Perspectives from Systemic Functional Linguistics 9781138237384, 9781315299877

365 43 4MB

English Pages [451] Year 2018

Report DMCA / Copyright


Table of contents :
List of Figures
List of Tables
1 Perspectives from Systemic Functional Linguistics
2 The Stance of Systemic Functional Linguistics Amongst Functional(ist) Theories of Language and Its ‘Systemic’ Purpose
3 Rethinking (Context of) Culture in Systemic Functional Linguistics
4 Pushing the Boundaries of Appliability: Conceptual Frameworks in Cognitive Science and Functional Linguistics
5 The Interface of SFL and Pragmatics: Reporting Strategies in Poe’s “William Wilson”
6 The Notion of a Multilingual Meaning Potential: A Systemic Exploration
7 Maintenance versus Shift in Literary Translation: SFL Perspectives on the Translation of User-Related Varieties in Socio-Cultural Contexts
8 Textual and Logical Choices in the Dramatic Monologue of Teahouse and Its English Translations
9 A Contrastive Description of Projection in English and Spanish Across Ranks: From the Clause Nexus to the Group
10 Towards a Systemic Functional Contrastive Analysis of Japanese and English Corporate Legal Discourse
11 Grammaticalising Attitude: Clause Juncture Particles and NEGOTIATION in Dagaare
12 Systemic Socio-Semantic Stylistics (SSS) as Appliable Linguistics: The Cases of Literary Criticism and Language Teaching/Learning
13 Reading and Writing Across the Curriculum
14 On the Relevance of the Textual Metafunction for Spanish Learners/Teachers of English
15 Nominal Groups in Arabic and English: Experiential Differences and Effect on Tunisian EFL Learners’ Translations
16 Cancer Care as an Integrated Practice: Consultations Between an Oncologist and Patients With Advanced, Incurable Cancer
17 The Process of Perception in the Early Academic Article
18 Construing Space in Academic Writing
19 Elaboration Postmodifiers in Academic and Popular Medical Articles
20 A Systemic Functional Analysis of the Use of Personal Reference in the Medrano Burglary Court Hearings
21 Hypotactic Enhancing Clauses in International Treaties
Recommend Papers

Perspectives from Systemic Functional Linguistics
 9781138237384, 9781315299877

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Perspectives from Systemic Functional Linguistics

This innovative collection brings together contributions from established and emerging scholars highlighting the ‘appliability’ of Systemic Functional Linguistics (SFL) and the ways in which theoretical and analytical conclusions drawn from its applications can inform and advance the study of language. The book discusses SFL’s theoretical foundations and development in recent years to demonstrate its evolution into a more effective analytical tool. Building on this theoretical framework, the volume showcases the theory’s applications in case studies exploring four sub-disciplines of language study: multilingual studies; translation studies; language learning and language teaching; and genre analysis. This all-inclusive volume demonstrates both SFL’s efficacy as a means of theoretical analysis, and also its value as a unique approach to the study of language and meaning, making this an indispensable resource for researchers and scholars in applied linguistics, discourse analysis, genre studies, translation studies, and multilingualism. Akila Sellami-Baklouti is a professor of English language and linguistics at Sfax University (Tunisia). She lectures mainly on generative syntax, semantics, Systemic Functional Grammar, academic writing and research methodology. She has published Ambiguity and Disambiguation (2007) and contributed to Choice in Language (2013) and Hybridity in Systemic Functional Linguistics (2016). Lise Fontaine is a senior lecturer at Cardiff University (Wales). She lectures mainly on functional grammar, word meaning, corpus linguistics and psycholinguistics. She has published Analyzing English Grammar: A Systemic-Functional Introduction (2012); Systemic Functional Linguistics: Exploring Choice (2013); and Choice in Language (2013).

Routledge Studies in Linguistics

Structure in Language A Dynamic Perspective Thomas Berg Metaphor and Reconciliation The Discourse Dynamics of Empathy in Post-Conflict Conversations Lynne J. Cameron Relational Semantics and the Anatomy of Abstraction Tamar Sovran Language Contact and the Origins of the Germanic Languages Peter Schrijver Metonymy and Language A New Theory of Linguistic Processing Charles Denroche A Forensic Linguistic Approach to Legal Disclosures ERISA Cash Balance Conversion Cases and the Contextual Dynamics of Deception James F. Stratman Conceptual Conflicts in Metaphors and Figurative Language Michele Prandi The Language of Pop Culture Edited by Valentin Werner Perspectives from Systemic Functional Linguistics Edited by Akila Sellami-Baklouti and Lise Fontaine For a full list of titles in this series, please visit

Perspectives from Systemic Functional Linguistics

Edited by Akila Sellami-Baklouti and Lise Fontaine

First published 2018 by Routledge 711 Third Avenue, New York, NY 10017 and by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2018 Taylor & Francis The right of the editors to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data A catalog record for this book has been requested ISBN: 978-1-138-23738-4 (hbk) ISBN: 978-1-315-29987-7 (ebk) Typeset in Sabon by Apex CoVantage, LLC


Figure 0.1 Geoff Thompson (1947–2015)

We would like to dedicate this volume to the memory of Professor Geoff Thompson. Geoff was a dear friend and colleague to us all. He was invited to contribute to this volume, following the plenary lecture he gave at the First Tunisian International Systemic Functional Linguistics Conference and Workshop, which was held in Hammamet, Tunisia, 26–28 March 2015. He had accepted, but sadly Geoff died very suddenly in November 2015. In his honour, we reserve this place for him. We remember him very fondly, and we all miss him greatly. Akila Sellami-Baklouti and Lise Fontaine


List of Figures List of Tables Acknowledgements 1 Perspectives from Systemic Functional Linguistics

x xii xv 1


2 The Stance of Systemic Functional Linguistics Amongst Functional(ist) Theories of Language and Its ‘Systemic’ Purpose



3 Rethinking (Context of) Culture in Systemic Functional Linguistics



4 Pushing the Boundaries of Appliability: Conceptual Frameworks in Cognitive Science and Functional Linguistics



5 The Interface of SFL and Pragmatics: Reporting Strategies in Poe’s “William Wilson”



6 The Notion of a Multilingual Meaning Potential: A Systemic Exploration



7 Maintenance versus Shift in Literary Translation: SFL Perspectives on the Translation of User-Related Varieties in Socio-Cultural Contexts A K I L A S E L L AMI- B AKL O UTI




8 Textual and Logical Choices in the Dramatic Monologue of Teahouse and Its English Translations



9 A Contrastive Description of Projection in English and Spanish Across Ranks: From the Clause Nexus to the Group



10 Towards a Systemic Functional Contrastive Analysis of Japanese and English Corporate Legal Discourse



11 Grammaticalising Attitude: Clause Juncture Particles and NEGOTIATION in Dagaare



12 Systemic Socio-Semantic Stylistics (SSS) as Appliable Linguistics: The Cases of Literary Criticism and Language Teaching/Learning



13 Reading and Writing Across the Curriculum



14 On the Relevance of the Textual Metafunction for Spanish Learners/Teachers of English



15 Nominal Groups in Arabic and English: Experiential Differences and Effect on Tunisian EFL Learners’ Translations



16 Cancer Care as an Integrated Practice: Consultations Between an Oncologist and Patients With Advanced, Incurable Cancer



17 The Process of Perception in the Early Academic Article DAV I D B A N KS


Contents 18 Construing Space in Academic Writing

ix 350


19 Elaboration Postmodifiers in Academic and Popular Medical Articles



20 A Systemic Functional Analysis of the Use of Personal Reference in the Medrano Burglary Court Hearings



21 Hypotactic Enhancing Clauses in International Treaties



Contributors Index

423 429


0.1 2.1 2.2 2.3 3.1 3.2 6.1

6.2 6.3 6.4


7.1 8.1 8.2 8.3 10.1


Geoff Thompson (1947–2015) Jackendoff’s Parallel Architecture (2002, p. 125) The system of Polarity (Halliday and Matthiessen, 2014, p. 23, Figures 2.1–2.8) Between functional motivations and linguistic structures: (a) DuBois (1985); (b) Newmeyer (1999) Relations between context and language: System and instance Matthiessen’s (2007) Cline of Variation The ordered typology of systems operating in different phenomenal realms with associated potentials—increasing complexity, decreasing domain of operation Typological variation in realizations of the term ‘negative’ in the system of polarity (‘positive’/‘negative’) Tense and aspect as complementary systems of construing process time Diversification in the relations between semantics and lexicogrammar in the representation of a multilingual meaning potential The stratal subsystems of a multilingual meaning potential consisting of two or more languages, and its context, where the functional complementarity of the languages constituting the multilingual meaning potential is specified Translation strategies on a shift-maintenance cline Theoretical framework of the present study Frequency of different types of Theme in the analysis Frequency of taxis and logico-semantic type Global organisation of a privacy policy under the GSP and the semantic system of RHETORICAL RELATIONS (Twitter privacy policy) Illustration of the interpersonal structure of the Dagaare clause

v 16 17 21 30 34

92 99 102


109 133 142 143 150

188 208

Figures 12.1 12.2 14.1 14.2 14.3 14.4 14.5 15.1 16.1 16.2 16.3 16.4 17.1 18.1 19.1 19.2 20.1 21.1

The double-articulation framework (based on Hasan, 1985 [1989], p. 99) The context of creation (our representation) Contentfulness scale Theme markedness Theme complexity Theme weight Theme reference Percentage of Arabic equivalents for English Qualifiers in participants’ performance FIELD network (Hasan, 2009, p. 183) A fragment of INTERACTANT RELATIONS network (Hasan, 2014, p. 33; Hasan, 2013) Butt’s (2003) network of GOAL-ORIENTATION A fragment of the MODE network (Hasan, 2014, p. 27) Types of process The coding scheme The relationship between genre, register and discipline (Bhatia, 2004) Coding scheme designed for corpus annotation Scheme of annotation Annotation scheme of hypotactic enhancing clauses

xi 232 234 276 278 279 281 282 302 318 323 325 333 340 354 373 375 387 409



2.2 2.3 2.4 5.1 5.2 5.3 7.1 7.2 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9 8.10 8.11 9.1 10.1

10.2 10.3

Eight factors outlining a cartography of the main interests of SFG, Dik’s FG and van Valin’s RRG (adapted from Butler, 2003) The dimensions (forms of order) in language and their ordering principles Three lines of meaning in the clause Reformulation of the system of Polarity as rules of alternating specifications Speech, writing and thought presentation scales (Semino and Short, 2004, p. 49) Level of projection in “William Wilson” Modality in projection Target Texts publication details Grammatical dialect markers in the Source Text Frequency of textual Themes Frequency of interpersonal Themes Frequency of topical Themes Frequency of marked and unmarked topical Themes Example 9 Example 10 Example 11 Example 12 Example 13 Example 14 Example 15 Main contrasts between English and Spanish projection in the clause nexus Frequency distribution of PROCESS TYPE in the Twitter privacy policy, by number and percentage (Japanese n = 180 clauses and English n = 246 clauses) Example 1a—English ST (effective: material) Example 1b—Japanese TT (effective: ‘attributing’ relational)

11 13 15 17 71 76 81 128 129 144 145 146 146 151 151 152 154 154 156 156 179

191 192 192

Tables 10.4 10.5 10.6 10.7 10.8

Example 2a Example 2b Example 3a Example 3b Summary of findings—contrastive analysis of English and Japanese privacy policies 10.9 Corpus of the study (n = 30 texts and 121,021 words: 41,406 words in English and 79,615 words in Japanese) 11.1 Negotiation particles and their general meanings across moods 12.1 Process types with Harry Potter and Lord Voldemort as Doers 12.2 Harry and his body-parts (excluding actions predicted by Voldemort) on a cline of dynamism (based on Hasan, 1985 [1989], p. 46) 12.3 Collocates for ‘truth’ in Foe 13.1 Participatory Action Research (PAR) cycle and the implementation of the programme 13.2 Elemental genres identified across the texts analysed 13.3 Stages in the narrative text 13.4 Stages of the research paper 13.5 Macro-report components 15.1 Experiential analysis of the NG in English in example (1) using SFG 15.2 Analysis of the NG in Arabic in example (1) using the grammar of constructs 15.3 Example of a pseudo-verbal construct in Arabic 15.4 Experiential description of example (12) 15.5 Analysis of the NG in Arabic following the grammar of constructs 15.6 Analysis of example (15) using the grammar of constructs 15.7 Analysis of example (20) using the grammar of constructs 15.8 Experiential division and structural realization of the NG in English in example (24) 15.9 Analysis and structural realization of the NG in Arabic in example (25) 15.10 Original NG versus translation preference in example (27) 16.1 Actions in the corpus instantiating the Field parameter [practical] and their textual realizations 16.2 Actions in the corpus instantiating the Field parameter [conceptual] and their textual realizations 16.3 Aspects of the facilitating role of the oncologist 16.4 Interactants’ share of discourse in texts I, II and III 17.1 Disciplines represented (1675) 17.2 Perception processes by discipline (Journal des Sçavans)

xiii 194 194 196 197 200 201 210 237

239 243 253 257 259 260 260 292 292 296 299 300 303 305 307 308 310 320 321 331 334 341 342

xiv Tables 17.3

Perception processes by discipline (Philosophical Transactions) 17.4 Perception verbs (Philosophical Transactions) 18.1 Semantic classification of space circumstantials 18.2 The corpus 18.3 Spatial references 18.4 Textual and experiential references 18.5 Experiential references of space across the sections and the disciplines 18.6 Concrete and abstract spatial references 18.7 PPs expressing concrete and abstract places 18.8 Functions of abstract and concrete spatial PPs 18.9 Concrete and abstract references in medical RAs 18.10 Rest and motion references across the disciplines 18.11 Prepositions expressing rest and motion 18.12 Grammatical functions of motion PPs 19.1 Basic analysis of the nominal group by various grammars (Fontaine, 2017) 19.2 Corpus description 19.3 Distribution of expansion qualifiers across genres 20.1 General distribution of pronouns across the corpus 20.2 Distribution of pronouns among the different participants in the hearing 20.3 A detailed distribution of personal pronouns among all the speakers in the hearing 20.4 The distribution of different process types and participant roles among the judge, the lawyers and the witnesses of the hearing 20.5 The distribution of processes and participant roles among the prosecutor and the defence lawyer 20.6 The distribution of different processes and participant roles among the witnesses of the hearing 21.1 Corpus description 21.2 Distribution of the semantic realisations of hypotactic enhancing clauses 21.3 Distribution of clausal hypotactic enhancement in the Preamble 21.4 Distribution of clausal hypotactic enhancement in the Core text

343 344 352 353 354 355 357 358 358 359 361 362 363 363 369 375 376 388 388 388

392 394 397 408 410 411 414


This work was completed with the support of the Laboratory Approaches to Discourse (LAD 03/UR/04-09) at the Faculty of Letters and Humanities, University of Sfax, Tunisia and the Systemic Functional Linguistics Association of Tunisia (SYFLAT: 2014S01520APSF1- JORT n°70). We would like to thank Ameni Hlioui and Najla Fki for their assistance in editing the volume.


