English Interlanguage Morphology: Irregular Verbs in Young Austrian EL2 Learners―Psycholinguistic Evidence and Implications for the Classroom 3031506162, 9783031506161

Citation preview

English Interlanguage Morphology Irregular Verbs in Young Austrian EL2 Learners— Psycholinguistic Evidence and Implications for the Classroom Thomas Wagner

English Interlanguage Morphology

Thomas Wagner

English Interlanguage Morphology Irregular Verbs in Young Austrian EL2 Learners—Psycholinguistic Evidence and Implications for the Classroom

Thomas Wagner Department of English Pädagogische Hochschule Oberösterreich Linz, Austria

ISBN 978-3-031-50616-1 ISBN 978-3-031-50617-8 (eBook) https://doi.org/10.1007/978-3-031-50617-8 © The Editor(s) (if applicable) and The Author(s), under exclusive licence to Springer Nature Switzerland AG 2023 This work is subject to copyright. All rights are solely and exclusively licensed 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. Cover credit: © Harvey Loake This Palgrave Macmillan imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland Paper in this product is recyclable.

Preface

At the start of the new millennium, a chapter in an edited volume (Plag, 2000) reported a small-scale but pioneering study on the potentially similarity-based organisation of irregular L21 verb morphology in German-speaking learners of English. Only two years later, a special issue of Studies in Second Language Acquisition presented state-of-the-art reviews on the general role of exemplars and their frequency (N. C. Ellis, 2002), grammar teaching (Biber & Reppen, 2002), and the potential of form-focussed instruction for the acquisition of implicit knowledge (R. Ellis, 2002). Eight years later, building on these insights, a monograph on irregular EL2 verb morphology in English and German further developed these ideas and looked into cross-linguistic differences and similarities (Wagner, 2010). The present volume builds on these publications and

1 L2 here refers to any language learned in addition to a person’s first language(s), regardless of whether it is a second, third, fourth, or any subsequent one. Analogous to the still prevalent concept of SLA, L2 is used here because it still is conventionally employed in the pertinent discourse. L2 thus pertains to what elsewhere is labelled L2s, non-native language, EAL/GAL, LX, or language learning and acquisition. There is similar disagreement as to the adequate name for the overarching research field, too, with labels ranging from second and foreign language acquisition, multiple language acquisition, multilingual acquisition, and third language acquisition to additional language acquisition. To this day, none of these terms, however, seem to have fully established itself, probably reflecting some general discord about their scope and limits (De Angelis, 2007). That is why SLA is used in the present volume.

v

vi

PREFACE

expands them in two ways. First, it addresses verb morphology in young EL2 learners, and second, it attempts to outline concrete ramifications for EL2 teaching and learning in a theory-guided, principled fashion. Exploring EL2 verb morphology in young learners is a timely and relevant contribution to psycholinguistics, and grounding the empirical findings in both a general and an SLA-specific learning theory marks a significant attempt at bridging the notorious gap between acquisitional research and its potential ramifications for the classroom. This volume thereby also contributes to R. Ellis’ (2002) proposal of explicit form-focussed teaching facilitating implicit knowledge acquisition. The idea for such an expansion was developed at the 48th annual meeting of BAAL, the British Association for Applied Linguistics, at Aston University, Birmingham, in 2015. It was after a paper presentation there and then that Palgrave Macmillan suggested turning all of this into a monograph. At the time, Palgrave’s new Pivot series, offering outlets for research published at its ‘natural length’, had just been into its third year, and with a digital version as the primary format, this prospect seemed truly exciting. Although data collection and first analyses were finished two years after (many thanks go to Christiane Dalton-Puffer, Monika Boniecki and Ulrike Podar, University of Vienna, for their invaluable help), the project had then been dormant for quite a while. Thanks to Palgrave’s executive linguistics editor Cathy Scott as well as her colleague Bhavya Rattan—but interrupted by COVID-19—the project came back into full swing about three years ago. I am indebted to Palgrave’s editors, whose tremendous help in shaping initial ideas as well as finalising the book’s structure eventually made this all possible. I am also indebted to the many young EL2 learners who courageously took part in the experiments and thus contributed stimulating and, at times, funny data. I very much appreciated comments on earlier versions of this text from Elizabeth Erling, Erwin Gierlinger, Gudrun Keplinger, and Harald Spann, as well as from the commissioned reviewers, whose critical comments and suggestions helped me strengthen my arguments and improve the manuscript. Special thanks also go to three dedicated students of mine, Elisabeth Frank, Hannah Schneeberger, and Jasmin Stadler, who helped with proofreading and the literature. And, of course, I extend the most heartfelt thanks to my family for their unconditional support. As usual, all remaining errors are mine. One objective of this BAAL conference in 2015, where it all started, was to engage the research community in creating new ways of imagining,

PREFACE

vii

theorising, and practising applied linguistics. I hope that, many years later, the present volume will serve as a contribution to this endeavour and merit further psycholinguistic as well as pedagogical work. This volume is dedicated to all language learners. Linz, Austria October 2023

Thomas Wagner

References Biber, D., & Reppen, R. (2002). What does frequency have to do with grammar teaching? Studies in Second Language Acquisition, 24(2), 199–208. https:// doi.org/10.1017/S0272263102002024 De Angelis, G. (2007). Third or additional language acquisition. Multilingual Matters. Ellis, N. C. (2022). Second language learning of morphology. Journal of the European Second Language Association, 6(1), 34–59. https://doi.org/10. 22599/jesla.85 Ellis, R. (2002). Does form-focussed instruction affect the acquisition of implicit knowledge? A review of the research. Studies in Second Language Acquisition, 24(2), 223–236. https://doi.org/10.1017/S0272263102002024 Plag, I. (2000). Irregular past tense formation in English interlanguage. In I. Plag & K. P. Schneider (Eds.), Language use, language acquisition, and language history. (Mostly) empirical studies in honour of Rüdiger Zimmermann (pp. 134–149). Wissenschaftlicher Verlag. Wagner, T. (2010). Interlanguage morphology. Irregular verbs in the mental lexicon of German-English interlanguage speakers. Narr-Francke-Attempto.

About This Book

This book combines an in-depth examination of L2 verbal morphology with a comprehensive discussion of its relevance for instructed EL2 teaching and learning. The first part of the book presents behavioural evidence for a similarity-based organisation of English irregular verbs in the mental lexicon of young Austrian EL2 learners. Based on two wellestablished experiments and comprehensive data analyses, a remarkable sensitivity of young Austrian learners to the morphophonological makeup of English irregular verbs can be illustrated. This psycholinguistic evidence is then discussed against potential accounts of L1 and L2 verbal morphology in the pertinent literature. In the second part of the book, ramifications for instructed EL2 teaching and learning are discussed, with reference to Processing Instruction and Variation Theory. This book thus tries to bridge the notorious gap between SLA research and EL2 classroom applications. It will therefore appeal to postgraduate students in Applied Linguistics and EL2, advanced level undergraduates as well as researchers, but also to teacher trainers across various disciplines.

Data Accessibility Statement Materials, data, and statistical analyses are available at https://osf.io/ 24n35/.

ix

Contents

1 8

1

Introduction References

2

Irregular Verb Morphology: Theoretical Accounts 2.1 English Irregular Verbs 2.2 German Irregular Verbs References

15 15 20 23

3

Irregular Verb Morphology: L1 and L2 Cognitive Accounts 3.1 Dual-Route Models of Irregular Verb Morphology 3.2 Single-Route Models of Irregular Verb Morphology References

31 32 36 45

4

The 4.1 4.2 4.3

Empirical Study General Method and Design Creation of Test Items Experiment 1 Materials and Procedure Participants Data Coding and Analysis Results Discussion Experiment 1 4.4 Experiment 2 Materials and Procedure

59 59 60 61 61 63 64 70 81 83 83 xi

xii

CONTENTS

Participants Data Coding and Analysis Results Discussion Experiment 2 4.5 General Discussion References

84 87 88 96 99 104

5

Implications for the Classroom 5.1 English Irregular Verbs and Variation Theory 5.2 English Irregular Verbs and Processing Instruction 5.3 A Roadmap Towards Concrete Ideas and Applications References

113 117 121 124 132

6

Conclusions and Limitations References

141 148

Index

151

About the Author

Thomas Wagner is a professor of Applied Linguistics and EFL at the University College of Education Upper Austria. Prior to his academic career, he had been a full-time teacher at a secondary school for ten years, where he taught English across many proficiency levels, and therefore became intimately familiar with the many facets as well as limitations of teaching English as a foreign language in instructed contexts. His current research interests, including grammar acquisition, foreign language aptitude as well as Variation Theory, are co-determined by his personal history as a language learner, language teacher, teacher educator, applied linguist, and father of two children who bring home, quite plainly, the seemingly never-ending challenges that come with instructed EL2 today.

xiii

List of Figures

Fig. 1.1 Fig. 2.1 Fig. 4.1 Fig. 4.2

Fig. 4.3 Fig. 4.4 Fig. 4.5 Fig. 4.6 Fig. 4.7 Fig. 4.8

Fig. 4.9

A rule-based mechanism producing default past tense inflection Structured lexical entry inheritance tree for the German verb werfen ‘to throw’ Sum-contrast-coding effects for Onset and Coda on the binary dependent variable Stacked barplots for random intercept variability in reply_type_binary by test item (left panel) and participants (right panel) Mosaicplots for type of inflection including repetition (left panels) and binary (right panels) by onset and coda Effect plot from the final generalised mixed model for the predictors Onset and Coda by L1 and proficiency Conditional inference tree for type of inflection by onset, coda, L1, and proficiency Mosaicplots for the influence of onset, coda, L1, and proficiency on the type of vowel change Conditional inference tree for vowel change patterns by onset, coda, L1, and proficiency Mosaicplots for type of inflection by all 16 items (left panel) and the items with prototypical segments (right panel) Mosaicplots for type of inflection by test item, onset, nucleus, coda, L1, and proficiency

4 22 66

69 73 76 78 79 80

89 90

xv

xvi

LIST OF FIGURES

Fig. 4.10

Fig. 4.11

Fig. 5.1 Fig. 5.2

Conditional inference trees for inflectional patterns by L1 and proficiency for given infinitives (left panel) and given participles (right panel) Morphophonological properties of an exemplary L2 lexical entry for the English verb sing in L1 German learners of English The place of PI in the language acquisition processes General roadmap for the application of Variation Theory and Input Processing in explicit instruction

94

102 122 131

List of Tables

Table 2.1 Table 2.2 Table 4.1 Table 4.2 Table 4.3 Table 4.4 Table 4.5 Table 4.6 Table 4.7 Table 4.8 Table 4.9 Table 4.10

Table 4.11

English apophony patterns from an input- and output perspective Common apophonic patterns in Modern High German Test items from experiment 1 with syllable structure and associated schema Crosstable of schools by proficiency level Crosstable for filler item and type of inflection as well as percentage of correctly inflected forms Percentages for the types of inflections provided by the participants Percentages for regular and irregular responses to the eight test items Percentages for regular, irregular, and repetition responses to the eight test items Percentages for types of vowel change responses to the eight test items Coefficients table from the generalised linear mixed model in experiment 1 Crosstable for the four permutations of experiment 1 and 2 with number of pupils per group Test items from experiment 2 with parts of speech, vowel change patterns, and hypothesised analogical inflections Coefficients table from the generalised linear mixed model in experiment 2

18 22 62 63 65 71 71 71 72 75 84

85 92

xvii

xviii

LIST OF TABLES

Table 4.12 Table 4.13 Table 5.1

Resulting vowel change patterns by response types in experiment 2 Resulting apophonic patterns by Expected apophonic patterns in experiment 2 Variational patterns for the teaching of irregular verb morphology

95 97 127

CHAPTER 1

Introduction

Abstract This chapter introduces the reader to a long-standing debate about the cognitive status of verb inflection. Numerous psycholinguistic studies have shown that, under experimental conditions, speakers and learners can extend irregular inflectional patterns to new words, an overgeneralisation strategy also observable in spontaneous L1 and L2 language production. On the one hand, such research gave rise to competing cognitive models, also known as the past tense debate, on the other, it has ramifications for EL2 teaching and learning of inflectional paradigms. The introduction outlines how, building on previous studies, the present volume extends existing EL2 empirical research to young learners at the European waystage level A2, and how the present empirical findings can be harvested for a principled approach to explicit grammar instruction. Keywords Irregular verb morphology · Mental lexicon · Past tense debate · Symbolic rule processing · Similarity-based processing · Analogy

The bane of every language student (Neubauer & Clahsen, 2009; Pinker, 1999) is what English irregular verbs have often been called in the context of teaching and learning. Although many of them are high-frequent tokens, denoting central human activities, their seemingly unpredictable © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 T. Wagner, English Interlanguage Morphology, https://doi.org/10.1007/978-3-031-50617-8_1

1

2

T. WAGNER

idiosyncratic inflections make them hard to acquire, the argument goes, demanding to be rote-learned in an arduous fashion (DeKeyser, 2005; Ellis, 2006). Psycholinguistic research, however, has been repeatedly suggesting that the organisation of irregular verb inflection, both in L1 and L2 English and German, is far from unpredictable and idiosyncratic (Bybee & Moder, 1983; Bybee & Slobin, 1982; Pinker & Prince, 1988; Plag, 2000; Wagner, 2010, 2017). And evidence of this nature would have substantial ramifications for the teaching and learning of irregular inflectional paradigms, an area which appears to be underresearched to this day. The present volume fills this gap in two significant ways. First, building on established psycholinguistic experimental designs, it extends prior findings to the area of young learners, and contextualises these within competing models of verbal inflection. Second, it succinctly relates the present empirical findings to one general and one SLA-specific learning theory. Such a principled synthesis forms a sound basis for further didactical deliberations around explicit instruction. While there is plenty of evidence around the general role of such instruction in EL2 (Benati, 2019; R. Ellis, 2002, 2005, 2006), the present synthesis helps clarify this role, integrate it into modern learner-centred pedagogy, and thus make it relevant for practitioners in the field. How, though, has psycholinguistic research managed to gain insights into cognitive storage and retrieval mechanisms around irregular verb morphology? It all probably started more than six decades ago with Berko’s (1958) pioneering experiment. In her study, four- to seven-yearold English-speaking children were prompted by means of picture cards to pluralise non-existing nouns (nonce-nouns), such as wug, and inflect nonce-verbs like bling 1 for past tense. Interestingly, when prompted to do so, the children willingly produced replies. Why did they not protest? After all, how could they inflect words they had never encountered before? Not only did they comply with this seemingly egregious task, but almost all the children seemed to know what to do. The study reports that 91% of the children produced grammatical plural inflections, like wugs, or other phonologically conditioned allomorph variants, and the great 1 Apparently, bling, and also bling-bling, are now words of English, listed, for instance, in the Oxford English Dictionary (OED, Oxford University Press, n.d.) or Cambridge Online Dictionary (Cambridge University Press, n.d.). The OED first listed bling in its New Word List from June 2012.

1

INTRODUCTION

3

majority of them created well-formed regular past tenses (blinged), too; in fact, quite a few children at the time were under the impression that they had just been taught new English words. Obviously, when inflecting nonexistent test material, children did not just regurgitate rote-learned and seemingly disconnected forms in a behaviourist manner (Skinner, 1957) but generated novel ones on demand in the course of speech production. In Berko’s data, however, and in many more ensuing psycholinguistic experiments (cf. Albright & Hayes, 2003, 2006; Bybee & Moder, 1983; Bybee & Slobin, 1982; Cuskley et al., 2015; Prasada & Pinker, 1993; Wagner, 2010, 2017), a different kind of generalisation capacity could also be observed. For one, some of Berko’s participants failed to extend the -ed allomorph from existing verbs like melt – melted to novel verbs in a consistent manner. Apparently, they had taken the voiceless alveolar suffix in the verb’s coda as a marker of past tense and perhaps concluded there was no need to provide more inflection. Even more intriguing, though, participants also produced non-default, irregular inflections for certain nonce-verbs, too. Test items such as gling or spling, for instance, would trigger glang, splang, or splung as responses. Both children and adult participants had thus demonstrated a sensitivity to the nonce-verbs’ morphophonology, using it to systematically build new past tense forms by analogy to existing verbs and their inflectional classes. They simply had taken real patterns as blue-prints and extended them to new input. For many test items, such responses were quite frequent, so they probably were generated with considerable ease and in a clearly non-random fashion. As later behavioural experiments demonstrated, such analogical extensions can operate both ways, from base form to inflection, but also in reverse, from inflected to base form. In Wagner (2010), for example, both speakers and learners of English created infinitives to given past tense participle nonces. When prompted to deal with test items such as skrung or sprut, a lot of their replies included skring or sprit. And in Wagner (2017), English-speaking learners of German, when prompted to inflect given German preterite forms for infinitive and participle, produced predominantly consistent and attested apophonic patterns. Both experiments suggest that extrapolating from the morphology of an inflection to a matching base form is something speakers and learners can obviously do.

4

T. WAGNER

Fig. 1.1 A rule-based mechanism producing default past tense inflection

Vpast V

suffiix

cook

-ed

Experimentally triggered extensions of rule-like inflections as well as overgeneralisations of past tense formations (go – *goed) were commensurate with the idea of a deterministic mechanism in a speaker’s language faculty, operating productively on symbolic representations of words. Symbolic here means that words are stored in the mental lexicon as abstract representations of morphosyntactic categories, like nouns or verbs, unconditionally accepted by the default mechanism as input for further computation, regardless of their morphophonological make-up. Rule-like means that this mechanism would predictably produce default output forms by means of concatenative processes, without ever requiring access to any lexical information. Figure 1.1 (cf. Pinker, 1999, p. 16) illustrates such a process. It was the mechanism’s symbolic and deterministic nature in particular, generative linguists would argue at the time (Chomsky & Halle, 1968; see also Pinker, 1999; Pinker & Ullman, 2002), that rendered its generalisation capacity virtually unbounded, affording “unlimited productivity” (Prasada & Pinker, 1993, p. 2). For it to work, it would take only two ingredients, namely words and rules (Pinker & Ullman, 2002). Such a rule-based generative mechanism would also explain typical overgeneralisation phenomena found in acquisitional processes (U-shaped learning) in both English and German. English children, for instance, are reported to produce forms such as *tooths or *goed in spontaneous speech (Marcus et al., 1992). Overirregularised past tense inflections such as they drunk or they sung are frequently attested in adult speech production, too (Anderwald, 2011). And German children frequently overgeneralise the default past tense rule in spontaneous participle production, like in *ge-geh-t ‘*goed’, instead of ge-gang-en ‘went’, *ge-schlaf-t ‘*sleeped’, instead of ge-schlaf-en ‘slept’, *ge-brech-t ‘*breaked’, instead of ge-broch-en ‘broke’ (Eisenbeiss, 2005; Marchman et al., 1997; Marcus et al., 1995; Plunkett & Marchman, 1991, 1993; Taatgen & Anderson, 2002). Very rarely,

1

INTRODUCTION

5

albeit attested in learner corpora (CHILDES, MacWhinney, 2000), do learners of German overirregularise, too, like in *geschlacht-en ‘slaughtered’, instead of geschlachte-t, where the non-default -en participle suffix was erroneously attached instead of the -t (cf. Fleischhauer, 2013). Viewing inflection as symbolic and mostly rule-governed comes with two theoretically motivated caveats, though. First, such a mechanism, like a pocket calculator, has no memory of its operations (Baayen, 2007). Whenever it processes irregular verbs, it does so on demand, but it does not learn from experience. Due to its symbolic nature, it actually does not need to. Storage, in this approach, is a mere idiosyncrasy (Baayen, 2007), with exceptions being confined to a small number of lexicalised forms, relegated to some sort of mental storage facility. If verbs, for instance, already have their tailor-made inflected forms, the rule-based mechanism would simply be blocked (Pinker, 1999), and stored exemplars would come to the rescue, processing these exceptions. This is probably why children seem to frequently overapply regular rather than irregular inflection in spontaneous speech (Xu & Pinker, 1995). Second, this mechanism is bound by directionality; unlike a parser, it can go from given input to output only, never backwards. In contrast, overirregularisations, like in Berko’s (1958) or Wagner’s (2010, 2017) data are difficult to account for by symbolic operations. Their frequency and systematicity did not suggest random inconsistencies or performance errors, so they were probably more than idiosyncratic deviations from the rule. Instead, they appeared to be driven by some analogical pattern association based on a verb’s (and paradigm’s) morphophonological properties, co-determined by token frequency (Bybee, 2001; Cuskley et al., 2015; Glass & Lau, 2003). One of the first computational implementations of such pattern association probably came from Rumelhart and McClelland (1986), who successfully taught a computer to form past tense inflections, using a rather simple so-called connectionist network. Not only did this computer model learn to correctly produce baked from bake, but it also mirrored the characteristic U-shaped overgeneralisation errors observable in young English-speaking children. Only two years later, though, Rumelhart and McClelland’s connectionist approach was seriously challenged by Pinker and Prince (1988) as well as Lachter and Bever (1988), who dismissed connectionist modelling as ill-conceived, sparking “a veritable brouhaha” (Kirov & Cotterell, 2018, p. 651). From then on, under the umbrella term past tense debate (Kirov & Cotterell, 2018; McClelland &

6

T. WAGNER

Patterson, 2002; N. C. Ellis, 2022; Pinker, 1999), numerous scientists devoted themselves to the exploration of English verb morphology, rendering it a petri-dish like Litmus test (Fodor & Pylyshyn, 1988), a new fruit fly of cognitive science (Ambridge, 2020; Pinker, 1999), and a long-awaited window into cognition. English verb morphology seemed particularly suited since it combines rule-like processes (walk – walked) with irregular but patterned, paradigmatically related exceptions (sing – sang – sung, go – went – gone, bring – brought – brought, hit – hit – hit, etc., Aronoff, 1994; Stump, 2001; Wunderlich & Fabri, 1995; for opposing views see Lieber, 1992). While in the beginning this past tense debate was limited to the area of the L1 (Bybee & Moder, 1983; Bybee & Slobin, 1982; Prasada & Pinker, 1993), it soon started bringing forth intriguing evidence from a second language acquisition point of view (Cuskley et al., 2015; Wagner, 2010, 2017). On the one hand, such learner data are crucial evidence for the debate around underlying psycholinguistic processes and human cognition in general. Thus, a lot of effort has gone into accounting for such inflectional behaviour, through rule-based, associationist, and hybrid cognitive models, albeit with conflicting findings (Beck, 1997; Birdsong & Flege, 2001; Broeder & Plunkett, 1997; Chandler, 1994; Colombo et al., 2006; Gor & Chernigovskaya, 2003; Hahne et al., 2006; Indefrey, 2006; Kırkıcı, 2007; Nakisa et al., 2000; Pinker & Ullman, 2002; Seidenberg & MacDonald, 1999; Sokolik & Smith, 1992). While most researchers would probably agree on human language involving some sort of associative storage device for the words’ sounds and meanings, there is still disagreement as to whether human language would also need symbolic processing. The present volume, however, is not concerned with supremacy issues around single- or dual-route approaches to language processing, be it rule-based, associative, or hybrid. Following N. C. Ellis (2002, 2022) and R. Ellis (2002), a role for associative learning is taken for granted, no matter if it is part of a single- or dual-route philosophy. Neither is the present volume concerned with the exact nature of associative language processing. The empirical findings presented in this volume would be consistent with exemplar-based models, relying on storage and similarity algorithms, prototypical models, abstracting over exemplars with statistics of central tendencies, or connectionist models, representing distributed information in a network of dynamic connections. All of these models are sensitive to patterns and their frequencies and thus reflect usage-based approaches to language learning.

1

INTRODUCTION

7

On the other hand, learner data could be utilised for the language classroom. The better we understand mechanisms governing storage and retrieval of irregular linguistic paradigms in learners’ mental lexicons, the more professional practitioners’ teaching strategies, materials, and handling of errors will be (Boers, 2021). Picking up on this potential for EL2 teaching and learning, this volume seeks to utilise empirical findings for practical applications in the EL2 classroom. Therefore, there are two research questions guiding this study. First, this research seeks to address how far irregular verb morphology in learners of English at the European waystage level A2 (Council of Europe, 2001) is governed by similarity-based and potentially prototype-driven pattern association. This research question will be followed through quantitative measurements in two empirical experiments, examining potential effects of the phonological segments onset and coda on the learners’ inflectional behaviour. Such segmental effects would prove that even young learners are sensitive to the verb’s—and inflectional paradigm’s—morphophonology and that it is indeed this morphophonology governing inflection, and not merely the lemma as a whole. Second, it is asked how two established teaching and learning theories, Input Processing and Variation Theory, could help utilise morphophonological generalisations for explicit instruction in the language classroom. This is an exploratory research question, and it is pursued hermeneutically and discursively by elaborating on the potential of these two theories with regard to the empirical findings. This is supposed to help gain a theorydriven, principled, and in-depth understanding of concrete implications for the classroom and make predictions about an appropriate approach to teaching irregular verbs in instructed EL2 settings. The remainder of this volume is structured as follows. Chapter 2 introduces the grammar of the phenomenon in question by summarising theoretical linguistic accounts of irregular verb morphology in both English and German. Starting out with classical approaches, like itemand-arrangement, (cf. Hockett, 1954), more recent theories such as natural morphology (Bertacca, 2010; Dressler, 1985, 1986) or minimalist morphology (Wunderlich, 1996; Wunderlich & Fabri, 1995) will be reviewed. Chapter 3 discusses different cognitive accounts of irregular verb morphology, both in terms of L1 and L2 processing. Chapter 4 contains the two empirical studies, their methodology, results, and a general discussion, synthesising current and previous findings along

8

T. WAGNER

with their practical implications for the classroom. This is supposed to segue into Chapter 5, which is concerned with the two learning theories that could inform a principled approach to teaching irregular verb morphology. Finally, Chapter 6 will present conclusions, discuss potential limitations, and give an outlook into avenues for further research. The data, the materials for the empirical experiments as well as the R-scripts from the statistical analyses can be found at https://osf.io/ 24n35/.

References Albright, A., & Hayes, B. (2003). Rules vs. analogy in English past tenses: A computational/experimental study. Cognition, 90(2), 119–161. https://doi. org/10.1016/S0010-0277(03)00146-X Albright, A., & Hayes, B. (2006). Modelling productivity with the gradual learning algorithm. The problem of accidentally exceptionless generalisations. In G. Fanselow, C. Féry, M. Schlesewsky, & R. Vogel (Eds.), Gradience in grammar (pp. 185–204). Oxford University Press. Ambridge, B. (2020). Against stored abstractions. A radical exemplar model of language acquisition. First Language. Special Issue, 40(5–6), 509–559. https://doi.org/10.1177/0142723719869731 Anderwald, L. (2011). Norm vs variation in British English irregular verbs: The case of past tense sang vs. sung. English Language and Linguistics, 15(1), 85–112. https://doi.org/10.1017/S1360674310000298 Aronoff, M. (1994). Morphology by itself: Stems and inflectional classes. The MIT Press. Baayen, R. H. (2007). Storage and computation in the mental lexicon. In G. Jarema & G. Libben (Eds.), The mental lexicon: Core perspectives (pp. 81– 104). Elsevier. Beck, M. L. (1997). Regular verbs, past tense and frequency. Tracking down a potential source of NS/NNS competence differences. Second Language Research, 13(2), 93–115. Benati, A. (2019). The role of instruction: An update. Language Teaching Research Quarterly, 11, 11–19. https://doi.org/10.32038/ltrq.2019.11.02 Berko, J. (1958). The child’s learning of English morphology. Word, 14, 150– 177. https://doi.org/10.1080/00437956.1958.11659661 Bertacca, A. (2010). Present-day English irregular verbs revisited. Poznan Studies in Contemporary Linguistics, 46(2), 127–154. Birdsong, D., & Flege, J. E. (2001). Regular-irregular dissociations in L2 acquisition of English morphology. In A. H. J. Do, L. Domínguez, & A.

1

INTRODUCTION

9

Johansen (Eds.), BUCLD 25. Proceedings of the 25th annual Boston University Conference on Language Development (Vol. 1, pp. 123–132). Cascadilla Press. Boers, F. (2021). Evaluating second language vocabulary and grammar instruction. A synthesis of the research on teaching words, phrases, and patterns. Routledge. Broeder, P., & Plunkett, K. (1997). Connectionism and second language acquisition. In N. C. Ellis (Ed.), Implicit and explicit learning of language (pp. 421–454). Harcourt Brace Jovanovich. Bybee, J. (2001). Phonology and language use. Cambridge University Press. Bybee, J., & Moder, C. L. (1983). Morphological classes as natural categories. Language, 59(2), 251–270. https://doi.org/10.2307/413574 Bybee, J., & Slobin, D. I. (1982). Rules and schemas in the development and use of the English past tense. Language, 58(2), 265–289. https://doi.org/ 10.2307/414099 Cambridge University Press. (n.d.). bling. In Cambridge Online Dictionary. Retrieved July 1, 2023, from https://dictionary.cambridge.org/dictionary/ english/bling Chandler, S. (1994, April). An exemplar-based approach to language acquisition. Paper presented at the Workshop on Cognitive Models of Language Acquisition, University of Tilburg, The Netherlands. Chomsky, N., & Halle, M. (1968). The sound pattern of English. Harper & Row. Colombo, L., Stoianov, I., Pasini, M., & Zorzi, M. (2006). The role of phonology in the inflection of Italian verbs. A connectionist investigation. The Mental Lexicon, 1(1), 147–181. https://doi.org/10.1075/ml.1.1.09col Council of Europe. (2001). Common European Framework of Reference for Languages: Learning, teaching, assessment. Cambridge University Press. Cuskley, C., Colaiori, F., Castellano, C., Loreto, V., Pugliese, M., & Tria, F. (2015). The adoption of linguistic rules in native and non-native speakers: Evidence from a Wug task. Journal of Memory and Language, 84, 205–223. https://doi.org/10.1016/j.jml.2015.06.005 DeKeyser, R. (2005). What makes learning second-language grammar difficult? A review of issues. Language Learning, 55(S1), 1–25. https://doi.org/10. 1111/j.0023-8333.2005.00294.x Dressler, W. U. (1985). On the predictiveness of natural morphology. Journal of Linguistics, 21(2), 321–337. https://doi.org/10.1017/S00222267000 1029X Dressler, W. U. (1986). Explanation in natural morphology, illustrated with comparative and agent-noun formation. Linguistics, 24, 519–548. https:// doi.org/10.1515/ling.1986.24.3.519

10

T. WAGNER

Eisenbeiss, S. (2005). Merkmalsgesteuerter Grammatikerwerb: Eine Untersuchung zum Erwerb der Struktur und Flexion von Nominalphrasen (Doctoral dissertation). Heinrich Heine University. https://docserv.uni-duesseldorf.de/ser vlets/DerivateServlet/Derivate-3185/1185.pdf Ellis, N. C. (2002). Frequency effects in language processing. A review with implications for theories of implicit and explicit language acquisition. Studies in Second Language Acquisition, 24(2), 143–188. https://doi.org/10.1017/ S0272263102002024 Ellis, N. C. (2022). Second language learning of morphology. Journal of the European Second Language Association, 6(1), 34–59. https://doi.org/10. 22599/jesla.85 Ellis, R. (2002). Does form-focussed instruction affect the acquisition of implicit knowledge? A review of the research. Studies in Second Language Acquisition, 24(2), 223–236. https://doi.org/10.1017/S0272263102002024 Ellis, R. (2005). Principles of instructed language learning. System, 33(2), 209– 224. https://doi.org/10.1016/j.system.2004.12.006 Ellis, R. (2006). Current issues in the teaching of grammar: An SLA perspective. TESOL Quarterly, 40(1), 83–107. https://doi.org/10.2307/40264512 Fleischhauer, E. (2013). Morphological processing in children. An experimental study of German past participles (Doctoral dissertation). Faculty of Human Sciences, University of Potsdam. Publication portal of University of Potsdam. https://publishup.uni-potsdam.de/opus4-ubp/frontdoor/ index/index/year/2014/docId/6807 Fodor, J. A., & Pylyshyn, Z. W. (1988). Connectionism and cognitive architecture: A critical analysis. Cognition, 28(1–2), 3–71. https://doi.org/10.1016/ 0010-0277(88)90031-5 Glass, A. L., & Lau, J. (2003). Grammatical intuitions about irregular verb inflections. The American Journal of Psychology, 116(1), 51–70. https://doi.org/ 10.2307/1423335 Gor, K., & Chernigovskaya, T. (2003). Mental lexicon structure in L1 and L2 acquisition: Russian evidence. Glossos, 4, 1–31. Hahne, A., Müller, J., & Clahsen, H. (2006). Morphological processing in a second language: Behavioural and event-related potential evidence for storage and decomposition. Journal of Cognitive Neuroscience, 18(1), 121–134. https://doi.org/10.1162/089892906775250067 Hockett, C. F. (1954). Two models of grammatical description. Word, 10(2–3), 210–234. https://doi.org/10.1080/00437956.1954.11659524 Indefrey, P. (2006). It is time to work toward explicit processing models for native and second language speakers. Applied Psycholinguistics, 27 (1), 66–69. https://doi.org/10.1017/S0142716406060103 Kırkıcı, B. (2007). The mental processing of L2 English lexical compounds: A developmental dual-mechanism account. In L. Roberts, A. Gürel, S. Tatar, &

1

INTRODUCTION

11

L. Marti (Eds.), Eurosla-Yearbook, 7 (pp. 7–25). John Benjamins. https:// doi.org/10.1075/eurosla.7.03kir Kirov, C., & Cotterell, R. (2018). Recurrent neural networks in linguistic theory: Revisiting Pinker and Prince (1988) and the past tense debate. Transactions of the Associations for Computational Linguistics, 6, 651–665. https://doi.org/ 10.1162/tacl_a_00247 Lachter, J., & Bever, T. (1988). The relation between linguistic structure and associative theories of language learning: A constructive critique of some connectionist learning models. Cognition, 28(1–2), 195–247. https://doi. org/10.1016/0010-0277(88)90033-9 Lieber, R. (1992). Deconstructing morphology: Word formation in syntactic theory. University of Chicago Press. MacWhinney, B. (2000). The CHILDES project: Tools for analyzing talk, volume 1. Transcription format and programs (3rd ed.). Lawrence Erlbaum Associates. Marchman, V. A., Plunkett, K., & Goodman, J. (1997). Overregularization in English plural and past tense inflectional morphology. A response to Marcus (1995). Journal of Child Language, 24(3), 767–779. https://doi.org/10. 1017/S0305000997003206 Marcus, G. F., Brinkmann, U., Clahsen, H., Wiese, R., & Pinker, S. (1995). German inflection: The exception that proves the rule. Cognitive Psychology, 29(3), 189–256. Marcus, G. F., Pinker, S., Ullman, M. T., Hollander, M., Rosen, T. R., Xu, F., & Clahsen, H. (1992). Overregularization in language acquisition. Monographs of the Society for Research in Child Development, 57 (4), 1–182. https://doi. org/10.2307/1166115 McClelland, J. L., & Patterson, K. (2002). Rules or connections in past-tense inflections: What does the evidence rule out? Trends in Cognitive Sciences, 6(11), 465–472. https://doi.org/10.1016/S1364-6613(02)01993-9 Nakisa, R. C., Plunkett, K., & Hahn, U. (2000). Single- and dual-route models of inflectional morphology. In P. Broeder & J. Murre (Eds.), Models of language acquisition (pp. 201–224). Oxford University Press. Neubauer, K., & Clahsen, H. (2009). Decomposition of inflected words in a second language: An experimental study of German participles. Studies in Second Language Acquisition, 31(3), 403–435. https://doi.org/10.1017/ S0272263109090354 Oxford University Press. (n.d.). bling. In Oxford English Dictionary (OED). Retrieved July 1, 2023, from https://www.oed.com/ Pinker, S. (1999). Words and rules. The ingredients of language. Phoenix. Pinker. S., & Prince, A. (1988). On language and connectionism. Analysis of parallel distributed processing of language acquisition. In S. Pinker & J. Mehler (Eds.), Connections and symbols (pp. 73–193). MIT Press.

