? Second Language Research, 24(3), 397–430. Schwartz, B., & Sprouse, R. A. (1996). L2 cognitive states and the full transfer/full access model. Second Language Research, 12(1), 40–72. Seidlhofer, B. (2011). Understanding English as a lingua Franca. Oxford: Oxford University Press. Slabakova, R. (2008). Meaning in the second language. Berlin: DeGruyter Mouton. Sohn, H.-M. (1999). The Korean language. Cambridge: Cambridge University Press. Turnbull, M., & Dailey-O’Cain, J. (2009). Introduction. In M. Turnbull & J. Dailey-O’Cain (Eds.), First language use in second and foreign language learning (pp. 1–14). Bristol: Multilingual Matters. Van Lier, L. (1996). Interaction in the language curriculum: Awareness, autonomy, and authenticity. London: Longman. Van Rijt, J., & Coppen, P.-A. (2017). Bridging the gap between linguistic theory and L1 grammar education – experts’ views on essential linguistic concepts. Language Awareness, 26(4), 360– 380. Wagner, S. (2004). ‘Gendered’ pronouns in English dialects: A typological perspective. In B. Kortmann (Ed.), Dialectology meets typology: Dialect grammar from a cross-linguistic perspective (pp. 479–496). Berlin: Mouton de Gruyter.

Grammatical Concepts for Pedagogical Grammar

41

White, L. (2006). Universal grammar and second language acquisition. Cambridge: Cambridge University Press. Whong, M., Gil, K.-H., & Marsden, H. (Eds.). (2013). Universal grammar and the second language classroom. Dordrecht: Springer. Widdowson, H. G. (2000a). On the limitations of linguistics applied. Applied Linguistics, 21(1), 3–25. Widdowson, H. G. (2000b). Object language and the language subject: On the mediating role of applied linguistics. Annual Review of Applied Linguistics, 20, 21–33. Widdowson, H. G. (2003). Defining issues in English language teaching. Oxford: Oxford University Press. Widdowson, H. G. (2016). ELF, adaptive variability and virtual language. In M. L. Pitzl & R. Osimk-Teasdale (Eds.), English as a lingua Franca: Perspectives and prospects (pp. 31–38). Berlin: De Gruyter. Zimmer, B. (2010). We. New York Times Magazine. October 1, 2010.

Part II

Native Language Settings

Teaching Word Order Variation with a Constraint-Based View on Grammar Daniel Gutzmann and Katharina Turgay

Abstract One prominent linguistic feature of German is that it exhibits rather flexible word order in the so-called middle field. And while much theoretical research has been carried out on this phenomenon, not much of the insights gained there has been carried over to teaching German in schools (in fact, in many school text books, word order is depicted as rather rigid). Using optimality theory as an example, this article sketches how modern linguistic theories can be used in the classroom to teach aspects of the Grammar of German and to raise the language awareness of the students. For this, we lay out the basics of optimality theory (OT) and how it applies to word order variation, before we sketch different ways how the basic of OT can be taught to younger and older students and why this would be a perfect fit for the officially supported curricula of the German school system. Keywords Word order · Optimality theory · Gradual grammaticality · Acceptability · Grammar teaching · Language awareness · German

1 Introduction The observation that word order in German is rather flexible is one of the most prominent features of German syntax and from a linguistic point of view, it is certainly one of the most studied phenomena. However, the knowledge that has been gained from this research does not seem to play any role when it comes to teaching German in schools. This contribution tries to close this gap a bit more by

D. Gutzmann () Institut für deutsche Sprache und Literatur I, Universität zu Köln, Köln, Germany e-mail: [email protected] K. Turgay Institut für Germanistik, Universität Landau, Landau, Germany e-mail: [email protected] © Springer Nature Switzerland AG 2020 A. Trotzke, T. Kupisch (eds.), Formal Linguistics and Language Education, Educational Linguistics 43, https://doi.org/10.1007/978-3-030-39257-4_4

45

46

D. Gutzmann and K. Turgay

suggesting that a modern linguistic theory like optimality theory (OT) can be put to use for modern language didactics and be used, to a certain extent, in the classroom.

1.1 Word Order in German One important and well-known feature of the syntax of German is that the word order of nominal constituents is relatively free. This is particulary true for the word order in the so-called “middle field”, i.e. the stratum that is demarcated on the left side by the finite verb in second position or, in case of subordinate clauses, subjunctions, and, on the right side, by non-finite verbal elements or finite verbs in final position. Reordering the constituents may lead to less acceptable sentences, but rarely leads to completely unacceptable structures. (1)

a.

b. c. d.

Gestern hat eine Schülerin der Lehrerin den Schlüssel yesterday has a student.FEM the teacher.FEM the key geklaut. stolen ‘Yesterday, a student stole the key from the teacher’. Gestern hat der Lehrerin eine Schülerin den Schlüssel geklaut. Eine Schülerin hat gestern der Lehrerin den Schlüssel geklaut. ...

There is an extensive body of research on the question of what guides the order of constituents in the middle field, coming from various theoretical background and often being backed-up by empirical research.1 This intensive research led to the identification of various factors that seem to influence word order and impact the relative acceptability of a given linearization, both morpho-syntactic and rather semantic-pragmatic factors, including but not limited to the following aspects: a. b.

morpho-syntactic factors: case, syntactic function, phrase type etc. semantic-pragmatic factors: agentivity, animacy, definiteness, focus, topicality

Beyond these, prosody also plays an important role, as do considerations about binding and scope.

1 Amongst

many others, see Lenerz (1977, 1981), Keller (2000), Pafel (2009), Müller (1999), Jacobs (1988), Audehm (2006), Zeman (2002), Turgay (2017), Abraham (1985), Büring (2001), Fanselow (2006), Frey (2004), Meinunger (2000), Molnárfi (2004), and Struckmeier (2014).

Teaching Word Order Variation with a Constraint-Based View on Grammar

47

However, while in linguistic research such a multi-factorial view on word order in German is at least widely accepted,2 these findings have not been taken over in language teaching contexts for the most part. As a look at textbooks for German classes in secondary/middle schools shows, word order variation in German is rarely discussed at all. Instead, the word order is often depicted as rather rigid in school text books, for instance as strictly being SVO (see the brief discussion of Turgay’s (2014) study in the next section). However, the linearization of constituents is not just something that warrants theoretical attention and analysis, it is – and this is something practical that the research made clear – an important tool of information structure in German and, as such, word order is for anchoring a sentence in the text/discourse context.3 Therefore, the intentional use of word order variations is an important skill for writing coherent texts. The following example illustrates this. One of the semantic-pragmatic factors involved in the middle-field word order is focus. A constituent that is in focus – for instance, because it provides the answer for a corresponding wh-question as in (2) and (3)– tends to follow the nonfocus constituent. Therefore, different word orders are favored depending on which question was asked (ignoring all other factors for the moment). (2)

A: B:

(3)

A: B:

Wem hat eine Schülerin gestern den Schlüssel geklaut? ‘From whom did the student steal the key yesterday?’ Gestern hat eine Schülerin den Schlüssel der LEHRerin geklaut. Was hat eine Schülerin gestern der Lehrerin geklaut? ‘What did the student steal from the teacher yesterday?’ Gestern hat eine Schülerin der Lehrerin den SCHLÜSSel geklaut.

This simple example already shows that varying the word order can help to tailor the sentence more to its context. This, of course, does not mean that there is something inherently “wrong” with sticking to the basic word order inside the middle field (which is subject ≺ indirect object ≺ direct object). Because it is the base order, it will fit most contexts, even if other word orders may be a better fit and therefore preferred. This is arguably true in particular for written language, since it lacks intonational information and hence putting the focused element to the end of the middle field as in (2) may help the reader to figure out the intended focus.4 This then ties back into language teaching. Since text production is one of the most important topics in German classes, it is important that students understand that they can adapt word order to structure their texts, rather than always going

2 For

different views, see for instance Struckmeier (2014) and Fanselow (2006), who argue that it is mostly prosody (together with semantics) that determines word order and that especially information-structural notions do not play a role for the syntactic linearization in German. 3 Again we should note here, that the idea that word order is used to optimize for information structure and coherence is challenged by, for instance, Fanselow (2006) and Struckmeier (2014). 4 Thanks to an anonymous reviewer for pointing this out. They also mention the production experiment conducted by Skopeteas and Fanselow (2010), who could not elicit non-base word orders from native speakers (based on givenness).

48

D. Gutzmann and K. Turgay

with a S(V)O order, in order to fully employ the discourse structuring potential that German offers them. In the following, we sketch how the factors interact with each other and impact word order and also present some suggestions for how the core ideas could be taught at school. Before doing so, we will however first show how word order is actually approached in textbooks for German classes.

1.2 Word Order in Textbooks In a previous study, Turgay (2014) looked at 18 current textbooks for teaching German in fifth grade for different school types and which were published between 2004 and 2013. Only fifth grade textbooks were looked at, due to the simple reason that in textbooks for higher grades, word order is not a topic at all. One crucial result of that study was that in most textbooks, word order is illustrated by tables and the example sentences are divided into Satzglieder which are sorted into these tables. There are two different ways that this is illustrated and which can be found independently of the school form.5 From the 18 textbooks, word order is illustrated in a rigid way, always in the order subject ≺ verb ≺ dative object ≺ accusative object.6 In the remaining 8 textbooks, word order variants are at least mentioned. In these cases, students are mostly supposed to change the word order by applying permutation tests; where it is sometimes even explicitly mentioned that this can improve their texts, even if no further details as to how are given. However, the word order variants that are given are still very limited: only subject and object are swapped in the examples given (i.e. object in first position, subject after verb second position). Scrambling inside the middle field (i.e. between subject and object or two objects) is not mentioned at all. In addition, also the order between objects is presented as being rigid. Another outcome of the study in Turgay (2014) is that more textbooks for Gymnasium address flexible word order than the books for other school forms. In addition, there seems to be a tendency that the newer the books, the more often flexible word order is illustrated. And in a few of the newest textbooks, there is actually a cross-linguistic comparison between the word order in German and English that contrasts the more flexible German word order with the rather strict one found in English.7

5 In

Germany there are the three different school forms after primary school (which usually encompasses first to fourth grade): “Hauptschule” (grade 5–9), “Realschule” (grade 5–10), and Gymnasium (grade 5–12/13). 6 It must be noted that the text books never explicitly state that the SVO order they present is the only viable option; which would be truly outrageous. However, by just presenting this option as the structure of a German sentence, they highly suggest that it is the only option, which is also highly problematic. 7 As we will discuss in Sect. 4, one aspect that gives us hope that flexible word order will be discussed in more detail in future textbooks is that the topological field model is now an official

Teaching Word Order Variation with a Constraint-Based View on Grammar

49

Surprisingly, even the textbooks from the study that mentioned the fact that word order in German is not rigid do not really give any reasons for why one would like to vary the word order.8 But, as already alluded to above, word order variation is an important tool to structure texts, especially due but not confined to the role word order variation plays for information structure, which, in turn, is important for structuring texts. We believe that if one adopts a more flexible view on word order which is based on the theoretical concept of linguistic constraints and their interaction with each other (in contrast to rigid rules that must be followed no matter what), this can be employed to teach students about the different ways they can structure their texts.

1.3 Outline In order to sketch out our idea how this may eventually be achieved, the remainder of this paper is structured as follows. In the next section, we present a small empirical study that we carried out for this paper and in which we had a look at texts produced by fifth and sixth graders to see which word orders they actually use in their writing. The results will then be compared to texts produced by higher classes (grades 11–13) to see if the school curriculum actually leads to any change in the respective word orders used in students’ texts. In Sect. 3, we will go into a bit more detail about the theoretical background that underpins the approach we suggest to pursue in the class room: a constraint-based approach to word order variations that is couched in the framework of optimality theory (OT). We also present ideas of how the core ideas of OT could be presented in a way that is approachable for students and connects to their everyday-life experiences. In Sect. 4 we then suggest how this can be used to teach word order variation in school and show that the ideas, concepts and skills to be acquired by teaching an OT-style view on word order are deeply anchored in the national education standards, called “Bildungsstandards” (standards of education). The final Sect. 5 concludes.

2 Word Order in Students’ Text: A Quick Corpus Study The study from Turgay (2014) that we sketched above shows that word order is a topic in text books for students in the fifth grade, but that the text books often misleadingly present word order as being rigidly SVO. However, our hope is that aspect in the new official curriculum of the federal state of Baden-Württemberg (KMBW 2016, 32), which lets us expect that other states will catch up with this as well. 8 As a reviewer suggested, this may be due to the fact that there is a gap between what linguistic theories have to offer to explain this and what kind of “language theories” teachers use when teaching. To lessen this gap even so slightly is one of the aims of the article, and the entire volume in which it appears in.

50

D. Gutzmann and K. Turgay

when this topic is actually taught, teachers will address the variability of word order in German. To get a better picture of how students actually write and which word order variations they actively use, we conducted a simple empirical study. We analyzed texts from students of the so-called “Orientierungsstufe” (5th and 6th grade) with respect to the word orders they use in the texts they produce. Do they build sentences mostly according to the rigid SVO word order as presented in the text books or do they vary the word order? In addition, we also analyze a corpus of texts from students from higher grades (“Sekundarstufe II”, grades 11–13) in order to see if there is a difference between the younger students and the older ones. This could give hints about whether or not we could assume that there is a progression towards more variation of word order. This may be expected, because students of the higher grades are supposed to be better at structuring their texts. To find out whether we can register a progression from the lower to higher grades years, we compare the results for the 5/6 graders with results of an analysis from texts which are written from students of the “Sekundarstufe II”, the highest school level.

2.1 Procedure The texts we could look at, both from the younger and the older students, were produced over the course of two school years. That is, the texts from the 5th and 6th grade students are from the same set of students, and the texts produced by the 11th to 13th year students are from a different set. There are 24 students in the younger group, and 23 students in the older group and we look at four texts from different points in time for each group (Table 1). From both groups (grade 5/6 vs. grade 11–13) we randomly choose one written page from one of their four texts irrespective of the genre of the text produced or the point in time. This gives us around 1,500 sentences9 from each group. Out of these sentences, we then determine the word order concentrating on the complements and the verb. That is, we checked the relative order between subject, verb and three objects: direct, indirect and genitive object,10 omitting adverbials and expressions of other syntactical functions like particles or negation.11

9 By

this, we basically mean every phrase with a finite verb, but also infinite verbal phrases. That is, when we have a main clause with one (finite or infinite) embedded clause, we coded these as two sentences. 10 We don’t consider prepositional objects, because they are rather strongly confined to the right side of the middle field directly before possible verbal elements in the so-called right sentence bracket. 11 As a reviewer pointed out, by not coding for negation, we cannot look at effects that negation may have on word order (i.e. due to scope reason, see Bobaljik and Wurmbrand 2012 and Wurmbrand 2008). However, as we will see, there is not much variation in word order to begin with, and we are aiming for a rather coarse quantitative impression here, we think it is safe to leave such detail questions aside for the purposes of this short study.

Teaching Word Order Variation with a Constraint-Based View on Grammar Table 1 Overview of the data

Grade 5/6 24 students Texts from 4 points in time 1,500 sentences

Table 2 Most frequent word order patterns used by the grade 5/6 group

subject ≺ verb ≺ direct object subject ≺ direct object ≺ verb (last) verb ≺ subject ≺ direct object

51

Grade 11–13 23 students Texts from 4 points in time 1,500 sentences 37.2% of the relevant sentences 20.2% of the relevant sentences 19.4% of the relevant sentences

186× 101× 97×

2.2 Results 2.2.1

Word Orders Used by the Grade 5/6 Group

The students from grade 5/6 produced a total of 32 different word orders in their texts. Out of these, a few are attested only once, like indirect object ≺ verb ≺ subject ≺ direct object or direct object ≺ subject ≺ verb (last). Fifteen word order variants occur only five times or less. The most frequent word order we attested in the corpus of the grade 5/6 group is subject ≺ verb, which amounts to a quarter of all sentences we looked at (27.4%). The reversed order verb ≺ subject is the second most frequent linearization and it has an occurrence of 17.1%.12 That these two are the most frequent word order patterns already shows that a large proportion of the sentences includes only the verb and one complement, in this case the subject. Altogether two-thirds (997 out of 1500) of the sentences include a subject or one object beside the verb, but not both. Since we are mostly interested in the relative ordering of the complements with respect to each other, these sentences are not relevant for our purposes. Hence, we end up with roughly 500 sentences in which there are at least two complements and the verb. Out of this set, the subject precedes the object(s) to 92.2% (461×). This is the word order illustrated as rigid in most textbooks. Table 2 gives an overview of the most frequent word order patterns we found in the data produced by the 5th/6th graders.

2.2.2

Word Orders Used by the Grade 11–13 Group

The students from grades 11–13 produce 32 different word order patterns as well, albeit a different set than the students from the younger subject group did. For each group there are nine linearization patterns that only one group but not the other uses. “verb ≺ subject” we are referring to a configuration in which the verb is in second position and the subject in the middle field, while a different constituent occupies the sentence initial position. In our data, there were no verb-first clauses, like yes-no questions. Thanks to a reviewer for asking for clarification.

12 By

52

D. Gutzmann and K. Turgay

Like the younger students, the older students produce 15 different structures that only occur only five times or less, for instance the ordering subject ≺ direct object ≺ genitive object ≺ verb (last) or verb ≺ direct object ≺ subject. The two most frequent patterns used by the older students are word orders in which the subject precedes the verb: subject ≺ verb (last), which occurs in 22.1% of all sentences, and subject ≺ verb, which can be attested in 19.1% of all sentences. More than two-thirds of the 1,500 sentences contain only a subject or one (or more) objects in addition to the verb, but not both, which again is not relevant for our investigation. There are 471 items left, in which the subject and at least one object occurs and in which we therefore can determine an ordering between subject and object. Out of these items, the subject precedes the object in 92.8% (437×) of the cases. Table 3 gives an overview of the most frequent word order patterns we found in the data produced by the 11th–13th graders.

2.2.3

Comparison

The brief look at some actual data produced from students from the “Orientierungsstufe” (grade 5/6) and from “Sekundarstufe II” (grade 11–13) shows that the most frequent word order patterns that involve a subject and an object are the same for the two groups and in all of them, the subject precedes the objects. This corresponds to the rigid word order illustrated in most textbooks. The differences between both groups regarding the preceding of the subject in contrast to the preceding of the object is statistically not significant,13 which means that there is no real difference between the two groups. This comes as a surprise because we expected the older students to produce more variable output than the younger students. Table 4 shows the absolute and relative frequency of the two relative orders between subject and object for both groups. Table 3 Most frequent word order patterns used by the grade 11–13 group

Table 4 Ordering between subject and object

subject ≺ direct object ≺ verb (last) subject ≺ verb ≺ direct object verb ≺ subject ≺ direct object

Grade 5/6 Grade 11–13

13 Chi-square

33.1% of the relevant sentences 29.7% of the relevant sentences 18.0% of the relevant sentences

subject ≺ object 461 92.2% 437 92.8%

object ≺ subject 39 7.8% 34 7.2%

156× 140× 85×

Total 500 471

with Yates correction. Chi squared equals .049 with 1 degrees of freedom. The twotailed P value equals 0.8246.

Teaching Word Order Variation with a Constraint-Based View on Grammar

53

This small study of a corpus of texts produced by younger and older students shows that they most often produce the rigid word order subject ≺ object in their texts. That is, they do not permute the order between subject and object very often. In addition, the results of both groups do not differ relevantly, so there is evidence of the older students choosing their word order more deliberately to structure their texts. However, as we sketched in the previous section, the relative flexibility of German word order is an important tool for writing coherent and information structurally well-executed texts. Again, the base order is an option that is viable for most contexts and configurations, but other variants may fit a particular context more specifically and, in turn, therefore provide more clues to the reader to figure out the intended discourse structure. For this reason, we believe it would be beneficial to the students’ text production skills to teach the flexibility of German word order explicitly, and definitively not in the rigid way in which it is still illustrated in some textbooks. In addition, especially at the higher grades, this would be a good opportunity to familiarize students with the idea of linguistic modeling, which could be integrated into lessons on word order. This may bolster their language awareness and help them develop their analytic skills with respect to language. As we will discuss later in Sect. 4, these are important aspects in the German curriculum, so that we think that teaching word order variation in a linguistically informed way would not only be beneficial for the students’ skill with respect to writing, but would also fit squarely into what is expected from them by the standards of education (“Bildungsstandards”) anyway (KMK 2003). In the next section, we therefore sketch the theoretical background of a constraint-based view on word order variation and its foundation in optimality theory (OT). Of course, we do not think that it is practical to directly teach OT in the class room, which is why we sketch two different problem sets that may be used to familiarize students with the ideas of constraint ordering and interaction, without going too deep into actual OT-terminology.

3 The OT-View on Word Order Variation in German 3.1 Optimality Theory Optimality theory (OT) has originally been developed for phonology (Prince and Smolensky 1993; McCarthy and Prince 1994), but has later been adapted to syntax as well (Grimshaw 1997; Pesetsky 1997; Müller 1999, 2000). In broad strokes, it can be described as a theory of constraint conflicts and their resolution by ordering (and weighting) them. That is, with the ideas of constraints and constraint ordering (and weighting), it involves at least two (three) concepts that diverge from a traditional rule-driven view on grammar. First, the idea of a constraint differs from that of a rule insofar as constraints, under the OT understanding, can be violated if there are system-internal reasons to do so. In contrast, the rules in a rule-based grammar are rather absolute and cannot be violated, at least not without a noticeable impact

54

D. Gutzmann and K. Turgay

on grammaticality and acceptability. In addition, rules in a traditional rule-based grammar are rather independent of each other and each has to be followed by any structure on its own merits. This leads us to the second core idea of OT. Not only are the constraints violable, they are also ordered with respect to each other so that some constraints rank higher than others. It is for this reason that some constraint is violated by a linguistic expression, but only in order that a higher ranked constraint is fulfilled, which is in conflict with the higher ranked one. That is, the constraint ordering is what is responsible for them to be violable in the first place, which contrasts with linguistic rules which are neither ranked nor violable. Finally, the third idea, which is a more recent addition to OT-based approaches to syntax (Keller 2000), is that constraints are not just ordered, but weighted which means that they receive an explicit weight which reflects their respective linguistic importance. This means that a higher constraint has more weight than a lower constraint, but, and this will be crucial, smaller constraints are able to “gang up” against a higher constraint, if their cumulative weight is higher than the weight of the higher constraint.14 The idea of constraint weighting is particular useful when it comes to word order variation, as a violation of many minor constraints could be less acceptable than one violation against a higher-ranked constraint. Moreover, even with this in place, it should be noted that under a standard view of OT only the most optimal candidate, which best fits the constraints, is viewed as the grammatical one (hence the name). However, as has been noted in the literature, this is not really the case with word order variations in the middle field as many of the possible linearizations of constituents are grammatical (otherwise, we would not speak of “variation” here). We will not go into the details here, but just note that this problem has been addressed in the literature and there are modifications of OT that are suitable to deal with more than one grammatical output ((e.g. Müller 2000; Pafel 2009; Keller 2000, 2006); see also Struckmeier (2014) for a related but slighlty different view). For the purposes of this paper, we will however not address the question of how and where to draw the cut-off point between an “acceptable but not so great” structure and one that should be considered altogether unacceptable. For now, and especially for the transfer of the core ideas of an OT-based approach to word order variation, it will suffice to just observe that some linearizations are “better” and more acceptable than others.

3.2 Constraints on Word Order As mentioned in the introduction, the literature on word order in the German middle field has identified various constraints that influence the word order. These constraints involve morpho-syntactic and semantic-pragmatic ones. The 14 Keller’s approach as been taken up and modified, for instance in the so-called decathlon model developed by Featherston (2005), who, like Keller (2000), derives his model from experimental data.

Teaching Word Order Variation with a Constraint-Based View on Grammar

55

following gives an overview of the most discussed ones, but it is probably not exhaustive.15 (4)

a. b. c. d. e. f. g.

subject ≺ indirect object ≺ direct object nominative ≺ dative ≺ accusative definites ≺ indefinites pronoun ≺ full NP animate ≺ inanimate agens ≺ patiens/theme background ≺ focus

In many cases, a lot of these constraints go together, since some of them tend to flock together. For instance, subjects are usually in nominative case and likely to be agentive. And since subjects tend to be already established in a longer discourse, they often are definite and realized pronominally as well. Let us illustrate this with an example. First, let us assume that the context of the target sentence is given by the following question. (5)

Was hat die Lehrerin gestern erklärt? ‘What did the teacher explain yesterday?’

Since this question sets expectations of what the focus of the answer should be – the direct object in this case – this context helps to determine what the focus of the target sentence should be. Now, in the context of (5), consider the following example. (6)

Gestern hat sie eine FORmel erklärt yesterday has she.NOM a.ACC formula.ACC explained ‘Yesterday, she explained a formula.’

When we now determine the categories relevant for the constraints in (4) for each of the two constituents in the middle field, we see that they completely align with each other. We put those values that, according to the constraints in (4), should go first, in bold face in order to highlight this alignment. (7)

Gestern yesterday

hat has

sie she subject nominative definite pronoun animate agens background

eine FORmel a formula object dative indefinite NP inanimate theme focus

erklärt explained

constraints maybe unstressed ≺ stressed, topic ≺ comment, and short ≺ long (Hawkins 1994).

15 Additional

56

D. Gutzmann and K. Turgay

Because the categories are distributed as they are in (6), they all pull in the same direction and there is no conflict between any of the applicable constraints. And in fact, reversing the word order between subject and object in (7) in the context of (5) is strongly marked. (8) ??Gestern hat eine FORmel sie erklärt yesterday has a.ACC formula.ACC she.NOM explained ‘Yesterday, she explained a formula.’ However, and this is why the word order, even for a sentence in a specific context, feels rather flexible: In many cases, at least some of the constraints are in conflict with each other so that no perfect ordering that obeys all constraints is possible. To illustrate this, imagine a context as set up by the following question. (9)

Was macht dem Studenten jetzt Sorgen? ‘What worries the student now?’

In this context, consider an utterance of the following example. (10)

Studenten eine schwierige Jetzt macht dem now makes the.DAT student.DAT a.NOM difficult.NOM GramMAtikfrage Sorgen grammar-question.NOM concerns ‘Now, a difficult grammar question worries the student.’

In order to show how the constraints are in conflict with each other, let us determine them for the two constituents in (10). Again, we bold-face those values that should make their expression go first. (11)

Jetzt now

macht makes

dem Studenten the student object dative definite patiens background

eine schwierige GramMAtikfrage a difficult grammar-question subject nominative indefinite agens focus

Sorgen worries

As we see, the constraints are in conflict with each other in a case like this in which the constraints do not line up perfectly. Hence, every ordering will violate at least some constraints. And if the two variants are equally guilty of that, both linearizations seem to be acceptable, as can be witnessed by the fact that the following example seems to be only a bit worse than the order in (10). (12) ??Jetzt macht eine schwierige GramMAtikfrage dem now makes a.NOM difficult.NOM grammar-question.NOM the.DAT Studenten Sorgen student.DAT concerns ‘Now, a difficult grammar question worries the student.’

Teaching Word Order Variation with a Constraint-Based View on Grammar

57

As mentioned above, in a traditional OT-framework, this should not be possible because, first one constraint should be ranked the highest (for instance, nominative ≺ accusative) and therefore, only one of the two examples in (10) and (12) could fulfill and the other one would be ruled out no matter what. This is because, in classic OT, constraints are absolutely ranked.16 However, for word order variation, a more fine-grained approach is in order. The suggestion that we would like to adopt here is that of a version of OT in which each constraint is assigned a certain weight. The acceptability of a structure then depends on how often and how many constraints it violates and how these constraints are weighted. The core idea is that a sentence gets “negative points”, so to speak, for each violation of a constraint equal to the weight of the constraint. This also enables for the observation that a sentence which violates a bunch of “minor” constraints can be less acceptable than a sentence that violates a single heavy weighted constraint (e.g. Keller 2000). Let us illustrate these ideas again with an example. Let us, for the sake of illustration, assume that we have the following four constraints with the given weight. (13)

a. b. c. d.

NOM: nominative ≺ accusative (4) DEF: definite ≺ indefinite (2) ANIM: animate ≺ inanimate (2) FOC: background ≺ focus (2)

In order to access the so-called “harmony value” (H-value) of an expression (Keller 2006), we simply check which constraints it violates and up the negative values for each violation. Let us apply this to our two examples from (10) and (12). We do this by using a tableau (as it is standard in OT-based analyses) and put an asterisk in each cell that represents a constraint a candidate violates. (14)

candidates C1: (10) C2: (12)

NOM 4 *

DEF 2

ANIM 2

FOC 2

*

*

*

H-value −4 −6

The first candidate, example (10), violates only the NOM-constraint and since the weight of that constraint is 4 (as stipulated for the sake of illustration), the H-value for (12) is −4. The alternative order of example (12) respects the NOM-constraint, but instead violates the other three constraints. Since these each have a stipulated weight of 2, the H-value for (12) is −6. Therefore, this sentence should come out less acceptable than the one in (10).17

16 By

“absolutely ranked” we mean that, if one constraint is ranked over another, any violation of the higher one will “trump” an arbitrary number of violations of lower constraints. Of course, two constraints may be tied. 17 While we stipulated the weights of the constraints here, it should be noted that there are ways to calculate the H-value from experimental data, as Keller (2000) shows.

58

D. Gutzmann and K. Turgay

It should be noted that, just from the H-values, it does not directly follow whether one candidate is unacceptable, only that one candidate is less acceptable than another candidate. But since the two sentences are rather close to each other, they can be both viewed as acceptable.18 This concludes our brief sketch of the (weighted) OT-view on word order variation in German. And we think that students are in principle able to grasp its core ideas. However, starting directly with teaching OT does not sound very promising. Instead, we believe a better way is to get students to think about ideas of constraint conflicts, ordering, and weighting in general – something they are familiar with from many everyday situations – before applying them to the linguistic problem of word order.

3.3 Scenarios for Introducing Constraint Interaction 3.3.1

Who Can Go First in Traffic?

One instance of an everyday life context in which we find constraints in conflict and constraint ordering are the rules that govern traffic. These have been used in a paper by Müller (2002) in order to introduce the core ideas of OT to an audience with some linguistic background. Thinking about a complex traffic situation at an intersection, Müller suggests that the question of who has the right of way and can go first can be thought of as being solved by a set of constraints and their respective ordering. For instance, in the German traffic rules – the so-called “Straßenverkehrsordnung” (StVO) – one constraint is that the car coming from the right can go before a car coming from its left (the “right-before-left” rule). However, this rule can be in conflict with traffic signs that explicitly assign the right of way to a particular street or take it away (like, e.g., stop signs). So, if we think of two constraints here, let us call them RIGHT-LEFT and TRAFFIC-SIGN, it is clear that TRAFFIC-SIGN ranks higher than RIGHT-LEFT. Similar thoughts apply to cases in which there is a traffic light (TRAFFIC-LIGHT) which trumps the other two constraints, and cases in which there is a police officer who regulates the traffic (POLICE-OFFICER), which trumps traffic lights and the other lower ranked constraints. (15)

a. b. c. d.

RIGHT- LEFT: Cars coming from the right have the right of way before cars coming from the left. TRAFFIC - SIGN : Cars whose street has a “right of way” sign, can drive. TRAFFIC - LIGHT: Cars with a green traffic light can drive; cars with red traffic light must not. POLICE - OFFICER : Cars that get right of way from a police officer can drive.

What is crucial here is that these rules can be in conflict with each other. You may come from the left, but have green light. In this case, the higher ranked rule, 18 Compare this to the two examples in (6) and (8), which have H-values of 0 and −10 respectively.

Teaching Word Order Variation with a Constraint-Based View on Grammar

59

TRAFFIC - LIGHT , applies and you can go, but only if there is not a police officer who does tell you otherwise and so on. In principle, Müller’s idea to use traffic rules and their interaction is a good idea to introduce the basic ideas of constraint conflicts and their resolution by constraint ordering. And maybe not so much for younger students, it is a well-known scenario for the older students of grade 11–13, because most of them are going to do their drivers license at that point (the younger students may be familiar with them from using their bike to go to school). However, the problem is that traffic situations should have only one “grammatical” output, the other ones all being ruled out by the traffic rule system. But, as illustrated above, this is not what we need for word order variability, because more than one word order should come out as acceptable, something that would be fatal in traffic. Moreover, with word order, we saw the phenomenon of ganging up: a violation of a number of minor constraints could be worse than the violation of one higher constraint. This we do not find in traffic either: even if you are coming from the right and a traffic sign that says you have the right of way, if a traffic light shows red to you, you must not go. Therefore, Müller’s (2002) scenario, as helpful as it certainly is for introducing the core ideas of the standard OT view, it is not suitable for the problem set at hand. However, it is not so difficult to come up with scenarios for which we can use weighted constraints and which are still representative for everyday scenarios.19 In the following, we will briefly sketch two such scenarios, one for younger students and one for older ones, to just give an idea of what kind of scenarios may be used to get students into thinking about constraints and their interaction.

3.3.2

Where to Sit in the Class Room?

For the younger students, the seating arrangement in the class room is a big deal and it leads to many discussions and adjustments (as most teachers can attest). Therefore, this may be a good scenario to introduce the core ideas of a weighted OT, because when students think about where they want to sit, they often take various considerations – think: constraints – into account, which may not always pull in the same direction. So, for the sake of illustration, let us assume a scenario in which a student, let’s say Anna, is new to the class and has to pick one of several available seats. Which seat she chooses depends on, for instance, if she can sit next to friends, or no boys, or if she can hide away a bit from the teacher’s view or has a good view on the blackboard. Depending on the actual situation, these constraints may be in conflict with each other, and they certainly are not all weighted equally. So let us assume the following constraints and weight.20

19 At

the end of introducing such scenarios to the students, it could be good exercise to let them come up with more contexts in which a weighted constraint approach can be helpful. 20 Of course, these constraints may be more or less realistic, but we think they are relatable enough for the younger students to get the example.

60

D. Gutzmann and K. Turgay

Fig. 1 Choosing-a-seat scenario

(16)

a. b. c. d. e.

I want to sit between two friends. (2 + 2) I want to sit next to my secret crush. (3) GIRLS : I want to sit between two girls. (1 + 1) HIDDEN : I want to be a bit hidden from the teacher. (2) VIEW: I want to have a good view towards the teacher. (1) FRIENDS :

CRUSH :

Note that, strictly speaking, (16-a) and (16-c) are combinations of two constraints, since you can have friends on both sides or only on one (or on neither). Same for (16-c) and girls. However, we combine them here and give them a split weight, so a specific seat can violate these constraints zero, one, or two times and correspondingly would get various negative points. To flesh out the idea of using seating order as an example for discussing weighted constraints, let us assume the class room situation illustrated in Fig. 1 with the indicated available seats, where here F indicates (female) friend, C is Anna’s crush, and B indicates (non-friendly) boys. The black circles represent obstacles, like a column or pillar, that block the view but also hide the seat behind it. Let us use “XY ” as the way to notate the features of a seat, where  are seats available for Anna to choose, and X and Y represent features of the neighboring seats. An overlined box  stands for a seat that is a bit hidden, but which has not a good view, and a box without an overline is not hidden, but has a good view. The four available seats can hence be represented as follows. (17)

F F

F C

BF

CB

Equipped with this, we can now calculate how good these four seats fit Anna’s preferences and which seat is the most optimal one to fulfill them. In order to do that, we assign negative values to the seats based on the constraints that it does not conform to. In case of FRIENDS and GIRL it is possible, that a seat can get the negative value twice, once for each side. If we do this for all constraints and all seats, we get the following tableau with the four “H-values” for each seat.

Teaching Word Order Variation with a Constraint-Based View on Grammar

61

(18) F F F C BF CB

FRIENDS

CRUSH

GIRLS

HIDDEN

VIEW

(2 + 2)

(3) −3

(1 + 1)

(2) −2

(1)

−2 −2 −4

−3

−1 −1 −2

−1 −1 −1

−5 −4 −7 −7

For this, we see that the most optimal seat for Anna is the one in which she sits between a friend and her crush and being hidden (H-value −4), followed closely by the seat between two friends with a good view (H-value −5). The other two seats are worse than those two and receive an H-value of −7. What is crucial here to observe is that the most optimal candidate of the four available seats does still violate some constraints. And in fact, given the constraints at work and the weight we assigned them, it is the best possible seat for Anna to choose. However, the fact that this is the most optimal candidate for Anna to choose, does not mean that, for instance, the F F is out of question for her (it is not “unacceptable”); another facet of the core ideas of the weighted OT view on word order. One interesting exercise for the students and a way to play around with these ideas is to let them change the weight of the preferences and let them see how this effects what the optimal seat for Anna is. For instance, if Anna dislikes boys (except for her crush) and hence the weight of GIRLS is increased to 2 + 2 and her wish to sit next to her crush is not as high and changed to 2 accordingly, then the values of the four seats change accordingly. (19) F F F C F B CB

FRIENDS

CRUSH

GIRLS

HIDDEN

VIEW

(2 + 2)

(2) −2

(2 + 2)

(2) −2

(1)

−2 −2 −4

−2

−2 −2 −4

−1 −1 −1

−4 −5 −7 −9

After this change, sitting between friends is the most optimal seat candidate. Other reassignments of values will lead to corresponding changes in how good the seats fit Anna’s preferences. Letting students change weights and calculate how the H-value of the candidates change may be a good way to get them used to thinking in terms of constraints and weights, which then will prepare them to transfer these concepts from such an everyday situation to the linguistic phenomenon of flexible word order. A quick note on the math: depending what the students were taught in math classes, it could be that they are not yet familiar with negative numbers. In this case, one can just assume that the seat candidates do not get negative values for constraint violations, but get positive points for adhering to them. Then the higher the resulting value for a seat will be, the more optimally it suits Anna’s preferences.

62

3.3.3

D. Gutzmann and K. Turgay

What to Do on Saturday Evening?

For older students, the seating arrangement scenario may not be the best to match their everyday experience at that age and hence it may be not a good way to grab their attention. However, it is easy to switch to a different scenario that suits them more. We suggest here the question of what to do on a Saturday evening. Again, different constraints, each important (= weighted) to a specific degree, have an influence on the Saturday evening plan. Imagine a context, in which a couple, say Alex and Chris, are trying to decide between different activities. Let us assume the following constraints and their respective weights. (20)

a. b. c. d. e.

MONEY:

We do not want to spend any money. (3) We want to have time for the two of us. (2) SLEEP : We want to get enough sleep. (1) FRIENDS : We want to see friends. (2) DANCE : We want to dance. (2) TIME - FOR - TWO :

Now, let us look at different activities and calculate their H-values based on the weights assigned in (20). (21)

MONEY TIME - FOR - TWO SLEEP FRIENDS DANCE

Cinema Movie at home Party at friend’s house Going out for dinner Dancing in the club Bowling

(3) −3

−3 −3 −3

(2)

(1) −1

−2

−1

−2 −2

−1 −1

(2) −2 −2

(2) −2 −2

−2

−2 −2

−8 −4 −3 −7 −6 −8

According to the weights we assigned to the constraints, going to a party at a friend’s house is the most optimal option for Alex and Chris, followed slightly by watching a movie together at home. As with the scenario we suggested for the younger students, the students could play around with the weight numbers or swap out constraints with other closer to their own preferences, to see how the results change thereby getting familiar with the interaction between constraint weighting and preferences. Of course, the two scenarios sketched here are just suggestions for scenarios that may be familiar to the students and therefore be a suitable starting point to encourage them to apply constraint-based thinking, which then can be transferred to the actual topic of word order.

Teaching Word Order Variation with a Constraint-Based View on Grammar

63

4 Word Order Variation in the Curriculum Having shown how the core ideas of a weighted OT-based view can be introduced to the students, which can then be used to actually talk about the constraints that influence word order in German, we will end this contribution by discussing how teaching all this is actually completely justified by the German curriculum and may serve an important role to achieve some of its goals. In general, it has been shown that language awareness and text production skills are closely related Neumann (2007: 82). This is why reflection about language and learning to think about it analytically is, to different extents, part of the curriculum for the lower and higher grades, as we will see. And the development of competency in text production is one of the most prominent aspects in the curriculum. This is mainly due to the fact that texts play a vital aspect in many aspects of social life and many students will be expected to be able to write texts of all kinds in their later life, including reports, protocols, applications, forms, emails and many more (T6). And even though the degree these writing skills will be required in their later jobs may vary from profession to profession, German school education seeks to give students an equal chance to pursue their goals for their later working life (T6). And in contrast to spoken language, written language is not acquired in the same way during language acquisition and the specific text production skills have to be learned separately (Wrobel 2014: 202). All these are reasons why increasing language awareness and developing text productions skills play a major role in the various curricula. And we think that word order is a topic that can be employed to foster them. For this reason, we will now have a look at where exactly this topic will fit into the standards of education for the younger and older students.

4.1 Fifth Grade As we already saw from the brief discussion of Turgay’s (2014) study on how word order is presented in textbooks, word order is a subject for the fifth grade, but in higher classes, it is not addressed again. So let us first have a look at how a more expanded treatment of word order variation based on the theoretical background presented in the previous section fits into the curriculum for German classes in the fifth grade. The official, federal view on the curricula for the different school forms and grades does not really specify topics that students should learn, but instead provides more general competencies that the students should acquire and which are sorted into different spheres of competency (“Kompetenzbereiche”). For instance, in the standards of education in the subject German for the intermediate school grade (“Bildungsstandards im Fach Deutsch für den Mittleren Schulabschluss”, KMK 2003), we find two spheres of competency that are relevant for our purposes and

64

D. Gutzmann and K. Turgay

into which the topic of word order variation fits. The first sphere is ANALYZING LANGUAGE AND LANGUAGE USE (“Sprache und Sprachgebrauch untersuchen”, KMK 2003: 8f), which, as a subcompetency includes thinking about language use and understanding it as a system (“über Verwendung von Sprache nachdenken und sie als System verstehen”). This aims at enabling students to analyze and produce texts according to functional, normative and, where applicable, aesthetical aspects. With respect to the topic area of “language as a system”, students are primarily supposed to look at grammatical phenomena and their functions and use them for text production and refinement. Moreover, the students are expected to be become able to correctly produce grammatical structures and correct them on demand. These last two aspects, text production and correction, also fall into the other sphere of competence relevant for this contribution, namely WRITING (“Schreiben” KMK 2003: 11f), which, as a subcompetency includes WRITING TEXTS . Here, the national standards state that students should be able to shape their text sophisticatedly and use linguistic means purposefully and deliberately. They should produce structures, comprehensibly, linguistically variable, and stylistically appropriate for their message. As a more concrete example, we looked at the concrete curriculum of the federal state Baden-Württemberg (“Bildungsplan 2016 Deutsch”, KMBW 2016). It is explicitly stated there stated that students should be able to analyze (and describe) the structure of simple sentences using the topological field model (including the notions of pre, middle, and post field as well as the two sentence brackets) and can determine the syntactic function of constituents (KMBW 2016: 32). All this shows that teaching word order is not only a good fit for what is expected from the national standards of education but is also explicitly expected, at least by the newer federal curriculum of Baden-Württemberg. This gives rise to the hope that the topic will be incorporated into the official education plans of other states as well and that it will also be a more important topic in future textbooks. Therefore, we think that putting the OT-based view sketched here to this phenomenon is a good way to approach this in the class room. Our didactical expectation for teaching word order by means of a weighted OTbased approach is that the students switch from a prescriptive perspective (which is somewhat reflected by the non-variable way in which SVO order is depicted in some of the textbooks) to a more descriptive view on grammar with respect to the word order of subject, predicate and objects. They are supposed to develop an awareness for word order variations and break up the dichotomy of grammatical vs. ungrammatical and switch to a gradual view of acceptability. Another aim, which goes beyond our scenario for teaching the constraint-based view, is to highlight the function of word order variations for improving their own writing, for example by permutation tests. Moreover, students get in contact with linguistic model building and, using the OT-based view, they can connect it to other non-linguistic problems of their everyday life. This directly can connect to the topological field model which is canonically used in many textbooks for the fifth grade. As suggested in the previous section, the students can get access to the rather abstract ideas of (weighted) OT and their linguistic application by introducing them via the seating arrangement

Teaching Word Order Variation with a Constraint-Based View on Grammar

65

scenario. This is best be done as a operational procedure, for instance by letting students vary and calculate their own variations of the original example, which fits the cognitive capacities of fifth graders.

4.2 Grade 11–13 As we already saw, word order is a topic in textbooks for the fifth grade and hence it is not really surprising that this topic fits into the official expectation (both national and federal). However, we noted that word order is completely absent from textbooks for higher grades. So let us have a look at the corresponding guidelines and standards for the “Sekundarstufe II”21 in order to see whether teaching word order variation and the core ideas of OT a second time. Again, we found a relevant sphere of competence in the corresponding standards of education in the subject German for the higher education entrance qualification (“Bildungsstandards im Fach Deutsch für die Allgemeine Hochschulreife”): REFLECTING ON LANGUAGE AND LANGUAGE USE (“Sprache und Sprachgebrauch reflektieren” KMK 2012: 20). There it is stated that students can analyze language as a system and, by doing so, extend their linguistic knowledge and awareness. At the so-called basic level, they should be able to analyze the conditions of successful communication; also on the basis of theoretical models. Teaching such models may be beneficial to students, because it helps increase their language awareness, which in turn has been proven to be related to their text production skills (Neumann 2007: 82). Of course, this is where our idea to teach word order using a constraint and competition-based view directly fits in. If, as outlined in the previous subsection, the core ideas of this concept are already established in the fifth grade, the older students should already have the foundational knowledge and the basic ideas and concepts and their awareness for word order variations simply have to be reactivated, as they already know word order as a gradient phenomenon. Building on this, the Saturday evening planning scenario sketched above – which fits their everyday life and may catch their interests – can be used to get the students back into thinking about constraints and their interaction and ordering via weights. This time, the idea of there being a linguistic theory behind all this can be addressed more explicitly and the analysis of word order variation by using such a theory can be used to introduce students to the core idea of linguistic modeling and the aims and development of linguistic theories. In addition, the information structural role of notions like focus and topic should be discussed in this context in order to show the students how word order variation can help them to better shape their text, “also on the basis of theoretical models” (KMK 2012: 20).

21 This

includes grade 11 to 12 or 13 of the upper schools, which must be completed to achieve the higher education entrance qualification (“Allgemeine Hochschulreife”).

66

D. Gutzmann and K. Turgay

5 Conclusion In this contribution, we suggested that the phenomenon of flexible word order in German should be taught in German classes, both to younger and older students. In contrast to the fact that word order is (still) represented in a rather rigid way in many textbooks, we think it would be beneficial to introduce students to the idea of flexible word order explicitly and make them aware of its function to structure their texts in a coherent way. The theoretical background which we think can – to a certain extent based on the age of the students – be put to use in order to achieve this is (a weighted version of) optimality theory. Having sketched the core ideas of this theoretical framework, we present two scenarios that may be used to get students familiar with constraint-based thinking and how the weighting of constraints influences the respective optimality of the candidates. These skills can then be transferred to word order so that students get first experience with linguistic modeling, to different degrees of explicitness depending on the age and knowledge of the students. We closed this article by examining how and where these topics can be anchored within the nationally and federally expected standards of education and showed that they are indeed a very good fit for both lessons in the fifth grade as well as for older students. We hope that this contribution has drawn an interesting connection between current linguistic theories and practical issues of language didactics. Of course, what we presented here are just ideas and suggestions and if they are actually viable for real teaching efforts must be proven in practice. Hence, an obvious next step for further research would be to set up an experiment to study if an approach as presented here can increase both the students’ awareness of the possibility and functions of word order variations in German and their ability to actively employ them to refine their writings. And, of course, in order to do that, a more explicit plan for structuring a sequence of lessons should be developed first, ideally together with actual teachers that will then put them to use in the class room.

References Abraham, W. (1985). Wortstellung und das Mittelfeld im Deutschen. In W. Abraham (Ed.), Wortstellung und das Mittelfeld im Deutschen (pp. 27–53). Tübingen: Narr. Audehm, B. (2006). Wort- und Satzgliedstellung im Mittelfeld des gesprochenen Deutsch. Eine korpusbasierte Untersuchung. Ph.D. thesis. Universität Hannover. Bobaljik, J., & Wurmbrand, S. (2012). Word order and scope. Transparent interfaces and the 3/4 signature. Linguistic Inquiry, 43(3), 371–421. https://doi.org/10.1162/LING_a_00094. Büring, D. (2001). Let’s phrase it!—Focus, word order, and prosodic phrasing in German double object constructions. In G. Müller & W. Sternefeld (Eds.), Competition in syntax. (Studies in genitive grammar, Vol. 49, pp. 101–137). Berlin/New York: de Gruyter. Fanselow, G. (2006). On pure syntax (uncontaminated by information structure). In P. Brandt & E. Fuß (Eds.), On pure syntax (uncontaminated by information structure). A Festschrift Presented to Günther Grewendorf on Occasion of His 60th Birthday (pp. 137–157). Berlin: Akademie Verlag.