Perspectives from Systemic Functional Linguistics Lise Fontaine and Akila Sellami-Baklouti

1. Introduction Since its advent in the 1960s, Systemic Functional Linguistics (SFL) has tried to “provide something to think with, a framework of related concepts that can be drawn on in many different contexts where there are problems that turn out to be, when investigated, essentially problems of language” (Halliday, 2009, p. viii). Like any other linguistic theory, SFL has endeavoured to seek answers to problems related to language. However, what distinguishes it from many other linguistic theories is that this endeavour has not been restricted to theoretical questions but has been extended to applications to other fields, offering thus new perspectives and trying to provide a framework that is appliable to other domains of study where language is involved. In his preface to the fourth edition of Halliday’s Introduction to Functional Grammar (2014: xviii), Matthiessen expresses his gratitude to Halliday for “daring to develop appliable linguistics when application was a sign of theoretical impurity”. Six decades after the development of SFL in the 1960s, this appliability is no longer seen as a sign of theoretical impurity, but rather as a sign of richness in linguistic studies. The continual progress and evolution of SFL as a fully developed linguistic theory has made of it more than just an efficient analytical tool as it has gained the power to contribute to the development of many domains. This volume has been motivated by the need to highlight some of the SFL theoretical perspectives which have made of it an ‘appliable’ theory, as it has been intended, and to show these applications at work. The compilation of chapters in this volume is an attempt at showing how theoretical and analytical perspectives from SFL have helped achieve major advances in some fields of study. In addition to theoretical perspectives (Section 1.1), this volume has thematized three other main perspectives on language, namely multilingual and translation studies (Section  1.2), language learning and teaching (Section 1.3) and genre analysis (Section 1.4).


Lise Fontaine and Akila Sellami-Baklouti

1.1. Theoretical Perspectives The first part of the volume explores the theoretical space created by the openness and functional orientation of SFL by considering its potential for interaction with other related theories. To begin this exploration, François (Chapter 2) situates SFL in its functional and systemic backgrounds by considering the historical place of SFL among functional theories of language and by evaluating ‘system’ vs ‘systemicity’ in terms of a theory of systemics. François argues that SFL, while generally systemic in terms of structural linguistics, is not compatible with general system theory due to its preoccupation with motivations which are primarily, or solely, inherent to the semiotic system rather than by its interactions with embodied cognition and the extraneous world. Following this, Bartlett (Chapter 3) examines the theoretical place of the context of culture within SFL by considering the complex interactions of language and society. He discusses how SFL theory might be able to account for perpetual socio-cultural change and argues for viewing texts as “the intersection of multiple semiotic histories”. His work provides an account of how SFL can contribute to and benefit from other models (e.g. critical theory and linguistic anthropological models), arguing that we should be “moving beyond our cultural silos”. In Chapter 4, Williams, Irwin and Russell similarly develop an approach that explores the interface between SFL and other theories, but here it concerns cognitive science. The authors build on the theoretical notions of stratification, instantiation, delicacy and rank to show how cognition can be modelled as a stratified semiotic system. In the last chapter of this section, Triki and Triki (Chapter 5) move into literary traditions by considering the interface between SFL, stylistics and pragmatics. Focusing on projection in particular, they propose a two-step analysis of text, drawing on these three theoretical approaches, which can offer an interpretation of literary texts in sociocultural, psychological and ideological terms.

1.2. Perspectives on Multilingual and Translation Studies Different from the quest for structural universals advocated by generative theory, the multilingual dimension in SFL has rather focused on the investigation of the collective human meaning potential. Although Halliday worked first on Chinese before adding English to his descriptive range of tasks, he was at the same time developing a general theory of grammar, and of language, that could be applied to any language under description, leading thus to a body of research providing accounts of the grammars of different languages. An attempt at sketching the theoretical foundations of such accounts is offered by Matthiessen in Chapter 6, where he locates the notion of meaning potential in a typology of systems and introduces its multilingual version, and then uses pairs and sets



of languages to show how this SFL perspective, combining a view ‘from below’ and a view ‘from above’ can be helpful in studying similarities and difference among languages. More examples of contrastive studies adopting this perspective are provided in the subsequent chapters. In Chapter 7, Sellami-Baklouti argues for the relevance of SFL notions of register and choice to the description and analysis of user-related varieties in a literary text and concludes that variation between languages can originate in the context, hence the need for a “contextual adaptation of meaning” (Matthiessen, this volume). The notion of choice in translation is also addressed in Chapter 8 by Wang and Ma, who discuss four different kinds of metafunctional translation shift related to the textual and logical metafunction in two English translations of the dramatic monologue in a Chinese drama entitled Teahouse. In Chapter 9, Arús Hita provides a contrastive description of projection across ranks in English and Spanish, highlighting similarities and dissimilarities. Another contrastive description—between English and Japanese—is provided in Chapter 10 by Chik, who draws on SFL perspectives to compare written corporate legal texts in a bilingual corpus (Japanese and English). In the last chapter in this section, Chapter 11, Mwinlaaru studies the role of clause initial and final particles in Dagaare and gives a profile of eighteen unique particles, identifying the various attitudinal meanings they enact in exchange and their systemic relationship with one another. 1.3. Perspectives on Learning and Teaching The third section of this volume, a second type of application of SFL is addressed; this section focuses on educational applications, which according to Halliday (1994) are “probably the broadest range of its applications; it includes experience in initial literacy, children’s writing, language in secondary education, classroom discourse analysis, teaching of foreign languages, analysis of textbooks, error analysis, teaching of literature and teacher education”. The chapters in this section address two main SFL applications in educational settings, namely models of register and register variation (Chapters 12 and 13) and models of functional grammar (Chapters 14 and 15). In Chapter 12, Miller and Luporini discuss the implications of teaching Systemic Socio-Semantic Stylistics in ESL literary criticism classes. In Chapter 13, Benítez, Barletta, Chamorro, Mizuno and Moss describe an institutional proposal implemented at Universidad del Norte (Barranquilla, Colombia) designed to provide teachers from various disciplines with theoretical and practical tools to guide their students’ comprehension and writing of discipline-specific texts. The descriptive models of the functional grammar are used to depict and address difficulties encountered by EFL learners when lexico grammatical realizations in English are different from those of their mother language. In Chapter 14, Martínez Insua studies Theme in the oral discourse


Lise Fontaine and Akila Sellami-Baklouti

of native and non-native users of English, arguing that gaining awareness and knowledge of the thematic structure of native English may help learners/teachers in the process of learning/teaching English. In Chapter 15, Moalla addresses the challenges that Tunisian learners face in the translation of nominal groups from English to Arabic, hypothesizing that these difficulties are mainly due to the lack of equivalence in experiential classification and patterns of modification across the two languages. 1.4. Perspectives on Genre Analysis The perspective taken in the fourth section of the volume concentrates on genre and discourse analysis. For Halliday (1994, p. xv) “the aim has been to construct a grammar for the purposes of text analysis”. He further argues that “a discourse analysis that is not based on grammar is not an analysis at all”. On the one hand, discourse analysts recognize the textualization of lexico grammar as one of the three phases of discourse analysis, along with organization of discourse and contextualization of discourse (Bhatia, 2004). Hyland (2011, p. 175) argues that the analysis of typical textual structures can help genre analysts reveal the preferences of disciplinary communities, and he recognizes the influence of “Halliday’s view of language as a system of choices which link texts to particular contexts through patterns of lexico grammatical rhetorical features”. The chapters in this section take this research direction and map generic typicality by exploring lexico grammatical recurrences in different genres pertaining to medical, scientific, academic and legal discourse. Karimi, Moore and Lukin (Chapter 16) open this section with their work on consultations between an oncologist and patients with advanced, incurable cancer. Their application of the SFL contextual system networks in the study of palliative oncology consultations reveals important aspects of consultation communication, including indicators of good practice for both expert practitioner and patient experience. Banks (Chapter 17) shows how the discourse analysis of historical texts can reveal ideological differences between exact sciences and the humanities in terms of the use of process types. In particular, Banks analyses the use of perception processes from the period 1665–1700 in texts taken from the Philosophical Transactions and the Journal des Sçavans, showing how his detailed grammatical analysis informs our understanding of these historical scientific texts. In Chapter 18, Benelhaj turns to academic writing and considers the way space is construed in comparative registers. Her work shows how a detailed analysis of spatial prepositions reveals different uses in different contexts. Continuing with academic writing as the focus, Ktari (Chapter 19) shows, through a detailed analysis of elaboration postmodifiers, how lexico grammatical structure is genre-driven. Her results shed light on the differing uses of elaboration in relation to generic variations. The legal discipline is taken up in Chapters 20 and  21. Hlioui



(Chapter 20) studies the use of personal reference in Medrano burglary court hearings. The lexico grammatical analysis both experientially and interpersonally shows how sensitive these expressions are as an index that reflects the ultimate goals of the speaker in terms of their role in the courtroom as well as the way in which the speaker chooses to represent their experience. Fki (Chapter 21) presents work on international treaties, where she provides detail about the function of hypotactic enhancing clauses. Her work shows how the communicative purposes of the specific sections of the treaties are construed by these constructions. 1.5. Future Perspectives The chapters in this volume testify to the richness of SFL perspectives and its appliability to other fields of study. The holistic model offered by SFL makes of it a suitable framework for the description of human languages, and such comprehensive descriptions can be fruitful in translation studies, educational settings and genre analysis. Although some of the applications have been investigated by contributors to this volume, the list is far from exhaustive, and more research should be pursued not only in each of the study fields outlined in the sections of the volume, but also in other study fields such as critical discourse analysis, semiotics, clinical linguistics and ecolinguistics, to cite a few. The breadth of these perspectives that SFL can offer both on and to other disciplines shows that it is moving forward with the objective it was first developed to fulfil: to “provide something to think with” (Halliday, 2009, p. viii).

References Bhatia, V. K. (2004) Worlds of Written Discourse. London and New York: Continuum. Halliday, M. A. K. (1994) An Introduction to Functional Grammar. London: Edward Arnold. Halliday, M. A. K. (2009) Preface. In J. J. Webster and M. A. K. Halliday (Eds.), Continuum Companion to Systemic Functional Linguistics. London and New York: Continuum. Halliday, M. A. K. and Matthiessen, C. M. I. M. (2014) Halliday’s Introduction to Functional Grammar. London and New York: Routledge. Hyland, K. (2011) Academic Discourse, in Ken Hyland and Brian Paltridge Continuum Companion to Discourse Analysis, London and New York: Continuum. pp. 171–184.


The Stance of Systemic Functional Linguistics Amongst Functional(ist) Theories of Language and Its ‘Systemic’ Purpose Jacques François

The purpose of this chapter is to evaluate the place of Systemic Functional Linguistics (SFL) against two different backgrounds, which offer two different but related perspectives on the theory. The first of these concerns the place of SFL within the broader history of functional theories of language. The second considers SFL in relation to the concept of ‘system’/‘systemicity’ as developed in von Bertalanffy’s General System Theory or systemics (1968, 1972). These two theoretical perspectives on SFL are presented in what follows over four main sections. The first section shall be devoted to an overview of nine decades of functional theories of language. Three periods are to be distinguished: first from the 1920s to the 1960s with four main centres of research in Prague, Vienna, Geneva and Paris; then the period of the 1970s and 1980s with a displacement of the main poles from Prague and Paris towards London, Amsterdam, Buffalo and the US West Coast; and finally from the 1990s to the 2010s with a wide spreading of research topics pertaining to two opposite variants of linguistic functionalism, on the one hand the theory of ‘internal functionalism’, taking only grammatical functions into account, and on the other hand that of ‘external functionalism’, dealing with pragmatic pressures as well. In the second section, I will investigate the stance of SFL amongst structural-functional language theories, first identifying the spectrum of present linguistic functionalism and, second, focusing on so-called structural-functional theories, that is, SFL, the Dutch Functional Grammar and Role and Reference Grammar. Then in section 3, I consider SFL from the following two perspectives, Jackendoff’s (structurally oriented) Parallel Architecture and Langacker’s (non-structurally oriented) Cognitive Grammar. It is only by comparison that we can appreciate the ambition of ‘systemicity’ in SFL. The fourth section explores the systemic ambition of SFL, first by evoking General System Theory in its relationship to cybernetics, and then by pointing to the genesis of linguistic systems in the theory of emergent grammar. The chapter ends with a brief reflection on the systemicity of SFL.

Functional(ist) Theories of Language


1. Nine Decades of Functional(ist) Theories of Language (1929–2014) 1.1. From the 1920s to the 1960s The first functionally oriented linguistic theories seem to appear in the early part of the 20th century with on the one hand the “thèses du cercle de Prague” in 1929 and in the same year the anticipatory work of the Swiss linguist Henri Frei entitled La grammaire des fautes with a very explicit sub-title: Introduction à la linguistique fonctionnelle, assimilation et différenciation, brièveté et invariabilité, expressivité.1 In the next decade, three linguists firmly established the domain of functional linguistics in three rather different directions, first Karl Bühler, a pioneering psycholinguist at the University of Vienna with his Sprachtheorie in 1934, which set down the foundations of linguistic pragmatics; then Charles Bally, a former colleague of Ferdinand de Saussure in Geneva, with Linguistique générale et linguistique française (1936), which developed the notion of ‘functionally equivalent expressions’ in every language (the topic of ‘internal linguistics’) and between languages (that of ‘external linguistics’); and Nicolaï Troubetzkoy in Prague with the groundbreaking Principles of Phonology (1939). In the 1950s, they were followed by three influential linguists interested in language comparison and evolution, namely Roman Jakobson, first in Prague and from the 1940s on at Harvard, studying child language and aphasia from the point of view of functional phonology (1941) and developing a linguistic theory of translation (1959); André Martinet, the author in 1955 of L’économie des changements phonétiques2 following Trubetzkoy’s functional phonology (then in New York and in the 1960s in Paris); and Eugenio Coseriu, teaching in Montevidoe and from the 1960s on in Tübingen, the proponent of a functional view of grammar diverging from Saussure’s doxa by introducing the concept of (usage) norm between system and speech (1952) and rehabilitating the study of the evolution of language structures between synchronic stages (1958). At the same time, J. R. Firth founded the London School of Linguistics, which was to be the cradle of SFL, moulding the crucial notions of ‘system’ and ‘restricted language’ (renamed as ‘register’ by M. A. K. Halliday) and proposing categories devoted to analysing “contextually situated language events” (Léon, 2007).