12

T. WAGNER

Pinker, S., & Ullman, M. T. (2002). Combination and structure, not gradedness, is the issue. Trends in Cognitive Sciences, 6, 472–474. https://doi.org/10. 1016/S1364-6613(02)02013-2 Plag, I. (2000). Irregular past tense formation in English interlanguage. In I. Plag & K. P. Schneider (Eds.), Language use, language acquisition, and language history. (Mostly) Empirical studies in honour of Rüdiger Zimmermann (pp. 134–149). Wissenschaftlicher Verlag. Plunkett, K., & Marchman, V. (1991). U-shaped learning and frequency effects in a multi-layered perceptron. Implications for child language acquisition. Cognition, 38(1), 43–102. https://doi.org/10.1016/0010-027 7(91)90022-v Plunkett, K., & Marchman, V. (1993). From rote learning to system building. Acquiring verb morphology in children and connectionist nets. Cognition, 48(1), 21–69. https://doi.org/10.1016/0010-0277(93)90057-3 Prasada, S., & Pinker, S. (1993). Generalization of regular and irregular morphological patterns. Language and Cognitive Processes, 8, 1–56. https://doi.org/ 10.1080/0169096930840694 Rumelhart, D. E., & McClelland, J. L. (1986). On learning the past tenses of English verbs. In D. E. Rumelhart, J. L. McClelland, & the PDP Research Group (Eds.), Parallel distributed processing: Explorations in the microstructures of cognition (Vol. 2, pp. 216–271). MIT Press. Seidenberg, M. S., & MacDonald, M. C. (1999). A probabilistic constraints approach to language acquisition and processing. Cognitive Science, 23(4), 569–588. https://doi.org/10.1207/s15516709cog2304_8 Skinner, B. F. (1957). Verbal behavior. Appleton-Century-Crofts. Sokolik, M. E., & Smith, M. E. (1992). Assignment of gender to French nouns in primary and secondary language: A connectionist model. Second Language Research, 8(1), 39–58. https://doi.org/10.1177/02676583920080 Stump, G. T. (2001). Inflectional morphology. A theory of paradigm structure. Cambridge University Press. Taatgen, N. A., & Anderson, J. R. (2002). Why do children learn to say “broke”? A model of learning the past tense without feedback. Cognition, 86(2), 123– 155. https://doi.org/10.1016/S0010-0277(02)00176-2 Wagner, T. (2010). Interlanguage morphology. Irregular verbs in the mental lexicon of German-English interlanguage speakers. Narr-Francke-Attempto. Wagner, T. (2017). L2 irregular verb morphology: Exploring behavioral data from intermediate English learners of German as a foreign language using generalized mixed effects models. Studies in Second Language Learning and Teaching, 7 (3), 535–556. https://doi.org/10.14746/ssllt.2017.7.3.9 Wunderlich, D. (1996). Minimalist morphology: The role of paradigms. In G. Booij & J. van Marle (Eds.), Yearbook of morphology 1995 (pp. 93–114). Kluwer Academic.

1

INTRODUCTION

13

Wunderlich, D., & Fabri, R. (1995). Minimalist morphology: An approach to inflection. Zeitschrift für Sprachwissenschaft, 14(2), 236–294. https://doi. org/10.1515/zfsw.1995.14.2.236 Xu, F., & Pinker, S. (1995). Weird past tense forms. Journal of Child Language, 22(3), 531–556. https://doi.org/10.1017/S0305000900009946

CHAPTER 2

Irregular Verb Morphology: Theoretical Accounts

Abstract This chapter discusses various theoretical approaches to both English and German irregular verb morphology. It outlines commonalities and differences between these approaches’ attempts to capture patterns and paradigmatic relationships. Section 2.1 is on English irregular verb morphology and briefly summarises early theoretical accounts, followed by a comprehensive analysis of the verbal paradigm from a current perspective. This analysis focusses on inflectional patterns and classes as well as challenging exceptions and idiosyncrasies. Section 2.2 looks at German irregular verb morphology, providing an in-depth analysis of this verbal paradigm. The chapter closes with an illustration of a minimalist morphology example. Keywords Morphology · Inflection · Apophony · Paradigm · Past tense · Minimalist morphology

2.1

English Irregular Verbs

English irregular verbs, a classic chestnut of morphology (Anderson, 1988), have not only challenged linguistic and psycholinguistic theories and models but also brought about terminological confusion and inconsistencies in the field’s jargon. Some studies, for instance, have called © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 T. Wagner, English Interlanguage Morphology, https://doi.org/10.1007/978-3-031-50617-8_2

15

16

T. WAGNER

English verbs like bake to be regular and default-like (Pinker, 1999), while verbs like sing or drive have been referred to as strong verbs (Paul, 1917), irregular (Augst, 1975; Hock, 1968; Wurzel, 1970), or non-default-like (Smolka et al., 2018). This distinction is problematic. What is termed strong verbs today often comprises historically weak verbs, which combine vowel alterations and participle suffixation (e.g. teach – taught ). Labelling strong verbs as completely irregular is also misleading, since they do form patterned and almost rule-like paradigms (Albright, 2009; Albright & Hayes, 2003; Regel et al., 2015). Furthermore, contrasts such as regular vs irregular as well as weak vs strong are easily conflated with affixal vs non-affixal, or concatenative vs non-concatenative processing. Such putative parallelisms are oversimplifications, though; some irregular verbs have suffixes (teach – taught ), while others do not show vowel change at all (hit – hit ). In addition to that, irregular does not automatically mean unproductive or arbitrary, as an abundance of systematic irregularisation in behavioural data shows (s. Chapter 1 for examples). In turn, seemingly lexicalised forms are often both paradigmatically connected and derivable (Bendjaballah, 2000). Rule-like paradigmatic and derivable information, coded through word-internal changes, is often referred to as apophony, and we can find it in languages like Celtic (Russell, 2017), Classic Arabic (Kuryłowicz, 1956; Ségéral & Scheer, 1998), Gongduk (Gerber, 2022), or Dinka (Ladd et al., 2009). For many Afro-Asiatic and Indo-European languages, some researchers even proposed a universal apophonic mechanism (Ségéral & Scheer, 1998), which would naturally go against the idea of randomly organised, unpredictable, and fully lexicalised forms. In sum, irregular verbs are not so irregular after all; they are variable, but structured. Complicating matters is the terminological confusion that comes with apophony. Despite its frequent occurrence across typologically different languages, there is surprisingly little consensus as to both its morphological status and terminological conventions. In the pertinent literature, it is frequently referred to as vowel change (Wagner, 2010), stem change (Fleischhauer, 2013), ablaut (Quirk et al., 1988), apophony (Ségéral & Scheer, 1998), Internal Vowel Alternation (IVA, Even-Simkin & Tobin, 2013) or gradation (Aarts, 2004; Wiese, 1996, 2008). Moreover, there is a bewildering discrepancy as to adequate categorisations of the verbs apophonic patterns (Wagner, 2010). Despite these issues, non-default inflectional processes will, in the remainder of this volume, be referred to as irregular, primarily because

2

IRREGULAR VERB MORPHOLOGY: THEORETICAL ACCOUNTS

17

this term is still used in countless studies and books. Irregular here means that paradigmatic information is patterned, predictable to some degree, and somewhat between idiosyncratic and freely applying rules. Irregular verbs are therefore considered to be organised in psycholinguistically real mental paradigms, but with restricted generalisation properties (Bybee & Moder, 1983; Bybee & Slobin, 1982; Marcus et al., 1995). And in the remainder of this volume, vowel-changing inflectional processes will be called apophonic; on the one hand, this reflects the paradigm’s nonrandom nature, on the other, it meaningfully captures no-change verbs such as hit or cut, which, strictly speaking, do not include vowel or stem alterations. Moreover, the present analysis will assume a certain psycholinguistic reality of what has been called inflectional classes or paradigms. In such classes, lexical entries contain the entire paradigm’s inflectional information beyond their single-word linguistic properties. This also means that the morphological organisation of such an inflectional paradigm would— at least partly—be sovereign with regard to other morphophonological and morphosyntactic components in the mental lexicon. So, what are the central characteristics of Modern English irregular verbs? There are between 180 and 200 verb types, with counts differing according to reference corpora, English varieties, or dialects (Olatoye, 2022), and because quite a few forms, like the verb to learn for instance, are in the process of being regularised (Bybee & Slobin, 1982; Carrol et al., 2012; Lieberman et al., 2007; Marslen-Wilson & Tyler, 1998; Pinker, 1991; for a contradicting view on regularisation see Peters, 2009). Diachronically, irregular verbs seem to be an endangered species (Juilland & Macris, 1973; Pyles & Algeo, 1992; Vennemann, 2000), with a shrinking type frequency. Since Old English, there have only been about a dozen new additions to this class, while around 50% either fell into disuse or underwent regularisation; this trend is still going on, which is why today we have competing forms such as dreamed and dreamt, or burned and burnt, leaving around 5% irregular verbs in modern English (Depraetere & Langford, 2020; Marcus et al., 1995; Marslen-Wilson & Tyler, 1998). Despite their low type frequency, they feature strongly on the token count (Bybee & Slobin, 1982; Henzen, 1965); they encode basic human activities (come, see, hear, go, get, or make, to mention but a few), are prominent in motherese and caretaker-speech, and they provide important English modals (be, have, do). In fact, according to the Longman Grammar of Spoken and Written English corpus (LGSWE),

18

T. WAGNER

from the twelve most common English lexical verbs eleven are irregular (Biber & Reppen, 2002, p. 205). It is their high token frequency that most probably helped them to survive universal regularisation tendencies (Gries & Ellis, 2015; Pinker & Prince, 1988). How does this class of verbs pattern? Consider Table 2.1 first. It presents examples of apophonic patterns for English past tense formation. The upper part of the table illustrates patterns starting out from infinitive forms, such as come and drive (input perspective), while the lower part demonstrates patterns when analysing how inflected forms can be associated with a base form (output or product-oriented perspective, cf. Bybee, 1995). Table 2.1 English apophony patterns from an input- and output perspective Input perspective

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

Verbs

Infinitive vowel

Past vowel

Patterns

come drive meet speak sing cut beat burn weep go

[2] [aI] [i:] [i:] [I] [2] [i:] [f:] [i:] [e*]

[eI] [e*] [e] [e*] [æ] [2] [i:] [f:] [e] [e]

a-b-a, e.g. [2]–[eI]–[2] a-b-c, e.g. [aI]–[ e*]–[I] a-b-b, e.g. [i:]–[e]–[e] a-b-b + -en, e.g. [i:]–[e*]–[e*] a-b-c, e.g. [I]–[æ]–[2] no-change, e.g. [2]–[2]–[2] no-change + -en, e.g. [i:]–[i:]–[i:] alveolar suffix, e.g. [f:]–[f:]–[f:] a-b-b + suffix, e.g. [i:]–[e]–[e] suppletion, [e*]–[e]–[A]

Output perspective

[aɪ]

[əʊ] [ɪ] [ʌ] [aʊ] [ɔ:] [u:]

drove, rode bit, hid struck bound, found bought, fought flew

2

IRREGULAR VERB MORPHOLOGY: THEORETICAL ACCOUNTS

19

As we can see in Table 2.1, upper part, there are numerous verbs that seem to undergo apophonic changes in a rather inconsistent fashion. Some verbs create patterns with two changes, some with three changes, others do not show any apophony, and some patterns combine with suffixation (build-built, weep-wept, cf. Kielar et al., 2008). While, from a synchronic point of view, this might seem unpredictable, diachronically these patterns are remnants of older phonologically conditioned processes. If, however, the paradigmatic nature is examined starting from the inflected forms, as illustrated in the lower part of Table 2.1, we can see that a great number of verbs can, for instance, be organised into some sort of [aI]-based inflectional class. This analysis is in line with productoriented generalisations, as suggested by Bybee and Moder (1983), Bybee (1995), or Wurzel (1984). Based on behavioural studies with Englishspeaking children and adults, Bybee and colleagues propose that speakers, despite rote-learning irregular verb forms, are able to generalise over such forms and use these generalisations (called schemas) for storage and retrieval of exemplars from their mental lexicon. English irregular verb classes could thus be regarded as the result of such generalisations over inflected forms, such as past tenses or participles, rather than over base forms, like infinitives. That is why they often are called productoriented schemas, since they organise paradigms without explicit recourse to related uninflected class members. Such a view has predictive qualities. Potentially new verbs with an infinitive rhyme [#_aInd] (skrined) could, with a certain probability, show a past tense rhyme [#_a*nd] (skround). A similar reasoning could be proposed for the class of verbs forming their past inflection with the rhyme [O:t] (bought ). In sum, paradigmatic patterns often do not reveal themselves when starting with their base forms; patterns are often more cohesive from an output- or productoriented point of view. And it is that kind of backward generalisation that repeatedly has come up in psycholinguistic experiments with both children and adults (Bybee, 1995, 2001; Bybee & Moder, 1983; Bybee & Slobin, 1982; Marcus et al., 1995; Prasada & Pinker, 1993; Xu & Pinker, 1995). Irregular paradigms like the ones in Table 2.1 are notoriously hard to account for linguistically. That is why the literature is replete with numerous and diverging categorisations (Wagner, 2010). The nature of the apophonic mutations, in particular, seems to challenge morphological analyses. How meaning-bearing is, for instance, a vocalic change from [I]

20

T. WAGNER

to [æ], as in sing and sang ? Neither segment carries or contains the grammatical information for infinitive or past tense, respectively, since nuclei such as [2] can also be found in infinitives (cut ), past tenses (clung ), and participles (sung ). Rather, it is the change itself that encodes grammatical meaning. That is why for instance classical item-and-arrangement approaches (Hoard & Sloat, 1973; Hockett, 1954; Jespersen, 1942) struggled with the non-concatenative nature of English verbal apophony, proposing, for lack of a better alternative, discontinuous and so-called zero allomorphs. Other approaches, like Chomsky’s famous SPE model (Chomsky & Halle, 1968) either employed underlying representations or exhaustive lists of idiosyncratic exceptions in order to account for the existing patterns. Eventually, though, the segmental and rule-based credo of such approaches not only failed to capture apophony as an inflectional process but also the word forms’ paradigmatic relationship (Bloomer, 1994; Halle, 1953; Hoard & Sloat, 1973; Jespersen, 1942; Quirk et al., 1988). Later word and paradigm approaches (Matthews, 1991), including natural morphology (Dressler, 1986), or minimalist morphology (Wunderlich, 1996), abandoned underlying forms and represented allomorphy explicitly, which probably rendered these analyses psycholinguistically more plausible (Cuskley et al., 2015; Pinker & Prince, 1988).

2.2

German Irregular Verbs

Like in Modern English, Modern High German (Pickl, 2023) has between 160 and 200 irregular verbs (Clahsen, 1999; Clahsen et al., 2001; Köpcke, 1998). And again, depending on the choice of reference corpora, varieties, and dialects, counts vary. Today’s type frequency of around 4% (Bittner, 1985, 1996; Köpcke, 1998) is also the result of a continuous decline, and like English, token frequency is high (Clahsen, 1997; Clahsen et al., 1997). Among the 1000 most frequent German word forms, there are 80 verbs, 33 of which are irregular (Augst, 1975; Ruoff, 1981; Weyerts & Clahsen, 1994). Traditionally, German verb morphology has been divided into the two categories strong and weak verb. The class of strong verbs, although diachronically misleading, is the class into which today’s irregular forms (singen ‘sing’, sprechen ‘speak’, rennen ‘run’) are normally grouped. Unlike English, apophony in German verb inflection also occurs with respect to number (ich spreche ‘I speak’ – er spricht ‘he speaks’); verbs

2

IRREGULAR VERB MORPHOLOGY: THEORETICAL ACCOUNTS

21

following this pattern are sometimes called irregular, while verbs showing apophony across tense formation only are called hybrid (Mailhammer, 2007; Trompelt et al., 2013) or subregular (Koch et al., 2023). And unlike English, Modern High German irregular verbs systematically combine apophony and affixation in the participle (singen – sang – gesung-en ‘sing’ – ‘sang’ – ‘sung’). Prefixation with ge- is prosodically conditioned, and suffixation involves the two allomorphs -t and -en, which occur with roughly similar token frequencies (Clahsen, 1997) and emerge in child language acquisition simultaneously (Clahsen & Rothweiler, 1993). The -t participle suffix, however, is considered the default; it is often argued to be highly productive, generalising to both newly coined real words and nonces, and it usually occurs with the stem vowel. The -en suffix, in contrast, generalises only based on phonological similarity (Marcus et al., 1995; Weyerts & Clahsen, 1994). Thus, German irregular participles usually feature the -en suffix. The combination of regular suffixation with apophonic patterns renders German participle formation a promising test case for cognitive models. While English verbs confound regular (walk-ed) with irregular and unpredictable suffixation (brough-t ) as well as frequency with inflectional type, German participles do not (Bybee, 1995; Clahsen, 1999; Pinker & Prince, 1988). As far as German inflectional classes are concerned, verbs combining apophony and suffixation in past tense are usually grouped separately. Unlike English, where buy – bought would traditionally be labelled irregular (+ suffixation), the class for German verbs like rennen – rann-te ‘run – ran’ is labelled a semi-weak or a mixed class (Bittner, 1996, pp. 99– 101; Marusch et al., 2012). Table 2.2 summarises common apophonic patterns in Modern High German. As we can see, there are numerous apophonic patterns across the German verb classes, and from an input perspective, they seem unpatterned, unpredictable, and therefore fully lexicalised (Bybee & Newman, 1995; Nübling et al., 2006; Wiese, 1996, 2000). And in German, too, we find an abundance of competing analyses of these patterns (Augst, 1975; Bittner, 1996; Fabricius-Hansen, 1977; Hempe, 1988). More recent approaches have repeatedly challenged classical analyses (Trompelt, 2010; Trompelt et al., 2013; Wiese, 2008; Wunderlich, 1996), also abandoning underlying forms in favour of cognitively grounded explicit allomorphic and paradigmatic representations. Figure 2.1 illustrates such representations, inspired by minimalist

22

T. WAGNER

Table 2.2 Common apophonic patterns in Modern High German Class

Pattern

Infinitive

Present 3rd pers.

Preterite 3rd pers.

Participle

(a)

Hybrid

(b)

Irregular

(c)

Mixed

(d)

Suppletion

a-b-c [I]-[2]-[*] a-b-b [aI]-[i:]-[i:] a-b-c [I]-[2]-[*] a-b-c [e]-[a:]-[O] a-b-b [e]-[a]-[a] a-b-c [aI]-[a:]-[e:]

trink-en ‘to drink’ bleib-en ‘to stay’ sprech-en ‘to speak’ fall-en ‘to fall’ renn-en ‘to run’ sein ‘to be’

trink-t ‘drinks’ bleib-t ‘stays’ sprich-t ‘speaks’ fall-t ‘falls’ renn-t ‘runs’ ist ‘is’

trank ‘drank’ blieb ‘stayed’ sprach ‘spoke’ fiel ‘fell’ rann-te ‘ran’ war ‘was’

ge-trunk-en ‘drunk’ ge-blieb-en ‘stayed’ ge-sproch-en ‘spoken’ ge-fall-en ‘fallen’ ge-rann-t ‘run’ ge-wesen ‘been’

[ verf ]v to throw

[… …]2./3.pers. pres. ind.

[…]imp.

[ …a…]pret.

[… … n]part.

[…y… ]subj.

Fig. 2.1 Structured lexical entry inheritance tree for the German verb werfen ‘to throw’

morphology (Wunderlich, 1996), for the German verb werfen ‘throw’ in a so-called inheritance tree. In such a tree-like lexical entry, phonological information is directly represented and connected to other paradigmatically related verbs. Starting from the head node, usually the infinitive, phonological information that is redundant is inherited in the lower-level nodes, illustrated through the bracketed dots. Thus, the apophonic change from infinitival [e] to [O] is represented by substitution, while redundant information, for instance the nucleus of the imperative form, is not. Eventually, all phonological information of related inflected forms is associated with each other

2

IRREGULAR VERB MORPHOLOGY: THEORETICAL ACCOUNTS

23

and the mother node on top of the diagram, as illustrated by the tree branches. Although structured lexical entries seem to provide a promising theoretical backdrop for German and English irregular verb morphology, the exact nature of the entries’ structure does not seem to be sufficiently specified; what exactly, for instance, do the lines in Fig. 2.1 represent? Are they always one-way or two-way connections? How is phonological information encoded? Is it through feature bundles or concrete phonemes? And how exactly would the entry for werfen ‘throw’ be associated with phonologically and apophonically similar verbs, like sterben ‘die’? Questions like that have been pursued in a number of behavioural L1 studies already, including children, adults (Clahsen et al., 2004; Smolka et al., 2007) as well as impaired speakers (Marusch et al., 2012; Penke et al., 2014). And there is evidence from electrophysiological and neuroimaging studies (Lück et al., 2006; Smolka et al., 2013) as well as corpus studies (Köpcke, 1998). Satisfactory answers, however, are scarce. German verb morphology, and some Romance languages (Marzi & Pirrelli, 2022; Meunier & Marslen-Wilson, 2004; Milin et al., 2009) admittedly advanced to a new test case within the past tense debate (Clahsen, 1999; Marcus et al., 1995) but ultimately failed to make a significant contribution to settling it. Overall, many accounts of verb morphology cannot capture the internal organisation or the apophonic processes in a satisfactory fashion. Models are either underspecified with regard to paradigmatic properties, or they propose cumbersome and highly abstract mechanisms that often lack behavioural and psycholinguistic underpinning (Bittner, 1996). In sum, current approaches appear to remain inconclusive to a considerable extent (Strobach & Schönpflug, 2011).

References Aarts, B. (2004). Modelling linguistic gradience. Studies in Language, 28(1), 1–49. https://doi.org/10.1075/sl.28.1.02aar Albright, A. (2009). Modeling analogy as probabilistic grammar. In J. Blevins & J. Blevins (Eds.), Analogy in grammar: Form and acquisition (pp. 185–213). Oxford University Press. Albright, A., & Hayes, B. (2003). Rules vs. analogy in English past tenses: A computational/experimental study. Cognition, 90(2), 119–161. https://doi. org/10.1016/S0010-0277(03)00146-X

24

T. WAGNER

Anderson, S. R. (1988). Morphological theory. In F. J. Newmeyer (Ed.), Linguistics: The Cambridge survey. Vol. I. Linguistic theory: Foundations (pp. 146–191). Cambridge University Press. Augst, G. (1975). Untersuchungen zum Morpheminventar der deutschen Gegenwartssprache. Gunter Narr. Bendjaballah, S. (2000). The negative preterite in Kabyle Berber. Folia Linguistica, 34(3–4), 185–224. https://doi.org/10.1515/flin.2000.34.3-4.185 Biber, D., & Reppen, R. (2002). What does frequency have to do with grammar teaching? Studies in Second Language Acquisition, 24(2), 199–208. https:// doi.org/10.1017/S0272263102002024 Bittner, A. (1985). Wie schwach sind die starken Verben? Überlegungen und Vorschläge zu einer Klassifizierung der neuhochdeutschen Verben. In W. U. Wurzel (Ed.), Studien zur Morphologie und Phonologie I . Studia Grammatica 21 (pp. 51–74). Akademie Verlag. Bittner, A. (1996). Starke ‚schwache’ Verben, schwache ‚starke’ Verben. Deutsche Verbflexion und Natürlichkeit. Stauffenburg. Bloomer, R. (1994). System-congruity and the participles of modern German and modern English. A study in natural morphology. Buske Verlag. Bybee, J. (1995). Regular morphology and the lexicon. Language and Cognitive Processes, 10(5), 425–455. https://doi.org/10.1080/01690969508407111 Bybee, J. (2001). Phonology and language use. Cambridge University Press. Bybee, J., & Moder, C. L. (1983). Morphological classes as natural categories. Language, 59(2), 251–270. https://doi.org/10.2307/413574 Bybee, J., & Newman, J. (1995). Are stem changes as natural as affixes? Linguistics, 33(4), 633–654. https://doi.org/10.1515/ling.1995.33.4.633 Bybee, J., & Slobin, D. I. (1982). Rules and schemas in the development and use of the English past tense. Language, 58, 265–289. https://doi.org/10. 2307/414099 Carrol, R., Svare, R., & Salmons, J. (2012). Quantifying the evolutionary dynamics of German verbs. Journal of Historical Linguistics, 2(2), 153–172. https://doi.org/10.1075/jhl.2.2.01car Chomsky, N., & Halle, M. (1968). The sound pattern of English. Harper & Row. Clahsen, H. (1997). The representation of German participles in the German mental lexicon: Evidence for the dual-mechanism model. In G. Booij & J. v. Marle (Eds.), Yearbook of morphology 1996 (pp. 73–96). Kluwer Academic. Clahsen, H. (1999). Lexical entries and rules of language: A multidisciplinary study of German inflection. Behavioral and Brain Sciences, 22(6), 991–1060. https://doi.org/10.1017/S0140525X99002228 Clahsen, H., Eisenbeiss, S., Hadler, M., & Sonnenstuhl, I. (2001). The mental representation of inflected words: An experimental study of adjectives and verbs in German. Language, 77 (3), 510–543. https://doi.org/10.1353/lan. 2001.0140

2

IRREGULAR VERB MORPHOLOGY: THEORETICAL ACCOUNTS

25

Clahsen, H., Eisenbeiss, S., & Sonnenstuhl-Henning, I. (1997). Morphological structure and the processing of inflected words. Theoretical Linguistics, 23, 201–249. https://doi.org/10.1515/thli.1997.23.3.201 Clahsen, H., Hadler, M., & Weyerts, H. (2004). Speeded production of inflected words in children and adults. Journal Child Language, 31(3), 683–712. https://doi.org/10.1017/S0305000904006506 Clahsen, H., & Rothweiler, M. (1993). Inflectional rules in children’s grammars: Evidence from German participles. In G. Booij & J. van Marle (Eds.), Yearbook of morphology 1992 (pp. 1–34). Springer. https://doi.org/10.1007/97894-017-3710-4_1 Cuskley, C., Colaiori, F., Castellano, C., Loreto, V., Pugliese, M., & Tria, F. (2015). The adoption of linguistic rules in native and non-native speakers: Evidence from a wug task. Journal of Memory and Language, 84, 205–223. https://doi.org/10.1016/j.jml.2015.06.005 Depraetere, I., & Langford, C. (2020). Advanced English grammar. A linguistic approach (2nd ed.). Bloomsbury Academic. Dressler, W. U. (1986). Explanation in natural morphology, illustrated with comparative and agent-noun formation. Linguistics, 24, 519–548. https:// doi.org/10.1515/ling.1986.24.3.519 Even-Simkin, E., & Tobin, Y. (2013). The regularity of the ‘irregular’ verbs and nouns in English. John Benjamins. Fabricius-Hansen, C. (1977). Zur Klassifizierung der starken Verben im Neuhochdeutschen. Deutsche Sprache, 5, 193–205. Fleischhauer, E. (2013). Morphological processing in children. An experimental study of German past participles (Doctoral dissertation). Faculty of Human Sciences, the University of Potsdam. Publication portal of University of Potsdam. https://publishup.uni-potsdam.de/opus4-ubp/frontdoor/ index/index/year/2014/docId/6807 Gerber, P. (2022). Verb stem alternation in Gongduk. Synchronic and diachronic analysis. Language and Linguistics, 23(2), 240–273. https://doi.org/10. 1075/lali.00105.ger Gries, S. T., & Ellis, N. C. (2015). Statistical measures for usage-based linguistics. Language Learning, 65(S1), 228–255. https://doi.org/10.1111/lang.12119 Halle, M. (1953). The German conjugation. Word, 9(1), 45–53. https://doi. org/10.1080/00437956.1953.11659455 Hempe, U. (1988). Die starken Verben im Deutschen und Niederländischen. Diachrone Morphologie. Niemeyer. Henzen, W. (1965). Deutsche Wortbildung. Niemeyer. Hoard, J., & Sloat, C. (1973). English irregular verbs. Language, 49(1), 107– 120. https://doi.org/10.2307/412105 Hock, D. (1968). Die Klassifizierung der starken Verben in der deutschen Gegenwartssprache. Deutsch als Fremdsprache, 5, 219–221.

26

T. WAGNER

Hockett, C. F. (1954). Two models of grammatical description. Word, 10(2–3), 210–234. https://doi.org/10.1080/00437956.1954.11659524 Jespersen, O. (1942). A modern English grammar on historical principles. Part IV: Morphology. Munksgaard. Juilland, A., & Macris, J. (1973). The English verb system. Mouton de Gruyter. Kielar, A., Joanisse, M. F., & Hare, M. L. (2008). Priming English past tense verbs: Rules or statistics? Journal of Memory and Language, 58(2), 327–346. https://doi.org/10.1016/j.jml.2007.10.002 Koch, E. M., Bulté, B., Housen, A., & Godfroid, A. (2023). The predictive processing of number information in subregular verb morphology in a first and second language. Applied Psycholinguistics, 1–34. https://doi.org/10. 1017/S014271642300022X Köpcke, K. M. (1998). Prototypisch starke und schwache Verben in der deutschen Gegenwartssprache. Sonderheft Germanistische Linguistik, 141–142, 45–60. Kuryłowicz, J. (1956). L’apophonie en Indo-Européen. Polska Academia Nauk. Ladd, D. R., Remijsen, B., & Manyang, C. A. (2009). On the distinction between regular and irregular inflectional morphology: Evidence from Dinka. Language, 85(3), 659–670. https://doi.org/10.1353/lan.0.0136 Lieberman, E., Michel, J. B., Jackson, J., Tang, T., & Nowak, M. A. (2007). Quantifying the evolutionary dynamics of language. Nature, 449(7163), 713– 716. https://doi.org/10.1038/nature06137 Lück, M., Hahne, A., & Clahsen, H. (2006). Brain potentials to morphologically complex words during listening. Brain Research, 1077 , 144–152. https://doi. org/10.1016/j.brainres.2006.01.030 Mailhammer, R. (2007). The Germanic strong verbs: Foundation and development of a new system. Trends in Linguistics. Studies and Monographs 183. de Gruyter. Marcus, G. F., Brinkmann, U., Clahsen, H., Wiese, R., & Pinker, S. (1995). German inflection: The exception that proves the rule. Cognitive Psychology, 29(3), 189–256. Marslen-Wilson, W., & Tyler, L. K. (1998). Rules, representations, and the English past tense. Trends in Cognitive Science, 2(11), 428–435. https:// doi.org/10.1016/s1364-6613(98)01239-x Marusch, T., von der Malsbur, T., Bastiaanse, R., & Burchert, F. (2012). Tense morphology in German agrammatism. The production of regular, irregular and mixed verbs. The Mental Lexicon, 7 (3), 351–380. https://doi.org/10. 1075/ml.7.3.05mar Marzi, C., & Pirrelli, V. (2022). Psycholinguistic research on inflectional morphology in the Romance languages. In Oxford research encyclopaedia of linguistics. Retrieved May 10, 2023, from https://oxfordre.com/linguistics/ display/10.1093/acrefore/9780199384655.001.0001/acrefore-978019938 4655-e-709

2

IRREGULAR VERB MORPHOLOGY: THEORETICAL ACCOUNTS

27

Matthews, P. H. (1991). Morphology (2nd ed.). Cambridge University Press. Meunier, F., & Marslen-Wilson, W. (2004). Regularity and irregularity in French verbal inflection. Language and Cognitive Processes, 19(4), 561–580. https:// doi.org/10.1080/01690960344000279 Milin, P., Filipovi´c Ðurd-evi´c, D., del Prado, M., & Martín, F. (2009). The simultaneous effects of inflectional paradigms and classes on lexical recognition: Evidence from Serbian. Journal of Memory and Language, 60(1), 50–64. https://doi.org/10.1016/j.jml.2008.08.007 Nübling, D., Dammel, A., Duke, J., & Szczepaniak, R. (2006). Historische Sprachwissenschaft des Deutschen: Eine Einführung in die Prinzipien des Sprachwandels. Gunter Narr. Olatoye, T. (2022). Irregular verb morphology in Nigerian English. World Englishes, 41(3), 1–19. https://doi.org/10.1111/weng.12595 Paul, H. (1917). Deutsche Grammatik (Vol. 2). Niemeyer. Penke, M., Wimmer, E., Hennies, J., Hess, M., & Rothweiler, M. (2014). Inflectional morphology in German hearing-impaired children. Logopedics Phoniatrics Vocology, 41(1), 9–26. https://doi.org/10.3109/14015439. 2014.940382 Peters, P. (2009). Irregular verbs. Regularization and ongoing variability. In P. Peters, P. Collins, & A. Smith (Eds.), Comparative studies in Australian and New Zealand English: Grammar and beyond. Varieties of English around the world (pp. 13–30). John Benjamins. https://doi.org/10.1075/veaw.g39. 02pet Pickl, S. (2023). (High) German. In Oxford research encyclopaedia of linguistics. Retrieved May 10, 2023, from https://oxfordre.com/linguistics/view/ 10.1093/acrefore/9780199384655.001.0001/acrefore-9780199384655e-939 Pinker, S. (1991). Rules of language. Science, 253(5019), 530–534. https://doi. org/10.1126/science.1857983 Pinker, S. (1999). Words and rules. The ingredients of language. Phoenix. Pinker, S., & Prince, A. (1988). On language and connectionism. Analysis of parallel distributed processing of language acquisition. In S. Pinker & J. Mehler (Eds.), Connections and symbols (pp. 73–193). MIT Press. Prasada, S., & Pinker, S. (1993). Generalization of regular and irregular morphological patterns. Language and Cognitive Processes, 8, 1–56. https://doi.org/ 10.1080/0169096930840694 Pyles, T., & Algeo, J. (1992). The origins and development of the English language (4th ed.). Harcourt Brace. Quirk, R., Greenbaum, S., Leech, G., & Svartvik, J. (1988). A comprehensive grammar of the English language. Longman.