Teaching Word Order Variation with a Constraint-Based View on Grammar

67

Featherston, S. (2005). The decathlon model of empirical syntax. In Linguistic evidence. Empirical, theoretical and computational perspectives (pp. 187–208). https://doi.org/10.1515/ 9783110197549.187. Frey, W. (2004). A medial topic position for German. Linguistische Berichte, 198, 153–190. Grimshaw, J. (1997). Projection, heads, and optimality. Linguistic Inquiry, 28, 373–422. Hawkins, J. A. (1994). A performance theory of word order and constituency. Cambridge: Cambridge University Press. Jacobs, J. (1988). Probleme der freien Wortstellung im Deutschen. Sprache und Pragmatik, 5, 8–37. Keller, F. (2000). Gradience in grammar. Experimental and computational aspects of degrees of grammaticality. Ph.D. thesis. Edinburgh: University of Edinburgh. http://homepages.inf.ed.ac. uk/keller/papers/phd.pdf. Keller, F. (2006). Linear optimality theory as a model of gradience in grammar. In G. Fanselow, C. Féry, R. Vogel, & M. Schlesewsky (Eds.), Gradience in grammar (pp. 270–287). Oxford: Oxford University Press. KMBW. (Ed.). (2016). Bildungsplan 2016. Deutsch. Ministerium für Kultur, Jugend und Sport, Baden-Württemberg. KMK. (Ed.). (2003). Bildungsstandards im Fach Deutsch für den Mittleren Schulabschluss. Beschlüsse der Kultusministerkonferenz. https://www.kmk.org/fileadmin/ veroeffentlichungen_beschluesse/2003/2003_12_04-BS-Deutsch-MS.pdf. KMK. (Ed.). (2012). Bildungsstandards im Fach Deutsch für die Allgemeine Hochschulreife. Beschlüsse der Kultusministerkonferenz. https://www.kmk.org/fileadmin/ veroeffentlichungen_beschluesse/2012/2012_10_18-Bildungsstandards-Deutsch-Abi.pdf. Lenerz, J. (1977). Zur Abfolge nominaler Satzglieder im Deutschen. Tübingen: Narr. Lenerz, J. (1981). Zum gegenwärtigen Stand der Wortstellungsforschung. Beiträge zur Geschichte der deutschen Sprache und Literatur, 103(1), 6–30. McCarthy, J., & Prince, A. (1994). The emergence of the unmarked: Optimality in prosodic morphology. Proceedings of NELS, 24, 333–379. Meinunger, A. (2000). Syntactic aspects of topic and comment. Amsterdam: John Benjamins. Molnárfi, L. (2004). On scrambling as defocusing in German and West Germanic. In A. Breitbarth & H. van Riemsdijk (Eds.), Triggers (pp. 331–386). Berlin: de Gruyter. https://doi.org/10.1515/ 9783110197433. Müller, G. (1999). Optimality, markedness, and word order in German. Linguistics, 37(5), 777– 818. https://doi.org/10.1515/ling.37.5.777. Müller, G. (2000). Elemente der optimalitätstheoretischen syntax. Tübingen: Stauffenburg. Müller, G. (2002). Verletzbare Regeln in Straßenverkehr und Syntax. Sprachreport, 3, 11–18. Neumann, A. (2007). Schreiben: Ausgangspunkt für eine kriteriengeleitete Ausbildung in der Schule. In H. Willenberg (Ed.), Kompetenzhandbuch für den Deutschunterricht (pp. 74–83). Baltmannsweiler: Schneider Hohengehren. Pafel, J. (2009). Zur linearen Syntax des deutschen Satzes. Linguistische Berichte, 217, 37–79. Pesetsky, D. (1997). Optimality theory and syntax: Movement and pronunciation. In D. Archangeli & D. Terence Langendoen (Eds.), Optimality theory. An overview (pp. 134–170). Oxford: Blackwell. Prince, A., & Smolensky, P. (1993). Optimality theory. Constraint interaction in generative grammar. Referenz auf die 2002er ROA-Version. https://doi.org/10.7282/T34M92MV. Skopeteas, S., & Fanselow, G. (2010). Focus types and argument asymmetries. A cross-linguistic study in language production. In C. Breul & E. Göbbel (Eds.), Contrastive information structure (pp. 169–197). Amsterdam: John Benjamins. Struckmeier, V. (2014). Prosodische, semantische und syntaktische Faktoren der deutschen Wortstellung. Prosodische, semantische und syntaktische Faktoren der deutschen Wortstellung. Berlin: Akademie Verlag. Turgay, K. (2014). Zur Variabilität der Wortstellung im Mittelfeld in Empirie und Unterricht. Zeitschrift für Angewandte Linguistik, 61(1). https://doi.org/10.1515/zfal-2014-0018.

68

D. Gutzmann and K. Turgay

Turgay, K. (2017). Frequenz vs. Akzeptabilität. Medialitätsabhängige Abfolgetendenzen im Mittelfeld. Trier: WVT. Wrobel, A. (2014). Schreibkompetenz und Schreibprozess. In H. Feilke & T. Pohl (Eds.), Schriftlicher Sprachgebrauch. Texte verfassen (pp. 85–100). Baltmannsweiler: Schneider Hohengehren. Wurmbrand, S. (2008). Word order and scope in German. Groninger Arbeiten zur germanistischen Linguistik, 46, 89–110. Zeman, J. (2002). Die deutsche Wortstellung. Wien: Edition Praesens.

Grammar Is Irrelevant – The Role of Epistemological Beliefs in Students’ Learning Success Daniela Elsner

Abstract Studies in the field of German linguistics reveal that the grammar proficiency of teacher trainees studying German is rather poor. Since research in educational sciences shows that epistemological beliefs play a major role in the learning success we want to test, (i) whether there is a positive correlation between epistemological beliefs and grammatical knowledge, and (ii) whether certain epistemological beliefs go along with a more sophisticated self-concept, more elaborate learning strategies as well as a higher motivation to deal with grammar. In this paper we report on the results of a questionnaire that investigates the relations between epistemological beliefs and the other variables of 201 students of German. Our results show that a poorer performance in the grammar test correlates with a lower motivation, a worse self-concept, and the use of repetitive learning strategies. Assuming that motivation is one of the key factors triggering learning processes, we conclude that it is necessary to review the academic curriculum and to implement different teaching methods in order to enhance the students’ motivation to deal with grammar. Keywords Epistemological beliefs · Grammatical content knowledge · Professional teacher competencies · Competence model · Motivation · Self-concept · Learning strategies

1 Introduction Research in educational sciences suggests that many teacher trainees consider knowledge of the academic subject matter that is taught at university classes to be irrelevant for their future occupation (cf. Hoppe-Graff et al. 2009). The following

D. Elsner () Institut für Sekundarbildung und Fachdidaktik, Pädagogische Hochschule Vorarlberg, Feldkirch, Austria e-mail: [email protected] © Springer Nature Switzerland AG 2020 A. Trotzke, T. Kupisch (eds.), Formal Linguistics and Language Education, Educational Linguistics 43, https://doi.org/10.1007/978-3-030-39257-4_5

69

70

D. Elsner

examples from evaluations of university grammar classes for prospective teachers point in the same direction: I could not benefit from this class a lot. I think the discussions were too detailed. Question: Why haven’t your expectations been measured up to? Answer: Because it was a linguistics class, I want to become a teacher’ ‘It is unlikely that I will be confronted with this subject matter in my professional life because I will become a teacher1

Additionally, large scale studies like TEDS-LT2 in Germany show that on completion of their studies, teacher trainees do not know more about grammar than when they began (cf. Bremerich-Vos and Dämmer 2013). Schäfer and Sayatz (2017) assess the grammar proficiency of students studying German at different stages of their studies. Though they find that students have a solid knowledge of grammar, they point out that studying German linguistics does not significantly improve the students’ grammar knowledge. How can this situation be accounted for? In this article we present the results of a survey that tests whether grammatical knowledge of German correlates with certain epistemological beliefs. If we understand learning not as a purely cognitive process, we can ask what impact conative and affective factors might have on the acquisition of knowledge. Learning success does not follow automatically from extensive learning and individual capabilities, but is also influenced by motivation and volition. Greenberg and Baron (2008: 248) understand motivation as a “set of processes that arouse, direct, and maintain human behaviour toward attaining a goal”. So in order to enrich their knowledge of grammar, students need intrinsic and/or extrinsic motivation as well as the volition to reach this goal. In addition, Baumert and Kunter (2006) point out that epistemological beliefs (i.e. subjective conceptions about the acquisition, structure, and validation of knowledge) play a major role in the learning success since they act as a filter for the processing of knowledge. Via motivation, learning strategies, and beliefs about learning and education, students’ epistemological beliefs affect their performance in the classroom (cf. Urhahne 2006; Hofer 2001). For mathematics, Dubberke et al. (2008) demonstrate a relation between the teachers’ beliefs and the pupils’ learning success. For domains as different as politics and biology, research has shown that there is a relation between epistemological beliefs and other relevant variables such as content knowledge, motivation, self concept (understood as awareness of one’s own competencies), and learning strategies (cf. Baumert et al. 2000, Hofer 2000, Urhahne 2006, Trautwein and Lüdtke 2008, Weschenfelder 2014). Neither the survey of epistemological beliefs nor the testing of relations between epistemological beliefs and other variables have been the focus of research in the

1 Translated

quotations from a feedback questionnaire of the class Morphologie (‘morphology’), WS 2015/16, University of Leipzig. 2 The Teacher Education and Development Study: Learning to Teach aims at identifying the development of the professional competencies of prospective teachers of German, English, and Mathematics.

Grammar Is Irrelevant – The Role of Epistemological Beliefs in Students’. . .

71

area of grammar education. Assuming that beliefs function as a predictor for other variables, as has been argued above, the following two hypotheses can be established against the background of existing research: (i) there is a positive correlation between epistemological beliefs and grammatical knowledge, and (ii) certain epistemological beliefs go along with a more sophisticated self-concept, more elaborate learning strategies as well as a higher motivation to deal with grammar. In this paper we want to test these two hypotheses by collecting and correlating data on grammatical knowledge and epistemological beliefs, as well as data on the students’ self-concept, learning strategies, and motivation to deal with grammar. To this end, a questionnaire was developed and will be elaborated on in Sect. 3. Since the author of this article teaches German linguistics, the language in focus will be German. However, in our opinion, the survey can be fruitfully adapted to other languages as well and thus enrich the research on the relation between epistemological beliefs and grammatical knowledge. Independent of the language, investigating students’ epistemological beliefs and the relations between beliefs and other variables might lead to a better understanding of learning and in the long term it might enhance teaching effectiveness at the university (cf. Hofer 2002). Beliefs and content knowledge are part of professional teachers’ competencies and need to be distinguished from other competencies like motivation and selfregulation. Based on the work of Baumert and Kunter (2006), we will first introduce a domain-specific competence model (Sect. 2), focusing on the modelling of professional content knowledge (Sect. 2.1) and epistemological beliefs (Sect. 2.2). In Sect. 3 we present and discuss the results of a questionnaire which surveys the grammatical content knowledge of 201 students of German, as well as the developmental level of their epistemological beliefs (based on the FREE3 questionnaire by Krettenauer 2005), their self-concept, their interest in grammar (based on scales by Köller et al. 2000), and their implemented learning strategies (based on scales by Kunter et al. 2002). Since our results (Sect. 3) show no correlation between the students’ epistemological beliefs and the other tested variables, we will discuss in Sect. 4 how this might be accounted for by methodological issues. Nonetheless, our data show that a poorer performance in the grammar test correlates with a lower motivation, a worse self-concept, and the use of repetitive learning strategies. Assuming that motivation is one of the key factors triggering learning processes (cf. Kunter and Trautwein 2013: 43), we conclude in line with Schäfer and Sayatz (2017) that it is necessary to review the academic curriculum and to implement different teaching methods in order to enhance students’ motivation to deal with grammar.

3 German: Fragebogen zur Erfassung des Entwicklungsniveuas epistemologischer Überzeugungen (questionnaire for measuring the level of epistemological beliefs).

72

D. Elsner

2 A Domain-Specific Competence Model 2.1 Grammatical Content Knowledge As a first step, we need to introduce a domain-specific competence model in order to determine the relevant competencies that are essential for successfully meeting the demands of teaching in a classroom (cf. Baumert and Kunter 2011). In contrast to other subjects, language teaching is special insofar as it is composed of different subdomains which also differ depending on whether a language is taught as a foreign language or whether the native language is the focus of instruction. Whereas in a foreign language class aspects like vocabulary, grammar, communication and culture play a major role, in native language classes the focus usually lies on literature and grammar. At this point we do not want to discuss why it is useful to learn about grammatical structures of one’s native language (see e.g. Heringer 1995), but since each subdomain differs in terms of the content knowledge to be acquired (e.g. grammatical content knowledge differs significantly from literary content knowledge), it is necessary to develop a domain-specific competence model tailored to grammar. Because the language in focus in this article is German, the following explanations will only refer to German grammar. In the context of COACTIV,4 a study which investigates teacher competence as a key determinant of instructional quality in mathematics, Baumert and Kunter (2006, 2011) develop a generic model that discerns four aspects of professional competence: professional knowledge (‘Professionswissen’), motivation (‘Motivationale Orientierungen’), beliefs, and self-regulation. Each aspect can be further differentiated, first into a number of spheres of knowledge (‘Wissensbereiche’) and next into a number of facets of knowledge (‘Wissensfacetten’) (Fig. 1). There is no consensus concerning the modelling of professional grammatical content knowledge. According to Shulman (1986), professional knowledge is divided into content knowledge, didactical knowledge, and pedagogical knowledge. The COACTIV model expands this division by organizational knowledge as well as counselling knowledge. Content knowledge constitutes the foundation for didactical knowledge, which among other things centers on defining the content of what should be taught in the classroom as well as on the question of how this content should be taught. Pedagogical knowledge encompasses the arrangement of the learning environment as well as knowledge about learning processes and performance assessment (Baumert and Kunter 2011: 32). Organizational and counselling knowledge are mostly independent of specific domains. The former comprises knowledge about the educational system and the educational institutions, whereas the latter concerns knowledge about negotiating with pupils and their parents (cf. Baumert and Kunter 2011: 40). Since in this article we are focusing on content knowledge as a dependent

4 Professional

Competence of Teachers, Cognitively Activating Instruction, and Development of Students’ Mathematical Literacy.

Grammar Is Irrelevant – The Role of Epistemological Beliefs in Students’. . .

73

Fig. 1 Generic model of professional competencies (focus: professional knowledge) (Baumert and Kunter 2006: 482)

variable, we will concentrate on modelling this sphere with respect to German grammar. It is not clear how the facets of grammatical content knowledge can be modelled and how much content knowledge a prospective teacher needs to have (cf. Pissarek and Schilcher 2017: 71, 80). Put differently, there is no consensus on what teacher trainees need to know about grammar for their future occupation. Moreover, the content of a grammar class at the elementary and secondary school levels differs substantially from the content of a university grammar class and students might have a different opinion of what grammar is than university lecturers. This could cause the students’ belief that the content of university grammar classes is irrelevant for their future occupation (see Sect. 1). In the following, we therefore briefly discuss the differences between school and university grammar. At schools in Germany, native language grammar plays a major role, especially in grades 5 and 6, and from a scientific point of view, the content that is taught then is mostly descriptive in nature and serves as a tool for the description of language structures. Grammar taught in school also often differs from university grammar in terms of the adequacy of criteria used for generating categories. For instance, at school, word classes are usually based on semantic criteria, which (i) do not play a role in German linguistics, and (ii) do not allow a thorough coverage of the matter (cf. Bredel 2013; Elsner 2019). Describing verbs as words that express activities is not adequate since this characteristic does not apply to words like to have, to own, or to be. Additionally, words from other word classes like explosion or sleep

74

D. Elsner

could mistakenly be categorized as verbs when mainly relying on semantic criteria. The aim of teaching grammar at school is on the one hand the acquisition of a meta-language for the description of linguistic structures and on the other hand the development of knowledge about the function of different syntactic categories, which should be implemented when revising or writing texts. Grammar in university classes requires school grammar but builds on it only rudimentarily. Depending on the theoretical framework, a considerable degree of abstraction takes place as well as a formalization of language. When studying German, linguistics is an independent domain beside other domains like contemporary literature. In contrast to teaching at schools, grammar at the university is not a means to an end (e.g. for the revision of texts), but is pursued as an end in itself. The ultimate goal is to discuss and develop models of linguistic phenomena within a theoretical framework. An example of such linguistic phenomena is ergative verbs, i.e. intransitive verbs like sterben ‘to die’ whose subjects show characteristics of direct objects. Besides describing the characteristics of ergative constructions, linguists attempt to explain why ergative subjects show syntactic characteristics of objects, but on the surface they appear as nominative NPs or subjects respectively (cf. Grewendorf 1989). Depending on the syntactic theory, different explanations can be brought forward. So the path from school grammar to university grammar is rather long and one can assume that students often do not see the relationship between the two. Against this background, it is not surprising that teacher students commonly consider grammar at the university to be irrelevant. We find a very similar situation in the domain of mathematics: students experience the transition from school to university as difficult and prospective teachers do not see the relevance of university mathematics for their future occupation (cf. Dreher et al. 2018). Empirical research within the domain of mathematics proves that in addition to academic mathematical content knowledge and knowledge about didactics, there is a third dimension called school-related content knowledge (Heinze et al. 2016a, 2016b). In terms of content, it comprises knowledge about mathematics at school and curricular knowledge as well as knowledge about the relationship between mathematics in academia and mathematics at school. If such a construct can be verified for the domain of grammar, it might provide an answer to the question of what teacher trainees of German need to know about grammar and how to develop more relevant grammar classes for these students. To date there has not been any research investigating possible separate dimensions of grammatical content knowledge and no domain-specific competence model has been developed for German grammar yet. However, there are proposals in the literature concerning what teacher trainees should know about German grammar: Eisenberg (2004: 5) claims that prospective teachers should have extensive knowledge about prototypical linguistic structures. Boettcher (1999: 247) is vague when postulating that students need to obtain a minimum of grammatical knowledge or orientational grammatical knowledge (“grammatisches Mindest- oder Orientierungswissen”) without further explanations. The Standing Conference of the Ministers of Education and Cultural Affairs of the federal states of Germany

Grammar Is Irrelevant – The Role of Epistemological Beliefs in Students’. . .

75

is among other things responsible for working towards securing quality standards at school and the university. In 2008 they issued a report summarizing for the first time which competencies prospective teachers should attain at the university. For the field of German grammar the report states that teacher trainees need to acquire basic knowledge of phonology, morphology, syntax, semantics, pragmatics, and text linguistics (KMK 2018: 27). These specifications are not very detailed but we can infer that teacher trainees need to know terms and methods for the description of linguistic structures. Hence prospective teachers should possess consolidated knowledge of school grammar and should be able to handle an inventory of linguistic notions in order to describe linguistic structures. However, this would have the dissatisfying result of putting school and university grammar on an equal footing. Additionally, it would then be unclear which contents should be relevant at the university level, since academic classes would only repeat what has been done at school. Evidence of a third knowledge dimension in addition to academic content knowledge and knowledge about didactics is a desideratum for the domain of grammar but that is beyond the scope of this paper. For the questionnaire, we therefore resort to conceptualizing grammatical content knowledge as consolidated knowledge of school grammar (see Pissarek and Schilcher (2017) for a similar approach).

2.2 Epistemological Beliefs Beliefs are part of the professional competencies of teachers and it can be assumed that they play a major role in predicting other variables, such as content knowledge or learning strategies (cf. Op’t Eynde et al. 2002). In general, they can be understood as implicit or explicit subjective conceptualizations that affect one’s actions (cf. Op’t Eynde et al. 2002, Baumert and Kunter 2011). The boundary between beliefs and knowledge is blurred. At the core beliefs are subjective individual constructs that are taken for granted by individuals but not necessarily by a greater community (cf. Op’t Eynde et al. 2002: 24). Knowledge, on the other hand, is a social construct that is not specific to individuals and has to meet certain criteria like consistency, justification and discursive validation (cf. Baumert and Kunter 2011: 41, Weschenfelder 2014: 80f.). Beliefs can be conceptualized in different ways. Op’t Eynde et al. (2002: 17) assess a large number of articles discussing beliefs about mathematics. They identify four topics that these articles deal with: epistemological beliefs, beliefs about teaching and learning mathematics, beliefs about the self in the context of mathematics learning and problem solving. In the following, we will concentrate only on epistemological beliefs since they are related to content knowledge. Section 2.2.1 elaborates on epistemological beliefs in general and introduces a model of epistemological beliefs developed by Hofer and Pintrich 1997). Section 2.2.2 discusses how epistemological beliefs of German grammar can be modelled.

76

2.2.1

D. Elsner

A General Model of Epistemological Beliefs

Epistemological beliefs are subjective conceptions of the structure, acquisition, and validation of knowledge (Baumert and Kunter 2006, 2011; Weschenfelder 2014). They influence our actions and thinking and therefore offer a vantage point for answering the questions of why teacher trainees might consider the content of university grammar classes to be irrelevant and why there is no significant progress in learning. Jehng et al. (1993) show that epistemological beliefs are dependent on the academic field: students in the ‘soft’ fields (e.g. arts, humanities) have a stronger tendency to believe that knowledge is uncertain whereas students in the ‘hard’ fields (e.g. mathematics, engineering) mostly believe that knowledge is absolute. Similar correlations have been pointed out by Trautwein et al. (2004). Most models conceptualizing epistemological beliefs can be traced back to Perry (1970), who outlines a developmental model of intellectual progression of college students which “postulates an ongoing qualitative reorganization of the making of meaning” (Hofer and Pintrich 1997: 91). The model emanates from a category Perry named dualism (= absolutistic, right-and-wrong view of the world) and via nine positions - ending in a category named commitment within relativism (relativism = understanding of knowledge as relative, contingent, and contextual). Perry’s model was criticized because of the biased selection of subjects (mainly male Harvard students), and because of difficulties concerning the measurement of the single levels (Hofer and Pintrich 1997: 93). Additionally, the model is one-dimensional and therefore relations between knowledge and learning cannot be investigated. Hofer and Pintrich (1997: 113ff.) advance a multidimensional view; they analyze six models of epistemological beliefs and cluster the different dimensions into two superordinate core categories: nature of knowledge (i.e. what one believes knowledge is) and nature of knowing (i.e. how one comes to know, cf. Hofer 2000: 380). Each category in turn consists of two dimensions (Fig. 2): The multidimensionality allows for the investigation of relations between the single dimensions and makes it possible for individuals to exhibit different values for the different dimensions (Hofer 2001: 359f.). The single dimensions can be understood as continua so that developments between the extremities are possible. The extremities typically correspond to a naïve and a mature belief (cf. Urhahne 2006: 190). In the following, we briefly explain the naïve and mature positions (i.e.

epistemological beliefs

nature of knowing

nature of knowledge certainty of knowledge

simplicity of knowledge

source of knowledge

justification for knowing

Fig. 2 Multidimensional view of epistemological beliefs (Hofer and Pintrich 1997)

Grammar Is Irrelevant – The Role of Epistemological Beliefs in Students’. . .

77

the extremities) of each of the four dimensions. The reader should keep in mind that this strongly reduces the complexity of the model and that there are numerous intermediate positions between the two extremities. By certainty of knowledge, Hofer and Pintrich (1997) refer to the fact that knowledge can be apprehended as a fixed/absolute (naïve position) or as a variable/relativistic (mature position) construct. Simplicity of knowledge relates to the fact that knowledge can be understood as an accumulation of facts (naïve position) or as a complex network (mature position). The source of knowledge can be an individual who conveys knowledge to another individual (naïve position) or the source can be the individual him- or herself, understood as the creator of knowledge (mature position). Justification for knowing refers to the evaluation of arguments and to the question of whether individuals make use of expertise. Typically, a naïve believer justifies opinions “on the basis of what feels right” (cf. Hofer 2000: 381). A mature believer integrates the views of experts and the results of scientific research when judging beliefs. For different domains correlations concerning the single dimensions have been detected: Baumert et al. (2000: 262) point out for mathematics that a static understanding of the subject correlates positively with memorizing techniques. Individuals with relativistic beliefs on the other hand are more interested in mathematics. Ryan (1984) states that college students with a dualistic level of beliefs indicate comprehension when they can replicate the facts (remembering) whereas students with a non-dualistic level indicate comprehension when they can apply their knowledge in new situations. Additionally, learning strategies correlate positively with students’ performance – more elaborate learning strategies go along with better marks. Likewise Schommer (1993) shows that beliefs about the source of knowledge are predictive for the students’ grade point average. The lower the assumption that learning is a fast process, the better the grade point average. Urhahne (2006) states a correlation between epistemological beliefs and other variables. More mature beliefs go along with a higher motivation, a better selfconcept, and more elaborate learning strategies.

2.2.2

Towards Modelling Epistemological Beliefs of German Grammar

As yet there is no proposal for the modelling of domain-specific facets of epistemological beliefs for our subject of interest, i.e. German grammar. In the following, we will develop initial ideas that can be understood as a basis for further conceptualizations and discussions. In line with Krettenauer (2005), three levels of certainty of knowledge can be differentiated. A naïve (absolutistic) view assumes that there is always one correct solution to each and every problem. Applied to grammar, individuals with an absolutistic level of beliefs would assume that grammatical knowledge is invariable and that an explanation and/or categorization is either true or false. For example, with regard to word classes it would be assumed that each and every lexeme can always be allocated to one correct word class. A relativistic level acknowledges different opinions but traces them back to subjective perspectives:

78

D. Elsner

basically, every opinion is equally good and the selection of one of these opinions is a purely subjective decision. For example, a lexeme like the German quitt (‘even’) can be categorized as an adjective or an adverb depending on the criteria employed, but these criteria and the argumentation are not assessed and therefore it does not matter which word class is chosen. A postrelativistic level accepts the possibility of different opinions as well, but also evaluates the arguments and based on this evaluation a certain opinion is selected. Since quitt cannot be inflected (1) and cannot appear between a determiner and a noun (2), we could analyze it as an adverb. However, adverbs can always function as adverbials, but quitt can only be used predicatively (3). Additionally, there are other adjectives as well that cannot be inflected (4) (e.g. super ‘great’). 1. 2. 3. 4.

*die quitten the.MASC.PL. even.MASC.PL. *die quitt the.MASC.PL. even Die Freunde sind the friends are *der supere the.MASC.SG. great.MASC.SG.

Freunde friends.MASC.PL. Freunde friends.MASC.PL. quitt. even. Film film.MASC.SG.

All in all, the adjective analysis seems to account for more facts and can be seen as the ‘better’ alternative compared to the adverb analysis. The dimension simplicity of knowledge oscillates between a static perspective of understanding grammar as an accumulation of facts and rules and a dynamic perspective understanding grammar as a complex network. A simple listing of possibilities of plural formations can be seen as a static perspective on the German plural system. A dynamic perspective would go further in trying to model the organization or rather the structure of the plural system and would try to find and explain generalizations like the complementary distribution of the German plural suffixes -n and -en. As source of knowledge, both the teacher as well as the learner can be taken into consideration. A naïve understanding assumes that grammatical knowledge is passed on from the teacher to the learner – then the source is outside of the individual. Increase of knowledge is in this case a consequence of repetitive practice and limited to the acquisition of context free knowledge (cf. Weschenfelder 2014: 114). For example, it is typically stated in school books that verbs are ‘tu-Wörter’ (‘do-words’) and that they can be identified by asking a certain question (Was tut X? ‘What is X doing?’). The identification of verbs is then practiced by the mindless application of the question test (cf. Elsner 2019). A deep understanding of which role verbs play in a sentence fails to be acquired by learners (cf. Mesch and Dammert 2015). A mature understanding assumes that individuals can construe grammatical knowledge themselves when presented with specific environments. For example, subject-predicate agreement in German clauses can be detected by learners

Grammar Is Irrelevant – The Role of Epistemological Beliefs in Students’. . .

79

comparing different sentences, or characteristics of adjectives can be discovered by using and comparing adjectives in different syntactic environments. Regarding justification for knowing, a naïve position would be to resort to one’s own belief of what is right. Linguistic cases of doubt (for example case alterations in PPs: [wegen [dem Wetter]NP-Dative ]PP vs. [wegen [des Wetters]NP-Genitive ]PP , ‘due to the weather’) often trigger the demand from language users to know which of the two options is the right one. Naïve believers would just follow their own feeling, whereas a more mature believer would integrate experts’ views (for example by using a reference work or consulting certain web pages). In this section we introduced a competence model and discussed the dimensions of professional content knowledge as well as epistemological beliefs domainspecifically. Grammatical content knowledge is conceptualized as consolidated knowledge of school grammar. With regard to epistemological beliefs the questionnaire (see next section) will concentrate on the dimension of certainty of knowledge, differentiating between an absolutistic, a relativistic, and a post-relativistic position.

3 Grammatical Content Knowledge, Beliefs, and Other Variables It was stated in Sect. 1 that prospective teachers of German often regard grammar as irrelevant and that there is hardly any learning progress during the course of their studies. Research in educational sciences has proven that epistemological beliefs function as a predictor for other variables (c.f. Sect. 2.2.1). Based on existing research in other domains than grammar we stated the following hypotheses in Sect. 1: (i) there is a positive correlation between epistemological beliefs and grammatical content knowledge, and (ii) more mature epistemological beliefs go along with a higher interest in grammar (i.e. a higher motivation to deal with grammar), a better self-concept as well as more elaborate learning strategies. Using a questionnaire, we want to test these hypotheses. In this section we first provide information about our sample (Sect. 3.1) and the questionnaire (Sect. 3.2) before presenting the results (Sect. 3.3).

3.1 Participants The sample consists of 201 students (144 female, 55 male, 2 n.s.) of German at different stages of their studies (mean number of semesters: 3.4, SD = 2.4). Threequarters of them (74.2%) are studying to become teachers in a primary or secondary school.

80

D. Elsner

3.2 Methods and Procedure For the conceptualization of the questionnaire, we employed validated scales by Krettenauer (2005), Köller et al. (2000), and Kunter et al. (2002) for collecting data on epistemological beliefs, self-concept, motivation, and learning strategies. The questionnaire was handed out in the academic year 2017/2018 and was filled in by the participants during the first sessions of different linguistics classes.

3.2.1

The Grammar Test

The first part of the questionnaire was a grammar test. Since measuring grammatical content knowledge is rather difficult, especially when it is unclear how to model the competencies (cf. Sect. 2.1), we resorted to testing syntax and morphology, i.e. two domains that are crucial especially for teaching grammar in grades 5 and 6 (cf. Rahmenlehrplan (‘core curriculum’) Berlin/Brandenburg part c (German), p. 37). The grammar test consisted of 15 items (α = .71): 8 of them tested declarative knowledge, 4 of them procedural knowledge, and 3 of them metacognitive knowledge. We understand declarative knowledge as knowledge about facts (i.e. knowing that, e.g. nouns inflect for case and number) and procedural knowledge as knowing how (e.g. the ability to perform syntactic transformations like deleting or moving syntactic units). Meta-cognitive knowledge is knowledge about knowing (e.g. when explaining how one detected nouns in a text). To ensure validity we used publicly available VERA-85 items (IQB 2009), but also items from questionnaires by Habermann (2007) and Mesch and Dammert (2015), which have not been piloted and validated. Figure 3 shows a sample item testing procedural and meta-cognitive knowledge of German grammar:

Fig. 3 Sample item procedural/meta-cognitive knowledge (a: underline all the verbs in the following sentences, b: How did you detect the verbs?)

5 VERA-8

(Vergleichsarbeiten 8. Jahrgangsstufe) is a large-scale state-wide comparison test in grade 8 for determining achievement levels of all students in that grade. It is conducted once a year in the spring and examines German and/or mathematical competencies. The test aims at enhancing educational quality (cf. https://www.iqb.hu-berlin.de/vera).

Grammar Is Irrelevant – The Role of Epistemological Beliefs in Students’. . .

81

The participants’ performance was assessed based on marks similar to school marks where 1 indicates a very good and 6 a very bad performance.

3.2.2

Level of Epistemological Beliefs

We employed Krettenauer’s (2005) FREE for measuring the level of epistemological beliefs because it has been validated in different studies and its subscales have been shown to be predictive for different variables (cf. Trautwein and Lüdtke 2008). The FREE captures the dimension of certainty of knowledge and consists of 10 issues.6 Each issue is followed by three statements, each reflecting one of the three levels of certainty of knowledge (i.e. absolutistic, relativistic, and post-relativistic). Subjects were asked to indicate their level of agreement with each and every statement based on a 6-point Likert scale (from “I fully agree” to “I fully disagree”) and also to indicate the statement that corresponds best to their own opinion. The following example illustrates one of the issues as well as the respective statements (statement 1 = relativistic, statement 2 = absolutistic, statement 3 = post-relativistic): Issue: There are scientists who think that genetically modified food does not pose a health threat to humans. Other scientists regard genetically modified food as hazardous. • Scientists simply describe their own preconceptions as scientific insights. After all, no one can have knowledge of the long-term consequences that the use of biogenetics has for humans. • If one would collect and evaluate all the facts, one could surely give a definite answer to this question. • Even if one cannot estimate the consequences of the use of biogenetics with absolute certainty, there could be good reasons both for and against this method of producing food. Based on the subjects’ answers four indices are generated: One value for each subscale (absolutistic, relativistic, post-relativistic) is generated by averaging over the sum of agreement to the respective statements. The W-Index conforms to the percentage of the selection of post-relativistic statements. The FREE can be regarded as satisfyingly reliable and valid (cf. Krettenauer 2005, Trautwein and Lüdtke 2008). However, the internal consistencies of the subscales were rather low in the present study (absolutism α = .59, relativism α = .65, post-relativism α = .64).

6 The

10.

original version consists of 14 issues. Due to the length of our questionnaire we only used

82

3.2.3

D. Elsner

Motivation, Self-Concept & Learning Strategies

Motivation was investigated using a scale based on Köller et al. (2000), which measures students’ interest in grammar as a form of intrinsic motivation to deal with grammatical phenomena. It consists of five items and on a 4-point Likert scale (from “I fully disagree” to “I fully agree”) subjects had to indicate their level of agreement with different statements (e.g. For me personally grammar is one of the most important things; α = .77). The self-concept is understood as the degree of awareness of one’s own grammar competencies and is based on a scale by Köller et al. (2000). Again, subjects were supposed to indicate their level of agreement on a 4-point Likert scale (from “I fully disagree” to “I fully agree”) to four items (e.g. Even though I am trying very hard, grammar is more difficult for me than for my fellow students; α = .82). For measuring learning strategies, we employed three scales by Kunter et al. (2002) that were developed for the PISA7 -study 2000. The first scale centers on strategies of repetition (e.g. When preparing for a grammar exam I try to learn everything by rote that could be relevant for the exam, four items; α = .80) where learning does not involve an integration of newly acquired knowledge with existing knowledge. The second scale queries strategies of elaboration (e.g. When preparing for a grammar exam I try to connect new contents with things I have learned before, four items; α = .81) where previous knowledge is activated and newly acquired knowledge is integrated into an existing network of knowledge. Contents are supposed to be better remembered this way than by mere rote learning (cf. Urhahne 2006). Finally, the third scale investigates control strategies (e.g. When preparing for a grammar exam I force myself to check whether I retain what I have learned, four items; α = .65). Subjects had to indicate how often they resort to the respective strategy (4-point Likert scale, from “almost never” to “almost always”).

3.3 Results Previous research has found that students studying ‘soft’ subjects have a stronger tendency to believe that knowledge is uncertain (cf. Jehng et al. 1993). We therefore expect to find low ratings for absolutistic and higher ratings for relativistic and postrelativistic statements since our participants are students of a ‘soft’ subject and this is exactly what our data shows. The subjects possess relative mature epistemic beliefs at least within the dimension of certainty of knowledge (M = 51.04, SD = 20.08). Absolutistic (18.43%) and relativistic (24.58%) statements were less frequently chosen as statements that correspond to the subjects’ own opinion

7 The Programme for International Student Assessment is a triennial survey that tests the skills and knowledge of 15-year-old students in mathematics, reading, and science (http://www.oecd. org/pisa/).

Grammar Is Irrelevant – The Role of Epistemological Beliefs in Students’. . .

83

than post-relativistic statements. Additionally, post-relativistic statements received higher rates (M = 3.81, SD = 0.63) than absolutistic (M = 2.65, SD = 0.50) and relativistic (M = 3.49, SD = 0.64) statements. Given previous work (see Sect. 1) and studies like Ivo and Neuland (1991) we expect the students’ motivation to deal with grammar to be fairly low. However, the subjects indicated that they are quite motivated, though the values are not extremely high (M = 2.60, SD = 0.61, a higher number represents a higher motivation). Ninety percent of Risel’s (1999: 57) subjects indicate that their grammatical knowledge is patchy. Schäfer and Sayatz (2017: 244) state that their subjects are able to assess their grammatical knowledge in relation to their performance in a grammar test. Since quite a few studies have shown that students often do not possess deep knowledge of German grammar (e.g. Schmitz 2003; Bremerich-Vos and Dämmer 2013), we expect rather low rates for the self-concept. Again, our results meet our expectations. The subjects are aware of the fact that their grammatical competencies are rather ordinary (M = 1.98, SD = 0.65, a higher number represents a worse selfconcept). As for the learning strategies, we can see that repetition as well as more elaborate learning strategies are employed and the subjects check whether they have retained what they have learned (Table 1). However, the subjects’ performance in the grammar test was rather bad (M = 3.25, SD = 0.88) if we keep in mind that grammatical content knowledge was modelled as deep knowledge of school grammar, and that in principle school children from grade six on should have been able to pass this test, as by then they usually have learned the fundamentals of word classes and syntactic functions. As has been stated before, we want to test whether there is a positive correlation between epistemological beliefs and grammatical content knowledge, and whether more mature epistemological beliefs go along with a higher interest in grammar (i.e. a higher motivation to deal with grammar), a better self-concept as well as more elaborate learning strategies. As can be seen in Table 2, neither the W-index nor any of the subscales correlate significantly with the relevant variables. The results therefore falsify both hypotheses. However, there are other correlations. Since the Table 1 Descriptive statistics for the investigated domains

W-Index Subscale absolutism Subscale relativism Subscale post-relativism Motivation/Interest in Grammar Self-concept Repetition Elaboration Control Performance grammar test

M 51.04 3.81 3.49 2.65 2.60 1.98 2.51 2.74 3.07 3.25

SD 20.08 0.63 0.64 0.50 0.61 0.65 0.73 0.72 0.54 0.88

Motivation Self-concept Repetition Elaboration Control W-Index Grammar test Subscale post Subscale Rel Subscale abs

Motivation 1 −.330** −.106 .206** −.003 −.103 −.209** −.003 .029 −.094

Self-concept −.330** 1 .222** −.135 .018 −.055 .472** .071 −.155* −.117

Rep. .134 .222** 1 .087 .078 .029 .149* −.022 −.068 −.002

Elab. .206** −.135 .087 1 .050 .027 −.131 −.056 .051 −.047

Control −.003 .018 .078 .050 1 .089 .005 .045 .099 .263**

Table 2 Bivariate correlations between investigated domains, *p < .05, **p < .01 W-Index −.103 −.055 .029 .027 .089 1 −.100 −.358** .169* .233**

Grammar Test −.209** .472** .149* −.131 .005 −.100 1 .139 −.090 −.090

Subscale Post −.003 .071 −.022 −.056 .045 −.358** .139 1 .163* −.065

Sub. Rel .029 −.155* −.068 .051 .099 .169* −.090 .163* 1 .039

Sub. Abs −.094 −.117 −.002 −.047 .263** .233** −.090 −.065 .039 1

84 D. Elsner

Grammar Is Irrelevant – The Role of Epistemological Beliefs in Students’. . .

85

W-Index conforms to the percentage of the selection of post-relativistic statements it is to be expected that it correlates negatively with the values for post-relativistic statements (since a low value for a statement means a high agreement) and positively with the values for relativistic and absolutistic statements (since a high value means refusal). This is exactly what the data shows: the higher the W-index, the lower the values for post-relativistic, and the higher the values for relativistic and absolutistic statements. Unsurprisingly, a higher motivation goes along with a better self-concept, the use of elaborate learning strategies, and a better performance in the grammar test. A poor self-concept correlates significantly with repetitive learning strategies and a worse performance in the grammar test. Finally, the use of repetitive learning strategies goes along with a poor performance in the grammar test. In addition, we also tested whether categorical variables (gender, subject of study,8 Latin at school) correlate with any of the variables under investigation. It turns out that students who had Latin at school achieve better results in the grammar test (U = 6.003, z = 3.55, r = .02, p < .001) and that students who do not want to become teachers show a better performance in the grammar test (U = 2760, z = −2.81, r = −.20, p < .005).

4 Discussion & Conclusion Based on existing research dealing with epistemological beliefs and their predictive qualities, we investigated whether epistemological beliefs can account for the fact that teacher trainees of German often do not see the relevance of grammar for their future occupation and do not acquire knowledge of grammar to a considerable extent during their course of studies. Contrary to previous studies, there seem to be no correlations between the students’ epistemological beliefs and their grammar competencies, their motivation to deal with grammar, their self-concept, and implemented learning strategies. How can this be accounted for? Krettenauer’s (2005) FREE has two characteristics that might be of relevance: It only measures the dimension of certainty of knowledge, and it is not domain specific. In the following we discuss in what way these two aspects might hold an explanation for the fact that we could not find correlations between the epistemological beliefs and the other tested variables. Urhahne (2006) investigates the role of domain-specific epistemological beliefs as a predictor for other variables and finds that certainty of knowledge only correlates with repetitive learning strategies. Biology students who believe that knowledge is certain and invariable tend to implement repetitive learning strategies. However, especially for the dimensions of simplicity and justification of knowledge, he finds numerous correlations with motivation, self-concept, and learning

8 This

not.

refers to the question of whether German is being studied in order to become a teacher or

86

D. Elsner

strategies. In addition to this, the FREE is domain-general since it contains issues covering different topics and domains like natural sciences (including medical science) as well as social sciences and humanities (cf. Krettenauer 2005). Whether epistemological beliefs need to be modelled as domain-specific or not is a topic of debate (cf. Hofer 2005; Urhahne 2006). As has been shown elsewhere (cf. Jehng et al. 1993), epistemological beliefs differ according to the field of study and Hofer (2000) additionally shows that one and the same individual may hold different beliefs about different domains. So it is likely that results differ as a function of the domain-general or domain-specific modelling of epistemological beliefs. Finally, as Peter et al. (2015) point out, it might not be fully adequate to understand sophisticated epistemological beliefs as a rejection of absolutism since both positions, relativism and post-relativism, would reject absolutism. Operationalizing epistemological beliefs in this way might hold an explanation for ,,inconsistent findings on the relationship between epistemic beliefs, and [ . . . ] learning [ . . . ]“ (Peter et al. 2015:218). For our case we would then need to resort to a domainspecific questionnaire which covers all four of the dimensions identified by Hofer and Pintrich (1997). Such an investigation is being undertaken at present (Opitz and Elsner in prep.). Nevertheless, our study did reveal that students do not possess grammatical content knowledge to a sufficient extent. A worse performance in the grammar test correlates with a lower motivation, a worse self-concept, and the use of repetitive learning strategies. Assuming that a higher motivation can among other things be triggered by realizing that the content of grammar classes is relevant when training to become a teacher, we propose that the academic curriculum needs to be adjusted to fit the teacher trainees’ needs better. In order for this to happen we need to know first what it is that teacher trainees should know, i.e. we have to think about the conceptualization of grammatical content knowledge. With reference to research conducted in other fields (especially mathematics), it has to be clarified whether a construct similar to school related content knowledge can be identified and how exactly it can be modelled for linguistics and grammar, respectively. Schlipphack (2012) suggests that formal linguistic approaches might counteract insufficient grammatical knowledge. She advocates implementing generative approaches to grammar (Government and Binding) in secondary school education. This idea is motivated mainly by referring to the importance of Chomsky’s achievements and the cognitive foundations of his theory, which provides precise answers to fundamental questions like ‘What is a language?’ or ‘How are languages acquired?’. Schlipphack (2012: 108) also mentions the usefulness of generative approaches for syntactic analyses. If generative grammar should be (re)introduced in schools, such concepts need to be implemented permanently in teacher education as well. But even if no such reintroduction takes place, it might be worthwhile thinking about implementing formal approaches to grammar in teacher education. Syntactic trees make hierarchical structures obvious and convey the fact that a language is structured in a certain way more clearly than a linear analysis of sentence constituents. Acquiring knowledge of formal grammar is tied to dealing with highly abstract concepts and formalizations, which try to explain what it is that native

Grammar Is Irrelevant – The Role of Epistemological Beliefs in Students’. . .

87

speakers actually know when they are able to speak a language. A traditional sentence analysis only provides a description of the sentence constituents, but hardly explains why the sentence is structured the way it is. In addition to reviewing the curriculum, we have to reflect on the way we teach grammar in academic classrooms. Boettcher (1994) states that university grammar classes do not differ significantly from school grammar classes. Maybe one way of influencing students’ motivation is to think about applying alternative methods to teaching grammar. Ideas can already be found in recent articles, e.g. Zepter (2015), which puts forward and discusses the concept of inventing a language, or Kleinbub (2015), which shows how to use transcripts of school lessons in academic classes. Acknowledgments I would to thank Tanja Kupisch, Steve Berman, and two anonymous reviewers for helpful comments and remarks on earlier versions of this paper. This work was presented at the “Modern Linguistics and Language Didactics” workshop, University of Konstanz, 15-16 March 2018. Many thanks go to the organizers, Tanja Kupisch and Andreas Trotzke, as well as the audience of the workshop for comments and remarks.

References Baumert, J., & Kunter, M. (2006). Stichwort: Professionelle Kompetenz von Lehrkräften. Zeitschrift für Erziehungswissenschaft, 9, 469–520. Baumert, J., & Kunter, M. (2011). Das Kompetenzmodell von COACTIV. In M. Kunter, J. Baumert, W. Blum, U. Klusmann, S. Krauss, & M. Neubrand (Eds.), Professionelle Kompetenz von Lehrkräften. Ergebnisse des Forschugnsprogramms COACTIV (pp. 29–54). Münster: Waxmann. Baumert, J., Köller, O., & Neubrand, J. (2000). Epistemologische Überzeugungen und Fachverständnis im Mathematik- und Physikunterricht. In J. Baumert, W. Bos, & R. Lehmann (Eds.), TIMSS/III. Dritte internationale Mathematik- und Naturwissenschaftsstudie. Mathematische und naturwissenschaftliche Bildung am Ende der Schullaufbahn. Band II: Mathematische und physikalische Kompetenzen am Ende der gymnasialen Oberstufe (pp. 229–269). Opladen: Leske + Budrich. Boettcher, W. (1994). Grammatikunterricht in Schule und Lehrerausbildung. Der Deutschunterricht, 46, 8–31. Boettcher, W. (1999). Der Kampf mit dem Präpositionalobjekt. In A. Bremerich-Vos (Ed.), Zur Praxis des Grammatikunterrichts. Mit Materialien für Lehrer und Schüler (pp. 193–252). Freiburg im Breisgau: Filibach Verlag. Bredel, U. (2013). Sprachbetrachtung und Grammatikunterricht. 2nd revised edition. Paderborn et al.: Schöningh. Bremerich-Vos, A., & Dämmer, J. (2013). Professionelles Wissen im Studienverlauf: Lehramt Deutsch. In S. Blömeke et al. (Eds.), Professionelle Kompetenzen im Studienverlauf. Weitere Ergebnisse zur Deutsch-, Englisch-, und Mathematiklehrerausbildung aus TEDS-LT (pp. 47– 76). Münster: Waxmann. Dreher, A., Lindmeier, A., Heinze, A., & Niemand, C. (2018). What kind of content knowledge do secondary mathematics teachers need? A conceptualization taking into account academic and school mathematics. Journal für Mathematik-Didaktik, 39, 319–341. Dubberke, T., Kunter, M., McElvany, N., Brunner, M., & Baumert, J. (2008). Lerntheoretische Überzeugungen von Mathematiklehrkräften: Einflüsse auf die Unterrichtsgestaltung und den Lernerfolg von Schülerinnen und Schülern. Zeitschrift für Pädagogische Psychologie, 22, 193– 206.