1.2. In the 1970s and 1980s This second period is characterized by the developing study of the functional structure of the sentence in Prague by Jan Firbas and Peter Sgall, whose conclusions about the distribution of theme, rheme, topic and focus become a significant part of Simon Dik’s Theory of Functional


Jacques François

Grammar in Amsterdam (1978, 1997) and are later best theorized in the work of Knud Lambrecht, Information Structure and Sentence Form, in 1996. At the same time in Paris, Martinet developed his conception of functional syntax (1975), whereas Michael Halliday published in 1973 his Explorations in the Functions of Language, soon followed by his seminal work on syntax at the text level, Cohesion in English (in 1976 together with Ruqaiya Hasan). Two years later, in 1978, Dik published the first outline of his Functional Grammar, thus founding the Dutch School of Functional Grammar, whose views were soon disseminated through Benjamin’s Functional Grammar Series and the Amsterdam Working Papers in Functional Grammar. Some years later, in 1984, William Foley and Robert Van Valin set down the foundation of Role and Reference Grammar in their work Functional Syntax and Universal Grammar. At this same time, Christian Lehmann, who was then the main collaborator of Hansjakob Seiler in the research team UNITYP (Language Universals and Typology) in Cologne, wrote in 1982 his Thoughts on Grammaticalization, a crucial milestone in the genesis of external (or non-structural) linguistic functionalism, whose main proponent in Germany was to be Bernd Heine in the next decade. In 1979 Talmy Givón noted in Understanding Grammar that “today’s morphology is yesterday’s syntax”, a view already suggested by Franz Bopp, the main founder of the comparative grammar of Indo-European languages in the first half of 19th century in his so-called agglutination theory. And five years later Givón established the founding principles of the functional typology of languages in his Syntax: A Functional Typological Introduction in two volumes (1984–1990), followed by a fully revised second edition in the next decade. 1.3. From the 1990s to 2010s: Internal vs. External Functionalist Theories of Language The last decade of the 20th century is the most exciting for the simultaneous strengthening of internal (or structural) functionalism, mostly in Europe, and the blossoming of external (or non-structural) functionalism, mostly on the US West Coast and in Germany. This period is marked by the publication of the second edition of Michael Halliday’s Introduction to Functional Grammar (First ed. 1985) for SFL and of Kees Hengeveld’s Non-Verbal Predication in 1992, a milestone in functional typology of languages in the FG vein, together with the posthumous publication by the same Hengeveld of the second volume of Dik’s Theory of Functional Grammar in 1997, and for RRG by the publication of Van Valin’s Advances in RRG and especially of Syntax: Structure, Meaning,

Functional(ist) Theories of Language


and Function together with Randy LaPolla in 1997, which quickly became the seminal work of the emerging community of RRG researchers centred in Buffalo. Concerning the external vein of linguistic functionalism, Hansjakob Seiler together with Walter Premper published a crucial work in 1991 which was devoted to the dimension of participation and explained how every language organizes the predicate frames (or argument structures) as a result of the competition and cooperation of linguistic and extraneous pressures, whereas his younger colleague in Cologne, Bernd Heine, began an exploration of The Cognitive Foundations of Grammar (1997a), studying first auxiliarization in 1993 and then the grammatical expression of possession in 1997b. At the same time (more exactly from 1986 on), the domain of cognitive linguistics was enriched by ‘construction grammars’ (by Charles Fillmore and Paul Kay, George Lakoff, Adele Goldberg (2006), William Croft (2001), etc.), while ‘usage-based theories of language’ (by Ronald Langacker, Bryan MacWhinney, Joan Bybee, etc.) flourished in the USA, especially at West Coast universities. A crucial publication at the beginning of the new millennium, in 2003, is Chris Butler’s Structure and Function—a Guide to Three Major StructuralFunctional Theories in two volumes, the first devoted to the SFL, FG and RRG approaches to the simplex clause, and the second to the complex sentence, the discourse dimension and the high or low integration of psycholinguistic and sociolinguistic aspects. Butler’s ambitious comparison is soon reinforced by the publication of the third edition in 2004 and the fourth in 2014 of Halliday’s Introduction to Functional Grammar, from then on in collaboration with Christian Matthiessen; by that of Hengeveld’s and MacKenzie’s Functional Discourse Grammar in 2008; and by Van Valin’s Exploring the Syntax-Semantics Interface in 2006. Conversely, these works urged Butler, together with Francisco GonzalvezGarcia, to the publication in 2014 of Exploring Functional-Cognitive Space, a work based on a questionnaire submitted to all the major representatives of functional linguistics about the modelling of language systems by universal cognitive features and thus exploring the overlapping and affinity between the functional typology of language on the one hand and cognitive linguistics on the other hand. Concerning the diachronic and typological orientation of external functionalism, Bernd Heine published in 2002 together with Tania Kuteva a Word Atlas of Grammaticalization that summarizes two decades of research on that topic across hundreds of languages, and in 2011 Heine edited in collaboration with Heiko Narrog a state of the art volume with their Oxford Handbook of Grammaticalization based on the fascinating works of Paul Hopper, Elisabeth Traugott, Talmy Givón (2002, 2009), Joan Bybee, William Croft (2001), etc.


Jacques François

2. The Stance of Systemic Functional Linguistics Amongst Structural-Functional Language Theories This section will briefly consider three main theoretical frameworks in terms of the domains of investigation they favour and those they presently neglect. Butler (2003) is the main reference3 for this multilevel comparison. Table 2.1, assembled according to Butler’s observations, inventories the posture of SFL, Dik’s FG and van Valin’s RRG related to eight basic factors. What ensures the comparability of these three theories representing internal functionalism is their similar stance concerning factors 1 (language viewed as a communicative rather than as a computational tool) and 8 (language acquisition envisaged from a constructionist point of view).4 Concerning factors 2 (the rejection of the autonomy, arbitrariness and/or self-containedness of language) and 4 (the centrality of meaning and the secondary place of syntax), the goals of Halliday, Dik and van Valin are rather similar. Three factors reveal a stronger similarity between FG and RRG against SFG: • • •

3 concerning the integration of communicative competence, against its rejection in FDG; 6 related to the significance of psychological adequacy, against SFG’s attempt to interpret cognition in terms of language; and 7 concerning that of typological adequacy, whereas SFG focuses on dealing with particular languages, English, French, Chinese, etc., without any commitment to typological generalizations.

And, finally, factor 5 (the textual orientation) brings SFG and FG closer only since DFG’s development in Hengeveld and MacKenzie (2008), whereas RRG contends with focusing on three special components of discourse: information structure, clause connection and reference tracking. A last distinctive factor must be highlighted: whereas the social background is a significant part of the description in SFG, it fails in FG and RRG. Dirk Geeraerts emphasizes this distinctive feature of SFL: Within the group of functionalist frameworks, Systemic Functional Linguistics is the one that most distinctly follows up on this SOCIAL CONCEPTION OF LANGUAGE. Thinking about language in social, interactional terms suggests that the systemic descriptive and theoretical framework might be particularly suited for socially oriented types of linguistic investigation. (Geeraerts, 2010: 80)

Functional(ist) Theories of Language


Table 2.1 Eight factors outlining a cartography of the main interests of SFG, Dik’s FG and van Valin’s RRG (adapted from Butler, 2003) [p. 48] Systemic Functional Grammar . . . 1 2





7 8

[p. 39–40] Dutch Functional Grammar . . .

[p. 43] Role and Reference Grammar . . .

• is centrally concerned with language as communication • rejects the • favours partial • rejects the autonomy arbitrariness or motivation by of the linguistic self-containedness communicative system and takes of the linguistic factors the strong view system that everything in the grammar is functionally motivated • rejects the concept • tries to account for the communicative of communicative competence of the language user, or, competence alternately, attempts to characterize communicative competence • places semantics • regards meaning • considers semantics and pragmatics as central (concept and pragmatics at the heart of the of a semantically as central to the model, but also oriented approach, but has a (semantically ‘lexicogrammar’) recognizes syntax motivated) as one means for syntactic the realization of component meaning • is concerned • is text-oriented • is concerned with information with discoursestructure, clause related matters in connection and recent Functional reference tracking Discourse Grammar (see Hengeveld and MacKenzie, 2008) • is committed to psychological adequacy resp. • interprets cognition cognitively oriented explanation in terms of language rather than vice versa • emphasizes typological adequacy • is little interested in typological adequacy • favours a constructionist approach to language acquisition (only a program in FG)

3. SFL, Parallel Architecture and Cognitive Grammar in Contrast The ambition of ‘systemicity’ in SFL may be appreciated only by comparing it to other theories that aim for systemicity too insofar as they focus on the linking between expression and content (in Hjelmslev’s


Jacques François

terminology) or between syntax and semantics, as with Ray Jackendoff’s theory of Parallel Architecture (1997, 2002) and Ronald Langacker’s theory of Cognitive Grammar (2008). The main difference between the two former views is that for SFL (see Halliday and Matthiessen, 2014), the whole linguistic apparatus builds a continuous hypersystem from phonemes to textemes, whereas Jackendoff extends the generative principle beyond the syntax upstream to phonology and downstream to conceptual semantics in the form of three generators operating in parallel, yielding a Parallel Architecture. The difference between these two views and that of Cognitive Grammar is that according to Langacker the structures of human cognition ‘inform’ the grammar. In this conception, the challenge consists in accounting for the diversity of world languages despite the universal cognitive features permeating every grammar. For, as Talmy Givón emphasizes (see 1995, pp. 439–443), the construal of a protolanguage is facilitated by iconicity, but, conversely, arbitrary linguistic signs are crucial for elaborating language structures having the power to free the working memory of the interlocutors and thus to allow fast and effective communication. However, the three theories share a negative feature, namely their trifling involvement in the study of language dynamics. Actually Halliday and Matthiessen (2014, p. 68) evoke the topic of grammaticalization that they propose to deal with at three levels: a) that of language acquisition (ontogenetic time) b) that of the evolution of language faculty (phylogenetic time), which they regard as unattainable for the linguist, and c) that of the discourse flow (logogenetic time), an original view but apparently restricted to the study of event anaphora. (see 2014, p. 68) And the authors conclude (p. 47): So when we talk of the ‘system’ of language, as the underlying potential that is instantiated in the form of text, we are in effect theorizing a language as the outcome of ongoing grammaticalization in all these three dimensions of time. As we will see in what follows, Jackendoff deals with the phylogenetic dimension suggested by Halliday and Matthiessen (see Jackendoff, 2002, pp. 231–266). It was this exploration that led in 2005 to a decisive article by Jackendoff together with Steven Pinker (see Pinker & Jackendoff 2005) as a reply to Hauser, Chomsky and Fitch (2002), who were trying to distinguish between Broad and Narrow Language Faculties. However,

Functional(ist) Theories of Language


concerning the dynamics of language, this is only explored in his work on the emergence of ‘Constructions’.5 In what follows, SFL is considered in a comparative perspective along with Jackendoff’s framework and Langacker’s cognitive theory. 3.1. Situating ‘System’ in SFL Before moving on to the comparison of SFL to Jackendoff’s (structural) Parallel Architecture and Langacker’s (non-structural) cognitive linguistics, this section provides a brief overview of where system is situated with SFL theory. Halliday and Matthiessen draw five dimensions of language, as illustrated in Table 2.2 (2014, p. 20). • •

• • •

structures on the syntagmatic axis, hierarchized in a series of levels; systems on the paradigmatic axis, categorized according to their delicacy with on the one hand the closed system of grammar and on the other hand the open system of lexis; stratifications, with various sorts of fulfilment (semantic, lexicogrammatical, phonological or phonetic); instantiations, specifying instances, types or potentials; and metafunctions, categorized as ideational (subdivided into logical and experiential), interpersonal or textual.

SFL is labelled as systemic because the speaker is regarded—in Willem Levelt’s terms (1989)—as (1) a conceptualizer, (2) a formulator and (3) an articulator:6 the organizing principle adopted is that of system: the grammar is seen as a network of interrelated meaningful choices. In other

Table 2.2 The dimensions (forms of order) in language and their ordering principles Dimension




structure (syntagmatic order)





system (paradigmatic order) stratification

clause ~ group/phrase ~ word ~ morpheme [lexicogrammar] ; tone group ~ foot ~ syllable ~ phoneme [phonology] grammar ~ lexis [lexicogrammar]








semantics ~ lexicogrammar ~ phonology ~ phonetics potential ~ subpotential/instance type ~ instance ideational [logical ~ experiential] ~ interpersonal ~ textual


Jacques François words, the dominant axis is the paradigmatic one: the fundamental components of the grammar are sets of mutually defining contrastive features. (p. 49)

These system choices have to be realized because “the system . . . is the potential that lies behind the text” (p. 49). Analysing a text means showing the functional organization of its structure and especially “that meaningful choices have been made, each one seen in the context of what might have been meant but was not” (p. 24). Realization is “one manifestation of a general relationship that pervades every quarter of language” because “a language is a stratified system” (ibid.). The basic architecture of SFL consists of three systems: •

The first is that of theme, concerned with the partition in theme vs rheme of the clause viewed as message (its textual metafunction). In the tradition of the “functional perspective of the sentence” elaborated by the Prague School of Linguistics (F. Danes, J. Firbas, P. Sgall), the theme is defined as “the orientation chosen for the message” (p. 43). The second is that of mood, concerned with the organization of the clause as an exchange instance. It is defined as “the primary interpersonal system of the clause—the grammaticalization of the semantic system of speech function” (2014, p. 142, footnote 5), and Halliday and Matthiessen comment on the various moods amongst which the speaker has to make a choice in a major clause (one involving a predicator in its structure): A major clause is either indicative or imperative in mood, if indicative, it has a Finite (operator) and a Subject. An indicative clause is either declarative or interrogative (still in mood), if declarative, the subject comes before the Finite. An interrogative clause is either yes/no type or WH-type, if yes/no type, the Finite comes before the Subject, if WH-type is has a Wh element. (2014, p. 23)

The third is that of transitivity, concerned with the linguistic expression of process types at the level of the clause viewed as representation (its experiential metafunction). The clause represents a process together with its participants and circumstances, if any. This system displays two viewpoints, that of the impact of an agent against a patient (what R. Langacker, 2008, calls his “energy transfer”), and that of its semantic domain, either material, mental or relational.