28

T. WAGNER

Regel, S., Opitz, A., Müller, G., & Friederici, A. D. (2015). The past tense debate revisited: Electrophysiological evidence for subregularities of irregular verb inflection. Journal of Cognitive Neuroscience, 27 (9), 1870–1885. https://doi.org/10.1162/jocn_a_00818 Ruoff, A. (1981). Häufigkeitswörterbuch gesprochener Sprache. Niemeyer. Russell, P. (2017). The evolution of Celtic. In J. Klein, B. Joseph, & M. Fritz (Eds.), Handbook of comparative and historical Indo-European linguistics: An international handbook (Vol. 2, pp. 1274–1297). Mouton De Gruyter. https://doi.org/10.1515/9783110523874-028 Ségéral, P., & Scheer, T. (1998). A generalized theory of ablaut: The case of modern German strong verbs. In R. Fabri, A. Ortmann, & T. Parodi (Eds.), Models of inflection (pp. 28–59). De Gruyter. https://doi.org/10.1515/978 3110919745.28 Smolka, E., Khader, P. H., Wiese, R., Zwitserlood, P., & Rösler, F. (2013). Electrophysiological evidence for the continuous processing of linguistic categories of regular and irregular verb inflection in German. Journal of Cognitive Neuroscience, 25(8), 1284–1304. https://doi.org/10.1162/jocn_a_00384 Smolka, E., Zwitserlood, P., & Rösler, F. (2007). Stem access in regular and irregular inflections: Evidence from German principles. Journal of Memory and Language, 57 , 325–347. https://doi.org/10.1016/j.jml.2007.04.005 Smolka, E., Zwitserlood, P., Wiese, R., Marslen-Wilson, W., & Rösler, F. (2018, revision). Priming effects of German participles—The past-tense debate is not over yet. Manuscript in preparation. Strobach, T., & Schönpflug, U. (2011). Can a connectionist model explain the processing of regularly and irregularly inflected words in German as L1 and L2? International Journal of Bilingualism, 15(4), 446–465. https://doi.org/ 10.1177/1367006911403205 Trompelt, H. (2010). Production of regular and non-regular verbs. Evidence for a lexical entry complexity account. Spektrum Patholinguistik Schriften (Vol. 2). Universitätsverlag Potsdam. Trompelt, H., Bordag, D., & Pechmann, T. (2013). (Ir)Regularity of verbs revisited: Evidence for lexical entry complexity. The Mental Lexicon, 8(1), 26–52. https://doi.org/10.1075/ml.8.1.02tro Vennemann, T. (2000). Zur Entstehung des Germanischen. Sprachwissenschaft, 25, 233–269. Wagner, T. (2010). Interlanguage morphology. Irregular verbs in the mental lexicon of German-English interlanguage speakers. Narr-Francke-Attempto. Weyerts, H., & Clahsen, H. (1994). Netzwerke und symbolische Regeln im Spracherwerb: experimentelle Ergebnisse zur Entwicklung der Flexionsmorphologie. Linguistische Berichte, 154, 430–460.

2

IRREGULAR VERB MORPHOLOGY: THEORETICAL ACCOUNTS

29

Wiese, R. (1996). Phonological versus morphological rules. On German ablaut and umlaut. Journal of Linguistics, 32(1), 113–135. https://doi.org/10. 1017/S0022226700000785 Wiese, R. (2000). The phonology of German. Oxford University Press. Wiese, B. (2008). Form and function of verbal ablaut in contemporary standard German. In R. Sackmann (Ed.), Explorations in integrational linguistics. Four essays on German, French, and Guaraní. Current Issues in Linguistic Theory 285 (pp. 97–151). John Benjamins. Wunderlich, D. (1996). Minimalist morphology: The role of paradigms. In G. Booij & J. van Marle (Eds.), Yearbook of morphology 1995 (pp. 93–114). Kluwer Academic. Wurzel, W. U. (1970). Studien zur deutschen Lautstruktur. Studia Grammatica 8. Akademie Verlag. Wurzel, W. U. (1984). Flexionsmorphologie und Natürlichkeit. Akademie-Verlag. Xu, F., & Pinker, S. (1995). Weird past tense forms. Journal of Child Language, 22(2), 531–556. https://doi.org/10.1017/S0305000900009946

CHAPTER 3

Irregular Verb Morphology: L1 and L2 Cognitive Accounts

Abstract This chapter contextualises associative approaches to L1 and L2 verb morphology within the two major camps competing in the past tense debate. After reviewing studies from the dual-route perspective, which views morphological processing as operating with both symbolic processing and pattern association, single-route approaches, including connectionism, analogical learners, prototype models, are discussed. Similarity-based prototype models, in particular, are reviewed in greater detail, since they provide the theoretical backdrop of the two empirical studies in Chapter 4. Therefore, this chapter closes synthesising evidence from such models with respect to stimuli creation, experimental design of the two empirical studies. Keywords Single-route and dual route · Connectionism · Analogy · Prototype · Schema · Apophony

Cognitive accounts of both English and German irregular verb morphology can be broken down into two main camps, reflecting one of the most controversial debates in cognitive science in the 1980s and 1990s (Leminen et al., 2019; Pinker, 1999). At the heart of this debate is the question as to whether human cognition relies on one or two cognitive mechanisms. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 T. Wagner, English Interlanguage Morphology, https://doi.org/10.1007/978-3-031-50617-8_3

31

32

T. WAGNER

On the one hand, single-route models propose that one cognitive architecture would be sufficient to account for both default and nondefault-like morphological processing. In many of such models, learning is conceptualised as linguistic representations gradually emerging from language usage and experience. This single-route camp includes various concrete applications, such as connectionist, associative analogical, and prototype models (Leminen et al., 2019; McClelland & Patterson, 2002; Seidenberg & Plaut, 2014; Westermann, 2000). A great number of computational studies have produced evidence in favour of such models, and many L1 behavioural studies, too, support this approach. On the other hand, there is a dual route or hybrid account, which proposes two separate and fundamentally different cognitive mechanisms. In such models, verb morphology would require both pattern association and morpheme-based symbolic rule processing (Pinker & Ullman, 2002). In terms of learning, such hybrid models would claim that linguistic representations in a speaker or learner are partly innate. Here, too, countless L1 behavioural studies present evidence in favour of a dual-route mechanism. And there is evidence for both approaches from learner languages, too. Despite L2 language processing being a “young but lively domain of scientific research, where the still limited results available cannot be easily fitted into a uniform model” (Giraudo & Dal Maso, 2018, p. 617), the past tense debate has brought forth numerous studies on L2 irregular verb morphology. Two leading journals even have published special issues on this topic in recent years (Rogers et al., 2016), including behavioural and neuroimaging studies (Roncaglia-Denissen & Kotz, 2016). Despite this growing body of empirical evidence, results are still mixed, though, and it remains unclear how categorically distinct L1 and L2 morphological processing actually are, and if L2 processing had best be explained by single- or dual-route accounts (see Shirai, 2019, for a recent review).

3.1 Dual-Route Models of Irregular Verb Morphology When critically reviewing Rumelhart and McClelland’s (1986) computer model, Pinker and Prince (1988) argued that rules were essential and indispensable for modelling past tense inflection. Instead of one unitary cognitive apparatus, they suggested a decompositional, dualroute approach, also known as dual mechanism, entries-plus-rules-model, words-and-rules-theory, or hybrid model (Clahsen, 1997, 1999; Pinker &

3

IRREGULAR VERB MORPHOLOGY: L1 AND L2 COGNITIVE …

33

Ullman, 2002; Wunderlich & Fabri, 1995). Such a hybrid is supposed to mirror two allegedly central characteristics of human cognition, namely symbolic and associative processing. While regular inflection would be computed by deterministic rules, irregular morphology would involve form-specific associative analogical processing. Thus, in dual-route models, regular forms would not feature lexical representation in the lexicon but would be generated on demand, while irregular forms are generated by an associative network outside the grammar. The last three decades have brought forward plenty of L1 evidence in favour of this approach, from a variety of sources and languages, and mostly from behavioural or brain imaging studies (Bozic et al., 2014; Faroqi-Shah, 2013; Leminen & Clahsen, 2014; Ullman, 2001a, 2001b, 2004; Veríssimo & Clahsen, 2014 among others; for a review of older studies see Wagner, 2010). There is disagreement, though, as to the interplay between the two routes. They could be conceptualised as operating parallelly, with a faster full-form lexical track (Caramazza et al., 1988; Chialant & Caramazza, 1995), or as competing (Baayen et al., 1997; Baayen & Schreuder, 1999; Schreuder & Baayen, 1995). As far as learner languages are concerned, there are also plenty of studies supporting the dual-route approach (Beck, 1997; Birdsong & Flege, 2001; Clahsen et al., 2010; Flege et al., 1999; Gor & Chernigovskaya, 2003; Hahne et al., 2006; Kırkıcı, 2007; Krause et al., 2015; Pliatsikas & Marinis, 2013). Empirical evidence from such studies usually illustrates a categorical and qualitative dissociation between symbolic and associative processing in behavioural data, which would be at odds with the gradual and form-related transition between regular and irregular processes, as advocated by single-route approaches. Flege et al. (1999), for instance, found such a dissociation in that age of arrival (AoA) was a predictor for the accuracy of L2 irregular verb but not regular verbs; Birdsong and Flege (2001) reported similar effects. And Safaie (2021) could show that, apart from strong communalities between L1 and L2 verb processing, for both regular and irregular verbs, there was no frequency effect for regulars in either group, lending further support to dual-route processing. Symbolic and rule-based morphological processing was further developed in a single-route fashion, sometimes referred to as generalised context models (Nosofsky, 1990; Racz et al., 2014), or multiple-rules models (Ambridge, 2010). They model even idiosyncratic morphological

34

T. WAGNER

operations by inductive stochastic micro-rules (Albright, 2009; Plunkett & Bandelow, 2006). Explicitly challenging associative and similaritybased processing, Albright and Hayes (2003), for example, developed a computer model that incrementally derives morphophonological microgeneralisations about English verb morphology from input, thus creating ever more complex and comprehensive inflectional rules for particular phonological environments. Each micro-rule is then assessed in terms of accuracy and receives a so-called confidence score, which in turn trains the model to process new input through ever more specific rules. When gauging their model’s performance against behavioural data (Prasada & Pinker, 1993), Albright and Hayes found that their micro-rules could successfully replicate gradient responses found in Prasada and Pinker’s original analysis. Unlike early generative rule-based models (Chomsky & Halle, 1968), and unlike the dual-route model, however, their micro-rules captured differing probabilities of inflectional patterns within both regular and irregular inflection, rendering classic default inflection gradient, as it were (see Albright, 2009 for similar findings in Italian; see Ernestus & Baayen, 2003 for Dutch). Consequently, both symbolic and similaritybased processing no longer appeared to represent an insurmountable contradiction. A second rule-like approach comes from universal apophony. Starting out with analyses of Classical Arabic inflection (Guerssel & Lowenstamm, 1993, 1996), universal apophony has been claimed to account for various linguistic paradigms across many languages (Bendjaballah & Haiden, 2002, 2005; Boyé, 2000; Gerber, 2022; Ségéral, 2000). At the very heart of this approach lies the so-called universal apophonic path, often spelled out as Ø → i → a → u → u. It contains three different vowel qualities and four context-free mono-directional mutations, illustrated by the →. The path’s mutations always operate from left to right, and while partial apophony is allowed, skipping steps in the path is illicit. Note that this path also illustrates increasing degrees of markedness, with Ø representing the most unmarked element. For irregular verb morphology in English and German, this would mean that a verb’s lexical entry in the mental lexicon would merely require the stem vowel’s phonological information. Any other vowel qualities in the verbal paradigm would not be explicitly represented, since they could be derived on demand in language use (Bendjaballah & Haiden, 2002, 2005; Boyé, 2000; Gerber, 2022; Guerssel & Lowenstamm, 1996; Ségéral, 2000). Looking more closely at concrete examples in English and

3

IRREGULAR VERB MORPHOLOGY: L1 AND L2 COGNITIVE …

35

German, however, it seems as though the path was insufficient; there are many more vowels in both languages than the path contains. English, for instance, has vowels such as [f:] or [O:], and in German we have umlaut vowels like [y:] or [E:], none of which are explicitly addressed in the apophonic path. Ségéral and Scheer, therefore, re-analyse German irregular verb morphology on an infra-segmental level, on which vowels can consist of so-called parasitic and apophonic elements (Kaye et al., 1985), only one of which—namely the apophonic element—can enter the apophonic mechanism. The nucleus [e] of the German verb bergen ‘recover’, for instance, would be re-analysed as consisting of parasitic [a] and the apophonic [i]. Speakers then implicitly feed the apophonic vowel quality [i] into the path and thus inflect the verb bergen for past tense as barg. Based on this re-analysis, the apophonic path could, they argued, account for German apophonic patterns such as singen-sang-gesungen, rinnen-rann-geronnen, biegen-bog-gebogen, and liegen-lag-gelegen. Further support for such universal apophonic relationships comes from children’s nursery rhymes, onomatopoeic expressions, and expletives. Among these, apophonic mutations such as [I-æ] and [I-2] are remarkably frequent. Moreover, the pattern [I-2-*] forms a triad of optimally discernible vowels (Martinet, 1955, pp. 263–264) and is in fact the most frequent apophonic pattern in German (Augst, 1975; Marchand, 1969, p. 430). If the apophonic path was psycholinguistically real, it would seriously challenge both single- and dual-route models. However, Ségéral and Scheer’s (1998) results remain inconclusive. First, their analysis of German verb morphology excluded mixed verbs while including archaic or regularised forms (dingen, erkiesen, kreischen, schnauben), amounting to more than 20 controversial cases. Second, they oversimplified phonological surface properties of German vowels, like the ATR contrast (i: versus I, like in bieten ‘offer’ versus bitten ‘request’, for example, distinguishing between what has traditionally been termed tense and lax in phonology). Third, only around 60% of German irregular verbs were fully predictable, leaving high-frequent verbs like kommen ‘come’ or rufen ‘call’ unaccounted for. In order to psycholinguistically corroborate this approach, empirical studies would need to show how inflectional behaviour in speakers actually mirrors predictions made by this path.

36

T. WAGNER

3.2 Single-Route Models of Irregular Verb Morphology Most single-route approaches share a long-harboured suspicion that generative symbolic rule processing was a mere by-product of linguistic enquiry, a convenient illusion that was not borne out by empirical or computational evidence, and thus not necessary (Aarts, 2004; Ambridge, 2010; Blything et al., 2018; Bompolas et al., 2017; Bybee, 2001; Fanselow et al., 2006; Givón, 1986; Hay & Baayen, 2005; Itkonen, 2005; Kielar et al. 2008; López Rúa, 2003; Post et al., 2008; Slioussar et al., 2014; Tabak et al., 2010; Vosters, 2012; Wanner, 2006). What linguistics traditionally called grammatical rules is now considered a mere epiphenomenon of associative similarity-based processing in action. Proponents of such an approach would model irregular verb morphology through massive storage and similarity-based generalisations only, which, they argue, would accurately reflect both acquisition and actual language use. Partly owing to what has been termed the computational turn (Divjak & Milin, 2023), single-route models received considerable attention throughout the past three decades. There are various manifestations and models within the single-route camp, including connectionist networks (Bechtel & Abrahamsen, 1990; Christiansen & Chater, 1999; Rumelhart & McClelland, 1986; Sokolik, 1990), analogical or memory-based learners (Arndt-Lappe et al., 2018; Christiansen & Chater, 1999; Daelemans et al., 1994, 1999; Daelemans & van den Bosch, 2005; Jones, 1996; Milin et al., 2023; Plag et al., 2023; Seidenberg & MacDonald, 1999; Skousen, 1989, 2005, 2009; Westermann, 2000), and prototype models (Bybee, 1995, 1988). For an overview see Ambridge (2020), Chandler (2017), or Naranjo (2019). In the remainder of this chapter, they will be discussed in turn, with reference to both L1 and L2 morphological processing. The past tense debate arguably commenced with a connectionist model (Rumelhart & McClelland, 1986), capturing both default and non-default English verbal inflection. Inspired by human cerebral structures, such models implement language behaviour in associative networks, which are sub-symbolic in nature. Such networks learn, unsupervised, by adapting parallel associations between verb stems (input units) and their related past tense forms (output units), thus generalising onto new input once it is similar enough in form (Rumelhart & McClelland, 1986). In addition to that, some later models sometimes contained hidden

3

IRREGULAR VERB MORPHOLOGY: L1 AND L2 COGNITIVE …

37

layers (Plunkett & Marchman, 1993) or tried to incorporate semantic information (Joanisse & Seidenberg, 1999). Nowhere in such networks, however, is rule-based manipulation explicitly programmed, and patterns and regularities, emerging when looking at the network’s performance, are a mere by-product of distributed memory activation. Therefore, such networks are often said to be able to “learn without rules” (Seidenberg & Plaut, 2014; Sokolik, 1990, p. 685). While for rule-based models language is nothing but “rules all the way down”, in connectionist models it is “memory all the way up” (Ambridge, 2020; Baayen, 2007; Feldman et al., 2010; Pinker, 1999, p. 97). Connectionist networks also dispense with full-form representations and instead distribute morphological information across the entire network. Such a unitary, transparent and parsimonious architecture is said to have a high degree of psycholinguistic validity. Rumelhart and McClelland’s pioneering work was followed by a proliferation of various connectionist models for a whole range of linguistic phenomena, like word recognition, reading, or syntactic and semantic processing (for recent overviews see Joanisse & McClelland, 2015; Kirov & Cotterell, 2018; Kodner, 2022). In nominal inflection, for instance, we find computational studies on German plurals (Hahn & Nakisa, 2000), French nouns (Sokolik & Smith, 1992), or noun marking in Serbian (Mirkovic et al., 2011). A whole plethora of evidence comes from verbal morphology studies across various languages (Desai et al., 2006; Engelmann et al., 2019; Justus et al., 2008; for older studies see Wagner, 2010; for comments on this development from an SLA perspective see N. C. Ellis, 2002; Shirai, 2019). Single-route approaches have not only been modelled in computational networks, however. Additional support for these approaches also comes from behavioural studies, like Strobach and Schönpflug (2011) on German irregular verbs, or Zaretsky et al. (2016) on German plural formation. Overall, connectionist models have stimulated a vast amount of research and accumulated a large body of supporting evidence, although even after three decades it is not quite clear if they have the upper hand in the past tense debate (Seidenberg & Plaut, 2010). Marusch et al. (2019), for instance, in a recent picture naming study on German participle inflection, concluded that their behavioural data would be inconclusive with respect to connectionist approaches. In two experiments, they did find morphological priming effects, but naming latencies were not influenced by verb type, which, they reasoned, could not be

38

T. WAGNER

accounted for by either dual- or single-route models. Behavioural data supporting single-route English past tense formation come from Westermann et al. (2008), Butler et al. (2012), as well as Blything et al. (2018). In the area of the L2, connectionist modelling has long been hailed as a promising alternative to generative approaches (Broeder & Plunkett, 1997; Feldman, 1985). Despite this initial euphoria, however, concrete computational applications are rather scarce. Blackwell and Broeder (1992) successfully modelled longitudinal data from Turkish and Moroccan learners of Dutch pronominal forms in a three-layered back propagation network, but many other L2 programmes work with simulations, and not concrete data, like in Gasser (1990), Elman (1991), or Sokolik and Smith (1992), for instance. It is challenging, after all, to imagine training a computer network with a realistic L1 or multilingual background on concrete new L2 language input. Despite this challenge, however, Nelson (2013) advocated a substantial role of connectionism for SLA theory. In contrast to connectionism, analogical or memory-based learners operate with massive storage of full-form representations of linguistic information and rather simple and transparent algorithms (Aha et al., 1991). Like connectionist models, they are operational even if linguistic input is sparse, noisy, corrupt, or not well-formed (Jurafsky, 2002). Over the past three decades, such models have been applied to a variety of linguistic phenomena in various languages. There are studies on Dutch and German compounding (Krott et al., 2001, 2007) and noun plural marking (Hahn & Nakisa, 2000; Keuleers et al., 2007), Arabic plurals (Dawdy-Hesterberg & Pierrehumbert, 2014), and, most recently, English derivational morphology (Plag et al., 2023). Using data from various behavioural studies, analogical models have successfully challenged competing non-associative approaches, too (Eddington, 2000). English verb morphology, in particular, has been repeatedly addressed, like in Chandler (1994), Derwing and Skousen (1994), Eddington (2000, 2004), Hickey and Martin (2001), and van Noord and Spenader (2015), among others. Quite a few analogical learners can be traced back to the simple analogical two-part equation A :: B as C :: D (Hock, 1991, p. 173; see Bybee, 2010, for critical reflections on this assumption). Such a rather simple analogical operation across stored exemplars is part of the so-called performance component in analogical computer models, while the full-form

3

IRREGULAR VERB MORPHOLOGY: L1 AND L2 COGNITIVE …

39

feature matrix representations of the exemplars are located in the learning component. The feature matrices would code phonetic information such as a verb’s onset, nucleus, coda, and its inflections, as well as its semantic meaning of pastness. The performance unit would take these exemplars and, using some sort of distance metrics, carry out classification tasks and suggest the most frequent inflectional pattern in the learning component to a new item. While analogical or memory-based learners have not yet been fully applied to the area of L2 verb morphology (cf. Chandler, 1994, for L1 verb morphology), Behrens (2017) clearly advocates a role for analogy in L2 morphology, too. And there are many behavioural L2 studies in favour of associative and similarity-based approaches (Feldman et al., 2010; Giraudo & Dal Maso, 2018; Nicoladis & Paradis, 2012, Paradis et al., 2011; Rogers et al., 2016; Romano, 2015; for an overview see Diessel, 2019). Analogical models, too, however, have their shortcomings. It is unclear, for instance, how they would manage similarity-based classifications if there are only very few exemplars to draw on for analogical computations (Albright, 2009; Itkonen, 2005), or how the output of analogical formations can be constrained to well-formed segment combinations only (Becker, 1990, p. 24; Jespersen, 1942, p. 33); ungrammatical formations would be bound to occur, after all, if, for instance, proportional analogy can create segment combinations without constraints, thereby producing potentially illicit syllables. This has been pointed out by Albright and Hayes (2003), who questioned analogy as a basis for verbal inflection, instead proposing rule-like micro-generalisations, which could theoretically operate on one single infinitive-past-tense association and would thus be independent of the size and complexity of analogical learning components. Despite these critical voices, some researchers recently pointed out advantages of analogical models over connectionist models when pitted against each other with the same data (Ambridge, 2020). Prototype models, which have a considerable history in linguistics (Mangasser-Wahl, 2000; Taylor, 1995), could be thought of as a similarity-based or analogical learner with the additional element of an abstracted, category-specific prototype. Such models propose analogy, gradience, family resemblance, and fuzzy categorisation as the driving force behind cognitive processing. They have been theorised and tested, amongst others, for phonology (Bybee, 2001), morphology (Berg, 2012;

40

T. WAGNER

Bybee & Eddington, 2006; Bybee & Moder, 1983; Bybee & Slobin, 1982), and syntax (Panther & Köpcke, 2008). Regarding verb morphology, such models would, based on morphophonological family resemblance between verbs (Rosch, 1973; Rosch & Mervis, 1975), assume a prototypical representative, or schema, for an entire inflectional class. Such a schema, formalising central tendencies of all its class members, is used as a basis for analogical operations. They generalise over uninflected base forms, like infinitives, as well as derived output or product-oriented forms, like past tense or participle inflections (Beach, 1964; Bybee, 1995; Reed, 1972; Rosch & Mervis, 1975). Several theoretical and behavioural studies have suggested a prototypical organisation of irregular verb morphology for both English and German speakers. As early as Berko’s study (1958), behavioural evidence has pointed towards a similarity-based organisation of irregular verbs. Based on apophonic changes like [# _ I _ ï] → [# _ æ _ ï] as well as [# _ I _ ï] → [# _ 2 _ ï] in her data, she alluded to the paradigmatic power behind phonetically similar and morphologically consistent words and how this could be the basis of generalisations to new words. Later on, Bybee and Slobin (1982) as well as Bybee and Moder (1983) conducted a series of behavioural elicitation experiments with Englishspeaking children and adults, in which they confronted participants with sentence completion frames stimulating the inflection of a given verb for past simple (This is a girl who knows how to VERB. She is VERBing. She did the same thing yesterday. What did she do yesterday? Yesterday she ___ _; cf. Bybee & Slobin, 1982, pp. 267–268). Their results, too, pointed towards a prototypical organisation of verbal morphology. For German, both theoretical and corpus studies suggest a comparable prototypical organisation (Bittner, 1996; Köpcke, 1998). A prototypical organisation of L2 verb morphology has long been suggested, too (DeKeyser, 1995; Godfroid, 2015; Godfroid & Uggen, 2013). And empirical corroboration comes from a pioneering study by Plag (2000), who, adapting Bybee and Moder’s (1983) design, observed prototypicality effects in behavioural data from advanced learners of English. Comparable effects can be found in behavioural studies with learners of English (Cuskley et al., 2015; Wagner, 2010) as well as English-speaking learners of German (Wagner, 2017). Thus, prototype

3

IRREGULAR VERB MORPHOLOGY: L1 AND L2 COGNITIVE …

41

models, with their emergent input- and product-oriented morphophonological schemas, seem to offer a promising account of L2 irregular verb morphology. Despite solid psycholinguistic evidence, however, prototype models, too, have been controversial for a long time (Bolinger, 1977; Geeraerts, 1988, 1989; Labov, 1973; Mangasser-Wahl, 2000; Schmid, 1998). The concept itself seems to be vague. First, there is disagreement as to whether both prototypes and prototypical categories are supposed to be fuzzy and gradient in nature (Geeraerts, 1989; Givón, 1986). Aitchison (1992), for instance, argued that the semantic category bird was organised around a prototype, with some birds being more prototypical than others; thus, birdiness would be gradient (Bolinger, 1992). However, the (biological) category itself is probably not prototypical, and class membership would therefore be discrete. After all, birds are either birds or not; they cannot be birds to some degree (Taylor, 2003). Second, some researchers treat a prototype as a cognitively manifest form, representing the best possible exemplar of a category with maximal cue validity (Aitchison, 1992; Bybee, 2001; Pinker, 1999). Others, in contrast, claim prototypes were a mere idealised abstract member of a class (Hampton, 1998; Rosch, 1975). Applied to morphological theory, such a form would not even have to explicitly exist in a mental lexicon of a speaker. This distinction is not trivial, since in both Bybee and Moder (1983) and Plag (2000) nonceverb triggering the highest proportion of apophony did not contain those segments that featured the statistically highest cue validity. For the two empirical studies in this volume, however, associative similarity-based models seem to provide the most appropriate theoretical basis. This has various reasons. First, the dual-route approach, with its integral complexity, still seems to suffer from psycholinguistic plausibility. On the one hand, the necessity of a rule-based mechanism complementing associationist storage can only be argued for from an idealised language competence perspective. On the other hand, the critical question of how such a mechanism is supposed to evolve during child language acquisition seems to be somewhat neglected (Newport, 2016). This dispute is still unresolved, and some researchers have even switched camps during their academic career. While Ambridge (2010), for instance, argued for the legitimacy of both single- and dual-route approaches, as long as pattern association was in the models, he emphatically dismissed any abstract and symbolic representation in human cognition in a more recent publication (Ambridge, 2020). This goes to show how the field

42

T. WAGNER

is still undecided about an appropriate cognitive architecture of verb morphology (Baayen, 2007). Second, empirical evidence from dual-route studies remains inconclusive. Alleged dissociations between two different kinds of cognitive processing, as argued by proponents, have not consistently been demonstrated in behavioural data. Instead, regular and irregular English verbal inflection have successfully been modelled in single-route applications, such as connectionist computer networks (Rumelhart & McClelland, 1986), stochastic rules (Albright & Hayes, 2003), and memory-based analogical learners (Chandler, 1994, 2017). Thus, even if there was some sort of a qualitative dissociation between regular and irregular inflection, it would not necessitate a second cognitive mechanism to account for it. Eventually, traditional conceptual boundaries between deterministic rules, weighted connections, productive schemas, and analogical generalisations might in fact be fluent, merely reflecting stochastic and probabilistic at the very heart of morphological processing (Albright, 2009; Baayen, 2003; Baayen & del Prado Martín, 2005; Baayen & Hay, 2005; Bod, 2009; Bod et al., 2003; Chater et al., 2006; Divjak & Arppe, 2013; Gor & Chernigovskaya, 2005; Hawkins & George, 2006; Jurafsky, 2002; Niu et al., 2022; Romberg & Saffran, 2010; Seidenberg & MacDonald, 1999). In this respect, Bybee (1988) argued long ago already that there was no discrete boundary between what traditionally has been called regular and irregular processing, or between rules, schemas, and prototypes. In fact, upon closer inspection, there are striking similarities between these concepts. In prototype models, for instance, class members reflect a varying degree of proximity in relation to an abstract prototype, which can be operationalised as probabilities for new input to be categorised according to this prototype. In memory-based analogical learners, it is the formal similarity of individual stored exemplars to new input, which generates probabilities for certain categorical treatment of new input, often operationalised in k-nearest neighbour algorithms. And Albright and Hayes’ (2003) confidence values form comparable similarity metrics, expressing gradually decreasing probabilities for a micro-rule to be applied to new input. In other words, Bybee and Moder’s (1983) prototypical English irregular verb schema [sCC _ I _ ï], Köpcke’s German input prototype [# _ I _ ï(k)] and Albright and Hayes’ micro-rule [I] → [2] / C(C)(C) _ ï [+past] seem to ultimately describe the same phenomenon in merely different terms. While Albright and Hayes keep emphasising that their continuum

3

IRREGULAR VERB MORPHOLOGY: L1 AND L2 COGNITIVE …

43

of confidence values was not to be interpreted as a similarity measure, it does nevertheless appear to reflect a gradually increasing degree of family resemblance between verbs subjected to their micro-rules. That is why in the remainder of this chapter, evidence from associative similarity-based models will be reviewed in detail and synthesised with respect to stimuli creation and experimental design in the two empirical studies. Recall first that Bybee and colleagues have repeatedly provided evidence for a similarity-based and potentially prototypical organisation in child and adult English L1 speakers, with schemas like [sCC _ I _ ï(k)] being associated with [I]-[2] apophony, for instance. Recall, too, that, for German irregular verbs, Bittner (1996), referring to concepts such as naturalness and iconicity in his theoretical treatise, analysed inflectional classes as being paradigmatically related, although his notion of prototypicality does not entirely overlap with what Köpcke (1998) or Bybee (1995) described with reference to corpus or behavioural data. In addition to that, Prasada and Pinker (1993), replicating Bybee and Moder’s (1983) behavioural experiments, also found prototypicality effects, and Köpcke (1998) then, in his corpus analysis, was able to distil German prototypical schemas, such as [# _ aI _ plosive], [# _ aI _ fricative], and [# _ I _ ï(k)]. Building on this evidence, Plag (2000), in another nonce-word elicitation experiment, reported prototypical organisation in advanced German-speaking learners of English, with an input prototypelike [sCC _ I _ ï(k)]. Building on and partly replicating Plag (2000), Wagner (2010) found similar but slightly different prototypes, like [sC(C) _ I _ ï(k)] and [sCC _ i: _ s] for German L1 speakers, [sC(C) _ I _ m/ n/ï(k)] for English L1 speakers, and [(s)C(C) _ I _ ï(k)] as well as [sC _ I _ m] for advanced German-speaking learners of English. In addition to that, Wagner’s (2010) data also suggested that analogical formations could be formed in both directions, from infinitive to inflection and vice versa, by both native speaker groups and learners of English. Thus, we can assume that both speakers and learners can generalise over input forms, such as infinitives, but also output forms, or inflected forms, such as past tense formations and participles. In turn, such bidirectional inflectional behaviour would be incompatible with predictions made by both the apophonic path (Becker, 1990; Ségéral & Scheer, 1998) and the dual mechanism (Pinker, 1999). Further evidence for associative and similarity-based L2 morphology in German comes from a study with English-speaking learners of German (Wagner, 2017). There, too, learners readily responded to nonce-verbs like spling with splang

44

T. WAGNER

and splung, thus revealing similarity-based effects for [I-a] apophony and prototypical schemas like [∫(C) _ I _ ï(k)], [# _ I _ ï(k)], and also [C_ fricative]; segments with the velar nasal favoured [a]-apophony, while [s/ ç] favoured [o:/O]. In this German L2 study, contrary to Penke (2006), both onset and coda showed pronounced prototypicality effects. And this study, too, provided evidence against the apophonic path. In the most recent study, Cuskley et al. (2015) could also show similarity-based effects for both speakers and learners of English. In sum, there is strong evidence that intermediate and advanced learners overirregularise non-default patterns under certain conditions. Learners are thus able to form non-default past tense—and partly participle—formations based on morphophonological analogies to existing verbs and verbal paradigms. Thus, their learner language must contain enough exemplars of non-default verbs for similarity-based analogical processing to be possible. They clearly are sensitive to the stored exemplars’ form and can use this sensitivity for analogy-based processing and generalisation. What we lack evidence for, however, is whether young EL2 learners, after just a few years of comparably limited exposure in instructed contexts, exhibit such generalisation properties. Accordingly, the research question for the empirical part of this volume asks in how far the organisation of semi-regular verbs in the mental lexicon of twelve- to 14-year-old learners of English is consistent with predictions made by associative, analogical single-route models. Two behavioural nonce-word elicitation tasks explored this question, testing three hypotheses. First, it was hypothesised that after three to four years of English, learners would be moderately sensitive to prototypical schemas and would use them to inflect given nonce-words in a non-default fashion. Second, it was hypothesised that there are both input- and product-oriented schemas in young learners’ mental lexicons, operating bi-directionally. The third hypothesis addresses the question of the extent to which the sensitivity to similaritybased schemas develops as a function of instructional input and exposure. That is why there is a small comparison group of 16-year-old learners, who went through the same experiments as the twelve- to 14-year-old learners.