88

D. Elsner

Eisenberg, P. (2004). Wieviel Grammatik braucht die Schule? Didaktik Deutsch, 17, 4–25. Elsner, D. (2019). Unterstreiche die Wörter, die sagen, wie etwas ist – Bildungsgangfokussierende Analyse von Lern- und Übungsaufgaben zur Wortartenkategorisierung. F&E Edition, 25, 71– 82. Greenberg, J., & Baron, R. (2008). Behavior in organizations (9th ed.). Upper Saddle River: Pearson Prentice Hall. Grewendorf, G. (1989). Ergativity in German. Dordrecht: Foris Publications. Habermann, M. (2007). Grammatiktest für Studienanfänger. Information available online: http:/ /www.presse.uni-erlangen.de/infocenter/presse/pressemitteilungen/2007/nachrichten_2007/ 04_07/64grammatik.shtml Heinze, A., Dreher, A., & Lindmeier, A. (2016a). Teachers’ mathematical content knowledge in the field of tension between academic and school mathematics. khdm-Report, 5, 21–26. Heinze, A., Dreher, A., Lindmeier, A., & Niemand, C. (2016b). Akademisches versus schulbezogenes Fachwissen – ein differenzierteres Modell des fachspezifischen Professionswissens von angehenden Mathematiklehrkräften der Sekundarstufe. Zeitschrift für Erziehungswissenschaft, 19, 329–349. Heringer, H. J. (1995). Grammatikunterricht – wozu? Mitteilungen des deutschen Germanistenverbandes, 2, 8–16. Hofer, B. K. (2000). Dimensionality and disciplinary differences in personal epistemology. Contemporary Educational Psychology, 25, 378–405. Hofer, B. K. (2001). Personal epistemology research: Implications for learning and teaching. Educational Psychology Review, 13, 353–383. Hofer, B. K. (2002). Personal epistemology as a psychological and educational construct: An introduction. In B. K. Hofer & P. R. Pintrich (Eds.), Personal epistemology: The psychology of beliefs about knowledge and learning (pp. 3–14). Mahwah: Lawrence Erlbaum Associates Inc. Hofer, B. K. (2005). The legacy and the challenges: Paul Pintrich’s contributions to personal epistemology research. Educational Psychologist, 40, 95–105. Hofer, B. K., & Pintrich, P. R. (1997). The development of epistemological theories: Beliefs about knowledge and knowing and their relation to learning. Review of Educational Research, 67, 88–140. Hoppe-Graff, S., et al. (2009). Mitten im BA-Studium: Studienmotive, Orientierungen, Erfahrungen und Überzeugungen von Studierenden im neuen Lehramtsstudiengang. In S. Hoppe-Graff et al. (Eds.), Das Lehramtsstudium an der Universität Leipzig: Voraussetzungen, Erfahrungen und Probleme aus der Sicht von Studierenden und Referendaren. Beiträge zur Professionalisierung der Lehrerbildung. Band 1 (pp. 33–68). Leipziger Universitätsverlag: Leipzig. IQB. (2009). VERA-8. Online: https://www.iqb.hu-berlin.de/vera Ivo, H., & Neuland, E. (1991). Grammatisches Wissen. In Diskussion Deutsch (Vol. 5, pp. 484– 506). Jehng, J.-C. J., Johnson, S. D., & Anderson, R. C. (1993). Schooling and students’ epistemological beliefs about learning. Contemporary Educational Psychology, 18, 23–35. Kleinbub, I. (2015). Grammatik unterrichten – Professionsorientierung in der Lehrerbildung. In U. Bredel & C. Schmellentin (Eds.), Welche Grammatik braucht der Grammatikunterricht? (2nd Rev ed., pp. 135–159). Hohengehren: Baltmannsweiler. KMK. (2018). Ländergemeinsame inhaltliche Anforderungen für die Fachwissenschaften und Fachdidaktiken in der Lehrerbildung. Available online: https://www.kmk.org/fileadmin/ Dateien/veroeffentlichungen_beschluesse/2008/2008_10_16-Fachprofile-Lehrerbildung.pdf Köller, O., Schnabel, K. U., & Baumert, J. (2000). Der Einfluß der Leistungsstärke von Schulen auf das fachspezifische Selbstkonzept der Begabung und das Interesse. Zeitschrift für Entwicklungspsychologie und Pädagogische Psychologie, 32, 70–80. Krettenauer, T. (2005). Die Erfassung des Entwicklungsniveaus epistemologischer Überzeugungen und das Problem der Übertragbarkeit von Interviewverfahren in standardisierte Fragebogenmethoden. Zeitschrift für Entwicklungspsychologie und Pädagogische Psychologie, 37, 69–79.

Grammar Is Irrelevant – The Role of Epistemological Beliefs in Students’. . .

89

Kunter, M., & Trautwein, U. (2013). Psychologie des Unterrichts. Paderborn et al.: Ferdinand Schöningh. Kunter, M., et al. (2002). PISA 2000: Dokumentation der Erhebungsinstrumente. Berlin: MaxPlanck-Institut für Bildungsforschung. Mesch, B., & Dammert, Y. (2015). Verbwissen in der Primarstufe. In B. Mesch & B. Rothstein (Eds.), Was tun mit dem Verb? Über die Möglichkeit und Notwendigkeit einer didaktischen Neuerschließung des Verbs (pp. 1–43). Berlin/Boston: de Gruyter. Op’t Eynde, P., de Corte, E., & Verschaffel, L. (2002). Framing students’ mathematics-related beliefs. A quest for conceptual clarity and a comprehensive categorization. In G. C. Leder, E. Pehkonen, & G. Törner (Eds.), Beliefs: A hidden variable in mathematics education? (pp. 13–37). Dordrecht: Springer. Opitz, A., & Elsner, D. (in prep). Fragebogen zur Wahrnehmung von Grammatik. To appear in: Döring, S., & Elsner, D. (eds.), Grammatik für die Schule. Perry, W. G. (1970). Forms of intellectual and ethical development in the college years: A scheme. New York: Holt, Rinehart & Winston. Peter, J., Rosman, T., Mayer, A.-K., Leichner, N., & Krampen, G. (2015). Assessing epistemic sophistication by considering domain-specific absolute and multiplistic beliefs separately. British Journal of Educational Psychology, 86, 204–221. Pissarek, M., & Schilcher, A. (2017). FALKO-D: Die Untersuchung des Professionswissens von Deutschlehrenden. Entwicklung eines Messinstruments zur fachspezifischen Lehrerkompetenz und Ergebnisse zu dessen Validierung. In S. Krauss et al. (Eds.), FALKO: Fachspezifische Lehrerkompetenzen. Konzeption von Professionswissen in den Fächern Deutsch, Englisch, Latein, Physik, Musik, Evangelische Religion und Pädagogik (pp. 67–111). Münster/New York: Waxmann. Rahmenlehrplan Berlin, & Brandenburg Teil C (Deutsch). Developed and edited by Landesinstitut für Schule und Medien Berlin-Brandenburg. Available online: https://bildungsserver.berlinbrandenburg.de/fileadmin/bbb/unterricht/rahmenlehrplaene/Rahmenlehrplanprojekt/ amtliche_Fassung/Teil_C_Deutsch_2015_11_10_WEB.pdf Risel, H. (1999): Schlaglichter auf Wissensbestände. Anmerkungen zu “Grammatikbiografien”. In: Klotz, Peter Peyer, Ann (eds.): Wege und Irrwege sprachlich-grammatischer Sozialisation. Bestandsaufnahme – Reflexionen – Impulse. Baltmannsweiler: Schneider Verlag Hohengehren, 53–60. Ryan, M. P. (1984). Monitoring text comprehension: Individual differences in epistemological standards. Journal of Educational Psychology, 76, 248–258. Schäfer, R., & Sayatz, U. (2017). Wie viel Grammatik braucht das Germanistikstudium? Zeitschrift für germanistische Linguistik, 42, 221–255. Schlipphack, I. (2012). Generative Grammatik für die Schule. Marburg: Tectum Verlag. Schmitz, U. (2003). Satzzeigen. Wie oft kann das grammatische Abendland noch untergehen? Mitteilungen des deutschen Germanistenverbandes, 50, 452–458. Schommer, M. (1993). Epistemological development and academic performance among secondary students. Journal of Educational Psychology, 85, 406–411. Shulman, L. S. (1986). Those who understand: Knowledge growth in teaching. Educational Researcher, 15, 4–14. Trautwein, U., & Lüdtke, O. (2008). Die Erfassung wissenschaftsbezogener Überzeugungen in der gymnasialen Oberstufe und im Studium. Validierung des Fragebogens zur Erfassung des Entwicklungsniveaus epistemologischer Überzeugungen (FREE). Zeitschrift für Pädagogische Psychologie, 22, 277–291. Trautwein, U., Lüdtke, O., & Beyer, B. (2004). Rauchen ist tödlich, Computerspiele machen aggressiv? Allgemeine und theorienspezifische epistemologische Überzeugungen bei Studierenden unterschiedlicher Fachrichtung. Zeitschrift für pädagogische Psychologie, 18, 187–199.

90

D. Elsner

Urhahne, D. (2006). Die Bedeutung domänenspezifischer epistemologischer Überzeugungen für Motivation, Selbstkonzept und Lernstrategien von Studierenden. Zeitschrift für Pädagogische Psychologie, 20, 189–198. Weschenfelder, E. (2014). Professionelle Kompetenz von Politiklehrkräften. Eine Studie zu Wissen und Überzeugungen. Wiesbaden: Springer. Zepter, A. (2015). Systemorientierter Grammatikunterricht: Sprachen erfinden und Grammatik entdecken. Linguistische Berichte, 244, 383–406.

Shaking Students’ Beliefs About Grammar: Some Thoughts on the Academic Education of Future Language Teachers Sandra Döring

Abstract Courses on grammar constitute a basic component of the first academic phase of language teacher education at German universities. However, it has often been reported that teacher trainees do not feel very confident about their knowledge of grammar. Furthermore, grammar is often considered as relevant, but at the same time boring, and (too) complicated. When it comes to planning their own lessons, they even seem to ignore what they (should) have learnt at university. Students come to university with beliefs about grammar and grammar teaching that influence their appraisal of the subject matter they are exposed to in seminars and lectures. Crucially, as beliefs work like filters, they determine the outcome of learning since no satisfactory learning success will be achieved if students fail to see the interesting aspects of a subject. Thus, they constrain what students turn their attention to. In this paper, I will first present the results of two explorative studies concerning beliefs about grammar (352 responses to questionnaire 1 from seven different German universities, and 37 responses to questionnaire 2 from one seminar at Leipzig University). Second, on these grounds, I shall discuss how we might get the essentials of grammar across by considering students’ beliefs. The focus of this paper is on teacher trainees of German as a subject language, the ambient language and often the (or one of the) L1 of the prospective teachers, and of their future pupils. Keywords Teacher trainees · Grammar knowledge · Grammar teaching · Epistemic beliefs · German · Language teacher · Academic teaching · Teacher training programme

S. Döring () Institut für Germanistik, Universität Leipzig, Leipzig, Germany e-mail: [email protected] © Springer Nature Switzerland AG 2020 A. Trotzke, T. Kupisch (eds.), Formal Linguistics and Language Education, Educational Linguistics 43, https://doi.org/10.1007/978-3-030-39257-4_6

91

92

S. Döring

1 Introduction Courses on grammar constitute a basic component of the first academic phase of language teacher education at German universities. However, teacher trainees do not feel very confident about their content knowledge of grammar. Grammar is often considered as relevant1 , but at the same time boring, and (too) difficult (cf. Ivo and Neuland 1991), and when it comes to applying their knowledge while planning their own lessons, teacher trainees even seem to ignore what they (should) have learnt in university courses. Students in teacher training programmes come to university with beliefs about grammar and grammar teaching. Epistemic (or epistemological or personal) beliefs, as studied in the (educational) psychology of the hypothesised link between beliefs and other cognitive variables (like motivation, self-regulation, learning strategies, cognitive processing, cognitive change; see Hofer and Bendixen 2012; Paulsen and Wells 1998), are beliefs about the nature of knowledge and knowing that vary from naive beliefs (certain, stable) to sophisticated beliefs (knowledge changes over time) (Kienhues et al. 2008). Epistemic beliefs play an important role in conceptual change learning, see Hofer and Bendixen (2012). Students’ epistemic beliefs might differ according to their studied subject (see Paulsen and Wells (1998) for support of this hypothesis). Beliefs determine the appraisal and the outcome of learning, as they work like filters, and teachers’ beliefs may influence their pedagogical practices and, hence, students learning (for an overview see for example Pajares 1992; Blömeke 2004; Hofer and Bendixen 2012). Teachers’ beliefs are important for the process of teacher education, and changes in teachers’ practices may be the result of changes in teachers’ beliefs, see Richards et al. (2001). Consequently, beliefs have to be considered in university classes for teacher trainees. The aim of this paper is twofold: first, I would like to investigate what today’s students think about grammar, thus, what are the beliefs of today’s students more than 25 years after the study of Ivo and Neuland (1991) and the study of Zimmermann (1995). Secondly, as considering beliefs is relevant for effective teaching, I want to tackle the question of how we can deal with teacher trainees’ beliefs in teacher training programmes at university. The focus of this paper is on teacher trainees of German as a subject language, the ambient language and often the (or one of the) L1 of the prospective teachers, and of their future pupils. This is important to keep in mind as foreign language teachers might have to deal with other difficulties, while pursuing other aims as well. The paper is structured as follows: In the first part, I present and discuss the data of two explorative studies (questionnaires) addressing the question of what students think about grammar (see Sects. 2.1, 2.2, and 2.3). On this basis, I pursue this research question with respect to three aspects: the essentials of grammar for prospective teachers (Sect. 3.2.1), how to get these essentials across by taking into account special needs of grammar teaching for future teachers (Sect. 3.2.1), by 1 Similarly,

foreign language learners consider grammar important, see Incecay and Dollar (2011), Richards et al. (2001), and foreign language learners believe grammar was an important aspect of the language.

Shaking Students’ Beliefs About Grammar

93

considering reflective teaching (see Sect. 3.2), and by sketching out some alternative ways of teaching (and learning) grammatical theory (see Sect. 3.2.3). I also outline some difficulties and hitches. A short summary concludes this article in Sect. 4.

2 What Do Students Today Think About Grammar? In this section, I will present the results of two different explorative studies to approach the question of what teacher students of German think about grammar. Both studies are based on questionnaires. While the first questionnaire was run at seven universities across Germany and therefore has a larger perspective on questions and results, the second questionnaire was run in one of my own linguistic classes with more detailed but at the same time more specific results. In this paper, I focus on a particular part of the two questionnaires, namely, I concentrate on results that seem relevant as a baseline for an approach to the research question of this paper of how we can shake (or even change) students’ beliefs about grammar. The section is structured as follows: in Sects. 2.1 and 2.2, I present the two questionnaires (the sample, the data collection, and the data interpretation), the results are summarised in Sect. 2.3.

2.1 Questionnaire 1 2.1.1

The Sample

The first study was run from January to July 2016 at seven universities in Germany. The questionnaire was handed out by colleagues as a hard copy in seminars or lectures at German departments.2 The chosen universities are spread over Germany (north/south as well as east/west). The questionnaire consists of 16 questions, preceded by seven questions concerning personal details such as semester, teaching subjects,3 type of school, university, course, sex, age, date. The total number of responses of teacher students is 352,4 the mean semester is 4 (second half of the second academic year), the median semester 4.9 (first half of the third academic year). 57% (202) of the responses come from students of the school type Gymnasium (secondary school), who are 23 years old (median).

2.1.2

Data Collection

For the purpose of this paper, I report the results of two different types of questions. The first question part (Question 1) consisted of four statements concerning the self2 Special

thanks to all colleagues. Germany, for secondary school teachers, two subjects are studied. 4 Flensburg = 68; Göttingen = 53; Jena = 47; Leipzig = 77; Mainz = 65; Rostock = 16; Tübingen = 26. 3 In

94

S. Döring

Jena

Lpz

Mainz

Rost

3.5 3.0 2.5 2.0 1.5

Ich würde gern mehr Grammatik im Studium machen Goe

1.0

3.5 3.0 2.5 2.0 1.5

Flens

4.0

I would like to have more grammar at university

4.0

I am very interested in grammar

1.0

Mein Interesse an grammatischen Inhalten im Studium ist sehr hoch

concept of students’ beliefs about grammar. Participants had to indicate their level of agreement with the statements on a 4-point Likert scale (from ‘I fully agree’ to ‘I totally disagree’). Alternatively, they could indicate the ‘I don’t know’ case. In the following figure, the results for the four statements in question 1 are reported in boxplots for each of the participating universities. Most of the participants are rather interested in grammar (mainly agreed), they would rather like to have more grammar courses at university (between 1 and 3), they find grammar moderately complicated (between 2 and 3), and they mostly do not feel insecure about the fact that there is often no right-wrong-distinction in grammar (2 and 3). For more details see Fig. 1. The second question I want to report on (Question 8) was referring to German lessons at school. In total, eight statements were given, the answers had to be indicated by the participants on a 4-point Likert scale (from ‘I fully agree’ to ‘I totally disagree’). Alternatively, they could, again, indicate the ‘I don’t know’ case. In Fig. 2, the ratings of the statements in question 8 are reported in boxplots for each of the participating universities. The participants agree (between 1 and 2) that the analysis of parts of speech and of constituents of the sentence (Satzglied) are highly important in German lessons at school. The participating students judge grammar knowledge from school slightly less useful for university purposes (mostly between 2 and 3) than grammar knowledge from university for school purposes.

Tueb

Flens

Goe

Goe

Jena

Lpz

Mainz

1= I fully agree, 4=I totally disagree

Lpz

Mainz

Rost

Tueb

Rost

Tueb

1.5

2.0

2.5

3.0

3.5

4.0

I feel insecure that there is often no right and wrong in grammar

1.0

3.0 2.5 2.0 1.0

1.5

Grammatik finde ich kompliziert

3.5

4.0

I find grammar complicated

Flens

Jena

1= I fully agree, 4=I totally disagree Verunsichert, dass es oft kein Richtig und Falsch in Grammatik zu geben scheint

1= I fully agree, 4=I totally disagree

Flens

Goe

Jena

Lpz

Mainz

Rost

Tueb

1= I fully agree, 4=I totally disagree

Fig. 1 Answers to statements in question 1, presented at each of the participating universities

Jena Lpz

4.0 3.5 3.0 2.5 2.0 1.5 1.0

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Grammatik aus der Schule nützlich an Uni

Grammar from school useful for university

Rost

Flens

Rost

Not more time than necessary for grammar

3.5 3.0 2.5 2.0 1.5 1.0

4.0 3.5 3.0 2.5 2.0

Jena Lpz

4.0

More grammar at school Für Gramm nicht mehr Zeit als notwendig

1= I fully agree, 4=I totally disagree

Rost

Flens

Jena Lpz

Rost

Grammar up to high school level

Grammar from university for school

3.5 3.0 2.5 2.0 1.5

Grammatikwissen aus Uni für Schule

1.0

3.5 3.0 2.5 2.0 1.5 1.0

Jena Lpz

4.0

1= I fully agree, 4=I totally disagree

4.0

1= I fully agree, 4=I totally disagree

Flens

Rost

Flens

Flens

Jena Lpz

Rost

1= I fully agree, 4=I totally disagree

Rost

1.5

2.0

2.5

3.0

3.5

4.0

Lectures about grammar of other languages

1.0

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Grammar in the curricular out of habit

Jena Lpz

1= I fully agree, 4=I totally disagree Zu Kursen über Grammatik anderer Sprachen würde ich gehen

1= I fully agree, 4=I totally disagree

Grammatik ist nur im Lehrplan aus Gewohnheit

Jena Lpz

1= I fully agree, 4=I totally disagree

1.5

Mehr Grammatik an der Schule

Flens

Flens

Grammatikunterricht bis zur 12./13. Klasse

95

Word classes, constituents very important

1.0

Hoher Stellenwert der WA− und SG−Analyse

Shaking Students’ Beliefs About Grammar

Flens

Jena Lpz

Rost

1= I fully agree, 4=I totally disagree

Fig. 2 Answers to statements in question 8, presented at each of the participating universities

96

S. Döring

The median is mostly at 2 for the statement that more grammar should be taught at school, and up to high school level (they rather agree). The participating students seem willing to invest more time than necessary for grammar (mean = 3), and they indicate that they would go to lectures about the grammar of other languages. All participating students strongly disagree with the statement that grammar is only part of the curriculum out of habit.

2.1.3

Data Interpretation

When compared to the literature, the ratings of the statements in questions 1 and 8 are partially surprising: students state that they are rather interested in grammar, they are not strongly against more grammar at university, they do find grammar just slightly complicated, and they do not feel insecure about the (often missing) rightwrong-distinction in grammar. Students seem to be willing to invest more time in grammar classes, and they state that they would go to lectures on grammar of other languages. The reason for the agreement with more grammar at school, and even up to high school level, is not clear. Either they wished to be more prepared for university grammar classes (participants indicated that grammar knowledge from school is at least slightly useful at university) or they want to teach more grammar at school when they are teachers. The questionnaire does not allow any definite conclusion in this regard. Do we have students who are interested in grammar, and are ready to teach more grammar at school, and lecturers underestimate this potential because they believe that students believe that grammar is dull and complicated?

2.2 Questionnaire 2 2.2.1

The Sample

The second questionnaire was handed out by myself as a hard copy in the very first session of an advanced linguistic course in a compulsory module at Leipzig University. This course was taught as a compact seminar, therefore, the semester itself was already in its last month. The course is recommended to students in their 7th semester (fourth academic year). Usually, they have completed at this time three linguistic modules, two of which contain courses on grammar. Furthermore, by that time, they have had their first practical school experience. The questionnaire consists of 8 open questions, followed by three questions concerning personal details as the second subject studied for teaching, type of school, and age. The total number of responses of teacher students is 37, the average semester is, as expected, the 7th. For the purpose of this paper, I am interested in the answers to two questions, namely associations with grammar lessons (Question 1), and what the participants believe what pupils think about grammar lessons (Question 5).

Shaking Students’ Beliefs About Grammar

2.2.2

97

Data Collection

The questionnaire consists of incomplete statements participants were asked to complete. Hence, lists of words are the result. Students gave different amounts of words, I noted them all, and present the result as word clouds. For the first statement (free associations with grammar lessons), the answers may be summarised by complicated (kompliziert), dry (trocken), confusing (unübersichtlich), dull (langweilig), unpopular (unbeliebt), see Fig. 3. Moreover, the answers to the incomplete sentence Pupils find grammar lessons . . . (Schüler_innen finden Grammatikunterricht . . . ) were mostly: dry (trocken), complicated (kompliziert), useless (unnötig), monotonous (eintönig), logic (logisch), confusing (unübersichtlich), complex (komplex), dull (öde), interesting (interessant), see Fig. 4.

2.2.3

Data Interpretation

Strikingly, teacher trainees think that pupils think in the same way about grammar as they do: dry (trocken), complicated (kompliziert), confusing (unübersichtlich) are often repeated attributes for both groups, for students’ estimation of pupils’ beliefs Fig. 3 Free associations with grammar lessons

Fig. 4 Pupils find grammar lessons . . .

98

S. Döring

even more so than for their own beliefs. Not even one student chose I don’t know for the pupils opinion about grammar. Future teachers believe that they do know how their (future) pupils think about a certain subject like grammar. To my knowledge, there is no independent survey about what pupils think about grammar in different schools, at different ages, at different times of the school year. Since it is assumed that teaching behaviour is influenced by what teachers think about pupils’ beliefs (in the same way as university teaching is influenced by what lecturers think about students’ beliefs), it is indispensable to make students aware of their beliefs, and also to make them aware of their beliefs about others’ beliefs (see Sect. 3.2.2 for practical details).

2.3 Results and Summary of the Explorative Studies In the last two sections, we have seen two different explorative studies. Do we know better what students’ beliefs about grammar are today? In the first questionnaire, the participants state that they find grammar complicated on a rank between 2 and 3, but in the seminar questionnaire with an open question, complicated is (besides dry) the most prominent word that comes to students’ mind for the stimulus grammar. The subjects rate, rather positively, that they wish for more grammar at school, even up to high school level, but in questionnaire 2 they state that pupils (therefore their future pupils) find this subject rather useless, confusing, monotonous. It seems clear that we cannot tell what students today think about grammar across the board. The differences between the results of the first and the second questionnaire might be tied to social expectations: it seems socially accepted that grammar is considered complicated and that someone has problems with grammar (similarly to mathematics). This might play a role for answering the question with free associations. Perhaps it is rather a sign of general beliefs than of individual ones (even if both are pieces of the puzzle). Moreover, we do not really know what the subjects actually mean by complicated: difficult to understand or to explain? Do they mean complex or sophisticated? By which criterion do they judge grammar as complicated? By their marks at school or at university, by their teachers’ capacity to explain, in comparison to which other subjects (literature, mathematics, or sports)? Is complicated always negatively connoted for the participants? Beliefs need to be identified before they can be changed, but after these two explorative studies we may sum up that it is more difficult than expected to classify general beliefs about grammar. In addition, individual experiences at school or university, personal interest in puzzling facts or structures might play a role as well as social expectations. Hence, dealing with students’ beliefs involves the recognition of individual situations.

Shaking Students’ Beliefs About Grammar

99

3 How Can We Deal with Students’ Beliefs? Before we discuss how we can deal with students’ beliefs about grammar in university classes, I think it is important to state what I think are the essentials or the areas of knowledge about grammar that we should try to get across to future language teachers (see Sect. 3.1). In Sect. 3.2, we address the research question of how we can shake or change students’ beliefs, concentrating on three points: the particularities of grammar teaching for future teachers, the method of reflective teaching, and suggestions for alternative ways of teaching. In that section, I shall report and comment on a range of activities that I tried out in my own teaching. In Sect. 3.3, I address some difficulties and hitches.

3.1 Essential Areas of Knowledge About Grammar for Future Language Teachers In this subsection, I would like to discuss essentials of knowledge about grammar for future language teachers. Certainly, future language teachers need to have explicit knowledge about the grammar of the language they teach. Fillmore and Snow (2018) (based on Fillmore and Snow 2000, 2002), distinguish five roles of teachers: the teacher as communicator, as educator, as evaluator, as educated human being, and as agent of socialisation (see Fillmore and Snow 2018: 14ff.), each of which demand an in-depth knowledge about language. By exploring these functions, Fillmore and Snow (2018) provide convincing arguments for integrating linguistics in the curriculum for teacher trainees. Especially as an educator and as an evaluator, teachers need profound knowledge of formal linguistics as a solid background to act on. Beside this rather serving function of formal linguistics, it is simply interesting to know more about human language because of its uniqueness. Besides the teaching of explicit knowledge about grammar, the training of linguistic competences seems to me even more important. Therefore, instead of giving a list of subjects to teach, I would like to explore the question of how students might be enabled to deal with language from a linguistic point of view. The major goal for students and pupils being concerned with linguistics is language awareness.5 Hence, the purpose consists of making them realise that language is structured, that we can speak about language and that categories are helpful for the discovery of generalisations and regularities. The aim is to make unconscious knowledge explicit and to enable students (and pupils) to speak about language. In 5 According

to the Association for Language Awareness (ALA): Language Awareness can be defined as explicit knowledge about language, and conscious perception and sensitivity in language learning, language teaching and language use, see Association for Language Awareness (ALA) (2018). Language awareness is a key term in all language teaching. For a short introduction see Carter (2003), Bolitho et al. (2003); for language awareness in primary and secondary education see Denham and Lobeck (2010).

100

S. Döring

the following, I will concentrate on three areas of competences: classification, regularities, and language varieties; and I will give concrete examples I used in seminars (or workshops).6 First, I consider it essential for teacher students to be able to organise, classify, and categorise items according to certain criteria. At first glance, it looks quite easy to do so, and even too simple to teach in linguistics classes at university. But, at a closer look, a full range of linguistic knowledge is necessary to deal with the task. Essential in this respect are the ability to rate items as linguistically relevant, organize and re-organize items, choose and justify criteria to organize items, and the experience that different criteria give different results. Hence, the teacher students’ choices depend on the goal of the analysis being carried out. This is fundamentally different from learning a list of terms by heart. When organising items, students learn that categories are not a given thing (what a list of items suggests) but the result of a process with its specific objective. In order to achieve an intended result, one necessarily has to deal with choices and with the fact that even the best analyses have either leftovers or work with criteria that are too soft. The classical way to train these competences is probably by classifying words into parts of speech. Words might be ordered by various criteria (e.g. semantic, morphological or syntactic criteria, or even starting letters, number of syllables and so on). Similarly, it is possible to sort constituents by heads, complements or alike. But these competences can be trained by organising clauses in terms of verb positions or by finite or infinite verbs or by number of subordination. Word formation may be another area to train these competences: by type of word formation, by type of affixes used, by type of basis. Or by valency or subcategorisation of verbs (organising verbs by number and/or type of their arguments). Second, in my view, another essential competence is the ability to discover regularities in language. To achieve this goal, students have on one hand to be taught the regularities of the studied language (here: German), and on the other hand, they have to learn how to discover regularities. Seeing regularities is one thing, verbalising regularities and learning how to verbalise regularities is another matter. All these competences are included in discovering regularities. In the example in (1), we see a simple German noun phrase consisting of an article, an adjective, and the noun in (1-a), and without the article in (1-b). Regularities can be discovered on different levels like word order, agreement, weak and strong inflection of adjectives depending on the presence of a determiner. (1)

a. b.

6 Similarly

das kleine Haus the small house kleines Haus small.NEUTR.SG. house

in SLA: some researchers in this wide field are more interested in properties of SLA, others more in teaching. The implications of formal approaches for language teaching might be very fruitful if both interests are working together, see Whong (2013), or the whole volume Whong et al. (2013).

Shaking Students’ Beliefs About Grammar

101

c. *das Haus kleine the house small d. *kleine Haus small.FEM.SG. house Another example for discovering regularities might come from a language which is unknown to the students. Experience shows that discovering and verbalising regularities in a foreign and/or unknown language might be easier for students, probably because of increased attention to unknown language material and the involvement of a conscious discovery. In (2), I report an exercise from Haspelmath (2002). In my experience, students have fun doing exercises of this kind. They have to find out about the hidden rules and to apply and to speak about them. The last part is often the most difficult one, even if the results are completely fine. (2)

Somali exhibits a great amount of allomorphy in the plural formation of its nouns. Four different allomorphs are represented in the following examples. Based on these examples, formulate a hypothesis about the phonological conditions for each of the plural allomorphs. [. . . ] (Haspelmath 2002: 36)

Singular awowe baabaco beed buug cashar fure ilmo miis

Plural awowayaal baabacooyin beedad buugag casharro furayaal ilmooyin miisas

Translation grandfather palm egg book lesson key tear table

Singular qado shabeel waraabe xidid tuulo tog albaab buste

Plural qadooyin shabeello waraabayaal xididdo

Translation lunch leopard hyena eagle village river door blanket

Third, a further essential competence is the ability to deal with language varieties in a broader sense. First, a look at diverging linguistic choices in non-standard varieties can reveal a lot about Standard German. Second, teachers are faced with the increasing multilingual reality of their classrooms, a linguistic potential to appreciate. (Spoken) varieties of German, different L1s of students, (modern) foreign languages taught at school have to be taken into account. Students should learn about other languages, be trained in comparing languages linguistically, be experts of their own variety(ies). Of course, due to time constraints, students cannot be prepared for all languages and varieties they might encounter in their future classrooms. But they may be equipped with the tools to deal with such situations in the future. For this, it is necessary to open linguistic classes to other languages, to see different structures. In this context, I present two different examples. In Table 1, the inflection of to have and to be in one variety of German (Kluftern) is presented in comparison to the Standard German inflection (which students of

102

S. Döring

Table 1 Inflection of the verbs to be and to have in the Swabian variety of Kluftern Rabanus (2008, 131) and in Standard German 1. Sg. 2. Sg. 3. Sg. 1. Pl. 2. Pl. 3. Pl.

to have (Kluftern) ho(-n) ho-scht ho-t ho-nt ho-nt ho-nt

to be (Kluftern) b¯ı bi-scht isch-t si-nt si-nt si-nt

to have (Standard German) habe hast hat haben habt haben

be (Standard German) bin bist ist sind seid sind

German can produce on their own). In a further step, they can discover regularities in the verb paradigms, compare the two varieties, and then speak about the discovered regularities. The second example involves examples from simple noun phrases in Romance foreign languages taught at school, such as Spanish, French, and Italian, and from Arabic. In order to compare the structures of these noun phrases, students have to learn how to read glosses (a task which is constantly under-represented in German studies curricula for German students). In most cases, students will have studied one of the Romance languages at school (if it is not one of those, it may be Latin). Therefore, they start with the most familiar language, analyze and describe the structure before they approach the lesser known structures. They may describe the structure of the noun phrase for languages individually, for Romance languages in general, in comparison to German, in comparison to Arabic, in comparison to other languages they know. (3)

[Spanish] a. la camisa azul the shirt blue the blue shirt b. un amigo viejo a friend old an old friend c. un viejo amigo a old friend a long-time friend

(5)

[Italian] il mio amico the my friend my friend

(6)

[Arabic] al

(4)

k¯atib al maschh¯ur writer ART.DEF famous the famous writer ART. DEF

[French] a. la chemise bleu-e the shirt blue-FEM the blue shirt b. un bâtiment ancien a building historical a historical building c. un ancien ami a former friend

Shaking Students’ Beliefs About Grammar

103

After having looked at areas of knowledge about grammar, I estimate essential for prospective teachers of German, I would like to address the next step, how to get these essentials across.

3.2 Shaking Students’ Beliefs: How to Get the Essentials Across or Grammar for Prospective Teachers The major goal of university teaching might be to get the essentials across. For teaching grammar in linguistics classes, it necessarily includes strategies of shaking students’ beliefs about grammar. For shaking or changing beliefs, it is important to separate at least two levels for beliefs: the factual and the emotional level. And, of course, both levels need to be called into awareness, and addressed. In Sect. 3.2.1, I’ll concentrate on particularities of grammar teaching for prospective teachers, in Sect. 3.2.2, I take up the subject of reflective teaching, namely how to integrate students’ beliefs in teaching. In the third Sect. 3.2.3, I shall sketch alternative ways of teaching grammar I tried out in higher education that differ from teaching practices that students experienced at school or in university classes in general.

3.2.1

Grammar Teaching for Future Teachers

In all teaching, the addressee should be kept in mind. In our case, it is the fact that the students are prospective teachers that should be kept in mind in university teaching. In my experience, such a perspective changes the focus in the choice of topics, examples, and methods. That does not mean that school by itself becomes the subject of the course. Especially in mixed groups, all students have to be addressed. To be sure, all students, including future teachers, should study scientific ways of looking at language, but in general, not all of them are future researchers but teachers. Nevertheless, they need a profound knowledge about the German language. It is up to the university teacher to choose good examples for exemplary learning, which is not an easy task, especially with the major goal that these examples should be good for school teaching and autonomous learning. In mathematics, it is claimed that there is a third kind of knowledge beside scientific knowledge and school knowledge of mathematics: school related content knowledge, see Heinze et al. (2016) and Dreher et al. (2018). For linguistics, we are still at the beginning of defining or discussing what might be a school related content knowledge for grammar, see Döring and Elsner (in prep). I believe, we should be thinking more about method than about content in grammar teaching, see Döring and Wöllstein (in prep). Moreover, in linguistics classes at university, we need time for questions, time for revision, time to redo things. That is due to the fact that our classes are a kind

104

S. Döring

of model class for grammar, the teacher may set an example which differs from school experience with grammar. In addition, I would argue that it is preferable to make sure that students have understood, that they are enabled to develop their own research questions, to see linguistic problems, to have ideas for solutions. This seems to me much more important than going through a lot of subject matters, without anything being really digested by the students. Another difficulty seems to be to verbalise the lines of thinking, the chains of thoughts in linguistic argumentation. Although for university teachers, the correct categorisation of a word seems in many cases quite obvious, that might not be the case for students. The same reasoning holds for teachers and pupils. Therefore, it is vital for students to be aware of the process (e.g. categorising words), and to be trained to verbalise this process. Otherwise, students are neither able to follow the reasoning, nor to explain it. As in mathematics, if a pupil asks the teacher how to find a proof and the answer is that the pupil should just think about it and if it does not work, that he should try again and think harder – the pupil still does not know what he should do exactly. Teaching grammar at university should be inquiry and problem based learning. The research on this type of learning has clearly shown positive effects on retention of information. Moreover, discovery based learning develops inquiring minds. This is a potential for life-long learning, especially required for prospective teachers, as they have to continue their studies in an autonomous way for all their professional life. Problem based learning is, for sure, time consuming. But it is probably the only way to show that grammar is fun and not limited to terms and labels. Many things may be discovered about what we unconsciously do. As the major target competence is language awareness for prospective teacher students as for pupils, it seems evident and unavoidable to me to invest and reserve time for autonomous discoveries in an inductive way instead of top-down teaching in a deductive way. Furthermore, as mentioned before, the time consuming argument is not a valid one if we earn a deeper understanding of language, and we can provide a model for grammar lessons at school. In my opinion, existing teaching material for schools should be considered in linguistics classes at university, not only from a didactic point of view. Students should also be enabled to use these materials in an appropriate manner. By teaching material, I mean textbooks, teacher manuals, workbooks, as well as online material such as platforms for teachers (e.g. 4teachers), youtube-videos, or learning apps. Teachers use these materials especially under time pressure and for subjects they are not very confident with. This constitutes a problem as they might not be able to properly judge these materials. For this reason, it is, I think, unavoidable, to do (not only to discuss, but to do) the exercises, to discuss the proposed solutions in the manual and assess them against a linguistic background. On the one hand, only by doing the exercises, students (and teachers) see the lurking problem(s), and on the other hand, they may be enabled to use the exercises more efficiently. This may include that the exercise itself can be used at school but maybe in an adapted way. Why should it be done in university classes? Is the support of a university teacher really necessary? Yes, it is. And it is a shortcoming of university grammar

Shaking Students’ Beliefs About Grammar

105

classes in general that there is often not enough time and space to discuss solutions, and procedures to find solutions. This is a highly demanding task, and part of the learning process, and not part of a revision phase. Including other languages and varieties (here: of German) in linguistics classes at university opens the possibility to learn more about Standard German by looking at other grammatical realisations (for examples see Sect. 3.1). Moreover, these data are often new to everyone, and therefore, definitely not boring. Additionally, the different varieties, L1s, and foreign languages taught at school may be appreciated without any ranking, just as equally interesting languages.

3.2.2

Reflective Teaching: Taking Beliefs into Account

Beside the choice of topics, beliefs by themselves should be subject of (university) teaching. As mentioned above, beliefs have to be identified first before one can shake (or change) them. Therefore, a culture of teaching that makes students aware of the beliefs they bring to class and that encourages them to evaluate their beliefs is necessary. This behaviour encourages the prospective teachers to establish such a culture in their classrooms because taking beliefs into account is essential for effective learning. Pupils, of course, also have beliefs, and they determine the outcome of their learning in the same way. Respecting beliefs creates, moreover, a positive learning environment: respectful, without fear, giving room for questions and discussions. One possibility of making beliefs part of the teaching are short questionnaires with rather open questions. Students may fill in the answers to questions of the type presented in Sect. 2.2. As mostly students and teachers do not know each other at the beginning of a university course, it is not always possible to speak about the answers immediately (if possible it should be done). I ask them to save this questionnaire for the whole length of the course. Part-way through the course, we start talking about the associations written at the beginning of the course, mentioning possible changes, discussing ways of changes of their beliefs. From this moment onwards, beliefs are part of the meta level discussion of the course, students thus have the chance to evaluate their beliefs during the course, or at least become aware of them. Such approach is interesting and instructive for both sides, teachers and students. In addition, the course benefits from the discussions. Another option is to disclose the meta level of the course structure, its content, why I teach what I do when and how. This is relevant for their own teaching in future but, according to my experience, it also improves the learning results in a noticeable way. Sure, it is time consuming, as all aspects of reflective teaching, I present here. But, as I said before, it is better to cover less content, when at the end of the course, students know possibilities of doing grammar, and they are willing to keep trying, and learning. Moreover, at some points of the seminar, increasingly towards the end, I distinctively point out the additional knowledge or competences they will have gained after the course. At the first glance, this seems unnecessary because it should be obvious. But after discussions with students, I realised that they

106

S. Döring

do not see the step they made. After grammar courses at different levels, they always have the same feeling of unsolved (or unsolvable) problems, and they do not see that the problems presented in advanced courses are on a completely different level from those in earlier courses. Hence, I decided to make this increase of knowledge visible. 3.2.3

Alternative Ways of Teaching

When teaching grammar at university, major attention should be given to regular teaching in core modules as all students have to participate and hence may be addressed. Nevertheless, I would like to report some alternative ways of teaching, I tried out myself. All activities are summed up on Linguistik & Schule (linguistics and school), a platform for experimental teaching, material and research questions, see LuS (2015ff.). Concerning teaching, I would like to sketch out special tutorials (Lernstudio), project weekends on grammar (GraWo), and lectures on the grammar of other languages. The addressees of these activities are students and participation is voluntary. Beside the voluntariness, it is important that all these activities are without grading, and hence without credits. The tutorials are offered to prospective teacher students from the third year onwards (in contrast to regular tutorials in the first academic year) on special subjects like parts of speech, constituents, linear sentence structure like topologisches Feld. The students have successfully completed the core modules, and they have often made their first own teaching experiences. This is often the moment, when they realise what they should know but believe that they do not know it. For reasons of time, there is too little space in the curriculum to restart again. Moreover, not all students realise the same gaps, nor at the same moment. Therefore, tutorials over a short time on special subjects (minimal commitment) are offered by other prospective teacher students, ideally, on a regular basis (depending on the available financial support). Project weekends on grammar are another option of such alternative learning activities. At the best, it is a weekend in a small group of approximately 12 to 15 prospective teacher students with different second subjects, such as various foreign languages taught at school, pursuing a future career in different school types. This group goes on a weekend-trip. For time-and-money-saving reasons, a youth hostel not far from the university town is completely fine. The advantage of being away is the full commitment and the time of discussion in pauses or in the two evenings spent together. Students need to bring a grammar book (whatever they have at the bookshelf at home) and one school book (workbook or alike). In various smaller groups they work on topics such as adjective or noun phrase, discover what is written in their material, what is written in the material of the others, try to make sense of it. Surprisingly, they work very hard in these units of approximately 90 min on a pure grammar topic, and they discuss it, discover differences, raise questions. Other languages in the classroom are another topic of this weekend. Everybody is invited to shortly present one grammatical aspect of another language of their choice (from a list), and the possible impact on the German competences. The preparation may be in advance, or even on the weekend itself. Besides the fact that students are

Shaking Students’ Beliefs About Grammar

107

presented with different languages, they learn how to deal with information about the structures of languages, and see a number of different languages at least in a short sketch-out. If financial support is available, this part may even be outsourced in a regular lecture on languages in school contexts, open for everybody, at least for teachers, prospective teachers of all subjects, and even everybody else such as future parents for example.7 Specialized literature in German (as it is the target language here) on linguistic topics for (future) teachers in a short format for which linguistic and educational work together is not easy to find.8 Furthermore, we need forums that bring interests of formal linguistics and teaching together. Beside practical experiences, research based exchange such as the DFG-Netzwerk Grammatik für die Schule (GrafüS), the scientific network grammar for school,9 or the conference series Linguistics and Didactics (LiDi) at which this paper was presented, is vital. Both initiatives show a broader interest and the up-todateness of the subject.

3.3 Difficulties, Hitches In this subsection, I would like to mention some difficulties and hitches in changing beliefs in a steady way. A major difficulty in university programmes is the lack of time. In the long term, it is no solution to offer important grammar teaching experiences in an optional side-programme. Another aspect is, that we, in general, struggle with far to big group sizes (especially in a subject like German). This is not beneficial for the learning progress of the students and for reflective teaching aiming to shake students’ beliefs. As we are talking about future teachers, this is in the end a rather political question. But my personal impression – and I think the results of the first questionnaire (see Sect. 2.1) suggest that too – is that the first academic phase of language teacher education at universities is, nevertheless, rather successful, or at least on the right track in this respect. I am not saying that we have reached our goal. It is definitely an ongoing process. The next problem to address is the route of our students into practical school experiences – the missing link between the first and the second phase of teacher training programmes in Germany, and the missing

7 Teachers

may also be addressed by workshops on grammar with new material. But motivating them to participate in such a workshop is not that easy, see also Sect. 3.3. 8 The book series LinguS might be interesting in this respect. The concept of the books takes different aspects into account: all books are written in teams of a linguist and an expert on language didactics and consist of exercises and proposals for solutions. The subjects of the series are chosen by school relevance: words, clauses, texts, lexicology, language history, spelling, pragmatics; at the same time heritage languages, German as second language, multilinguism, language acquisition, German sign language, linguistic education, variational linguistics, see Döring and Gallmann (2017ff.) 9 For reference in G EPRIS see Döring and Elsner (2016–2019), for publications see Döring and Elsner (in prep), Döring and Wöllstein (in prep).

108

S. Döring

link from practising teachers back to university. In the second phase of language teacher education, students encounter beliefs of experienced teachers, mentors. In this training phase, they are in a very dependent relationship, and completely out of university’s scope. It would not only be in the interest of grammar, or in the interest of teachers of German, but for the benefit of all subjects, all teachers, all schools, and, eventually, of all pupils and therefore our society if lecturers of the first and second phase of teacher training programmes started working together, and if we started organising academic exchanges between schools and universities, and kept, on the one hand, teachers updated about relevant research results, and, on the other hand, learned about what students should know when finishing the academic phase of the teacher training programme. I am looking forward to the benefit such a cooperation will generate for both sides.

4 Summary and Outlook In this article, I addressed the question of how we can shake or change students’ beliefs. In the first part, I presented data of two explorative studies about what today’s students think about grammar (see Sects. 2.1, 2.2, and 2.3). It is rather uncontroversial that beliefs have to be considered in (university) teaching if we want to get the essentials across. In the second part, hence, I discussed areas of knowledge about grammar for prospective teachers I consider essential before addressing the possibilities of how to get these essentials across whilst taking students’ beliefs into account. Grammar teaching for prospective teachers must differ in some points from grammar teaching for future researchers, although both should definitely acquire scientific methods of looking at language (see Sect. 3.2.1). Reflective teaching makes students aware of their beliefs and gives time and space for discussing them (see Sect. 3.2.2). In Sect. 3.2.3, I presented alternative teaching and learning experiences. In Sect. 3.3, I mentioned some difficulties, but in total, I am rather optimistic about the teacher trainees at the end of their first academic phase of teacher education at university. The academic education of future language teachers is, of course, an ongoing process. Beside keeping this objective in mind, the next target (and challenge) is how to get teachers (back) in a constant dialogue with universities. Just a note for not losing courage, it is a rather lengthy process: to settle a new generation of teachers at all levels (university lecturers, mentors, school teachers, teacher trainers, textbook authors, other persons in charge of education) – it will take at least 30 years, I think. Acknowledgements Thanks be to the DFG for funding the scientific network DO 2041/1-1 Grammatik für die Schule (GrafüS) grammar for school, and to the Zentrum für Lehrerbildung und Schulforschung (ZLS) Leipzig for financial support to edit the collected data. Moreover, thanks be to Kristin Börjesson, Jochen Geilfuß-Wolfgang, Marcel Guhl and Fabian Heck for very helpful comments and remarks, to the editors and organisers of the conference Linguistics and Didactics (LiDi) in March 2017 in Konstanz, and, last but not least, to all teacher trainees sharing and discussing experiences in grammar (teaching) at university and at school.