Finally, Table 2.3 from Halliday and Matthiessen (2014, p. 83) summarizes the metafunctional and structural status of the three basic systems.

Functional(ist) Theories of Language


Table 2.3 Three lines of meaning in the clause Metafunction

Clause as . . .



textual interpersonal

message exchange





Theme ^ Rheme Mood [Subject + Finite] + Residue [Predicator (+ Complement) (+ Adjunct)] process + participant(s) (+ circumstances), e.g. Process + Actor + Goal

By integrating the three textual, interpersonal and experiential metafunctions, SFL excludes any separation between semantics and pragmatics. For example, Halliday considers the information structure of the sentence as a major concern, as does the ‘functional sentence perspective’. This differs from Jackendoff’s ‘Parallel Architecture’ because Jackendoff, following Chomsky, holds a restrictive view of linguistics as the exploration of language competence, and leaves the study of language performance to the researchers in functional grammar, discourse analysis and sociolinguistics. 3.2. ‘System’ in Jackendoff’s Parallel Architecture (1997, 2002) Jackendoff set the foundations of his theory in 1983 (Semantics and Cognition) by introducing a semantic generator beside the syntactic one. He referred to his bigenerative architecture as “X-bar semantics” in 1990 (Semantic Structures). Then in 1997 he completed his model (The Architecture of the Language Faculty) by adjoining a third phonological generator. Figure 2.1, adapted from his 2002 work, Foundations of Language, delivers the three ‘integrative processors’, interrelated through three internal interfaces between phonological, syntactic and conceptual structures. By mentioning the place of working memory and introducing dynamic external interfaces from audition, to vocalization and to perception and action, Jackendoff aims at implementing the parallel grammar as a ‘processing architecture’. One major difference between this Parallel Architecture and Cognitive Grammar, which will be discussed below, is that whereas Jackendoff has incorporated a priori relations of mutual information between the three generative components (phonology, syntax and conceptual semantics), Langacker conceives of the lexicogrammatical system as essentially informed by universal cognitive structures, like those of spatiality, temporality, subjectivation or energy transfer modelling the systems of semantic roles, etc.). However, we will now return to the structural aspects


Jacques François

Interface processor from audition

Phonological integrative processor

Syntactic integrative processor

Syntactic structures

Phonological structures

Conceptual integrative processor

Conceptual structures

LINGUISTIC WORKING MEMORY Interface processor to vocalization



interface processor(s)

interface processor(s)

Interface processors to perception and action


interface processor(s)

Figure 2.1 Jackendoff’s Parallel Architecture (2002, p. 125)

of ‘system’ in SFL before considering the comparison with cognitive grammar. 3.3. Back to the Structuration of the ‘Systems’ in SFL In relation to the ‘integrative processors’ of Jackendoff’s Parallel Architecture, one must notice that the ‘systems’ of SFL turn out to be generative processors as well. However, they do not operate on the syntagmatic axis as in generative syntax (with A rewritten as B + C) but on the paradigmatic axis (with A specified as either B or C). In SFL, a system is a set of options, where one option is selected in an environment where the other option could have been selected but was not; each system includes an entry condition and an associated probability of occurrence for each option.7 As an example, the Polarity system (a functional and not a syntactic category) shown in Figure 2.2 illustrates two parallel systems, NOMINAL GROUP FUNCTION and CLAUSE FUNCTION, which are both available if the option ‘specialized’ is selected. Note that the values 0.9 and 0.1 attached to ‘positive’ and ‘negative’ respectively indicate the general probability frequency; i.e. most clauses are expressed with positive polarity. It is possible to reformulate the system as rules of alternating specification as shown in Figure 2.2. As suggested by the combinatorial notations am, an, bm and bn, the functional category of Polarity has various instantiations across different places in the general lexicogrammatical system. What we can take

Functional(ist) Theories of Language


positive 0.9 clause


generalized negative 0.1

NEGATIVE TYPE specialized

am : none no + N ; neither (+N) an : at no time ; under no circumstances ; for no reason ; in no way bm : no-one ; nobody ; nothing

as Deictic (a) NOMINAL GROUP as Thing (b) FUNCTION


in participation (m) in circumstance (n)

bn : never ; nowhere ; nowise ; seldom

Figure 2.2 The system of Polarity (Halliday and Matthiessen, 2014, p. 23, Figures 2.1–2.8)

Table 2.4 Reformulation of the system of Polarity as rules of alternating specifications 1. 1.2. 1.2.2.

polarity negative polarity [type] negative polarity [specialized type]

negative polarity [specialized type] [NP-function] negative polarity [specialized type] [Clause-function]

→ positive | negative → generalized | specialized → [NP-function] & [Clause-function] → deictic | thing-person → in participation | in circumstance

from this is that the main ‘systemic’ feature of SFL serves therefore to elaborate graphs devoted to the progressive specification of major functional categories whose instantiations are conceived of as lexical or morphosyntactic. 3.4. Langacker’s Lexicogrammatical Systems as Reflexes of Perceptual and Cognitive Systems At this point, we turn our attention to Langacker’s theory of language in order to consider lexicogrammatical systems in a non-structural or external orientation in contrast to SFL’s more structural, internal orientation. In his admirable work of linguistic metatheory, Structure and Function, Butler (2003, pp. 56–57) defines Cognitive Grammar as: (1) a usage-based theory (like other Construction Grammars), (2) viewing human language as a communication tool (like all functional grammars), (3) regarding communication devices like metaphor as crucial (unlike most functional grammars, but like Lakoff’s Construction Grammar), (4) rejecting the autonomy of the language system (like Jackendoff), (5) postulating cognitive motivations (as in Hopper’s ‘emergent grammar’) and (6) conceiving


Jacques François

of syntax as a symbolic link between semantics and phonology (like Jackendoff). In short, Langacker views Cognitive Grammar as construing differentiated lexicogrammatical systems from universal cognitive features by weighting cross-cultural preferences. As an illustration one may mention the two main ways of syntactically combining a motion and a change of place caused by that motion according to Talmy (1985). Romance languages pick up as the finite verb the predicate conveying the caused change of place and express the causing motion as a gerundive or a prepositional phrase as in example (1) from French. Conversely, Germanic languages, including English, pick up as the finite verb the predicate of the causing motion and express the change of place with the means of a directional preposition or adverb as in (2) in English (across) and (3) in German (über . . . hinüber). A transitive expression is possible in German as shown in (4) by incorporating the preposition durch (synonymous with über in this context) in the morphological substance of the verb. (1) (2) (3) (4)

Paul a traverse la Manche en nageant/à la nage. Paul swam across the channel Paul schwam über den Ärmelkanal hinüber Paul durchschwamm den Ärmelkanal

While there is perhaps a similar motivation in Cognitive Grammar as compared to SFL in terms of accounting for a relationship between content and expression (see Figure 2.1), the orientation of this motivation is fundamentally different. For Halliday it is that of socially oriented meanings, and for Langacker it is that of cognitively oriented meanings. In other words, Langacker is interested in motivating the lexicogrammar in terms of cognitive processes, whereas Halliday is interested in motivating it through social interaction.

4. From ‘Systemic’ to ‘Systemics’ In this section I would like to question the compatibility between the two senses of ‘system(ic)’ in SFL and in ‘systemics’ or ‘General System Theory’ (section 4.1), as well as SFL’s ability to conceive of natural languages as dynamic systems as both DuBois (1985) and Hopper (1987) have (section 4.2). 4.1. The General System Theory of Ludwig von Bertalanffy In 1947 the biologist Ludwig von Bertalanffy founded a new large ‘theory of everything’. He wrote: “We postulate a new discipline called general system theory. General System theory is a logico-mathematical field whose task is the formulation and derivation of those general principles

Functional(ist) Theories of Language


that are applicable to ‘systems’ in general” (von Bertalanffy 1947, 1955; reprinted in Bertalanffy, 1968, pp. 32, 253). As a matter of fact, the birth of General System Theory had to do with the success in the same time of Norbert Wiener’s (1948) new ‘cybernetics’ (the general theory of government). The motivation of the two views was different: for Wiener it was the development of self-directing missiles during World War II, and for Bertalanffy it was the modalization of the evolution of species following the so-called New synthesis merging Darwin’s natural selection and Mendel’s genetic inheritance in the 1930s. But the main common idea was that of ‘homoeostasis’ (or more commonly of ‘feedback loop’), that is, “the tendency towards a relatively stable equilibrium between interdependent elements, especially as maintained by physiological processes” (Oxford Compact Dictionary). The whole excerpt borrowed from an article of 1972 in which Bertalanffy recalls the genesis of General System Theory or systemics with regard to cybernetics merits careful reading: In the meantime a different development had taken place. Starting from the development of self-directing missiles, automation and computer technology, and inspired by Wiener’s work, the cybernetic movement became ever more influential. Although the starting point (technology versus basic science, especially biology) and the basic model (feedback circuit versus dynamic system of interactions) were different, there was a communality of interest in problems of organizations and teleological behavior. Cybernetics too challenged the ‘mechanistic’ conception that the universe was based on the ‘operation of anonymous particles at random’ and emphasized ‘the search for new approaches, for new and more comprehensive concepts, and for methods capable of dealing with the large wholes of organisms and personalities’. Although it is incorrect to describe modern systems theory as ‘springing out of the last war effort’—in fact, it had roots quite different from military hardware and related technological developments—cybernetics and related approaches were independent developments which showed many parallelisms with general system theory. (Bertalanffy, 1972, pp. 413–414) 4.2. Emerging Linguistic Systems The crucial question is now whether some present linguistic theories may be ‘systemic’ in the sense of Bertalanffy’s ‘systemics’. It cannot be said to be the case with SFG, because SFG is a structural or internal theory of language functions in which external motivations or pressures able to shape the architecture of natural languages are not questioned. Therefore, we must pay attention to theories integrating such motivations, in other words external theories of language functions. At least two


Jacques François

pioneers of that view are noteworthy, namely John DuBois with his paper of 1985 entitled “Competing Motivations” in John Haymann’s Iconicity in Syntax, and Paul Hopper with his contribution to the publications of the Berkeley Linguistic Society in 1987 entitled “Emergent Grammar”. On his homepage,8 DuBois summarizes his view in “Competing motivations” as follows: A simple exchange of utterances between two speakers contains a virtual microcosm of meaning, structure, prosody, pragmatics, interpretation, interaction, cognition—all the issues that linguists have found interesting enough to build disciplines and theories around. Understanding the organization of complexity in language provides deep intellectual challenges. I find it interesting to ask how grammars coordinate different layers of function—expressing semantic relations and managing information, for example—as they co-exist and compete for control of the organization of linguistic structures, like the clause. I see grammar as resolving competing motivations in systematic ways, thus driving the self-organization of grammatical systems and the emergence of complexity in linguistic structure—a really exciting new perspective for linguistics today. Please notice first the topic of “self-organization of grammatical systems” that obviously derives from the concept of the feedback loop born in Wiener’s cybernetics and developed in Bertalanffy’s systemics, and then that the idea of competing and cooperating motivations has been exploited in a branch of generative linguistics, so-called Optimality Theory, first in phonology and later in syntax (see Geraldine Legendre and Paul Smolensky’s ‘Harmonic Grammar’ from 1990 on). As to Paul Hopper, he develops similar ideas in his paper of 1987: I meant to suggest that structure, or regularity, comes out of discourse and is shaped by discourse as much as it shapes discourse in an on-going process. Grammar is hence not to be understood as a pre-requisite for discourse, a prior possession attributable in identical form to both speaker and hearer. Its forms are not fixed templates, but are negotiable in face-to-face interaction in ways that reflect the individual speakers’ past experience of these forms, and their assessment of the present context, including especially their interlocutors, whose experiences and assessments may be quite different . . . The notion of EMERGENCE is a pregnant one. It . . . takes the adjective emergent seriously as a continual movement towards structure . . . a view of structure as always provisional, always negotiable, and in fact as epiphenomenal, that is, at least as much an effect as a cause. (p. 142)

Functional(ist) Theories of Language


As a matter of fact, the concept of emergent grammar is not persuasive always and everywhere, but only in social configurations where strong innovative external pressures challenge weak conservative pressures. For instance, that is not the case in present French because the prestige of metropolitan French inhibits the influence of Belgian, Swiss, Canadian or African French, but it begins to be the case for English since there exist bilingual dictionaries of British vs Singaporean English strongly influenced by Malay. And it was obviously the case in the fifth to eighth centuries in western Europe with the fast disintegration of spoken Latin and the emergence of various Romance languages: French, Italian, Spanish, Portuguese, Occitan, Catalan, Rheto-romance and, at the eastern side, Romanian. A decade later, in 1999, Frederick Newmeyer, an influential proponent of a generative syntax integrating functional features, made a counterproposition against DuBois’ (1985) and Hopper’s (1987) views in a paper entitled “Some Remarks on the Functionalist Formalist Controversy in Linguistics”. Newmeyer (1999, p. 474) acknowledges the reality of functional motivations, but he does not conceive of them as able to directly fashion linguistic structures. Therefore, trying to offer a structuralized revision of DuBois’ idea of competing motivations, he inserts, between motivations and structures, a structural system shaped by these very motivations and conversely shaping these structures. Newmeyer’s view is obviously persuasive when the conservative pressures successfully challenge the innovative and therefore disintegrating pressures (see Figure 2.3),9 whereas DuBois’ and Hopper’s view of direct competition between grammatical structures and functional motivations is convincing when the conservative pressures are no longer able to balance the innovative ones. That happened, for instance, at the end of fifth century, when the collapse of the Western Roman Empire led to

a. Structurea

Functional motivationa


Functional motivationb


Functional motivationc


Functional motivationd

b. Structurea Structureb Structurec Structured

Functional motivationa Structural system

Functional motivationb Functional motivationc Functional motivationd

Figure 2.3 Between functional motivations and linguistic structures: (a) DuBois (1985); (b) Newmeyer (1999)


Jacques François

the disintegration of the entire Roman administration and school system. Therefore, a catastrophic level of illiteracy lasted until the eighth century. In these three centuries the linguistically unified Romania fragmented into several dialects which finally became independent languages (in conformity with Hopper’s view), while Byzantine Greek remained the common language of the Eastern Roman Empire until its collapse one millenary later, because the administration of the empire was robust and ensured the conservation of the language structures from generation to generation through a well-organized education system (in conformity with Newmeyer’s view).