3

IRREGULAR VERB MORPHOLOGY: L1 AND L2 COGNITIVE …

45

References Aarts, B. (2004). Modelling linguistic gradience. Studies in Language, 28(1), 1–49. https://doi.org/10.1075/sl.28.1.02aar Aha, D. W., Kibler, D., & Albert, M. K. (1991). Instance-based learning algorithms. Machine Learning, 6, 37–66. https://doi.org/10.1007/BF0015 3759 Aitchison, J. (1992). Good birds, better birds and amazing birds: The development of prototypes. In P. J. L. Arnaud & H. Bejoint (Eds.), Vocabulary and applied linguistics (pp. 71–84). Macmillan. Albright, A. (2009). Modeling analogy as probabilistic grammar. In J. Blevins & J. Blevins (Eds.), Analogy in grammar: Form and acquisition (pp. 185–213). Oxford University Press. Albright, A., & Hayes, B. (2003). Rules vs. analogy in English past tenses: A computational/experimental study. Cognition, 90(2), 119–161. https://doi. org/10.1016/S0010-0277(03)00146-X Ambridge, B. (2010). Children’s judgments of regular and irregular novel pasttense forms. New data on the English past-tense debate. Developmental Psychology, 46(6), 1497–1504. https://doi.org/10.1037/a0020668 Ambridge, B. (2020). Against stored abstractions. A radical exemplar model of language acquisition. First Language. Special Issue, 40(5–6), 509–559. https://doi.org/10.1177/0142723719869731 Arndt-Lappe, S., Plag, I., Koch, K., & Krott, M. (2018). Transparent Analogical Modelling (TrAML). GitHub. https://github.com/SabineArndtLappe/ TrAML Augst, G. (1975). Untersuchungen zum Morpheminventar der deutschen Gegenwartssprache. Gunter Narr. Baayen, R. H. (2003). Probabilistic approaches to morphology. In R. Bod, J. Hay, & S. Jannedy (Eds.), Probabilistic linguistics (pp. 229–288). MIT Press. Baayen, R. H. (2007). Storage and computation in the mental lexicon. In G. Jarema & G. Libben (Eds.), The mental lexicon: Core perspectives (pp. 81– 104). Elsevier. Baayen, R. H., & del Prado Martín, F. M. (2005). Semantic density and pasttense formation in three Germanic languages. Language, 81(3), 666–698. https://doi.org/10.1353/lan.2005.0112 Baayen, R. H., Dijkstra, T., & Schreuder, R. (1997). Singulars and plurals in Dutch. Evidence for a dual parallel route model. Journal of Memory and Language, 37 (1), 94–117. https://doi.org/10.1006/jmla.1997.2509 Baayen, R. H., & Hay, J. B. (2005). Shifting paradigms: Gradient structure in morphology. Trends in Cognitive Sciences, 9(7), 342–348. https://doi.org/ 10.1016/j.tics.2005.04.002

46

T. WAGNER

Baayen, R. H., & Schreuder, R. (1999). War and peace: Morphemes and full forms in a noninteractive activation parallel dual-route model. Brain and Language, 68(1–2), 27–32. https://doi.org/10.1006/brln.1999.2069 Beach, L. R. (1964). Cue probabilism and inference behavior. Psychological Monographs: General and Applied, 78(5), 1–20. https://doi.org/10.1037/h00 93853 Bechtel, W., & Abrahamsen, A. (1990). Connectionism and the mind. An introduction to parallel processing in networks. Blackwell. Beck, M. L. (1997). Regular verbs, past tense and frequency. Tracking down a potential source of NS/NNS competence differences. Second Language Research, 13(2), 93–115. Becker, T. (1990). Analogie und morphologische Theorie. Wilhelm Fink. Behrens, H. (2017). The role of analogy in language processing and acquisition. In M. Hundt, S. Mollin, & S. Pfenninger (Eds.), The changing English language: Psycholinguistic perspectives (pp. 215–239). Cambridge University Press. https://doi.org/10.1017/9781316091746.010 Bendjaballah, S., & Haiden, M. (2002). Meaningful vowels. In P. Bottari (Ed.), Proceedings of the XXVIII Incontro di Grammatica Generativa. Congedo Editore. Bendjaballah, S., & Haiden, M. (2005). Die Grammatik der Präpositionen im Berber. Folia Linguistica, 39(3–4), 319–361. Berg, T. (2012). The cohesiveness of English and German compounds. The Mental Lexicon, 7 (1), 1–33. https://doi.org/10.1075/ml.7.1.01ber Berko, J. (1958). The child’s learning of English morphology. Word, 14, 150– 177. https://doi.org/10.1080/00437956.1958.11659661 Birdsong, D., & Flege, J. E. (2001). Regular-irregular dissociations in L2 acquisition of English morphology. In A. H. J. Do, L. Domínguez, & A. Johansen (Eds.), BUCLD 25. Proceedings of the 25th annual Boston university conference on language development (vol. 1, pp. 123–132). Cascadilla Press. Bittner, A. (1996). Starke ‚schwache’ Verben, schwache ‚starke’ Verben. Deutsche Verbflexion und Natürlichkeit. Stauffenburg. Blackwell, A., & Broeder, P. (1992, May). Interference and facilitation in SLA: A connectionist perspective. Seminar on Parallel Distributed Processing and Natural Language Processing, San Diego, UCSD. Blything, R. P., Ambridge, B., & Lieven, E. V. (2018). Children’s acquisition of the English past-tense: Evidence for a single-route account from novel verb production data. Cognitive Science, 42(2), 621–639. https://doi.org/ 10.1111/cogs.12581 Bod, R. (2009). From exemplar to grammar: Integrating analogy and probability in language learning. Cognitive Science, 33(5), 752–793. https://doi.org/10. 1111/j.1551-6709.2009.01031.x

3

IRREGULAR VERB MORPHOLOGY: L1 AND L2 COGNITIVE …

47

Bod, R., Hay, J., & Jannedy, S. (Eds.). (2003). Probabilistic linguistics. MIT Press. Bolinger, D. L. (1977). Meaning and form. Longman. Bolinger, D. L. (1992). About furniture and birds. Cognitive Linguistics, 3(1), 111–117. https://doi.org/10.1515/cogl.1992.3.1.0 Bompolas, S., Marzi, C., Pirrelli, V., Ferro, M., & Cardillo, F. A. (2017). For a performance-oriented notion of regularity in inflection: The case of modern Greek conjugation. Italian Journal of Computational Linguistics, 3(1), 77–92. https://doi.org/10.4000/ijcol.435 Boyé, G. (2000). Problèmes de morpho-phonologie verbale en français, en espagnol et en italien. Linguistique (Doctoral dissertation). Université Paris-Diderot, Paris VII. Thèse de doctorat en Linguistique. https://theses.hal.science/tel00276756/document Bozic, M., Fonteneau, E., Su, L., & Marslen-Wilson, W. D. (2014). Grammatical analysis as a distributed neurobiological function. Human Brain Mapping, 36(3), 1190–1201. https://doi.org/10.1002/hbm.22696 Broeder, P., & Plunkett, K. (1997). Connectionism and second language acquisition. In N. C. Ellis (Ed.), Implicit and explicit learning of language (pp. 421–454). Harcourt Brace Jovanovich. Butler, R., Patterson, K., & Woollams, A. M. (2012). In search of meaning: Semantic effects on past-tense inflection. The Quarterly Journal of Experimental Psychology, 65(8), 1633–1656. https://doi.org/10.1080/17470218. 2012.661441 Bybee, J. (1988). Morphology as lexical organisation. In M. Hammond & M. Noonam (Eds.), Theoretical approaches to morphology (pp. 119–141). Academic Press. Bybee, J. (1995). Regular morphology and the lexicon. Language and Cognitive Processes, 10(5), 425–455. https://doi.org/10.1080/01690969508407111 Bybee, J. (2001). Phonology and language use. Cambridge University Press. Bybee, J. (2010). Language, usage and cognition. Cambridge University Press. Bybee, J., & Eddington, D. (2006). A usage-based approach to Spanish verbs of ‘becoming.’ Language, 82(2), 323–355. https://doi.org/10.1353/lan.2006. 0081 Bybee, J., & Moder, C. L. (1983). Morphological classes as natural categories. Language, 59(2), 251–270. https://doi.org/10.2307/413574 Bybee, J., & Slobin, D. I. (1982). Rules and schemas in the development and use of the English past tense. Language, 58(2), 265–289. https://doi.org/ 10.2307/414099 Caramazza, A., Laudanna, A., & Cristina, R. (1988). Lexical access and inflectional morphology. Cognition, 28(3), 297–332. https://doi.org/10.1016/ 0010-0277(88)90017-0

48

T. WAGNER

Chandler, S. (1994, April). An exemplar-based approach to language acquisition. Paper presented at the Workshop on Cognitive Models of Language Acquisition, University of Tilburg, The Netherlands. Chandler, S. (2017). The analogical modelling of linguistic categories. Language and Cognition, 9(1), 52–87. https://doi.org/10.1017/langcog.2015.24 Chater, N., Tenenbaum, J. B., & Yuille, A. (2006). Probabilistic models of cognition: Conceptual foundations. Trends in Cognitive Science, Special Issue, 10(7), 287–291. https://doi.org/10.1016/j.tics.2006.05.007 Chialant, D., & Caramazza, A. (1995). Where is morphology and how is it processed? The case of written word recognition. In L. B. Feldman (Ed.), Morphological aspects of language processing (pp. 55–76). Lawrence Erlbaum Associates. Chomsky, N., & Halle, M. (1968). The sound pattern of English. Harper & Row. Christiansen, M. H., & Chater, N. (1999). Connectionist natural language processing. The state of the art. Cognitive Science, 23(4), 417–437. https:// doi.org/10.1016/S0364-0213(99)00010-5 Clahsen, H. (1997). The representation of German participles in the German mental lexicon: Evidence for the dual-mechanism model. In G. Booij & J. v. Marle (Eds.), Yearbook of morphology 1996 (pp. 73–96). Kluwer Academic. Clahsen, H. (1999). Lexical entries and rules of language: A multidisciplinary study of German inflection. Behavioral and Brain Sciences, 22(6), 991–1060. https://doi.org/10.1017/S0140525X99002228 Clahsen, H., Felser, C., Neubauer, K., Sato, M., & Silva, R. (2010). Morphological structure in native and nonnative language processing. Language Learning, 60, 21–43. https://doi.org/10.1017/S1366728912000648 Cuskley, C., Colaiori, F., Castellano, C., Loreto, V., Pugliese, M., & Tria, F. (2015). The adoption of linguistic rules in native and non-native speakers: Evidence from a Wug task. Journal of Memory and Language, 84, 205–223. https://doi.org/10.1016/j.jml.2015.06.005 Daelemans, W., Gillis, S., & Durieux, G. (1994). The acquisition of stress: A data-oriented approach. Computational Linguistics, 20, 421–451. Daelemans, W., Zavrel, J., van der Sloot, K., & van den Bosch, A. (1999). TiMBL: Tilburg Memory Based Learner, version 2.0, reference guide. Induction of linguistic knowledge technical report. ILK Research Group. Daelemans, W., & van den Bosch, A. (2005). Memory-based language processing. Studies in natural language processing. Cambridge University Press. Dawdy-Hesterberg, L. G., & Pierrehumbert, J. B. (2014). Learnability and generalisation of Arabic broken plural nouns. Language, Cognition and Neuroscience, 29(10), 1268–1282. https://doi.org/10.1080/23273798. 2014.899377

3

IRREGULAR VERB MORPHOLOGY: L1 AND L2 COGNITIVE …

49

DeKeyser, R. (1995). Learning second language grammar rules: An experiment with a miniature linguistic system. Studies in Second Language Acquisition, 17 (3), 379–410. https://doi.org/10.1017/S027226310001425X Derwing, B. L., & Skousen, R. (1994). Productivity and the English past tense. Testing Skousen’s analogy model. In S. D. Lima, R. L. Corrigan, & G. K. Iverson (Eds.), The reality of linguistic rules (pp. 193–218). John Benjamins. Desai, R., Conant, L. L., Waldron, E., & Binder, J. R. (2006). FMRI of past tense processing. The effects of phonological complexity and task difficulty. Journal of Cognitive Neuroscience, 18(2), 278–297. https://doi.org/ 10.1162/089892906775783633 Diessel, H. (2019). The grammar network. How linguistic structure is shaped by language use. Cambridge University Press. https://doi.org/10.1017/978110 8671040 Divjak, D., & Arppe, A. (2013). Extracting prototypes from exemplars. What can corpus data tell us about concept representation? Cognitive Linguistics, 24(2), 221–274. https://doi.org/10.1515/cog-2013-0008 Divjak, D., & Milin, P. (2023). Using computational cognitive modeling in usage-based linguistic. In M. Diaz-Campos & S. Balasch (Eds.), The handbook of usage-based linguistics. Wiley Blackwell. Eddington, D. (2000). Analogy and the dual-route model of morphology. Lingua, 110(4), 281–298. https://doi.org/10.1016/S0024-3841(99)000 43-1 Eddington, D. (2004). Issues in modelling language processing analogically. Lingua, 114(7), 849–871. https://doi.org/10.1016/S0024-3841(03)000 63-9 Ellis, N. C. (2002). Frequency effects in language processing. A review with implications for theories of implicit and explicit language acquisition. Studies in Second Language Acquisition, 24(2), 143–188. https://doi.org/10.1017/ S0272263102002024 Elman, J. L. (1991). Distributed representations, simple recurrent networks, and grammatical structure. Machine Learning, 7 (2), 195–225. https://doi.org/ 10.1007/BF00114844 Engelmann, F., Granlund, S., Kolak, J., Szreder, M., Ambridge, B., Pine, J., Theakston, A., & Lieven, E. (2019). How the input shapes the acquisition of verb morphology: Elicited production and computational modelling in two highly inflected languages. Cognitive Psychology, 110, 30–69. https://doi.org/ 10.1016/j.cogpsych.2019.02.001 Ernestus, M., & Baayen, R. H. (2003). Predicting the unpredictable: Interpreting neutralized segments in Dutch. Language, 79(1), 5–38. https://doi. org/10.1353/lan.2003.0076

50

T. WAGNER

Fanselow, G., Féry, C., Schlesewsky, M., & Vogel, R. (Eds.). (2006). Gradience in grammar: Generative perspectives (pp. 185–204). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199274796.003.0010 Faroqi-Shah, Y. (2013). Selective treatment of regular versus irregular verbs in agrammatic aphasia. Efficacy data. Aphasiology, 27 (6), 678–705. https://doi. org/10.1080/02687038.2012.751577 Feldman, J. A. (1985). Connectionist models and their applications: Introduction. Cognitive Science. A Multidisciplinary Journal, 9(1), Special Issue, 1–2. https://doi.org/10.1207/s15516709cog0901_1 Feldman, L. B., Kosti´c, A., Basnight-Brown, D. M., Durd-evi´c, D. F., & Pastizzo, M. J. (2010). Morphological facilitation for regular and irregular verb formations in native and non-native speakers: Little evidence for two distinct mechanisms. Bilingualism: Language and Cognition, 1(13), 119–135. https://doi.org/10.1017/S1366728909990459 Flege, J. E., Yeni-Komshian, G. H., & Liu, S. (1999). Age constraints on secondlanguage acquisition. Journal of Memory and Language, 41(1), 78–104. https://doi.org/10.1006/jmla.1999.2638 Gasser, M. (1990). Connectionism and universals of second language acquisition. Studies in Second Language Acquisition, 12(2), 179–199. https://doi.org/10. 1017/S0272263100009074 Geeraerts, D. (1988). Where does prototypicality come from? In B. RudzkaOstyn (Ed.), Topics in cognitive linguistics (pp. 207–229). John Benjamins. Geeraerts, D. (1989). Introduction. Prospects and problems of prototype theory. Linguistics, 27 , 587–612. https://doi.org/10.1515/ling.1989.27.4.587 Gerber, P. (2022). Verb stem alternation in Gongduk. Synchronic and diachronic analysis. Language and Linguistics, 23(2), 240–273. https://doi.org/10. 1075/lali.00105.ger Giraudo, H., Dal Maso, S. (2018). Towards a constructional approach of L2 morphological processing. In G. Booij (Ed.), The construction of words. Studies in morphology (vol. 4, pp. 603–622). Springer. https://doi.org/10.1007/ 978-3-319-74394-3_21 Givón, T. (1986). Prototypes: Between Plato and Wittgenstein. In C. Craig (Ed.), Noun classes and categorization. Proceedings of a symposium on categorization and noun classification (pp. 77–102). John Benjamins. Godfroid, A. (2015). The effects of implicit instruction on implicit and explicit knowledge development. Studies in Second Language Acquisition, 38(S2), 177–215. https://doi.org/10.1017/S0272263115000388 Godfroid, A., & Uggen, M. (2013). Attention to irregular verbs by beginning learners of German. An eye-movement study. Studies in Second Language Acquisition, 35(2), 291–322. https://doi.org/10.1017/S02722 63112000897

3

IRREGULAR VERB MORPHOLOGY: L1 AND L2 COGNITIVE …

51

Gor, K., & Chernigovskaya, T. (2003). Mental lexicon structure in L1 and L2 acquisition: Russian evidence. Glossos, 4, 1–31. Gor, K., & Chernigovskaya, T. (2005). Formal instruction and the acquisition of verbal morphology. In A. Housen & M. Pierrard (Eds.), Investigations in instructed second language acquisition (pp. 131–166). Mouton de Gruyter. Guerssel, M. & Lowenstamm, J. (1993). Classical Arabic apophony. Ms., UQUAM & Université Paris 7. Guerssel, M. & Lowenstamm, J. (1996). Ablaut in classical Arabic measure 1 active verbal forms. In J. Lecarme, J. Lowenstamm & Ur Shlonsky (Eds.), Studies in Afroasiatic grammar (pp. 123–134). Holland Academic Graphics. Hahn, U., & Nakisa, R. C. (2000). German inflection: Single route or dual route? Cognitive Psychology, 41(4), 313–360. https://doi.org/10.1006/cogp. 2000.0737 Hahne, A., Müller, J., & Clahsen, H. (2006). Morphological processing in a second language: Behavioural and event-related potential evidence for storage and decomposition. Journal of Cognitive Neuroscience, 18(1), 121–134. https://doi.org/10.1162/089892906775250067 Hampton, J. A. (1998). Similarity-based categorization and fuzziness of natural categories. Cognition, 65(2–3), 137–165. https://doi.org/10.1016/S00100277(97)00042-5 Hawkins, J., & George, D. (2006). Hierarchical temporal memory. Concepts, theory and terminology. Numenta Technology, 3(27), 1–20. http://www.num enta.com/technology.php Hay, J., & Baayen, R. H. (2005). Shifting paradigms: Gradient structure in morphology. Trends in Cognitive Sciences, 9(7), 342–348. https://doi.org/ 10.1016/j.tics.2005.04.002 Hickey, R. J., & Martin, R. G. (2001). An instance-based approach to pattern association learning with application to the English past tense domain. Knowledge-Based Systems, 14(3–4), 131–136. https://doi.org/10. 1016/S0950-7051(01)00089-2 Hock, H. H. (1991). Principles of historical linguistics. Walter de Gruyter. Itkonen, E. (2005). Analogy as structure and process. Approaches in linguistics, cognitive psychology and philosophy of science. John Benjamins. https://doi.org/ 10.1075/hcp.14 Jespersen, O. (1942). A modern English grammar on historical principles. Part IV: Morphology. Munksgaard. Joanisse, M. F., & McClelland, J. L. (2015). Connectionist perspectives on language learning, representation and processing. WIREs Cognitive Science, 6(3), 235–247. https://doi.org/10.1002/wcs.1340 Joanisse, M. F., & Seidenberg, M. S. (1999). Impairments in verb morphology after brain injury: A connectionist model. Proceedings of the National Academy

52

T. WAGNER

of Sciences of the United States of America, 96(13), 7592–7597. https://doi. org/10.1073/pnas.96.13.7592 Jones, D. (1996). Analogical natural language processing. UCL Press. Jurafsky, D. (2002). Probabilistic modeling in psycholinguistics: Linguistic comprehension and production. In R. Bod, J. Hay, & S. Jannedy (Eds.), Probabilistic linguistics (pp. 39–96). MIT Press. Justus, T., Larsen, J., de Mornay Davies, P., & Swick, D. (2008). Interpreting dissociations between regular and irregular past-tense morphology: Evidence from event-related potentials. Cognitive, Affective, and Behavioral Neuroscience, 8(2), 178–194. https://doi.org/10.3758/cabn.8.2.178 Kaye, J. D., Lowenstamm, J., & Vergnaud, J. R. (1985). The internal structure of phonological representations: A theory of charm and government. Phonology, 2(1), 305–328. https://doi.org/10.1017/S0952675700000476 Keuleers, E., Sandra, D., Daelemans, W., Gillis, S., Durieux, G., & Martens, E. (2007). Dutch plural inflection: The exception that proves the analogy. Cognitive Psychology, 54(4), 283–318. https://doi.org/10.1016/j.cogpsych. 2006.07.002 Kielar, A., Joanisse, M. F., & Hare, M. L. (2008). Priming English past tense verbs: Rules or statistics? Journal of Memory and Language, 58(2), 327–346. https://doi.org/10.1016/j.jml.2007.10.002 Kırkıcı, B. (2007). The mental processing of L2 English lexical compounds: A developmental Dual-Mechanism account. In L. Roberts, A. Gürel, S. Tatar, & L. Marti (Eds.), Eurosla-Yearbook, 7 (pp. 7–25). John Benjamins. https:// doi.org/10.1075/eurosla.7.03kir Kirov, C., & Cotterell, R. (2018). Recurrent neural networks in linguistic theory: Revisiting Pinker and Prince (1988) and the past tense debate. Transactions of the Associations for Computational Linguistics, 6, 651–665. https://doi.org/ 10.1162/tacl_a_00247 Kodner, J. (2022). Computational models of morphological learning. In Oxford research encyclopaedia of linguistics. https://doi.org/10.1093/acrefore/978 0199384655.013.378 Köpcke, K. M. (1998). Prototypisch starke und schwache Verben in der deutschen Gegenwartssprache. Sonderheft Germanistische Linguistik, 141–142, 45–60. Krause, H., Bosch, S., & Clahsen, H. (2015). Morphosyntax in the bilingual mental lexicon: An experimental study of strong stems in German. Studies in Second Language Acquisition, 37 (4), 597–621. https://doi.org/10.1017/ S0272263114000564 Krott, A., Baayen, R. H., & Schreuder, R. (2001). Analogy in morphology. Modeling the choice of linking morphemes in Dutch. Linguistics, 39(1), 51–93. https://doi.org/10.1515/LING.2001.008

3

IRREGULAR VERB MORPHOLOGY: L1 AND L2 COGNITIVE …

53

Krott, A., Schreuder, R., Baayen, R. H., & Dressler, W. U. (2007). Analogical effects on linking elements in German compound words. Language and Cognitive Processes, 22(1), 25–57. https://doi.org/10.1080/016909605003 43429 Labov, W. (1973). The boundaries of words and their meanings. In C. J. N. Bailey & R. W. Shuy (Eds.), New ways of analyzing variation in English (pp. 340–373). Georgetown University Press. Leminen, A., & Clahsen, H. (2014). Brain potentials to inflected adjectives: Beyond storage and decomposition. Brain Research, 16(1543), 223–234. https://doi.org/10.1016/j.brainres.2013.10.038 Leminen, A., Smolka, E., Duñabeitia, J. A., & Pliatsikas, C. (2019). Morphological processing in the brain: The good (inflection), the bad (derivation) and the ugly (compounding). Cortex, 116, 4–44. https://doi.org/10.1016/j.cor tex.2018.08.016 López Rúa, P. (2003). Birds, colours and prepositions. The theory of categorization and its applications in linguistics. Lincom. Mangasser-Wahl, M. (Ed.). (2000). Prototypentheorie in der Linguistik. Anwendungsbeispiele, Methodenreflexion, Perspektiven. Stauffenburg. Marchand, H. (1969). The categories and types of present-day English wordformation. Beck. Martinet, A. (1955). Économie des changements phonétiques. Traité de phonologie diachronique. Francke. Marusch, T., Jäger, L. A., Neiß, L., Burchert, F., & Nickels, L. (2019). Overt language production of German past participles: Investigating (ir-)regularity. Language, Cognition and Neuroscience, 34(3), 289–308. https://doi.org/10. 1080/23273798.2018.1527936 McClelland, J. L., & Patterson, K. (2002). Rules or connections in past-tense inflections: What does the evidence rule out? Trends in Cognitive Sciences, 6(11), 465–472. https://doi.org/10.1016/s1364-6613(02)01993-9 Milin, P., Tucker, B. V., & Divjak, D. (2023). A learning perspective on the emergence of abstractions. The curious case of phone(me)s. Language and Cognition, 1–23. https://doi.org/10.1017/langcog.2023.11 Mirkovic, J., Seidenberg, M. S., & Joanisse, M, F. (2011). Rules vs. statistics. Insights from a highly inflected language. Cognitive Science, 35(4), 638–681. https://doi.org/10.1111/j.1551-6709.2011.01174.x Naranjo, M. G. (2019). Analogical classification in formal grammar. Empirically oriented theoretical morphology and syntax (Vol. 5). Language Science Press. Nelson, R. (2013). Expanding the role of connectionism in SLA theory. Language Learning, 63(1), 1–33. https://doi.org/10.1111/j.1467-9922. 2012.00699.x

54

T. WAGNER

Newport, E. L. (2016). Statistical language learning: Computational, maturational, and linguistic constraints. Language and Cognition, 8(S3), 447–461. https://doi.org/10.1017/langcog.2016.20 Nicoladis, E., & Paradis, J. (2012). Acquiring regular and irregular past tense morphemes in English and French. Evidence from bilingual children. Language Learning, 62(1), 170–197. https://doi.org/10.1111/j. 1467-9922.2010.00628.x Niu, D., Yang, L., Cai, T., Li, L., Wu, X., & Wang, Z. (2022). A new hierarchical temporal memory algorithm based on activation intensity. Computational Intelligence and Neuroscience. https://doi.org/10.1155/2022/6072316 Nosofsky, R. M. (1990). Relations between exemplar-similarity and likelihood models of classification. Journal of Mathematical Psychology, 34(4), 393–418. https://doi.org/10.1016/0022-2496(90)90020-A Panther, K. U., & Köpcke, K. M. (2008). A prototype approach to sentences and sentence types. Annual Review of Cognitive Linguistics, 6, 83–112. https:// doi.org/10.1075/arcl.6.05pan Paradis, J., Nicoladis, E., Crago, M., & Genesee, F. (2011). Bilingual children’s acquisition of the past tense: a usage-based approach. Journal of Child Language, 38(3), 554–578. https://doi.org/10.1017/S0305000910000218 Penke, M. (2006). Flexion im mentalen Lexikon. Linguistische Arbeiten. Niemeyer. Pinker, S. (1999). Words and rules. The ingredients of language. Phoenix. Pinker, S., & Prince, A. (1988). On language and connectionism. Analysis of parallel distributed processing of language acquisition. In S. Pinker & J. Mehler (Eds.), Connections and symbols (pp. 73–193). MIT Press. Pinker, S., & Ullman, M. T. (2002). Combination and structure, not gradedness, is the issue. Trends in Cognitive Sciences, 6, 472–474. https://doi.org/10. 1016/S1364-6613(02)02013-2 Plag, I. (2000). Irregular past tense formation in English interlanguage. In I. Plag & K. P. Schneider (Eds.), Language use, language acquisition, and language history. (Mostly) Empirical studies in honour of Rüdiger Zimmermann (pp. 134–149). Wissenschafticher Verlag. Plag, I., Kawaletz, L. Arndt-Lappe, S., & Lieber, R. (2023). Analogical modeling of derivational semantics. Two case studies. In S. Kotowski & I. Plag (Eds.), The semantics of derivational morphology. Theory, methods, evidence (pp. 103– 141). De Gruyter. https://doi.org/10.1515/9783111074917 Pliatsikas, C., & Marinis, T. (2013). Processing of regular and irregular past tense morphology in highly proficient second language learners of English: a selfpaced reading study. Applied Psycholinguistics, 34(5), 943–970. https://doi. org/10.1017/S0142716412000082

3

IRREGULAR VERB MORPHOLOGY: L1 AND L2 COGNITIVE …

55

Plunkett, K., & Bandelow, S. (2006). Stochastic approaches to understanding dissociations in inflectional morphology. Brain and Language, 98(2), 194– 209. https://doi.org/10.1016/j.bandl.2006.04.014 Plunkett, K., & Marchman, V. (1993). From rote learning to system building: Acquiring verb morphology in children and connectionist nets. Cognition, 48(1), 21–69. https://doi.org/10.1016/0010-0277(93)90057-3 Post, B., Marslen-Wilson, W. D., Randall, B., & Tyler, L. K. (2008). The processing of English regular inflections: Phonological cues to morphological structure. Cognition, 109(1), 1–17. https://doi.org/10.1016/j.cognit ion.2008.06.011 Prasada, S., & Pinker, S. (1993). Generalization of regular and irregular morphological patterns. Language and Cognitive Processes, 8, 1–56. https://doi.org/ 10.1080/0169096930840694 Racz, P., Beckner, C., Hay, J. B., & Pierrehumbert, J. B. (2014). Rules, analogy and social factors co-determine past-tense formation patterns in English. In Ö. Çetino˘glu, J. Heinz, A. Maletti, & J. Riggle (Eds.), Proceedings of the joint workshop between the Sigmorphon and Sigfsm Morphfsm (pp. 55–63). Association for Computational Linguistics. Reed, S. K. (1972). Pattern recognition and categorization. Cognitive Psychology, 3(3), 382–407. https://doi.org/10.1016/0010-0285(72)90014-x Rogers, J., Révész, A., & Rebuschat, P. (2016). Implicit and explicit knowledge of inflectional morphology. Applied Psycholinguistics, 37 (4), 781–812. https://doi.org/10.1017/S0142716415000247 Romano, F. (2015). Salience and proficiency as determinants in the acquisition of L2 morphology. Italian Journal of Linguistics, 27 (2), 75–110. Romberg, A. R., & Saffran, J. R. (2010). Statistical learning and language acquisition. WIREs: Wiley Interdisciplinary Reviews: Cognitive Science, 1(6), 906–914. https://doi.org/10.1002/wcs.78 Roncaglia-Denissen, M. P., & Kotz, S. A. (2016). What does neuroimaging tell us about morphosyntactic processing in the brain of second language learners? Review article. Bilingualism: Language and Cognition, 19(4), 665–673. https://doi.org/10.1017/S1366728915000413 Rosch, E. (1973). Natural categories. Cognitive Psychology, 4(3), 328–350. https://doi.org/10.1016/0010-0285(73)90017-0 Rosch, E. (1975). Cognitive representations of semantic categories. Journal of Experimental Psychology: General, 104(3), 192–232. doi.apa.org/journals/ xge/104/3/192.pdf Rosch, E., & Mervis, C. B. (1975). Family resemblance: Studies in the internal structure of categories. Cognitive Psychology, 7 , 573–605. https://doi.org/ 10.1016/0010-0285(75)90024-9 Rumelhart, D. E., & McClelland, J. L. (1986). On learning the past tenses of English verbs. In D. E. Rumelhart, J. L. McClelland & the PDP

56

T. WAGNER

Research Group (Eds.), Parallel distributed processing: Explorations in the microstructures of cognition (Vol. 2, pp. 216–271). MIT Press. Safaie, E. (2021). Sensitivity to regular and irregular past tense morphology in native speakers and second language learners of English: Evidence from intermediate-to-advanced Persian speakers of L2 English. Journal of Psycholinguist Research, 50(5), 1107–1135. https://doi.org/10.1007/s10936-02109790-3 Schmid, H. J. (1998). Zum kognitiven Kern der Prototypentheorie. In F. Ungerer (Ed.), Kognitive Lexikologie und Syntax (pp. 9–28). Universitätsverlag. Schreuder, R., & Baayen, R. H. (1995). Modeling morphological processing. In L. B. Feldman (Ed.), Morphological aspects of language processing (pp. 131– 154). Lawrence Erlbaum Associates. Ségéral, P. (2000). Théorie de l´apophonie et organisation des schèmes en sémitique. In J. Lecarme, J. Lowenstamm, & U. Shlonsky (Eds.), Research in Afroasiatic grammar (pp. 263–299). John Benjamins. Ségéral, P., & Scheer, T. (1998). A generalized theory of ablaut: The case of modern German strong verbs. In R. Fabri, A. Ortmann, & T. Parodi (Eds.), Models of inflection (pp. 28–59). De Gruyter. https://doi.org/10.1515/978 3110919745.28 Seidenberg, M. S., & MacDonald, M. C. (1999). A probabilistic constraints approach to language acquisition and processing. Cognitive Science, 23(4), 569–588. https://doi.org/10.1016/S0364-0213(99)00016-6 Seidenberg, M. S., & Plaut, D. C. (2010). Progress in understanding word reading. Data fitting versus theory building. In S. Andrews (Ed.), From inkmarks to ideas. Current issues and lexical processing (pp. 55–79). Routledge. Seidenberg, M. S., & Plaut, D. C. (2014). Quasiregularity and its discontent: The legacy of the past tense debate. Cognitive Science, 38(6). https://doi. org/10.1111/cogs.12147 Shirai, Y. (2019). Connectionism and second language acquisition. Routledge. Skousen, R. (1989). Analogical modeling of language. Kluwer Academic. Skousen, R. (2005). Quantum analogical modeling: A general quantum computing algorithm for predicting language behavior. Papers from the 2007 AAAI Spring Symposium. https://arxiv.org/ftp/quant-ph/papers/ 0510/0510146.pdf Skousen, R. (2009). Expanding analogical modeling into a general theory of language prediction. In J. P. Blevins & J. Blevins (Eds.), Analogy in grammar. Form and acquisition (pp. 164–184). Oxford University Press. Slioussar, N., Kireev, M. V., Chernigovskaya, T. V., Kataeva, G. V., Korotkov, A. D., & Medvedev, S. V. (2014). An ER-fMRI study of Russian inflectional

3

IRREGULAR VERB MORPHOLOGY: L1 AND L2 COGNITIVE …

57

morphology. Brain and Language, 130, 33–41. https://doi.org/10.1016/j. bandl.2014.01.006 Sokolik, M. E. (1990). Learning without rules. PDP and a resolution of the adult language learning paradox. TESOL Quarterly, 24(4), 685–696. https://doi. org/10.2307/3587115 Sokolik, M. E., & Smith, M. E. (1992). Assignment of gender to French nouns in primary and secondary language: A connectionist model. Second Language Research, 8(1), 39–58. https://doi.org/10.1177/02676583920080 Strobach, T., & Schönpflug, U. (2011). Can a connectionist model explain the processing of regularly and irregularly inflected words in German as L1 and L2? International Journal of Bilingualism, 15(4), 446–465. https://doi.org/ 10.1177/1367006911403205 Tabak, W., Schreuder, R., & Baayen, R. H. (2010). Producing inflected verbs. A picture naming study. The Mental Lexicon, 5(1), 22–46. https://doi.org/ 10.1075/ml.5.1.02tab Taylor, J. R. (1995). Linguistic categorization. Prototypes in linguistic theory. Clarendon Press. Taylor, J. R. (2003). Linguistic categorization. Prototypes in linguistic Theory (2nd ed.). Oxford University Press. Ullman, M. T. (2001a). A neurocognitive perspective on language: The declarative/procedural model. Nature Reviews. Neuroscience, 2(10), 717–726. https://doi.org/10.1038/35094573 Ullman, M. T. (2001b). The declarative/procedural model of lexicon and grammar. Journal of Psycholinguistic Research, 30(1), 37–69. https://doi.org/ 10.1023/a:1005204207369 Ullman, M. T. (2004). Contributions of memory circuits to language: The declarative/procedural model. Cognition, 92(1–2), 231–270. https://doi. org/10.1016/j.cognition.2003.10.008 van Noord, R., & Spenader, J. K. (2015). Modeling the learning of the English past tense with memory-based learning. Computational Linguistics in the Netherlands Journal, 5, 65–80. https://clinjournal.org/clinj/article/view/58 Veríssimo, J., & Clahsen, H. (2014). Variables and similarity in linguistic generalization: Evidence from inflectional classes in Portuguese. Journal of Memory and Language, 76, 61–79. https://doi.org/10.1016/j.jml.2014.06.001 Vosters, R. (2012). Geolinguistic data and the past tense debate. Linguistic and extralinguistic aspects of Dutch verb regularization. In G. De Vogelaer & G. Seiler (Eds.), The dialect laboratory: Dialects as a testing ground for theories of language change (pp. 227–248). John Benjamins. Wagner, T. (2010). Interlanguage morphology. Irregular verbs in the mental lexicon of German-English interlanguage speakers. Narr-Francke-Attempto. Wagner, T. (2017). L2 irregular verb morphology: Exploring behavioral data from intermediate English learners of German as a foreign language using

58

T. WAGNER

generalized mixed effects models. Studies in Second Language Learning and Teaching, 7 (3), 535–556. https://doi.org/10.14746/ssllt.2017.7.3.9 Wanner, D. (2006). The power of analogy. An essay on historical linguistics. Trends in Linguistics. Studies and Monographs 170. Mouton de Gruyter. Westermann, G. (2000). Constructivist neural network models of cognitive development (Doctoral thesis). Division of Informatics, University of Edinburgh, Edinburgh, UK. Westermann, G. Kovic, V., & Ruh, N. (2008). Mechanisms of verb inflection— Regular vs. irregular or easy vs. hard? In B. C. Love, K. McRae, & V. M. Sloutsky (Eds.), Proceedings of the annual meeting of the Cognitive Science Society (Vol. 30, pp. 739–744). Cognitive Science Society. https://escholars hip.org/content/qt4zq8p3xn/qt4zq8p3xn.pdf Wunderlich, D., & Fabri, R. (1995). Minimalist morphology: An approach to inflection. Zeitschrift für Sprachwissenschaft, 14(2), 236–294. https://doi. org/10.1515/zfsw.1995.14.2.236 Zaretsky, E., Lange, B., Euler, H., & Neumann, K. (2016). Factors considered and ignored in plural acquisition: Frequency rules? Language and Cognition, 8(2), 283–313. https://doi.org/10.1017/langcog.2014.51

CHAPTER 4

The Empirical Study

Abstract This chapter presents two psycholinguistic experiments exploring irregular verb morphology in the mental lexicon of young German-speaking learners of English. First, the general method and experimental design are summarised. This includes test item creation based on existing evidence from the pertinent literature. Then, experiment 1, a behavioural single-word elicitation study, is introduced, with materials, procedure, sample, data coding, results, and a brief discussion. Afterwards, experiment 2, a behavioural multi-word elicitation study follows the same structure as experiment 1. Bridging into implications for the classroom, this chapter closes with a general discussion of the empirical evidence and its relevance for EL2 teaching and learning. Keywords Nonce-words · Repeated-measure design · Mixed regression modelling · Conditional inference tree · Random forest analysis

4.1

General Method and Design

The aim of the present empirical study is to explore A2 learners’ sensitivity to morphophonological details of English irregular verbs and their capabilities to generalise over this knowledge in both an inputand product-oriented fashion. By means of two paper-and-pencil tasks © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 T. Wagner, English Interlanguage Morphology, https://doi.org/10.1007/978-3-031-50617-8_4

59

60

T. WAGNER

with nonce-word stimuli, grammatical inflections were elicited from the learners (Berko, 1958; Bybee & Moder, 1983; Lemhöfer & Radach, 2009; Plag, 2000; Prasada & Pinker, 1993; Wagner, 2010). In experiment 1, eight nonce-words along with four distractors, embedded in a specially prepared reading and listening task, had to be inflected for past tense. In the second experiment, 16 sentence completion tasks (Marusch et al., 2012) prompted pupils to inflect given infinitives for past tense and participle as well as given participles for past tense and infinitive (Albright & Hayes, 2003; Gries et al., 2005). This task tested if pupils could produce apophonic vowel changes in both directions, from given base form (infinitive) to inflected forms as well as from given output form (participle) to past tense inflection and base form (infinitive). Such backward formations were supposed to test the learners’ product-oriented generalisations and in turn the psycholinguistic reality of the apophonic path’s directionality.