Shaking Students’ Beliefs About Grammar

109

References Association for Language Awareness (ALA). (2018). Last access 31 Dec 2018. https://lexically. net/ala/la_defined.htm, http://www.languageawareness.org Blömeke, S. (2004). Empirische Befunde zur Wirksamkeit der Lehrerbildung. In S. Blömeke, P. Reinhold, G. Tulodziecki, & J. Wildt (Eds.), Handbuch Lehrerbildung (pp. 59–91). Braunschweig: Westermann. Bolitho, R., Carter, R., Hughes, R., Ivanic, R., Masuhara, H., & Tomlinson, B. (2003). Ten questions about language awareness. ELT Journal, 57, 251–259. Carter, R. (2003). Language awareness. ELT Journal, 57(1), 64–65. Denham, K., & Lobeck, A. (2010). Linguistics at school: Language awareness in primary and secondary education. Cambridge: Cambridge University Press. Döring, S., & Elsner, D. (2016–2019). GrafüS – Grammatik für die Schule. DFG-Netzwerk, http:// gepris.dfg.de/gepris/projekt/320262413 Döring, S., & Elsner, D. (in prep). Grammatik für die Schule. Basisartikel. Döring, S., & Gallmann, P. (Ed.). (2017ff.). LinguS. Sprachbewusstheit fördern. Tübingen: Narr. Döring, S., & Wöllstein, A. (in prep). Besser als gedacht. Kommentar zum Basisartikel. Dreher, A., Lindmeier, A., Heinze, A., & Niemand, C. (2018). What kind of content knowledge do secondary mathematics teachers need? Journal für Mathematik-Didaktik, 39(2), 319–341. https://doi.org/10.1007/s13138-018-0127-2 Fillmore, L. W., & Snow, C. E. (2000). What teachers need to know about language. Washington, DC: Center for Applied Linguistics/ERIC Clearinghouse on Languages and Linguistics. Fillmore, L. W., & Snow, C. E. (2002). What teachers need to know about language. In C. T. Adger, C. E. Snow, & D. Christian (Eds.), What teachers need to know about language (pp. 7– 54). Washington, DC: Center for Applied Linguistics/Delta Systems. Fillmore, L. W., & Snow, C. E. (2018). What teachers need to know about language. In C. T. Adger, C. E. Snow, & D. Christian (Eds.), What teachers need to know about language (Language Education, 2nd ed., pp. 8–51). Washington, DC: CAL (Center for Applied Linguistics). Haspelmath, M. (2002). Understanding morphology. London: Arnold. Heinze, A., Dreher, A., Lindmeier, A., & Niemand, C. (2016). Akademisches versus schulbezogenes Fachwissen – ein differenzierteres Modell des fachspezifischen Professionswissens von angehenden Mathematiklehrkräften der Sekundarstufe. Zeitschrift für Erziehungswissenschaft, 19(2), 329–349. Hofer, B. K., & Bendixen, L. D. (2012). Personal epistemology: Theory, research, and future directions, chap. 9. In K. R. Harris, S. Graham, & T. Urdan (editors-in chief) (Eds.), APA educational psychology handbook: Vol. 1 Theories, constructs, and critical issues (pp. 227– 256). American Psychological Association. Incecay, V., & Dollar, Y. K. (2011). Foreign language learners’ beliefs about grammar instruction and error correction. Procedia – Social and Behavioral Sciences, 15, 3394–3398. Ivo, H., & Neuland, E. (1991). Grammatisches Wissen. Skizze einer empirischen Untersuchung über Art, Umfang und Verteilung grammatischen Wissens (in der Bundesrepublik). Diskussion Deutsch, 121(436), 437–493. Kienhues, D., Bromme, R., Stahl, E. (2008). Changing epistemological beliefs: The unexpected impact of a short-term intervention. British Journal of Educational Psychology, 78(4), 545– 565. LuS. (2015ff). Projekt: Linguistik und Schule. lus.philol.uni-leipzig.de, Organisation: Sandra Döring. Pajares, M. F. (1992). Teachers’ beliefs and educational research: Cleaning up a messy construct. Review of Educational Research, 62(3), 307–332. https://doi.org/10.3102/00346543062003307 Paulsen, M. B., & Wells, C. T. (1998). Domain differences in the epistemological beliefs of college students. Research in Higher Education, 39(4), 365–384. Rabanus, S. (2008). Morphologisches Minimum. Distinktionen und Synkretismen im Minimalsatz hochdeutscher Dialekte (ZDL-Beiheft 134). Stuttgart: F. Steiner.

110

S. Döring

Richards, J. C., Gallo, P. B., & Renandya, W. A. (2001). Exploring teachers’ beliefs and the processes of change. PAC Journal, 1(1), 41–62. Whong, M. (2013). Applied generative SLA: The need for an agenda and a methodology. In M. Whong, K.-H. Gil, & H. Marsden (Eds.), Universal grammar and the second language classroom (Band 16 of Educational Linguistics, pp. 231–247). Berlin/Heidelberg: Springer Science+Business Media. Whong, M., Gil, K.-H., Marsden, H. (Ed.). (2013). Universal grammar and the second language classroom. (Band 16 of Educational Linguistics). Berlin/Heidelberg: Springer Science+Business Media. Zimmermann, G. (1995). Einstellungen zu Grammatik und Grammatikunterricht. In C. Gnutzmann & F. Königs (Eds.), Perspektiven des Grammatikunterrichts (Tübinger Beiträge zur Linguistik, Vol. 404, pp. 181–200). Tübingen: Gunter Narr.

Do Linguistic Landscapes Influence the Spelling Competence of Orthographic Beginners? Two Case Studies Björn Rothstein

Abstract The paper discusses whether linguistic landscapes influence the spelling competence of orthographic beginners. Two case studies are presented. In case study 1, a specifically designed corpus is presented: it includes an exemplary spelling phenomenon from one linguistic landscape (N-N compounds), which is analysed with respect to the frequency of wrong spellings. In the course of a second study, a field experiment, we analyse if wrong spellings in a linguistic landscape actually influence orthographic beginners. The test participants were third graders (N = 92), as it may be assumed that the effect of linguistic landscapes on spelling is particularly significant among orthographic beginners who, according to the curriculum, have already acquired basic reading skills but have not yet completed their spelling acquisition. According to the curriculum, the relevant orthographic phenomenon (N-N compound spellings with hyphen) has not yet been acquired in this form. Both studies presented in the paper do not rule out that wrong spellings in linguistic landscapes may affect the spelling performance of orthographic beginners. Keywords Linguistic landscapes · Orthography · Spelling · Reading · Learning · N-N compounds

1 Introduction Written language is our constant companion, due to the mobile terminals, billboard captions, traffic signs, timetables, newspapers, books and magazines we read when out and about. Thus, we might assume that reading permanently influences our cognitive and, more specifically, our orthographic skills. The present paper empirically investigates the spelling performance of orthographic beginners regarding German N-N compounds by referring to the currently much researched concept of linguistic

B. Rothstein () Germanistisches Institut, Ruhr-Universität Bochum, Bochum, Germany e-mail: [email protected] © Springer Nature Switzerland AG 2020 A. Trotzke, T. Kupisch (eds.), Formal Linguistics and Language Education, Educational Linguistics 43, https://doi.org/10.1007/978-3-030-39257-4_7

111

112

B. Rothstein

Fig. 1 Wrong spellings in public spaces

landscapes (roughly: “writing in public spaces”). In particular, wrong spellings in public spaces and their possible influence on spelling performance will be analysed. Two examples of this are provided in Fig. 1. The image on the left is a professionally made advertisment for German Postbank, the one on the right is a sign that we may suppose was made rather hastily. Both cases violate the official regulations (Amtliche Regelungen) for German spelling: they are incorrect. “riesich” (a blend of gigantic and I) should be spelled “riesig” (gigantic), and “Taschen Bücher” (paperbacks) should be spelled “Taschenbücher” or “Taschen-bücher”. The two different colours of the caption in the first image presents “riesich” as an intended spelling, thus resulting in the paronomasia “ich bin riesig (I am gigantic)”. Such spellings are usually called alternative spellings (German: “Anderssschreiben“): Andersschreiben [ . . . ] [ist] zwar als Abweichung zu verstehen, jedoch erfolgt das Abweichen von einem Muster, einer Norm, einer Regel oder einer Konvention zumeist nicht aus Unvermögen, sondern ist intendiert. (Schuster and Tophinke 2012: 13) Alternative spelling [ . . . ] should indeed be understood as a deviation, however this deviation from a pattern, a norm, a rule or a convention is in most cases not a result of inability but is intentional.

Wrong spellings are unintentional deviations that are the result of an inability to write correctly according to an orthographic pattern, norm, rule or convention. From the point of view of orthographic beginners, alternative and wrong spellings are initially spellings that are thought to be correct spellings, insofar as they do not recognize the difference from the officially correct spelling. To be able to distinguish alternative from wrong spellings, learners would have to know the correct spelling and recognize the intentional nature of alternative spellings. When dealing with the acquisition of orthography in the following, we subsume cases like “riesich” and “Taschen Bücher” under the term wrong spelling. In the context of wrong spellings it is interesting to bear the increasingly deteriorating spelling performance of students in mind (e.g. Steinig et al. 2011). The conditions of socialisation, teacher training, general performance at school and the teaching of spelling itself are blamed: social-scientific and psychological studies

Do Linguistic Landscapes Influence the Spelling Competence?

113

identify the negative influences of parents’ educational levels and language skills, teacher training is criticized, and the teaching of spelling is said to apply ineffective methods/didactics. Frequently poor reading abilities and not being in the habit of reading are also held responsible for bad spelling performance (for an overview, see Tacke 2000; Steinig et al. 2011; Siekmann 2011). If it is true that wrong spellings are frequently found in public spaces and if they are frequently read, we should address the following question: Do they influence the spelling performance of orthographic beginners? In other words: Will beginners who read “riesich” and “Taschen Bücher” once spell these words like this themselves? In the context of sociolinguistic research on multilingualism and urbanisation, the writing in public spaces is usually referred to as the linguistic landscape. The classic definition of linguistic landscape is as follows: The language of public road signs, advertising billboards, street names, place names, commercial shop signs, and public signs on government buildings combines to form the linguistic landscape of a given territory, region, or urban agglomeration. (Landry and Bourhis 1997: 25)

The research on linguistic landscapes originates from the ethno- and sociolinguistic research on language contact which developed in the 1960s and 1970s. Initially it was rather oriented towards semiotics (Römer 1968), later predominantly towards sociolinguistics. It analysed, for example, place-name signs and traffic signs in multilingual regions (Flaschberger and Reiterer 1980). The strong growth of multilingualism research at the end of the 1970s (Gal 1979) and of sociological research on urbanisation and metropolises (Ziegler et al. 2018) resulted in a growing interest in the use of written language in public spaces. This interest led to linguistic landscape research, to the analysis of the entirety of written language in a defined public area (Landry and Bourhis 1997). Linguistic landscape research investigates the functional-pragmatic meaning of public writing and the social negotiation cultures it is based on and which are verbalised using comparatively little language in the quantitative sense (cf. Schiedermeier 2015: 73). So far, there is almost no research on the influence of linguistic landscapes on spelling competence. The paper is organised as follows: we first discuss relevant previous research (Sect. 2). We then present a specifically designed corpus that includes an exemplary spelling phenomenon from one linguistic landscape (N-N compounds), which is analysed with respect to the frequency of wrong spellings. In the course of a second study, a field experiment, we analyse if wrong spellings in a linguistic landscape actually influence orthographic beginners (Sect. 3). We argue in favour of such an influence and discuss it in Sect. 4.

2 Linguistic Landscapes and Spelling Performance Commonly, poor reading competence and the habit of reading are not taken into account (for an overview, see Tacke 2000; Steinig et al. 2011; Siekmann 2011).

114

B. Rothstein

Indeed, early psycholinguistic studies from English-speaking countries claim that children who read well also show good spelling performance (Perfetti 1997). In the German research space, however, studies on the didactics of the German language (Siekmann 2011; Eichler 1986) and psychological studies (Scheerer-Neumann 1986; Wimmer et al. 2000) show that there is no reciprocity between reading and spelling performance. Other research at least suggests that reading may be one of several factors influencing spelling performance if the focus is not on the content of the text being read but on its spelling: Jacoby and Hollingshead (1990) provided evidence that reading may affect spelling performance by asking their test persons to correctly spell a number of words they had just read in a list in which these words were correctly or incorrectly spelled. Previously read wrong spellings affected the spelling of the words in question. Mecklenbräukler et al. (1995) provided evidence that even reading a word just one time may influence its later spelling if this word is embedded in a continuous text and if the focus of the test persons was on each individual word and not primarily on content. Accordingly, it seems that meaningextracting reading, i.e. reading with a focus on the propositional content of a text, does not affect spelling performance. Spelling has so far been marginalised by linguistic landscape research. With the exception of Herold et al. (2018), there is no research on how the reading of linguistic landscapes affects spelling performance. They provided evidence that the reading of texts in the linguistic landscape does affect spelling performance. Starting from the language-didactic question of whether wrong spelling in public spaces affects the spelling performance of orthographic beginners, they investigated whether linguistic landscapes influence the spelling performance of Germanspeaking third and sixth graders by means of a non-representative study. Herold et al. (2018) used correctly spelled loanwords that were topically known but orthographically unknown to the students. More specifically, they investigated the spelling of the words “Empathie” (empathy), “parallel” (parallel), “Skifahren” (skiing), “Rhythmus” (rhythm) and “Revier” (district) after the participants of the tests had been exposed to a linguistic landscape including all these words for four weeks. Participants in the tests were an intervention group of 22 third graders and 19 sixth graders, and a reference groups consisting of 10 third graders and 21 sixth graders. The intervention period was four weeks and included the following intervention material, which was displayed next to the blackboard (Fig. 2). There was no intervention for the control group. The cooperating teachers of both groups were instructed not to discuss the materials and to report all related classroom-activies. For the third graders, a significant difference concerning the acquisition of spelling skills in the cases of Revier (p = 0.7, d = .67) and Empathie (p = .05, d = .60) was found for the two groups of this form. Concerning the sixth graders, a significant difference between test and control group could only be stated concerning the spelling of Empathie (p = .06, d = .61). Based on this, Herold et al. (2018) concluded that effects of linguistic landscapes on the spelling performance of third and sixth graders cannot be ruled out, although the above-mentioned research does not identify any correlation between reading and spelling (Thomé 2008; among others).

Do Linguistic Landscapes Influence the Spelling Competence?

115

Fig. 2 Intervention material from Herold et al. (2018)

However, the study by Herold, Schuttkowski and Rothstein does not allow for statements on long-term effects because, for reasons of school logistics, a follow-up test was not possible. Also, the number of test persons, test items and orthographic phenomena is clearly too small for generalising statements. At best we can speak of a tendency: it seems that the spelling skills of orthography learners may be influenced by their reading of writing in public spaces, which may perhaps be explained by the frequency of reading particular landscape items. Usually, linguistic landscapes consist of extremely short texts which – due to their locations – are somewhat prominent in the public space and are perceived accordingly: perhaps the writings are read more frequently than texts outside the linguistic landscapes. In the following section, we therefore investigate frequency effects of texts inside linguistic landscapes on the spelling performance of orthographic beginners.

3 How Linguistic Landscapes Affect Spelling Performance 3.1 Hypothesis According to our hypothesis, wrong spellings in linguistic landscapes affect spelling performance. Thus, two assumptions must be proven: (i) that the occurrence of

116

B. Rothstein

wrong spellings in linguistic landscapes is sufficiently frequent (see Study 1 in Sect. 3.3), and (ii) that linguistic landscapes affect spelling performance (see Sect. 3.4: Study 2). First, however, in Sect. 3.2 we will present an orthographic case example, which is suitable for both studies.

3.2 Orthographic Case Example: Linking Positions of N-N Compounds The spelling of linking positions in N-N compounds is well suited for the issue pursued here. Due to rather inconsistent official regulations, they provide a great deal of potential for wrong/alternative spellings. In German, we find two ways of spelling the linking positions of N-N compounds. Spelling German N-N compounds without a hyphen or any other orthographic marker is the default (Autotür ‘car door’), but a hyphen can be used optionally: writers are free to use a hyphen in compounds with a loanword as the first constituent (Drogerie-Kette ‘drugstore chain’) and in those with one letter appearing three times in a row (Schiff-Fahrt ‘boat cruise’). To facilitate their reading, optional hyphens are also possible in N-N compounds in general, especially with neologisms and occasionalisms (Baumhaus-Tür ‘tree house door’). The hyphen becomes obligatory in the case of acronyms as the first constituent of a compound (U-Bahn ‘underground’) and in N-N compounds whose first element is a number (Fünf-Sekunden-Intervall ‘five-second-interval’). According to the official regulations (“Regeln und Wörterverzeichnis”), there must be a hyphen between all elements of multiple compounds including a word group or a compound with a hyphen (Fünf-Sekunden-Intervall-Angelegenheit ‘five-second-interval-issue’). The hyphen is obligatory in phrasal compounds such as Schau-mir-in-die-Augen-Blick ‘Look-me-in-the-eyes-look’ and compouns with acronyms (type: U-Bahn). Nonregular groups of compound spelling are: spelling with a dash (*Auto–tür ‘car door’), spelling with an inner-word majuscule (*AutoTür ‘car door’), spelling with a space (*Auto Tür ‘car door’) and spelling with an apostrophe (*Auto’tür ‘car door’), see Starke (1993), Barz (1993), Jacobs (2005), Buchmann (2015). To sum up, we find the following possibilities for spelling linking positions of N-N compounds in German. The distribution in Table 1 can be explained as follows: In German, the space and the dash are only permitted between but not within words (Jacobs 2005; Buchmann 2015). Capital letter writing is regulated by syntax, because majuscules only occur in sentence-initial position and on the heads of noun phrases. Taking a traditional perspective in separating morphology from syntax, the spelling of linking positions of German N-N compounds is sensitive to the morphological inner-word level, where syntax can’t interfere: N-N compounds form morphological words, the linking position trivially being part of them. This rules syntactic markers such as space, dash and inner-word majusculewriting out. In German, the hyphen is a morphological marker: it separates lexemes

Do Linguistic Landscapes Influence the Spelling Competence?

117

Table 1 Ways of spelling linking positions of German N-N compounds N-N compound spelling Without hyphen or other orthographic marker Space Optional hyphen

Obligatory hyphen

Dash Inner-word majuscule Apostrophe

Explanation Default (except cases with obligatory hyphen) Not regular In single/multiple compounds to provide easier understanding After acronyms/numbers; in multiple compounds with 1 obligatory hyphen; in phrasal compounds

Not regular Not regular Not regular

Example Autotür ‘car door’ *Auto Tür ‘car door’ Baumhaus-Tür ‘tree house door’

U-Bahn (underground), Fünf-Sekunden-Intervall (five second interval), Fünf-Sekunden-Intervall-Angelegenheit (five second interval issue), Schau-mir-in-die-Augen-Blick (look-me-in-the-eyes-look) *Auto–tür ‘car door’ *Auto Tür ‘car door’ *Auto’tür ‘car door’

belonging to the same lexical node. This explains all optional spellings with the hyphen. It also accounts for hyphenated writing of phrasal compounds including compounds with acronyms and numbers as first constituents of the compound. Phrasal compounds are compounds with maximal phrases as the first constituent, where the hyphen is used to indicate the different lexical parts of the compound. Thus, phrasal compounds are words that contain syntactic phrases. It also nicely explains why compounds with numbers as the first part need to be written with a hyphen: originally, they are phrases and somewhat comparable to phrasal compounds: Fünf-Sekunden-Intervall is based on fünf Sekunden. Hyphenated spelling with acronyms as the first constituent of a compound results from their long forms often being phrases themselves: ZDF-Nachrichten (‘ZDF News’) originates from Zweites-Deutsches-Fernsehen-Nachrichten (‘ News from the German TV channel ZDF’). But this also predicts that using a hyphen in the linking positions of N-N compounds would be the default. Why shouldn’t writers always use the hyphen instead of not using any orthographic marker in the linking position? Being optional, the hyphen seems to be restricted pragmatically in these cases: writers do have the choice between using or leaving out a hyphen. Consider that other concatenative word formation types in German like derivation normally do not make use of the hyphen to mark morpheme boundaries. Therefore, using no orthographic marker to indicate inner-word morpheme boundaries is the default. Writers may use the hyphen to avoid unclear readings of N-N compounds (e.g. Druckerzeugnis that is ambiguous between Drucker-Zeugnis (‘printer certificate’) vs. Druck-Erzeugnis (‘printed product’)). One way to think of this is to consider Gricean pragmatics: Traditionally speaking, N-N compounds with a hyphen are interpreted with reference to the maxim of modality (‘Be perspicuous. Avoid obscurity of expressen.

118

B. Rothstein

Avoid ambiguity.’). Using the hyphen to indicate morpheme boundaries instead of using no orthographic marker at all makes the lexical composition of N-N compounds more precise and – in cases of ambiguity – less ambiguous, because its lexical constituents become more easily recognisable to the reader: the presence of a hyphen underlines the interpretation of constituents as lexical parts of N-N compounds. This reference to the reader is also found in the official regulations (“Regeln und Wörterverzeichnis”): The hyphen provides the writer with the possibility of marking each element as such, of separating them from each other and in this way emphasising them for the reader, instead of writing them in one word as is otherwise common with compounds and derivations. (Regeln und Wörterverzeichnis § 45)

Consider, for instance, the ambiguous multiple N-N compound Rotweinglaskiste. It can be understood as a box for red wine glasses or as a glass box for red wine. By using a hyphen (Rotwein-Glaskiste oder Rotweinglas-Kiste), writers avoid this ambiguity.

3.3 Study 1: Frequency of Wrong Spelling in Linguistic Landscapes 3.3.1

Hypothesis

We hypothesize that the reading of wrong spellings in linguistic landscapes influences the spelling performance of orthographic beginners, if the reading happens sufficiently.

3.3.2

Method

As it is our general assumption that in particular the spelling performance of orthography learners (students of primary level) is influenced by wrong spellings in linguistic landscapes, our study simulated the daily routines of students outside school: among these are the way to school, TV programmes and Internet portals. We stipulated a way to school of 750 metres, that in the context of our study the students watched programmes of the German-language TV channels RTL and Sat 1, and that all the children’s visits to the Internet happen at the toggo.de portal for children and youths.1 These activities are hypothetical, but nevertheless plausible. We stipulated that the material we compiled covers about 20 school days. For our survey we applied the documentary method of linguistic landscape research, which

1 Toggo.de is a webpage run by the German television channel Super RTL. It particularly addresses

children and youths by presenting relevant television material and games. Toggo.de also contains commercials.

Do Linguistic Landscapes Influence the Spelling Competence?

119

includes the systematic collection of public, visually perceivable linguistic items in the form of digital photographs. All items were stored in a corpus.

3.3.3

Data Collection and Data Processing

There were three corpora: way to school, TV commercials and online portal. For the sub-corpus way to school, public writing with N-N compounds in the area under analysis were documented by means of a digital camera, were processed using Photoshop, and then included in the corpus in May 2016. The area under analysis consisted of two inner-city shopping streets in a randomly chosen, medium-sized city in the Ruhr region. Street A is about 550 metres long and is part of a pedestrian area where there only runs a tram line. Whereas in the lower part of the street there are in particular privately owned shops and takeaways, the upper part also hosts some bigger chain stores. Of Street B, only the first 200 meters bordering on Street A were covered. Similar to the lower part of Street A, there we also find mainly sole proprietorship shops. The data collection for the sub-corpus TV commercials took place in April and May, 2016, at different times of the day. 20 h each of programmes of two private TV channels (RTL and Sat 1) were recorded. The recording was done by means of an Apple Mac-Book-Air (Internet-based TV set allowing for recordings using Quick Time-Player) or a TV receiver (model: DIFITALBOX IMPERIAL HD 5 twin), from where the data were transferred to an external hard disc. Using Quick TimePlayer the TV commercials were cut out of the recordings. Screenshots allowed photographing of the relevant video material. Unreadable elements – due to a commercial running at too high a speed or to their being too small – were not included in the data. Also for the sub-corpus online portal screenshots were made and edited. The survey looked at www.toggo.de and covered a total of 4 days in April and May, 2016, at times of the day when there was no school. Also included were online commercials at toggo.de. The toggo.de portal was randomly chosen. All data were incorporated into a four-columned, table-like corpus with eight categories: in the first column the respective examples are consecutively numbered. The second column provides information about the fields of context, expression, correct, category, date, time, place and observer. Based the chosen field of data recording and the orthographic phenomenon to be analysed, the entries in the categories of context, category and place are field-specific. The context category refers to the respective environment where the linguistic material was identified. Date and place refer to the time and location of the photographic documentation. The observer is the person who documented the linguistic phenomenon. Expression gives the N-N compound as copied from the photograph. Under correct the compound is entered according to its regular spelling. If the example already represents a regular compound or if it is an unmarked hyphenated spelling, this column is marked with three asterisks (***). The last column gives the photographic documentation.

120

B. Rothstein

Table 2 N-N compound spellings in the orthographic landscape N-N compound spelling Without hyphen Space Optional hyphen Obligatory hyphen Dash Inner-word majuscule Total

3.3.4

Way to school 255 118 56 25 5 14 473

TV commercials 136 33 35 23 7 0 234

Toggo 161 13 50 4 26 0 254

Total 552 (57%) 164 (17%) 141 (15%) 52 (5%) 38 (4%) 14 (2%) 961

Data Assessment

The assessment of the corpus was conducted using the above-defined types of N-N compound spellings: without hyphen, space, optional hyphen, obligatory hyphen, dash and inner-word majuscule. Spellings with an apostrophe were not observed. The total number of N-N compounds is 961. The distribution of the respective subcorpuses as well as the overall corpus is structured as follows (Table 2). If we add up the wrong spellings, this means the groups dash, inner-word majuscule and space, 23% of the compounds are wrong spellings (that is almost every fourth compound). If we also add optional hyphens, the final share of wrong spellings or of those requiring an explanation is 38%. From this we can conclude that wrong spellings of N-N-compounds are frequent in the linguistic landscape we investigated. This takes us to the second study and to the question of whether and in which way any effects of wrong spellings in linguistic landscapes can actually be proven.

3.4 Study 2: Field Experiment on Perceiving Linguistic Landscapes 3.4.1

Hypothesis

I hypothesize that the reading of wrong spellings in linguistic landscapes affects the spelling performance of orthographic beginners.

3.4.2

Method

We applied parts of the method by Herold et al. (2018) to N-N compounds: We carried out a six-week field experiment according to the control group design, including pre-, post-, and follow-up tests: in the classrooms of the intervention group a poster with hyphenated spellings of N-N compounds was shown for 2 weeks, and in the classrooms of the control group a poster with the same N-N compounds,

Do Linguistic Landscapes Influence the Spelling Competence?

121

Fig. 3 Poster for test groups

however without hyphens, was shown over the same period of time. The teachers were instructed not to discuss the posters during the entire period of the project, if possible, and to inform us of anything noteworthy. The DIN A 1 posters were put up next to the blackboard after the pre-test and were taken away immediately before the post-test. As test material we used the words “Müll-Trennung” (waste separation), “Restmüll-Tonne” (residual waste bin), “Bio-Tonne” (organic waste bin), “PapierTonne” (paper waste bin), “Gelber Sack” (plastic waste bin) and “Glas-Container” (bottle bank) / “Mülltrennung”, “Restmülltonne”, “Biotonne”, “Papiertonne”, “Gelber Sack” and “Glascontainer” on the following posters. We are aware of the fact that the number of test items is very small, however due to the limited number of posters we could use more words were not possible. All the spellings we tested are regular ones (see Sect. 3.2), however the hyphenated spellings are unusual and require explanation: they might be considered alternative and thus wrong spellings (Figs. 3 and 4). Pre-, post- and follow-up tests were identical and consisted of the following dictation exercise which the cooperating teachers dictated without hyphens: “In Deutschland wird Müll getrennt. Es gibt Biotonnen, Restmülltonnen, Papiertonnen, den gelben Sack und Glascontainer.”

122

B. Rothstein

Fig. 4 Poster for control groups

In Germany, there is waste separation. There are bins for organic waste, for residual waste, for paper waste, for plastic waste and for bottles.

The test was anonymous. At the request of the cooperating schools, no personal data were recorded. We will come back to this in the discussion.

3.4.3

Test Persons, Data Recording and Data Processing

The test participants were third graders, as it may be assumed that the effect of linguistic landscapes on spelling is particularly significant among orthographic beginners who, according to the curriculum, have already acquired basic reading skills but have not yet completed their spelling acquisition. According to the curriculum, the relevant orthographic phenomenon (N-N compound spellings with hyphen) has not yet been acquired in this form. 128 test persons from 6 third form classes of two different primary schools in the Ruhr region took part in the experiment. Only 92 test persons were considered, as the other 36 tests persons had missed at least one test or – according to information from the teachers – had insufficient German language skills (N = 92). As the students were not

Do Linguistic Landscapes Influence the Spelling Competence?

123

coincidentally randomised but the distribution happened according to classes, the two groups at the two schools were not of the same size (34 test persons for intervention, 58 students for the comparison). The test took place in May, June and July, 2018. There were no incidents.

3.4.4

Data Assessment

Although we are interested in hyphens, the spelling of the complete words was also considered to see whether there are significant effects on word spelling as a whole. Following Herold et al. (2018), a scoring system was worked out for the assessment of the dictations which gives the degree of correct spelling in terms of numbers, allowing us to identify any differences between the tests. The letter was chosen as the relevant linguistic entity; syllables and graphemes were left out of consideration. The principles were as follows (Table 3). In this context, not only letters not present in the orthographically correct spelling of the word were counted as wrong but also those being correctly used for spelling the word but whose positions were wrong. The following maximum scores for each word result from the total number of its letters plus one point for capitalisation/use of small letters (Table 4). With α = .879, the alpha coefficient proved to be good or even excellent. Hyphenated spellings were counted separately, as 0 (no hyphen) or as 1 with a hyphen. On average, the tests persons achieved the following values for word spelling and hyphenated spelling given maximum scores of 51 points respectively 1 point (Table 5). Table 6 shows total numbers of hyphenated spelling. Only the intervention group used hyphens. To calculate differences between the test results, the difference between the individual scores of the results achieved for individual words or hyphenated spellings during the different tests was calculated. Table 7 shows the mean differences for the

Table 3 Scoring system for the assessment Principles 1 2 3 4

Score 1 point plus 1 point plus 1 point less 1 point less

Table 4 Maximum score for test item

Explanation For each correct letter in the correct sequence For the correct capitalisation or the correct use of small letters For each wrong letter For each wrong separate spelling Word Biotonnen Restmülltonnen Papiertonnen Glascontainer

Points Max. 10 points Max. 15 points Max. 12 points Max. 14 points

124

B. Rothstein

Table 5 Average score of the pre-, post- and follow-up tests Pre-test M (SD) 42,99 (5,7) 0 (0)

Word spelling Hyphenated spelling

Post-test M (SD) 45,04 (5,1) 0,35 (0,96)

Follow-up test M (SD) 45,48 (5,00) 0,38 (1,01)

Table 6 Distribution of hyphens in the intervention group in total numbers Pre-test Post-test Follow-up test

Bio-Tonne 0 8 11

Restmüll-Tonne 0 5 11

Papier-Tonne 0 6 9

Glas-Container 0 5 2

Table 7 Mean differences between the test results on word spellings Control group (N = 58)

Intervention Group (N = 34)

Total (N = 92)

Difference Post-pre

Biotonnen Restmülltonnen Papiertonnen Glascontainer 0,53 0,38 0,65 1,62

Follow-up-post 0,15 Follow-up-pre 0,68 Post-pre 0,31 Follow-up-post Follow-up-pre Post-pre Follow-up-post Follow-up-pre

0,02 0,33 0,39 0,07 0,46

0,65 1,03 0,21

0,18 0,82 0,14

−0,09 1,53 0,74

−0,28 −0,07 0,27 0,07 0,34

0,02 0,16 0,33 0,08 0,40

0,41 1,16 1,07 0,29 1,29

test results on word spellings, both for the intervention and control group students. Consider, for instance, the mean difference between the post-test and the pre-test test for the spelling of “Biotonnen”. It was calculated by substracting the spelling score that students achieved for “Biotonnen” in the pre-test from the spelling score that the same students achieved for “Biotonnen” in the post-test. Table 8 gives the mean differences for the test results on hyphens. Due to the limited size of the random sample and lacking standard distribution, for the comparative analysis of the mean values of both groups we made use of the Mann-Whitney-U test with an error rate α of 5% to assess the difference between the significance of each group (Table 9). The only significant improvements occurred in the case of “Glascontainer” when the post- was compared to the pre-test (p < 0.00964) and in the case of “Restmülltonne” with the follow-up compared both to the post-test and to the pretest (p < .01684 and p < .01828, respectively). Concerning hyphenated spelling, there were no significant effects.

Do Linguistic Landscapes Influence the Spelling Competence?

125

Table 8 Mean differences between the test results on hyphens Difference Control group Post-pre (N = 58) Follow-up-post Follow-up-pre Intervention Post-pre group Follow-up-post (N = 34) Follow-up-pre Total (N = 92) Post-pre Follow-up-post Follow-up-pre

Bio-Tonnen Restmüll-Tonnen Papier-Tonnen Glas-Container 0 0 0 0 0 0 0,17

0 0 0,16

0 0 0,10

0 0 0,12

0,05 0,22 0,11 0,03 0,14

0,03 0,19 0,10 0,02 0,12

0,05 0,16 0,07 0,03 0,10

−0,09 0,03 0,08 −0,05 0,022

4 Discussion and Summary Study 1 investigated the frequency of wrong spellings in linguistic landscapes. To investigate this, we made use of a corpus which was collected using the documentary method of studying linguistic landscapes. It is certainly a problem that the structure of the corpus that Study 1 is based on is stipulated (we may only guess the actual leisure time activities of the concerned group of test persons). Due to the restricted structure of the corpus, focusing on just one orthographic phenomenon, and to the comparably small amount of data, Study 1 is not representative. It is remarkable, however, that almost every fourth N-N compound spelling in the analysed corpus was wrong. However, on the basis of Study 1 it could not be investigated whether these wrong spellings were noticed at all. Therefore, we conducted study 2, a six-week field experiment with control group design and pre-, post-, and follow-up tests: the intervention group was exposed during two weeks to a poster with hyphenated spellings of the following compounds: “Müll-Trennung” (waste separation), “Restmüll-Tonne” (residual waste bin), “Bio-Tonne” (organic waste bin), “Papier-Tonne” (paper waste bin), “Gelber Sack” (plastic waste bin) and “Glas-Container” (bottle bank). In the classrooms of the control group a poster with the same N-N compounds, however without hyphens, was shown over the same period of time. Contrary to expectations, only in the case of one of the tested words did there occur any effects immediately after intervention (Glascontainer, when the post-test was compared to the pretest (p < 0.00964)). However, these effects did not last beyond the post-test. Significant effects could be proven for Restmülltonne when comparing the followup to the post-test (p < .01684) and when comparing the follow-up to the pre-test (p < 001828), however not when comparing the pre- to the post-test. Although there was a sufficient frequency of wrong spellings in the linguistic landscape, it does not seem to have played any significant role. Does this mean that – contrary to Herold et al. (2018) – there is no evidence that linguistic landscapes affect the spelling

U-value Z-value p-value

U-value Z-value p-value

U-value Z-value p-value

Biotonnen Bio-Tonnen Restmüll tonnen Difference between follow-up and post-test 909 935 690 −0.62881 0.4085 −23,903 .53526 .6818 .01684 Difference between follow-up and pre-test 800,5 782 693,5 −149,647 164,611 −236,199 .13362 .09894 .01828 Difference between post- and pre-test 821,5 833 870 −13,266 123,357 −0,932,428 .09176 .10935 .17619

Table 9 Mann-Whitney-U test, Z-values and p-values

867 0,95,855 .16853

799 15,086 .13104

918 0.54601 .58232

Restmüll-Tonnen

748,5 −19,171 .02743

760 −182,407 .6876

950,5 −0.28312 .77948

Papier tonnen

884 0,82,104 .20611

833 123,357 .2187

935 0.4085 .6818

Papier-Tonnen

696,5 −233.773 .00964

851,5 −108,393 .28014

859,5 101,922 .30772

Glas container

884 0,82,104 .20611

952 0,27,098 .78716

918 −0.54601 .58232

Glas-Container

126 B. Rothstein

Do Linguistic Landscapes Influence the Spelling Competence?

127

performance of orthographic beginners? To answer this question, the following three points have to been considered. Firstly, it may be that the posters we used were not liked by the test persons for topical and/or design reasons. In other words, it may be that these posters did not attract the attention of the test persons or did not invite them to read: here, we would need a data-set that would allow motivation, interest, actual reading of the linguistic landscape, and reading skills to be included in the calculation of the results. It is generally known from studies on spelling acquisition that interest in particular supports the development of spelling performance (Brüggelmann 1990; Richter 1998). Unfortunately, these data could not be recorded in the context of Study 2 due to a request of the schools. Secondly, in the case of future tests, the multi-modality of linguistic landscapes, i.e. roughly their combinations of images and texts, would have to be taken into consideration: multi-modal texts are read differently from those with exclusively linguistic components. Ethnological applications of multi-modal text-linguistics provide evidence for non-lingual text components being highly relevant for the understanding and interest-guided perception of multi-modal texts (Staubach 2017); however, for the time being an investigation of their effects on spelling performance is still a research desideratum which is indeed urgent given the above-mentioned text composition, which has changed towards multi-modality, and the reading behaviour of young people, who increasingly read multi-modal digital texts. That multi-modal texts are particularly interesting for students is shown by most recent text-linguistic studies (Staubach 2017). Accordingly, it would be necessary to check whether the image elements attract the students’ attention and perhaps make the reading of the text elements redundant (this may be the case, for example, with the posters used here, so that perhaps the design of the texts will have to be changed). Thirdly, the texts were not connected to any action context, i.e. the authentic feature of reading a text, which is so typical for linguistic landscapes, is missing: typically, the texts of linguistic landscapes are of an inviting nature; in the case of regulatory texts this refers, for instance, to writing informing about traffic rules which must be followed, in the case of commercial texts it is “calls to buy”. The texts used for Study 2 were presented within a context which was not authentic: typically, one would expect information about waste separation to be presented by linguistic landscapes in the vicinity of the appropriate waste bins; the classrooms where Study 2 was carried out, however, did not allow for waste separation by means of different bins. As a consequence, the texts used for Study 2 have no function and are not authentic. It is generally known from language-didactic studies on the function and authenticity of linguistic signs that these are factors influencing attention (Betz et al. 2016). For these three reasons we believe the results of Study 2 neither clearly confirm the effect of linguistic landscapes on spelling performance, nor do they really refute it. To sum up, the two studies presented here do not rule out that wrong spellings in linguistic landscapes may affect the spelling performance of orthographic beginners. With respect to our introductory example: it is not impossible that orthographic

128

B. Rothstein

beginners reading “riesich” and “Taschen Bücher” are influenced when it comes to spelling these words. In this regard our results neither confirm nor refute the results of Herold et al. (2018). Given the massive evidence against correlations between reading and spelling performance (Thomé 2008; among others), this fact alone is a result worthy of being published. However, it has become obvious from the discussion that for future studies on the influence of linguistic landscapes, the variables of interest, motivation, reading skills, multi-modality, function and authenticity will have to be taken into consideration if we want to gain final insights into what is going on within this domain. While interest, motivation, and reading skills have already been analysed as factors influencing the spelling performance of different age groups, a general assessment of the influence of multi-modality, authenticity, and function on spelling performance is still lacking. However, as these are essential features of linguistic landscapes (large parts of linguistic landscapes consist of multi-modal texts belonging to authentic action contexts with an informative function) and as those learning how to spell are constantly surrounded by linguistic landscapes, taking these features into consideration is a true research desideratum.

References Barz, I. (1993). Graphische Varianten bei der substantivischen Komposition. Deutsch als Fremdsprache, 30, 167–172. Betz, A., Flake, S., Mierwald, M., & Vanderbeke, M. (2016). Modelling authenticity in teaching and learning contexts: A contribution to theory development and empirical investigation of the construct. In C.-K. Looi, J. Polman, U. Cress, & P. Reimann (Eds.), Transforming learning, empowering learners: The International Conference of the Learning Sciences (ICLS) 2016 (Vol. 2, pp. 815–818). Singapore: International Society of the Learning Sciences. Brüggelmann, H. (1990). Rechtschreibung. Kinder lernen in qualitativen Sprüngen. Spektrum der Wissenschaft, 1, 26–29. Buchmann, F. (2015). Die Wortzeichen im Deutschen. Heidelberg: Winter. Eichler, W. (1986). Zu Uta Frith’s Dreiphasenmodell des Lesen (und Schreiben) Lernen. Oder: Lassen sich verschiedene Modelle des Schrifterwerbs aufeinander beziehen und weiterentwickeln? In G. Augst (Ed.), New trends in graphemics and orthography (pp. 234–247). Berlin: de Gruyter. Flaschberger, L., & Reiterer, A. F. (1980). Der tägliche Abwehrkampf. Kärntens Slowenen. Wien: Braunmüller. Gal, S. (1979). Language shift: Social determinants of linguistic change in bilingual Austria. San Francisco: Academic. Herold, I., Schuttkowski, C., & Rothstein, B. (2018). Orthographie und Linguistische Landschaften. Wie beeinflussen öffentliche Fehlschreibungen die Rechtschreibleistungen? Deutsch als Fremdsprache, 3, 154–159. Jacobs, J. (2005). Spatien. Zum System der Getrennt- und Zusammenschreibung im heutigen Deutsch. Berlin: de Gruyter. Jacoby, L., & Hollingshead, A. (1990). Reading student essays may be hazardous to your spelling: Effects of reading incorrectly spelled words. Canadian Journal of Psychology, 44, 345–358. Landry, R., & Bourhis, R. (1997). Linguistic landscape and ethnolinguistic vitality: An empirical study. Journal of Language and Social Psychology, 19, 23–49.

Do Linguistic Landscapes Influence the Spelling Competence?

129

Mecklenbräukler, S., Wippich, W., & Kary, R. (1995). Beeinflußt das Lesen richtig und falsch geschriebener Wörter die spätere Rechtschreibung? In S. Filipp, H. Mandl, et al. (Eds.), Zeitschrift für Entwicklungspsychologie und Pädagogische Psychologie. (27/4) (pp. 323–338). Göttingen: Hogrefe. Perfetti, C. A. (1997). The psycholinguistics of spelling and reading. In C. A. Perfetti et al. (Eds.), Learning to spell: Research, theory and practice across languages (pp. 6–44). Hillsdale: Erlbaum. Regeln und Wörterverzeichnis: Deutsche Rechtschreibung. Regeln und Wörterverzeichnis. Entsprechend der Empfehlung des Rats für deutsche Rechtschreibung. Überarbeitete Fassung des amtlichen Regelwerks 2004. München, Mannheim. http://www.rechtschreibrat.com/. Last Accessed on 6 Mar 2017. Richter, S. (1998). Interessenbezogenes Rechtschreiblernen. Methodischer Leitfaden für den Rechtschreibunterricht in der Grundschule. Westermann: Braunschweig. Römer, R. (1968). Die Sprache der Anzeigenwerbung. Düsseldorf: Schwann. Scheerer-Neumann, G. (1986). Wortspezifisch: Ja – Wortbild: nein. In H. Brüggelmann (Ed.), ABC und Schriftsprache. Rätsel für Kinder, Lehrer und Forscher (pp. 171–185). Konstanz: Faude. Schiedermeier, S. (2015). Überlegungen zur Kulturvermittlung im Fach Deutsch als Fremdsprache. Linguistic Landscapes und Erinnerungsorte. In C. Badstübner-Kizik & A. Hille (Eds.), Kulturelles Gedächtnis und Erinnerungsorte im hochschuldidaktischen Kontext. Perspektiven für das Fach Deutsch als Fremdsprache (pp. 65–81). Frankfurt: Lang. Schuster, B.-M., & Tophinke, D. (2012). Andersschreiben als Gegenstand der linguistischen Forschung. In B.-M. Schuster & D. Tophinke (Eds.), Andersschreiben: Formen, Funktionen, Traditionen (pp. 13–18). Berlin: Erich Schmidt. Siekmann, K. (2011). Der Zusammenhang von Lesen und (Recht-)Schreiben. Frankfurt: Lang. Starke, G. (1993). Droht uns eine Bindestrich-Inflation? In Muttersprache (Vol. 103, pp. 50–60). Staubach, K. (2017). Multimodale Sehflächen lesen. Eine semiotische Analyse jugendlicher Bekleidung. Zeitschrift für Angewandte Linguistik, 66, 31–58. Steinig, W., et al. (2011). Schreiben von Kindern im diachronen Vergleich – Texte von Viertklässlern aus den Jahren 1972 und 2002. Münster: Waxmann. Tacke, G. (2000). Die Lese-Rechtschreibleistungen werden immer schlechter – Tatsache oder Legende? Unterrichtswissenschaft, 28, 304–333. Thomé, G. (2008). Transfer-Effekt überschätzt. Ergebnisse aus der DESI-Studie zum Zusammenhang von Lesen und Schreiben. Praxis Deutsch, 210, 58–59. Wimmer, H., Mayringer, H., & Landerl, K. (2000). The double deficit hypothesis and difficulties in learning to read a regular orthography. Journal of Educational Psychology, 90, 668–680. Ziegler, E., Schmitz, U., & Uslucan, H.-H. (2018). Attitudes towards visual multilingualism in the linguistic landscape of the Ruhr Area. In M. Pütz & N. Mundt (Eds.), Expanding the linguistic landscape: Multilingualism, language policy and the use of space as a semiotic resource. Bristol: Multilingual Matters.

Part III

Second/Third Language Settings

The Present Tense in English, Again Amber Dudley and Roumyana Slabakova

Abstract We report on an experimental study examining what aspectual tense forms we use to convey aspectual meanings when talking about present events in English. We test the effect of structural priming on the use of aspectual tense morphosyntax in the English present tense by native speakers, upper-intermediate and advanced L2 learners of English with French as their native language. Comparative production data from a video retell task is used. Aspectual choices from two previous studies by Liszka are compared with our partial replication. While Liszka primed participants to use the progressive tense, our instructions were neutral in this respect. Findings for native speakers point to a high level of individual variation in the use of present progressive and present simple to denote events simultaneous with the speech moment. Not only are choices variable, but they are also influenced by priming. We argue that this variability creates difficulties for learners of English that teachers should know about. Keywords Aspect · L1 influence · Priming · Production · Present tense · English

1 Introduction The number of studies investigating the second language (L2) development of the Tense-Aspect-Mood system has grown exponentially over the past two decades, with the vast majority focusing on past temporality (Salaberry et al. 2013; Ayoun 2015). Much less consideration has been given to present tenses and their meanings

A. Dudley () University of Southampton, Southampton, UK R. Slabakova University of Southampton, Southampton, UK NTNU The Norwegian University of Science and Technology, Trondheim, Norway e-mail: [email protected] © Springer Nature Switzerland AG 2020 A. Trotzke, T. Kupisch (eds.), Formal Linguistics and Language Education, Educational Linguistics 43, https://doi.org/10.1007/978-3-030-39257-4_8

133

134

A. Dudley and R. Slabakova

(Ayoun 2015). In this chapter, we will turn our attention to present tense sentences such as Jill is eating a sandwich and Jill eats a sandwich in the morning, in order to explore the mapping between aspectual meaning and its morphosyntactic representation (i.e., aspectual tense morphosyntax). In this paper, the term “aspectual meaning” refers to semantic information about the temporal structure of an event or state relative to its temporal location (Comrie 1976). It is perhaps a truism accepted by most teachers of English that Jill is eating a sandwich denotes an ongoing event while Jill eats a sandwich in the morning denotes a habitual event, repeated with some frequency. We will show that the mapping between these meanings and forms is far from uniform in the language use of native speakers of English and learners of English. In addition, we will show that this usage is variable and discuss the subsequent pedagogical implications. Researchers have used a wide range of experimental techniques, including the use of comprehension and production data, in order to investigate this phenomenon. Relatively few studies, however, have directly considered the influence of certain elicitation techniques on both native speaker and L2 production and the implications for questions of L2 development and ultimate attainment. Thus, this chapter sets out to address the current gap in the literature by comparing production data from two studies: Liszka (2009, 2015), which we designate as Study One, and our partial replication of that study, which we call Study Two. The crucial difference between the two studies is the inclusion of structural priming (that is, the use of the present progressive form in the pre-task instructions) in Study One and its exclusion in Study Two. More specifically, we examine the distribution of forms and meanings when talking about present events in English, and the effect of structural priming on the use of aspectual tense morphosyntax in the English present tense by native speakers, upper-intermediate and advanced L2 learners of English, whose first language is French. This chapter is structured as follows. In Sect. 2, we describe the linguistic property in question and its formal expression in English. In Sect. 3, we outline the equivalent forms’ expression in the participants’ L2. In Sect. 4, we review previous studies which explore the L2 development of aspectual tense morphology and sensitivity to structural priming. Then, in Sect. 5, we outline the methodology used in the present study before presenting the statistical analyses of our results. Finally, we discuss our findings in relation to previous studies and the wider theoretical debate in the field of L2 acquisition.