5. Conclusion In conclusion, I regard SFG/SFL as being actually ‘systemic’ in the usual sense of structural linguistics. But this view of systemicity is irreconcilable with that of Bertalanffy’s systemics or General system theory, because Halliday’s view is a structural one, assuming that the linguist is concerned only by motivations inherent to the semiotic system and not by motivations between this semiotic system, the embodied cognition and the extraneous world. Conversely, the view of external linguistic functionalism illustrated by the works of DuBois (1985), Hopper (1987), Langacker (2008), Lakoff (1987) and Croft and Cruse (2004) illustrates Bertalanffy’s systemics applied to the linguistic realm. I recall as evidence in favour of that conclusion two extracts from DuBois and Hopper: I see grammar as resolving competing motivations in systematic ways, thus driving the self-organization of grammatical systems and the emergence of complexity in linguistic structure.10 PAUL HOPPER: The notion of Emergent Grammar is meant to suggest that structure, or regularity, comes out of discourse and is shaped by discourse in an ongoing process . . . The notion of emergence . . . takes the adjective ‘emergent’ seriously as a continual movement towards structure. (1987, p. 142) JOHN DUBOIS:

Notes 1. The grammar of errors: introduction to a functional linguistics, assimilation and diversification, brevity and invariability, expressivity. 2. The economy of sound changes 3. See especially vol.1, p. 48 for SFG, p. 39–40 for Dik’s FG and p. 43 for van Valin’s RRG. 4. Another (negative) factor brings the three theories close, namely their common lack of interest in dischronic development. 5. See Jackendoff and Goldberg’s paper in 2004 about the so-called family of resultative constructions in English

Functional(ist) Theories of Language


6. The authors distinguish between instantiation and realization of a text: “The defining criterion is instantiation: text as instance. But realization comes in because what becomes accessible to us is the text as realized in sound or writing” (2014, p. 51). They evoke at the same place the impossible access to “instances of language at higher strata—as selections in meaning, or even in wording”. The analogy with Levelt’s psycholinguistic architecture is striking: selections in meanings are related to Levelt’s Conceptualization level, selections in wording to the first layer of his Formulation level, instantiation (involving word order and inflections) to its second layer, and realization to its Articulation level. Here, realization is seen as an “additional property” (ibid.), like articulation in Levelt’s view. 7. See Halliday (2013) for a detailed account of systems, including the notation and what systems represent. 8. 9. A similar concern is at the root of the extension of the theory of phonological optimality to syntax: diverging syntactic priorities are put in competition until the most effective syntactic structure emerges. 10. See

References Bally, C. (1936) Linguistique générale et linguistique française, 4th ed. Berne: Francke. Bertalanffy, L. (1968) General System Theory: Foundations, Development, Applications. New York: George Braziller. Bertalanffy, L. (1972) The history and status of general system theory. The Academy of Management Journal, 15(4), pp. 407–426. Bühler, K. (1934) Sprachtheorie (Theory of Language). Frankfurt/Main: UTB. Butler, C. (2003) Structure and Function: A Guide to Three Major StructuralFunctional Theories. Amsterdam: Benjamins. Butler, C. and Gonzalvez-Garcia, F. (2014) Exploring Functional-Cognitive Space. Amsterdam: Benjamins. Coseriu, E. (1952) Sistema, Norma y Habla. In Teoría del lenguaje y lingüística general. Madrid: Gredos. Coseriu, E. (1958) Syncronia, Diacronia y Historia—El Problema del cambiolingüistico. Ciencias XV, pp. 211–355, Montevideo. Croft, W. (2001) Radical Construction Grammar. Chicago: Chicago University Press. Croft, W. and Cruse, D. A. (2004) Cognitive Linguistics. Cambridge, UK: Cambridge University Press. Dik, S. (1978) Functional Grammar. Dordrecht: Foris. Dik, S. (1997) The Theory of Functional Grammar (2 vol.). Berlin: De Gruyter. Dubois, J. (1985) Competing motivations. In J. Haiman (Ed.), Iconicy in Syntax. Amsterdam: Benjamins, pp. 343–366. Foley, W. and Van Valin, R. (1984) Functional Syntax and Universal Grammar. Cambridge: Cambridge University Press. Frei, H. (1929) La grammaire des fautes. Rennes: Ennoïa. Reprinted 2007. Geeraerts, D. (2010) Recontextualizing grammar: Underlying trends in thirty years of cognitive linguistics. In E. Tabakowska and M. L. Łukasz Wiraszka (Eds.), Cognitive Linguistics in Action: From Theory to Application and Back. Berlin and New York: De Gruyter, pp. 71–102.


Jacques François

Givón, T. (1979) Understanding Grammar. Chicago: Chicago University Press. Givón, T. (1984–1990) Syntax: A Functional Typological Introduction. Amsterdam: Benjamins. Givón, T. (1995) Functionalism in Grammar. Amsterdam: Benjamins. Givón, T. (2002) Bio-Linguistics: The Santa-Barbara Lectures. Amsterdam: Benjamins. Givón, T. (2009) The Genesis of Syntactic Complexity. Amsterdam: Benjamins. Goldberg, A. (2006) Constructions at Work: The Nature of Generalization in Language. Oxford: Oxford University Press. Halliday, M. A. K. (1973) Explorations in the Functions of Language. London: Arnold. Halliday, M. A. K. (1985) An Introduction to Functional Grammar, 2nd ed. London: Arnold. Halliday, M. A. K. (2013) Meaning as choice. In L. Fontaine, T. Bartlett and G. O’Grady (Eds.), Systemic Functional Linguistics: Exploring Choice. Cambridge: Cambridge University Press, pp. 15–36. Halliday, M. A. K. and Hasan, R. (1976) Cohesion in English. London: Longman. Halliday, M. A. K. and Matthiessen, C. M. I. M. (2014) Introduction to Functional Grammar, 4th ed. London: Routledge. Hauser, M., Chomsky, N. and Fitch, T. (2002) The faculty of language: What is it, who has it, and how did it evolve? Science, 298, pp. 1569–1579. Heine, B. (1993) Auxiliaries: Cognitive Forces and Grammaticalization. Oxford: Oxford University Press. Heine, B. (1997a) The Cognitive Foundations of Grammar. Oxford: Oxford University Press. Heine, B. (1997b) Possession: Cognitives Sources, Forces, and Grammaticalization. Cambridge, UK: Cambridge University Press. Heine, B. and Kuteva, T. (2002) Word Atlas of Grammaticalization. Cambridge, UK: Cambridge University Press. Heine, B. and Narrog, H. (Eds.). (2011) The Oxford Handbook of Grammaticalization. Oxford: Oxford University Press. Hengeveld, K. (1992) Non-verbal Predication. Amsterdam: Benjamins. Hengeveld, K. and Mackenzie, L. (2008) Functional Discourse Grammar. Cambridge, UK: Cambridge University Press. Hopper, P. (1987) Emergent grammar. Proceedings of the Thirteenth Annual Meeting of the Berkeley Linguistics Society, pp. 139–157. Jackendoff, R. (1983) Semantics and Cognition. Cambridge, MA: MIT Press. Jackendoff, R. (1997) The Architecture of the Language Faculty. Cambridge, MA: MIT Press. Jackendoff, R. (2002) Foundations of Language: Brain, Meaning, Language, Evolution. Oxford: Oxford University Press. Jackendoff, R. and Goldberg, A. (2004) The English resultatives as a family of constructions. Language, 80(3), pp. 532–568. Jakobson, R. (1941) Child Language, Aphasia, and Phonological Universals. The Hague and Paris: Mouton. Jakobson, R. (1959) On linguistic aspects of translation. In L. Venuti (Ed.) (2000) Translation Studies Reader, 2nd ed. New York: Routledge, pp. 113–119. Lakoff, G. (1987) Women, Fire, and Dangerous Things: What Categories Reveal About the Mind. Chicago: Chicago University Press.

Functional(ist) Theories of Language


Lambrecht, K. (1996) Information Structure and Sentence Form. Cambridge: Cambridge University Press. Langacker, R. (2008) A Basic Introduction to Cognitive Grammar. Cambridge: Cambridge University Press. Lehmann, C. (1982) Thoughts on Grammaticalization. Munich: Lincom. Léon, J. (2007) From linguistic events and restricted languages to registers. Firthian legacy and Corpus linguistics. The Henry Sweet Society Bulletin, 49, pp. 5–26. Levelt, W. J. W. (1989) Speaking: From Intention to Articulation. Cambridge, MA: MIT Press. Martinet, A. (1955) L’économie des changements phonétiques. Bern: Francke. Martinet, A. (1975) Studies in Functional Syntax. Munich: Fink. Newmeyer, F. (1999) Some remarks on the functionalist-formalist controversy in linguistics. In M. Darnell, E. Moravcsik, M. N. Newmeyer and K. Wheatley (Eds.), Functionalism and Formalism in Linguistics, vol. 1. Philadelphia: John Benjamins, pp. 469–486. Pinker, S. and Jackendoff, R. (2005) The faculty of language: What’s special about it? Cognition, 95, pp. 201–236. Seiler, H. and Premper, W. (Eds.). (1991) Partizipation: das sprachliche Erfassen von Sachverhalten. Tübingen: Narr. Talmy, L. (1985) How language structures space. In H. L. Pick and L. P. Acredolo (Eds.), Spatial Orientation: Theory, Research, and Application. New York: Planum Press, pp. 225–282. Troubetzkoy, N. (1939) Grundzüge der Phonologie. Travaux du Cercle Linguistique de Prague 7, Prag. Engl. transl. Principles of Phonology (1969). Berkeley and Los Angeles, CA: University of California Press. Van Valin, R. (Ed.). (1993) Advances in Role and Reference Grammar. Amsterdam: Benjamins. Van Valin, R. (2006) Exploring the Syntax-Semantics Interface. Cambridge: Cambridge University Press. Van Valin, R. and Lapolla, R. (1997) Syntax: Structure, Meaning and Function. Cambridge: Cambridge University Press. Wiener, N. (1948) Cybernetics, or Control and Communication in the Animal and the Machine. Cambridge, MA: MIT Press.


Rethinking (Context of) Culture in Systemic Functional Linguistics1 Tom Bartlett

All social sciences today are tackling the need to understand globalisation as a set of forces and changes—political, social, economic, environmental and cultural. The causes and effects of problems regularly cross boundaries that were previously treated as fixed and absolute. So do the resources to deal with them. With hindsight, it is now clear that hose boundaries were always less absolute than they seemed, and the flows were always there. Hodge (2017, p. xii)

1. Introduction Culture or, more specifically, the context of culture, is a central concept within the “extravagant architecture” of Systemic Functional Linguistics (SFL) for two principal reasons, one of which relates to the idea of language as a social semiotic (Halliday, 1978), the theoretical cornerstone of SFL, and the other to SFL as an appliable linguistics and its use within discourse analysis (Bartlett, in press). From the theoretical side, linguistic behaviour is seen as the realisation of, or the means of performing, the abstract meanings that comprise a particular culture, with every semantic alternation corresponding to a culturally meaningful distinction. Within applied work, the meaning of each instance of linguistic behaviour, whether the unit of analysis is an utterance, a text or a register, is only understood against the backdrop of the other options that were available but not chosen. Ultimately, these two perspectives are complementary, simply two sides of the same coin as the distinction between langue and parole disintegrates. This social semiotic approach to language is further developed in SFL in line with Halliday’s classical Marxist view that the material conditions of a society determine its organisation and development, with different social groups displaying systematically different behaviours as a function of their relationship to the means of production—in other words, their social class. From these ideas it logically follows that different social groups will consistently put language to different uses and develop different repertoires of use from within the

Rethinking (Context of) Culture


total set of options offered by the linguistic system. In SFL terminology, register is the term for linguistic variation according to use, and code is the term for linguistic variation according to user (Halliday, 2007a [1964]). However, the differential access of social groups to distinct registers means that the two concepts are inextricably bound up. The accepted corollary of this perspective is that an entire social system, or culture, is the set of relations between the different social groups that comprise it, while, in SFL terms, the language system is a similarly interconnected set of meanings. And from this it follows that an internally differentiated but ultimately unitary linguistic system (or language) can be mapped onto an internally differentiated but ultimately unitary cultural system. This is the standard SFL model of the relationship between language and culture, largely unchanged since the 1960s: an essentially structuralist model that is represented in Figure 3.1 and discussed in more depth below. However, developments in sociology and in social linguistics, often with poststructural and/or neo-Marxian orientations, have challenged the notion that either a language or a culture can ultimately be closed or bounded in this way. From the sociological perspective, crosscutting social groups can be formed around temporary or enduring affiliations according to such variables as gender, social values and ethnicity as well as class, while, from the linguistic perspective, individuals bring the social and linguistic resources they have accrued from past encounters and negotiate a linguistic common ground accordingly. In an age typified by accelerations in global communication and transnational flows, creating a context of superdiversity (Vertovec, 2007), the social and linguistic resources that come into contact within individual situations are derived from increasingly diverse sources. Through the situated negotiation of these variables, the participants create unique, though non-arbitrary, microcultures which relate to each other not in a hierarchical part-whole arrangement, but through multiple interlocking chains that never reach closure but orient, according to the demands of the situation, to more generalised or higher order systems at a variety of scales and with respect to diverse centres of normativity at these different scales. Such a view of cultures and microcultures, therefore, potentially presents a problem for SFL as a linguistic theory and, by implication, as an appliable linguistics. However, in this chapter I shall take the view that the standard SFL model of the language-culture relation as summarised above and outlined in more detail below lags behind existing theoretical work in the discipline so that, through following the trajectory of existing ideas, a modified and compatible model can be developed. As with the more formal aspects of language description, we can expect that the model developed should be descriptively adequate, typologically adequate and theoretically adequate (cf. Wray’s [2014] discussion of Butler’s [2009] criteria for a functional theory of language). And given the importance SFL attaches to being an appliable theory, we can