4.2

Creation of Test Items

The test items in experiments 1 and 2 were designed to be highly prototypical triggers for similarity-based, analogical processing. The starting points for the creation of the test items were prototypicality effects reported for certain morphophonological schemas (Bybee & Moder, 1983; Bybee & Slobin, 1982; Plag, 2000; Prasada & Pinker, 1993; Wagner, 2010). Thus, prototypical constituents, such as [ŋ], [ŋk], [m], [n], [g], or [k] as well as schemas such as [sCC _ I _ ŋ] were operationalised as templates for the creation of new nonce-verbs. Unlike the classical wug-study, and unlike some previous studies (Cuskley et al., 2015), nonce-verb selection centred around prototypical segments mostly. Thus, no distance metric, like in Nerbonne and Heeringa (1997), for instance, was employed. That is why the nucleus, a segment with a low cue validity, was limited to the most prototypical input-oriented schema segment [I] in experiment 1. Moreover, in Bybee and Moder’s (1983) experiment, English speakers systematically avoided no-change responses when producing apophonic past tense forms including [2] and [æ]. Thus, output-oriented schemas containing either [2] or [æ] would disproportionally increase inflections with other vowels. In other words, nonce-verbs with [2] could artificially inflate responses with [æ], biasing the results in favour of one particular apophonic pattern.

4

THE EMPIRICAL STUDY

61

In a second step, the most prototypical schemas from these studies were cross-checked with 79 irregular verbs from a widely used EFL coursebook (Gerngross et al., 2017). Checks revealed that in these materials irregular verbs in past simple and present perfect had been explicitly and repeatedly addressed. This step was supposed to make sure that the syllable segments in the test items would have been represented in the book’s irregular verb inventory and that the learners would have very likely been exposed to these schemas in their instructed input. Interestingly, the phonotactically most complex onset structure [s CC] was the only one that did not occur in the course materials; it was included nevertheless, since it had a high cue validity in various behavioural studies (Plag, 2000; Wagner, 2010), and since consonant clusters with extrasyllabic s are acoustically salient, rendering them sensitive to frequency effects (Marecka & Dziubalska-Kołaczyk, 2014). Thus, the prototypical constituents eventually taken for test item creation were CC, s C, s CC, [I], [m], [t], [ŋ], and [ŋk]. Since not every combination of onset, nucleus, and coda is represented in the course book, and since the combinations of the three constituents have unequal observations or sample sizes, the present experiments have a non-orthogonal and unbalanced design.

4.3

Experiment 1

Materials and Procedure Experiment 1 contained a listening and reading comprehension task modelled on materials found in many common European EL2 textbooks for lower secondary schools. The actual task was preceded by a survey on pupils’ biodata, their appreciation of the current course book, their school type, gender, and L1s. Some of these variables served as covariates in the statistical analyses. The main purpose of the experiment’s introduction was to reinforce the illusion that the entire procedure was a trial run for future classroom material design, and that the pupils would be expert reviewers, as it were, of this new material. Introduction, instructions, biodata, and the actual listening and reading comprehension task lasted around 15 minutes. Prior to the actual listening and reading comprehension task, two items for practice demonstrated what the pupils were supposed to do. Then, pupils had to listen to a travel diary about a Madagascar trip. They were

62

T. WAGNER

supposed to follow the transcribed text in the test sheet and simultaneously inflect 16 given verbs for past tense, eight of which were real test items. Making this task a combination of listening comprehension and language in use with accompanying audio was supposed to both standardise the timing for gap filling across all participants (5 seconds) and disambiguate the eight nonce-words’ phonetic form. All test items were presented to the pupils in complementary sequences in order to counterbalance possible order effects and patterning (Prasada & Pinker, 1993; Ramscar, 2002). In experiment 1, the real verbs make, go, sell, stand, step, stop, arrive, and follow were used as fillers, serving a twofold purpose. First, they were supposed to make the task look natural and authentic, providing existing regular and irregular verbs that pupils could be assumed to be familiar with; second, these fillers tested if participants understood the timeframe setting and applied inflections in a meaningful way. The real test items and their syllable structure are summarised in Table 4.1. As we can see in Table 4.1, onsets are always filled with at least two consonants, nucleus vowel [I] is constant, and coda varies between single and double consonants, containing either nasals or the voiceless alveolar stop. After this task, the pupils completed a fake grammar task and stated Table 4.1 Test items from experiment 1 with syllable structure and associated schema Syllable structure Syllable constituents

Test item

pling tring spim sking sprit skrit skrink splink

Templatea

Onset

Nucleus

Coda

Schema

CCVC CCVC s CVC s CVCC s CCVC s CCVC s CCVCC s CCVCC

[pl] [tr] [sp] [sk] [spr] [skr] [skr] [spl]

[I] [I] [I] [I] [I] [I] [I] [I]

[ŋ] [ŋ] [m] [ŋ] [t] [t] [ŋk] [ŋk]

[CC_I_ŋ] [CC_I_ŋ] [s C_I_m] [s C_I_ŋ] [s CC_I_t] [s CC_I_t] [s CC_I_ŋk] [s CC_I_ŋk]

Note Items are sorted according to increasing number of syllable segments a Following conservative approaches to syllabification, word-edge coronal obstruent -s is treated as extra-syllabic

4

63

THE EMPIRICAL STUDY

their opinion on how suitable they thought such a task was for future classroom material. This again was supposed to distract the pupils away from the experiment’s real intention. Participants 260 secondary school EL2 learners from four different Austrian schools took part in both experiments. Missing values in the data set (MNAR) were replaced using a polytomous logistic regression method (van Buuren & Groothuis-Oudshoorn, 2011), so that all 260 cases could be retained. Table 4.2 crosstables schools and proficiency levels according to the Common European Framework of Reference (Council of Europe, 2001). The subsample of 19 pupils from a 6th grade, proficiency level B1, served as a comparison group and entered the statistical modelling as a grouping factor. Among all 260 pupils, 124 were female, 136 were male, and none of them declared themselves non-binary. Their age ranged between 12 and 16 years (M = 13.82, SD = 0.93). All subjects were unimpaired learners, and participation was entirely voluntary, without any financial remuneration. Among the 260 participants, 221 reported to be L1 speakers of German, while 39 cited multilingual backgrounds. Among these 39 learners, first and additional languages included Albanian, Arabic, Armenian, BCS (Bosnian, Croatian, Serbian), Czech, Dari, Italian, Persian, Polish, Slovakian, Tagalog, Tulu, and Turkish. All learners self-reported various languages they were learning at the time of data collection; 206 cited two languages, such as English plus French, Spanish, or Latin, 54 cited even more. The learners’ L1 entered the statistical models as a binary covariate, with German as the baseline level and all the other languages Table 4.2 Crosstable of schools by proficiency level

Proficiency levels A2 Schools

Grammar school 1 Middle school 1 Grammar school 2 Middle school 2

115 10 78 38

B1

19

64

T. WAGNER

as the comparison. Exposure to English outside class was not assessed separately, but according to the English teachers involved there was no exceptional extra-mural English (Smit & Schwarz, 2019). Written consent for the learners’ participation in the experiment was secured prior to data collection. Data collection was anonymised, and the entire empirical study conformed to both the American Psychological Association (APA) ethical principles (APA, 2017) as well as the recommendations for good practice in applied linguistics research by the British Association for Applied Linguistics (BAAL, 2021). A data-driven power analysis (t = 2.00, α = 0.05) revealed that, for the smallest effect sizes of interest with onset sCC and coda ï(k), the study would reach power ≥ 0.80 for N ≥ 200. For the other segments, power ranged around 0.50 for N ≥ 200. In other words, the current sample size can be regarded as being sufficient to avoid Type-II errors and detect effects sizes of a considerable magnitude. Data Coding and Analysis Response Variable All responses given by the learners were logged as provided on the test sheets and computerised. After data cleaning, faithful response behaviour was checked by inspecting distractor inflection in the first part of the experiment. The assumption was that serious engagement with the test sheets would be visible in highly accurate inflections of the given existing English irregular verbs. Table 4.3 shows such accuracy in per cent for all eight fillers. Judging by these percentages, it seems as though learners did engage seriously with the task. Interestingly, while for most distractors accuracy is above 90%, it is much lower for items sell and stand; this could be a frequency effect, meaning that exposure to these two words both in- and outside class was limited, and that the learners were still not certain about appropriate inflections. The verb stand does in fact have a comparably low relative frequency in the BNC corpus (seventh lowest value from 160 irregular verbs, BNC World, 2001), but sell ranks in the middle. The participants’ responses to the test items were the unit of analysis and thus formed the dependent variable; responses were coded as a categorical variable in three ways. First, each word provided was coded for the three-factor levels ‘regular’, ‘irregular’, and ‘repetition’ (reply_type). Repetition was a separate factor level in some analyses, since around 9%

4

THE EMPIRICAL STUDY

65

Table 4.3 Crosstable for filler item and type of inflection as well as percentage of correctly inflected forms Type of inflection

Fillers

(1) (2) (3) (4) (5) (6) (7) (8)

arrive follow go make sell stand step stop

Irregular

Regular

6 1 253 234 199 205 14 0

254 259 7 26 61 55 246 260

Accuracy in % 97.69 99.62 97.31 90.00 76.54 78.85 94.62 100.00

of all observations avoided both overt regular past tense suffixation and vowel change. In a second step, replies were coded binary, as ‘irregular’ versus ‘regular’ (reply_type_binary); here, repetitions were collapsed with the other irregular replies and thus treated as if they reflected nondefault past tense formation strategies. Third, participants’ replies were coded for the type of vowel change they produced (vowelchange), with the levels ‘[I-2]’, ‘[I-æ]’ ‘[I-O]’, and ‘no-change’, along with a residual level ‘none’. The coding [O] included both [O:] and [A], since potential ATR contrasts (cf. Chapter 3) are irrelevant for this analysis (Wiese, 2000). The difference between levels ‘no-change’ and ‘none’ is that for some test items similarity-based no-change formations could be expected, like with item skrit (in analogy to hit ), for instance. For other items, however, a lack of any morphophonological change was interpreted as plain repetition. Predictor Variables In experiment 1, the test items’ onset and coda structures entered the regressions models as within-subjects fixed effects; recall that the nucleus was constant. After removing all distractors, onset variants were coded as ‘CC’, ‘s C’, and ‘s CC’, while coda was broken down into ‘[m/t]’, ‘[ŋ]’, and ‘[ŋk]’. Both onset and coda entered final regression models sumcontrast-coded, since there was no natural default reference level; thus, what was modelled here is the difference between their conditional effects and the unweighted grand mean effect. In other words, this coding would

66

T. WAGNER

show us how differently one particular syllable segment level behaved compared to the average behaviour across all segment levels. In order to test neighbouring factor levels against each other, both onset and coda were also coded as repeated contrasts (Schad et al., 2020). This coding would allow examining possible hierarchical effects within the segment levels; for coda, for instance, one could test whether ‘[m/t]’ behaves differently compared to ‘[ŋ]’, and whether ‘[ŋ]’ would differ from ‘[ŋk]’. Figure 4.1 illustrates potential sum-contrast-coded effects of onset and coda on binary reply_type. Figure 4.1 contains two mosaic plots that illustrate the contingency table of each variable pair as shaded bar segments. To the left, we can see the proportions of regular and irregular replies by onset, to the right we can see these proportions by coda. The dashed line represents the mean probability of irregular past tense formations estimated in a generalised linear intercept model. Against this mean, we can see that, for instance, the onset structure s CC produced fewer irregular replies than on average (below the line), while coda structure [ŋ] seems to attract more irregular formations than on average (above the line). Additional covariates were the between-subjects factors proficiency level and participants’ L1. Proficiency level was treatment-coded binary, with the levels ‘A2’ and ‘B1’ (proficiency). Level ‘A2’ was the reference level here. Thus, L1 effects can be interpreted as the difference between lower A2 and higher B1. L1 was treatment-coded binary (Li_ bi), with the levels ‘German’ and ‘other’. ‘German’ coded pupils who

sC

sCC

onsets

0.8 0.6 0.4 0.2 0.0

0.0 CC

irregular

0.2

0.4

0.6

regular

regular irregular

irregularity

0.8

1.0

B 1.0

A

m/t

ng

ngk

codas

Fig. 4.1 Sum-contrast-coding effects for Onset and Coda on the binary dependent variable

4

THE EMPIRICAL STUDY

67

cited German as their sole L1 or as their first language in some bior multilingual setting; ‘other’ was used for pupils with L1s other than German or another first language in bi- or multilingual settings. Analyses For statistical computing, R version 4.2.2 (R Core Team, 2022) was used. The corresponding R-scripts can be found at https://osf.io/ 24n35/. Potential influences of the four predictors onset, coda, Li_ bi and proficiency on reply_type_binary and vowelchange were first examined using χ 2 -tests and Cramer’s V (bias-corrected). In a second step, the influence of each single predictor as well as their combined effects were examined. Monofactorial models were supposed to gauge individual main effects and control potential collinearity in the multifactorial models (Belsey et al., 1980; Friedman & Wall, 2005). Binary outcomes were modelled using generalised linear mixed regression models (lme4 package, version 1.1.31, Bates et al., 2015), which are particularly suited to the repeated-measure design of experiment 1 and 2 as well as non-orthogonal and unbalanced data sets (Cunnings & Finlayson, 2015; Cunnings & Linck, 2015; Gries, 2015; Jaeger, 2008). What is more, model assumptions are comparably relaxed, as these models can deal with sphericity and homoskedasticity issues, and they allow simultaneous modelling of both random and fixed effects (Baayen et al., 2008; Clark, 1973). To estimate parameters, maximum likelihood was used instead of penalised quasi-likelihood, since the latter is reported to produce biased estimates with binary response variables (Thiele & Markussen, 2012). The optimiser used was the one bound by quadratic approximation (bobyqa, Luo et al., 2007) instead of the heuristic default Nelder-Mead method, since some authors of the lme4 package suggested that switching to bobyqa optimisers for logistic models seems to generally improve model fits and convergence. The final model was built bottom-up using likelihood tests but also checked post-hoc for parsimony (Calcagno, 2020; Calcagno & de Mazancourt, 2010). When models did not converge, the random effects structure with the smallest variance was excluded first, followed by the exclusion of the fixed-effects interactions and random slopes for participants. Although there is little consensus as to convergence problems, models that did not converge are not reported in the results section (Cunnings & Finlayson, 2015). Predictors’ effect sizes are reported through logits, odds ratios, as well as a pseudo Cohen’s d value (Sánchez-Meca et al., 2003). For visualisations of the models, the

68

T. WAGNER

packages effects (Fox, 2003) and visreg (Breheny & Burchett, 2017) were used. Multinomial outcomes (three-part inflections and vowel changes) were modelled by means of non-parametric conditional inference trees, fitted by binary recursive partitioning. Such models provide an established statistical procedure for the analysis of behavioural data in a usage-based context (Bybee & de Napoleão de Souza, 2019; Klavan & Divjak, 2016; Lin, 2021), and, contrary to some multinomial regression procedures, they produce easily interpretable statistics (partykit package, Hothorn et al., 2006a, 2006b). The algorithm’s stopping criterion was based on Bonferroni corrected p-values, and the α-level was set to 0.05, which is appropriate for small to moderately-sized data sets. Thus, such inference trees contain only those predictors that are significantly dependent on the target variable and growing the trees can dispense with additional (post-)pruning or cross-validation (see also Schlosser et al., 2019). The combination of mixed models and inference trees appears to complement each other fruitfully (Bybee & Napoleão de Souza, 2019). While mixedeffects models can show general predictor trends when all other influences are held equal, inference trees and random forest models illustrate, in an accessible way, combinatorial effects of predictors as well as effects for particular subtests of the data. Binary conditional inference tree analyses were complemented by socalled model-based recursive partitioning based on generalised linear mixed models, as implemented in the package glmertree (version 0.2.0, Fokkema et al., 2018). These algorithms split the present data set so that the terminal nodes are associated with the four predictor coefficients while adjusting for random effects. Such trees are reported to sometimes have higher predictive accuracy than linear mixed-effects models with prespecified interaction effects. In addition to that, random forest analyses assessed a predictor’s importance in a model (Genuer & Poggi, 2020). The R package used here was randomForest (Liaw & Wiener, 2002). Random Effects Both participants and test items, fully crossed and correlated, entered the mixed models as random-effect factors since their levels were randomly sampled from a larger population of potential EL2 learners and potential test items. Random intercept variance for both variables is illustrated in Fig. 4.2.

4

A

B

by−subject−variability for inflection 1.0

regular

0.6

0.8

0.8

251

231

241

221

201

211

191

181

171

151

161

141

131

121

0.0

irregular 1

0.0 tring

sprit

splink

spim

skrit

skrink

sking

pling

test items

0.2

0.2

0.4

0.4

0.6

regular irregular

choice of inflection

1.0

by−item−variability for inflection

69

THE EMPIRICAL STUDY

participants

Fig. 4.2 Stacked barplots for random intercept variability in reply_type_ binary by test item (left panel) and participants (right panel)

As we can see in Fig. 4.2, only mild random variability is visible for the eight test items in experiment 1 (left panel). Item tring, for instance, the last bar in the graph, shows around twice as many irregular formations than item sprit. In the right-hand panel, in contrast, we can see considerable variation in the participants’ choice of type of inflection; while some participants show almost even distributions between the two types of inflection, others have clear preferences for either one. Thus, our participants apparently show personal preferences when providing inflections, and it seems advisable to statistically control for this effect. Likelihood tests (p < 0.001) suggested including both random variance for participants’ intercepts (s 2 = 4.04, SD = 2.01) and slopes, though. Likelihood tests also recommended the inclusion of random intercepts and slopes for the eight test items, though, their intercept variability is in fact minute, with s 2 = 0.06, SD = 0.24. After trimming those models that showed singularity issues and did not converge properly (Barr et al., 2013), the final model contained a random intercept for participants only. Collinearity and Model Criticism Model fit was also assessed through the inspection of R 2 -values (but see Hosmer et al., 2013 for critical comments on this practice), residuals as well as potential multicollinearity, variance inflation, and overfitting (Baayen, 2008). While multicollinearity and variance inflation relate to skewed estimation of effects due to correlated predictors, overfitting (Babyak, 2004; Draper & Smith, 1998) refers to an excess of a model’s variables in relation to the observations in the data set; an abundance

70

T. WAGNER

of terms and variable levels might prevent a model from adequately estimating a predictor’s effect. Collinearity between predictor variables was checked before running simultaneous regression models. Our predictors onset and coda were significantly and medium strongly associated (χ 2 (4) = 1733.3, p < 0.001, Cramer’s V = 0.65, 95% CI [0.64, 0.65]). With VIFs of 4.82, 95% CI [4.47, 5.21] for both onset and coda, however, variance inflation did not appear to be a substantial problem; while classic approaches suggested VIFs threshold levels smaller than 1.50–3.00 for main effects (Booth et al., 1994; Zuur et al., 2011) and values smaller than 20.00 for interactions, more recent studies advocate thresholds of 10.00 for the main effects (Montgomery et al., 2021). Deviance residuals and homoscedasticity were checked using the ggResidpanel package (version 0.3.0; Goode & Rey, 2019) as well as the DHARMa package (version 0.4.6; Hartig, 2022). Detailed output from these model criticism procedures can be found in the corresponding OSF repository at https://osf.io/24n35/. Results When looking at the inflections that the learners produced in experiment 1, we can see that almost 77% looked like regular default inflections, with the -ed suffix attached to the test item. This count includes plain repetitions. Such regular replies would conform to theories suggesting that novel verbs would automatically be inflected by the default mechanism, since nonce-verbs have no lexicalised and stored form available for retrieval from the mental lexicon. In turn, however, almost one-fourth of the replies appear to run against such a default mechanism, with around 23% showing overt apophony. Consider Table 4.4 now. It shows the percentages of regular and irregular inflections in experiment 1 split up by the learners’ L1 and their proficiency level. We can see that the proportion of irregular formations is higher with learners quoting a mother tongue other than German (4.71% out of 15% compared to 18.37% out of 85%). We can also see that while for the A2 learners irregular responses occur around three times less frequently than regular formations (20.82% vs. 71.88%), the B1 learners’ irregular formations amount to almost 50% of their replies (2.26% vs. 5.04%). This is the first indication of a potential influence from proficiency level. Consider Table 4.5 now for a breakdown of all responses into regular and irregular patterns crosstabulated by test item.

4

THE EMPIRICAL STUDY

71

Table 4.4 Percentages for the types of inflections provided by the participants L1

Type of inflection

Total

Other

German

Irregular

4.71

18.37

Regular Total

10.29 15.00

66.63 85.00

Proficiency

Total

A2

B1

23.08

20.82

2.26

23.08

76.92 100.00

71.88 92.70

5.04 7.30

76.92 100.00

Table 4.5 Percentages for regular and irregular responses to the eight test items Test items pling

sking

skrink skrit

spim

splink sprit

tring

Total

Inflections Irregular 2.98 2.84 2.93 2.26 3.03 3.08 1.83 4.13 23.08 Regular 9.52 9.66 9.57 10.24 9.47 9.42 10.67 8.37 76.92 Total 12.50 12.50 12.50 12.50 12.50 12.50 12.50 12.50 100.00

Table 4.6 Percentages for regular, irregular, and repetition responses to the eight test items Test items pling

sking skrink skrit

spim

splink sprit

tring

Total

Inflections Irregular 2.98 2.84 2.93 2.26 3.03 3.08 1.83 4.13 23.08 Regular 8.99 8.41 8.99 8.56 8.65 8.70 8.17 7.45 67.92 Repetition 0.53 1.25 0.58 1.68 0.82 0.72 2.50 0.92 9.00 Total 12.50 12.50 12.50 12.50 12.50 12.50 12.50 12.50 100.99

What we can see in Table 4.5 is that for test item tring, for instance, a third of all responses were in fact irregular (4.13% out of 12.50%). This is the first indication for how readily such a test item lends itself to non-default processing in the learners’ mental lexicon. In contrast, item sprit scores comparably low on irregulars (1.83%). If we separate regular responses from plain repetitions, however, another pattern emerges (Table 4.6).

72

T. WAGNER

When looking at repetitions as a separate and potentially similaritybased reply strategy, we can see that more than 30% (23.08% + 9.00%) of the learners’ replies constitute non-default inflections. For test items such as skrit and sprit, in particular, the number of plain repetitions is relatively high. For sprit, repetitions amount to 20% out of the entire responses to this item. This indicates that some nonce-verbs attract irregular nonchange apophonic patterns, as can be found in the verb class containing hit, quit, or cut. That is why patterns with plain repetitions to test items skrit and sprit are coded as irregular inflections later on, since the omission of overt apophony co-occurring with a lack of past tense suffixation is akin to some of the no-change classes of irregular verbs. When looking at the vowel change patterns produced by the learners, we can see the following distribution (Table 4.7). Table 4.7 illustrates the test items’ preferences for certain types of vowel change. Item skrink, for example, attracts a considerable amount of [I-æ] changes (more than 14%), as do pling, spim, splink, and tring. Thus, what we can see is variation of vowel-changing patterns by the phonological make-up of the test items. Let us now look at syllable segments’ individual influence on reply_ type, with the levels ‘regular’, ‘irregular’ and ‘repetition’ (Fig. 4.3, panels A and B) as well as reply_type_binary, contrasting ‘irregular’ versus ‘regular’ replies (panels C and D to the right). Mosaicplots like those in Fig. 4.3 are often used to visualise categorical data (Friendly, 1994; Hartigan & Kleiner, 1981). The size of the tiles in Table 4.7 Percentages for types of vowel change responses to the eight test items Test items pling Vowel [I-2] changes [I-æ] [I-O] Nochange None Total

0.53 2.21 0.24 0.58

sking 0.77 1.59 0.34 1.15

skrink skrit 1.11 1.83 0.14 0.48

0.19 1.39 0.38 1.78

spim 0.1 2.74 0.14 0.58

splink sprit 0.77 2.26 0.19 0.62

0.19 1.15 0.48 2.26

tring 0.91 2.36 0.91 0.77

Total 4.57 15.53 2.82 8.22

8.94 8.65 8.94 8.76 8.94 8.66 8.42 7.55 68.86 12.50 12.50 12.50 12.50 12.50 12.50 12.50 12.50 100.00

4

ngk

inflection

D regular

irregular

inflection

coda

m/t

regular

χ2 (2) = 11.05, p < .01 V = 0.07, CI = [0.00, 0.89]

χ2 (2) = 11.61, p < .01

sCC

V = 0.07, CI = [0.00, 0.64]

73

V = 0.07, CI = [0.00, 0.90]

ngk

χ2 (4) = 34.90, p < .01

onset

coda

ng

m/t

CC sC

onset

irregular

repetition

V = 0.09, CI = [0.00, 0.66]

χ2 (4) = 18.78, p < .01

sCC

C regular

CC

irregular

repetition

ng

B regular

sC

A irregular

THE EMPIRICAL STUDY

inflection

inflection

Fig. 4.3 Mosaicplots for type of inflection including repetition (left panels) and binary (right panels) by onset and coda

such plots are proportional to the counts from the corresponding contingency table. Apart from the tile’s size, however, such plots also visualise possible dependencies between the categorical variables. This visualisation is done through both horizontal and vertical asymmetries. Looking at Fig. 4.3, we can see horizontal asymmetries in all panels, both for threepart (A and B) as well as binary (C and D) inflections. For onset variant CC, for instance, panels A and C illustrates relatively more irregular than regular formations; similar effects can be observed for coda [ï], panel B and D, but, interestingly, not for coda (ïk). Repetition seems to be associated with onsets variant sCC (panel A) and coda [m/t] (panel B). Potential effects visible in Fig. 4.3 were examined using conditional inference trees and mixed regression models. Individual effects of onset CC and coda [ŋ] towards irregular replies against regular ones and repetitions, as illustrated in panels A and B, were confirmed in conditional inference trees. And while two corresponding monofactorial generalised mixed models for potential effects of onset and coda on the binary type of inflection revealed an insignificant main effect for onset (Type-II Wald statistics, χ 2 (2) = 3.17, p = 0.16), a significant partial effect for the structure CC against s CC (logit = 0.50, SE = 0.26, z = 1.92, p = 0.05) emerged. For coda, the main effect is not significant either (Type-II Wald statistics, χ 2 (2) = 4.64, p = 0.10), but the effect for [ŋ] against [m/ t] is (logit = 0.47, SE = 0.24, z = 1.98, p = 0.048). When examining the combined effect of both syllable segments in a conditional inference tree, coda turns out to be the strongest predictor, followed by onset. In such a model, the highest amount of irregular responses can be found with structure [s C(C) _ I _ ŋ(k)]; interestingly, the algorithm could not differentiate between effects for onsets s CC and CC here, as suggested

74

T. WAGNER

in Fig. 4.3. This indicates that the presence of the extra-curricular s does not appear to be influential in the multifactorial analysis. In the final step of the analysis, the two covariates L1_ binary and proficiency were added to the statistical models in order to account for their potential effects, too. In the corresponding generalised regression model for binary inflection, we can see significant main effects for onset (Type-II Wald statistics, χ 2 (2) = 15.61, p < 0.001), coda (Type-II Wald statistics, χ 2 (2) = 14.95, p < 0.001), and participants’ L1 (Type-II Wald statistics, χ 2 (2) = 5.24, p < 0.05). The predictor proficiency is almost significant (Type-II Wald statistics, χ 2 (1) = 3.54, p = 0.06). Table 4.8 summarise the model’s coefficients and associated statistics based on marginal (least square) means contrasts, averaged over the levels of all other predictors, along with their effect sizes (Cohen’s d approximations). Table 4.8 illustrates medium but significant fixed effects for onset CC and coda [ŋk] towards irregular past tense formation (positive logits and odds ratios > 1, Cohen’s d above 0.30). In addition to that, there is a significant fixed effect for onset s CC towards regular past tense formation (negative logit and an odds ratio < 1, Cohen’s d at −0.37). Interesting coefficients can be seen for the covariates learner’s L1 and their proficiency level, too. Apparently, learners with German as their L1 behave significantly differently compared to learners with other languages; they are associated with a 38% lower irregular inflection odds ratio than multilingual learners. In turn, bilingual learners show a clear trend towards irregular inflection, with a significant positive logit and an odds ratio of 61% above the monolingual German-speaking learners. Although the influence of proficiency is not quite significant, it comes with a medium effect size of 0.30. Here, B2 learners show a trend towards irregular formations; thus, it seems as if the higher proficiency level raises the odds for irregular formations by 70%. In a parallel model with repeated contrast coding for onset and coda, estimates show that both contrasts for onset are significant. The logit for s C versus CC is negative, and the odds for s C triggering irregular formations are only around 60% of CC’s odds. Likewise, the estimate for s CC against s C is also negative; here, the odds for s CC triggering irregular formations are less than 50% of s C’s odds. In sum, there seems to be a declining degree of prototypicality for onsets triggering irregular replies, going from CC to sC to s CC. For coda, only one neighbouring contrast is significant. Comparing [ŋk] to [ŋ], we get a positive logit, and the

4.15

Variance

−2.07 0.58 0.09 −0.68 −0.18 −0.38 0.56 −0.48 0.48 −0.53 0.53

Logit

2.04

SD

−2.47 0.18 −0.16 −1.09 −0.44 −0.83 0.19 −0.95 0.01 −1.16 −0.10

LL

95% CI

−1.67 0.99 0.35 −0.26 0.08 0.06 0.94 −0.01 0.95 0.10 1.16

UL

0.20 0.17 0.11 0.17 0.11 0.19 0.16 0.21 0.21 0.28 0.28

SE

−10.14 3.45 0.87 −3.92 −1.65 −2.06 3.58 −2.29 2.29 −1.88 1.88

z

1.23 1.79 1.10 0.51 0.84 0.68 1.75 0.62 1.61 0.59 1.70

OR

−1.14 0.32 0.05 −0.37 −0.10 −0.21 0.31 −0.26 0.26 −0.29 0.29

d

[… ɔ:n …]Vpast

[… ʊn …] en Vpart

[br]o + [[ɪ] ]r

[… ɔ:n …]Vpart.