2 Linguistic Property and Its Treatment The current study focuses on the obligatory aspectual contrast between the simple and progressive forms in the English present tense. In (1), the simple verb pattern (V-s) gives rise to the habitual/generic interpretation of the eventuality, ‘to work at home’, whereas in (2), the complex verb pattern involving the auxiliary verb to be and the participle V-ing form triggers a progressive situational interpretation of the

The Present Tense in English, Again

135

same eventuality. The sentences introduced by # are considered inappropriate with the supplied adverbial. (1) (2)

She works at home on Mondays. #She is working at home on Mondays. She is working at home (right now). #She works at home (right now).

Numerous theories have been advanced to account for this aspectual distinction from multiple theoretical frameworks, including the generative, discourse representation and cognitive grammar paradigms, to name just a few. The current study adopts one of many generative approaches to the analysis of the aspectual distinction in the English present tense. In this section, we provide a brief introduction to the generative approach before discussing its application to the analysis of aspectual forms in the English present tense. Readers who are already familiar with the generative framework are encouraged to skip Sect. 2.2.

2.1 Introduction to Generative Terminology and Its Application to Tense This sub-section aims to provide a brief outline of the relevant terminology for readers unfamiliar with the generative paradigm. Within the generative framework, it is argued that verbs carry several types of morphosyntactic features. These features are properties of words and can explain not only morphological and semantic differences, but also syntactic differences (Adger 2003). One example would be the tense feature [past]. This feature is typically represented morphologically by adding the affix -ed, or in some cases -t, to the verb stem. On a semantic level, it is interpreted as describing an event that took place at a time prior to the utterance. Current generative theory maintains that the aspectual contrast between the progressive and simple forms in the English present tense is the result of the interpretability of these features and the semantic effects triggered by verb raising (i.e. v-to-T movement), an operation of narrow syntax (Hawkins et al. 2008; Déchaine and Manfredi 2000; Adger 2003). Interpretability in this context refers to whether a morphosyntactic feature conveys semantic meaning or not. The plural -s morpheme on English nouns is said to be an expression of an interpretable feature because it generates semantic meaning (i.e., more than one), whereas uninterpretable features, such as gender agreement for determiners and adjectives, do not express any semantic meaning (Adger 2003). v-to-T movement refers to the syntactic operation whereby a finite main verb from the head v position of vP moves into the head I position of IP (Adger 2003). However, as we will see, not all languages allow such an operation due to the strength of the uninterpretable inflectional feature, [uInfl:] of v.

136

A. Dudley and R. Slabakova

Feature interpretability thus plays a critical role in the derivation of a sentence; that is, the process of generating a sentence in line with the syntactic rules of a specific language. For a derivation to converge (i.e., to generate a well-formed expression) at the logical form (LF), uninterpretable features must be checked and deleted from the derivation and interpretable features must be valued. If uninterpretable features are not deleted from the derivation, then it will crash (i.e., fail to generate a well-formed expression).

2.2 Generative Analysis of the Present Simple Form As we briefly mentioned previously, some languages, such as English, do not allow v-to-T raising. Adger (2003) explains that this is due to the syntactic operation of agreement, involving the interpretable features [present], [past], and [Prog(ressive)], and an uninterpretable feature [uInfl:] associated with v (Adger 2003). When the interpretable features are valued, the [uInfl:] feature is deleted from the derivation. Déchaine and Manfredi (2000) maintain that in English, T has two uninterpretable features: [V] and [AGR]. Thematic verbs in English cannot raise from v-to-T because v has a weak [uInfl:] feature, which means that v does not have to be checked locally and is valued in situ (Liszka 2015). The instantiation of the uninterpretable feature, [AGR], thus prevents T(ense) from being interpreted on the basis of the aspectual properties of the verb (Déchaine and Manfredi 2000). Because thematic v-to-T raising is not licensed in English, this triggers the generic/habitual interpretation of the present simple form (Hawkins et al. 2008). Notice the link between lack of main verb movement, a syntactic property, and habitual interpretation, a semantic property. However, this is just one possible analysis. Other authors, such as Adger (2003) a.o., do not link verb movement with interpretation.

2.3 Generative Analysis of the Present Progressive Form Déchaine and Manfredi’s present tense analysis contrasts with the progressive form analysis. In minimalist syntax, the progressive involves two simultaneous components: a deictic component (i.e. the auxiliary to be) and an aspectual component (the participle, V-ing) (Liszka 2009, 2015). The deictic component expresses an event-in-progress at Speech Time, whereas the aspectual component conveys the semantic property of ongoingness. According to Adger (2003), the functional category [Prog(ressive)] underlies the progressive form in English. [Prog(ressive)] has a strong [uInfl:] feature, which causes the auxiliary verb to raise to T since strong features must be checked locally (Adger 2003). The obligatory v-to-T raising of the auxiliary verb is argued to give rise to the event-in-progress interpretation of the progressive form (Hawkins et al. 2008; Liszka 2009).

The Present Tense in English, Again

137

Hawkins et al. (2008) stress that a habitual/generic interpretation of the progressive form is not possible since the interpretable feature [Prog] sets the value of the [uInfl:] feature of v as [uInfl:Prog]. The semantic properties of the aspectual interpretive component, ‘ongoingness’, is encoded morphosyntactically using the participle V-ing, where V represents a thematic verb. In this and the previous subsections, we presented one analysis which links verb movement and interpretation, thereby explaining the prototypical form–meaning mappings.

2.4 Non-prototypical Interpretations of Present Tense Forms in English However, the prototypical meaning–form mappings are not the end of the story. In addition to the prototypical interpretations, there are also highly-marked contexts, such as (oral) narrative commentaries, where the two surface forms can overlap in aspectual meaning. In certain types of oral narratives, namely narrative (sports) commentaries, the semantic dichotomy between the simple and progressive form is cancelled out (Williams 2002). Consequently, the simple form often adopts some of the functional-semantic domain of the progressive form (Vraicu 2015). This use is often referred to as “the reportive use” of the present simple. For example, in (3a), the present simple is used to describe a series of events that are more or less concomitant to the utterance time. In (3b), the described events are complete and happened in the recent past. This usage is termed “the perfective use” of the present. (3) a. “He hits it into the hole. Jeter makes a nice stop. He fires to first, and gets him by a step.” (Langacker 2011: 60) b. “So I go to pay for my sandwich, and the guy asks me for two pounds – but I don’t have any money on me!”

REPORTIVE PERFECTIVE

These additional meanings have been analysed from different theoretical perspectives. Boundedness theories of aspect use aspectual operators in order to refer to temporal boundaries of eventualities. For example, Cowper (1998) refers to the reportive reading of the present simple as the instantaneous present. She argues that the eventuality the present simple describes should not be understood as concomitant with the moment of speech, as this would oppose the principle of nonsimultaneity of points, that is, the concept that tense or any other functional element in a language cannot enforce simultaneity on two temporal points. Instead, these eventualities should be understood as occurring over an interval that acts as the discourse anchor for the event. The moment of speech can only act as an interval when both the speaker and the hearer adopt a close view of the speech event. This means that the speaker allows the hearer to experience the speech event and the eventualities conveyed by the utterance “with the same immediacy that the speaker experienced” (Cowper 1998 cited in Vraicu 2015: 295). It is exactly this closeness to

138

A. Dudley and R. Slabakova

the eventuality, or vividness, that is characteristic of the reportive use of the present simple, particularly in narrative commentaries. Narrative commentaries refer to situations where the speaker cannot control the action that is unfolding before their eyes. In such contexts, the present simple form performs two roles (Vraicu 2015; Williams 2002): (1) to heighten “the dramatic immediacy of the narrative material” and (2) to advance the storyline (Vraicu 2015: 297). This means that if the speaker were to use the progressive, it would suggest the speaker is only interested in describing a sequence of simultaneous events, but not in heightening the dramatic immediacy of the narrative material. According to Vraicu (2015), the use of the progressive form would mean that the situation the progressive describes is unbounded, which thus anchors it in the deictic present of the speaker. This in turn instantiates a subjective viewpoint of the situation. As Williams (2002) points out, in other commentaries, such as play-by-play descriptions of sporting as in sporting events, commentators use the progressive form for subjective comments because it highlights that the speaker’s value judgments or the eventualities it describes do not form part of the anticipated events. This consequently suggests that, when a speaker uses the present simple to narrate a series of narrative or sporting events, the role of the speaker is backgrounded, allowing the reader to experience the events as though they were present. Chuquet (1994) maintains that the timelessness of the present simple means that English present-based narratives are not attached to “an identifiable deictic reference”, which in turn instantiates “an anaphoric-type of linkage with respect to a fictitious enunciative origo” (Vraicu 2015: 298). The present simple consequently imposes a holistic viewpoint, which is congruous with relations of sequentiality that exist between the events that advance the storyline. In this context, the progressive is used in present-based narratives to trigger an internal viewpoint which halts the chronology of the storyline, thus bringing about a close-up effect at a particular point in the storyline. In summary, the present simple has a reportive and a perfective meaning in addition to the habitual meaning. We do not discuss the perfective meaning further in this study. Although describing seemingly ongoing events, the reportive is also a present perfective, in the sense that a perfective construal of an eventuality or state implies that the speaker views the described situation in its entirety at the time of speaking (Smith 1991). Sportscaster events are of that (very short) granularity. This meaning is triggered by a special context and has special pragmatic functions. The form–meaning mappings of the present tenses in English appear to be as in Fig. 1. One form, the present simple, has three meanings; at the same time, the ongoing meaning can be denoted by two forms. This learning situation is bound to create difficulty for learners of English.

The Present Tense in English, Again

139

Fig. 1 Relationship between form and meaning for English present tenses

3 Grammatical Aspect in French 3.1 French Aspectual Tenses In order to better understand the acquisition task at hand, we must first discuss how the aforementioned interpretations are expressed in French, the L2 learners’ L1. The main aspectual distinction in French is between perfectivity and imperfectivity in the past tense, marked morphologically using the passé composé and the imparfait, respectively (Smith 1997). Unlike English, progressive aspect is not grammaticalised in French; it is typically expressed using the simple present or the progressive periphrastic expression, être en train de ‘to be in the middle of doing something + infinitive’ (Mortier 2008). This periphrasis carries a similar meaning to the English progressive form since it expresses semantic information about a dynamic event in progress (Ayoun 2005). Unlike the English progressive, this periphrasis can only be used with eventive predicates, not stative predicates (Mortier 2008). Since its use is not in diametrical opposition to other viewpoints, the progressive periphrasis does not form part of the French aspectual viewpoint system (Smith 1991). As we have previously established, English uses the progressive to refer to an event that is simultaneous with the speech time. In contrast to its English counterpart, the simple present, the présent, can also express semantic information about an event that is in progress (M.A. Jones 1996). As Mortier (2008) points out, the distinction between the progressive periphrastic expression and the présent is often minimal. For example, the use of the periphrasis in (5) serves to highlight the durative nature of an event and thus avoid any aspectual ambiguity resulting from the présent (Mortier 2008). Example (4) may be interpreted as describing an action simultaneous with the speech time or a general habit. French, however, has several temporal adverbs, such as en ce moment in (6), it uses to avoid a habitual interpretation.

140

(4)

(5)

(6)

A. Dudley and R. Slabakova

Elle joue du she play.3SING.PRS some ‘She plays/is playing the piano’ Elle est en she is.3SING.PRS in ‘She is playing the piano.’ En ce moment, elle joue at the moment, she play.3SING.PRS ‘At the moment, she is playing the piano.’

piano. piano. train process du some

de of

jouer play

du some

piano. piano.

piano. piano.

3.2 Narrative Commentaries in French As discussed previously, the English present simple sometimes overlaps in functional-semantic scope with the progressive form in present-based narratives, where the simple form can refer to an event that is simultaneous with the utterance time (Williams 2002; Vraicu 2015). French present-based narratives, however, contrast with their English counterparts. Chuquet (1994) argues that the principal distinction between the English simple present and the French présent lies in the perfective aspectual value of the English simple present. According to Chuquet (1994), the timelessness of the English simple present establishes a holistic view on the eventuality it describes, consequently resulting in the perfective aspectual interpretation. Events are typically anchored deictically through the use of the progressive form given its imperfective aspectual meaning. As discussed in the preceding section, the French présent lacks the progressive/simple aspectual distinction, which means the présent’s primary function is to serve as a tense form since it anchors eventualities in time in relation to the speech time. Given the semantic indeterminacy of the présent, French present-based narratives therefore opt for deictic anchoring. Such deictic anchors include the semantic properties of the predicates, conjunctions such as donc, temporal adverbs and subordinates (Vraicu 2015). As we have already mentioned, the progressive/simple aspectual contrast enables the speaker to establish a hierarchy between plot-advancing and collateral material in the narrative and develop a framing effect. This is not possible in French given the open-endedness of the default interpretation of eventualities in the présent (Vraicu 2015). Consequently, speakers must use other devices, such as conjunctions, temporal adverbials and the semantic properties of verbal predicates, to create a similar framing effect. Moreover, French does not have access to the same internal viewpoint due to the absence of an overt aspectual contrast as in English (Vraicu 2015). While Chuquet (1994) argues that the progressive periphrastic expression être en train de ‘be in the process of’ allows a more subjective viewpoint, its distribution is much more limited and less systematic than the English progressive (Vraicu 2015).

The Present Tense in English, Again

141

4 L2 Acquisition of Aspect in the English Present Tense This section provides an overview of previous studies looking at the L2 development of the aspectual distinctions in the English present tense, focusing in particular on the key theoretical debates and previous L2 studies on the development of acquisition and the effect of structural priming.

4.1 The Interpretability Hypothesis As we will see in Sect. 4.2, several previous studies investigating the L2 development of this linguistic property (Hawkins et al. 2008; Liszka 2009, 2015; Al-Thubaiti 2015) have adopted the representational deficit view as a standpoint. The representational deficit view of L2 acquisition maintains that the lack of convergence L2 learners demonstrate during the advanced stages of L2 acquisition results from an impairment to the language faculty. Numerous theories, including the Fundamental Difference Hypothesis (Bley-Vroman 1989, 1990), the Representational Deficit Hypothesis (Hawkins 2001) and the Failed Functional Features Hypothesis (Franceschina 2001; Hawkins and Chan 1997) have been subsumed under this approach. The most recent instantiation is the Interpretability Hypothesis (Hawkins and Hattori 2006; Tsimpli and Dimitrakopoulou 2007). Proponents differentiate between interpretable and uninterpretable features, as defined above. The Interpretability Hypothesis states that L2 learners will experience learnability issues when acquiring uninterpretable features not activated in their L1. Interpretable features, on the other hand, are not predicted to be problematic. Empirical support for this hypothesis, however, remains divided with some studies providing supporting evidence (Tsimpli and Dimitrakopoulou 2007; Franceschina 2001, 2005) and others opposing evidence (e.g., Leal and Slabakova 2014; Dominguez 2007; Foucart and Frenck-Mestre 2012; Hopp 2012; Sagarra and Herschensohn 2013). Slabakova (2015: 680) argues that the primary linguistic input to which L2 learners are exposed plays a crucial role in the acquisition of features and that learners must be exposed to a significant level of unambiguous input for successful acquisition to take place.

4.2 Previous Research on Aspectual Tenses Many studies have examined the L2 development of the progressive/simple aspectual contrast in the English present tense (Slabakova 2003; Gabriele and Canales 2011), with several studies using it as a test case for the much-debated Interpretability Hypothesis (Hawkins et al. 2008; Liszka 2009, 2015; Al-Thubaiti 2015). A common theme in most of these studies is the observation that advanced L2 learners,

142

A. Dudley and R. Slabakova

despite initial development delays, can ultimately acquire the distributional and interpretational properties of the progressive/simple aspectual contrast. In contrast to the aforementioned studies, Liszka (2009, 2015) argues that advanced Frenchspeaking L2 learners of English suffer from a representation deficit in this domain, due to a purported inability to map the ongoing semantic interpretation to the progressive form. Liszka’s experiment involved two oral tasks and one written task. Task One consisted of an untimed picture description of a busy street scene; Task Two was a 9-min video clip description and Task Three––a contextualized dialogue task. Of particular interest to the current study is Task Two, the video clip description, given that it is the most likely of the three tasks to generate the reportive interpretation. Furthermore, findings from this task contributed significant support to Liszka’s proposal of a syntactic deficit. In this task, participants were asked to simultaneously watch and (orally) describe a video clip from the popular TV series, Mr Bean. In order to ensure that participants did not use the historic present form, participants were primed before the task with the following instructions: “You are going to watch a TV programme and I would like you to describe what is happening on the screen at the same time as you are watching”. Findings from the (2009) study showed that while L2 learners could use both forms productively, they made a significant proportion of errors when producing the progressive in ongoing contexts. Liszka consequently argued that this residual optionality may be the result of a permanent L2 deficit stemming from the parametric differences between English and French. According to Liszka, the French native speakers were unable to revalue the uninterpretable feature, [Infl:], from strong to weak, meaning that their mental grammars allowed English thematic verbs to raise to [Infl:], thus triggering an event-in-progress interpretation of the present simple. Liszka (2015) later replicated the 2009 study using an identical battery of test instruments. The crucial difference between the 2009 and 2015 was that the L2 learners in the 2015 study were resident in the target language community (i.e. the UK) and those in the 2009 study were primarily instructed learners resident in France. The L2 learners were classed as very advanced L2 English speakers on the basis of an independent proficiency test. Mirroring the 2009 study, the 2015 findings suggest that very advanced L2 learners demonstrated difficulty and optionality when matching the progressive form to the correct semantic interpretation in the online production task. The 2015 participants, however, performed markedly better in the offline task in comparison to the 2009 participants. We must, however, interpret Liszka’s (2009, 2015) findings with caution given the experimental design of the study. Firstly, when coding the data from the video clip description task, Liszka specifically stated that the use of the present progressive with thematic verbs was coded as appropriate, but the use of the present simple with thematic verbs was only coded as appropriate when the context was established with an adverbial expression. Such an analysis assumes a one-to-one mapping between form and meaning, that is, an ongoing semantic interpretation can only be mapped onto the present progressive form, but not the present simple.

The Present Tense in English, Again

143

As discussed above, the present simple is often used in present-based narratives and (sports) commentaries to denote an event that is simultaneous with the speech time. Although Liszka (2015: 71) recognises this highly-marked use of the simple form, she argues the instructions given to informants were “clear and unambiguous [enough] in order to prime the production of descriptions” and not reportive narratives. In other words, Liszka used syntactic priming to ensure that speakers would produce the progressive form (more on this in the next section). Liszka (2009, 2015) interpreted these findings as evidence to suggest that the persistent form-to-meaning mismatches in progressive contexts were due to a (permanent) syntactic deficit. We, however, contend that such a proposal may be too strong a formulation. Slabakova (2015: 680) argues that the primary linguistic input to which L2 learners are exposed plays a crucial role in acquiring grammatical features, and that learners must be exposed to an abundancy of “unambiguous evidence in the input” to successfully reset parametric values. In the case of progressive semantics, the input is inconsistent. For example, in certain contexts, native speakers can alternate between the progressive and simple form to denote an event that is simultaneous with the time of utterance. Therefore, it is not surprising that despite the syntactic priming, L2 learners demonstrated persistent form-tomeaning mismatches given the inconsistent input.

4.3 Structural Priming Research In light of these contradictory findings, relative to previous research, we must further investigate the extent to which syntactic priming has an effect on speakers’ aspectual choices. The phenomenon of syntactic priming has gained widespread interest in the psycholinguistics of language production. Structural priming (or syntactic) priming refers to “the phenomenon whereby comprehension or production of an utterance is facilitated by previous comprehension or production of an utterance with the same (or a related) structure” (Flett et al. 2013: 752). For example, Bock (1986), the seminal priming study, showed that the likelihood of participants describing a picture with an active construction (e.g., Lightning is striking the church) increased when participants echoed an unconnected active sentence (e.g., One of the fans punched the referee), compared with a situation where they echoed the corresponding passive sentence (e.g., The referee was punched by one of the fans). Such behaviour suggests that speakers have access to abstract syntactic representations within a language and that using a particular syntactic construction when uttering a sentence can facilitate the utterance of that construction in subsequent production (Flett et al. 2013: 752). In addition, the effect of syntactic priming has also found empirical support among the L2 population (e.g., Bernolet et al. 2007; Kantola and van Gompel 2011; Schoonbaert et al. 2007; Flett et al. 2013; Gries and Kootstra 2017). However, while the effects of priming are robust and widespread

144

A. Dudley and R. Slabakova

among L1 speakers, the L2 research is unfortunately not as extensive and thus necessitates further testing with a wider range of structures.

5 The Present Study 5.1 Research Questions and Methodology In light of the literature reviewed above, this study aims to explore the following research questions: • What aspectual tense morphosyntax do native English speakers use when asked to report an event simultaneous with the moment of speech? In other words, what is the linguistic input French learners of English are exposed to, in order to learn this form–meaning correspondence? • To what extent does the presence or absence of structural priming in the instructions influence L1 and L2 speakers’ subsequent production of aspectual tense morphosyntax? • To what extent can L2 learners’ optionality in aspectual use be considered an artefact of a representational deficit? In order to address the above research questions, we combine data from the video description task in Liszka’s (2009, 2015) studies (Study One) as well as our own partial replication of that task (Study Two). In Study One, participants were told: “You are going to watch a TV programme and I would like you to describe what is happening on the screen at the same time as you are watching”. (Italics were not included in the actual instructions). In Study Two, however, the participants watched the same 9-min video clip, but this time the instructions were simplified in that participants were told: “Describe the events orally at the same time as the video.” No progressive forms appeared in the instructions. During the clip, the audio from the original recording was muted. Since the task was designed to be an online, real-time processing task, participants were not allowed to operate the recording equipment or pause the extract at any point during the task. The 9-min clip, taken from the well-known British television series, The Return of Mr Bean, was identical to the one used in Liszka (2009, 2015). In the comedy sketch, Mr Bean is celebrating his birthday at a French restaurant. At the beginning, we see Mr Bean writing a birthday card to himself. After looking at the menu and checking how much money he has, Mr Bean chooses the cheapest item on the menu, the steak tartare. Steak tartare is a French dish consisting of finely chopped raw beef. Mr Bean is a little surprised at first, but it soon becomes apparent that he does not like the meat despite his best attempts to eat it. Mr Bean then tries to hide the meat in various places, including a bread roll, a sugar pot and a lady’s handbag. The clip ends with the manager apologising to Mr Bean after a waiter trips and spills food over his table.

The Present Tense in English, Again

145

Data analysis involved multiple stages. Firstly, the data was transcribed and coded for verbal morphology. Then we followed the coding procedure described in Liszka’s (2009, 2015) by differentiating between obligatory present progressive contexts (OC Present Progressive) and obligatory present simple contexts (OC Present Simple). In OC Present Progressive, the use of the present progressive with dynamic thematic verbs was coded as acceptable, whereas the use of the present simple when not accompanied by licensing adverbial expressions, such as then, suddenly, as soon as, etc., was coded as unacceptable. In OC Present Simple, the use of the present simple with stative verbs (e.g., know, want, feel), copular forms (e.g., am/are/is + adjective) and/or dynamic thematic verbs when used with the aforementioned licensing adverbial expressions was coded as appropriate, whereas the use of the present progressive in these contexts was coded as inappropriate. All verbs were coded in this task, but verbal forms other than the present progressive and the present simple were excluded from the analysis. Finally, we calculated the percentage scores for the distribution of the two forms in OC Present Simple and OC Present Progressive for individual participants. The individual percentage scores were then conflated to provide group means for each speaker group. Samples of speech from the Oral Production Task are included in (7) and (8) to give the reader a flavour of the data and coding procedure. (7) Native Speaker of English: and the man sat at the table is continuing to tap the glasses with his knife (present progressive) and then the waiter reappears (present simple) and he is taking a napkin from the table (present progressive) and laying it over the man’s lap (present progressive) and then the man picks up the napkin (present simple) (8)

French-speaking L2 Learner: and the man sits down (#present simple) he pulls a card from his pocket (#present simple) opens it (#present simple) and writes [.] something on it (#present simple) he closes it (#present simple)

5.2 Participants 5.2.1

Participants: Study One

Participants in Study One were taken from Liszka’s (2009, 2015) studies and included 16 advanced French-speaking L2 learners of English and 5 native speakers of British English. All L2 learners completed an in-house proficiency test at the author’s university and were considered to be of advanced, and in some cases near-

146

A. Dudley and R. Slabakova

native, proficiency with varying degrees of naturalistic exposure and input from instruction. Although the 2015 L2 participants were slightly more advanced and had received more naturalistic exposure than the 2009 L2 participants, we decided to conflate the two L2 participant groups given their comparable aspectual choices in the video description task. For ease of comparability with Study Two’s measure of proficiency, we entered these L2 learners’ proficiency scores as 36/40 in order to account for their advanced proficiency based on the independent measure used in Study Two.

5.2.2

Participants: Study Two

Participants in Study Two were taken from our partial replication and included 15 upper-intermediate-to-advanced French-speaking L2 learners of English and 15 native speakers of British English. All L2 participants completed an independent proficiency test (Slabakova 2003) with an average score of 29.60/40 and a standard deviation of 6.06. While we recognise that participants in Study One and Two were not matched in proficiency and this may ultimately weaken the strength of any resulting conclusions, we took this into account when running statistical analyses. At the time of testing, the L2 participants were either resident in France as students or teachers of English or had been resident in the United Kingdom for no more than 3 months.

6 Results In this section, we examine participants’ aspectual choices (present progressive, present simple) in each obligatory context and the effect of aspectual priming in the task instructions. Section 5 above describes the obligatory present progressive and present simple contexts used in the coding in more depth. Results from the two studies are presented visually in Fig. 2 and summarised in Table 1. In Fig. 2, the left-hand column in each cell represents aspectual tense choice under no priming conditions, while the column on the right represents the same choices under priming with the present progressive (Liszka’s studies). A simple visual inspection of Fig. 2 suggests that most speakers, regardless of L1 or task type, used the present simple in obligatory present simple contexts (the top two cells), with very low rates of the present progressive relative to the present simple. However, we find an observable interaction between L1 groups and task type in obligatory present progressive (the bottom two cells). This difference is particularly notable among L1 English speakers, whose present simple use dramatically increases in the absence of syntactic priming with the progressive. For example, the L1 controls used the present progressive in just 48.53% of obligatory progressive contexts when not primed, compared with 99.14% when primed with it.

The Present Tense in English, Again

147

Fig. 2 Aspectual use in production

In order to analyse these observations further and explore statistically whether aspectual choice (present progressive, present simple) depends on L1 (English, French), proficiency, obligatory context (present progressive, present simple) and/or task (presence of priming, absence of priming), we ran a binomial logistic regression analysis, using the glm() function in R. We used the main effect structure and a model comparison tool, the likelihood ratio test (ANOVA() in R), in order to construct a statistical model that best described speakers’ aspectual use. A positive estimate (β) indicates an increase in the use of the present progressive relative to the present simple, whereas a negative value suggests a decrease. The critical values from our logistic regression model are summarised above in Table 2. We found a main effect for all four predictors (L1, task, context and proficiency) as well as several reliable two-way interactions (L1: Context, L1: Context, Task: Context) and a significant three-way interaction (L1: Task: Context). Note that positive B values indicate an increase in the likelihood of using the present progressive over the present simple and a negative value represents a decrease. The presence of a three-way interaction (L1: task: context) supersedes the significance of the previously reported individual predictors and two-way interactions. Our results thus suggest that L1, task and context all interact in order to modulate aspectual

148

A. Dudley and R. Slabakova

Table 1 Summary of aspectual tense choice Task Priming

Context OC present simple

Form Present simple Present progressive

OC present progressive

Present progressive Present simple

Non-priming

OC present simple

Present simple Present progressive

OC present progressive

Present progressive Present simple

First Language (L1) English French 280/281 499/526 (99.64%) (94.87%) 1/281 27/526 (0.36%) (5.13%) 347/350 304/555 (99.14%) (54.77%) 3/350 251/555 (0.86%) (45.23%) 591/613 456/461 (96.41%) (98.92%) 22/613 5/461 (3.59%) (1.08%) 645/1329 377/777 (48.53%) (48.52%) 684/1329 400/777 (51.47%) (51.48%)

Table 2 Regression model for whole dataset Intercept L1: English vs. French Task: Non-priming vs. priming Proficiency Context: OC present simple vs. OC present progressive L1: Task L1: Context Task: Context L1: Task: Context

B(SE) −1.83 (0.50) −1.63 (0.52) −2.34 (1.03) −0.04 (0.01) 3.23 (0.22) 4.13 (1.14) 1.24 (0.51) 7.15 (1.18) −8.52 (1.28)

z value −3.63 −3.15 −2.29 −3.22 14.43 3.62 2.45 6.07 −6.65

p-value < 0.001 < 0.005 < 0.05 < 0.005 < 0.001 < 0.001 < 0.05 < 0.001 < 0.001

tense use. In other words, when primed in OC Present Progressive, French speakers’ use of the present progressive increases. Given the contrast in findings between the priming and non-priming task specifically in OC Present Progressive, we decided to run L1-specific models on OC Present Progressive, as presented in Table 3. We find that English speakers’ use of the present progressive increases after priming, and so does French speakers’, but not to the same extent. In addition, proficiency also modulates French speakers’ aspectual tense use, but not English speakers’, resulting in a significant decrease in present progressive use. Finally, in order to explore the individual variation further, Fig. 3 plots the proportion of present progressive use in OC Present Progressive (the simultaneous event

The Present Tense in English, Again

149

Table 3 Regression model for OC present progressive by L1

Intercept Task: Non-priming vs. priming Proficiency

B(SE) −0.06 (0.05) 4.81 (0.58)

English French z-value p-value B(SE) z-value p-value −1.07 0.29 1.03 (0.35) 2.96 < 0.005 8.26 < 0.001 0.42 (0.12) 3.38 < 0.001 −0.04 (0.01) −3.20 < 0.005

Fig. 3 Individual proportions of present progressive use in simultaneous event contexts

meaning) in the non-priming task (Study Two). The x-axis shows proficiency scores while the y-axis shows the percentage of present progressive use. Of particular note in Fig. 3 is the distinct variation in aspectual tense use among the English native speakers. It is clear that there are some native speakers who produce limited tokens of the progressive, whereas others use the present progressive predominantly. In addition, we also find speakers who freely alternate between the two forms. As we will explore in the discussion section, this has important implications for language acquisition, as it does not provide a consistent native speaker input for L2 learners.

150

A. Dudley and R. Slabakova

7 Discussion This study set out to investigate native and learner use of the present progressive and the present simple in describing simultaneous events in a video description task. The actual, and not the traditionally taught, usage of the aspectual tenses constitutes the linguistic input to the learners who are acquiring aspectual tense meanings in English. A secondary goal was to establish the extent to which structural priming in task instructions modulates native and non-native speakers’ choices of aspectual tense forms. A third objective was to examine the validity of Liszka’s (2009, 2015) claim that a representational deficit could account for L2 learners’ difficulty in mapping the simultaneous interpretation to the corresponding progressive form; that is, whether the underlying syntactic structure of an L2 learner’s L1 inhibits her from acquiring a target-like interpretation of the linguistic structures under investigation. To this end, we replicated the video description task from Liszka (2009, 2015), but this time removing the tokens of the present progressive in the task instructions to eliminate any effects of structural priming on subsequent production. We then compared our dataset with that of Liszka (2009, 2015). To summarise the findings, this study provides evidence to suggest that when describing ongoing events simultaneous with the moment of speech, that is, present events, L1 as well as L2 speakers of English are likely to use both the present simple and the present progressive tense. However, when describing habitual events, speakers predominately use the present simple. These form-meaning mappings are thus in contrast: one allows optionality of form while the other does not. A particularly striking aspect of our dataset was the level of individual variation among speakers in OC Present Progressive when not primed. It appears that native speakers of English license both the present simple and the present progressive in this context, with some speakers exclusively using the present simple or the present progressive and others a combination of the two forms. Such a finding has important implications for the role of input in L2 acquisition because it suggests that in naturally occurring conversation, native speakers of English license both aspectual tenses, almost in free alternation, to describe ongoing or simultaneous events in the present tense. This finding thus runs contrary to traditional English language teaching, which assumes a simple one-to-one mapping between form and meaning. Moreover, many L2A scholars have argued that the primary linguistic data to which L2 learners are exposed is crucial and that exposure to abundant and unambiguous evidence is essential for successful acquisition (e.g., Rankin and Unsworth 2016; Slabakova 2013). In the context of our findings, this suggests that the input to which L2 learners are exposed is not sufficient in order to fully reset the required form–meaning mappings, as they are continually faced with inconsistent input, thus further complicating the already challenging learning task. In addition, in our present experiment, there was not a consistent baseline with which to compare the L2 speaker data, given the significant individual variation among L1 speakers.

The Present Tense in English, Again

151

Evaluating our second objective, we found that structural priming significantly modulated both native and non-native use of aspectual tense forms in oral production, especially in OC Present Progressive, but also in OC Present Simple contrary to expectations. However, the presence of structural priming alone does not fully explain speakers’ aspectual tense choice, as we will see in the proceeding discussion. Our statistical analyses revealed that French speakers in OC Present Progressive were significantly less likely to use the present progressive over the present simple when compared with English native speakers. A more in-depth analysis of aspectual tense use within L1 groups in OC Present Progressive revealed that English native speakers use of the present progressive increases dramatically in the presence of priming relative to the absence of priming. The L2 learners’ use of the present progressive also increased in OC Present Progressive, but not to the same effect as evidenced by the smaller coefficient estimate. It is also worth briefly turning our attention to aspectual tense use in OC Present Simple. An unexpected aspect of our dataset was the fact that the presence of structural priming also significantly modulated aspectual tense choice in OC Present Simple, albeit not to the same degree. In those contexts, we found that in the presence of priming, the likelihood of a native speaker using the present progressive actually decreased. In addition, we found a significant difference between English and French speakers; our results suggest that in OC Present Simple, French speakers are more likely to use the present progressive over the present simple. Such a finding suggests that French speakers are influenced by priming, and are more conservative with their use of the present progressive in present simple contexts. This is to be expected because, while the present progressive can combine quite naturally with stative predicates such as love and hate, traditional pedagogical grammars do not tend to advocate this use as much as the more prototypical uses. Overall, our findings, specifically the significant effect of structural priming and speakers’ L1, have important implications for current debates in the field of L2 acquisition, in particular L2 sensitivity to structural priming, ultimate attainment, language pedagogy and experimental design. With respect to L2 sensitivity to structural priming, our study provides further empirical evidence in support of L1 and L2 speakers’ sensitivity to structural priming. It is important, however, to view sensitivity not as a dichotomous concept, but rather operating on a spectrum. In the context of our study, L2 speakers, like L1 speakers, were sensitive to the structural priming, but not to the same extent. This sensitivity was also modulated by additional factors, such as proficiency. This finding is thus in line with previous studies attesting L2 sensitivity to structural priming (e.g., Bernolet et al. 2007; Kantola and van Gompel 2011; Schoonbaert et al. 2007; Flett et al. 2013). A related objective of the current study was to investigate the validity of Liszka’s (2009, 2015) claim that L2 speakers’ residual optionality in their aspectual tense choice was indicative of a representational deficit. Based on our findings, we argue that there is no substantial evidence to suggest that the L2 optionality can be attributed to a representational deficit that stems from the structural differences between English and French. Instead, we propose that several additional factors, in

152

A. Dudley and R. Slabakova

particular L2 sensitivity to structural priming as well as the nature of the L2 input, provide a more robust explanation for why L2 speakers’ aspectual tense use is not always target-like. Based on the findings presented in this chapter and the resulting discussion, we conclude that aspectual present tense expression can constitute a significant difficulty for learners of English. The traditionally instructed one-to-one mapping between form and meaning is not supported by the usage on the ground. The present simple tense corresponds to at least three distinct meanings: habitual, perfective and reportive. When describing a series of events simultaneous with the present moment, native speakers use both present simple and present progressive. Teachers of the language should be aware of this situation. In this study, we probed the reportive meaning, while we leave the perfective for future research. In conclusion, together with the optional expression of the reportive, structural priming had a significant impact on the subsequent production of both native and non-native speakers. Taking into account the experimental task in question and additional contributing factors, we argue that there is no sufficient evidence that a representational deficit is the source of L2 speakers’ residual optionality in Liszka’s (2009, 2015) study. Instead, we believe that L2 speakers’ optional usage of the progressive and simple forms results from their reduced (but not lack of) sensitivity to structural priming as well as the ambiguous evidence of the form–meaning mapping in English. It is important to underline, however, that learners, in parallel with native speakers, fluctuate with the two tenses much more in the reportive than in the habitual contexts. Finally, we encourage future researchers to exercise caution when designing oral production tasks, especially those targeting the L2 population, that include instructions featuring structural priming. As our study has shown, minor adjustments, e.g., the inclusion of three tokens of the present progressive, can significantly impact both native and non-native speaker production.

References Adger, D. (2003). Core syntax: A minimalist syntax. Oxford: Oxford University Press. Al-Thubaiti, K. A. (2015). L2 acquisition of English aspect by L1 Arabic speakers: The role of interpretable features at the syntax-semantics interface. In D. Ayoun (Ed.), The acquisition of the present (pp. 185–214). Amsterdam: John Benjamins Publishing Company. Ayoun, D. (2005). The acquisition of tense and aspect in L2 French from a universal grammar perspective. In D. Ayoun & R. Salaberry (Eds.), Tense and aspect in romance languages: Theoretical and applied perspectives (pp. 79–127). Amsterdam: John Benjamins. Ayoun, D. (2015). The acquisition of the present. Amsterdam: John Benjamins Publishing Company. Bernolet, S., Hartsuiker, R. J., & Pickering, M. J. (2007). Shared syntactic representations in bilinguals: Evidence for the role of word order repetition. Journal of Experimental Psychology: Learning, Memory, and Cognition., 33, 931–949. Bley-Vroman, R. (1989). What is the logical problem of foreign language acquisition? In S. Gass & J. Schachter (Eds.), Linguistic perspectives on second language acquisition (pp. 41–64). Cambridge: Cambridge University Press.

The Present Tense in English, Again

153

Bley-Vroman, R. (1990). The logical problem of foreign language learning. Language Analysis, 20, 3–49. Bock, J. K. (1986). Syntactic persistence in language production. Cognitive Psychology., 18, 355– 387. Chuquet, H. (1994). Le présent de narration en anglais et en français. Paris: Ophrys. Comrie, B. (1976). Aspect: An introduction to the study of verbal aspect and related problems. Cambridge: Cambridge University Press. Cowper, E. (1998). The simple present tense in English: A unified treatment. Studia Linguistica, 52(1), 1–18. Déchaine, R., & Manfredi, V. (2000). Interpreting null tense. Paper presented at the Round Table on the syntax of tense and aspect, November 2000, University of Paris. Dominguez, L. (2007). Knowledge of features in fossilized second language grammars. Second Language Research, 23(2), 243–260. Flett, S., Branigan, H. P., & Pickering, M. J. (2013). Are non-native structural preferences affected by native language preferences? Bilingualism: Language and Cognition, 16, 751–760. https:// doi.org/10.1017/s1366728912000594. Foucart, A., & Frenck-Mestre, C. (2012). Can late L2 learners acquire new grammatical features? Evidence from ERPs and eye-tracking. Journal of Memory and Language, 66, 226–248. Franceschina, F. (2001). Morphological or syntactic deficits in near-native speakers? An assessment of some current proposals. Second Language Research, 17(3), 213–247. Franceschina, F. (2005). Fossilized second language grammars: The acquisition of grammatical gender. Amsterdam: John Benjamins Publishing Company. Gabriele, A., & Canales, A. (2011). No time like the present: Examining transfer at the interfaces in second language acquisition. Lingua, 121, 670–687. Gries, S. T., & Kootstra, G. J. (2017). Structural priming within and across languages: A corpusbased perspective. Bilingualism: Language and Cognition., 20(2), 235–250. Hawkins, R. (2001). Second language syntax: A generative introduction. Oxford: Blackwell. Hawkins, R., & Chan, C. (1997). The partial availability of universal grammar in second language acquisition: The ‘failed functional features hypothesis’. Second Language Research, 13, 187– 226. Hawkins, R., & Hattori, H. (2006). Interpretation of English multiple wh-questions by Japanese speakers: A missing uninterpretable feature account. Second Language Research, 22(3), 269– 301. Hawkins, R., Casillas, G., Hattori, H., Hawthorne, J., Husted, R., Lozano, C., Okamoto, A., Thomas, E., & Yamada, K. (2008). The semantic effects of verb raising and its consequences in second language grammars. In J. Liceras, H. Zobi, & H. Goodluck (Eds.), The role of formal features in language acquisition (pp. 333–355). Mahwah: Lawrence Erlbaum. Hopp, H. (2012). Grammatical gender in adult L2 acquisition: Relations between lexical and syntactic variability. Second Language Research, 29(1), 33–56. Kantola, L., & van Gompel, R. P. (2011). Between- and within- language priming is the same: Evidence for shared bilingual syntactic representations. Memory & Cognition, 39, 276–290. Langacker, R. W. (2011). The English present: Temporal coincidence vs. epistemic immediacy. In A. Patard & F. Brisard (Eds.), Cognitive approaches to tense, aspect and epistemic modality (pp. 45–86). Amsterdam: John Benjamins Publishing Company. Leal, T. M., & Slabakova, R. (2014). The interpretability hypothesis again: A partial replication of Tsimpli and Dimitrakopoulou 2007. International Journal of Bilingualism, 18(6), 537–557. Liszka, S. A. (2009). Associating meaning to form in advanced L2 speakers: An investigation into the acquisition of the English present simple and present progressive. In N. Snape, Y.-K. I. Leung, & M. Sharwood Smith (Eds.), Representational deficits in SLA: Studies in honor of Roger Hawkins (pp. 229–246). Amsterdam: John Benjamins Publishing Company. Liszka, S. A. (2015). The L2 acquisition of the English present simple – Present progressive distinction: Verb-raising revisited. In D. Ayoun (Ed.), The acquisition of the present (pp. 57– 85). Amsterdam: John Benjamins Publishing Company.

154

A. Dudley and R. Slabakova

Mortier, L. (2008). An analysis of progressive aspect in French and Dutch in terms of variation and specialization. Language in Contrast, 8(1), 1–20. Rankin, T., & Unsworth, S. (2016). Beyond poverty: Engaging with input in generative SLA. Second Language Research., 32(4), 563–572. Sagarra, N., & Herschensohn, J. (2013). Processing of gender and number agreement in late Spanish bilinguals. International Journal of Bilingualism, 17(5), 607–627. Salaberry, M. R., Comajoan, L., & González, P. (2013). Integrating the analyses of tense and aspect across research and methodological frameworks. In M. R. Salaberry & L. Comajoan (Eds.), Research design and methodology in studies on L2 tense and aspect (pp. 423–450). Berlin: De Gruyter Mouton. Schoonbaert, S., Hartsuiker, R. J., & Pickering, M. J. (2007). The representation of lexical and syntactic information in bilingauls: Evidence from syntactic priming. Journal of Memory and Language, 56(2), 153–171. Slabakova, R. (2003). Semantic evidence for functional categories in interlanguage grammars. Second Language Research, 19(1), 42–75. Slabakova, R. (2013). Adult second language acquisition: A selective overview with a focus on the learner linguistic system. Linguistic Approaches to Bilingualism, 3(1), 48–72. Slabakova, R. (2015). The effect of construction frequency and native transfer on second language knowledge of the syntax-discourse interface. Applied Psycholinguistics, 36(3), 671–699. Smith, C. S. (1991). The parameter of aspect. Dordrecht: Kluwer. Smith, C. S. (1997). The parameter of aspect. Revised second edition. Dordrecht: Kluwer. Tsimpli, I. M., & Dimitrakopoulou, M. (2007). The interpretability hypothesis: Evidence from wh-interrogatives in second language acquisition. Second Language Research, 23, 215–242. Vraicu, A. (2015). The simple present and the expression of temporality in L1 English and L2 English oral narratives: When form meets discourse. In D. Ayoun (Ed.), The acquisition of the present (pp. 289–334). Amsterdam: John Benjamins Publishing Company. Williams, C. (2002). Non-progressive aspect in English in commentaries and demonstrations using the present tense. Journal of Pragmatics, 34, 1235–1256.

Post-instruction Processing of Generics in English by Japanese L2 Learners Neal Snape

Abstract Over the past couple of decades there has been a concentrated effort to look at the acquisition of articles by adult second language (L2) learners of English. Many studies have tended to focus on whether adult L2 learners can successfully acquire articles, whilst other studies have conducted interventions to determine if instruction can be an effective way to facilitate acquisition though explicit learning. However, more often than not, in studies that implement a pre-test > post-test > delayed post-test design, there is no real way of ascertaining whether any of the explicit knowledge gained through instruction can change underlying linguistic competence. The current study addresses the question of whether explicit instruction can affect real-time reading times by administering a self-paced reading task. The aim of using a self-paced reading task is to find out whether instruction leads to faster or slower reading times for the instruction group of definite kind sentences, e.g., The dodo is extinct, compared with the non-instruction group. The findings reveal that the instruction group is faster in their reading times compared with the non-instruction group, from the critical region (kind predicate, e.g., be extinct) onwards, for the definite kind condition. We discuss the implications of our findings for the Competing Systems Hypothesis (Rothman J, Lang Contrast 8:74–106, 2008). Keywords L2 acquisition · Competing Systems Hypothesis · On-line processing · Intervention · Post-instruction · Kind · Generic · Articles

1 Introduction Different methods are available to researchers to try and assess whether second language (L2) learners can comprehend complex semantics, such as off-line tasks like forced-choice elicitation tasks, acceptability judgement tasks, picture-

N. Snape () Department of English Communication, Faculty of International Communication, Gunma Prefectural Women’s University, Sawa-gun, Gunma Prefecture, Japan e-mail: [email protected] © Springer Nature Switzerland AG 2020 A. Trotzke, T. Kupisch (eds.), Formal Linguistics and Language Education, Educational Linguistics 43, https://doi.org/10.1007/978-3-030-39257-4_9

155

156

N. Snape

sentence matching tasks and truth-value judgement tasks (Montrul and Ionin 2012; Snape 2018). Alternatively, on-line tasks are increasingly being used to measure judgements in real-time, such as self-paced reading, self-paced listening and eyetracking. The current study is a continuation of an intervention study conducted by Snape et al. (2016) and Umeda et al. (2019) who administered an off-line task to Japanese L2 learners of English and provided several weeks of instruction on articles. For this additional study, an on-line self-paced reading task was given to the same participants reported in Snape et al. (2016) and Umeda et al. (2019) to test whether instructed L2 learners can improve in their judgements of a certain type of generic known as ‘kind’, e.g., The dodo is extinct, after instruction. In light of the recent resurgence on whether instruction can be of benefit to L2 learners, there have been a number of studies that have examined the acquisition of articles, adjectival modifiers, null subjects, verb movement by L2 learners (Whong et al. 2013) and an increasing focus on instruction of morphology and semantics with regards to the articles the and a (Lopez and Sabir 2019; Snape and Yusa 2013). Typically, when instruction in the L2 classroom is discussed, many language teachers and second language researchers alike will raise the issue of whether any negative evidence in the form of explicit instruction can facilitate acquisition somehow (Trahey and White 1993). Whether explicit instruction can change underlying linguistic competence is debatable (see VanPatten 2016; VanPatten and Rothman 2015) since there is no way of knowing if such a transition has taken place.1 Rothman (2008) argues that the Competing Systems Hypothesis predicts differences in L2 learners’ interlanguage grammars, contingent upon how they acquired the L2. If an L2 learner acquired the L2 via pedagogical rules, it is possible that the resulting interlanguage grammar produces target-deviant L2 performance (depending on the domain acquired) as it is in competition with the generative system. However, it is not clear whether some type of intervention would coursecorrect any target-deviant path during the acquisition process. With this in mind, the current study aims to show whether instruction can lead to short-term advances and tries to, in part, address the Competing Systems Hypothesis. The way we test the Competing Systems Hypothesis is by administering an on-line task in the form of a self-paced reading task after instruction had already been delivered. This paper is structured in the following approach. Section 2 discusses the role explicit instruction can offer L2 learners and whether the knowledge gained changes underlying linguistic structure. Section 3 reviews definitions of genericity and L2 off-line studies of genericity as this was the focus of instruction given to our participants. Section 4 presents findings from on-line studies of articles as we implement a self-paced reading task to measure reading times in our study. Section 5 outlines some key intervention study findings of article instruction that are relevant to our intervention. Section 6 describes the motivation for the current small-scale

1 There

are certain tools that can be used to try and detect implicit knowledge such as eye tracking or fMRI technologies. Chamorro et al. (2016) argue that online and offline comprehension data can lead to different results and it is important to study both to understand the whole picture.