28 Tom Bartlett add to these the requirement that the model developed should have applicational validity. Taking descriptive and typological validity together, such a mapping must satisfy two pairs of competing parameters: it must be stable enough to allow generalisations across individual events and diverse contexts while dynamic enough to allow for variation across situation, time and space; and it must be comprehensive enough to account, in some way, for the totality of instances while also being open-ended enough to admit that no such totality exists. With regard to theoretical adequacy, we can consider a different complementarity: on the one hand, that functional theories of grammar, including SFL, should accommodate and respond to insights and developments across a range of what can be called social-interactional approaches to language; and on the other, that these social-interactional approaches cannot afford to ignore SFL’s conception of language as a metastructural system—that is, not as a set of structures, but as the underlying means by which such structures emerge through the articulation of functionally significant elements. In this chapter I will take as my starting point the SFL notion of (context of) culture and, through a dialogue with other linguistic approaches, interrogate the model in terms of stability/dynamicity and comprehensiveness/ open-endedness in an attempt to produce a more descriptively adequate model for applied linguistic analysis that combines the insights of recent developments in interactional sociolinguistics with the foundational SFL conception of language as a stratified and functionally diverse semiotic system. Such a combination of theoretical frameworks and methodological approaches offers an opportunity for analysing not only novel situations as they unfold in discourse but also the means and trajectories by which these instances potentially become types as new ‘cultures’ develop and stabilise or disarticulate and atrophy at different scales. I will end the chapter with a brief consideration of applicational validity and how linguists can use this model to relate the linguistic features which comprise each text, and which encode social beliefs and relationship systems, outwards to the social conditions of the text’s production and uptake and the meanings that are ‘at risk’ (i.e. most likely to be realised) within the specific juncture and according to the differential social positions of the speakers.

2. The SFL Model of Language and Culture, Outline and Problems In this section I outline the model of language and society as developed in SFL, chiefly by Halliday in the 1960s and 1970s, but with important later additions by Hasan. I will point to what I see as the problems with the model on the basis of developments in linguistics, social theory and society itself. Many of the ideas I pick up on here, as suggested above, are already apparent in SFL writings, most notably those of Hasan, that have

Rethinking (Context of) Culture


problematised the existing architecture. The suggestions I put forward in Section 3, drawing on recent work in interactional sociolinguistics on scales and centres and Laclau and Mouffe’s (1985) neo-Gramscianism, are therefore seen as in many ways the logical continuation of this theoretical trajectory rather than a departure from it. Consequently, in that section I hope not only to demonstrate how the architecture of SFL can benefit from these theoretical innovations but also to suggest that these other perspectives and theoretical orientations can benefit from a consideration of the SFL architecture. The SFL modelling of language and culture drew inspiration from (and, I will argue, is ultimately limited by) a classical Marxist materialdialectic conception of a society as a systemically integrated and ultimately bounded whole onto which differences in language, according to both use and user (Halliday, 2007a [1964]), can be isomorphically mapped. Differences according to use are labelled registers, and cover such diverse categories as legal documents and greetings cards (Halliday, 2007a [1964], p. 24); differences according to user, or at least those beyond the superficial differences of dialect, are labelled codes (Halliday, 2007b [1971], pp. 58–59; Bernstein, 2000), with the class system and the complementarities of the relation between labour and capital seen as the principal regulating factor of social differentiation. Given this goal of mapping correlations between variation in language behaviour with these different uses and users from the materialist dialectical perspective, the model is characterised by a focus on regularities over deviations, and categories over instances (Halliday, 2007a [1964], p. 18), with the categories subsumed into a unitary context of culture. The model therefore masks fundamental differences and incompatibilities and underplays the degree of diversity within a ‘language community’, if such a bounded entity could ever be claimed to exist. Figure 3.1 captures in two-dimensional form the essentials of the model of language and culture that underpins SFL theory and description. It operates along two axes: horizontally, a continuum from the underlying system to specific instances, with scales of recognisable types along the way; and vertically, from culture to the lexicogrammar, with semantics (what we can do through language) as the interface between the social context and language form. In this way, moving along the horizontal axis at the stratum of culture, situation types are seen as recognisable generalisations from individual situations, as instances of cultural activity, and the cultural system is seen as the sum of such potential types (Halliday, 2007a [1971], p. 18, echoed by Martin and Rose, 2008, Chapter 6). At each point along this cline the cultural can be related to the linguistic, such that individual situations are realised as texts; situation types are realised as registers; and the cultural system, the entire behavioural potential of a language community, is realisable through the language system as the entire language potential.

30 Tom Bartlett

Figure 3.1 Relations between context and language: System and instance

It is important to note at this point that the relations set out along the horizontal axes do not represent typologies: the cultural system is not made up of discrete and insulated situation types nor the language system of discrete and insulated registers. Rather, as we move along the cline from system to instance we specify and hence restrict the features that are at risk, or potentially relevant to that type, such that more delicately specified cultural activities will ‘activate’ more restricted but not mutually exclusive sets of linguistic features. Both situation types and registers are therefore topologically rather than typologically related: individually identifiable, but sharing diverse arrays of features to different extents. The elegance of this model masks two potentially damaging underlying assumptions, however: firstly, it was originally based on a rather deterministic conception of social structure and linguistic behaviour; and, secondly, it suggests that there is ultimately a unitary language community as the point at which the entire range of behavioural options can be mapped (relatively) neatly onto the entire range of language options. The complementarity of these two assumptions is captured in the following quote (Halliday, 2007b [1971], p. 59): It is the social context that defines the limits of the options available; the behavioural alternatives are to this extent context-specific. But the total range of meanings that is embodied in and realized through the language system is determined by the context of culture—in other words by the social structure. The deterministic element evident in Halliday’s earlier writings is in line with Marxist theorising of the time, in which social and cultural systems are superstructural byproducts of the economic base. Halliday’s model therefore represents a materialistic conception of language variation as

Rethinking (Context of) Culture


tied to class formations based on relationships to the means of production. While variation according to these parameters is defined as differences in code, or coding orientation (e.g. Bernstein, 2000), which are not included in the above model, it is argued that coding orientations are a product of the contexts of situation, and hence registers, to which speakers are differentially exposed and into which they are socialised (Halliday, 2007c [1974], pp. 90–93). Conversely, access to specific registers is also a function of social class. In later developments of the model of language and culture, code was added to the cline of instantiation (Matthiessen, 2007, p. 539; see Figure 3.2), but this representation creates further problems, as will be discussed. Determinism in this crude form is a problem in that it eliminates, or at least severely delimits, the room for agency and bottom-up change to the system through discursive innovation and intervention at the individual and group level. That is not to say that the model should reflect our aspirations rather than the realities of social stricture, but that the model should be able to account for evidenced examples of change and the conditions under which such change was and is possible. Later refinements to the model address this problem through the reformulation of the realisation relationship between culture and language as a bidirectional relationship in which the context activates the language behaviour, which in turn construes the context (Hasan, 2013). In these terms, the context is not predetermined but is a function of meaningful choices in language which are simultaneously constrained by the variables of context while recalibrating them. This formulation allows for wiggle room and agency without ruling out the importance of sociopolitical structures and the limits to agency. While there have been developments to the model in terms of determinism and agency, the conception of culture as ultimately a unitary construct has remained more or less unchanged, as evidenced in Matthiessen’s (2007) version in Figure 3.2. That this enduring conceptualisation is not accidental, an effect of the limitations of two-dimensional heuristic models, is apparent in Halliday’s (2007a [1964], p. 7) rather surprising claim that, despite differences in language use between, for example, Americans and British, “it is a mistake to exaggerate this distinction, or to conclude therefrom that there is no unified Englishspeaking language community”. This raises the question of the extent to which such a single community extends: would it, for example, encompass bilingual Indians, speakers of Guyanese Creole, and second language speakers of all levels and stripes as well as educated Americans and the working class speakers of Yorkshire dialects? And how would the non-English languages that comprise an essential part of these speakers’ repertoires, and against which their use of English attains its valeur, fit into the culture of a unitary language community? If such a comment stretched the conception of a unified culture even at the time of writing,

32 Tom Bartlett it seems even less plausible in this era of accelerated transcultural flows and heightened linguistic contact. Such contact can result in a localised hybridisation of the diverse which at once points to the increased interconnection of cultures but also to the impossibility of closure, as is captured in the following quote from Prabh Deep, a rapper from Delhi in India (Singh, 2017): So I don’t bother like if people can’t understand the language. But I know they will enjoy the flow. Like you enjoy the flow and the rhythm and all the things. There is no separation between the languages. So there is one little message in it. Like (.) people think you and me are different. No we are not different. You’ve seen the rap. You’ve seen the eight bars. JASPAL: Uhum PRABH DEEP: English and Panjabi is not a separate thing. JASPAL: Ah PRABH DEEP: It’s (.) it’s a combined. It’s the SAME thing. JASPAL: Ya PRABH DEEP: You never feel the difference. PRABH DEEP:

What this quote suggests is that, for Prabh Deep and his audience, English and Panjabi, or at least some aspects of both, come together to form a single culture, the elements of which therefore gain their valeur in relation to other English-language cultures and Panjabi cultures simultaneously—it cannot be contained within either one or the other and therefore renders unclosable and non-unitary the very conception of “an English-speaking language community”. The quote from Hodge that serves as the epigraph for this chapter seems particularly relevant here, as the changes to society captured in the above conversation have brought to salience features such as originality, spontaneity and hybridity that were always there but previously passed over or taken for granted. This is, in part, a consequence of the emphasis on regularities in the SFL approach and the goal of relating language variation to social variation by user and by use in a codified and comprehensive way. Following the combined logic of variation according to use and an ultimately unitary cultural system, Halliday (2007a [1964], p. 18) makes the important claim, worth repeating in full, that registers are not marginal or special varieties of language. Between them they cover the total range of our language activity. It is only by reference to the various situations and situation types in which language is used that we can understand its functioning and its effectiveness. Language is not realised in the abstract: it is realised as the activity of people in situations, as linguistic events which are manifested in a particular dialect and register.

Rethinking (Context of) Culture


While not denying the general thrust of this point, that language is always realised in relation to context, there is a suggestion here, parallel to both the deterministic and unitary orientations discussed above, that the situation types in which language occurs are pre-known and comprise a closed set. This is an idea that is repeated in Martin and Rose (2008), where culture is mapped as “a system of genres”,2 with the corollary that all texts belong to one genre or another.3 This totalising view of genre/register derives in theoretical terms from the emphasis on seeking regularities across instances and in practical terms from the thrust in much SFL work towards increasing access to prestigious registers in educational contexts, particularly with marginalised groups. This work, as important as it is as an application of the theory, has nonetheless led to a skewing of the theory towards descriptions of “homogeneous, congruent and reified” relations between text and context (Bartlett, 2015, 2017a, p. 388): that is to say, language behaviour in which each participant adopts the role expected of them from the wider context and which, through repetition, has become the fixed way of behaving within such a situation across an entire group. What this approach does not account for, but which should be an essential element of the theoretical/descriptive analysis, is those cases where participants challenge the role expected of them or where, owing to the newness of the situation and the array of social features that comprise it, there is no preset linguistic template for whatever the activity is or may turn out, in hindsight, to have been (e.g. Bartlett, 2012). In all such situations, as with Prabh Deep’s transcultural rap, language users will bring with them semiotic histories which inform the way they act according to the situation, but in each case, following Derrida and pace Martin and Rose, they cannot not mix genres. This is a point explicitly acknowledged by Hasan (2000, p. 44) and a significant direction in her later work on the permeability of context (e.g. Hasan, 2000, 2016), which I shall return to later. The mixing of cultural codes within a single genre type, or register, exemplified by Prabh Deep, problematises the relationship between the two within the standard SFL model. As variation according to use and according to user respectively, both register and code should be accounted for in the overall model of language and culture, but, as we can see from Figure 3.1, codes were not included in earlier models and it is Matthiessen (2007, p. 539) who includes them, placing them at the less delicate end of the cline of instantiation from registers (see Figure 3.2). However, this is problematic for a couple of interconnected reasons. Firstly, Matthiessen argues that his cline of instantiation is based on the degree and type of variety displayed, and the greater the variation, the closer to the instance end of the cline. In this argument, codal differences display contextual consistency and semantic variance whereas registers display variance along both parameters and are thus further along the cline. There is a theoretical flaw with this argument, however, which has significant

34 Tom Bartlett

Figure 3.2 Matthiessen’s (2007) Cline of Variation

implications for theorising the culture-language relationship. It is a basic tenet of SFL that higher strata are construed by the stratum below them, such that context is construed by the semantics, which are construed by the lexicogrammar. Semantic differences therefore entail contextual differences (and vice versa), so that Matthiessen’s characterisation of code does not hold up. This is not just angels on pinheads: it is essential to the SFL view of language as a social semiotic, particularly given the bidirectional relation described above, that the context is defined by the ideational, interpersonal and textual meanings made within it. This leads us to question how and, more importantly, if we can represent two different types of semantic variation, by use and by user, along a single cline. Leaving aside Matthiessen’s criteria of contextual and semantic variation, let’s consider what it means to have register further along the cline of instantiation than code. As explained above, the cline does not represent a typological relationship, so such an ordering would not imply that the culture was the superordinate sum of discrete codes which were individually the superordinate sum of discrete registers. This would clearly be an untenable situation as it would entail a total lack of semantic overlap between codes, between registers within codes and between registers from different codes. As explained above, the model assumes that different contexts can share semantic features, but that they must by definition show variety, so that the cline of instantiation simply suggests a diminishing level of sharedness, not discreteness. From this perspective, placing code closer to the system end of the cline than register does not imply that codes are composed of a set of discrete and insulated registers; but it does imply that all registerial variation can be subsumed or generalised away within specific codes, just as variation between instances can be subsumed or generalised away within specific registers. Now, while this is a more complex position than the typological perspective, it still implies a certain degree of boundedness at the superordinate level: while individual