< bring >

[… ɪn …t]Vpast

[w]o + [[ɪ] k]r

[… ɪn …t]Vpart.

< wink >

V

V

V

V

V

[… æn …] e Vsubj.

[… ɪn …] Vimp.

V

…

[s]o + [ɪƞk]ənr

…

…

[st]o + [ɪƞk]ənr

…

…

[tr]o + [ɪƞk]ənr

…

[… a:n …]Vpast

[br]o + [ɪƞ]ənr

[… a:n …]Vpart.

[… ɪn …tə]Vpast

[w]o + [ɪƞk]ənr

[… ɪn …tə]Vpart.

V

V

V

V

V

Fig. 4.11 Morphophonological properties of an exemplary L2 lexical entry for the English verb sing in L1 German learners of English

4

THE EMPIRICAL STUDY

103

Note, too, that in the rounded boxes, each main entry consists of a phonological and written form as well as the syntactic category information verb (V). For the phonological form, onset and rhyme are explicitly coded, marked through the subscripted o and r. Each main entry is connected to its morphosyntactic categories tense and mood as well as its morphosyntactic properties past tense (preterite for German), participle, and subjunctive (grey boxes). The full entry would probably contain all tense properties as well as aspectual information. Apophonic changes are coded whenever they occur (subscripted n for nucleus); they might in fact be coded—and inter-connected—in all representations as redundant but still accessible information, like in Wunderlich’s (1996) inheritance trees, with percolating and redundant elements being represented by the bracketed dots “[…]”. The lexical entry for sing is also linked to verbs following regular inflections (see wink and winken ‘wave’). Unlike generative symbolic representations, this type of inflection is phonologically encoded, too, with its properties connected to all other apophonic variants (Ambridge, 2010). In sum, what is represented explicitly in Fig. 4.11 is a verb’s morphophonological (and orthographic ) base form, its morphosyntactic category along with its properties, its intralingual connections to phonologically similar verbs and their inflections, as well as cross-linguistic connections to parallel entries in the L2. All of this would reflect an emergent bilingual verbal paradigm. The lexical entry’s explicit representation of morphophonological and syntactic information differs from traditional representations in two ways. First, it almost completely dispenses with symbolic representations. Second, unlike the dual-route model, which comprises an innate domain-specific cognitive mechanism, this usage-based entry would be domain-general and thus highly parsimonious. Instead of principles and parameters, or a universal grammar, the entry illustrates cognitive principles such as categorisation, storage, analogy, and association. As a consequence, the organisation of semi-structured and gradient grammatical phenomena, such as irregular verbs, would be ideally conceived of as derived from domain-general architecture, as an adaptive and emergent process (Ellis & Larsen-Freeman, 2006). As Bybee (2010) once put it, perhaps those phenomena “are more like sand dunes than like planned structures, such as a building” (Bybee, 2010, p. 2). Gradience, use, and emergence are thus not descriptive categories for linguistic analysis, they constitute the essence of the phenomenon in question. Likewise, rules

104

T. WAGNER

and regularities, often the ultimate objective in classical linguistic analyses, have no primacy; they are the result of the same mechanism that produces gradience and variation. L2 grammars would thus be nothing but the sum of a learner’s experiences with an L2, very much in line with approaches such as construction grammar (Fillmore et al., 1999), cognitive grammar (Langacker, 2008), image-schema-based-instruction (Hwang, 2023), or the cognitive linguistics-based instruction (CLI) in language teaching (Yang, 2023).

References Albright, A., & Hayes, B. (2003). Rules vs. analogy in English past tenses: A computational/experimental study. Cognition, 90(2), 119–161. https://doi. org/10.1016/S0010-0277(03)00146-X Ambridge, B. (2010). Children’s judgments of regular and irregular novel pasttense forms. New data on the English past-tense debate. Developmental Psychology, 46(6), 1497–1504. https://doi.org/10.1037/a0020668 Ambridge, B. (2020). Against stored abstractions. A radical exemplar model of language acquisition. First Language. Special Issue, 40(5–6), 509–559. https://doi.org/10.1177/0142723719869731 APA, American Psychological Association. (2017). Ethical principles of psychologists and code of conduct (2002, amended effective June 1, 2010, and January 1, 2017). https://www.apa.org/ethics/code/ BAAL, British Association for Applied Linguistics. (2021). Recommendations on good practice in applied linguistics. https://www.baal.org.uk/wp-content/upl oads/2021/03/BAAL-Good-Practice-Guidelines-2021.pdf Baayen, R. H. (2008). Analyzing linguistic data: A practical introduction to statistics using R. Cambridge University Press. Baayen, R. H., Davidson, D. J., & Bates, D. M. (2008). Mixed-effects modelling with crossed random effects for subjects and items. Journal of Memory and Language, 59(4), 390–412. https://doi.org/10.1016/j.jml.2007.12.005 Babyak, M. (2004). What you see may not be what you get: A brief, nontechnical introduction to overfitting in regression-type models. Psychosomatic Medicine, 66(3), 411–421. https://doi.org/10.1097/01.psy.0000127692.23278.a9 Barr, D. J., Levy, R., Scheepers, C., & Tily, H. J. (2013). Random effects structure for confirmatory hypothesis testing: Keep it maximal. Journal of Memory and Language, 68(3), 255–278. https://doi.org/10.1016/j.jml. 2012.11.001 Bates, D., Liegl, R., Vasishth, S., & Baayen, H. (2015). Parsimonious mixed models. Unpublished manuscript. https://arxiv.org/abs/1506.04967

4

THE EMPIRICAL STUDY

105

Belsey, D. A., Kuh, E., & Welsch, R. E. (1980). Regression diagnostics. Identifying influential data and sources of collinearity. Wiley. Benati, A. (2019). Classroom-oriented research. Processing Instruction (findings and implications). Language Teaching, 52(3), 67–93. https://doi.org/ 10.1017/S0261444817000386 Berko, J. (1958). The child’s learning of English morphology. Word, 14, 150– 177. https://doi.org/10.1080/00437956.1958.11659661 Bernard, A. (2015). An onset is an onset: Evidence from abstraction of newlylearned phonotactic constraints. Journal of Memory and Language, 78, 18–32. https://doi.org/10.1016/j.jml.2014.09.001 Blom, E., Bosma, E., & Heeringa, W. (2021). Regular and irregular inflection in different groups of bilingual children and the role of verbal short-term and verbal working memory. Languages, 6(1), 1–23. https://doi.org/10.3390/ languages6010056 Booth, G. D., Niccolucci, M. J., & Schuster, E. G. (1994). Identifying proxy sets in multiple linear regression: An aid to better coefficient interpretation (US Department of Agriculture, Forest Service research Paper INT). U.S. Department of Agriculture, Forest Service, Intermountain Research Station. Breheny, P., & Burchett, W. (2017). Visualization of regression models using visreg. The R Journal, 9, 56–71. https://doi.org/10.32614/RJ-2017-046 Bybee, J. (2001). Phonology and language use. Cambridge University Press. Bybee, J. (2010). Language, usage and cognition. Cambridge University Press. Bybee, J., & Moder, C. L. (1983). Morphological classes as natural categories. Language, 59(2), 251–270. https://doi.org/10.2307/413574 Bybee, J., & Napoleão de Souza, R. (2019). Vowel duration in English adjectives in attributive and predicative constructions. Language and Cognition, 11(4), 1–27. https://doi.org/10.1017/langcog.2019.32 Bybee, J., & Newman, J. (1995). Are stem changes as natural as affixes? Linguistics, 33(4), 633–654. https://doi.org/10.1515/ling.1995.33.4.633 Bybee, J., & Slobin, D. I. (1982). Rules and schemas in the development and use of the English past tense. Language, 58, 265–289. https://doi.org/10. 2307/414099 Calcagno, V. (2020). glmulti: Model selection and multimodel inference made easy. R package version 1.0.8, https://CRAN.R-project.org/package=glmulti. Calcagno, V., & de Mazancourt, C. (2010). glmulti. An R package for easy automated model selection with (generalized) linear models. Journal of Statistical Software, 34(12), 1–29. https://doi.org/10.18637/jss.v034.i12 Clahsen, H., Felser, C., Neubauer, K., Sato, M., & Silva, R. (2010). Morphological structure in native and nonnative language processing. Language Learning, 60, 21–43. https://doi.org/10.1017/S1366728912000648 Clark, H. (1973). The language-as-fixed-effect fallacy. A critique of language statistics in psychology research. Journal of Verbal Learning and Verbal

106

T. WAGNER

Behaviour, 12(4), 335–359. https://doi.org/10.1016/S0022-5371(73)800 14-3 Council of Europe. (2001). Common European framework of reference for languages: Learning, teaching, assessment. Cambridge University Press. Cunnings, I., & Finlayson, I. (2015). Mixed effects modelling and longitudinal data analysis. In L. Plonsky (Ed.), Advancing quantitative methods in second language research (pp. 159–181). Routledge. Cunnings, I., & Linck, J. A. (2015). The utility and application of mixed-effects models in second language research. Language Learning, 65(1), 185–207. https://doi.org/10.1111/lang.12117 Cuskley, C., Colaiori, F., Castellano, C., Loreto, V., Pugliese, M., & Tria, F. (2015). The adoption of linguistic rules in native and non-native speakers: Evidence from a wug task. Journal of Memory and Language, 84, 205–223. https://doi.org/10.1016/j.jml.2015.06.005 Davies, M. (2012). Expanding horizons in historical linguistics with the 400million word Corpus of Historical American English. Corpora, 7 (2), 121–157. https://doi.org/10.3366/cor.2012.0024 Draper, N. R., & Smith, H. (1998). Applied regression analysis (3rd ed.). Wiley Series in Probability and Statistics. Wiley. Ellis, N. C. (2002). Frequency effects in language processing. A review with implications for theories of implicit and explicit language acquisition. Studies in Second Language Acquisition, 24(2), 143–188. https://doi.org/10.1017/ S0272263102002024 Ellis, N. C. (2006). Selective attention and transfer phenomena in L2 acquisition: Contingency, cue competition, salience, interference, overshadowing, blocking, and perceptual learning. Applied Linguistics, 27 (2), 164–194. https://doi.org/10.1093/applin/aml015 Ellis, N. C., & Larsen-Freeman, D. (2006). Language emergence. Implications for applied linguistics—Introduction to the special issue. Applied Linguistics, 27 (4), 558–589. https://doi.org/10.1093/applin/aml028 Fillmore, C. J., & Kay, P. (1999). Grammatical constructions and linguistic generalisations. The What’s X doing Y? construction. Language, 75(1), 1–33. https://doi.org/10.2307/417472 Fokkema, M., Smits, N., Zeileis, A., Hothorn, T., & Kelderman, H. (2018). Detecting treatment-subgroup interactions in clustered data with generalized linear mixed-effects model trees. Behavior Research Methods, 50(5), 2016– 2034. https://doi.org/10.3758/s13428-017-0971-x Fox, J. (2003). Effect displays in R for generalised linear models. Journal of Statistical Software, 8(15), 1–27. https://doi.org/10.18637/jss.v008.i15 Friedman, L., & Wall, M. (2005). Graphical views of suppression and multicollinearity in multiple regression. The American Statistician, 59(2), 127–136. https://doi.org/10.1198/000313005X41337

4

THE EMPIRICAL STUDY

107

Friendly, M. (1994). Mosaic displays for multi-way contingency tables. Journal of the American Statistical Association, 89, 190–200. https://doi.org/10.1080/ 01621459.1994.10476460 Genuer, R., & Poggi, J. M. (2020). Random forests with R. Springer. Gerngross, G., Puchta, H., Holzmann, C., Lewis-Jones, P., & Stranks, J. (2017). More! Student’s book and workbook. General and Enriched Course. Helbling Languages. Giraudo, H., & Dal Maso, S. (2018). Towards a constructional approach of L2 morphological processing. In G. Booij (Ed.), The construction of words (pp. 603–622). Studies in Morphology (Vol. 4). Springer. https://doi.org/ 10.1007/978-3-319-74394-3_21 Goode, K., & Rey, K. (2019). ggResidpanel. Panels and interactive versions of diagnostic plots using ggplot2. R package version 0.3.0. https://CRAN.Rproject.org/package=ggResidpanel Gries, S. T. (2015). The most under-used statistical method in corpus linguistics: Multi-level (and mixed-effects) models. Corpora, 10(1), 95–125. https://doi. org/10.3366/cor.2015.0068 Gries, S., Hampe, B., & Schönefeld, D. (2005). Converging evidence: Bringing together experimental and corpus data on the association of verbs and constructions. Cognitive Linguistics, 16(4), 635–676. https://doi.org/10. 1515/cogl.2005.16.4.635 Hartig, F. (2022). DHARMa. Residual diagnostics for hierarchical (multi-level / mixed) regression Models. R package version 0.4.6. https://CRAN.R-pro ject.org/package=DHARMa Hartigan, J. A., & Kleiner, B. (1981). Mosaics for contingency tables. In W. F. Eddy (Ed.), Computer science and statistics: Proceedings of the 13th symposium on the interface (pp. 268–273). Springer. Hosmer, D. W., Lemeshow, S., & Sturdivant, R. X. (2013). Applied logistic regression (2nd ed.). John Wiley. Hothorn, T., Hornik, K., Van de Weil, M. A., & Zeileis, A. (2006a). A Lego system for conditional inference. The American Statistician, 60(3), 257–263. https://doi.org/10.1198/000313006X118430 Hothorn, T., Hornik, K., & Zeileis, A. (2006b). Unbiased recursive partitioning: A conditional inference framework. Journal of Computational and Graphical Statistics, 15(3), 651–674. https://doi.org/10.1198/106186006X133933 Hwang, H. B. (2023). Image-schema-based-instruction enhanced L2 construction learning with the optimal balance between attention to form and meaning. International Review of Applied Linguistics in Language Teaching, online first, 1–25. https://doi.org/10.1515/iral-2023-0015 Jaeger, T. (2008). Categorical data analysis: Away from ANOVAs (transformation or not) and towards logit mixed models. Journal of Memory and Language, 59(4), 434–446. https://doi.org/10.1016/j.jml.2007.11.007

108

T. WAGNER

Klavan, J., & Divjak, D. (2016). The cognitive plausibility of statistical classification models: Comparing textual and behavioral evidence. Folia Linguistica, 50(2), 355–384. https://doi.org/10.1515/flin-2016-0014 Langacker, R. (2008). Cognitive grammar. A basic introduction. Oxford University Press. 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 (pp. 106–140). Routledge. Lardiere, D. (2009). Some thoughts on the contrastive analysis of features in second language acquisition. Second Language Research, 25(2), 173–227. https://doi.org/10.1177/0267658308100283 Lemhöfer, K., & Radach, R. (2009). Task context effects in bilingual nonword processing. Experimental Psychology, 56(1), 41–47. https://doi.org/ 10.1027/1618-3169.56.1.41 Leung, Y.-K. I. (2006). Verb morphology in second language versus third language acquisition. In S. H. Foster-Cohen, M. M. Krajnovic & J. M. Djigunovi´c (Eds.), EuroSLA yearbook 6 (pp. 27–56). John Benjamins. https://doi.org/10.1075/eurosla.6.05leu Lewis, M. (1993). The lexical approach. The state of ELT and a way forward. Language Teaching Publications Series. Thomson & Heinle. Liaw, A., & Wiener, M. (2002). Classification and regression by randomForest. R News, 2(3), 18–22. Lin, Z. (2021). A methodological review of machine learning in applied linguistics. English Language Teaching, 14(1), 74–85. https://doi.org/10.5539/elt. v14n1p74 Lüdecke, D. (2018). ggeffects. Tidy data frames of marginal effects from regression models. Journal of Open Source Software, 3(26), 772, 1–5. https://doi. org/10.21105/joss.00772 Luo, Z. Q., Sidiropoulos, N. D., Tseng, P., & Zhang, S. (2007). Approximation bounds for quadratic optimization with homogeneous quadratic constraints. Society for Industrial and Applied Mathematics Journal (SIAM), 18(1), 1–28. https://doi.org/10.1137/050642691 MacWhinney, B. (1992). Transfer and competition in second language learning. In R. J. Harris (Ed.), Cognitive processing in bilinguals (pp. 371–390). NorthHolland. https://doi.org/10.1016/S0166-4115(08)61506-X MacWhinney, B., & O’Grady, W. (Eds.). (2015). The handbook of language emergence. Wiley-Blackwell. https://doi.org/10.1002/9781118346136 Marcus, G. F. (1995). The acquisition of the English past tense in children and multilayered connectionist networks. Cognition, 56(3), 271–279. https://doi. org/10.1016/0010-0277(94)00656-6

4

THE EMPIRICAL STUDY

109

Marecka, M., & Dziubalska-Kołaczyk, K. (2014). Evaluating models of phonotactic constraints on the basis of sC cluster acquisition data. Language Sciences, 46(A), 37–47. https://doi.org/10.1016/j.langsci.2014.06.002 Marusch, T., von der Malsbur, T., Bastiaanse, R., & Burchert, F. (2012). Tense morphology in German agrammatism. The production of regular, irregular and mixed verbs. The Mental Lexicon, 7 (3), 351–380. https://doi.org/10. 1075/ml.7.3.05mar Montgomery, D. C., Peck, E. A., Vining, G. G. (2021). Introduction to linear regression analysis (6th ed.). Wiley Series in probability and Statistics. Wiley. Muroya, A. (2018). L1 transfer in L2 acquisition of English verbal morphology by Japanese young instructed learners. Languages, 4(1), 1. https://doi.org/ 10.3390/languages4010001 Nerbonne, J., & Heeringa, W. (1997). Measuring dialect distance phonetically. In R. Mitkov & B. Boguraev (Eds.), Proceedings of the third meeting of the ACL special interest group in computational phonology (pp. 11–18). Association for Computational Linguistics. Odlin, T. (2000). Language transfer: Cross-linguistic influence in language learning (7th ed.). Cambridge University Press. Penke, M. (2006). Flexion im mentalen Lexikon. Linguistische Arbeiten. Niemeyer. Plag, I. (2000). Irregular past tense formation in English interlanguage. In I. Plag & K. P. Schneider (Eds.), Language use, language acquisition, and language history. (Mostly) Empirical studies in honour of Rüdiger Zimmermann (pp. 134–149). Wissenschafticher Verlag. Prasada, S., & Pinker, S. (1993). Generalization of regular and irregular morphological patterns. Language and Cognitive Processes, 8, 1–56. https://doi.org/ 10.1080/0169096930840694 R Core Team. (2022). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/ Ramscar, M. J. A. (2002). The role of meaning in inflection: Why the past tense does not require a rule. Cognitive Psychology, 45(1), 45–94. https://doi.org/ 10.1016/S0010-0285(02)00001-4 Räsänen, S. H., Ambridge, B., & Pine, J. M. (2016). An elicited-production study of inflectional verb morphology in child Finnish. Cognitive Science, 40(7), 1704–1738. https://doi.org/10.1111/cogs.12305 Rogers, J., Révész, A., & Rebuschat, P. (2016). Implicit and explicit knowledge of inflectional morphology. Applied Psycholinguistics, 37 (4), 781–812. https://doi.org/10.1017/S0142716415000247 Sánchez-Meca, J., Marín-Martínez, F., & Chacón-Moscoso, S. (2003). Effectsize indices for dichotomized outcomes in meta-analysis. Psychological Methods, 8(4), 448–467. https://doi.org/10.1037/1082-989X.8.4.448

110

T. WAGNER

Schad, D. J., Vasishth, S., Hohenstein, S., & Kliegl, R. (2020). How to capitalize on a priori contrasts in linear (mixed) models: A tutorial. Journal of Memory and Language, 110, 104038. https://doi.org/10.1016/j.jml.2019.104038 Schielzeth, H., Dingemanse, N. J., Nakagawa, S., Westneat, D. F., Allegue, H., Teplitsky, C., Réale, D., Dochtermann, N. A., Zsolt Garamszegi, L., & ArayaAjoy, Y. G. (2020). Robustness of linear mixed-effects models to violations of distributional assumptions. Methods in Ecology and Evolution, 11, 1141–1152. https://doi.org/10.1111/2041-210X.13434 Schlosser, L., Hothorn, T., & Zeileis, A. (2019). The power of unbiased recursive partitioning: A unifying view of CTree, MOB, and GUIDE. arXiv:1906.10179, arXiv.org e-print archive. https://arXiv.org/abs/1906. 10179 Ségéral, P., & Scheer, T. (1998). A generalized theory of ablaut: The case of modern German strong verbs. In R. Fabri, A. Ortmann, & T. Parodi (Eds.), Models of inflection (pp. 28–59). De Gruyter. https://doi.org/10.1515/978 3110919745.28 Shinohara, Y., Han, C., & Hestvik, A. (2022). Discriminability and prototypicality of nonnative vowels. Studies in Second Language Acquisition, 44(5), 1260–1278. https://doi.org/10.1017/S0272263121000978 Smit, U., & Schwarz, M. (2019). English in Austria: Policies and practices. In R. Hickey (Ed.), English in the German-Speaking World. Studies in English Language (pp. 294–314). Cambridge University Press. https://doi.org/10. 1017/9781108768924.015 Stärk, K., Kidd, E., & Frost, R. L. A. (2022). The effect of children’s prior knowledge and language abilities on their statistical learning. Applied Psycholinguistics, 43(5), 1045–1071. https://doi.org/10.1017/S01427164 22000273 The British National Corpus, version 2 (BNC World). (2001). Distributed by Oxford University Computing Services on behalf of the BNC Consortium. http://www.natcorp.ox.ac.uk/ Thiele, J., & Markussen, B. (2012). Potential of GLMM in modelling invasive spread. CAB Reviews, 7 , 1–10. https://doi.org/10.1079/PAVSNNR20 127016 Thompson, C. (2018). The role of practice within second language acquisition. In C. Jones (Ed.), Practice in second language learning (pp. 30–52). Cambridge University Press. https://doi.org/10.1017/9781316443118.004 van Buuren, S., & Groothuis-Oudshoorn, K. (2011). mice: Multivariate imputation by chained equations in R. Journal of Statistical Software, 45(3), 1–67. https://doi.org/10.18637/jss.v045.i03 Wagner, T. (2010). Interlanguage morphology. Irregular verbs in the mental lexicon of German-English interlanguage speakers. Narr-Francke-Attempto. Wiese, R. (2000). The phonology of German. Oxford University Press.

4

THE EMPIRICAL STUDY

111

Wu, Z., & Juffs, A. (2022). Effects of L1 morphological type on L2 morphological awareness. Second Language Research, 38(4), 787–812. https://doi.org/ 10.1177/0267658321996417 Wunderlich, D. (1996). Minimalist morphology: The role of paradigms. In G. Booij & J. van Marle (Eds.), Yearbook of morphology 1995 (pp. 93–114). Kluwer Academic. Yang, L. C. (2023). The effects of a cognitive linguistic approach on college students’ learning English tenses in Taiwan. Language Teaching Research. Advanced online publication. https://doi.org/10.1177/136216 88231178196 Zuur, A. F., Ieno, E. N., & Elphick, C. S. (2011). A protocol for data exploration to avoid common statistical problems. Methods in Ecology and Evolution, 1(1), 3–14. https://doi.org/10.1111/j.2041-210X.2009.00001.x

CHAPTER 5

Implications for the Classroom

Abstract Based on the empirical evidence, this chapter will outline the potential implications of this evidence for the EL2 classroom. These implications will be theoretically grounded by means of two well-established learning theories, namely Variation Theory and Input Processing. Both theories can be regarded as explicit, learner-centred approaches to teaching, learning, systematically exploiting cognitive principles in order to guide learners’ attention, awareness, processing strategies. After each learning theory is discussed in turn, important communalities are outlined. Based on this synthesis, a potential roadmap towards concrete ideas for EL2 classroom applications is discussed and graphically illustrated. Keywords Grammar teaching · Explicit and implicit knowledge · Form-focussed instruction · Variation Theory · Input Processing · CEFR

There has long been a controversial debate around causal connections between observable L2 phenomena and their psychological or cognitive reality (Fukuta et al., 2023). More than 20 years ago, however, R. Ellis (2002a) claimed that psycholinguistic evidence for usage-based processing would have profound implications for theories of language acquisition, and, successively, for approaches to teaching second or © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 T. Wagner, English Interlanguage Morphology, https://doi.org/10.1007/978-3-031-50617-8_5

113

114

T. WAGNER

foreign languages. In his view, language learning would basically consist of gradual expansions of associations between patterned linguistic elements and the statistics behind such associations; and L2 language use would be grounded on harnessing such probabilistic information in comprehension and production. It thus is the cognitive representations as such that help speed acquisition in a learner’s grammar (see also Ellis & Schmidt, 1998). In a similar vein, Bybee (2006) argued that what we usually call a learner’s grammar is in fact the result of similarity-based categorisation of language experience, enabling future generalisations. Given that the psycholinguistic evidence from the present two empirical studies is very much in line with this kind of reasoning, the teaching of irregular verb morphology might indeed provide as suitable test case. Recall that it has been referred to as the bane of every language student (Pinker, 1999, IX). And inflectional morphology in general has been described as one of the bottlenecks of L2 development (Slabakova, 2008), a “fundamental problem” for L2 users (Dekeyser, 2005, p. 6). Practitioners would therefore be well-advised to try and exploit underlying cognitive principles in order to guide learners’ attention and processing strategies and thus facilitate learning and acquisition. The question is how practitioners in European classrooms, driven by curricular and logistic constraints, with a couple of contact hours per week, can exploit these underlying cognitive principles. The Common European Framework of Reference for Languages (Council of Europe, 2001, 2018/2020), considerably shaping language learning and teaching in over 40 countries, does not offer much in the way of EL2 pedagogy and methodology (Hulstijn, 2014) in order to support practitioners. While it meticulously describes language proficiency levels and provides a standardised framework for assessing and comparing language abilities, it does not delve into specific teaching methodologies or instructional approaches. Emphasising outcomes and goals of language teaching, it does not offer guidance on how to teach. Educators and institutions must therefore supplement the CEFR with additional resources and methodologies to effectively design and implement language teaching strategies. Despite promising attempts in immersion programmes (Cummins, 2009), and content and language integrated learning (CLIL, Goris et al., 2019; Lasagabaster & Sierra, 2009), regular EL2 school classes, to this day, offer rather limited opportunities for massive exposure and incidental learning. Instead, standard EL2 teaching needs to incorporate

5

IMPLICATIONS FOR THE CLASSROOM

115

more explicit and form-focussed strategies, too. What is more, research shows that mere exposure through immersion and language baths does not guarantee intake, simply because learners can easily misdirect attention, overlook and overhear cues, or employ inadequate form-meaning matching strategies. Inevitably, didactic deliberations geared towards real EL2 classroom situations need to incorporate explicit types of instruction (Norris & Ortega, 2000; Schenck, 2017). This does not suggest that grammatical patterns or morphological paradigms are to be crammed into learners’ minds in a rote-like fashion. Quite the contrary, explicit grammar teaching is supposed to harness those cognitive mechanisms that are operational anyway and hence are bound to promote acquisitional processes, perfectly compatible with inductive and communicative language teaching (Ellis, 2002b; Newby & Pölzleitner, 2022; Terrell, 1991). And there is a renewed interest in the interplay between explicit teaching on the one hand, and explicit learning and implicit, subconscious acquisition, on the other (VanPatten & Smith, 2022). Both types of learning appear to be separable but cooperative (Ellis, 2022), while neither is sufficient. Both types of resulting knowledge can develop simultaneously from instruction (DeKeyser, 2009), and some type of explicit knowledge can even be transformed into implicit knowledge through proceduralisation and automatisation (DeKeyser, 2015), albeit with extensive deliberate practice. Thus, implicit acquisition can be advanced by explicit instruction. While it is still controversial, in how far exactly form-focussed or feature-focussed instruction has beneficial effects on the acquisition of grammatical structures (Ellis, 2002a, 2012), countless studies suggest that form-focussed explicit instruction can definitely rival, and sometimes outperform implicit types, and that its effects are durable (Biber & Reppen, 2002; de Graaff & Housen, 2009; Ellis, 2002a, 2006). In instructed contexts in particular, the facilitative role of raised consciousness and attention, for instance, has been demonstrated repeatedly. For German, there are various EL2-focussed studies from the 1990s (Grabowski, 1995; Grabowski & Mindt, 1994, 1995; Schmidt, 1995) pointing towards the facilitative role of noticing and awareness. More recently, Godfroid and Uggen (2013) could again confirm Schmidt’s attentional constraints on processing vowel-changing irregular German verbs in English-speaking learners of German. They could show that young English-speaking learners of German, at a time when they had not been exposed to non-default German inflection, showed a sensitivity towards German present simple apophony (tragen – trägt ‘carry – he

116

T. WAGNER

carries’). Also investigating German apophony in an L2, Koch et al. (2023) found attentional effects in the eye gaze data of their Dutchspeaking learners of German. In the area of derivational morphology, Wu and Juffs (2022) show that attention to the morphology of derivational inflection in L1 could transfer to L2 processing. In an L2 eyetracking study on verb morphology, Cintrón-Valentín and Ellis (2016), for instance, demonstrated how perceptual salience of verbal morphology guided the learners’ attention, and how instructional focus positively affected success in their acquisition. They highlight the importance of attention to morphophonological and typological cues in form-focussed teaching. Mere focus on form, though, does not always suffice to raise learners’ awareness of semi-regular inflectional patterns, as they might fail to notice the relevant dimensions in the linguistic input, as Schierloh (2011) could show in an eye-movement study with adult beginner learners of German as a foreign language. She concluded that the learners in her study would probably be unlikely to incidentally acquire the correct irregular verb morphology of L2 German. Thus, if learners are left to their own resources, unguided, even extensive temporal attention to morphological details alone might not result in intake, because the threshold of noticing would not be crossed (DeKeyser, 2005; Sharwood Smith & Truscott, 2011). Such attentional effects are expedient for instructed EL2 settings. Teaching the paradigmatic organisation of the English irregular verb paradigm can hardly make use of lexical cues in the learners’ input. Unlike past tense formation in general, which is often connected to lexical tense markers, irregular verbs are a purely morphological, or grammatical phenomenon. Morphological forms, however, are much less salient than lexical forms (Cintrón-Valentín & Ellis, 2016). Given the abundance of evidence outlined so far, one would like to believe that teaching practices have readily implemented this research and made it an integral part of EL2 teaching practices. This, however, does not quite seem to be the case. As Larsen-Freeman (2015, p. 263) prominently pointed out a couple of years ago, “not much second language acquisition or applied linguistics research on grammar has made its way into the classroom” yet. It thus appears as though more support in translating SLA findings into concrete classroom implementations is needed, in particular for grammar teaching. That is why, in the remainder of this chapter, two theories, which help implement explicit and form-focussed instruction in a principled fashion, are discussed. The two theories in question are Input Processing (IP, VanPatten, 1993, 2002; VanPatten &

5

IMPLICATIONS FOR THE CLASSROOM

117

Cadierno, 1993) and Variation Theory (VT, Lo, 2012; Marton & Booth, 2009). Input Processing is a learner-centred SLA theory concerned with the deliberate guidance of learner attention and learner strategies through manipulated input. Variation Theory is a general learner-centred theory of learning, theorising how learners understand objects of learning through the discernment of the objects’ critical features and aspects. One of Variation Theory’s decisive advantages is a complementary action research model, called Learning Study (Lo, 2012), which implements its principles through continuous, systematic, and cyclic reflections of teachers’ approaches and practices. What both Input Processing and Variation Theory share is an explicit but inductive—and thoroughly learner-centred—approach to EL2 teaching, in which universal cognitive principles are exploited in order to guide learners towards the most efficient and comprehensive understanding of what is to be learned (Erlam, 2003; Glaser, 2014; Shaffer, 1989). The core idea behind both approaches is that exposure to examples is not enough but must be complemented by principled guidance towards the appropriate perception of the examples’ critical properties and how these interact. Also, both approaches always start with the identification of potential processing problems in the learners, and both systematically use contrasts and similarities for inductive learning processes. Input Processing and Variation Theory will first be discussed in turn in the following two sections and then synthesised by means of a roadmap towards classroom applications. A discussion of specific teaching materials or lesson plans would, however, go beyond the scope of this chapter.