Post-instruction Processing of Generics in English by Japanese L2 Learners

157

study, which is part of a larger study by Snape et al. (2016) and Umeda et al. (2019). Lastly, Sects. 7 and 8 provide discussion and conclusion.

2 Explicit Learning and Underlying Linguistic Competence There has been a great deal of discussion by various scholars about implicit/explicit learning, instruction and knowledge over the years in relation to acquisition (see DeKeyser 2003; Ellis et al. 2009; Rothman 2008). However, rather than reviewing the debate over implicit/explicit learning, we focus on discussion about how pedagogical rules L2 learners are taught in a classroom through the use of negative evidence cannot restructure interlanguage grammars (Schwartz and Gubala-Ryzak 1992) by reviewing a study by Rothman (2008). Rothman’s (2008) study relates, in part, to our current study since we examine the way pedagogical rules shape adult L2 learner interlanguage grammars. Rothman (2008) examined the performance in two production tasks by tutored (classroom), and naturalistic L1 English learners of L2 Spanish acquiring the preterite and imperfective distinction. The study set out to test the Competing Systems Hypothesis, as advanced by Rothman (2008), in order to account for differences found between tutored and untutored L2 learners. The tutored group had all received at least 5 years of instruction and in some cases were teachers of Spanish, thus they had an advanced understanding of Spanish grammar. The untutored group were living in Spanish-speaking countries at the time of data collection. However, the tutored learners had been given a number of simplified rules to follow through textbooks and formal classroom instruction, and as a result, Rothman (2008) claims that this may promote ‘faulty intuitions’ about the preterite and imperfective distinction. The results clearly showed that the tutored learners differed from the untutored naturalistic learners as the latter group performed much more like the Spanish native speakers. The combined results are in line with the predictions of the Competing Systems Hypothesis. Rothman (2008) argues “that only natural positive evidence leads to grammatical competence. In other words, while explicit positive evidence (pedagogical rules), constitute an indispensable facet of the most common situation of second language learning (the classroom), these rules do not lead to underlying linguistic competence” (p. 99). In other words, negative evidence (feedback from the teacher) is very different to positive evidence and as a result, L2 learners who learn rules in a classroom cannot internalize explicit knowledge, and as Schwartz and Gubala-Ryzak (1992) argue, learners fail to restructure their interlanguage grammars. In sum, it is important to note that the studies by Rothman (2008) tested competence and performance using an off-line method. In the current study we plan to use self-paced reading to see if an on-line method can provide further insight

158

N. Snape

into instructed L2 learners’ underlying linguistic competence of articles.2 In the following section we outline articles and bare plurals and how they can be used generically.

3 Genericity and L2 Acquisition 3.1 Kind vs. General Statement A distinctive type of generic occurs in English and it is known by the term ‘kind’. ‘Kind’ is used to define the example used in the Introduction, The dodo is extinct. It is typically used to refer to ‘the kind whose members have the property of being dodos’, which means the definite singular refers to ‘the unique taxonomic entity/species of dodo (a type of bird)’ (Dayal 2004; Krifka et al. 1995). A kind predicate like be extinct, be common, be widespread, indicates that the subject noun phrase (NP), i.e., The dodo, must denote a kind: an individual or a group of individuals cannot be extinct, but a kind can be (Ionin et al. 2011).3 Examples of how a kind predicate as subject NP functions in sentences are provided in (1). (1)

a. b. c. d. e.

The dodo is extinct. Dodos are extinct. *A dodo is extinct. #The dodos are extinct. *Dodo is extinct.

With the kind predicate be extinct, only (1a, b) are possible sentences with subject NPs that can refer to a group who have the property of being dodos; (1c) is ungrammatical with an indefinite singular and (1e) is ungrammatical as it is a bare singular count noun, which English does not usually license. (1d) is pragmatically odd as definite plurals in English typically do not have a generic interpretation and do not appear with kind predicates. Only a unique or maximal denotation is possible with definite plurals (Heim 1991), for instance, The lemurs live in Madagascar. In other words, (1d) subject NP The dodos must be a general statement and contextually salient within the discourse context for it to be a felicitous sentence. The other type of generic that is more common is a statement about a referent, as in (2).

2 See

Marsden (2018) for a review of the current research into how generative SLA can be useful to inform language instructors about the development of the mental architecture in L2 acquisition. 3 Bare plurals can also denote ‘kind’ such as e.g., Dodos are extinct, but for the purpose of the current study, we will solely focus on the definite singular ‘kind’.

Post-instruction Processing of Generics in English by Japanese L2 Learners

(2)

a. b. c. d. e.

159

A lion usually hunts alone. The lion usually hunts alone. Lions usually hunt alone. #The lions usually hunt alone. *Lion usually hunts alone.

(adapted from Ionin et al. 2011, p. 246) The subject NPs in (2a-c) head generic or general statements about lions in that they usually hunt alone. (2d), like (1d) above, does not have a generic interpretation, only a unique or maximal denotation is possible that refers to a particular group of lions. (2e) is ungrammatical like (1e) as it is a bare singular count noun. This means that for most native speakers of English, the statements in (2a-c) share the same meaning as each statement is a general statement about lions. To summarize, this section has introduced the notion of genericity as marked by articles and the bare plural. However, an important distinction to note is that there is a certain type of generic, namely ‘kind’, a definite singular subject NP that appears with a kind predicate like be extinct. Conversely, the more frequent generic is a general statement that occurs without a kind predicate, and thus does not restrict the selection of subject NP; the subject NP can be a definite, indefinite or bare plural to refer to general properties of a referent.

3.2 L2 Acquisition of Generics In L2 studies of article acquisition from the eighties and nineties (Huebner 1985; Murphy 1997; Parrish 1987; Thomas 1989; Young 1996), scholars used Bickerton’s (1981) semantic framework to determine whether L2 learners were accurately supplying articles for generic statements in spoken and written production tasks. The findings largely showed that learners omitted articles, using only the bare singular count noun, i.e., no suppliance of the definite or indefinite singular. But, one drawback of using production tasks, especially free production, was that generic statements were rarely produced due to the infrequent use of generics in everyday conversation (Biber et al. 1999; Yoo 2009). Recently, there has been a renewed focus on generics in adult L2 acquisition (Ionin and Montrul 2009; Snape 2013; Snape 2018; Snape et al. 2013a, b) and bilingualism (Kupisch 2012; Serratrice et al. 2009) to see whether L2 learners and bilinguals alike can acquire generics. No longer do studies limit themselves to production tasks – increasingly forced-choice elicitation tasks, acceptability judgement tasks (AJT), picture matching tasks and truth-value judgement tasks are used to illicit judgements about generic statements. The study by Ionin et al. (2011) used the kind vs. general distinction outlined in Sect. 3.1 above to investigate whether Russian L2 learners and Korean L2 learners of English were able to make native-like judgements of kind and general statements by providing ratings for sentences using an AJT. An example from the AJT is in (3).

160

N. Snape

(3) Kind condition in an AJT Lead-in sentences: I have been studying biology today and I found out that many species are no longer alive. For example, I found out . . . . a. b. c. d. e.

the dodo is extinct. dodos are extinct. a dodo is extinct. the dodos are extinct. dodo is extinct.

1 1 1 1 1

2 2 2 2 2

3 3 3 3 3

4 4 4 4 4

Each short lead-in is followed by five sentences. Participants were instructed to judge how appropriate each sentence is in the context of the lead-in, and rate it on a scale from 1 to 4. They did not need to rank the sentences with respect to one another: they were told to give the same rating to two or more of the sentences that follow a lead-in if they felt more than one sentence was possible. If a sentence is very appropriate, rate it as a 4; if it is completely inappropriate, rate it as a 1. If its appropriateness is somewhere in-between, rate it as a 2 or 3. They were told to base their ratings on sentence appropriateness in the context of the lead-in. (3a, b) should be the two kind sentences that receive a rating of 4 since they are grammatical with definite and bare plural subject NPs and the kind predicate be extinct. (3c–e) should each receive a rating of 1 as they are either ungrammatical or do not have a kind interpretation. Ionin et al. (2011) found that all L2 learners performed poorly at rating the definite singular highly for the kind condition. The learners tended to rate the definite, indefinite and bare count singular noun the same on the AJT as they were unsure which article (or bare noun) was appropriate for sentences like those in (3). Snape (2013) tested Japanese and Spanish L2 learners of English using a similar AJT to the one implemented in Ionin et al.’s study. Snape’s findings replicated the results of Ionin et al. (2011) with the Japanese L2 group, but the Spanish L2 group performed much more like the native speaker control group in that they rated the definite singular between 3 and 4 and rated the indefinite singular and bare singular count noun between 1 and 2. Ionin et al. (2011) and Snape (2013) argue that their results are largely attributable to the L2 learners’ first language (L1); if the L1 lacks articles, there is no direct way to map features from the L1 to the L2 morphology (Lardiere 2004). Further evidence to suggest that Japanese L2 learners, even at advanced levels, continue to experience difficulty with articles for generic contexts comes from a study by Snape et al. (2013a) who compared article choices by different L1 groups. A forced-choice elicitation task was employed. The task had several test dialogues that included contexts with count and mass nouns, but each test condition was designed so that a generic interpretation would be readily available, as in (4).

Post-instruction Processing of Generics in English by Japanese L2 Learners

(4)

161

Kind condition in a forced-choice elicitation task A: This book gives interesting facts about South America. B: Like what? A: For example, ____ potato was first cultivated in South America. Ø an a the

The correct article is the since the potato refers to the ‘kind’ potatoes, not one specific type of potato. There is also the use of a kind predicate – be cultivated. The findings revealed different rates of accuracy according to what transferred from the L1 to the L2. The Japanese L2 learners were the least accurate at selecting the definite singular with a kind predicate compared with the Turkish L2 learners and Spanish L2 learners. The fact that Japanese and Turkish lacks a definite singular that can refer to kinds in part explains why even at advanced levels, Japanese and Turkish speakers of English cannot identify the appropriate article for generics like (4). Interestingly, the patterns of article choice differed between the Turkish and Japanese learners. The Turkish groups tended to select no article Ø whilst the Japanese groups preferred to select a. Snape (2018) administered a picture matching task (PMT) to Japanese L2 learners and native English speaker controls. As in previous studies by Snape (2013) and Snape et al. (2013a), the aim of the study was to examine L2 learners’ understanding of generic sentences. A PMT was implemented rather than an AJT. One advantage a PMT has compared with other types of tasks is that it simplifies things for the participant. An example is given in (5). (5) You may be surprised to learn the snow monkey is indigenous to mountainous areas of Japan.

The participant must read the sentence in (5) and then match the picture or pictures with the sentence. In (5), there are two possible options, picture b. and

162

N. Snape

c, as both one monkey and more than one monkey are acceptable choices to refer to kind. But, compare (5) with (6), where there is no kind predicate.4 (6) You can see the snow monkey is small and cute.

In (6), the only acceptable picture that matches the sentence is b. since the snow monkey is not being referred to as a group but rather one single monkey. The results showed that Japanese learners could correctly pick out both b. and c. pictures for (5) and select only picture b. for (6). Furthermore, the results from the PMT were compared with results from a forced-choice elicitation task completed by a different group of Japanese L2 learners. Participants performed significantly better with generics on the PMT compared with article choice on the forced-choice elicitation task (see (4) above). Learners selected the wrong article for the definite kind condition compared with the definite unique condition in (7). (7) ✗ ✗ 

Definite unique condition in a forced-choice elicitation task Kelly received a tulip in May. ____ tulip is beautiful. a no article the don’t know

In (7), the referent ‘tulip’ is introduced into the discourse through the use of the indefinite article a. As ‘tulip’ is referred to again in the following sentence, learners

4 An

anonymous reviewer points out that the sentence ‘The snow monkey is small and cute’ can have a general interpretation. It is possible to obtain a general interpretation, but the crucial point is that ‘The snow monkey is indigenous . . . ’ can only have a kind interpretation, whereas ‘The snow monkey is small and cute’ may be interpreted as general or unique.

Post-instruction Processing of Generics in English by Japanese L2 Learners

163

should select the not a as the correct article choice since the definite article has a unique interpretation, meaning that it refers to the same tulip. In sum, studies on testing L2 learners’ understanding of genericity and the notion of ‘kind’, have been revealing because even though learners struggle to choose or decide which article is acceptable on a forced-choice elicitation task or how to rate the definite singular on an AJT, Japanese L2 learners can match a kind statement to the correct picture(s) on a PMT. The findings from L2 studies of genericity are in line with a generative grammar proposal by Lardiere (2004, 2008) that features present in the L1 need to be remapped onto L2 morphology. Features such as definiteness, anaphoricity, specificity, genericity may be realized differently than by means of articles in languages like Russian, Korean and Japanese. These features in the L1 grammars need to be reassembled, and if successful, are mapped onto the articles in L2 English (see Cho 2017; Hawkins et al. 2006). The current study attempts to find out if instruction can lead to underlying linguistic competence of articles.

4 On-Line Processing Studies of Articles On-line methods are an accepted psycholinguistic tool to investigate implicit processing of language that is not production oriented (Mitchell 2004). Our study investigates how instructed L2 learners process different types of generics, so it is worth looking at the results of studies that have investigated the on-line processing of articles by adult second language (L2) learners (Chondrogianni et al. 2015; Kim and Lakshmanan 2008; Snape et al. 2013b; Trenkic et al. 2014). The study by Chondrogianni et al. (2015) explored on-line processing of definite articles in Turkish-speaking sequential bilingual children acquiring English and Dutch as an L2 in the UK and in the Netherlands, respectively. They focused on the semantic functions for discourse referents, anaphoric and bridging definites. Examples of anaphoric and bridging are provided in examples (9–10). (9)

Anaphoric There was a cat in the room. The cat was sitting comfortably on the cushions.

(10) Bridging Mary wanted to eat a banana. She peeled the skin and cut the fruit into slices. In (9), the italicized NP a cat is introduced into the discourse for the first time with the indefinite article and then referred to again in the proceeding sentence with definite article the cat. This is known as an anaphoric definite as the anaphor is the indefinite NP. In (10), the italicized NP a banana is introduced into the discourse but the following definite NP the skin refers to a banana. This extra-linguistic operation is called ‘bridging’ and relies on making inferences regarding the availability of shared world knowledge (Avrutin 1999). Chondrogianni et al. (2015) administered

164

N. Snape

a self-paced listening task to assess the bilingual children’s sensitivity to omission of definites in anaphoric and bridging conditions. Half the sentences in the task were grammatical (see examples 11–12) and the other half had omissions of definites. (11) Anaphoric condition The bear bought a book and a magazine. The bear/read/(the) book/in his garden/at lunchtime. (12) Bridging condition The monkey wanted to eat a banana but first he had to remove something. The monkey/removed/(the) skin/of the banana/at lunchtime. The results revealed that in both English and Dutch the bilingual children showed longer response times when there was an ungrammaticality caused by omission compared to grammatical sentences in both the anaphoric condition and the bridging condition. The authors argue that the children are sensitive to the ungrammaticality of omitted articles and conclude that despite the lack of the syntactic functional category DP in Turkish, it does not indicate an impairment in underlying linguistic competence (Hawkins and Chan 1997).

5 Previous Article Interventions Before reviewing the two papers that are part of the current study, we provide some background into articles-related interventions. Articles are an area of grammar that do not usually get a great deal of coverage in the language classroom, and textbooks devote little space to covering the range of functions articles have in English (Yoo 2009). As a result, L2 learners may have a limited understanding of how articles are used. One way to address the issue of insufficient instruction and, in some cases, lack of primary linguistic data (e.g., as in the case of English as a Foreign Language (EFL)) is to offer some kind of intervention in the form of explicit instruction (see Akakura 2012 for an overview). Snape and Yusa (2013) administered a small-scale study that included the instruction of generics to one group of Japanese L2 learners. Before instruction took place, the instruction group and a control group completed a pre-test in the form of an AJT, like the one used by Ionin et al. (2011). The instruction was delivered in English by a native English instructor over the course of 3 weeks. The results showed that between the pre- and two post-tests, no significant differences were found, meaning that there was no improvement in the participants’ ratings of the definite singular for kind sentences. Snape et al. (2016) and Umeda et al. (2019) conducted a follow-up intervention for 21 Japanese L2 learners of English using the same pre-test > post-test > delayed post-test design in Snape and Yusa (2013). The AJT consisted of 56 items in total: 16 items were kinds and 16 items were non-kinds, 16 items were related to definiteness and specificity, and 8 items were

Post-instruction Processing of Generics in English by Japanese L2 Learners

165

definite anaphoric singular and definite anaphoric plural fillers. Two versions of the task were created as we could then give participants one version for the pre-test and the other version for post-test 1. For post-test 2 and post-test 3 we employed the same pattern. That way, we avoided having the participants repeat the same version of the task for each test session. Twenty-one learners (instruction group) received instruction for 7 weeks and 16 learners (non-instruction group) took part in a pretest and four post-tests (3 weeks, 9 weeks, 21 weeks and 1 year after instruction). A definition for generics was devised by two native speakers of English, one of who is familiar with the definitions given by Krifka et al. (1995) and Dayal (2004) and the other who is an EFL instructor. The metalinguistic explanation regarding generics was divided into two types; generics expressed with the definite singular and bare plural (kind), and generics expressed with the indefinite singular and bare plural (general).5 For the kind generic, the participants were told that for kind, a singular count noun with the definite singular, the, together with a kind predicate such as be extinct can express genericity. In addition, they were told that the bare plural is also acceptable with this type of generic. The results for the instruction group are illustrated in Fig. 1. Figure 1 shows the mean ratings on the AJT for the definite singular kind, calculated over participants, with error bars showing the standard deviations. The legend provides information of each test. The instruction group made initial gains in post-tests 1–3, but this was not sustained over a longer period. The instruction group dropped down to almost the pre-test level, 1 year later, on post-test 4. Umeda et al. (2019) provide an explanation for their findings, based on factors such as a lack of ongoing instruction, frequency of definite singular kind in the input and the difficulty with mapping the feature [+kind] to the definite singular. Thus, intervention studies seem to be helpful for L2 learners in the short-term, but what do these results mean for explicit learning and underlying linguistic competence? We try to address this question in the following section by administering a self-paced reading (SPR) task to show whether the effects of instruction are reflected in instructed L2 learners’ reading times.

6 The Present Study The current small-scale study is part of the Snape et al. (2016) and Umeda et al. (2019) studies discussed in Sect. 5. Snape et al. (2016) and Umeda et al. (2019) have already reported differences between the instruction and non-instruction groups on 5 Krifka

et al. (1995) provide examples of generics that only allow an indefinite singular or bare plural as subject NP; they are characterizing sentences that are used to make a general statement about properties like, e.g., An orange is full of vitamin C/Oranges are full of vitamin C. These characterizing sentences are classed as infelicitous when a definite singular or definite plural is the subject NP as a preferred interpretation of, e.g., The orange is full of vitamin C/The oranges are full of vitamin C, is a unique or specific orange or a maximal set of oranges.

166

N. Snape

4

3.8 3.8

3.7

3.6

3.4

3.5 Mean ratings of AJT scale

3.2 2.9

3

2.8 Pre-test

2.5

2.5

Post-test 1 Post-test 2

2.1

Post-test 3

2

Post-test 4 1.5

1 definite singular kind

bare plural

Fig. 1 Instruction group results of definite singular kind and bare plural. (Adapted results from Umeda et al. 2019)

kind and general statement generics in post-test 2 from an AJT, but the results cannot inform us on whether instruction in the generic use of the leads to changes in the underlying linguistic system. The main aim of our study is to look at instruction effects after participants were given instruction to determine whether, in terms of processing, a difference could be observed between the instruction and noninstruction groups using a SPR task. The difference does not include within-groups.6 Our second aim is to address the issue of task; we explore the use of on-line techniques for testing grammatical sensitivity that are traditionally used for testing processing and parsing as self-paced reading is a good method to tap linguistic competence (VanPatten et al. 2012).7

6 Ideally,

if time had permitted, we would have liked to have administered the self-paced reading task after the pre-test so we could compare pre-instruction and post-instruction. 7 VanPatten et al. (2012) use the term ‘implicit knowledge’ rather than the term linguistic competence, but in order to refer to Universal Grammar and competence in a language, we stick to use the latter term.

Post-instruction Processing of Generics in English by Japanese L2 Learners

167

Table 1 Mean scores and range of scores for Test of English for International Communication of L2 learners (equivalent to B2 on the Common European Framework of Reference) Instruction group (n = 20) Non-instruction group (n = 14)

Mean scores 721 741

Range of scores 590–945 590–935

6.1 Participants Participants in our small-scale study were recruited at a university in Japan. In total, 34 learners agreed to participate in our SPR task, 20 participants from the instruction group and 14 participants from the non-instruction group. The participants are all female university students, and they are native speakers of Japanese majoring in English. All participants continued to attend EFL and content classes (between 6 and 8 h per week) during our study. Details about the two groups are provided in Table 1. We predict that there will be a difference between the instructed and noninstructed groups after 7 weeks of instruction. We recruited the instruction group immediately after post-test 2 and measured reading times.

6.2 Research Questions Our research questions for the current study is as follows: RQ1: Does the instruction group outperform the non-instruction group in real time on the definite singular kind? RQ2. Does the instruction group display underlying linguistic knowledge of definite singular kind after instruction? Research question 1 addresses whether instruction has an impact on the performance of a task in real time. Research question 2 sets out to discover whether slower reading times indicate additional processing costs (Juffs and Harrington 1995) in relation to the kind condition (definite singular).

6.3 SPR Stimulus Set and Experimental Design The stimulus material was based on what the participants had received instruction on. Two versions of the SPR task were created. The stimulus set with examples is provided in Table 2. There are six kind generics. Each kind generic features a different species (e.g., bear, dinosaur) accompanied with a kind predicate (e.g., widespread, extinct).

168

N. Snape

Table 2 SPR stimulus set with examples Stimulus set Kind grammatical condition (definite and bare plural) Kind ungrammatical condition (indefinite) General grammatical condition (indefinite and bare plural) General distractor condition (definite) Distractors (definite plural)

Examples from SPR task the/mosquito/is/common/in/many/places mosquitos/are/common/in/many/places a/mosquito/is/common/in/many/places

Number in each condition n=3 n=3 n=3

a/mosquito/is/annoying/for/lots/of/people

n=3

mosquitos/are/annoying/for/lots/of/people the/mosquito/is/annoying/for/lots/of/people

n=3 n=3

the/mosquitos/are/common/in/many/places n=6 the/mosquitos/are/annoying/for/lots/of/people

Table 2 shows the same grammatical and ungrammatical sentences for each condition for ease of reference. However, it is important to note here that no participant saw the same grammatical and ungrammatical kind and general condition in each version of the task.

6.4 Procedure Each participant was seated in front of a MacBook Pro computer equipped with a Cedrus response pad and SuperLab version 5.5 software. SuperLab was used for the self-paced reading task and to record reading times. For the SPR task, participants were instructed to read a short lead-in sentence presented on the screen. Once the participant had read the lead-in sentence, she was instructed to push any button on the response pad so that the lead-in sentence would disappear and the target sentence would appear on a new screen, one segment at a time, using the moving window procedure (Just et al. 1982). All participants (instructed and non-instructed groups) were told to make a judgement about the target sentences as to whether each sentence was appropriate or inappropriate with the lead-in before the experiment started. One button was marked ‘Yes’ for appropriate and another button was marked ‘No’ for inappropriate. Each ‘Yes’ and ‘No’ button was assigned a specific colour and labelled ‘Yes’ and ‘No’ to avoid any confusion. The setup of the ‘Yes, No’ buttons was switched for the two versions of the task. Reading times (measured in milliseconds) on critical segments of matched sentences serve as measures of linguistic sensitivity to grammatical violations.

Post-instruction Processing of Generics in English by Japanese L2 Learners

169

6.5 Results Firstly, we report the accuracy of responses to the definite singular kind and general conditions, then the reading times (RTs) along with the residual reading times (RRTs). Figure 2 reports the responses to the definite singular kind condition with standard error bars. The results in Fig. 2 show that the instruction and non-instruction groups performed similarly on the definite singular kind condition. The legend indicates the correct response (definite) and the incorrect response (indefinite). A chi-square test reveals that there is no significant difference between the instruction and noninstruction groups for the definite singular kind condition (Pearson χ2 = 2, p > .05). Figure 3 describes the responses for the general condition. The legend in Fig. 3 indicates both possible responses for the general condition with standard error bars. A chi-square test reveals that there is no significant difference between the instruction and non-instruction groups for the general condition (Pearson χ2 = .02, p > .05). The statistical software package SPSS version 17 was used to perform all the statistical tests. All outliers were removed before the calculation of raw reading times. The raw reading times with means and standard deviations (SDs) are reported in Table 3 for regions 4–6, the key regions. Table 3 shows that for the critical region (R4), there are no big differences in reading times for the instructed and non-instructed groups for the definite singular kind condition. For the general condition, the instruction group produced slower

1.00 0.90

Means per condition

0.80

0.70 0.60 definite

0.50

indefinite

0.40 0.30 0.20 0.10 0.00 Instruction

Non-instruction

Fig. 2 Means per condition over participant responses for the definite singular kind condition

170

N. Snape

1.00 0.90

Means per condition

0.80 0.70 0.60 definite

0.50

indefinite

0.40 0.30 0.20 0.10 0.00 Instruction

Non-instruction

Fig. 3 Means per condition over participant responses for the general condition Table 3 Mean raw RTs and SD (in brackets) per condition over participant for critical, postcritical/spillover and wrap-up regions Participants Instructed

Non-instructed

Condition 1 = def sing kind 2 = *indef sing kind 3 = def general 4 = indef general 1 = def sing kind 2 = *indef sing kind 3 = def general 4 = indef general

Raw reading times per region Region 4 (R4) Region 5 (R5) 680 (394) 552 (319) 614 (444) 584 (354) 587 (390) 567 (340) 651 (422) 553 (310) 683 (458) 529 (274) 681 (365) 531 (183) 544 (372) 532 (297) 544 (297) 489 (194)

Region 6 (R6) 580 (374) 527 (365) 551 (390) 716 (407) 566 (389) 643 (389) 692 (468) 819 (531)

An anonymous reviewer suggested that the control conditions (bare plurals and definite plurals) which served as distractors in our experiment be included in the analysis. But, to due to word limitations, we elected to only present details of the experimental conditions

reading times for indefinite general; this difference is not observed for the noninstruction group. The reading times for the post-critical/spillover region (R5) show that the instruction group is slower at reading the indefinite singular kind (the ungrammatical sentence) compared with definite singular kind (the grammatical sentence); this difference is not observed in the non-instruction group. There are differences for both groups for the general condition as the indefinite general shows

Mean RRTs

Post-instruction Processing of Generics in English by Japanese L2 Learners

350 300 250 200 150 100 50 0 -50 -100 -150 -200 -250 -300 -350

171

Instruction

Non-Instruction

R1 / R2 The (*A ) / koala

Regions / R3 / R4 / R5 / R6 / is / rare / in / Japan

Fig. 4 RRTs on definite minus indefinite mismatch conditions for the kind condition

longer reading times than the definite general. It may be that part of the reason for this is that the definite general is the preferred reading. The wrap-up region (the final region R6) comes before the participants have to decide if each sentence is appropriate or inappropriate with the lead-in sentence. From the responses shown in Figs. 2 and 3, it seems both groups of participants are unsure of the appropriateness for the definite singular kind and indefinite singular kind, and this is mirrored in the reading times in Table 3. The instruction group has slower reading times for the definite singular kind compared with the indefinite singular kind, but the difference is not significantly greater. Figures 4 and 5 report the RRTs for both groups for all regions with standard error bars. The legend refers to the instruction group as ‘Instruction’ and to the non-instruction group as ‘Non-instruction’. To control for the difference in length between the critical segments in the grammatical (definite singular kind, general) and ungrammatical (indefinite singular kind) conditions, raw RTs were transformed into residual RTs by normalizing RTs of each participant by her mean reading time per letter in each word for each condition.8 Outliers were trimmed at ±2.5 standard deviations (SDs) or more from the mean subject

8 To factor out effects of word length we used linear regression to calculate expected RTs per subject

on the basis of all their reading times on all segments (incl. fillers, but not practice). We removed any excessively high or low RTs. We then produced a RT data excel file with the length of each segment. Next, in SPSS, we computed a linear regression per subject to get the expected speed per subject depending on how long the words/length of segment are, for that particular person. To do this:

Mean RRTs

172

N. Snape

350 300 250 200 150 100 50 0 -50 -100 -150 -200 -250 -300 -350

Instruction Non-Instruction

Regions R1 / R2 / R3 / R4 / R5 / R6 (A) The / koala / is / cute / and / cuddly Fig. 5 RRTs on definite minus indefinite mismatch conditions for the general condition

RTs before the RRT analysis was conducted. The trimming of extreme data points resulted in the loss of 6% or fewer trials. The mean residual RTs in Fig. 4 show that for the instruction group, they are faster overall compared with the non-instruction group, at reading the kind condition. In comparison, there is no difference between the two groups for the general condition, except for the final wrap-up region (R6). The wrap-up region may have been read faster by the instruction group participants in anticipation of making a judgement about the appropriateness of the sentence. To examine whether the instruction group was sensitive to ungrammaticality of the kind condition, repeated-measures ANOVAs were performed. Condition (kind, general) and Grammaticality (definite, indefinite) were the within participants (F1) factors and Group (instruction, non-instruction) were the between participants (F2) factors for each segment (critical, post-critical/spillover and wrap-up regions) 1. We put all the RT and length of segment data in SPSS. 2. We split the file and organised into groups (to get an output per subject). 3. We ran a linear regression to obtain plot coefficients. This produced expected overall RTs per person and the coefficient. We put expected RTs and coefficients per subject from the linear regression data in a syntax file and ran the syntax file on our experimental RT file in SPSS to compute every RT per subject, the expected RT per item/segment, per condition. Finally, we subtracted the expected RTs from the raw RTs to give us residual reading times.

Post-instruction Processing of Generics in English by Japanese L2 Learners

173

separately. For the critical region (R4) there was no significant differences within Condition (F(1, 73) = 1.512, p > .05) and Grammaticality (F(1, 73) = 0.029, p > .05). A significant difference was found between the instruction group and the non-instruction group (F(1, 73) = 39.45, p < .001, Cohen’s d = −0.7). For the post-critical/spillover region (R5), there was no significant differences within Condition (F(1, 87) = 0.659, p > .05) and Grammaticality (F(1, 87) = 0.845, p > .05). For the wrap-up region (R6), there was no significant differences within Condition (F(1, 47) = 1.713, p > .05) and Grammaticality (F(1, 47) = 0.258, p > .05). Between group comparisons revealed significant differences in R5 (F(1, 87) = 33.22, p < .001, Cohen’s d = 0.6 and R6 F(1, 47) = 18.05, p < .001, Cohen’s d = 0.6). Pairwise comparisons for the instruction group participants revealed no significant differences in the critical region (R4) for RTs in the definite vs. the indefinite condition for both the kind condition (t(42) = .676, p > .05) and the general condition (t(50) = .333, p > .05). For the non-instruction group participants, no significant differences in R4 were found in each condition, kind (t(34) = .098, p > .05) and general (t(36) = −.181, p > .05). No significant differences were found in the post-critical /spillover region (R5) for the instruction group in the definite vs. the indefinite kind condition (t(51) = 1.114, p > .05) and in the general condition, there were no significant differences between definite and indefinite (t(54) = .0557, p > .05) and no differences for the non-instruction group (kind = t(39) = .0224, p > .05; (general = t(41) = −.866, p > .05). For the final wrap-up region (R6), no significant difference was revealed for both kind and general conditions for the instruction group (t(47) = 0.975, p > .05; t(30) = −.1676, p > .05). For the non-instruction group, a significant interaction was found between definite singular kind vs. indefinite singular kind (ungrammatical) (t(31) = 2.222, p < .04), but no interaction was found for the general (definite vs. indefinite) condition (t(21) = 0.265, p > .05).

7 Discussion The results of the current study show that the instruction group performed differently from the non-instruction group on the self-paced reading task for the kind condition (see Fig. 4). We revisit the research questions below: RQ1: Does the instruction group outperform the non-instruction group in real time on the definite singular kind? The answer to RQ1 is ‘Yes’ since it is clear that there are differences between the instruction and non-instruction group in performance on the definite singular kind. The instruction group were faster in their reading times compared with the noninstruction group. Most notably, one observation is that in the post-critical/spillover region for the kind condition, the instruction group slowed down when there was a subject NP-kind predicate mismatch (see Table 3 above), but this difference was not significant. No difference was observed in the general condition between the

174

N. Snape

two groups (see Fig. 5). Despite faster reading times for the kind condition by the instruction group, their accuracy responses for kind and general conditions were no better than the non-instruction groups’ responses (see Figs. 3 and 4). Perhaps due to the short amount of time between the instruction participants received and the time of testing means that the only observable differences to be found are that the instruction group are quicker to process meaning once they read the critical region (R4), and the processing cost is evident in the post-critical/spillover region (R5) in Fig. 4. These results cannot be directly compared with the findings from previous studies since our study involved instruction whereas Trenkic et al. (2014) and others tested an L2 group that received no specific instruction on articles. It is likely that if the instruction group in our study had not received an intervention, there would be no differences between the two groups in our study. RQ2. Does the instruction group display underlying linguistic knowledge of definite singular kind after instruction? The answer to RQ2 is unclear as to whether underlying linguistic competence is more native-like as a result of instruction. Studies reported in Larsen-Freeman and Long (1991) all point to, at best, what VanPatten (2017) believes as ‘trivial’ effects (if any) of instruction. “For example, instruction can affect how learners perform on certain kinds of form-focused tests and it can affect acquisition in the short term but disappears in the longer term. Fundamental changes in learners’ linguistic systems due to instruction were non-evident” (p. 49). Our results certainly suggest that there are short-term effects as they echo the findings of the off-line results from post-test 2 reported in Snape et al. (2016) and Umeda et al. (2019). Nonetheless, we cannot address the issue of long-lasting effects of instruction as one of the limitations of this study is that we were unable to include follow-up testing after post-test 4 (1 year after instruction). Perhaps the best way the results can be interpreted is that instructed learners simply get faster with explicit knowledge (VanPatten 2016), but over time any explicit knowledge gained through instruction gradually fades unless there is ongoing, systematic instruction offered. The implications for L2 pedagogy are what make our small-scale study of interest to instructors. For L2 instruction to be effective and beneficial, each area of grammar taught in the classroom needs to be revisited over and over again, otherwise there is a chance that any explicit knowledge gained will retreat and slowly wane. This is important in relation to morphology, as Slabakova (2016) refers to functional morphology as the bottleneck of L2 acquisition. Instruction may be even more important in the case of articles, because articles in English have a number of functions, generics being just one of many. In comparison, third person singular – s is far less complex to learn as a rule, e.g., add 3rd pers sing –s to a main verb in present simple when the subject is third person singular subject, he, she or it. It is not enough to provide instruction on articles a few times, at early stages of L2 acquisition – it must be sustained and ongoing so L2 learners can restructure their interlanguage grammars. The combined results from Snape et al. (2016), Umeda et al. (2019) and the current study are suggestive but seem to be consistent with the Competing Systems

Post-instruction Processing of Generics in English by Japanese L2 Learners

175

Hypothesis. In order to be able to provide a direct comparison with Rothman’s (2008) findings, naturalistic L2 learners would need to be tested, using the same conditions, kind and general generics.

8 Conclusion To conclude, the current study found that the instructed L2 learners improved in article choice between the pre- and post-tests 1 and 2. We believe it is useful to include an on-line task like a self-paced reading or listening task9 to find out whether learners’ underlying linguistic competence is native-like or non-native-like. Our findings suggest that even though instruction can lead to gains in faster reading times, it is unclear whether explicit knowledge is retained over the long term. In fact, results from Umeda et al. (2019) show that the instruction group failed to retain explicit knowledge after one-year post intervention in an off-line AJT (post-test 4); The self-paced reading task was not administered. VanPatten (2017) questions what it means to be non-native exactly. For instance, in performance, suppliance of definites could be high, which suggests L2 learners acquire the functional structure Determiner Phrase, but learners may continue to demonstrate non-native distribution of definite and indefinite articles as a result of discourse-related issues. In the case of generics, due to the relative infrequency of positive evidence coupled with the lack of negative evidence in L2 classrooms and textbooks, it may be more of a challenge for L2 learners to acquire subtle differences between kind and general generics. Acknowledgments I wish to thank Mari Umeda for her help with planning the current study. I am grateful for the editors’ and anonymous reviewers’ comments and suggestions. Any remaining errors are solely my own.

References Akakura, M. (2012). Evaluating the effectiveness of explicit instruction on explicit and implicit L2 knowledge. Language Teaching Research, 16(1), 9–37. Avrutin, S. (1999). Development of the syntax-discourse interface. Dordrecht/Boston: Kluwer Academic Publishers. Biber, D., Johansson, S., Leech, G., Conrad, S., & Finegan, E. (1999). Longman grammar of spoken and written English. London: Longman. Bickerton, D. (1981). Roots of language. Ann Arbor: Karoma Press.

9 We

chose to implement a self-paced reading task rather than a self-paced listening task (Chondrogianni et al. 2015) because there are possible issues with the perception of articles for L2 learners from article-less L1s in a listening task (see Pierce and Ionin 2011).

176

N. Snape

Chamorro, G., Sorace, A., & Sturt, P. (2016). What is the source of L1 attrition? The effect of recent L1 re-exposure on Spanish speakers under L1 attrition. Bilingualism: Language and Cognition, 19(3), 520–532. Cho, J. (2017). The acquisition of different types of definite noun phrases in L2-English. International Journal of Bilingualism, 21(3), 367–382. Chondrogianni, V., Vasi´c, N., Marinis, T., & Blom, E. (2015). Production and on-line comprehension of definiteness in English and Dutch by monolingual and sequential bilingual children. Second Language Research, 31(3), 309–341. Dayal, V. (2004). Number marking and (in)definiteness in kind terms. Linguistics and Philosophy, 27(4), 393–450. DeKeyser, R. (Ed.). (2003). Implicit and explicit learning. In C. Doughty & M. H. Long (eds.), The handbook of second language acquisition (pp. 313–348). Oxford: Blackwell. Ellis, R., Loewen, S., Elder, C., Erlam, R., Philp, J., & Reinders, H. (Eds.). (2009). Implicit and explicit knowledge in second language learning, testing and teaching. Bristol: Multilingual Matters. Hawkins, R., & Chan, Y. (1997). The partial availability of UG in SLA: The failed functional features hypothesis. Second Language Research, 3(13), 187–226. Hawkins, R., Al-Eid, S., Almahboob, I., Athanasopoulos, P., Chaengchenkit, R., Hu, J., Rezai, M., Jaensch, C., Jeon, Y., Jiang, A., Leung, I., Matsunaga, K., Ortega, M., Sarko, G., Snape, N., & Velasco-Zarate, K. (2006). Accounting for English article interpretation by L2 speakers. In S. H. Foster-Cohen, M. M. Kranjnovic, & J. M. Djgunovic (Eds.), EUROSLA yearbook (Vol. 6, pp. 7–25). Amsterdam: John Benjamins. Heim, I. (1991). Articles and definiteness. In A. Stechow & D. Wunderlich (Eds.), Handbook of semantics. An international handbook of contemporary research (pp. 487–535). Berlin: W de Gruyter. Huebner, T. (1985). System and variability in interlanguage syntax. Language Learning, 35(2), 141–163. Ionin, T., & Montrul, S. (2009). Article use and generic reference: Parallels between L1- and L2-acquisition. In M. D. P. G. Mayo & R. Hawkins (Eds.), Second language acquisition of articles: Empirical findings and theoretical implications (Vol. 49, pp. 147–174). Amsterdam: John Benjamins. Ionin, T., Montrul, S., Kim, J.-H., & Philippov, V. (2011). Genericity distinctions and the interpretation of determiners in second language acquisition. Language Acquisition, 18, 242– 280. Juffs, A., & Harrington, M. (1995). Parsing effects in second language sentence processing: Subject and object asymmetries in wh-extraction. Studies in Second Language Acquisition, 17(4), 483– 516. Just, M. A., Carpenter, P. A., & Woolley, J. D. (1982). Paradigms and processes in reading comprehension. Journal of Experimental Psychology: General, 111, 228–238. Kim, K., & Lakshmanan, U. (2008). L2 article semantics and second language processing. In R. Slabakova, J. Rothman, P. Kempchinsky, & E. Gavruseva (Eds.), Proceedings of the 9th generative approaches to second language acquisition conference (GASLA 2007) (pp. 103– 117). Somerville: Cascadilla Proceedings Project. Krifka, M., Carlson, G. N., & Pelletier, F. J. (Eds.). (1995). The generic book. Chicago: The University of Chicago Press. Kupisch, T. (2012). Specific and generic subjects in the Italian of German–Italian simultaneous bilinguals and L2 learners. Bilingualism: Language and Cognition, 15(4), 736–756. Lardiere, D. (2004). Knowledge of definiteness despite variable article omission in second language acquisition. In A. Brugos, L. Micciulla, & C. E. Smith (Eds.), The 28th Boston University conference on language development (pp. 328–339). Somerville: Cascadilla Proceedings Project. Lardiere, D. (2008). Feature-assembly in second language acquisition. In J. M. Liceras, H. Zobl, & H. Goodluck (Eds.), The role of formal features in second language acquisition (Second

Post-instruction Processing of Generics in English by Japanese L2 Learners

177

language research acquisition series. Theoretical and methodological issues) (pp. 106–141). New York: Lawrence Erlbaum Associates. Larsen-Freeman, D., & Long, M. H. (1991). An introduction to second language acquisition research. London: Longman. Lopez, E., & Sabir, M. (2019). Article pedagogy: Encouraging links between linguistic theory and teaching practice. RELC Journal, 50(1), 188–201. Marsden, H. (2018). Generative second language acquisition and language teaching. In J. Cho, M. Iverson, T. Judy, T. Leal, & E. Shimanskaya (Eds.), Meaning and structure in second language acquisition: In honor of Roumyana Slabakova (pp. 283–308). Amsterdam: John Benjamins. Mitchell, D. (2004). On-line methods in language processing: Introduction and historical overview. In M. Carreiras & C. J. Clifton (Eds.), The on-line study of sentence comprehension (pp. 15– 32). New York: Psychology Press. Montrul, S., & Ionin, T. (2012). Dominant language transfer in Spanish heritage speakers and second language learners in the interpretation of definite articles. The Modern Language Journal, 96(1), 70–94. Murphy, S. (1997). Knowledge and production of English articles by advanced second language learners. (PhD thesis), University of Texas, Austin. Parrish, B. (1987). A new look at methodologies in the study of article acquisition for learners of ESL. Language Learning, 37, 361–383. Pierce, L., & Ionin, T. (2011). Perception of articles in L2 English. In L. Plonsky & M. Schierloh (Eds.), Selected proceedings of the 2009 second language research forum (pp. 121–128). Somerville: Cascadilla Proceedings Project. Rothman, J. (2008). Aspect selection in adult L2 Spanish and the competing systems hypothesis: When pedagogical and linguistic rules conflict. Languages in Contrast, 8(1), 74–106. Schwartz, B., & Gubala-Ryzak, M. (1992). Learnablity and grammar reorganization in L2 acquisition: Against negative evidence causing the unlearning of verb movement. Second Language Research, 8(1), 1–38. Serratrice, L., Sorace, A., Filiaci, F., & Baldo, M. (2009). Bilingual children’s sensitivity to specificity and genericity: Evidence from metalinguistic awareness. Bilingualism: Language and Cognition, 12(2), 239–257. Slabakova, R. (2016). Second language acquisition. Oxford: Oxford University Press. Snape, N. (2013). Japanese and Spanish adult learners of English: L2 acquisition of generic reference. Studies in Language Sciences: Journal of the Japanese Society for Language Sciences, 12, 70–94. Snape, N. (2018). Definite generic vs. definite unique in L2 acquisition. Journal of the European Second Language Association, 2(1), 83–95. Snape, N., & Yusa, N. (2013). Explicit article instruction in definiteness, specificity, genericity and perception. In M. Whong, K.-H. Gil, & H. Marsden (Eds.), Universal grammar and the second language classroom (pp. 161–183). Dordrecht: Springer. Snape, N., García Mayo, M. P., & Gürel, A. (2013a). L1 transfer in article selection for generic reference by Spanish, Turkish and Japanese L2 learners. International Journal of English Studies, 13(1), 1–28. Snape, N., Hirakawa, M., Hirakawa, Y., Hosoi, H., & Matthews, J. (2013b). The role of genericity in online grammar processing by Japanese adult L2 learners and Japanese child L2 learners of English. In J. C. Amaro, T. Judy, & D. P. Cabo (Eds.), Proceedings of the 12th generative approaches to second language acquisition conference (GASLA 2013) (pp. 193– 202). Somerville: Cascadilla Proceedings Project. Snape, N., Umeda, M., Wiltshier, J., & Yusa, N. (2016). Teaching the complexities of English article use and choice for generics to L2 learners. In D. Stringer, J. Garrett, B. Halloran, & S. Mossman (Eds.), Proceedings of the 13th generative approaches to second language acquisition conference (GASLA 2015) (pp. 208–222). Somerville: Cascadilla Proceedings Project. Thomas, M. (1989). The acquisition of English articles by first- and second-language learners. Applied PsychoLinguistics, 10, 335–355.

178

N. Snape

Trahey, M., & White, L. (1993). Positive evidence and preemption in the second language classroom. Studies in Second Language Acquisition, 15(2), 181–204. Trenkic, D., Mirkovic, J., & Altmann, G. T. M. (2014). Real-time grammar processing by native and non-native speakers: Constructions unique to the second language. Bilingualism: Language and Cognition, 17(2), 237–257. Umeda, M., Snape, N., Yusa, N., & Wiltshier, J. (2019). The long-term effect of explicit instruction on learners’ knowledge on English articles. Language Teaching Research, Special Issue: Grammatical Meaning and the Second Language Classroom, 23(2), 179–199. VanPatten, B. (2016). Why explicit knowledge cannot become implicit knowledge. Foreign Language Annals, 49(4), 650–657. VanPatten, B. (2017). Situating instructed language acquisition: Facts about second language acquisition. Instructed Second Language Acquisition, 1(1), 45–60. VanPatten, B., & Rothman, J. (2015). What does current generative theory suggest about the explicit–implicit debate? In P. Rebuschat (Ed.), Explicit and implicit learning of languages (pp. 91–116). Amsterdam: John Benjamins. VanPatten, B., Keating, G. D., & Leeser, M. J. (2012). Missing verbal inflections as a representational problem: Evidence from self-paced reading. Linguistic Approaches to Bilingualism, 2(2), 109–140. Whong, M., Gil, K.-H., & Marsden, H. (Eds.). (2013). Universal grammar and the second language classroom. Dordrecht: Springer. Yoo, I. W. (2009). The English definite article: What ESL/EFL grammars say and what corpus findings show. Journal of English for Academic Purposes, 8, 267–278. Young, R. (1996). Form-function relations in articles in English interlanguage. In R. Bayley & D. R. Preston (Eds.), Second language acquisition and linguistic variation (pp. 135–175). Amsterdam: John Benjamins.

Explicit and Implicit Knowledge of Article Semantics in Belarusian Learners of English: Implications for Teaching Anders Agebjörn

Abstract The study investigates explicit and implicit knowledge of article semantics in 26 second language learners of English. Their first languages, Russian and Belarusian, lack articles. The research question is whether explicit knowledge, operationalised as an ability to explain what principle governs the choice between the definite and the indefinite article in a multiple-choice test, is associated with target-like use of articles in an oral, communicative task, assumed to tap mainly into implicit knowledge. No correlation between explicit and implicit knowledge, as measured with the two tasks, was found, suggesting that explicit knowledge is not necessary to acquire article semantics in this group of learners. The result is discussed within a modular, cognitive framework, and implications for teaching are considered. Keywords Explicit knowledge · Implicit knowledge · Articles · Definiteness · Second language acquisition · Russian · English · Language teaching · Belarusian

1 Introduction The relation between explicit and implicit knowledge, that is, knowledge available respectively unavailable to consciousness (Truscott 2015), is extensively debated within the fields of Second Language Acquisition (SLA) and Applied Linguistics (e.g. Hulstijn 2005; Rebuschat 2015; Sanz and Leow 2011; Spada 2015). This is not surprising given that the way we understand the distinction has direct bearing on both the practical question How should a language be taught? and the theoretically fundamental question What is language?