Rethinking (Context of) Culture


instances may show overlaps between them, in contrast to the typological perspective, the set of instances that comprise a register is bounded. Extending this argument, while registers may share overlapping features, Matthiessen’s model implies that the set of registers that comprise a code is similarly bounded. This would suggest, on the one hand, that a register cannot belong to two codes at once and, on the other, the even starker idea that people with different coding orientations do not interact within the same register (either through reciprocal roles or simple ‘mingling’). This may not be the intention, but it is the implication, even with the more sophisticated conception of a cline as degrees of commonality: texts share features, but each text belongs to a single register, so, by extension, while registers share features, they belong to a single higher level category, which in this representation is code. As is often the case, this implication could be an effect of the limitations of two-dimensional representation (and see Matthiessen, 2015, for a more detailed consideration); however, representations get reified and used in derivative work, including application, so it is important to overcome the limitations as far as possible. From what I have argued so far, a more accurate picture (though not an unproblematic one; see below) would be that the culture is both the sum of the codes and the sum of the registers with no differential ordering of the two within that relationship: just as codes are variation according to subsets of users putting language to different uses, so registers are variation according to subsets of uses being performed by different users. They are at the same level of abstraction, just different in kind, and as a result there are multiple overlaps between and within them. As the Prabh Deep example shows, specific registers can be considered in light of the different codes within them just as much as codes can be characterised according to the different registers which comprise them. Moving to a more abstract level, though, there remains the problem that for a single culture to be the sum of both register and code simultaneously it would have to be defined as the limit at which the overlaps between the two are ultimately bounded into a single coterminous space. As already suggested, this would seem to be an unlikely scenario, particularly in an age of superdiversity. Let me take a further example from my own work. In my study of interaction between indigenous groups and governmental and non-governmental bodies in Guyana (Bartlett, 2012), I focused on the North Rupununi District Development Board (NRDDB) as a forum bringing the two groups together. As set out in the book, these two (or more) culturally distinct groups had, at a certain degree of generalisation and abstraction, different voices and different coding orientations; yet they came together into a single situation, bringing with them their semiotic histories and resultant ‘generic expectations’, and, in the event, the register that developed over time within the NRDDB as an institution showed features from these different coding orientations as

36 Tom Bartlett well as from other registers. In other words, it shared features from registers from within the indigenous culture, which overlapped with other registers within the indigenous culture, and so on. Are these other registers, therefore, part of the same context of culture as the NRDDB register and, by extension, so also the registers of the non-governmental organisation and government participants that overlap with the NRDDB register but are not part of it? And where would the boundary of these overlaps be? In other words, at what scale of generality does the context of culture operate? At what scale can we refer to a ‘unified English-language community’ or any other? Only in the ultimate limiting case where all groups are included in the network? Such a possibility may have existed for the Trobriand Islands, but it is doubtful that the concept of bounded diversity could be extended beyond such cases. We can develop the same line of argument from a slightly different direction. Building on Derrida’s not-not-mixing and Hasan’s concept of registerial permeability, we can suggest that just as a text can belong to two registers simultaneously, so a register can belong within two codes simultaneously (as my Guyanese data suggests). From here we are in a position to extend the logic to suggest that a code can belong to two cultures simultaneously. Combining these ideas, which are really the perspectives from above and from below on the same phenomenon, we can say both that a register may be realised by elements from various codes and cultures and also that a register or code can belong to two cultures simultaneously, and at that point, the idea of an ultimately unitary (context of) culture collapses. Two further points are worth making here before exploring how this idea might reshape the model and why this is important for analysis. Firstly, the idea that a register can be realised by interacting elements from different coding orientations is an important condition for social change as it creates the conditions for the potential colonisation and appropriation of that register by marginalised groups, as suggested by my Guyanese data (cf. Bernstein, 2000, p. 15, on strength of framing and my elaboration on this in Bartlett, 2012). And this relates to the second point, that a text only belongs to two registers simultaneously, or a register to two codes simultaneously, as hybrids, until that mix is recognised as a new register or code in its own right—registers and codes are social conventions, after all, and we are all mongrels in this respect. And this is where the study of non-homogeneous, non-reified and non-congruent contexts comes in as a necessary complement to the study of the more stable and recognised forms. In an age of increasing intercultural connectivity and the hugely significant social, political and economic implications of this, we need to understand, across a whole range of discourse-informed disciplines, the social and linguistic processes of change through which stable forms become hybridised and hybrids become stabilised as new sociolinguistic configurations take hold and the marginalised move to the centre. To facilitate this we need an appropriate heuristic model, and while the

Rethinking (Context of) Culture


concept of a unitary context of culture as the ultimate reference point no longer seems viable, we can continue to develop Halliday’s emphasis on contextual contingency and articulate this with recent conceptual developments in interactional sociolinguistics and linguistic ethnography to reorientate the SFL model. Accordingly, in the following section I will take as a starting point the idea that it is not the whole space-time of a ‘language community’ that provides the cultural context to individual texts or even registers and relate this to recent uptakes of Bakhtin’s notion of chronotope (Blommaert, 2015; Cloran, 2010), the unit of space-time that is simultaneously instantiated by the text while providing the framework of relevance within which that text is to be interpreted (Singh and Bartlett, 2018). From there I will argue that the rearrangement of the SFL architecture along these lines is at once consistent with core SFL principles while developing the theoretical model in ways that will enhance future applications. As this chapter is directed at an SFL audience, the emphasis will therefore be on modifications to the SFL model. However, an important corollary of this approach is that SFL theory can also make important contributions to other interactional and linguistic anthropological work whose own models can be similarly adapted or complemented to afford the important theoretical insights SFL has to offer about how (rather than just that) languages, contexts and cultures are constructed and change.

3. Scales of Context and the Limits of Climate Change Following from the argument developed so far, we can say that each communicative event construes, or indexes (e.g. Silverstein, 1976), its own context of situation in terms of the spatio-temporal extent of its relevance: that is, the degree of shared experience it is drawing on or manipulating. Such an idea tallies with Hasan’s (e.g. 2013) concept of the context of situation as the context made relevant by the participants; where it develops Hasan’s concept is through a reimagining of the relationship between the individual context of situation and those to which it is related no longer as an instance of a superordinate culture, but seen instead as the joint making sense of an intersection of multiple semiotic histories, the commonalities between which cluster together at different scales of abstraction. In this section, therefore, I will attempt to reinflect core elements of SFL theory in light of related theoretical developments in interactional sociolinguistics, namely the concepts of superdiversity (Vertovec, 2007) and scales (Blommaert, 2015; Bartlett, 2008), and the neo-Marxist (or, better, neo-Gramscian) discourse theory of Laclau and Mouffe (1985). Taking the context of situation as our starting point, we can say that any such instance brings together a group of individuals who have been uniquely socialised through their individual experiences in interacting

38 Tom Bartlett with others and who therefore bring with them their own coding and registerial repertoires, which may be extensive or limited at the individual level and more or less shared at the group level, ideas that resonate with the alternative architecture of Gregory’s communication linguistics (de Villiers and Stainton (Eds.), 2001). This diversity should be neither overplayed nor underplayed. Given the Marxist materialist goals of SFL, although individual differences in socialisation and hence in texts were explicitly recognised, what was deemed important in the theory was to find commonalities and to produce generalised categories that could be compared and correlated. That mission remains very much alive, but can and should be complemented (especially if the theory is to be holistic) with accounts of the instance as unique (though set against the wider background) and with each participant bringing different experiences to bear on the context, to make relevant if and when possible, and so to use as resources in forming the small culture (Holliday, 2013; Singh, Kantara and Cserző (Eds.), 2016) that frames the instance. From this perspective it is both the overlaps and the differences in codal and registerial repertoires that come into focus and the degree to which the relevances made in this microcontext are relevant and legitimate at higher scales. Putting flesh onto these abstract bones, we can relate them to research I am currently undertaking into negotiations over local management of resources in the Western Isles of Scotland (e.g. Singh and Bartlett, in press). At the local level, islanders, with divergent viewpoints and experiences, come together as a group and reach a degree of consensus with regard to the policies they would like to see implemented. This discourse operates at the local scale with the island community as the centre. This means that although the various participants bring with them diverse experiences across a range of contexts as semiotic resources, they orient these to the needs of the immediate event and the normative expectations of culture at the islands’ level (potentially altering these in the process). However, in order for their plans to be heard and legitimated, they have to function at various higher scales, each with different normative centres: the Western Isles Council, the Scottish government, the UK government and the European Union. For each of these centres the coding and registerial norms that are in force are different so that, while they may all be superordinate to the local context in terms of scale, they do not correspond to the clines of instantiation of the SFL model in two important regards. Firstly, the norms in operation at these higher scales (the backdrops against which instances have valeur) are not necessarily abstractions from the multiplicity of lower-scale activities that feed into them; and, secondly, as a consequence, they do not follow the topological but ultimately hierarchical scalarity of SFL but are dispersed, or polycentric, each with their own interacting system of higher and lower level scales and centres which have to be negotiated individually and multiply. From here we can return to the question ‘at what scale does (the context of) culture operate?’ and answer

Rethinking (Context of) Culture


‘at any and every scale’: they are the chronotopic backgrounds brought into play at each point and against which each act makes sense (and here Silverstein’s indexicality coalesces with Hasan’s realisation as a bidirectional relationship between strata). A common culture is a matter of negotiation, agreement and convention, an interactional alliance between individual ‘members’ who coarticulate a contingent common ground, a temporarily ring-fenced system of complex interactions, from the stock of their diverse semiotic histories and for the purposes of activity at that scale. Each individual, of course, belongs, or can belong, to many cultures, and the choices made within each will be dependent on both their own semiotic histories within these cultures and the interactive dynamics of the (context of) situation. Such cultures can range in scale, generality and persistence as a construct, from microcultures such as the NRDDB and upwards along different parameters through the local Makushi culture and institutional Iwokrama culture and upwards again to the macro level of Guyanese, Amerindian and international development culture, all of which are intersecting but separate. This takes us back to Halliday’s oft-aired comparison of system and instance in language to the climate and the weather, with each concrete instance of weather perturbing the climate as an abstract system (Halliday, 2003, p. 27; see also Singh, 2016, pp. 26–27 for a further discussion of this metaphor). Continuing the thrust of this chapter, I would say that this analogy can be taken as presenting too totalising a view of the language system and one which is real for no one but the linguist. The concept of ‘the British climate’, for example, is no more than the arbitrary superimposition of a geographical frame over a hugely diverse range of phenomena. The climate in the Hebrides is more distinct from that of Cornwall than that climate is from that of the Channel Islands, which in turn connects with but differs from that of Northern France, and so on. So also with language cultures, though the parameters of variation are often social as opposed to geographical: determining where one language culture ends and the next begins is not a social reality but an act of imagining communities (Anderson, 1983) that can vary according to the purpose of the analyst. Every instance of use does perturb the system, but only at the local level, only for those who come into contact with that change. For these, the system of use in context may be perturbed, but for other, more distant users of language, nothing changes: the backdrop of past semiosis against which their actions take meaning as valeur is unchanged. Or, rather, they are in their own turn perturbed by instances of use at their own local level. At times these changes remain highly localised, as within a family; at others they operate at higher scales with more abstract centres such as an institution, region or social grouping. Systemic change is local and scalar, not global, and time does not always serve to spread this change throughout the system but at times to divide it, to push contexts apart.

40 Tom Bartlett

4. Concluding Remarks I have tried to show over the course of this chapter that the idea of a unitary (context of) culture is both unfeasible and unhelpful. From the theoretical standpoint it posits a level of abstraction and generalisation that underplays the complex interactions of language and society; and from the applied perspective, therefore, such an overarching context of culture cannot form the backdrop against which language use at any scale gains its valeur. Meanings are made locally, at the intersection of meaning systems, and spontaneous acts are as important as regularities, the hitherto impermissible as important as the standard. As a corollary, change happens locally and does not spread throughout the system but creates new subsystems at different scales and with different centres. This is not the picture of homeostasis presented in SFL theory, where the system keeps changing to stay the same; this is more akin to the neoMarxist/neo-Gramscian discourse theory of Laclau and Mouffe, where antagonisms are necessarily built into the system, where there is always an element of the other acting as a grain of sand within the system and preventing closure. In the Western Isles example above, while local microcultures are contingently formed and locally unique, they will always carry within them elements of hegemonic discourse from outside, while, conversely, discourse at higher scales will be continuously challenged by the individual and group histories of the participants. The result is not homeostasis but perpetual socio-cultural change. This is important for discourse analysis, not least in educational linguistics, Critical Discourse Analysis (CDA) and Positive Discourse Analysis (PDA). In each of these disciplines it is necessary to evaluate an instance in terms of what was possible, what could have been said but was not, and in terms of the differential resources open to the different participants as individuals or groups and the potential for articulating these in radically new formations. In this regard, CDA has tended to focus on a critique of existing structures as an awareness-raising first step in realising alternative discourses as forms of social action. However, such studies have often suffered from a disconnect between micro- and macroanalysis: in text-based studies, detailed linguistic analysis tends to be rather loosely related to the political structures at the macro level; while in context-based studies, individual texts are rather unconvincingly chosen as exemplars of the sociopolitical phenomenon under discussion. An approach which examines individual texts as the intersection of multiple semiotic histories and the macrocontext as the interconnecting and legitimation of discourses at multiple scales would provide a much richer and more nuanced picture of the workings of power and resistance within and across contexts. More specifically, the equal emphasis built into this approach between the reified and the new provides a framework for CDA to analyse “emergent discourses of hybrid identity” (Luke, 2002, pp. 106–107) which may challenge the status quo, and so