5.1 English Irregular Verbs and Variation Theory Variation Theory (VT) is a pedagogical learning theory grounded in phenomenography (Bowden & Marton, 1998; Marton & Booth, 2009). In this approach, learning always means learning ‘something’, and learning ‘something’ in turn means developing a thorough understanding of this ‘something’, perhaps even developing an entirely new way of seeing it (Lo, 2012, p. 68). VT is thus compatible with the concept of school subjects being carriers of content that in turn serves as the foundation for the acquisition of competencies and skills. As a pedagogical theory,

118

T. WAGNER

VT also synthesises constructivist cognitive approaches with explicit intervention. While it acknowledges the cognitively motivated constructivist learning paradigm, it views teachers as pro-active and engaging co-authors of the learning process (Bransford et al., 2000). VT is also genuinely inductive (Glaser, 2014; Shaffer, 1989). The ultimate purpose always is for learners to arrive at appropriate conclusions based on the input given; by no means does teaching according to Variation Theory license mindless memorisation, unmotivated rote-learning, or grammar translation. It thus aligns with what is often called guided inductive learning (Cerezo et al., 2016; Robinson, 1996; for counter arguments see Erlam, 2003). Based on discerning variational patterns (differences and similarities) in the learners’ input, VT conceptualises learning as a dynamic path towards understanding so-called objects of learning. Traditionally, there are three variational patterns on such a path. The first type of variation, V1, is the variation within the learners’ comprehension of the object of learning prior to instruction. This is mainly about diagnostic strategies in order to assess what learners actually bring to the classroom and how this informs appropriate teaching strategies and materials. Variation 2 (V2) refers to the variation within teachers’ perception and comprehension of what an object of learning actually is, its critical aspects and features, as well as its didacticisation (Thorsten, 2015). This means that teachers need to start out with reflecting on their particular perception of a potential object of learning, how this perception might differ from learners, and how critical aspects and features are best exploited in order to promote learning. Finally, variation 3 (V3), then, is the actual didactic variation of critical aspects and features in the teaching as a means to promote a profound comprehension of the object of learning. Such objects of learning, however, are not traditional outcomes as specified in national curricula or textbooks. In VT, those policy-driven outcomes would rather be referred to as educational objectives. Objects of learning, in contrast, are dynamic, ever-changing individualised stages on a general path to comprehension (Lo, 2012, p. 41); they are the imminent next step on such a learner’s path, co-determined by each previous step. An Austrian B1 curricular learning outcome, for instance, specifies that learners can talk about their experiences, dreams, hopes, and aims, that they can justify their opinion, and that they can tell stories and reflect on book or film plots (BGB, 2023). The learning path towards such competences would include mastering various tense formations, such as present simple, present perfect, and past tense as well as a good understanding

5

IMPLICATIONS FOR THE CLASSROOM

119

of what exactly these tenses express. A classic VT object of learning for German-speaking learners of English would include the conceptual and formational mastery of English present perfect and its apparent but misleading morphological equivalence in the German Perfekt. Thus, a concrete VT lesson’s object of learning could be the learners’ awareness of the discrepancy between cross-linguistic morphological similarities (auxiliaries and participles) on the one hand, and their stark conceptual differences (German Perfekt for narrated past events, English present perfect for time spans up to the present) on the other. Teaching, then, means using variation in order to make such a path accessible and pave the way towards ultimate comprehension. And VT teaching always starts from the awareness of learners’ demands and what variational patterns individual learners need at one point in time; it thus is a designated learner-centred approach (Lo, 2012). Providing such access is done via disclosing an object’s so-called critical aspects and critical features. Not quite unlike Chomsky’s principle and parameters (Chomsky, 1986), critical aspects are dimensions of variation (Chomsky’s principles), while features are particular values, or states, which these aspects can adopt (Chomsky’s parameters). In order to understand the concept of colour, for instance, one would need to be exposed to its variants (red, blue, and many others). If everything in a learner’s input were red, however, colour, as a superordinate concept, would be incomprehensible. It is the differences and similarities coming with colour exemplars that disclose the principle of colour. In other words, a critical feature becomes visible by opening up the dimension of variation in which the feature can adopt a specific value (Holmqvist et al., 2007). In turn, dealing with critical features in isolation, without the awareness of the underlying critical aspect, would not be beneficial for the learning process. This is like music classes, where learners are required to memorise the notes of the C-major scale without a notion of Western diatonic tonal systems. The crucial aspect of the objects of learning is their dynamic variability. It is not to be understood as monolithic and pre-defined content. Objects of learning, usually focussing on the demands of a learner, emerge and change in the learning process and should be continuously re-assessed and utilised by the teacher (Marton & Booth, 2009; Miechie et al., 2019). This dynamic nature of the object of learning is often referred to as intended, enacted, and lived variants of it. Such objects of learning appear to be self-evident for declarative knowledge, like the laws of physics, or

120

T. WAGNER

deterministic procedures in mathematical reasoning (Marton & Booth, 2009, p. 84). For EL2 teaching and learning, however, they seem to be much harder to define, in particular against strictly competence- and skill-based educational objectives, like in most European EL2 teaching. In Austria, for instance, objects of learning are defined as what learners communicatively can do in the language to be learned (Alderson, 2007; Council of Europe, 2001; Figueras, 2012; Jones, 2011). Declarative content knowledge can thus merely be a vehicle for building these competencies, remaining random and interchangeable to some extent. In a Chomskian sense, what we are left with in such an approach is an individual’s performance in real-life communicative situations; explicit and verbalisable knowledge of grammar would be entirely immaterial. However, although variation theory allows objects of learning to be framed as skills or indirect objects (Bransford et al., 2000; Marton & Booth, 2009), their conceptualisation in EL2 teaching and learning is much less straightforward than for traditional declarative knowledge teaching. Despite those and other conceptual problems (Cheung & Wong, 2014; Gierlinger et al., 2016), however, we have been witnessing a rapid proliferation of VT across different parts of Europe (Gierlinger et al., 2016; Ott, 2017; Pind-Roßnagl, 2015; Spann et al., 2020). There are also numerous studies with a focus on language learning from China and Hong Kong, among others, trying to implement VT in instructed contexts (Lo, 2012). Zhang (2009, 2012), for instance, explored the potential of VT for EL2 teaching and learning in general. Other EL2 studies range from grammar acquisition (Holmqvist & Lindgren, 2009; Lo, 2012) to reading (Johnston, 2014; Tong, 2012) and speaking (Selin, 2014). All of these studies report positive effects of their VT based interventions. And the power of variation for learning processes has been alluded to outside phenomenography, too. Schenck (2017), for instance, discussed scope and limits of form-meaning matching in terms of skilful variations in form, meaning, or both, in order to enable this matching process. Some proponents of VT even claim that the strikingly uniform and universal child language acquisition rests on unconscious variational learning (Lo, 2012, p. 30).

5

IMPLICATIONS FOR THE CLASSROOM

121

5.2 English Irregular Verbs and Processing Instruction Thirty years ago, Input Processing (IP), an input-oriented SLA-based teaching and learning theory, was proposed as a vital component of instructed foreign language teaching (VanPatten & Cadierno, 1993). At the time, it complemented current theories on L2 input, acknowledging that the language which learners are exposed to usually is flooded with lexical and grammatical cues, some of which, however, are only moderately salient, while some are redundant, and yet others run against potentially L1-biased processing strategies. That is why cues are not necessarily picked up, why information sometimes goes missing, and why erroneous form-meaning mappings are easily produced, especially once the sole focus is on communication. IP promised relief and remedy. Adopting and appreciating both the fundamental role of input in language acquisition and the natural limits of input in instructed contexts, IP puts the meaning of grammar centre stage (Boers, 2021), and while it is built around solid cross-linguistic insights from the cognitive sciences, it also offers concrete and hands-on strategies as well as routines for practitioners in the field. Thus, from its inception, it has been trying to link second language acquisition theory and research with pedagogical issues (VanPatten, 1996, p. 1). IP also rests on the assumption that there is an interface between explicit and implicit knowledge. While this was rejected repeatedly in the 1980s (Krashen, 1983), there is broad agreement today that explicit knowledge does have beneficial effects for implicit acquisition, either within the paradigm of the weak (Ellis, 1993) or strong interface position (DeKeyser, 1998). For a recent overview of this topic, see Godfroid (2022). In IP, with its two didactic implementations Processing Instruction (PI) and Structured Input activities (SI), teachers are encouraged to manipulate input in a way that learners are prevented from misguided form-meaning matches and instead are directed towards those critical features that are necessary for meaning-making and input comprehension. IP thus refers to a somewhat universal process according to which learners of a language would always attempt to match forms and meaning in their input (VanPatten, 2004). PI, in turn, reflects an instructor’s attempt to manipulate input so that inappropriate matching strategies are abandoned in favour of more effective ones, ultimately transforming input into intake (DeKeyser, 1998). Finally, SI is the product of such

122

T. WAGNER

input

(I) PI focussed practice

intake

(II)

learner language

(III)

output

traditional explicit grammar instruction focussed practice

Fig. 5.1 The place of PI in the language acquisition processes

PI efforts, it is the sum of all the resources and materials used, but also the classroom practices and teaching approaches which accomplish this transformation. In other words, IP describes a universal human set of cognitive meaning-making principles, PI operationalises these principles for instructed contexts, and SI examples ideally inform teachers how to translate all this into concrete classroom practices (VanPatten, 2004). Figure 5.1, adapted from VanPatten and Cadierno (1993, p. 46), illustrates the decisive departure of Input Processing from traditional explicit grammar instruction. As we can see in Fig. 5.1, second language acquisition is argued to consist of three fundamental steps, marked by the Roman numbers and the horizontal arrows. Input data can become intake (I), which then can enrich the learner language (II) and be used for output (III). This development is mono-directional (cf. arrows) and usually bound by certain constraints. This is why not all input automatically becomes intake, and why, based on the intake, a learner still has to actively develop and enrich their acquired system. While traditional explicit grammar instruction has usually focussed on a learner’s output by eliciting and refining language production (right box), both written and spoken, through explanation, drill, and repetition, PI is concerned with step I, in which input is converted into intake and in which adequate form-meaning connections are made by a learner (left box). Thus, rather than trying to manipulate output, PI attempts to change the way in which input is perceived and processed by a learner. A concrete example for such input manipulation can be found in an experimental IP study (Marsden, 2006), utilising SI activities around the teaching of the French passé compose in the experimental conditions. Parallel to English verb morphology, but semantically different, French passé compose is formed with the auxiliary avoir / être ‘have’/‘be’ as well as the lexical verb’s participle. The sentence somebody has eaten therefore translates into quelqu’un a mangé, with the critical a-auxiliary, which

5

IMPLICATIONS FOR THE CLASSROOM

123

is not present in the present tense equivalent (quelqu’un mange). Apart from the accent, this is the only discernible morphological difference. The addition of this auxiliary, however, Marsden argues, is often overlooked by learners, who end up saying things like il mangé (French imparfait) or il mange (French present tense). Thus, her SI material involves the following steps. First, teachers need to be aware that students’ output might be missing the auxiliary for L2 processing reasons. The Lexical Preference Principle, for instance, explains why learners tend to over-rely on lexical time adverbials, like le weekend dernier ‘last weekend’, which sufficiently express the past time frame, or on pronouns like il, which sufficiently disambiguate the subject. The auxiliary verb might also be missing, because some verbs actually begin with this sound (il achète ‘he buys’ vs il a acheté ‘he bought’). Thus, teachers are supposed to first point out that in French, if you want to talk about what somebody did in the past, an auxiliary, like a, needs to be added before the main verb regardless of other lexical items expressing pastness. Then, examples like quelqu’un mange vs quelqu’un a mange are given to the learners, illustrating the time frame contrast and hence the auxiliary’s crucial role in this tense formation. Marsden then reports several referential and affective tasks that are supposed to help match morphological and semantic meaning information regarding the passé compose. Ever since its inception, IP has been applied to various grammatical phenomena (Henry, 2022), languages (Ito & Wong, 2021; Marsden, 2006), and different educational sectors (Chan, 2018), facilitating many aspects of EL2 learning (for reviews see Benati, 2019; DeKeyser & Botana, 2015; Shintani, 2015). Regarding the acquisition of lexis, Kim and Nam’s (2017) study on idiom acquisition in Cantonese learners of English, for instance, suggests that IP activities were more efficient than comprehension-based activities for language production, and that explicit information on the issue in question facilitated acquisition. Regarding the acquisition of grammar, Benati (2020), in a study on English causative forms, showed that learners using IP activities often outperformed, or at least matched learners going through traditional instruction. Likewise, Wong and Ito (2018) showed a similar effect in the context of the acquisition of French causative forms. Verbal L2 morphology, in particular, has been examined in a number of experimental studies, like English future tense formation (Benati,

124

T. WAGNER

2001), present continuous (Buck, 2000), as well as passive voice (Benati, 2020). English past tense has been addressed in Benati (2005), Benati and Angelovska (2015), and Chan (2018, 2019). Chan’s studies on Cantonese learners acquiring English irregular verbs, for example, pointed towards the benefits of IP for the acquisition of the English past tense. In her intervention studies, IP teaching outperformed implicit and classical explicit interventions in interpretation tasks and implicit interventions in production tasks. Chan concluded that explicit instruction in general, and IP in particular, was more effective than implicit instruction. Quite a few reported IP effects, however, appear to be overly enthusiastic at times. In Benati (2020), for instance, the acquisitional effects with regard to English causative passive emerged in comparison to what was called ‘traditional instruction’. This comparison condition, however, artificially suppressed form-meaning relationships and thus rendered this type of instruction unrealistic and ecologically invalid, which in turn probably inflated the observed effects. In Benati (2005), too, the comparison group was exposed to mechanical tasks, which did not exactly foster formmeaning associations. In addition to that, some post-test items in this study imitated what the experimental group had been doing anyway, which conflated potential intervention effects, too. In sum, though, despite methodological and conceptual caveats (DeKeyser et al., 2002), IP appears to have its merits, and many studies have shown that at least comprehension of grammatical phenomena, such as tense formation, voice, or noun inflection can be improved by SI activities. Added values certainly come with learners realising that grammar, too, conveys meaning and that they can revise their processing strategies if need be.

5.3 A Roadmap Towards Concrete Ideas and Applications Synthesising VT and IP, this section tries to outline a roadmap towards concrete ideas and applications for instructed teaching of EL2 irregular verb morphology. At first sight, this roadmap does perhaps not look like it was informed by decades of research. Experienced practitioners might in fact view it as rather commonsensical, if not self-evident. It does constitute, however, a logical synthesis of cognitive and educational research. Note that this roadmap, while centred around learners making formmeaning connections, is strictly limited to the morphophonological properties of the paradigm. There are, of course, many more important aspects

5

IMPLICATIONS FOR THE CLASSROOM

125

of English verb morphology, as pointed out in Larsen-Freeman’s lexicogrammatical approach to teaching grammar (Larsen-Freeman, 2003, 2015), which are not explicitly addressed in this roadmap. For more such ideas and guidelines around potential IP applications, see also Boers (2021, pp. 201–202), Lee and VanPatten (2003), Nava and Pedrazzini (2018, chapter 4), VanPatten (1996), as well as Wong (2004). For more concrete applications within VT, mostly outside language teaching and learning, though, see Lo (2012). The first phase in both VT and IP is diagnostics. The decisions on what exactly to teach and how to best teach it are infused with diagnostics relating to both the learners (V1) and the teachers (V2). In V1, teachers would run diagnostic pre-tests, or brief communicative tasks, and thus determine potential obstacles for the learners. This also means identifying those critical aspects and features that could cause learners difficulties and possibly prevent them from fully comprehending the objects of learning. In this process, teachers would also realise that they might see the same object of learning differently compared to their learners. This commonly leads to V2, the teacher’s reflections of their own content knowledge and its variation. In this process, they might, for example, re-discover certain aspects and previously overlooked details of irregular verb morphology. On the one hand, this step already highlights crucial critical aspects and features that need attention; on the other, such reflections could involve consciously overriding dearly held practices, and, if need be, expanding the teacher’s in-depth expertise around the objects of learning. Finally, as part of V3, teachers would delineate appropriate potential objects of learning, along with their critical aspects and features. This could include the semantic concept of past tense, its morphophonological formation, the contrast between default suffixation and apophony, and paradigmatic patterns as well as potential cross-linguistic transfer and interference effects. In IP, teachers would recognise, based on the lexical preference principle, the inherent challenge English tense marking poses. Learners, confronted with past tense inflection, are bound to first look out for lexical tense markers, such as yesterday or last week, and only then turn their attention towards potential grammatical markers, like the -ed suffix (with its phonetically conditioned allomorphs), or apophonic changes, respectively. Teachers would also notice that the irregular verbs’ status as content words partly facilitates the acquisition, since they would be

126

T. WAGNER

processed in a learner’s input before grammatical markers such as suffixation or apophony. In addition to that, they would realise that apophony in irregular verbs does not carry redundancy. Form-meaning connections between sang, for instance, and its past tense meaning of sing, are more likely to be accomplished than, for instance, the interpretation of a third person -s, simply because the latter is mostly redundant, but past tense apophony is not. Second, in both VT and IP, suitable, tailor-made input would need to be created through PI and SI. This basically entails two steps. On the one hand, succinct explicit explanations around the target pattern would be given; this, in turn, is supposed to prevent faulty processing strategies from being applied from the outset. Importantly, the diagnostics from phase 1 would help shape these explanations and co-determine the selection and creation of PI materials and courses of action. On the other hand, input would be manipulated (SI) so that appropriate formmeaning matches can be elicited. This contrived yet authentic input, while providing opportunities for focus on form, must be driven by meaning-making tasks and the learners’ desire for authentic communicative exchange. Implementing VT strategies, the input manipulation would be guided by V3, that is, variational patterns highlighting aspects of irregular verb morphology in order to make them, like input enhancement, more salient and thus more accessible (Sharwood Smith, 1993). Such systematic variational patterns in the input would push learners towards awareness of essential differences and similarities in the form and meaning of English irregular verbs. In VT research, this is often presented in tabular forms, like in Table 5.1. This table shows six potential variational patterns for the teaching of irregular verb morphology. The first pattern, line 1, starts with the grammatical target. Here, it is the meaning of pastness. In terms of medium, such a grammatical target could be taught by means of simplified readers and various standard coursebook topics that relate to past events. Variation is avoided for the text type, register, and topic, while systematic variation is introduced for time references and propositions. This means that the texts ought to contain formal time references (tense formations) for present and past, along with their propositions of currentness and pastness. Through these contrasts, learners are supposed to perceive the critical aspect time reference through tense formation. These binary critical features, admittedly simplified for the current purpose, are the possibility in English to express

Simplified (PI, SI) texts, course book topics

Brief sentences Subjects, objects, (PI, SI), dialogues, complements, course book topic adverbials

Pastness (form, regular formation)

Pastness (form, irregular formation

Subjects, objects, complements, adverbials

Regular verbs, irregulars from one verb class

Tense formation of regular lexical verbs

Time references and propositions

Varied

Variation

Text type, register, topic

Simplified (PI, SI) texts, course book topics

Pastness (meaning)

Invariant

Medium

Variational patterns for the teaching of irregular verb morphology

Grammatical target

Table 5.1 Critical features

IMPLICATIONS FOR THE CLASSROOM

(continued)

Utterances can refer to the present (regular, repeated, habitual) and the past (finished, no time span, no reference to present) Morphological No endings in contrast of present present simple, simple and past except 3rd-person simple formation -s, -ed ending in written and three allophones in spoken regular past simple formation Traditional duality of Regular formations English past tense as the default, formation, regular generally versus irregular unproductive irregular past tense formation, consistent patterns within one verb class

Possibilities of time reference through tense formation

Critical aspect

5

127

Medium

Simplified (PI, SI) texts, short dialogues, letters, diary entries

Worksheets, tables and lists

Parallel text translations with cognates, corpus concordances, worksheets

Verb classes

Product-orientation

L1-L2 correspondences

(continued)

Grammatical target

Table 5.1

Tense formation of regular and irregular lexical verbs

Varied

Orientation of the irregular verbs’ presentation The original text Language of the and its propositions parallel translations

Selection of lexical verbs

Time references and propositions

Invariant

Variation

Categorisation principles for verb classes Synchronic communalities in English and German, the role of etymology

Range of past tense morphemes

Critical aspect

Past tense morphs, regular ending, apophony and its patterns, alveolar suffixation, apophony and alveolar suffixation, suppletives, no-change verbs Input- versus output-oriented verb categorisation Parallelisms and differences in apophonic processes, diachronic developments, language typology

Critical features

128 T. WAGNER

5

IMPLICATIONS FOR THE CLASSROOM

129

regular, repeated, and habitual actions in contrast to finished actions, points in time, with no explicit reference to the present. Additional variational patterns would, slowly and gradually, introduce more complex morphological and semantic aspects of irregular past tense to the learners and thus facilitate the discernment of further critical aspects and features. Pragmatic issues (Larsen-Freeman, 2003) are intentionally left out in order to keep this overview accessible. PI and SI in the column headed “Medium” indicate that textual input should be carefully manipulated in order to meet the learners’ processing demands. In addition to the suggestions in Table 5.1, variational patterns could also compare and contrast traditional alphabetic irregular verb lists with linguistically motivated ones (Bybee & Moder, 1983) or frequency-based approaches (Grabowski, 1995; Grabowski & Mindt, 1995). Complementing the above variational patterns, SI activities would further contrast familiar form-meaning mappings of morphological patterns with those that are new to the learners. Such activities often start out with so-called referential activities, followed by affective tasks (VanPatten, 2002). Referential activities, in particular, should be brief and concise, focussing on one linguistic feature at a time so that proper processing is possible. In such activities, attention to form and meaning is usually required, and they are supposed to support the learners’ adequate mapping of form and meaning. For past tense formation, a simple referential activity could provide simplified sentences containing either past simple (target) or present simple (null form) verb phrases, without any further co-indexed lexical tense markers, like temporal adverbials. The task would consist, for instance, of categorising each sentence according to temporal reference. Correct form-meaning mapping would thus depend on the verb inflection only. Other referential activities could prompt learners to read a text in which somebody reports both habitual and past activities, and the learners need to categorise these according to time reference and thus demonstrate that they can associate habitual with present simple and past with past simple. The juxtaposition of both tenses would force learners to recognise how the two different temporal references are morphologically realised. In affective activities learners usually encounter more instances of targeted forms in full texts while they are asked to personally relate to utterances, situations, or experiences, and express their own opinions. Learners could, for instance, be confronted with celebrity reports of some common past event and then prompted to agree or disagree with the

130

T. WAGNER

person’s behaviour or attitude and express how they personally reacted in similar situations. Learners replies would not be assessed in terms of right or wrong but rather how appropriate and complex they are (see Marsden, 2006 as well as Marsden & Chen, 2011, for more examples). Both types of SI need to encourage learners to actively do something, and both types need to provide contexts that are purposeful from a learner’s point of view and thus invite communicative responses. Maintaining the form-meaning connection is essential here, though; once learners are not pushed towards form-meaning associations anymore, input processing can naturally not help enhance them. The learners’ responses can be rudimentary at first, like yes–no-replies, selection or numbering of elements, decisions between alternatives, etc., and then gradually prompt more elaborate and complex responses. For teachers, such activities can have diagnostic merits in turn, since they often reveal if or in how far learners have already made appropriate form-meaning connections. Finally, and perhaps contrary to classic mainstream IP protocol, affective SI activities should provide a plethora of opportunities for pushed output, too (Swain, 1985). Thus, once diagnostics indicate awareness of formal features and appropriate meaning-making strategies, extensive communicative tasks should govern classroom activities. This would include rich and authentic input, creating opportunities for automatised explicit, and, eventually, implicit knowledge development and consolidation (Godfroid & Kim, 2021). Figure 5.2 summarises the general steps of such a roadmap graphically. The roadmap starts in the top-left corner with two circularly related VT steps. In V1, the learners’ needs are acknowledged as a prime basis for material design and teaching strategies. In the next step, teachers critically revisit their own knowledge around the topic or grammar point in question; this is a vital step since the successful definition of objects of learning, along with their critical features and aspects, presuppose a profound understanding of the subject matter. The next two boxes in Fig. 5.2 illustrate how IP would consider the two main principles and related subprinciples in order to address the topic or grammar point in question in a learner-oriented fashion. Here IP rests on the assumption that materials and teaching strategies can be effective only once they cater towards successful matching and meaningmaking strategies. The first three boxes in the middle of Fig. 5.2 show how VT delineates appropriate objects of learning in order to make critical features and aspects discernible. Parallelly, IP would now provide

5

IMPLICATIONS FOR THE CLASSROOM

VT

VT

IP

DIAGNOSTICS (V1)

(V3) define central critical aspects and features of the object(s) of learning

PREPARATION

revisit teacher expertise and acknowledge variation in expertise

IP DIAGNOSTICS

consider the two main IP principles and their ramifications for a grammar topic

VT (V2) consider and acknowledge variation in learner competences DIAGNOSTICS

PREPARATION

VT (V3) delineate preliminary objects of learning based on V1 and V2 PREPARATION

VT

(V3) create variational patterns that make the critical aspects and features discernible PREPARATION

131

select or create suitable PI activities, first referential, then affective

VT TEACHING

use variation to further develop the enacted and dynamic objects of learning

IP TEACHING

use SI materials and strive for more affective pushed output

IP

IP

VT

DIAGNOSTICS

PREPARATION

TEACHING

consider the nine subprinciples and their ramifications for a grammar topic

select or create succinct explanations and contrived yet authentic input

use Learning Study to improve teaching in a cyclic action research fashion

Fig. 5.2 General roadmap for the application of Variation Theory and Input Processing in explicit instruction

succinct explanations of the grammar point in question, so that ill-advised meaning-making strategies in the learners are not adopted in the first place. On the right-hand side, we can see how preparations culminate in the actual teaching of the grammar point in question. In VT, objects of learning would be revisited and changed accordingly in this process, while in IP, SI materials would foster appropriate meaning-making through both referential and affective activities. Finally, in line with VT, the entire teaching process would be systematically and critically reflected, thus providing the basis for further teaching. In sum, this chapter outlined how the large body of empirical evidence around explicit instruction in grammar teaching can be translated into theoretically grounded and principled classroom applications. With respect to the teaching of English irregular verb morphology, the principles and ideas outlined in this chapter could, without departing from

132

T. WAGNER

the overarching communicative teaching paradigm, enrich EL2 teaching practices and thus foster learning.

References Alderson, J. C. (2007). The CEFR and the need for more research. The Modern Language Journal, 91(4), 659–663. https://doi.org/10.1111/j.1540-4781. 2007.00627_4.x Benati, A. (2001). A comparative study of the effects of Processing Instruction and output-based instruction on the acquisition of the Italian future tense. Language Teaching Research, 52(2), 95–127. https://doi.org/10.1191/136 216801678766877 Benati, A. (2005). The effects of processing instruction, traditional instruction and meaning-output instruction on the acquisition of the English past simple tense. Language Teaching Research, 9(1), 67–93. https://doi.org/10.1191/ 1362168805lr154oa Benati, A. (2019). Classroom-oriented research. Processing Instruction (findings and implications). Language Teaching, 52(3), 67–93. https://doi.org/ 10.1017/S0261444817000386 Benati, A. (2020). The effects of structured input and traditional instruction on the acquisition of the English causative passive forms: An eye-tracking study measuring accuracy in responses and processing patterns. Language Teaching Research, Online First. https://doi.org/10.1075/aals.9.08ch5 Benati, A., & Angelovska, T. (2015). The effects of Processing Instruction on the acquisition of English simple past tense: Age and cognitive task demands. International Review of Applied Linguistics in Language Teaching, 53(2), 249–269. https://doi.org/10.1515/iral-2015-0012 BGB. (2023). Verordnung des Bundesministers für Unterricht und Kunst vom 14. November 1984 über die Lehrpläne der allgemeinbildenden höheren Schulen. Aktuelle Fassung BGBl. II Nr. 1/2023. Teil 6, Lehrpläne der einzelnen Unterrichtsgegenstände, Pflichtgegenstände, Unterstufe, lebende Fremdsprachen (erste, zweite), Kompetenzniveaus des Europäischen Referenzrahmens (GER). https://www.ris.bka.gv.at/eli/bgbl/II/2023/1 Biber, D., & Reppen, R. (2002). What does frequency have to do with grammar teaching? Studies in Second Language Acquisition, 24(2), 199–208. https:// doi.org/10.1017/S0272263102002024 Boers, F. (2021). Evaluating second language vocabulary and grammar instruction: A synthesis of the research on teaching words, phrases, and patterns. Routledge. Bowden, J., & Marton, F. (1998). The university of learning: Beyond quality and competence in higher education. Kogan Page.

5

IMPLICATIONS FOR THE CLASSROOM

133

Bransford, D. J., Brown, A. L., & Cocking, R. R. (2000). How people learn. Brain, mind, experience, and school (Expanded ed.). National Academy Press. Buck, M. (2000). The effects of Processing Instruction on the acquisition of English progressive aspect. Estudios de Lingüistica Aplicada, 24(43), 77–95. https://www.redalyc.org/pdf/588/58804305.pdf Bybee, J. (2006). From usage to grammar: The mind’s response to repetition. Language, 82(4), 711–733. https://doi.org/10.1353/LAN.2006.0186 Bybee, J., & Moder, C. L. (1983). Morphological classes as natural categories. Language, 59(2), 251–270. https://doi.org/10.2307/413574 Cerezo, L., Caras, A., & Leow, R. P. (2016). The effectiveness of guided induction versus deductive instruction on the development of complex Spanish ‘gustar’ structures. Studies in Second Language Acquisition, 38(2), 265–291. https://doi.org/10.1017/S0272263116000139 Chan, M. (2018). Processing instruction in helping map forms and meaning in second language acquisition of English simple past. The Journal of Educational Research, 111(6), 720–732. https://doi.org/10.1080/00220671.2017.141 1879 Chan, M. (2019). The role of classroom input: Processing Instruction, traditional instruction, and implicit instruction in the acquisition of the English simple past by Cantonese ESL learners in Hong Kong. System, 80, 246–256. https:// doi.org/10.1016/j.system.2018.12.003 Cheung, W. M., & Wong, W. Y. (2014). Does lesson study work? A systematic review on the effects of lesson study and learning study on teachers and students. International Journal for Lesson and Learning Studies, 3(2), 137–149. https://doi.org/10.1108/IJLLS-05-2013-0024 Chomsky, N. (1986). Knowledge of language. Its nature, origin, and use. Praeger. Cintrón-Valentín, M. C., & Ellis, N. (2016). Salience in second language acquisition: Physical form, learner attention, and instructional focus. Frontiers in Psychology, 29(7), 1–21. https://doi.org/10.3389/fpsyg.2016.01284 Council of Europe. (2001). Common European Framework of Reference for Languages: Learning, teaching, assessment. Cambridge University Press. Council of Europe. (2018/2020). Common European Framework of Reference for Languages: Learning, teaching, assessment (Companion volume). Council of Europe Publishing. Cummins, J. (2009). Bilingual and immersion programs. In M. Long & C. Doughty (Eds.), The handbook of language teaching (pp. 159–181). Wiley-Blackwell. De Graaff, R., & Housen, A. (2009). Investigating the effects and effectiveness of L2 instruction. In M. H. Long & C. J. Doughty (Eds.), The handbook of language teaching (pp. 726–755). https://doi.org/10.1002/978144431 5783.ch38

134

T. WAGNER

DeKeyser, R. (2005). What makes learning second-language grammar difficult? A review of issues. Language Learning, 55(S1), 1–25. https://doi.org/10. 1111/j.0023-8333.2005.00294.x DeKeyser, R. (2009). Cognitive-psychological processes in second language learning. In M. Long & C. Doughty (Eds.), Handbook of second language teaching (pp. 119–138). Blackwell. DeKeyser, R. (2015). Skill acquisition theory. In B. VanPatten & J. Williams (Eds.), Theories in second language acquisition: An introduction (pp. 94–112). Routledge. DeKeyser, R., & Botana, G. P. (2015). The effectiveness of Processing Instruction in L2 grammar acquisition: A narrative review. Applied Linguistics, 36(3), 290–305. https://doi.org/10.1093/applin/amu071 DeKeyser, R., Salaberry, R., Robinson, P., & Harrington, M. (2002). What gets processed in processing instruction? A commentary on Bill VanPatten’s ‘Processing instruction: An update.’ Language Learning, 52(4), 805–823. https://doi.org/10.1111/1467-9922.00204 DeKeyser, R. M. (1998). Beyond focus on form. Cognitive perspectives on learning and practising second language grammar. In C. Doughty & J. Williams (Eds.), Focus on form in classroom second language acquisition (pp. 42–63). Cambridge University Press. Ellis, R. (1993). Second language acquisition and the structural syllabus. TESOL Quarterly, 27 (1), 91–113. https://doi.org/10.2307/3586953 Ellis, R. (2002a). The place of grammar instruction in the second/foreign language curriculum. In E. Hinkel & S. Fotos (Eds.), New perspectives on grammar teaching in second language classrooms (pp. 14–34). Routledge. Ellis, R. (2002b). Does form-focussed instruction affect the acquisition of implicit knowledge? A review of the research. Studies in Second Language Acquisition, 24(2), 223–236. https://doi.org/10.1017/S02722 63102002024 Erlam, R. (2003). The effects of deductive and inductive instruction on the acquisition of direct object pronouns in French as a second language. Modern Language Journal, 87 (2), 242–260. https://doi.org/10.1111/1540-4781. 00188 Ellis, R. (2006). Current issues in the teaching of grammar: An SLA perspective. TESOL Quarterly, 40(1), 83–107. https://doi.org/10.2307/40264512 Ellis, N. C. (2022). Second language learning of morphology. Journal of the European Second Language Association, 6(1), 34–59. https://doi.org/10. 22599/jesla.85 Ellis, N. C., & Schmidt, R. (1998). Rules or associations in the acquisition of morphology? The frequency by regularity interaction in human and PDP learning of morphosyntax. Language & Cognitive Processes, 13(2&3), 307–336. https://doi.org/10.1080/016909698386546

5

IMPLICATIONS FOR THE CLASSROOM

135

Figueras, N. (2012). The impact of the CEFR. ELT Journal, 66(4), 477–485. https://doi.org/10.1093/elt/ccs037 Fukuta, J., Nishimura, Y., & Tamura, Y. (2023). Pitfalls of production data analysis for investigating L2 cognitive mechanism: An ontological realism perspective. Journal of Second Language Studies, 6(1), 95–118. https://doi. org/10.1075/jsls.21013.fuk Gierlinger, E. M., Spann, H., & Wagner, T. (2016). Variation theory in Austrian initial EFL teacher education: Potentials and challenges. International Journal for Lesson and Learning Studies, 5(2), 130–141. https://doi.org/10.1108/ IJLLS-12-2015-0045 Glaser, K. (2014). The neglected combination. A case for explicit-inductive instruction in teaching pragmatics in ESL. TESL Canada Journal, 30(7), 150–163. https://doi.org/10.18806/tesl.v30i7.1158 Godfroid, A. (2022). Hypotheses about the interface between explicit and implicit knowledge in second language acquisition. In A. Godfroid & H. Hopp (Eds.), The Routledge handbook of second language acquisition and psycholinguistics (pp. 294–307). Routledge. https://doi.org/10.4324/978 1003018872 Godfroid, A., & Kim, K. M. (2021). The contributions of implicit statistical learning aptitude to implicit second-language knowledge. Studies in Second Language Acquisition, 43(3), 606–634. https://doi.org/10.1017/S02722 63121000085 Godfroid, A., & Uggen, M. S. (2013). Attention to irregular verbs by beginning learners of German. An eye-movement study. Studies in Second Language Acquisition, 35(2), 291–322. https://doi.org/10.1017/S02722 63112000897 Goris, J., Denessen, E., & Verhoeven, L. (2019). Effects of content and language integrated learning in Europe: A systematic review of longitudinal experimental studies. European Educational Research Journal, 18(6), 675–698. https://doi.org/10.1177/1474904119872426 Grabowski, E. (1995). Unregelmäßige Verben im authentischen Englisch und im Englischunterricht. Praxis des neusprachlichen Unterrichts, 42(2), 147–157. Grabowski, E., & Mindt, D. (1994). Die unregelmäßigen Verben des Englischen: eine Lernliste auf empirischer Grundlage. Die Neueren Sprachen, 93(4), 334– 353. Grabowski, E., & Mindt, D. (1995). A corpus-based learning list of irregular verbs in English. ICAME Journal, 19, 5–22. Henry, N. (2022). The offline and online effects of processing instruction. Applied Psycholinguistics, 43(4), 945–971. https://doi.org/10.1017/S01427 16422000200

136

T. WAGNER

Holmqvist, M., Gustavsson, L., & Wernberg, A. (2007). Generative learning. Learning beyond the learning situation. Educational Action Research, 15(2), 181–208. https://doi.org/10.1080/09650790701314684 Holmqvist, M., & Lindgren, G. (2009). Students learning English as a second language. An applied linguistics learning study. Problems of Education in the 21st Century, 10, 86–96. Hulstijn, J. H. (2014). The Common European Framework of Reference for Languages: A challenge for applied linguistics. International Journal of Applied Linguistics, 165(1), 3–18. https://doi.org/10.1075/itl.165.1.01hul Ito, K., & Wong, W. (2021). Sometimes less is more. The effects of phonetically variable input on auditory processing instruction for L2 French. Studies in Second Language Acquisition, 44(4), 1–26. https://doi.org/10.1017/S02 7226312100084X Johnston, N. (2014). Understanding the information literacy experiences of EFL (English as a Foreign Language) students (Doctoral dissertation). Faculty of Science and Engineering, Queensland University of Technology. Jones, N. (2011). Linking assessments to international frameworks of language proficiency (Special Issue 2, pp. 42–47). The Common European Framework of Reference. Research Matters. A Cambridge Assessment publicahttps://www.cambridgeassessment.org.uk/Images/567584-linkingtion. assessments-to-international-frameworks-of-language-proficiency-the-com mon-european-framework-of-reference.pdf Kim, J. E., & Nam, H. (2017). The pedagogical relevance of processing instruction in second language idiom acquisition. International Review of Applied Linguistics in Language Teaching, 55(2), 93–132. https://doi.org/10.1515/ iral-2015-0027 Koch, E. M., Bulté, B., Housen, A., & Godfroid, A. (2023). The predictive processing of number information in subregular verb morphology in a first and second language. Applied Psycholinguistics, 1–34. https://doi.org/10. 1017/S014271642300022X Krashen, S. (1983). Second language acquisition and second language learning. Pergamon Press. Larsen-Freeman, D. (2003). Teaching language. From grammar to grammaring. Heinle. Larsen-Freeman, D. (2015). Research into practice: Grammar learning and teaching. Language Teaching, 48(2), 263–280. https://doi.org/10.1017/ S0261444814000408 Lasagabaster, D., & Sierra, J. M. (2009). Immersion and CLIL in English: More differences than similarities. ELT Journal, 64(4), 367–375. https://doi.org/ 10.1093/elt/ccp082 Lee, J., & VanPatten, B. (2003). Making communicative language teaching happen (2nd ed.). McGraw-Hill.