A. Agebjörn () Department for Swedish, University of Gothenburg, Gothenburg, Sweden e-mail: [email protected] © Springer Nature Switzerland AG 2020 A. Trotzke, T. Kupisch (eds.), Formal Linguistics and Language Education, Educational Linguistics 43, https://doi.org/10.1007/978-3-030-39257-4_10

179

180

A. Agebjörn

The present study addresses both questions by investigating explicit and implicit knowledge of article semantics in second language (L2) learners of English, lacking articles in their first languages (L1s), Russian and Belarusian. Article semantics is known to pose persistent difficulties to this group of learners (e.g. Ionin et al. 2004), and the research question is whether explicit knowledge, operationalised as an ability to explain what principle governs the choice between the definite and the indefinite article, is associated with a more target-like use of articles in an oral, communicative task, assumed to tap mainly into implicit knowledge. Experimental studies have shown both that explicit instruction on article semantics can affect the performance on tasks assumed to tap into implicit knowledge (Akakura 2012) and that article semantics can be learned without awareness of what was being learned (Williams 2005). Thus, with regard to L2 acquisition of article semantics, it is unclear what the relation between explicit and implicit knowledge is. The paper seeks to further explore the question and discuss its implications for teaching. Section 2 presents a background to the explicit–implicit discussion, introduces a modular, cognitive framework, and defines, within this framework, article semantics and the learning task for the participants in the study. Previous research is also reviewed here. Section 3 presents the method; Sect. 4 the results. Section 5 discusses the result and its implications for L2 teaching.

2 Theoretical Background In a theoretically neutral definition, explicit respectively implicit knowledge refer to stored information that is available respectively unavailable to consciousness. The notions consciousness and awareness are often used synonymously and refer to the subjective experience we have when we are awake (Truscott 2015). As pointed out by Andringa and Rebuschat (2015, p. 187), the way we understand the explicit–implicit distinction “is largely determined by one’s views of how language is represented in our minds.” The literature often refers to an interface and a noninterface position. According to the interface position, typically represented by Schmidt (1990, 2012) and DeKeyser (2003), the explicit–implicit distinction is a matter of degree: explicit knowledge, when used, slowly becomes proceduralised, that is, implicit. In Truscott’s (2015) wording, the interface position is based on “the nothing-special assumption” according to which knowing a language is like knowing anything else. The non-interface position, typically associated with Krashen (e.g. 1981) and Paradis (2004, 2009), holds that explicit and implicit knowledge are fundamentally different mental entities: explicit knowledge resides in declarative memory and is available to consciousness while implicit knowledge resides in procedural memory and is unavailable to consciousness; explicit processes are, even when speeded-up, controlled, while implicit processes are automatic, that is, uncontrolled; explicit knowledge is the result of learning while implicit knowledge is the result of acquisition. With training, implicit knowledge might substitute explicit knowledge, but explicit knowledge can never become implicit (Paradis 2004, 2009). Both positions make claims about language teaching.

Explicit and Implicit Knowledge of Article Semantics

181

The opposition between the two positions, which has its roots in the different views represented by generative and usage-based approaches to language, might have been exaggerated. In fact, while Schmidt’s (1990, 2012) well-known Noticing Hypothesis takes consciousness of surface forms to be necessary for L2 acquisition, it is less clear with regard to the acquisition of grammatical meaning; and while Krashen (1981) and Paradis (2004, 2009) take linguistic knowledge to be primarily implicit, they also acknowledge a potential role of explicit knowledge in L2 learning and teaching. The Modular On-line Use and Growth of Language framework (MOGUL) (Sharwood Smith 2017; Sharwood Smith and Truscott 2014), used in the present paper, seeks to account for insights from both generative and usagebased approaches to language learning and use in multilingual minds. While it is a modular model, taking language to depend on several autonomous but interconnected modules (e.g. Fodor 1983; Jackendoff 2002; Paradis 2004), it also considers the role of consciousness.

2.1 The MOGUL Framework The Modular On-line Use and Growth of Language framework (MOGUL) is based on Jackendoff’s (2002) Parallel Architecture with its three generative modules: Syntactic Structures (SS) and Phonological Structures (PS) that constitute the purely linguistic system, and Conceptual Structures (CS) that represents our thoughts. The goal of a module is to activate or manipulate structures to match whatever input it receives, via interfaces, from other modules. Each time a structure is activated, its resting level increases; structures activated simultaneously are assigned the same index. In competition with other structures, a high resting level and a high degree of co-indexation increase the chance to be selected in future processing. Creation of new structures, increased resting levels and extended co-indexation constitute what we call learning, or acquisition, or growth. Growth is “the lingering effect of processing” (Sharwood Smith and Truscott 2014, p. 93). In MOGUL, consciousness is the result of extensive activation. In practice, only Perceptual Out-put Structures (POpS) – structures that represent the output from for instance the auditory and the visual systems – ever receive the amount of activation needed to become conscious.1 This is in line with Paradis’ (2009, p. 37) observation that we can only be aware of “perceivable input.” It might sound counterintuitive that CS representations – our thoughts – are unavailable to consciousness. Thinking of a concept like WALK, though, you can be aware of the visual or auditory representation of the word walk, of the sound of shoes touching the path, of an image of someone walking, or of the feeling of muscles working in your legs, but you cannot be aware of the concept’s argument structure or of its event kind. However, thanks to the CS–POpS interface, CS representations might

1 Affectional

Structures also reach consciousness.

182

A. Agebjörn

CS

SS

PS

POpS Fig. 1 The MOGUL framework. Phonological Structures (PS) and Syntactic Structures (SS) constitute the purely linguistic system. Conceptual Structures (CS) represents our thoughts. Only representations in Perceptual Out-put Structures (POpS) ever become conscious. Arrows represent interfaces between modules

be “projected into conscious awareness” (Sharwood Smith 2017, p. 125), which yields “an experience of fringe consciousness” (Sharwood Smith and Truscott 2014, p. 292) of the conceptual structure itself. A model of the framework is seen in Fig. 1.

2.2 Definiteness In line with the modular model presented above, there is a clear distinction between definiteness as, on the one hand, a semantic concept and, one the other hand, a syntactic category. As for semantics, the referent of a definite noun phrase (NP) is, from the listener’s point of view, in some sense identifiable, for instance by virtue of being contextually present (deictic reference) or previously mentioned in discourse (anaphoric reference). The identifiability can also stem from encyclopaedic knowledge – we know that there is only one plausible referent for the NP the sun – or a combination of deictic or anaphoric reference and encyclopaedic knowledge: in the presence of a bus, the speaker can say the driver on the assumption that the listener knows that a bus normally has a driver (e.g. Hawkins 1991; Lyons 1999). MOGUL does not specify what theories best describe the internal structures and processes of individual modules. As for CS, Jackendoff (2002) proposes several subcomponents, which he calls tiers. What we normally call concepts belong to the propositional tier whereas definiteness is represented in the referential tier. Because there is no POpS counterpart to representations found in the referential tier, they cannot be projected into consciousness. This is in line with Lucas’ (2011) claim that articles has no conceptual content but encodes procedures, unavailable to consciousness.2 For the sake of simplicity, the present paper will speak of semantic definiteness as a concept: IDENTIFIABILITY (cf. Lyons 1999).

2 The

fact that semanticists, after more than a century of research, can still not agree on what articles mean illustrates that this meaning is unavailable to consciousness. If introspection into CS was possible, there would be no discussion.

Explicit and Implicit Knowledge of Article Semantics

183

Syntactic definiteness is taken to be a syntactic position, D, designated for determiners (Lyons 1999). In article languages like English, an article is an obligatory default determiner. Historically, this syntactic category is the grammaticalisation of the semantic concept IDENTIFIABILITY (Lyons 1999, p. 278), but in several languages, including interlanguages, D can also become associated with the concept SPECIFICITY, specifying whether the NP refers or not. Several studies have suggested that L2 learners sometimes associate articles with SPECIFICITY rather than with IDENTIFIABILITY (e.g. Ionin et al. 2004, but see also Trenkic 2008). In an L1 speaker of English hearing the, POpS builds a phonetic representation, [ð@], which is available to consciousness. As soon as PS receives this structure via the POpS–PS interface, a matching phonological structure, /ð@/, is built, which SS matches by activating the syntactic category D, which CS matches by activating the IDENTIFIABILITY concept. These are abstract representations of which we cannot be aware. The process is automatic, that is, uncontrolled: when hearing a language you know, you cannot prevent PS from activating phonological structures, SS from activating syntactic structures, or CS from activating conceptual structures. In speaking, the order is the opposite but the processes are equally uncontrolled and unavailable to consciousness.

2.3 The Learning Task There is some consensus that, as suggested by Trenkic (2004, p. 1402), “definiteness as a meaning category is an element of interpretation in all languages” (cf. Lyons 1999). For article-lacking Slavic languages like Russian and Belarusian (Mayo 2003; Sussex and Cubberly 2006), ample evidence suggests that speakers systematically represent NPs as definite or indefinite despite there being no obligatory morphosyntactic encoding of the concept (e.g. Brun 2001; Cho and Slabakova 2014). For instance, NPs with definite reference are ungrammatical in the existential construction in Russian (1a) just as in English (1b), indicating that the semantic concept IDENTIFIABILITY is associated also with Russian NPs. (The examples are from White et al. 2011.) (1)

koshelek a. * V ofise est’ tvoj in office exist your-NOM purse-NOM ‘There is your purse in the office’ b. * There seems to be the fly in my soup.

Opinions diverge on whether such facts should be taken as evidence that the syntactic category D is instantiated in article-lacking languages like Russian and Belarusian. Trenkic (2004) takes the adjectival nature of Slavic determiners as

184

A. Agebjörn

evidence for this not being the case.3 Others argue that D is indeed instantiated also in these languages, despite their lack of articles (e.g. Pereltsvaig 2007; Progovac 1998; Rutkowski 2002). Suffice it to say that Russian and Belarusian speaking learners of English articles must do two things. First, they need to adjust their NP structure: either D must be instantiated or, if D is already there, its value must change so that it is obligatorily filled with overt material. Second, they need to co-index the PS representation /ð@/ with the SS representation D and the CS representation IDENTIFIABILITY . This semantic concept itself, however, needs not to be learned. This natural growth of language takes time. Meanwhile, an L2 learner likely develops explicit knowledge about the language. In fact, since CS tries to make sense of the world as it is represented in POpS, and since the world includes language, all language users develop metalinguistic systems in CS. For instance, small children have concepts like WORD while more sophisticated metalinguistic systems might include schematic visualisations of possible NP structures. Such CS representations can be projected into consciousness via POpS, as described above. Typically, L2 learners explicitly know that articles are used obligatorily, that they are placed before the noun, that the definite article is used to “refer back” etcetera. Obviously, such CS representations are not the same as the abstract structures found in SS or the abstract IDENTIFIABILITY concept. However, if a metalinguistic systems in CS is speeded up, it can possibly mimic the automatic CS–SS–PS chain. Moreover, if the metalinguistic system works well, it can indirectly facilitate the natural growth of the linguistic system by feeding it with more consistent input (Paradis 2009; Sharwood Smith and Truscott 2014; Truscott 2015). Thus, MOGUL acknowledges a possible role of consciousness in L2 acquisition. Regarding article semantics, it is an empirical question what this role is.

2.4 Previous Research Butler (2002) investigated what explicit knowledge of article semantics made Japanese learners of English select incorrect articles in a multiple-choice test. The participants (n = 80) first completed the test and were then interviewed about their choices of articles. A qualitative analysis showed that the specificity–definiteness confusion was the most frequent reason for choosing an incorrect article. Similar results were reported in Yang and Ionin (2009), where the learners, instead of being interviewed, wrote explanations for their choices of articles during the completion of the multiple-choice task. This misconception among L2 learners might be explained by the fact that textbooks of English as a foreign language, as shown by Ionin 3 Lyons

(1999) and Trenkic (2004) take D to be instantiated in the pronominal system in articlelacking Slavic languages, but not in lexical NPs.

Explicit and Implicit Knowledge of Article Semantics

185

(2003, pp. 239–240), fail to provide their readers with comprehensive and accurate explanations of article semantics. Would better explicit instruction improve L2 teaching? Proponents of an explicit–implicit interface often refer to Norris and Ortega’s (2000) meta-analysis, showing positives effects of explicit instruction interventions. As acknowledged by the authors themselves, though, most of the reviewed studies say nothing about the implicit–explicit distinction because the intervention outcomes were measured with tasks most likely tapping into explicit knowledge; such studies demonstrate only that explicit teaching results in explicit knowledge. The same problem pertains to several studies investigating the effect of explicit instruction on article semantics (e.g. Master 2002; Lopez 2017; Sheen 2007; Snape and Yusa 2013). Akakura (2012), however, used tasks designed specifically to tap into implicit knowledge – elicited imitation and oral production – to investigate the effect of explicit instruction on articles and genericity. Learners of English (n = 94) were divided into a control group and a treatment group, and a pre–post–delayed test design revealed a significant and durable advantage for the treatment group. Such demonstrated effects of explicit instruction do not necessarily exclude the possibility of implicit learning of article semantics. In a series of experiments, Williams and colleagues showed that semantic constraints on article use can be learned from exposure without awareness of what is being learned (Williams 2005; see also Paciorek and Williams 2015). Native speakers of English were exposed to English sentences with four artificial articles, two definite and two indefinite ones. As shown with reaction times, they learned that the choice between the two definite and the two indefinite articles depended on whether the referent was animate or not; this was also the case for the participants who could not verbalise this knowledge. Paciorek and Williams (2015, p. 81–82) conclude that, “Linguistic contexts make available a large amount of semantic information that remains implicit in our understanding of events, and which is available to guide inferencing, and, we would argue, participate in learning.” Moreover, Leung and Williams (2012, 2014) found that associations between articles and semantic distinctions that are not encoded by natural grammars, like the relative size of the referent, were not learned in the same kind of experiments (see also Chen et al. 2011). Even though the authors themselves discuss several possible explanations, these results clearly suggest that implicit learning of grammatical meaning is constrained by universal semantic structures. This goes well with the modular view on language and the idea that semantics – the structure of thoughts – is universal and language-independent (e.g. Fodor 1998; Jackendoff 2002; Slabakova 2008).

186

A. Agebjörn

3 Method 3.1 Participants The relation between explicit and implicit knowledge of article semantics was investigated in 26 Belarusian learners of English.4 Their L1s were Russian and, for 13 of them, Russian and Belarusian; both languages lack articles (Mayo 2003; Sussex and Cubberly 2006). Their mean age was 18.8 (sd = 4.4; range = 17–39). They had all started to learn English in primary or secondary school with an average age of onset of 7.4 (sd = 3.1; range = 5–17). Later, 12 of them had studied one or two other article languages, like German, Portuguese or Italian. With one exception, English was their first article language and their non-native language of highest proficiency; one of them had started to learn German before English and was also more proficient in German than in English. Self-reported English proficiency, assessed according to the Russian version of the global scale of the Common European Framework of Reference (Council of Europe 2001), ranged from A2–C1 with a mean in between B1 and B2. In addition, 22 L1 speakers of English, who had learned English before age 3 in an English speaking environment, were recruited to a control group. Their mean age was 20.1 (sd = 8.0; range = 15–55).

3.2 Tasks and Procedures Two tasks were carried out in the study. The first one tested the participants’ use of English articles in a communicative situation; this will be referred to as the implicit task. The second one, a multiple-choice test where they had to both choose articles in given contexts and write explanations for their choices, tested their metalinguistic knowledge of article semantics; this will be referred to as the explicit task. To what extent these tasks actually tap into explicit and implicit knowledge will be discussed in Sect. 5.

3.2.1

The Implicit Task

The implicit task is a map task, a semi-structured elicitation method (Eisenbeiss 2010) loosely inspired of Trenkic (2000). It is designed to elicit NPs in definite and indefinite contexts. Because it forces the participant to focus on the communicative situation, the performance is assumed to reflect mainly the learner’s implicit knowledge.

4 They

were recruited from the Belarusian State University and Centre for Swedish Studies in Minsk for a project on third language acquisition of Swedish. When their English was tested, they had studied Swedish for 4–5 weeks, a few hours per week.

Explicit and Implicit Knowledge of Article Semantics

187

Fig. 2 The implicit task. The participants received a map and had to instruct the researcher where to place the blocks on the board

Between the participant and the researcher, there is a board depicting a town with a main street, a park and several buildings with labels on them – School, Restaurant, Bank etcetera – and there are also blocks with pictures of chairs, cars and buses of different colours; see Fig. 2. The participant gets a map showing were these blocks are supposed to be in the town and must instruct the researcher where to place them accordingly. When this is done, there are new blocks depicting girls, boys and a police officer, and the participant gets a new map showing how these people must be placed in relation to the chairs, cars and buses that are now in the town. The procedure is repeated a third time, now with blocks with cats and dogs. Thus, the participant has to both introduce new, non-identifiable referents and refer to unique locations in the town (“There is a chair in the park”; “Put a white car on the street by the restaurant”) but must also refer to items that have already been introduced (“The girl in the school has a white cat”; “A boy is standing on the chair in the park”). Learners and controls were tested under similar conditions, sometimes in an empty classroom, other times in a quiet library area. The learners had access to a wordlist with nouns, adjectives, verbs and prepositions needed to solve the task; this was to enable them to focus on the communicative situation without having to worry about finding the right words. All participants solved the task in 5–15 min.5

3.2.2

The Explicit Task

After the implicit task, the participants completed a 35-item multiple-choice test (mean number of days between the tasks = 4.9; sd = 3.9) inspired by Ionin et al. (2004), Trenkic (2008), and Yang and Ionin (2009). Each test item was a short 5 It

should be pointed out that the learners had completed the same task once before, a few days up to two weeks earlier, but this time in Swedish, a language they had studied for only three–four weeks up to that point. This might have affected the efficiency with which they solved the task but hardly their use of English articles.

188

A. Agebjörn

dialogue and for each dialogue, the participant had to choose the appropriate article for an NP. For the last seven dialogues, the learners were instructed to write explanations for their choices, as in Yang and Ionin (2009).6 Six of these dialogues, provided in (2), are included in the paper.7 (2)

a. On the phone: (expected answer: a) –Where are you? –Actually, I’m in Madrid. –Wow, what are you doing there? –Right now, I’m sitting in {a, the, ∅} bar with an old friend of mine. Her name is Anna-Lucia. I think you’ve met her once, haven’t you!? b. On the phone: (expected answer: a) –Where are you? –Actually, I’m in Dublin. –Wow, what are you doing there? –Right now, I’m having a beer in {a, the, ∅} pub. It’s a really nice place. It’s called The Dubliner. I like it a lot. c. At a dinner: (expected answer: the) –Do you have kids? –Yes, I have a boy and a girl. {A, The, ∅} girl lives in London now. She’s a designer. d. At a friend’s home: (expected answer: the) –Do you have any pets? –Yes, I have a cat and a dog. –Wow, can I please see {a, the, ∅} cat!? What’s her name? –It’s a he. His name is Mayakovsky. Yes, of course you can see him! e. In the office: (expected answer: the) –We have great problems now, don’t we? What are we going to do? –I’ll call {a, the, ∅} boss. He’s smart. He’s always got a plan. –Ok, great.

6 The controls were not instructed to explain their choices because there is no reason to hypothesise

a relation between implicit and explicit knowledge in this group. for one of the seven dialogues were excluded from the analysis because it turned out that the speaker in this dialogue does not really have to take the listener’s perspective when choosing article. Thus, for the present purpose, these explanations are irrelevant.

7 Explanations

Explicit and Implicit Knowledge of Article Semantics

189

f. At home: (expected answer: the) –What are you reading? –A very interesting book. –Who is {an, the, ∅} author? Do you like him? –Her. Her name is Sara Stridsberg. She’s really good. The dialogues are designed to test whether the learners know that choosing between the definite and the indefinite article requires the speaker to take the listener’s perspective. In dialogues a and b, the bar and the pub are obviously known to the speaker but not to the listener. In the first one, however, the speaker does not intend to say anything about the bar, whereas in the second one, the NP introduces a new discourse referent; thus, only in the latter, the indefinite article could felicitously be substituted with the vernacular English specificity marker this (cf. Ionin 2006; Trenkic 2008). In dialogues c and d, the reference is definite and anaphoric. The speaker in c, contrary to the speaker in d, knows the referent of the NP, a factor that has been shown to potentially affect L2 learners’ article choices (Trenkic 2008). In dialogues e and f, finally, the reference is non-anaphoric, and the definite article must be selected on the assumptions that in an office there is normally a unique boss and a book normally has a unique author. Again, the speaker in e, contrary to the one in f, knows who the referent is.

3.3 Analysis 3.3.1

Analysing the Implicit Task

The recordings of the participants solving the implicit task were transcribed and all singular NPs referring to locations in the town and to the blocks that should be placed in the town were coded for article and reference.8 As for articles, the, this and that were considered definite articles; a, another and one were considered indefinite articles. As for reference, what was considered a definite respectively indefinite context was defined by the behaviour of the controls. Four main types of reference were identified: references to (i) locations in the town (e.g. “the bank”) and to (ii) blocks that had already been placed in the town (e.g. “the girl in the church”) were definite, whereas reference to (iii) non-unique blocks that had not yet been placed in the town (e.g. “take a chair”) was clearly indefinite. Reference to (iv) unique blocks that had not yet been placed, however, were rarely appreciated as definite despite their potential identifiability. This was because most participants

8 Basically,

only the NPs needed to actually solve the task are included in the analysis. NPs like “my mistake” and “the right side” were excluded. This was because the controls used such NPs much more frequently than the learners did. Thus, including them would make a comparison of the groups misleading.

190

A. Agebjörn

focused on the map rather than on the table with the blocks: they said “Take a white bus” independently of whether there were several white buses or only one among the blocks. In other words, reference type (iv) is ambiguous with respect to definiteness. Two variables were calculated: Accuracy and Suppliance * Accuracy. Accuracy refers to the number of articles produced in accurate contexts – that is, definite articles with reference type (i) or (ii) and indefinite articles with reference type (iii) – divided by the total number of articles produced in these non-ambiguous contexts. Because learners who produce very few articles could still get a high Accuracy value, the second variable, which controls for Suppliance, was also calculated. First, the Suppliance rate was calculated. Here, NPs with reference type (i) were excluded, because the labels on the board – “School”, “Park” etcetera – seemed to prime bare NPs in both learners and some controls. Thus, the Suppliance rate was calculated by dividing to total number of non-bare NPs with reference (ii), (iii) or (iv) with the total number of NPs in these obligatory contexts. Finally, the Suppliance rate was multiplied with the Accuracy rate to obtain the Suppliance * Accuracy variable, which indicates to what extent the participant produced many articles in accurate contexts.

3.3.2

Analysing the Explicit Task

The explanations, which were almost exclusively written in Russian, were translated to English by a linguist and native speaker of Russian. The coding of the explanations focused on whether the learners were aware that the choice of article requires the speaker to take the listener’s perspective. Explanations explicitly referring to the role of the listener scored 2; explanation that could be interpreted that way, even though it was not explicitly stated, scored 1; explanation that did not meet these criteria scored 0. Examples of 2-point explanations are given in (3). The explanation in (3a) is faultless. As for (3b), saying that “it was mentioned before” was not considered a sufficient explanation, but because the learner added “for the listener” the explanation still scored 2. Also explanations like (3c), where the learner stated that both speakers know the referent, scored 2. (3)

a. The interlocutor doesn’t know what bar it is. (Собеседник не знает, о каком баре идет речь.) b. for the listener, it [the bar] is mentioned the first time (слушатель впервые сталкивается с упоминанием о нем) c. both understand about whom they are speaking (оба понимают о ком именно идет речь)

Examples of 1-point explanations are given in (4). One common type is represented by (4a–c) where the learners used the pronoun we, seemingly taking the listener’s perspective. Other explanations that scored 1 include words like неизвестный ‘unknown’ or непонято ‘incomprehensible’, as in (4d); here, it was

Explicit and Implicit Knowledge of Article Semantics

191

presumably meant that the bar is unknown to the listener. The same holds for (4e): when writing “your cat,” the learner apparently took the listener’s perspective since the cat belongs to the listener. In (4f), it was stated that there were too little information about the referent, and this was interpreted as there being too little information from the listener’s point of view. Finally, in a few cases, as in (4g), the learner pointed to the uniqueness of the referent – these explanations also scored 1. (4)

a. we haven’t heard about this bar before (мы не слышали ничего об этом баре ранее) b. we know about the cat (written in English) c. In the previous sentence, a girl has already been mentioned. That is, we already know who we are talking about. (В предыдущем предложении уже упоминалась девочка. То есть мы уже знаем, о ком говорим.) d. it’s impossible to understand what bar it is (непонято какой паб) e. definite, your cat (определенный, ваш кот) f. It is not specified in which concrete bar the friends are sitting. (Нет уточнения, конкретики о том, в каком конкретно баре сидят друзья.) g. they can have only one concrete boss (этих людей может быть только один конкретный начальник)

Examples of 0-point explanations are given in (5). In most cases, these are explanations like (5a), based merely on specificity. Frequently used words are неопределенном ‘indefinite’, именно ‘exactly’, любому ‘anyone’, определенному ‘certain’ and, most commonly, конкретный ‘specific, concrete’. Also, as mentioned above, explanations based only on whether or not the referent was mentioned before, like (5b), were not considered sufficient, not least since such explanations in several cases coincided with an incorrect choice of article. Finally, for dialogue e and f, where the definite article should be chosen based on the associations book–author and office–boss, simply pointing to this association was not considered a sufficient explanation; thus, cases like (5c) also scored 0. (5)

a. it’s about some indefinite bar (идет речь о каком-то неопределенном баре) b. the word is mentioned the first time (слово упоминается впервые) c. The concrete author of the book. (Конкретный автор книги.)

An SLA researcher coded all 156 explanations (26 * 6) a second time according to the criteria given above. The 3-point rating scale constitutes an ordinal scale and

192

A. Agebjörn

the inter-rater reliability was accordingly tested with Spearman’s rho. A correlation coefficient of .76 indicates a good, albeit not excellent, inter-rater reliability, showing that this kind of rating is difficult.9 Cases where the two raters made different decisions where discussed. For example, an explanation for choosing the definite article in dialogue f, provided in (6), was considered a 0-point explanation by one of the raters but a 2-point explanation by the other. Of course, saying that it is a “specific book” was not a sufficient explanation, but because the learners added “which he has read,” where “he” can only refer to the listener, it was agreed upon that this was indeed a 2-point explanation; the learner was seemingly aware that the choice of article depended on the state of the mind of the listener. (6)

because it is the author of a specific book, which he has read (потому что автор конкретной книги, которую он прочитал.)

Finally, for each learner, the scores were summarised and divided by 12, the maximum value. This yielded the Explicit knowledge variable, potentially ranging from 0–1.

3.3.3

Analysing the Relation Between Explicit and Implicit Knowledge

Finally, to answer the research question, correlations were tested between the Explicit knowledge variable and each of the two variables obtained from the implicit task: Accuracy and Suppliance * Accuracy. The three variables are treated as continuous and the statistic used was Pearson’s product-moment coefficient.

4 Results 4.1 Results for the Implicit Task Table 1 summarises the total number of elicited NPs included in the analysis by group. These are NPs both with and without adjectival modifiers. However, 86 NPs from the L1 group and 16 NPs from the L2 group, with inherently definite adjectives like first, second, next and same, were excluded from the analysis because these socalled selectors might affect the use of articles (Dahl 2004, p. 153). Table 2 provides the mean Accuracy by group, and mean values for the variables on which this rate was calculated: number of NPs with unambiguously definite reference (type i and ii) respectively indefinite reference (type iii), and number of definite respectively indefinite articles in these contexts. Table 3 provides the mean Suppliance rate, mean values for the variables on which this rate was calculated – number of obligatory 9 Compare

Akakura (2012), where explanations of article errors were rated as either correct or incorrect, and the inter-rater reliability (Cronbach’s alpha) was .64.

Explicit and Implicit Knowledge of Article Semantics

193

Table 1 Raw number of elicited NPs by context and group Group L1 (n = 22) L2 (n = 26)

Def. contexts 1,298 1,034

Indef. contexts 508 521

Obligatory contexts 1,250 1,421

Table 2 Accuracy and the data on which this variable is calculated. All numbers are means; standard deviations are in brackets Group L1 (n = 22) L2 (n = 26)

Def. contexts 59.0 (12.9) 39.8 (16.8)

Indef. contexts 23.1 (2.2) 20.0 (6.2)

Def. articles 60.0 (12.5) 42.7 (21.9)

Indef. articles 22.1 (3.4) 17.1 (10.5)

Accuracy 0.97 (0.04) 0.82 (0.20)

Table 3 Suppliance * Accuracy and the data on which this variable is calculated. All numbers are means; standard deviations are in brackets Group L1 (n = 22) L2 (n = 26)

Obl. contexts 56.8 (5.3) 54.7 (4.4)

Art. in obl. contexts 56.0 (4.4) 44.8 (12.3)

Suppliance 0.99 (0.02) 0.82 (0.22)

Suppl. * Acc. 0.96 (0.05) 0.68 (0.25)

contexts for articles (reference type ii, iii and iv) and number of articles in these contexts – and mean values for the Accuracy * Suppliance variable by group. Welsh two sample t-tests show that the L2 group on average scored significantly lower than the L1 group with regard to both Accuracy (t = 3.84; p = 0.001) and Suppliance * Accuracy (t = 5.64; p = 0.000). Moreover, for both variables, the standard deviation is considerably larger in the L2 group than in the L1 group. A Pearson’s product-moment correlation test shows that, for the L2 group, there is no correlation between Accuracy and Suppliance (r(24) = .26; p = 0.200). In other words, it is not the case that those who produce many articles are also more accurate in choosing which article to use; thus, it makes sense to multiply these two variables to obtain the Accuracy * Suppliance variable.

4.2 Results for the Explicit Task As mentioned above, the controls did not explain their answers in the multiplechoice test. However, to ensure that the contexts for which the learners had to explain their article-choices were indeed either definite or indefinite, the controls also completed the multiple-choice test. Number of correct answers by groups are provided in Table 4. As can be seen, the controls unanimously chose the definite article for dialogues c, d, e and f. For the dialogues a and b, 1 respectively 3 controls diverged from the rest by choosing the definite article. Despite this, the general consistency among the controls indicates that the six dialogues constitute unambiguously definite or indefinite contexts. As for the learners, some of them chose an incorrect article both for the indefinite contexts (a and b) and for the

194

A. Agebjörn

Table 4 Number of expected answers for each dialogue. Accuracy rate within brackets Group L1 (n = 22) L2 (n = 26)

a 21 (0.96) 21 (0.81)

b 19 (0.86) 17 (0.65)

c 22 (1.00) 26 (1.00)

d 22 (1.00) 26 (1.00)

e 22 (1.00) 23 (0.79)

f 22 (1.00) 21 (0.81)

definite contexts where the referent had not been mentioned before (e and f). When the referent was mentioned before (c and d), all learners chose the definite article.10 Their explanations rarely indicated that the learners were aware that the choice of article requires the speaker to take the listener’s perspective. Instead, the explanations most frequently referred to whether or not the referent was specific and whether or not it was mentioned before. Whereas the few accurate explanations were always accompanied by a correct choice of article, many correct choices of articles were accompanied by faulty explanations. Of the 156 explanations, 59.6 percent scored 0, 30.7 percent scored 1 and 9.6 percent scored 2. Only eight of the 26 learners produced at least one 2-point explanation. One learner stood out in scoring 2 for all six explanations (Explicit knowledge = 1.00), but the average Explicit knowledge value was 0.25 (sd = 0.22).

4.3 The Relation Between Explicit and Implicit Knowledge There is no correlation between the Explicit knowledge variable and any of the two variables obtained from the implicit task: Accuracy (r(24) = .10; p = 0.628) and Suppliance * Accuracy (r(24) = .22; p = 0.292). This lack of association is visualised in Fig. 3. As can be seen, a few learners seem to have some explicit knowledge that they apparently do not use in the implicit task. On an interface account, these learners have not yet automatised their explicit knowledge. As can also be seen, several learners use articles fairly accurately despite no obvious explicit knowledge. This suggests that explicit knowledge of article semantics is not a prerequisite for implicit knowledge in this group of learners.

5 Discussion The study has investigated explicit and implicit knowledge of article semantics in 26 learners of English, lacking articles in their L1s, Russian and Belarusian. The research question was whether explicit knowledge, operationalised as an ability to

10 It

should be pointed out that, in the multiple-choice test, the learners chose the zero-article in only 2 percent of the cases. Thus, the average Suppliance rate was considerably higher in the multiple-choice test (0.98) than in the implicit test (0.82). Apparently, the learners explicitly know that articles are used obligatorily. The average accuracy rate, on the contrary, was almost the same for the two tests (0.86 and 0.82 respectively).

0.6 0.0

0.0

0.2

0.4

Explicit task

0.8

0.8 0.6 0.4 0.2

Explicit task

195 1.0

1.0

Explicit and Implicit Knowledge of Article Semantics

0.2

0.4

0.6

0.8

Implicit task: Accuracy

1.0

0.0

0.2

0.4

0.6

0.8

1.0

Implicit task: Suppliance * Accuracy

Fig. 3 There is no significant correlation between the Explicit knowledge variable and any of the two implicit task variables: Accuracy to the left and Suppliance * Accuracy to the right

explain that the choice between the definite and the indefinite article requires the speaker to take the listener’s perspective, was associated with implicit knowledge, operationalised as target-like use of articles in an oral, communicative task. For both tasks, there was a considerable variation within the group, as shown by large standard deviations. Importantly, there was no significant correlation between the variables obtained from the two tasks: some learners seemed to know the meaning of articles despite not being able to verbalise this knowledge; others seemed to have explicit knowledge that they did not apply in the communicative situation. This suggests that, for learners of English lacking articles in their L1s, explicit knowledge of article semantics is neither enough nor necessary for an accurate use of articles in a communicative situation. As discussed by many (e.g. Paradis 2004; Schmidt 1990; Truscott 2015), one cannot know to what extent verbal reports tap into explicit knowledge and to what extent the behaviour in oral, communicative tasks tap into implicit knowledge. As for the verbal reports in the present study, they were admittedly very short. On the one hand, it is possible that another method, for example interviews with follow-up questions or think aloud-protocols, would have revealed more about the participants’ explicit knowledge. On the other hand, there is also the possibility that long answers and follow-up questions would make the learners realise, through online analysis, that article use depends on the listener’s state of the mind, even though they did not actually have this explicit knowledge when the communicative task was solved. Indeed, if the learners explicitly knew that the definite article is used because the listener can understand what the reference is, why would they spontaneously answer that the definite article is used because the reference is a concrete or specific object? Regarding the implicit task, even though it forces the participants to focus on communication, there is nothing in principle that prevents them from employing explicit knowledge in their use of articles (cf. Spada 2015, p. 77; DeKeyser 2003,

196

A. Agebjörn

p. 320). For sure, to some extent they probably did so. Some of them, for instance, started to use articles a few minutes into the task as if they suddenly recalled the typically explicit rule “When speaking English, use articles!” No doubt, the learners in the present study were aware of the existence of articles, and – as shown by the fact that the zero-article was chosen extremely rarely in the multiple-choice test – they were aware that articles are used obligatorily. Moreover, as seen in their explanations, they were also aware that the choice of article has something to do with reference. Crucially, however, most of them were not able to verbalise that this choice depends on the state of the mind of the listener. The fact that there was no significant correlation between the learners’ ability to accurately explain how articles are used and the accuracy of their use of articles in the communicative task clearly indicates that the two tasks measure different things. A plausible conclusion is that most of the learners, just like most native speakers of article languages, did not have explicit knowledge that the choice of article requires the speaker to take the listener’s perspective. It is also likely that whether or not they were aware of this fact had little to do with their implicit knowledge of article semantics. Notice however that the result does by no means exclude the possibility of a relation between explicit and implicit knowledge of article semantics in this group of learners. The study thus suggests that the results from Williams’ laboratory experiments can be generalised to real-world L2 learning. In the experiments, reviewed in Sect. 2.4, native speakers of English learned to associate artificial articles with the animacy distinction without being aware of this association. In English, animacy is a feature of some pronouns, but it is not extensively encoded in the grammar. The same holds for definiteness in Russian and Belarusian: it is a feature in for instance the pronominal system, but it is not obligatorily expressed by articles. It is plausible that Russian and Belarusian speaking learners of English associate article form and meaning, from input and without awareness, just like the participants did in Williams’ experiments. Recall also that implicit semantic learning seemed to be constrained by universal semantic structures (Leung and Williams 2012, 2014); one can speculate that the implicit learning of article semantics happens because semantic definiteness is already present in the learners. Indeed, Fodor (1998, p. 9) writes that “[. . . ] English has no semantics. Learning English isn’t learning a theory about what its sentences mean, it’s learning how to associate its sentences with the corresponding thoughts.” Semantic definiteness is part of thinking, independently of whether a particular grammar encodes this meaning, so the learners in the present study did not actually learn the semantics, but had to associate it with the article form. While explicit knowledge did not turn out to be crucial in the present study, MOGUL and other modular accounts (e.g. Paradis 2004, 2009) acknowledge a possible role of consciousness in L2 learning. For a study like Akakura (2012), which found positive effects of explicit instruction on tasks assumed to tap into implicit knowledge of article semantics, the MOGUL framework offers two possible explanations. According to the first one, the learners might speed up their explicit knowledge to the extent that it can mimic the automatic CS–SS–PS chain (see

Explicit and Implicit Knowledge of Article Semantics

197

Sect. 2.2). According to the second one, it might be that explicit knowledge indirectly boosts the implicit association between the article form and its abstract meaning. To be more precise, since NPs universally are definite or indefinite, each time a definite article is included in an NP with definite reference, the likelihood that the article and the abstract semantic concept will be co-indexed increases. Consequently, if a high-quality explicit (metalinguistic) system – of which one can be aware – can help the learner to produce and interpret articles correctly, it might facilitate the development of the CS–SS–PS chain by feeding the implicit (linguistic) system with better input. “There is, then,” Sharwood Smith and Truscott (2014, p. 302) write, “a clear logic to the monitoring function of metalinguistic knowledge in the MOGUL framework [. . . ].” The result from the present study, however, suggests that explicit knowledge of article semantics is by no means necessary for such growth.

5.1 Implications for Teaching Some SLA researchers and L2 teachers might have taken works by Krashen (e.g. 1981) as a pretext for arguing against “grammar” or Focus on Form and for “communicative” teaching. Today, however, few would take that position. For instance, despite the fact that their interventions studies showed no significant effect of explicit instruction on article semantics, both Lopez (2017) and Snape and Yusa (2013) argued for linguistically informed L2 instruction. This is also the case for Rothman (2008), who compared the acquisition of tense by tutored and naturalistic L2 learners of Spanish. He found that “oversimplified pedagogical rules taught to L2 learners form a system of linguistic knowledge that they use to monitor their output and, thus, affects their performance” and concluded that “pedagogical explanations would greatly benefit from understanding and keeping in mind the grammatical underpinnings to the preterit and imperfect and, to the extent that this is possible, incorporate linguistic rules into teaching” (Rothman 2008, pp. 98–100). In the present study, it happened that learners explained an incorrect use of the indefinite article in the multiple-choice test with an inaccurate rule that they had probably been taught: “Use the indefinite article when the referent has not been mentioned before.” Nevertheless, also this paper will refrain from concluding that teaching should only focus on communication. If explicit rules are provided by teachers and textbooks, these rules should be correct, but as the present study has shown, knowing the correct rules for article use does not automatically enable the learner to use articles correctly. In MOGUL terms, during processing, new structures are created, resting levels are increased, and associations between structures are strengthened. An obvious implication of this is that learning requires practice. The need for practice is underscored by many. Paradis (2009, p. 100) writes that, “Perception is a requisite in providing the target, but it is the repetition of the imitation of the target, not the knowledge, that produces competence.” Hulstijn’s (2015) best advice to L2 learners is to “practice, practice,

198

A. Agebjörn

and practice speaking the language in a variety of communicative situations.” (Slabakova 2008, p. 14) concludes that “learners should be trained systematically on functional morphology, while at the same time being exposed to a wealth of naturalistic, meaningful input, which will allow them to fully engage the linguistic structures involved in language comprehension.” Thus, just like providing the correct rule is not enough, only providing L2 learners with input, or forcing them to communicate in different situations, is an equally inefficient way of teaching. The crucial question is how practicing specific phenomena can be combined with communication. Krashen (1981, p. 114) took the semantics of English articles as an example of phenomena that cannot be learned as simple rules, and he called for improved teaching methods – meaningful and communicative drills: “These are activities in which students can really communicate or in which communication is simulated” (Krashen 1981, p. 104). If the goal is to establish an association between the article and the abstract IDENTIFIABILITY concept, the practice – or “drilling” – must happen in meaningful contexts, because NPs have unambiguously definite or indefinite reference only in meaningful contexts. One concrete way of combining drilling and simulated communication is to play communicative games like the one used to elicit NPs in the present study. First, with help from a teacher or textbook, the learners have to analyse what articles should be used in the what contexts. Here, explicit knowledge comes in handy. Second, playing the game during a few minutes, the learners will hear and produce hundreds of articles in unambiguously definite and indefinite contexts. This way, the association between the article and its abstract meaning will strengthen. This is a crystal clear example of how explicit knowledge – indirectly – might boost the unconscious development of co-indexations between abstract representations in interconnected modules. According to Spada (2015), theories that make a clear distinction between explicit and implicit knowledge, like MOGUL, might very well be accurate, but she calls for caution when it comes to applying such theories in L2 teaching. If explicit processes can be speeded-up so that their output is indistinguishable from the output of the implicit system, she writes, it is neither relevant nor possible for L2 learners and teachers to worry about the two types of knowledge. A similar argument is advanced in Paradis (2009, p. 103): “The distinction between automatic and speeded-up is important as long as it is a theoretical question, but for practical questions, successful L2 speakers do not mind [footnote: Any more than most car drivers worry about how the engine functions.] how and by what means they are able to communicate.” The present paper, however, has argued that a thorough and comprehensive understanding of how linguistic knowledge is acquired, stored and processed is crucial for L2 teachers. Indeed, Hulstijn (2005, p. 130) writes that the explicit–implicit distinction might be important for decision making among “[c]urriculum planners, material designers, teachers, and learners [. . . ]”; Marinis and Cunnings (2018) underscore that the distinction is crucial also in language assessment. According to Whong (2007, p. 148), MOGUL “provides a principled basis from which decisions about pedagogy can be made.” She argues that language teaching can benefit from a modular view on language, and that the main advantage

Explicit and Implicit Knowledge of Article Semantics

199

of MOGUL is that it defines the distinction between explicit and implicit knowledge without downplaying the role of the former: MOGUL “maintains a generative view of language while accommodating broader notions of language development in order to provide an accessible framework of language relevant to adult second language teaching” (Whong 2007, p. 143). Often-used notions like noticing, as well as contemporary teaching theories and methods like genre pedagogics with its focus on form, can be coherently couched within the framework. In a nutshell, the framework can help teachers conceive of the many different types of structures and processes involved in language use and learning. Acknowledgements The paper has benefitted from critique from Åsa Wengelin, Susan Sayehli, an anonymous reviewer and the audience at the Modern Linguistics and Language Didactics workshop in Konstanz, March 2018. Anna Ransheim translated materials to and from Russian; she also contributed to the design of the multiple-choice test. Frida Splendido was the second rater of the qualitative material. Richard Emery and Anika Agebjörn checked the English at different points. The author himself is responsible for all remaining mistakes and faults. Many thanks to all the participants, and many thanks to Nastassia Maiskaya, Mariya Sakovets, Nina Shpakouskaya, Annika Helander, Graham Bowers, Öivind Linnerud and Melanie Lilja for letting me into their classrooms. A travel grant from Stiftelsen Stipendiefonden Viktor Rydbergs minne (DS2016-0766) enabled the data collection in Minsk. Kylskåpsoesi © provided the participating learners with wonderful gifts.

References Akakura, M. (2012). Evaluating the effectiveness of explicit instruction on implicit and explicit L2 knowledge. Language Teaching Research, 16(1), 9–37. Andringa, S., & Rebuschat, P. (2015). New directions in the study of implicit and explicit learning: An introduction. Studies in Second Language Acquisition, 37(2), 185–196. Brun, D. (2001). Information structure and the status of NP in Russian. Theoretical Linguistics, 27(2/3), 109–136. Butler, Y. G. (2002). Second language learners’ theories on the use of English articles: An analysis of the metalinguistic knowledge used by Japanese students in acquiring the English article system. Studies in Second Language Acquisition, 24(3), 451–480. Chen, W., Guo, X., Tang, J., Zhu, L., Yang, Z., & Dienes, Z. (2011). Unconscious structural knowledge of form–meaning connections. Consciousness and Cognition, 20(4), 1751–1760. Cho, J., & Slabakova, R. (2014). Interpreting definiteness in a second language without articles: The case of L2 Russian. Second Language Research, 30(2), 159–190. Council of Europe. (2001). Common European framework of reference for languages: Learning, teaching, assessment. Cambridge: Cambridge University Press. Dahl, Ö. (2004). Definite articles in Scandinavian: Competing grammaticalization processes in standard and non-standard varieties. In B. Kortmann (Ed.), Dialectology meets typology: Dialect grammar from a cross-linguistic perspective (pp. 147–180). Berlin: Walter de Gruyter. DeKeyser, R. (2003). Implicit and explicit learning. In C. J. Doughty & M. H. Long (Eds.), The handbook of second language acquisition, chap 11 (pp. 313–348). Malden: Blackwell Publishing Ltd. Eisenbeiss, S. (2010). Production methods in language acquisition research. In E. Blom & S. Unsworth (Eds.), Experimental methods in language acquisition research (pp. 11–34). Amsterdam: John Benjamins Publishing Company.

200

A. Agebjörn

Fodor, J. A. (1983). The modularity of mind: An essay on faculty psychology. Cambridge, MA: MIT Press. Fodor, J. A. (1998). Concepts: Where cognitive science went wrong. New York: Oxford University Press. Hawkins, J. A. (1991). On (in)definite articles: Implicatures and (un)grammaticality prediction. Journal of Linguistics, 27(2), 405–442. Hulstijn, J. H. (2005). Theoretical and empirical issues in the study of implicit and explicit secondlanguage learning: Introduction. Studies in Second Language Acquisition, 27(2), 129–140. Hulstijn, J. H. (2015). Explaining phenomena of first and second language acquisition with the constructs of implicit and explicit learning. In P. Rebuschat (Ed.), Implicit and explicit learning of languages (vol. 48, pp. 25–46). Amsterdam: John Benjamins Publishing Company. Ionin, T. (2003). Article semantics in second language acquisition. Ph.D. thesis, Massachusetts Institute of Technology. Ionin, T. (2006). This is definitely specific: Specificity and definiteness in article systems. Natural Language Semantics, 14(2), 175–234. Ionin, T., Ko, H., & Wexler, K. (2004). Article semantics in L2 acquisition: The role of specificity. Language Acquisition, 12(1), 3–69. Jackendoff, R. (2002). Foundations of language: Brain, meaning, grammar, evolution. Oxford: Oxford University Press. Krashen, S. D. (1981). Second language acquisition and second language learning. Oxford: Oxford University Press. Leung, J. H., & Williams, J. N. (2012). Constraints on implicit learning of grammatical formmeaning connections. Language Learning, 62(2), 634–662. Leung, J. H., & Williams, J. N. (2014). Crosslinguistic differences in implicit language learning. Studies in Second Language Acquisition, 36(4), 733–755. Lopez, E. (2017). Teaching the English article system: Definiteness and specificity in linguisticallyinformed instruction. Language Teaching 23(2), 1–18. Lucas, C. (2011). Definiteness, procedural encoding and the limits of accommodation. In V. Escandell-Vidal, M. Leonetti, & A. Ahern (Eds.), Procedural meaning: Problems and perspectives (pp. 157–182). Bingley: Emerald Group Publishing Limited. Lyons, C. (1999). Definiteness. Cambridge: Cambridge University Press. Marinis, T., & Cunnings, I. (2018). Using psycholinguistic techniques in a second language teaching setting, second language acquisition. In C. Wright, T. Piske, & M. Young-Scholten (Eds.), Mind matters in SLA (chap 10, pp. 185–202). Blue Ridge Summit: Multilingual Matters. Master, P. (2002). Information structure and English article pedagogy. System, 30(3), 331–348. Mayo, P. (2003). Belorussian. In G. Corbett & B. Comrie (Eds.), The Slavonic languages (chap 16, pp. 887–946). New York: Routledge. Norris, J. M., & Ortega, L. (2000). Effectiveness of L2 instruction: A research synthesis and quantitative meta-analysis. Language Learning, 50(3), 417–528. Paciorek, A., & Williams, J. N. (2015). Semantic implicit learning. In P. Rebuschat (Ed.), Implicit and explicit learning of languages (pp. 69–89). Amsterdam: John Benjamins. Paradis, M. (2004). A neurolinguistic theory of bilingualism. Amsterdam: John Benjamins Publishing. Paradis, M. (2009). Declarative and procedural determinants of second languages. Amsterdam: John Benjamins Publishing. Pereltsvaig, A. (2007). The universality of DP: A view from Russian. Studia Linguistica, 61(1), 59–94. Progovac, L. (1998). Determiner phrase in a language without determiners (with apologies to Jim Huang 1982). Journal of Linguistics, 34(1), 165–179. Rebuschat, P. (Ed.). (2015). Implicit and explicit learning of languages. Amsterdam: John Benjamins Publishing Company. Rothman, J. (2008). Aspect selection in adult L2 Spanish and the Competing Systems Hypothesis: When pedagogical and linguistic rules conflict. Languages in Contrast, 8(1), 74–106.