5

IMPLICATIONS FOR THE CLASSROOM

137

Lo, M. L. (2012). Variation Theory and the improvement of teaching and learning. Gothenburg Studies in Educational Sciences 323. Gothenburg University. Marsden, E. (2006). Exploring input processing in the classroom. An experimental comparison of Processing Instruction and enriched input. Language Learning, 56(3), 507–566. https://doi.org/10.1111/j.1467-9922.2006.003 75.x Marsden, E., & Chen, H. Y. (2011). The roles of structured input activities in Processing Instruction and the kinds of knowledge they promote. Language Learning, 61(4), 1058–1098. https://doi.org/10.1111/j.1467-9922.2011. 00661.x Marton, F., & Booth, S. (2009). Learning and awareness. Routledge. Miechie, L., Tan, Y. S. M., & Adler, D. J. (2019). Towards a conceptualbased, student-centred pedagogy. Teacher candidates’ experiences of crafting the objects of learning. International Journal for Lesson and Learning Studies, 8(4), 334–347. https://doi.org/10.1108/IJLLS-04-2019-0038 Nava, A., & Pedrazzini, L. (2018). Second language acquisition in action. Principles from practice. Bloomsbury. Newby, D., & Pölzleitner, E. (2022). Teaching communicative grammar. In J. Schumm Fauster & U. Fürstenberg (Eds.), Teaching English in Austria: From theory to the classroom and beyond (pp. 151–168). Graz University Library Publishing. Norris, J., & Ortega, L. (2000). Effectiveness of L2 instruction: A research synthesis and quantitative meta-analysis. Language Learning, 50(3), 417–528. https://doi.org/10.1111/0023-8333.00136 Ott, G. (2017). Exploring Variation Theory in form-focused language teaching. Teaching the present perfect in upper secondary EFL. CELT Matters, 1, 9–29. Pind-Roßnagl, S. (2015). Lesson Studies und Learning Studies als Schulentwicklungsberater/in initiieren und belgeiten. Erziehung und Unterricht, 165(1–2), 164–171. Pinker, S. (1999). Words and rules. The ingredients of language. Phoenix. Robinson, P. (1996). Learning simple and complex second language rules under implicit, incidental, rule-search, and instructed conditions. Studies in Second Language Acquisition, 18(1), 27–67. https://doi.org/10.1017/S02722631 00014674 Schenck, A. (2017). Learning to improve grammar instruction through comprehensive analysis of past research. International Review of Applied Linguistics in Language Teaching, 55(2), 165–195. https://doi.org/10.1515/iral-20150038 Schierloh, M. (2011). Attention to irregular verbs by beginning learners of German: An eye-movement study. Dissertation. Michigan State University.

138

T. WAGNER

Schmidt, R. (1995). Consciousness and foreign language learning: A tutorial on the role of attention and awareness in learning. In R. Schmidt (Ed.), Attention and awareness in foreign language learning (pp. 1–63). Second Language Teaching and Curriculum Center, University of Hawaii. Selin, P. (2014). Developing strategic competence in oral interaction in English as a foreign language. A classroom study (Doctoral thesis). Department of Pedagogical, Curricular, and Professional Studies, Faculty of Education, University of Gothenburg, Gothenburg. http://hdl.handle.net/2077/37801 Shaffer, C. (1989). A comparison of inductive and deductive approaches to teaching foreign languages. Modern Language Journal, 73(4), 395–403. https://doi.org/10.2307/326874 Sharwood Smith, M. A. (1993). Input enhancement in instructed SLA: Theoretical bases. Studies in Second Language Acquisition, 15(2), 165–179. https:// doi.org/10.1017/s0272263100011943 Sharwood Smith, M. A., & Truscott, J. (2011). Input, intake, and consciousness: The quest for a theoretical foundation. Studies in Second Language Acquisition, 33(4), 497–528. https://doi.org/10.1017/S0272263111000295 Shintani, N. (2015). The effectiveness of Processing Instruction and productionbased instruction on L2 grammar acquisition: A meta-analysis. Applied Linguistics, 36(3), 306–325. https://doi.org/10.1093/applin/amu067 Slabakova, R. (2008). Meaning in the second language. Studies in Language Acquisition (Vol. 34). Mouton de Gruyter. Spann, H., Wagner, T., & Gierlinger, E. M. (2020). Learning Study im fremdsprachlichen Unterricht Englisch – eine Fallstudie an einem oberösterreichischen Gymnasium zu Potenzialen und Herausforderungen aus Lehrer/innen/ perspektive. In K. Soukup-Altrichter, G. Steinmair, & C. Weber (Eds.), Lessonstudies in der Lehrerbildung. Gemeinschaftliche Planung und Evaluation von Unterricht im Lehramtstudium (pp. 241–260). Springer. Swain, M. (1985). Communicative competence: Some roles of comprehensible input and comprehensible output in its development. In S. Gass & C. Madden (Eds.), Input in second language acquisition (pp. 235–253). Newbury House. Terrell, T. D. (1991). The role of grammar instruction in a communicative approach. The Modern Language Journal, 75(1), 52–63. https://doi.org/ 10.2307/329834 Thorsten, A. (2015). How teachers’ practice knowledge is used and challenged in a learning study using variation theory as a tool. International Journal for Lesson and Learning Studies, 4(3), 274–287. https://doi.org/10.1108/ IJLLS-08-2014-0030 Tong, S. Y. A. (2012). Applying the theory of variation in teaching. Australian Journal of Teacher Education, 37 (10), 1–19. https://doi.org/10.14221/ajte. 2012v37n10.3

5

IMPLICATIONS FOR THE CLASSROOM

139

VanPatten, B. (1993). Grammar instruction for the acquisition-rich classroom. Foreign Language Annals, 26(4), 433–450. https://doi.org/10.1111/ j.1944-9720.1993.tb01179.x VanPatten, B. (1996). Input processing and grammar instruction. Theory and research. Ablex Publishing. VanPatten, B. (2002). Processing instruction: An update. Language Learning, 52(4), 755–803. https://doi.org/10.1111/1467-9922.00203 VanPatten, B. (2004). Processing instruction: Theory, research, and commentary. Erlbaum. VanPatten, B., & Cadierno, T. (1993). Explicit instruction and input processing. Studies in Second Language Acquisition, 15(2), 225–243. https://doi.org/ 10.1017/S0272263100011979 VanPatten, B., & Smith, M. (2022). Explicit and implicit learning in second language acquisition. Elements in second language acquisition series. Cambridge University Press. Wong, W. (2004). The nature of processing instruction. In B. VanPatten (Ed.), Processing instruction. Theory, research, and commentary (pp. 33–63). Lawrence Erlbaum Associates. Wong, W., & Ito, K. (2018). The effects of Processing Instruction and traditional instruction on L2 online processing of the causative construction in French. An eye-tracking study. Studies in Second Language Acquisition, 40(2), 241– 268. https://doi.org/10.1017/S0272263117000274 Wu, Z., & Juffs, A. (2022). Effects of L1 morphological type on L2 morphological awareness. Second Language Research, 38(4), 787–812. https://doi.org/ 10.1177/0267658321996417 Zhang, Y. (2009). Variation for the improvement of teaching and learning project: An English case on teaching personal pronouns of English language. School Partnership and Field Experience Office, the Hong Kong Institute of Education. Zhang, Y. (2012). Enhancing English language teaching via variation. Journal of Research in Education, 8, 35–39.

CHAPTER 6

Conclusions and Limitations

Abstract This chapter offers tentative conclusions, some limitations that come with the empirical study, and avenues for future research. The conclusion first describes the potential origins of a lingering scepticism towards explicit instruction and deliberate practice in European EL2 pedagogy. Then, the general relevance of empirical evidence for classroom practice is discussed, reviewing classic prejudices against such evidence. Afterwards, the role of teacher education in principled and evidenceinformed teaching is critically examined. After discussing the limits of the present study, a close cooperation between practitioners and researchers is suggested for future research. The final appeal underlines the importance of empirical research for modern foreign language teaching, not least to counteract the further spread of outdated clichés and classical dichotomies. Keywords Evidence-based pedagogy · Explicit instruction · Applied linguistics · Teacher training

Recall that Chapter 5 outlined a roadmap towards explicit inductive teaching of English irregular verb morphology. This map rests on the assumption that if L2 verb morphology is cognitively organised by a similarity-based and frequency-sensitive pattern associator, it is this very © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 T. Wagner, English Interlanguage Morphology, https://doi.org/10.1007/978-3-031-50617-8_6

141

142

T. WAGNER

mechanism that didactic deliberations might as well start out from. This seems immediately obvious for German-speaking learners of English due to the typological proximity of English and German and the ensuing similarity in terms of verb morphology. It is not restricted to such learners, though. As experiment 1 demonstrated, non-German learners turned out to be even more willing to produce apophonic responses. Thus, it is not exclusively the morphological L1-L2 communalities or pairings that merit such didactic deliberations, it is much rather the learners’ similarity-based representation of the evolving L2 morphological paradigms, irrespective of the learners’ L1s. In other words, the learners’ cognitive architecture and acquisitional mechanisms provide the window into EL2 processing, which can and should be taken into account by explicit and form-focussed instruction, as argued in Chapter 5. This is, admittedly, by no means a new idea; attempts at linguistically motivated didactic preparations of L2 irregular verb morphology go back more than half a century (Kingdon, 1951; Rand, 1966). And it needs to be pointed out here that cognitive or linguistic perspectives are not to be misunderstood as one-size-fits-all solutions. EL2 teaching has been, after all, more or less successful in many contexts in which it takes place. However, what we seem to be witnessing ever since the communicative turn in the 1970s, and its ensuing communicative language teaching (Larsen-Freeman, 2000; Schenck, 2017), is some enduring scepticism towards explicit teaching and deliberate practice (Rousse-Malpat et al., 2022). The role of explicit instruction has been controversially discussed for a long time (Richards & Rodgers, 2001), but ever since Norris and Ortega’s (2000) seminal study, and a follow-up meta-study by Goo et al. (2015), explicit teaching was getting more advocates. This probably boosted the role of explicit instruction in SLA research. As Norris and Ortega (2000) concede, however, there was limited generalisability from their 49 sample studies at the time, and House and De Graaf (2009) pointed out that a general assumption of instruction facilitating learning was still controversially discussed among SLA researchers almost ten years later. Perhaps it is the alleged echoes of the audiolingual or grammar translation method that still cast explicit teaching in a poor light and sometimes makes it look outdated. Even worse, there are voices in the profession claiming that empirical research and scientific evidence around such teaching approaches would not have to say much about classroom applications anyway. Even respectable journals in English language teaching

6

CONCLUSIONS AND LIMITATIONS

143

every now and then provide outlets for such rather ideologic stances. Just a couple of years ago, an applied linguistics researcher claimed that empirical research was largely irrelevant and that teachers could well do “without outsiders’ intervention” (Medgyes, 2017, p. 491). In such a view, one must assume, that inspiration for teaching would probably have to exclusively come from ‘within’, from practitioner to practitioner, worthy by virtue of the classroom baptism of fire. Showing how volatile such an approach can be, Lethaby et al. (2021) discuss a rather illustrative case in point. For decades, the story goes, practitioners had been witnessing language learners’ negative associations with red pens used for marking. Thus, Scrivener (2005), in a popular textbook, advised practitioners to use a green or blue one instead. There is, however, scarce evidence for such a claim. Is this a universal phenomenon, valid in all cultures? Rather unlikely. Is this true for societies where the teaching profession is held in high esteem? Also rather unlikely, since the red pen might as well be perceived as authoritative. And if, as a consequence of Scrivener’s advice, teachers switched to green, wouldn’t then green adopt the same associations and necessitate yet another change? Existing research in this field is still inconclusive (Dukes & Albanesi, 2013; Rutchick et al., 2010) but it certainly does not support Scrivener’s simplified and anxiety-inducing dogma. In fact, there are more powerful factors associated with the effects of corrective feedback marking than the colour of the pen. Practitioners, however, usually lacking time and resources to explore such a field of research, might pick up on such advice, for understandable reasons, to the best of their knowledge, and in good faith. This goes to show how powerful and at the same time dangerous such quick-fix advice can be. Overall, it is astounding how categorically and vehemently sometimes both researchers and practitioners deny any beneficial connection between cognitive research, of which we have an abundance these days, and classroom application. In Europe, potential scepticism towards explicit teaching has potentially been aggravated by the paradigmatic change towards communicative and competence-based outcome orientation, as propagated by the CEFR (Council of Europe, 2001) and the new companion volume (2018/2020). While the principled orientation towards communicative and competence-based learning outcomes certainly marked a huge step forward—after all, very few people will find themselves verbalising linguistic meta-knowledge as part of their language careers—the paths towards such outcomes seem underrepresented, or even obfuscated, in

144

T. WAGNER

the CEFR. This is not to say that the CEFR proactively discourages explicit teaching; in fact, the companion website makes it clear that the CEFR will not tell practitioners “what to do, or how to do it”, or what methodology they should employ (Council of Europe, 2023). The communicative and implicit nature of the outcomes, however, might have fostered the assumption that such objectives can be reached through implicit teaching only. This might be misguided, and, if anything, merely reflects certain doctrines of the day. In general, dichotomous thinking, motivated by long-harboured and unquestioned beliefs, pitting one approach against another, appears rather bizarre. While it is understandable that researchers often strive to isolate interventional effects, it is, especially from a practitioner’s perspective, questionable why theories should contrast communicative and grammar teaching, explicit and implicit teaching, meaning-focussed and formfocussed instruction, or inductive and deductive learning. This perpetuates the impression that these approaches are mutually exclusive, while in fact, they are complementary (Boers, 2021; Doughty & Williams, 1998). Not only does this do injustice to empirical evidence, but it also limits classroom practices and thus learning opportunities. While it seems prudent to approach language acquisition from cultural, sociological, philosophical, and educational viewpoints—since the mind is not just in the head (Ellis, 2019, p. 43)—neither approach can claim supremacy; they just represent different ways of looking at the same phenomenon. Communicative teaching and naturalistic environments, for example, offer valuable input as well as opportunities to connect form and meaning in pushed output. This provides direly needed exposure to the language for usage-based knowledge to emerge. A disdain for form-focussed and consciousness-raising activities, however, jeopardises the learners’ accuracy and control of output, rendering non-salient features, albeit communicatively central, invisible to them. Grammar, after all, fulfils communicative functions. Therefore, explicit teaching can both raise consciousness and awareness and also provide opportunities for inductive and communicatively oriented learning and acquisition. So why not offer learners the best of both worlds, with a reasonable mix of attention to form and frequency, hypothesis revision, and memory consolidation, but also plenty of spontaneous communication, automatisation, and self-expression? One reason for the ongoing disregard of research findings from cognitive sciences, by both acquisition researchers and practitioners, could be its inaccessibility. More often than not, research findings appear cloistered,

6

CONCLUSIONS AND LIMITATIONS

145

a quixotic ivory tower, dodging the reality of language classrooms. And what applied linguistics researchers deem worthy of investigation does not always seem to resonate with what teachers and textbook writers find useful to practise their craft (Boers, 2021). Why is that so? Are applied linguistics results simply not as helpful as researchers might think? Is applied linguistics, contrary to linguistics applied, merely a mediating scholarly activity, trying to scientifically recast real-world problems, without being genuinely relevant to the reality of EL2 teaching and learning (Widdowson, 2000)? Under this assumption, applied linguistics would indeed be nothing more than an ivory tower, and it would urgently need to strive for real practical relevance. If practical relevance is to be found in the evidence, however, teacher education plays a pivotal role for evidence-based or evidence-informed teaching (Graus & Coppen, 2016; Lethaby et al., 2021). On the one hand, research output needs to be made more digestible. Some years ago, Marsden and Kasprowicz (2017, p. 613) urged editors of academic journals to “promote a more international, systematic, and sustainable flow of research” towards the practitioners in the field. One way of doing this are brief summaries of research findings for non-specialists, like in the OASIS project (Open Access Summaries in Language Studies, https://oasis-database.org/), run by the International Association of Applied Linguistics. Another approach comes from recently founded academic journals, such as Benjamins’ Pedagogical Linguistics (https:// benjamins.com/catalog/pl, ISSN 2665-9581), founded in 2020, or Educational Linguistics (https://www.degruyter.com/journal/key/edu ling/html, ISSN 2748-9329), which try to synthesise linguistic evidence and its potential pedagogical relevance. On the other hand, teacher trainers would need to advocate and utilise already existing researchand evidence-based inspiration for grammar teaching (Aarts et al., 2019; Bauer & Nation, 2020; Behrens, 2014; Griffee & Gorsuch, 2024; Keck & Kim, 2014; Larsen-Freeman, 2003; Park-Johnson & Shin, 2020; Rankin & Whong, 2020). This means that linguistics as well as language courses in teacher training programmes should genuinely attempt to synthesise cognitive and psycholinguistic evidence with their implications for the classroom. This very synthesis would help render both disciplines meaningful and complementary within the study programme. The significance of psycholinguistics, however, does not lie in excessive language-dependent linguistic expertise; in line with what Rankin and Whong (2020) termed Virtual Grammar, it would be much more

146

T. WAGNER

useful for future teachers to be aware of general linguistic properties and principles involved in verbal morphology, such as word versus rule processing, or irregularisation. Verbal inflection, while differing radically across languages, is universal, after all, and comes with a constrained range of forms in the world’s languages. Just as Widdowson (2016) argues for the benefit of learners exploring the nature of language in general prior to learning one particular language, EL2 teachers would certainly benefit from a profound understanding of various communicative possibilities for the expression of tense when trying to teach the specifics of English, with all its potential for translanguaging (García & Wie, 2014). Such an understanding and such an awareness would be like a tool, enabling teachers to assess materials, classroom strategies, but also their learners output from different perspectives (McManus & Marsden, 2017, 2018). Theories isolated and detached from real-life applications, in turn, will not inspire future teachers to become reflective practitioners, capable of objectifying their decisions and practices; nor does a pragmatic and self-absolving methodology permeated by unquestioned assumptions. The present study, naturally, comes with a number of limitations. Cross-sectionally investigating English L2 morphology, it cannot illuminate longitudinal developments. What is more, the two experiments did not allow dissociating purely analogical or similarity-based effects from a prototype-based mechanism. Therefore, no conclusions can be drawn as to the degree of abstraction in the young learners’ mental lexicon. As a consequence of such limitations, cognitive and psycholinguistic research would need to consider a longitudinal perspective of L2 morphology, how it is constrained in young learners, and how it develops, perhaps as a function of exposure, or based on growing meta-linguistic awareness. Data from modern learner corpora could be incorporated in both analogical learners and connectionist networks in order to further explore the organisation of L2 morphology. Complementing computational modelling and behavioural studies, like in Clahsen (1999), could investigate L2 English verb morphology with various learners’ L1s and thus explore specifically cross-linguistic and potentially universal organisational principles. Avenues for future acquisitional and didactic research, though, had best be pursued in tandem, with close co-operations between practitioners and researchers, on an equal footing, and as “two-way-traffic” (Boers, 2021, p. 209). While didactic applications should ideally be informed by research, acquisitional research would be well-advised to start out

6

CONCLUSIONS AND LIMITATIONS

147

from actual real-life problems in the classroom as well as the practitioner’s perspective. This would give experimental studies credible ecological validity. To this end, researchers should probably consult with practitioners before designing studies and thus make sure their investigations are of real relevance beyond some theoretical or empirical controversy. In turn, this would make experimental conditions both replicable in future studies and transferable to real teaching situations. Thus, while rigorously relating psycholinguistics evidence to theory-driven approaches to EL2 teaching and learning is desirable and promising, this process does not always have to start with an academic dispute; it could be driven by practical necessities and demands, too. One approach utilising this link between acquisitional research and its associated classroom applications is VT’s so-called Learning Study (Marton et al., 2019), which is always conducted by a team of practitioners and researchers. In this regard, IP and VT could, for instance, employ experimental studies in order to isolate effects of variational patterns and PI routines when teaching irregular verbs. EL2 classroom research, in turn, could implement these patterns and routines in concrete learning scenarios and assess how well they work in the real-life ensemble of conditions, approaches, and methodologies. IP and VT could also theorise an efficient preparation of irregular verb morphology in classroom materials, analyse popular textbooks in this regard, and then have classroom research put their findings to the test and assess whether they are actually conducive to learning. And classical issues in language teaching and learning, such as the language learner’s psychology, with all its individual differences (Gregersen & Mercer, 2022) would need to be more closely connected to questions of explicit and implicit knowledge development (Roehr-Brackin, 2022) in language learners. In conclusion, the past tense debate is far from being over yet (Ellis, 2022; Smolka et al., 2018), and we should probably continue exploiting its valuable output to inform our teaching practices, and to safeguard us from a growing number of pied pipers in the language teaching profession. “The plural of anecdotes is not data”, as the British physician and Guardian columnist purportedly remarked, and “Trust me, I have done this for years” is no substitute for scientific evidence. The present volume tried to utilise such scientific evidence and thus provide impulses for both academics and practitioners; while it has merely cast new lights through old windows, it hopefully sparked a healthy curiosity to further explore

148

T. WAGNER

the teaching and learning of irregular verbs; they do no longer have to be the bane of every language student.

References Aarts, B., Cushing, I., & Hudson, R. (2019). How to teach grammar. Oxford teaching guides. Oxford University Press. Bauer, L., & Nation, I. S. P. (2020). English morphology for the language teaching profession. Routledge. Behrens, S. (2014). Understanding language use in the classroom. A linguistic guide for college educators. Multilingual Matters. Boers, F. (2021). Evaluating second language vocabulary and grammar instruction: A synthesis of the research on teaching words, phrases, and patterns. Routledge. Clahsen, H. (1999). Lexical entries and rules of language: A multidisciplinary study of German inflection. Behavioral and Brain Sciences, 22(6), 991–1060. https://doi.org/10.1017/S0140525X99002228 Council of Europe. (2001). Common European Framework of Reference for Languages: Learning, teaching, assessment. Cambridge University Press. Council of Europe. (2018/2020). Common European Framework of Reference for Languages: Learning, teaching, assessment (Companion volume). Council of Europe Publishing. Council of Europe. (2023, April). Common European Framework of Reference for Languages (CEFR). The framework. Purposes of the CEFR. https://www.coe.int/en/web/common-european-framework-refere nce-languages/uses-and-objectives Doughty, C., & Williams, J. (Eds.). (1998). Focus on form in classroom second language acquisition. Cambridge University Press. Dukes, R. L., & Albanesi, H. (2013). Seeing red: Quality of an essay, color of the grading pen, and student reactions to the grading process. The Social Science Journal, 50(1), 96–100. https://doi.org/10.1016/j.soscij.2012.07.005 Ellis, N. C. (2019). Essentials of a theory of language cognition. The Modern Language Journal, 103(1), 39–60. https://doi.org/10.1111/modl.12532 Ellis, N. C. (2022). Second language learning of morphology. Journal of the European Second Language Association, 6(1), 34–59. https://doi.org/10. 22599/jesla.85 García, O., & Wie, L. (2014). Translanguaging. Language, bilingualism and education. Palgrave Pivot Series. Palgrave Macmillan. Goo, J., Granena, G., Yilmaz, Y., & Novella, M. (2015). Implicit and explicit instruction in L2 learning: Norris & Ortega (2000) revisited and updated. In P. Rebuschat (Ed.), Implicit and explicit learning of languages (pp. 443–482). John Benjamins.

6

CONCLUSIONS AND LIMITATIONS

149

Graus, J., & Coppen, P. A. (2016). Student teacher beliefs on grammar instruction. Language Teaching Research, 20(5), 571–599. https://doi.org/10. 1177/1362168815603237 Gregersen, T., & Mercer, S. (2022). The Routledge handbook of the psychology of language learning and teaching. The Routledge handbooks in applied linguistics. Routledge. Griffee, D. T., & Gorsuch, G. (2024, forthcoming). Using theories for second language teaching and learning. Guidebooks for Language Teachers. Bloomsbury. Housen, A., & De Graaff, R. (2009). Investigating the effects and effectiveness of L2 instruction. In M. Long & C. Doughty (Eds.), The handbook of language teaching (pp. 736–755). Wiley Blackwell. Keck, C., & Kim, Y. (2014). Pedagogical grammar. John Benjamins. Kingdon, R. (1951). The irregular verb. ELT Journal, 5(8), 214–224. https:// doi.org/10.1093/elt/V.8.214 Larsen-Freeman, D. (2000). Techniques and principles in language teaching (2nd ed.). Oxford University Press. Larsen-Freeman, D. (2003). Teaching language. From grammar to grammaring. Heinle. Lethaby, C., Mayne, R., & Harries, P. (2021). An introduction to evidence-based teaching in the English language classroom. Pavillion. 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(4), 613–642. https://doi.org/ 10.1111/modl.12426 Marton, F., Cheung, W. M., & Chan, W. Y. (2019). The object of learning in action research and learning study. Educational Action Research, 27 (4), 481–495. https://doi.org/10.1080/09650792.2018.1489873 McManus, K., & Marsden, E. (2017). L1 explicit information can improve L2 online and offline performance. Studies in Second Language Acquisition, 39(3), 1–34. https://doi.org/10.1017/S027226311600022X McManus, K., & Marsden, E. (2018). Online and offline effects of L1 practice in L2 grammar learning. A partial replication. Studies in Second Language Acquisition, 40(2), 459–475. https://doi.org/10.1017/S0272263117000171 Medgyes, P. (2017). The (ir)relevance of academic research for the language teacher. ELT Journal, 71(4), 491–498. https://doi.org/10.1093/elt/ccx034 Norris, J., & Ortega, L. (2000). Effectiveness of L2 instruction: A research synthesis and quantitative meta-analysis. Language Learning, 50(3), 417–528. https://doi.org/10.1111/0023-8333.00136 Park-Johnson, S. K., & Shin, S. J. (2020). Linguistics for language teachers. Lessons for classroom practice. Routledge.

150

T. WAGNER

Rand, E. (1966). The irregular verb: A classification for teachers. English Teaching Forum, 4(2), 14–18. Rankin, T., & Whong, M. (2020). Grammar. A linguists’ guide for language teachers. Cambridge University Press. https://doi.org/10.1017/978110862 3360 Richards, J., & Rodgers, T. S. (2001). Approaches and methods in language teaching. Cambridge University Press. https://doi.org/10.1017/CBO978 0511667305.021 Roehr-Brackin, K. (2022). Explicit and implicit knowledge and learning of an additional language: A research agenda. Language Teaching, First View, 1–19. https://doi.org/10.1017/S026144482200026X Rousse-Malpat, A., Steinkrauss, R., Wieling, M., & Verspoor, M. (2022). Communicative language teaching. Structure based or dynamic usage based? Journal of the European Second Language Association, 6(1), 20–33. https:// doi.org/10.22599/jesla.86 Rutchick, A. M., Slepian, M. L., & Ferris, B. (2010). The pen is mightier than the word: Object priming of grading standards. European Journal of Social Psychology, 40(5), 704–708. https://doi.org/10.1002/ejsp.753 Schenck, A. (2017). Learning to improve grammar instruction through comprehensive analysis of past research. International Review of Applied Linguistics in Language Teaching, 55(2), 165–195. https://doi.org/10.1515/iral-20150038 Scrivener, J. (2005). Learning teaching. A guidebook for English language teachers. Macmillan Education. Smolka, E., Zwitserlood, P., Wiese, R., Marslen-Wilson, W., & Rösler, F. (2018, revision). Priming effects of German participles—The past-tense debate is not over yet. Manuscript in preparation. Widdowson, H. (2016). ELF, adaptive variability and virtual language. In M. L. Pitzl & R. Osimk-Teasdale (Eds.), English as a lingua franca: Perspectives and prospects: Contributions in honour of Barbara Seidlhofer (pp. 31–38). Mouton De Gruyter. https://doi.org/10.1515/9781501503177-008 Widdowson, H. G. (2000). On the limitations of linguistics applied. Applied Linguistics, 21(1), 3–25. https://doi.org/10.1093/applin/21.1.3

Index

A allomorph, 2, 3, 20 analogical, 3, 5, 32, 33, 36, 38–40, 42–44, 146 analogy, 3, 39 apophonic. See apophony apophonically. See apophony apophony, 3, 16–23, 34, 35, 40, 41, 43, 44, 60, 70, 72, 81–83, 88, 91, 93, 96–98, 101, 103, 115, 125, 126, 128, 142 applied linguistics, 143, 145 association, 39 associative, 32–34, 36, 38, 39, 41, 43, 44, 99, 100 ATR contrasts, 65 attention, 114, 115, 117, 125, 129 awareness, 115, 119, 126, 130

C coda, 3, 7, 39, 61, 62, 64–67, 70, 73–77, 81, 82, 87, 91–93, 96 codas. See coda

Common European Framework of Reference for Languages, 63, 114, 143, 144 communicative turn, 142 computational turn, 36 concatenative, 4, 16, 20 conditional inference tree, 68, 73, 77, 82, 87, 88, 93 connectionist, 5, 6, 32, 36–39 contrast coding, 74, 82 critical aspects, 118, 119, 125, 129 critical features, 117–119, 121, 125–130 cue validity, 41, 60, 61

D default, 3–5, 16, 21, 32, 34, 36, 44, 65, 67, 70–72, 81, 82, 87–89, 93, 96, 115, 125, 127 deliberate practice, 142 deterministic, 4, 33, 42 dual-route, 32–35, 41, 42, 100

© The Editor(s) (if applicable) and The Author(s), under exclusive licence to Springer Nature Switzerland AG 2023 T. Wagner, English Interlanguage Morphology, https://doi.org/10.1007/978-3-031-50617-8

151

152

INDEX

E ecological validity, 147 emergent, 100, 103 evidence-based, 145 evidence-informed, 145 exemplar, v, 5, 6, 38, 39, 41, 44, 99, 100 explicit knowledge, 115, 121

F family resemblance, 39, 43 fixed effects, 65, 67, 74 focus on form, 116, 126 form-focussed, 115, 116, 142, 144

G generalised mixed models, 73, 87 gradient, 34, 39, 41, 103

H homoskedasticity, 67 hybrid, 6, 21, 22, 32, 33

I implicit knowledge, v, vi, 115, 121, 130, 147 incidental learning, 114 inductive learning, 117, 118 inheritance tree, 22, 103 Input Processing, 7, 116, 117, 121–126, 130, 131, 147 intake, 115, 116, 121, 122 interface, 121 item-and-arrangement, 7, 20

K K-nearest neighbour, 42

L learner corpora, 146 Learning Study, 117, 147 lexical entry, 102, 103

M mental lexicon, 7, 34, 41, 44, 70, 71, 81, 99, 100, 146 micro-generalisations, 39 micro-rule, 34, 42, 43 minimalist morphology, 7, 20, 22 mixed regression model, 67, 73, 82, 89 mixed verbs, 21, 35 morphology, v morphophonological, 34, 40, 41, 44, 65, 99, 102, 103, 116, 124, 125 morphosyntactic, 17, 103

N natural morphology, 7, 20 nuclei. See nucleus nucleus, 35, 39

O object of learning, 117–120, 125 onset, 7, 39, 61, 64–67, 70, 73–77, 81, 82, 87, 89, 92, 99, 103 overgeneralisation, 4, 5

P paradigm, 16, 17, 20 paradigmatic. See paradigm past tense, 5, 18–21, 23, 32, 35–37, 40, 43, 44, 116, 118, 124, 125, 127–129 past tense debate, 32, 36, 147 pattern association, 32, 41, 141 practitioners, 143–147

INDEX

probabilistic, 42 Processing Instruction, 121, 129, 147 product-oriented, 18, 19, 59, 98 prototype, 7, 32, 36, 39–44, 60, 61, 74, 81, 82, 88, 91, 146 prototypical. See prototype R random forest, 68, 77, 81, 87 reflective practitioners, 146 rule-based, 34, 37, 41 S schema, 42–44, 60, 61, 82, 84 segment, 7, 44, 60–62, 64, 66, 72, 73, 82, 88, 89, 91 semi-regular, 44, 116 semi-weak, 21 similarity-based, 7, 36, 39, 40, 44, 99, 114, 141, 142, 146 single-route, 32, 36 statistical learning, 99 stochastic, 34, 42 strong verbs, 16, 20 Structured Input, 121–124, 126, 129–131 structured lexical entries, 23

153

suffixation, 16, 19, 21, 65, 70, 72, 86, 87, 93, 125, 126, 128 sum-contrast-coded, 65, 88 symbolic, 4, 6, 32–34, 36, 41, 103

T translanguaging, 146 treatment-coded, 66, 88

U usage-based, 68, 99, 100, 103, 113 u-shaped learning, 4

V Variation Theory, 7, 117–120, 124–126, 131, 147 Virtual Grammar, 145

W weak verbs, 16, 20

Y young learners, 2, 7, 44, 81, 82, 89, 98–100, 146