Explicit and Implicit Knowledge of Article Semantics

201

Rutkowski, P. (2002). Noun/pronoun asymmetries: Evidence in support of the DP hypothesis in Polish. Jezikoslovlje, 3(1–2), 159–170. Sanz, C., & Leow, R. P. (Eds.). (2011). Implicit and explicit language learning: Conditions, processes, and knowledge in SLA and bilingualism. Washington, DC: Georgetown University Press. Schmidt, R. (2012). Attention, awareness, and individual differences in language learning. In W. M. Chan, K. N. Chin, S. K. Bhatt, & I. Walker (Eds.), Perspectives on individual characteristics and foreign language education (vol. 6, pp. 27–50). Boston, MA: De Gruyter Mouton. Schmidt, R. W. (1990). The role of consciousness in second language learning. Applied Linguistics, 11(2), 129–158. Sharwood Smith, M. (2017). Introducing language and cognition: A map of the mind. Cambridge: Cambridge University Press. Sharwood Smith, M., & Truscott, J. (2014). The multilingual mind: A modular processing perspective. Cambridge: Cambridge University Press. Sheen, Y. (2007). The effect of focused written corrective feedback and language aptitude on ESL learners’ acquisition of articles. TESOL Quarterly, 41(2), 255–283. Slabakova, R. (2008). Meaning in the second language. Berlin: Walter de Gruyter. Snape, N., & Yusa, N. (2013). Explicit article instruction in definiteness, specificity, genericity and perception. In M. Whong, K. H. Gil, & H. Marsden (Eds.), Universal grammar and the second language classroom (pp. 161–183). Dordrecht: Springer. Spada, N. (2015). SLA research and L2 pedagogy: Misapplications and questions of relevance. Language Teaching, 48(1), 69–81. Sussex, R., & Cubberly, P. (2006). The Slavic languages. Cambridge: Cambridge University Press. Trenkic, D. (2000). The acquisition of English articles by Serbian speakers. Ph.D. thesis, University of Cambridge. Trenkic, D. (2004). Definiteness in Serbian/Croatian/Bosnian and some implications for the general structure of the nominal phrase. Lingua, 114(11), 1401–1427. Trenkic, D. (2008). The representation of English articles in second language grammars: Determiners or adjectives? Bilingualism: Language and Cognition, 11(1), 1–18. Truscott, J. (2015). Consciousness and second language learning. Bristol/Buffalo: Multilingual Matters. White, L., Belikova, A., Hagstrom, P., Kupisch, T., & Özçelik, Ö. (2011). There aren’t many difficulties with definiteness: Negative existentials in the L2 English of Turkish and Russian speakers. In M. Pirvulescu, M. C. Cuervo, A. T. Pérez-Leroux, J. Steele, & N. Strik (Eds.), Selected proceedings of the 4th conference on generative approaches to language acquisition North America (GALANA 4) (pp. 266–276) Whong, M. (2007). Seeking consensus: Generative linguistics and language teaching. Leeds Working Papers in Linguistics and Phonetics, 12, 143–155. Williams, J. N. (2005). Learning without awareness. Studies in Second Language Acquisition, 27(2), 269–304. Yang, M., & Ionin, T. (2009). Specificity and speaker knowledge in the second language acquisition of English articles. Master’s thesis, Guangdong University of Foreign Studies & University of Illinois at Urbana–Champaign.

L2 Influence in L3 Acquisition: The Role of the L3 Rosalinde Stadt, Aafke Hulk, and Petra Sleeman

Abstract With the present study, we aim at investigating to what extent the L3 affects transfer from background languages. To this end, we investigate two L3s – L3 German and L3 French – while keeping the L1 (Dutch) and the L2 (English) constant. To examine the two L3s in contrast, we compare data from two previous studies on third-year Dutch/English bilingual stream students (L3 French) to new data gathered amongst the same type of learners taking German as an L3. Data were gathered by means of a gap-filling task and a grammaticality judgement task. The results indicate significantly more L2 English influence on L3 French than on L3 German. We argue that this is due to perceived typological and structural resemblance between L1 Dutch and L3 German. Keywords Third language acquisition · Foreign language learning · Secondary school · L1 Dutch · L2 English · L3 French

1 Introduction In the field of morphosyntactic third language acquisition (L3A) research, studies are generally focused on investigating the role of the first language (L1) vs the role of the second language (L2) in L3 learning. In the last decades or so, various language combinations have been studied and empirical evidence has been found for different scenarios, including preferred influence of the L1 (L1 transfer scenario, Na Ranong and Leung 2009; Jin 2009; Hermas 2010, 2014a, b) or the L2 on the L3 (L2 status factor hypothesis, Bardel and Falk 2007, 2012; Falk and Bardel 2011) or transfer from either the L1 or the L2 into an L3 as a result of general typological grouping (Typological Primacy Model [TPM], Rothman 2011, 2013, 2015), or on (abstract) structural resemblance between one of the background languages and the L3 target

R. Stadt () · A. Hulk · P. Sleeman Department of Literary Studies and Linguistics, ACLC (Amsterdam Centre of Language and Communication), Amsterdam, The Netherlands e-mail: [email protected] © Springer Nature Switzerland AG 2020 A. Trotzke, T. Kupisch (eds.), Formal Linguistics and Language Education, Educational Linguistics 43, https://doi.org/10.1007/978-3-030-39257-4_11

203

204

R. Stadt et al.

language (Linguistic Proximity Model [LPM], Mykhaylyk et al. 2015; Westergaard et al. 2016).1 However, in morphosyntactic L3 studies on transfer, the research focus is generally on comparing the influence of L1 transfer and/or L2 transfer in relation to one and the same L3. With the present study, we aim at investigating to what extent transfer from similar background languages will be the same with a different L3. To this end, the research goal of this study is to compare two L3s – L3 German and L3 French – while keeping the L1 (Dutch) and the L2 (English) constant. This study derives from previous work (Stadt et al. 2016, 2018a, b, c) in which we both compared the role of L1 Dutch and L2 English in L3 French and investigated development with respect to the influence of Dutch (L1) and English (L2) in L3 French. Data were collected in the first 4 years of a secondary school in the Netherlands offering both a Dutch/English bilingual stream and a mainstream programme. In the bilingual stream programme, students receive at least 50% of their subjects in English whereas in the mainstream programme students follow the Dutch curriculum: classes are taught in Dutch, and English is taught as a subject. Results of these studies suggested that transfer occurs from both L1 Dutch and L2 English depending on the amount of L2 vs L1 exposure in the daily context and through the years (Stadt et al. 2016, 2018c) and on the developmental stage the L3 French learner is in (Stadt et al. 2018a, b). Even though in these studies different variables have been the object of investigation, the L3 (French) remained constant. We did not investigate what happens when we look at a different L3 with the same L1 and L2. By comparing the data on L3 French to new data on L3 German while keeping other variables as constant as possible, we aim at learning more about the interplay between background languages and different target L3 languages. A remarkable result from two of our previous studies (Stadt et al. 2016, 2018c) was the substantial role of L2 English in L3 French, especially in third-year bilingual stream students, when learners are exposed to a great deal of English in the daily school context and have been exposed to English through the years. The purpose of this study is to investigate whether this strong L2 influence that we found in the third-year bilingual stream context in the language combination L1 Dutch– L2 English–L3 French, also applies to the language combination L1 Dutch–L2 English–L3 German. To examine the two L3s in contrast, we therefore decided to investigate the extent to which the same type of intermediate L3 learners (in the same secondary school) transfer from English into German. To this end, the research question we address is: To what extent does L2 English play a similar role in L3 German as compared to L3 French? The paper is structured as follows. Considering that we build on previous studies, we start in Sect. 2 by presenting the linguistic description of the word order in declarative root clauses in the languages involved (Dutch, English, French, and German). In Sect. 3, we will address our previous research findings (Sect. 3.1) and will discuss some other studies that have been conducted on the acquisition

1 The

TPM builds on Kellerman’s idea on the influence of unconsciously perceived similarity between background and target language in SLA (Kellerman 1983, 1986).

L2 Influence in L3 Acquisition: The Role of the L3

205

of German as an L3 (Sect. 3.2). In Sect. 4 we first present the context of study (Sect. 4.1) followed by the design of the present study: first, the research question and the prediction relevant to this study (Sect. 4.2), followed by the participants (Sect. 4.3), the experiments (Sect. 4.4), and the procedure (Sect. 4.5). In Sect. 5, we present the results of the L3 German group and compare them to the results from our previous studies on L3 French. This is followed by a discussion of the findings (Sect. 6) and some concluding remarks (Sect. 7).

2 Linguistic Description: V-to-T Movement As the present study builds on our previous findings, in which transfer from the L2 was studied using the same construction, viz. verb placement construction in the language combination L1 Dutch–L2 English–L3 French, in this section we first concentrate on the linguistic description of verb placement with respect to all languages involved in the present study and earlier studies. The position of the adverb in Dutch, English, French, and German main clauses depends on whether the verb moves out of the VP (to T) (Pollock 1989). The asymmetry between English and Dutch, German, and French is due to a difference with respect to this type of verb movement: in German, Dutch, and French declarative main clauses, the verb moves to T, which results in the adverb appearing post-verbally (Verb-Adverb word order) in the surface structure (as illustrated in 1, 2, and 3). In English there is no such verb movement (as illustrated in 4), which results in the adverb appearing preverbally (Adverb-Verb word order) in declarative root clauses (Pollock 1989). (1)

Manon geht manchmal in den Zoo. *Manon goes sometimes to the zoo. ‘Manon sometimes goes to the zoo.’

(German)

(2)

Manon gaat soms naar de dierentuin. *Manon goes sometimes to the zoo. ‘Manon sometimes goes to the zoo.’

(3)

Manon va parfois au zoo. (French) *Manon goes sometimes to the zoo. ‘Manon sometimes goes to the zoo.’

(4)

Manon sometimes goes to the zoo. *Manon manchmal geht in den Zoo. *Manon soms gaat naar de dierentuin. *Manon parfois va au zoo. ‘Manon sometimes goes to the zoo.’

(Dutch)

206

R. Stadt et al.

Table 1 Dutch, English, French and German word order in declarative root clauses Verb movement Word Order Dutch English French German V-to-T movement Verb-Adverb Yes No Yes Yes

Concentrating on this type of verb movement offers two advantages to our study: (1) English differs in this respect from all the other languages involved, and (2) the students under investigation are intermediate L3 learners who have already received a great deal of L3 exposure (see Sect. 4.3 for more detailed information). By concentrating on errors in the L3 – that is, on negative transfer – we avoid the risk that transfer is simply increased L3 knowledge. In Table 1, verb placement in declarative root clauses is illustrated in all four languages. In the next section, we will give a summary of the results of our previous studies followed by other studies on L3 German conducted in the L3 field of research.

3 Background 3.1 Results of Our Previous Studies In Stadt et al. (2016, 2018a, b, c), we considered syntactic transfer both from L1 Dutch and from L2 English in L3 French in various stages of acquisition. Data were gathered by means of two offline tests: a grammaticality judgement test (henceforth GJT) to measure receptive knowledge and a gap-filling test (henceforth GFT) to measure guided production. Transfer from the L2 was detected via English Adv-V word order errors in French (as described above in Sect. 2), while transfer from the L1 was detected via negative transfer of the Dutch V2 rule, i.e. XVS(O) word order in French. In Dutch declarative root clauses with a sentence-initial adverb, the verb appears in the second position of the clause (in sentences such as Vandaag doet Manon haar examen *Aujourd’hui passe Manon son examen, *Today takes Manon her exams ‘Today Manon takes her exam’). The tests were very easy to take for all students, and the students were familiar with the vocabulary because we aimed at avoiding L3 proficiency being a variable. In what follows, we will summarise the main findings from previous studies in order to set the starting point of this study. In Stadt et al. (2016), we compared the acceptance of Dutch (L1) and English (L2) word order in French (L3) in 16 third-year bilingual stream students and in 11 third-year mainstream students by means of a GJT. The bilingual stream students accepted the English (L2) Adv-V word order more often than the Dutch (L1) XVS(O) word order (42.4% vs 24.6%, p = 0.005). In mainstream students, however, we found an equal proportion of Adv-V and XVS(O) errors in French (34.4% vs 37%, p = 0.742). L2 was thus found to have a substantial role in L3 French learning only in the bilingual stream students, who were exposed to and used the L2 to a greater extent in their day-to-day school environment. Additionally, findings of the same study also demonstrated a significant difference between mainstream and

L2 Influence in L3 Acquisition: The Role of the L3

207

Table 2 Results of Adv-V and XVS(O) constructions in GJT for third-year bilingual stream and mainstream students Y3 bilingual group Y3 mainstream group

N 16 11

Table 3 Results of Adv-V constructions in GFT for firstand third-year bilingual stream students

Adv-V errors 95/224 (42.4%) 53/154 (34.4%) p = 0.261

XVS(O) errors 55/224 (24.6%) 65/154 (37%) p = 0.033

First-year bilingual group Third-year bilingual group

N 21 16

p = 0.005 p = 0.742

Adv-V errors 10/168 (6%) 41/192 (21.4%) p = 0.003

bilingual stream students with regards to transfer from the L1, i.e. the mainstream students accepted the Dutch (L1) word order significantly more often than their bilingual stream counterparts (37% vs 24.6%, p = 0.033).2 We interpreted these results by stating that in a bilingual stream context (as compared to a mainstream context), when the L2 is more activated in the daily school context, the L1 is suppressed to a greater extent than in the mainstream group. In a bilingual stream context, L2 English thus seems to have an important influence on L3 French in intermediate learners. To give a clear overview, we repeat the results of this study in Table 2. In a cross-sectional study (Stadt et al. 2018c), we compared the number of AdvV errors in first-year bilingual stream students (who were in the first regular week of secondary school and were therefore on the threshold of being enrolled in the bilingual programme) to those of third-year bilingual stream students by means of a GFT. In that study, we found significantly more L2 English influence in intermediate stages than in initial stages of L3 French acquisition in the bilingual stream context: third-year bilingual stream students displayed more Adv-V errors in French than first-year bilingual stream students (p = 0.003). We repeat the results of Stadt et al. 2018c in Table 3. In a longitudinal study (Stadt et al. 2018b) we focused solely on bilingual stream students. In this study, in which we were also concerned with XVS(O) errors (L1 transfer) and Adv-V errors (L2 transfer) in L3 French, we tested the students three times: at the start of the first, second, and third years. The data were gathered by means of a GJT and a GFT. The results of the first-year students – in the initial stages of L3 acquisition – suggested that the L1 is the most popular source of transfer (which is in line with Na Ranong and Leung 2009; Jin 2009; Hermas 2010, 2014a, b). The first-year students incorrectly used and misjudged the XVS(O) word order in L3 French significantly more often than the second-year students (in the GJT:

2 The

difference between mainstream and bilingual stream students regarding transfer from the L2 was not significant. Nonetheless, the bilingual stream students did show a tendency to misjudge the English word order more often than the mainstream students (42.4% vs 34.4%, p = 0.261).

208 Table 4 Results of Adv-V and XVS(O) constructions in GJT for first- and second-year students

Table 5 Results of Adv-V and XVS(O) constructions in GFT for first- and second-year students

R. Stadt et al. N 18 18

First-year group Second-year group

First-year group Second-year group

N 18 18

Adv-V errors 34/98 (34.7%) 32/98 (32.7%) p = 0.68 Adv-V errors 21/144 (14.6%) 22/144 (15.3%) p = 0.88

XVS(O) errors 83/126 (65.9%) 41/126 (32.5%) p < .001 XVS(O) errors 115/144 (79.9%) 19/144 (13.2%) p < .001

65.9% vs 32.5%, p = .002 and in the GFT: 79.9% vs 13.2% p < .001). We explained the decrease in L1-errors in the students by stating that they ‘unlearned’ the V2 rule. Since English and French display the same word order in this type of sentence (XSV(O) word order), we also interpreted the decrease of XVS(O) errors as due to an increasing (positive) influence on English (L2) in French or to an increased L3 proficiency. The findings of this study showed that in bilingual stream secondary school students, in the language combination L1 Dutch–L2 English–L3 French, the L1 plays a substantial role in the initial stages of L3 acquisition. However, the L2 remains stable in later stages of acquisition (when students are surrounded by English in the daily context). In Tables 4 and 5, we summarise the main points that we discussed with respect to the longitudinal study. In summary, data from our previous studies suggested that in the language combination L1 Dutch–L2 English–L3 French, L2 transfer was found especially in bilingual stream students who were in intermediate stages of L3 acquisition. In what follows, we address some studies in which transfer from the L1/L2 into German (L3) was the object of investigation.

3.2 Other Studies on L3 German In the field of syntactic L3A, several studies have been conducted on L3 German. For instance, Falk and Bardel (2011) tested 44 intermediate level L3 German learners with either L1 French and L2 English as background languages (n = 22) or L1 English and L2 French as background languages (n = 22) on the acquisition of object-pronoun placement in main clauses (‘je le vois’, ‘I see him’, ‘Ich sehe ihn’ – L1/L2 French = L1/L2 English = L3 German) and subordinate clauses (‘You know that I see him’, ‘Tu sais que je le vois’, ‘Du weisst dass ich ihn sehe’– L1/L2 English = L1/L2 French = L3 German) (Falk and Bardel 2011: p. 60). Data were gathered by means of a judgement and correction task (GJCT). Half of the German sentences were grammatical, while the other half were ungrammatical. The results of this study showed that the judgements were most influenced by the L2s. Bohnacker (2006) examined verb placement in L3 German and concentrated on the use of the XVS(O) word order in an oral production task. The participants in

L2 Influence in L3 Acquisition: The Role of the L3

209

this study were six Swedish (+V2 language) initial-state learners of German (+V2 language), three of whom had prior knowledge of English (–V2 language) as an L2 and three who did not.3 The learners were old-age pensioners and elementary learners of German. The data suggested that although learners did positively transfer the XVS(O) word order from Swedish into L3 German, L2 knowledge of a non-V2 language (English) did influence transfer negatively. The studies of Rutgers (2013) and Rutgers and Evans (2015) also focus on L3 German, and these studies were – just as the present study – conducted in a partially L1 Dutch/L2 English bilingual school in the Netherlands. Amongst other things, Rutgers (2013) investigated the effect of L1 Dutch/L2 English bilingual education on the acquisition of L3 German in fourth-year secondary school students (15/16 years old). The results of her study suggested that in both the bilingual stream and non-bilingual stream group, Dutch took on a strong role with regards to the pupils’ processing of German and in relation to their (implicit and explicit) understanding of the structure of the German language. English, on the other hand, provided a considerably smaller role in the processing of German than Dutch did. However, it had a bigger role in the bilingual stream than in the non-bilingual stream group. For instance, in the bilingual stream the students’ writing skills were influenced by the English bilingual stream lessons, that is it affected word order, resulting in errors in the target language. In a longitudinal study, Rutgers and Evans (2015) focused on language processing and the role of metalinguistic skills in L3 German learning. Six bilingual stream and six mainstream stream fourthyear students (15/16 years old) were followed over a period of 6 months. They were interviewed, observed during their German classes, and monitored during think-aloud tasks in L3 German. Although it was difficult to measure levels of metalinguistic awareness, the findings did show a difference between bilingual stream and mainstream students with respect to L3 processing. With regards to transfer from the L2, this study found that the bilingual stream students used their more advanced L2 English writing skills and strategies in L3 German to a greater extent than the mainstream students. In essence, these studies all show that L2 English can affect L3 German acquisition (Bohnacker 2006; Falk and Bardel 2011; Rutgers 2013; Rutgers and Evans 2015), especially in a bilingual stream context (Rutgers 2013; Rutgers and Evans 2015). Before we proceed to our research question and prediction, let us describe the educational environment, which is relevant to set the context of the present study.

3 The

+/− indicates whether a feature is present (+) or not (−).

210

R. Stadt et al.

4 Design of This Study 4.1 Context of Study The data used in the present study were collected at the same partially bilingual secondary school as in previous studies. At this school, students can opt for a fouryear bilingual L1 Dutch L2 English educational programme – the Middle Years Programme (MYP) of the international baccalaureate – in which several school subjects, such as ‘Sciences’, ‘Humanities’, ‘Technology’, ‘Physical Education’, and ‘Performing Arts’ are taught in English. The MYP students receive over 50% of the school subjects in English. In the third year of the programme, this even rises to approximately 58%. After 3 years of bilingual stream education, students have received approximately 1942 h of input in English.4 The students in this school have homogeneous socioeconomic backgrounds (SES). In the region in the Netherlands where we collected our data, SES scores are high compared to other regions. Moreover, the students’ overall learning ability is more or less the same because of the Dutch school system, in which students are divided into three different main tracks after elementary school: a four-year pre-vocational programme (VMBO), a five-year intermediate track (HAVO), and a six-year pre-university track (VWO). At the end of most Dutch elementary school programmes, students take a nationwide standardised test to measure their overall learning ability. On the basis of this test and in accordance with the recommendation of the primary school teacher, the pupils are placed in one of the three tracks in secondary school. To enter this school’s Middle Years Programme, students need to be placed in either of the two more academically oriented tracks (HAVO or VWO).

4.2 Research Question and Prediction Several studies on L3 German have found a substantial role for the L2 in L3 learning (Bohnacker 2006; Falk and Bardel 2011) and more L2 influence in bilingual stream students than in mainstream students (Rutgers 2013; Rutgers and Evans 2015). The general conclusion we drew from our previous results in which we examined the language combination L1 Dutch–L2 English–L3 French is that the L2 may take on a strong role in the L3 in intermediate bilingual learners (Stadt et al. 2016, 2018c). The goal of the present paper is to further extend our knowledge of the L2 status in L3A by examining whether the L2 transfer that we found in the language combination L1

4 Considering

that a school year consists of about 35 weeks per year and that a class takes 45 min, the bilingual stream students have received 1627 h of instruction in English after 3 years: 17 classes per week (446 h) in the first year, 21 classes per week (551 h) in the second year, and 24 classes per week (630 h) in the third year. Additionally, the students also receive four 45-min classes of English as a subject in each year (315 h by the end of Y3).

L2 Influence in L3 Acquisition: The Role of the L3

211

Dutch–L2 English–L3 French also occurs in the L1 Dutch–L2 English–L3 German combination. To this end, we repeat the research question of this paper here: To what extent does L2 English play a role in L3 German similar to the one it plays in L3 French? Based on previous studies in which we found L2 transfer into L3 French in third-year bilingual stream students (Stadt et al. 2016, 2018c) and based on other studies in which L2 English transfer into L3 German was found (Bohnacker 2006; Falk and Bardel 2011) we predict a comparable number of Adv-V word order errors in L3 German as in L3 French in bilingual stream L1 Dutch/L2 English secondary school students.

4.3 Participants The participants consist of 22 Dutch third-year secondary school students learning German. These students are compared to the L3 French third-year bilingual stream students from Stadt et al. (2016, 2018c), in which data were gathered on 16 students (see Tables 2 and 3). The L3 German students are enrolled in the bilingual MYP educational programme and are therefore exposed to the same amount of English as the third-year L3 French students (Stadt et al. 2016, 2018c).5 German is taught in a formal school environment. German as a school subject starts in the second year of secondary school. By the end of year three, the students have received 131 h of L3 input and their level of German is approximately A2 according to the European Framework of Reference (CEFR). French is also taught in a formal school environment. Although French is taught from the first year on, the students do not necessarily have a higher level in French than in German at the end of year three; their CEFR level in French level at the end of their third year is approximately A2 (Council of Europe 2001). To be included in the analysis the students had to meet several criteria. Dutch had to be their only L1, and English the subsequent L2. We therefore excluded all simultaneous bilinguals. Furthermore, we also excluded students with French/English families, and students who were near native in a language other than Dutch.

4.4 Experiments To gather data, we used the same type of offline tasks as in the L3 French studies: a GJT and a GFT. Since we wanted to keep the dependent variables as constant as possible as compared to previous studies, the German tasks were based on the

5 It

is also important to bear in mind that English is ubiquitous in the Netherlands; all youngsters also receive a great deal of English input outside the school context due to music, films, television, the Internet, social networks, and apps on smartphones and tablets (Verspoor et al. 2007, 2010).

212

R. Stadt et al.

content of the French tasks. We used offline tasks so that we could gather our data during regular school hours and because reaction time was not the main interest of this study. It was more important that the students did not feel too much pressure. For both tasks the test items were simple, and consisted of short sentences with vocabulary that the students were familiar with. We handed out a vocabulary list before testing and used a lot of cognates so that we were absolutely sure that the vocabulary would not stand in the way of comprehending the test. The number of items was kept constant in comparison to our previous studies on third-year students. The GJT contained 45 items, of which 14 were test items testing the Adv-V word order. Half of the test items were grammatical, and the other half were ungrammatical. However, in the German experiments, we only looked at Adv-V word order in German whereas in the French tests, we looked at both Adv-V word order and at XVS(O) word order. Since German and Dutch are both V2 languages, we did not take this construction into account. The students had to indicate whether they deemed a sentence correct or incorrect. When they judged a grammatical sentence incorrect or an ungrammatical sentence correct, they got a ‘miss’, and when they judged an ungrammatical sentence incorrect or a grammatical sentence correct, they got a ‘hit’. The GFT contained 36 items, of which 12 items tested the Adv-V word order. The students had to fill in the given verb in one of the two gaps. In both tasks, distractors were used to check whether the students understood the test and whether they had taken it seriously. Just as with the test items, the fillers in both tests were simple constructions with a subject, a finite verb, and a direct and/or indirect object. According to the curriculum, the students had not received explicit instruction on verb placement in declarative root clauses with a manner/frequency adverb. Example from the German GJT: (5) Wir essen wirklich viele Bonbons! *Wir wirklich essen viele Bonbons! ‘We really eat a lot of bonbons.’

c/i c/i

Example from the German GFT: (6) Susan und Jan . . . . . . ...oft . . . . . . eine Serie. ‘Susan and Jan often watch a series.’

schauen

To be able to detect transfer from this specific English Adv-V word order construction, it was essential that the students be familiar with verb placement in these types of English sentences. Therefore, after the German tests, we also presented them with an initial English gap-filling task (as illustrated in 7). Example English gap-filling task: (7) John . . . . . . . . . . . . .... often . . . . . . . . . . . . . .too much sugar. ‘John often eats too much sugar.’

eats

The students had to fill in the gap while using the given verb. The English gap-filling task contained 24 items, eight of which tested the English Adv-V word order. We

L2 Influence in L3 Acquisition: The Role of the L3

213

excluded all students who made >3 errors in the gap-filling task. We started out with 30 students taking German. On the basis of this test and a background questionnaire containing language related questions, we had to exclude eight L3 German students.

4.5 Procedure and Analysis The testing took place during school hours. The German data were collected at the end of May 2017, and the French data had been gathered in May 2015. We started out with the German tests (GJT and GFT respectively) followed by the English GFT because we did not want to bias the students with respect to influence from English. After the linguistic tests, the students filled in the background questionnaire. Since reaction time was not the focus of this study, we stressed that there was no need to hurry. However, we also instructed them that concentration was very important, and we emphasised that they should not hesitate too much and that there were no ‘wrong’ answers. All data were controlled for normality using a Shapiro–Wilk test. Since the samples were small, we used a non-parametric Mann–Whitney U test to calculate the difference in Adv-V errors in French and in German. The dependent variable was the number of errors.

5 Results In this section, we present the results. To be able to compare the L1 Dutch–L2 English–L3 German group to the L1 Dutch–L2 English–L3 French group, we also repeat the data from the French grammaticality judgement task (Stadt et al. 2016) and the French gap-filling task (Stadt et al. 2018c). Before presenting the statistical analysis, we will first examine the distribution of the number of Adv-V errors, in boxplots representing the distribution of the number of Adv-V errors in German and in French in the GJT (Fig. 1) and in the GFT (Fig. 2). The L3 German and L3 French boxplots are presented in the same figure. In both figures, the left boxplot (1,00) represents the German distribution of the data, and the right boxplot (2,00) represents the French distribution of the data. On the y-axis, the number of Adv-V errors are listed. The lower the boxplots, the fewer Adv-V errors the students made. Figure 1 shows that with respect to the Adv-V judgement errors in German (the left boxplot), the median – i.e. the centre of the data – is one. The interquartile range box – the tinted area that represents 50% of all the scores – shows that half of the scores fall between zero and three errors and the top quartile of scores – representing the highest 25% of the scores – shows that the highest 25% fall within the range between three and seven errors. 25% of the students (the lowest quartile of

214

R. Stadt et al.

Fig. 1 Number of Adv-V errors in German and French in the grammaticality judgement task

scores) did not make any errors. The upper and lower whiskers show that minimum value is 0 and the maximum value is 7. The right boxplot revealing the distribution of the Adv-V errors in French shows that the median number of errors is six. It also reveals that 50% of the scores – the interquartile range box – fall between four and seven and a half errors and that the highest 25% of the scores (the top quartile of scores) fall within the range between seven and a half and eleven errors. The boxplot also shows that the lowest 25% of scores (the lowest quartile of scores) fall within the range of two and four and a half errors. The upper whisker shows that the maximum value is eleven, while the lower whisker shows that the minimum value is two. Figure 2 presents the boxplots of the gap-filling task. The left boxplot (1,00) represents the number of Adv-V errors in German. The boxplot shows that there is one outlier: one student (participant number 1) made 8 Adv-V errors in German. It also reveals that 75% of the scores (the interquartile range box and the lowest quartile of scores) fall between zero and one error. The boxplot also shows that the top quartile of scores – i.e. the highest 25% – falls within the range between one and two errors and that the median is zero. The boxplot revealing the distribution of the Adv-V errors in French (right boxplot in Fig. 2) shows that 50% of the scores fall between one and four errors and that the highest 25% falls within the range between

L2 Influence in L3 Acquisition: The Role of the L3

215

Fig. 2 Number of Adv-V errors in German and French in the gap-filling task Table 6 Results of Adv-V errors in French and German in the GJT French group German group

N 16 22

Adv-V errors 95/224 (42.41%) 44/308 (14.29%)

mean 5.94 2.00

SD 2.24 2.23

Min. 2 0

Max. 11 7

Min. 0 0

Max. 7 2

Table 7 Results of Adv-V errors in French and German in the GFT French group German group

N 16 21

Adv-V errors 41/192 (21.35%) 10/252 (3.97%)

mean 2.56 0.48

SD 2.19 0.75

four and seven errors. The lowest quartile, i.e. 25% of scores, falls between zero and one error. The median, however, is two. In Table 6, we display the results of the GJT: the total number of Adv-V errors and the corresponding percentages in both French and German, the standard deviation (SD) and the accuracy scores. In Table 7, the same information is given with regards to the GFT. As the boxplot that represents the distribution of the AdvV errors in the GFT showed, one student in the German group was an outlier. Therefore, this student is not considered in the statistical analyses.

216

R. Stadt et al.

To calculate the difference between French and German, we used a nonparametric Mann–Whitney U test. The results of the GJT revealed a significant difference between the number of Adv-V errors in French and German (U = 39.00 z = −4.08, p < .001). The same non-parametric Mann–Whitney U test also revealed a significant difference between the number of Adv-V errors in French and German in the GFT (U = 63.50 z = −3.40, p < .001).6

6 Discussion 6.1 Interpretation of the Results The research question we addressed was: In L1 Dutch students, to what extent does L2 English play a role in L3 German similar to the one it plays in L3 French? We predicted a comparable number of English Adv-V word order errors in both L3 groups based on previous results in which we found substantial L2 influence in L3 French bilingual stream students (who receive a considerable amount of L2 exposure through the years and in the bilingual stream daily context) and based on our other studies in which syntactic transfer from L2 English on L3 French was found in different types of learners. The experimental data in this study, however, suggest that third-year bilingual stream secondary school students transfer English (L2) significantly more often into French (L3) than into German (L3): from the results presented in Tables 6 and 7 (GJT and GFT respectively), we can conclude that Adv-V errors are more easily made in L3 French than in L3 German. In other words, we did not find evidence for the prediction that the role of L2 English is the same in both L3s. The fact that English is an important source of negative transfer into L3 French and less so in L3 German could be explained by various factors. Keeping in mind that other variables important to this study – such as the experimental conditions, the construction tested, the school context, and the amount of L2 exposure – were kept as constant as possible, we claim that the (amount of) L2 effect depends on the L3 and more specifically on the typology of the language combinations involved. With regards to our results, we argue that the fact that German (L3) is less affected by English (L2) than French (L3) is due to the influence of Dutch (L1); that is, the (typological) relation between the L3 and the L1 is a relevant factor for (the amount of) L2 transfer. The low number of Adv-V errors in German is due to the strong (positive) influence of Dutch in the language combination L1 Dutch–L2 English–L3 German. The Typological Primacy Model [TPM], Rothman 2011, 2013, 2015), claims that L3 transfer can occur from either the L1 or the L2 depending

6 Including

the outlier in the GFT, the mean rises to 0.82 and the SD rises to 1.76. The difference between the number of Adv-V errors in French and German in the GFT including the outlier is also significant (U = 79.50 z = −3.01, p = 0.003).

L2 Influence in L3 Acquisition: The Role of the L3

217

on perceived typological resemblances between background and target languages. This model accounts for the initial stages of L3 transfer. The L3 learner copies the full grammar of the background language that is perceived to be the most similar to the L3. The results of the present study show heavy influence of L1 Dutch in L3 German in intermediate learners, two languages that are typologically very closely related. Therefore, we could argue that the TPM may predict transfer in later stages of L3 development in the language combination L1 Dutch–L2 English– L3 German. However, since the students under investigation in this study are intermediate learners and not learners in the initial stages of acquisition, and since it is difficult to measure ‘wholesale transfer’ in these stages of acquisition, it would be rather ambitious to state that the results of this study are in line with the TPM. Furthermore, results from a previous study revealed that in the initial stages of L3 French acquisition students massively use the Dutch word order (Stadt et al. 2018b). Dutch and German are similar at a structural level and with respect to overall typological grouping (both belonging to the Germanic language family), which is not the case in the language combination L1 Dutch–L2 English–L3 French (French being a Romance language). Not only do Dutch and German share the V-to-T word order (as Dutch and French do as well), they generally display more structural similarity, i.e. they share the same word order (V2) in other main clause constructions as well. As such, these results could be explained by The Linguistic Proximity Model [LPM] (Mykhaylyk et al. 2015; Westergaard et al. 2016). According to this L3 model, similarity of (abstract) linguistic properties may cause transfer into the L3. When there is structural overlap between one of the background languages and the L3, transfer occurs. Since according to the LPM, transfer is ‘property by property’, the model also predicts transfer in later stages of acquisition. Apparently, in this case, the L3 German students – who are intermediate learners of German – are aware of these structural resemblances in main clauses and perceive Dutch as similar to German and therefore the most suitable language for transfer. Consequently, they copy the Dutch V-to-T word order onto L3 German at the same time avoiding Adv-V errors. However, they copy the Dutch V-to-T word order onto L3 French less often, which could explain why in the language combination L1 Dutch–L2 English–L3 French, Adv-V errors are avoided significantly less often. Thus, L1 Dutch does not get overruled by L2 English in the acquisition of L3 German in intermediate learners, contrary to what happens in the acquisition of L3 French, in which we found that in bilingual stream learners the L1 is more suppressed in third-year students than in their first-year counterparts. The results of the present study scarcely showed any Adv-V errors in German (in 10/252, 3.97% (GFT) and in 44/308, 14.29% (GJT) of the cases). Therefore, the findings of the present study do not support the L2 status Hypothesis (Bardel and Falk 2007, 2012; Falk and Bardel 2011). Our results are not in line with studies on L3 German that have found substantial influence from L2 English on L3 German (Falk and Bardel 2011; Bohnacker 2006; see Section 2.4). Falk and Bardel (2011) found both positive and negative transfer from the L2 on L3 German regardless of the ‘nature’ of the L2 (L1 Swedish–L2 English–L3 German or L1 Swedish–L2 French–L3 German). In Bohnacker’s study (2006), negative L2 English influence

218

R. Stadt et al.

on L3 German was found in the language combination L1 Swedish–L2 English–L3 German.7 For future research, it would be interesting to compare syntactic transfer with different L1’s while keeping the L2 (English) and L3 (German) constant, such as with L1 Swedish and L1 Dutch. Our findings are in line with Rutgers (2013), who also found that L3 German bilingual stream students transferred more Dutch (L1) than English (L2) into German (L3). In Rutgers’ (2013) study, students use Dutch as a preferred background language over English in L3 German processing and in the understanding of the structure of the German language. The findings in Rutgers’ study also suggested more L2 English transfer on L3 German in the bilingual stream group than the mainstream group. Future research could compare the findings of the present study to learners in mainstream Dutch education. In summary, considering all our studies conducted in different groups of learners while focusing on the language combinations involved, we found that L3 language transfer from previously acquired languages (L1/L2) is a complex matter and that various factors are involved. Neither transfer based on general typological grouping nor transfer based on structural similarities can fully explain all our results. In earlier work, we found that in the initial stages of L3 processing, learners preferred the L1 as a suitable language for transfer even though Dutch (L1) and French (L3) show no resemblance in general typological grouping and show little structural resemblance. A substantial role for the L2 was found in later (intermediate) stages of acquisition, but only in the language combination L1 Dutch–L2 English–L3 French and only amongst learners enrolled in a bilingual educational programme, i.e. when learners are sufficiently exposed to the L2 in the daily context and when they have (had) enough L2 exposure. In other words, we did find a special L2 status in intermediate learners, but only when the L1 is not a more suitable language for transfer (as is the case in the language combination L1 Dutch–L2 English–L3 German). Therefore, in general, it seems that transfer depends on the combination of the L1-L2 and L3 and that there is no unique answer with regards to the role of background languages in L3A.

6.2 Implications for Third Language Teaching With respect to implications for third language teaching, it is clear that both background languages can serve as transfer languages in intermediate learners in L3A and that in some language combinations this is more so than in others. In this study, we concentrated on negative transfer: to investigate the secondary school students’ behaviour, influence was measured by the number of errors they made in

7 Regarding

the findings of Bohnacker (2006), we have to stress that the differences could also be due to proficiency in L3 German: the participants in Bohnacker’s study were elementary learners of German, while our participants were intermediate learners.

L2 Influence in L3 Acquisition: The Role of the L3

219

the target language. Nonetheless, L2 influence can just as well be positive. For future research, it would be relevant to investigate to what extent secondary school students can also benefit from the fact that they have more languages available for transfer in third language learning. In the Netherlands, where English is more and more ubiquitous as an L2 and where students are more or less bilinguals, a suggestion for third language teaching could be a more multilingual approach in language learning. In the daily (traditional) Dutch school practice of foreign language learning in secondary schools, the foreign languages in school are approached as separate subjects and are always taught from the mother tongue (in the Dutch curriculum as well as in the MYP curriculum). If language students were made aware of resemblances (and differences) between languages, such as with respect to word order constructions, this could lead to advantages in L3 learning.

7 Conclusion The present article examined the extent to which language combinations can affect language transfer from the L2 into the L3. To this end, findings from our previous studies – which suggested a strong supportive role for L2 English in L3 French (when there is enough L2 exposure) – were compared to another L3, viz. L3 German, while other variables were kept as constant as possible (the background languages remained Dutch (L1) and English (L2)) and data were gathered in a comparable context amongst third-year bilingual stream students in the intermediate stages of L3 learning. Influence from the L2 was detected using the V-to-T movement construction found in Dutch, French, and German, but not in English. Our research question was: To what extent does L2 English play a role in L3 German similar to the one it plays in L3 French? We found that L2 English plays a significantly smaller role in intermediate learners in L3 German as compared to its role in L3 French. On the basis of the present data, as well as data from previous studies, we conclude that the supportive role of L2 English depends not only on the amount of L2 exposure the students receive in the daily context and through the years (as we found in Stadt et al. 2016, 2018c), but also on the combination of all three languages involved. Our suggestion as to why the L2 plays a more significant role in L3 French than in L3 German is related to both the typological relatedness and the strong overall structural similarity between L1 Dutch and L3 German. The findings of this study shed new light on how cross-linguistic influence works and how empirical findings should be analysed. In future L3 research, empirical findings should be evaluated while bearing closely in mind the influence of the language combination involved. In other words, it is relevant to emphasise the importance of language combinations with regards to L3 processing, which leads to the recommendation to always include the L3 as a possible variable. Furthermore, in the L3 debate a clear distinction should be made between general typological

220

R. Stadt et al.

groupings and structural resemblances: the combination of the two seem to be of importance with respect to L1/L2 transfer in intermediate stages of L3 learning. The results of this study raise many new questions for future research. The findings are not in line with our predictions or with Falk and Bardel (2011) and Bohnacker (2006), according to which L2 English can play a substantial role in L3 German. In our group of L3 German learners, there is too much L1 influence for the L2 to play a role. Therefore, more studies are needed, including on elementary learners of German, to investigate to what extent the role of L2 English in L3 German changes over the years in L3 learning. A longitudinal study should be undertaken to learn more about language transfer in different stages of L3 German development. Another relevant question concerns the difference between bilingual stream and mainstream students with respect to transfer from L2 English in L3 German. It would be interesting to examine whether in this context, mainstream students transfer even less from English than bilingual stream students, as Rutgers (2013) has suggested.

References Bardel, C., & Falk, Y. (2007). The role of the second language in third language acquisition: The case of Germanic syntax. Second Language Research, 23(4), 459–484. Bardel, C., & Falk, Y. (2012). The L2 status factor and the declarative/procedural distinction. In J. Cabrelli Amaro, S. Flynn, & J. Rothman (Eds.), Third language acquisition in adulthood (pp. 61–78). Amsterdam: John Benjamins. Bohnacker, U. (2006). When Swedes begin to learn German: From V2 to V2. Second Language Research, 22(4), 443–486. Council of Europe. (2001). Common European framework of reference for language learning, teaching and assessment. Cambridge: Cambridge University Press. Falk, Y., & Bardel, C. (2011). Object pronouns in German L3 syntax: Evidence for the L2 status factor. Second Language Research, 27(1), 59–82. Hermas, A. (2010). Language acquisition as computational resetting: Verb movement in L3 initial state. International Journal of Multilingualism, 7(4), 343–362. Hermas, A. (2014a). Multilingual transfer: L1 morphosyntax in L3 English. International Journal of Language Studies, 8(2), 1–24. Hermas, A. (2014b). Restrictive relatives in L3 English: L1 transfer and ultimate attainment convergence. Australian Journal of Linguistics, 34(3), 361–387. Jin, F. (2009). Third language acquisition of Norwegian objects: Interlanguage transfer or L1 influence? In Y.-K. I. Leung (Ed.), Third language acquisition and universal grammar (pp. 144–161). Clevedon: Multilingual Matters. Kellerman, E. (1983). Now you see it, now you don’t. In S. Gass & L. Selinker (Eds.), Language Transfer in Language Learning (pp. 112–134). Rowley, MA: Newbury House. Kellerman, E. (1986). An eye for an eye: Crosslinguistic constraints on the development of the L2 lexicon. In E. Kellerman & M. S. Smith (Eds.), Crosslinguistic Influence in Second Language Acquisition (pp. 35–48). Oxford: Pergamon Press. Mykhaylyk, R., Mitrofanova, N., Rodina, Y., & Westergaard, M. (2015). The linguistic proximity model: The case of verb-second revisited. In E. Grillo & K. Jepson (Eds.), Proceedings of BUCLD 39 (pp. 337–349). Somerville: Cascadilla Press.

L2 Influence in L3 Acquisition: The Role of the L3

221

Na Ranong, S., & Leung, Y.-k. I. (2009). Null objects in L1 Thai-L2 English-L3 Chinese: An empiricist take on a theoretical problem. In Y.-k. I. Leung (Ed.), Third language acquisition and universal grammar (pp. 162–191). Multilingual Matters: Clevedon. Pollock, J.-Y. (1989). Verb movement, universal grammar, and the structure of IP. Linguistic Inquiry, 20(3), 365–424. Rothman, J. (2011). L3 syntactic transfer selectivity and typological determinacy: The typological primacy model. Second Language Research, 27(1), 107–127. Rothman, J. (2013). Cognitive economy, non-redundancy and typological primacy in L3 acquisition: Evidence from initial stages of L3 romance. In S. Baauw, F. Drijkoningen, L. Meroni, & M. Pinto (Eds.), Romance languages and linguistic theory (pp. 217–248). Amsterdam: John Benjamins. Rothman, J. (2015). Linguistic and cognitive motivations for the Typological Primacy Model (TPM) of third language (L3) transfer: Timing of acquisition and proficiency considered. Bilingualism: Language and Cognition, 18(2), 179–190. Rutgers, D. (2013). Multilingualism and metalinguistic development in context: A comparative analysis of metalinguistic mediation in the learning of German as a foreign language by pupils following a Dutch-English bilingual education programme and pupils following a regular programme in the Netherlands. PhD thesis. Cambridge: Clare College. Rutgers, D., & Evans, M. (2015). Bilingual education and L3 learning: Metalinguistic advantage or not? International Journal of Bilingual Education and Bilingualism, 20(7), 788–806. Stadt, R., Hulk, A., & Sleeman, P. (2016). The influence of L2 English and immersion on L3 French in the Netherlands. In J. Audring & S. Lestrade (Eds.), Linguistics in the Netherlands 33 (pp. 152–165). Amsterdam: John Benjamins. Stadt, R., Hulk, A., & Sleeman, P. (2018a). The role of L2 exposure in L3A: A comparative study between third- and fourth-year secondary school students in the Netherlands. In J. Berns, H. Jacobs, & D. Nouveau (Eds.), Romance languages and linguistic theory (Vol. 13, pp. 279–297). Amsterdam: John Benjamins. Stadt, R., Hulk, A., & Sleeman, P. (2018b). The influence of L1 Dutch and L2 English on L3 French: A longitudinal study. In S. Liszka (Ed.), Journal of the European Second Language Association (Vol. 1(1), pp. 63–71). Stadt, R., Hulk, A., & Sleeman, P. (2018c). The influence of L2 English on L3 French acquisition in bilingual education. In T. Mearns & R. de Graaff (Eds.), Dutch Journal of Applied Linguistics 7(2) (pp. 228–250). Amsterdam: John Benjamins. Verspoor, M., de Bot, K., & van der Heiden, F. (2007). Engels in het voortgezet onderwijs, de rol van buitenschools taalcontact. Levende Talen Tijdschrift, 8(3), 3–10. Verspoor, M., de Bot, K., & van Rein, E. (2010). Binnen- en buitenschools taalcontact en het leren van Engels. Levende Talen Tijdschrift, 11(4), 14–33. Westergaard, M., Mitrofanova, N., Mykhaylyk, R., & Rodina, Y. (2016). Crosslinguistic influence in the acquisition of a third language: The linguistic proximity model. International Journal of Bilingualism, 21(6), 666–682.