Rethinking (Context of) Culture


to move beyond a focus on ideology critique and to document ‘other’ forms of text and discourse—subaltern, diasporic, emancipatory, local, minority, call them what we may—that may mark the productive use of power in the face of economic and cultural globalization. (Luke, 2002, p. 98) In such analysis there is always a dialectic between text and context as the conditions of production motivate the features of the text while the instances of text realise the social context and potentially perturb the underlying system of political relations. For PDA (Martin, 2012 [2004]; Bartlett, 2012, 2017b), an offshoot of CDA, it is the form of such challenges that is the focus of description along with the conditions that enable them—what I have elsewhere labelled the perturbation potential of the alignment of situational features, both social and linguistic, which operate at the local level but are constrained within the wider system of interrelations between levels (Bartlett, 2013; and see Bartlett, 2017b, for a discussion of the different elements within and approaches towards emancipatory discourses). It is a small step from here to political science, where evolving theories of hegemony, conviviality and radical democracy all demonstrate a focus on discourse as researchers seek not simply to catalogue differences between the language behaviours of different cultural groups but to explain these with relation to the socio-cultural structures within which they emerged and which constrain them. The goals of these disciplines are, on the one hand, to remove barriers to integration and cooperation as these groups come into contact and, on the other, to explore how the discourses of different factions within progressive politics (embracing diverse interests such as class, gender, ethnicity and environmentalism) can form a united bloc around a discursively created common cause in opposition to the hegemonic discourse of the neoliberal status quo. This is the domain of Laclau and Mouffe’s (1985) Discourse Theory and their conception of counter-hegemony as a construct of semiotic alliances that are contingent and, developing classical Marxism and Gramsci’s own conception of hegemony, no longer based on class. In this view successful counter-discourses will be the result of the multiple alliances into which we enter clustering together around concepts, or social imaginaries, such as religious affiliation, ethnicity and environmentalism, each with their own codes. Such alliances are possible when the different meaningful elements within each are successfully articulated as moments in a coherent whole, establishing the social imaginary as the ideological horizon towards which the previously differentiated groups can orient as a single bloc or force. However, while Discourse Theory has been able to chart the shifts in key political signifiers and the alliances formed around them at a fairly macro level, it has not generally accounted for the interaction of micro- and macrocontexts as change permeates the

42 Tom Bartlett system. Moreover, in its development of the political role of signifiers, Discourse Theory has focused on the ideational without a consideration of the wider role of language and the role of alternative interpersonal and textual systems in fomenting change. Within all these disciplines, therefore, I would suggest that the scalar model of context outlined above, and in combination with key conceptions from SFL, would represent an essential development. I have stressed here the need to view each event as a microcultural engagement, the context of which is, on the one hand, the past interactions within that setting and, on the other, the individual semiotic histories of the participants that radiate out from that centre in different directions, providing a second level of valeur to each utterance while delimiting the potential for alternative arrangements. As argued above, neither a unitary context of culture nor the concept of registers as standardising abstractions provides all that is the necessary here. This does not imply, however, a linguistic free-for-all, either for participants or for analysts. While anything is possible, in the Chomskyan sense, not everything is either likely or legitimate (by which I mean effective at the local level). This means that, as analysts, we need to understand how language works as a system and within society, without ignoring either aspect. SFL is uniquely positioned to achieve this double goal through its understanding of language as a stratal, metafunctionally diverse and social semiotic system in which meanings across an array of scales and variables are articulated. Through its conception of the lexicogrammar, semantics and context as separate strata of meaning, with the lower strata realising the higher, the SFL architecture allows for the vertical slippage and realignment of connections between the strata as the resources of language are deployed in novel ways and different social groups come into contact. And through its conception of metafunctional diversity, with ideational, interpersonal and textual meanings operating simultaneously, the architecture allows for the realignment of existing meanings into novel social signifiers at each stratum. To supplement interactional work on scales and microcultures, therefore, the theoretical architecture of SFL can be used to account for the mechanisms by which the diverse social and linguistic resources brought to individual situations are gradually articulated both horizontally and vertically to create new and shared systems of meaning-making. Developing key ideas from Discourse Theory and adapting them to all levels of language, I would claim that analysing language use in novel situations needs to go beyond the cataloguing of the diverse meanings made as floating signifiers in unstable contexts and to account for the steady articulation of these meanings—at the grammatical, semantic and contextual strata— as interconnected moments in developing social systems with different centres and scales of reference. As Hodge (2017, p. ix) points out, the advent of poststructuralism did not signal the end of structuralism: “both tendencies are alive and well in different incarnations”, and while “structuralism needs to incorporate

Rethinking (Context of) Culture


more of what it excluded: dynamism, process, causality, uncertainty . . . poststructuralism has discovered more structures not fewer, more forms and functions at more levels.” In this sense SFL can be seen not as a structuralist approach imposing the analyst’s order on a diverse world, but as a metastructural approach that establishes the mechanisms by which contingent structures are created and recreated at diverse scales within a constant dynamic flux of social and linguistic interaction and change. There has only been room here for a brief consideration of these last ideas. As I have said above, this chapter is written primarily for an SFL audience and as such I have focused on the challenges to the theory while backgrounding a lot of common understanding. There is thus a need for a complementary paper in which the assumptions of interactional and ethnographic approaches to language are backgrounded and the metastructural theory of SFL elaborated in greater depth. It is essential to keep moving beyond our cultural silos, to interact at different scales, orienting to different and transdisciplinary centres. In these terms, to repeat my earlier point, I see this chapter not as challenging SFL perspectives on language and culture, but as developing them along the trajectories set out by Halliday and Hasan while accommodating and assimilating to more recent trends in critical theory and interactional sociolinguistics. In doing so I have emphasised the SFL concepts of permeability and hybridity (Hasan, 2016; Miller and Bayley (Eds.), 2016) over regularity and classification, while not denying the role of both perspectives. The sociopolitical world in which we find ourselves operating, as both citizens and linguists, is in a far-from-equilibrium state (Hodge, 2017, p. 227) whereby complex forms have systematic relations between elements which produce non-linear causality in which relative wholes have effects that cannot be predicted from the properties of individual parts, and individual elements may have different properties and effects as a result of being part of a complex system. As Hodge points out, such diversity was always part of the system though we were a little blind to it. In these times of superdiversity we have a role as social linguists in uncovering the means to foster conviviality and an opportunity to promote a more inclusive and radical democracy that moves beyond but does not ignore class as the principle on which alliances of political interest and identity are founded.

Notes 1. My thanks to Chris Butler, Gerard O’Grady and Jaspal Singh for comments on earlier drafts of this chapter. 2. There is well-documented terminological slippage between the terms register and genre in the writings of Halliday and Martin. 3. This corollary has been explicitly stated on the SFL discussion board Sysfling.

44 Tom Bartlett

References Anderson, B. (1983) Imagined Communities: Reflections on the Origin and Spread of Nationalism. London: Verso. Bartlett, T. (2008) Wheels within wheels or triangles within triangles: Time and context in positioning theory. In F. M. Moghaddam, R. Harré and N. Lee (Eds.), Global Conflict Resolution Through Positioning Analysis. New York: Springer, pp. 169–188. Bartlett, T. (2012) Hybrid Voices and Collaborative Change: Contextualising Positive Discourse Analysis. London and New York: Routledge. Bartlett, T. (2013) ‘I’ll manage the context’: Context, environment and the potential for institutional change. In L. Fontaine, T. Bartlett and G. O’Grady (Eds.), Systemic Functional Linguistics: Exploring Choice. Cambridge: Cambridge University Press, pp. 342–364. Bartlett, T. (2015) Multiscalar modelling of context: Some questions raised by the category of mode. In W. Bowcher and J. Y. Liang (Eds.), Essays in Honour of Ruqaiya Hasan: Society in Language, Language in Society. London and New York: Palgrave, pp. 166–183. Bartlett, T. (2017a) Context in SFL: Towards scalar supervenience? In T. Bartlett and G. O’Grady (Eds.), The Routledge Handbook of Systemic Functional Linguistics. London and New York: Routledge, pp. 375–390. Bartlett, T. (2017b) Positive discourse analysis. In J. Richardson and J. Flowerdew (Eds.), The Routledge Handbook of Critical Discourse Analysis. London and New York: Routledge. Bartlett, T. (In press) Models of discourse in SFL. In G. Thompson, W. Bowcher, L. Fontaine and D. Schönthal (Eds.), The Cambridge Handbook of Systemic Functional Linguistics. Cambridge: Cambridge University Press. Bernstein, B. (2000 [Revised Edition (1996)]) Pedagogy, Symbolic Control and Identity: Theory, Research, Critique. Lanham, Boulder, New York and Oxford: Rowman and Littlefield Publishers Inc. Blommaert, J. (2015) Chronotopes, scales and complexity in the study of language and society. Annual Review of Anthropology, 44, pp. 105–116. Butler, C. (2009) Criteria of adequacy in functional linguistics. Folia Linguistica, 43(1), pp. 1–66. Cloran, C. (2010) Rhetorical unit analysis and Bakhtin’s chronotope. Functions of Language, 17(1), pp. 29–70. de Villiers, J. and Stainton, R. J. (2001) Communication in Linguistics Volumes 1 and 2. Toronto: Editions du GREF. Fontaine, L., Bartlett, T. and O’Grady, G. (Eds.). (2013) Systemic Functional Linguistics: Exploring Choice. Cambridge: Cambridge University Press. Halliday, M. A. K. (1978) Language as Social Semiotic: The Social Interpretation of Language and Meaning. London: Hodder Arnold. Halliday, M. A. K. (2003) On the ‘architecture’ of human language. In J. J. Webster (Ed.), On Language and Linguistics: Volume 3 in the Collected Works of M.A.K. Halliday. London: Continuum, pp. 1–29. Halliday, M. A. K. (2007a [1964]) The users and uses of language. In J. J. Webster (Ed.), Language and Society: Volume 10 in the Collected Works of Michael Halliday. London: Continuum, pp. 5–37.

Rethinking (Context of) Culture


Halliday, M. A. K. (2007b [1971]) Language in a social perspective. In J. J. Webster (Ed.), Language and Society: Volume 10 in the Collected Works of Michael Halliday. London: Continuum, pp. 43–64. Halliday, M. A. K. (2007c [1974]) Language and social man. In J. J. Webster (Ed.), Language and Society: Volume 10 in the Collected Works of Michael Halliday. London: Continuum, pp. 65–130. Hasan, R. (2000) The uses of talk. In S. Sarangi and M. Coulthard (Eds.), Discourse and Social Life. Harlow: Longman. Hasan, R. (2013) Choice, system, realisation: Describing language as a meaning potential. In L. Fontaine, T. Bartlett and G. O’Grady (Eds.), Systemic Functional Linguistics: Exploring Choice. Cambridge: Cambridge University Press, pp. 269–299. Hasan, R. (2016) The nature of language: Reflections on permeability and hybridity. In D. Miller and P. Bayley (Eds.), Hybridity in Systemic Functional Linguistics: Grammar, Text and Discursive Context. Sheffield, UK and Bristol, CT: Equinox, pp. 337–382. Hodge, B. (2017) Social Semiotics for a Complex World. Cambridge: Polity Press. Holliday, A. (2013) Understanding Intercultural Communication: Negotiating a Grammar of Culture. London and New York: Routledge. Laclau, E. and Mouffe, C. (1985) Hegemony and Socialist Strategy. London and New York: Verso. Luke, A. (2002) Beyond science and ideology critique: Developments in critical discourse analysis. Annual Review of Applied Linguistics, 22, pp. 96–110. Martin, J. R. (2012 [2004]) Positive discourse analysis: Solidarity and change. In W. Zhenhua (Ed.), CDA/PDA: Volume 6 in the Collected Works of J.R. Martin. Shanghai: Shanghai Jhao Tong University Press. Martin, J. R. and Rose, D. (2008) Genre Relations: Mapping Culture. London: Equinox. Matthiessen, C. M. I. M. (2007) The ‘architecture’ of language according to systemic functional theory. In R. Hasan, C. M. I. M. Matthiessen and J. Webster (Eds.), Continuing Discourse on Language, vol. 2. London and Oakville: Equinox, pp. 505–561. Matthiessen, C. M. I. M. (2015) Register in the round: Registerial cartography. Functional Linguistics, 2(9). Miller, D. and Bayley, P. (Eds.). (2016) Hybridity in Systemic Functional Linguistics: Grammar, Text and Discursive Context. Sheffield, UK and Bristol, CT: Equinox. Silverstein, M. (1976) Shifters, linguistic categories and cultural description. In K. H. Basso and H. A. Selby (Eds.), Meaning in Anthropology. Albuquerque: University of New Mexico Press, pp. 11–55. Singh, J. N. (2016) The journey is its own reward: Downscaling culture in intercultural communication research. In J. N. Singh, A. Kantara and D. Cserző (Eds.), Downscaling Culture: Revisiting Intercultural Communication. Newcastle: Cambridge Scholars Publishing. Singh, J. N. (2017) Transcultural Voices: Narrating Hip Hop Culture in Complex Delhi. Unpublished PhD thesis, Cardiff University. Singh, J. N. and Bartlett, T. (2018) Negotiating sustainability across scales: Community organising in the Outer Hebrides. In L. P. Moita-Lopes and M. Baynham (Eds.), AILA Review 2017: Meaning-making in the Periphery.

46 Tom Bartlett Singh, J. N., Kantara, A. and Cserző, D. (Eds.). (2016) Downscaling Culture: Revisiting Intercultural Communication. Newcastle: Cambridge Scholars Publishing. Vertovec, S. (2007) Super-diversity and its implications. Ethnic and Racial Studies, 30(6), pp. 1024–1054. Webster, J. J. (Ed.). (2007) Language and Society: Volume 10 in the Collected Works of Michael Halliday. London: Continuum. Wray, A. (2014) Developing comprehensive criteria of adequacy: The challenge of hybridity. In María de los Ángeles Gómez González, Francisco José Ruiz de Mendoza Ibáñez and Francisco Gonzálvez García (Eds.), The Functional Perspective on Language and Discourse. Amsterdam and Philadelphia: John Benjamins, pp. 19–36.


Pushing the Boundaries of Appliability Conceptual Frameworks in Cognitive Science and Functional Linguistics Jamie Williams, Derek Irwin and Noah Russell

1. Introduction Uncovering the nature and workings of the mind is one of the most important outstanding research areas in modern science, and the emergence of cognitive science during the mid-20th century advanced attempts to explain this phenomenon. Arising out of the demise of behaviourist psychology in the 1950s, this new discipline was explicitly multi-disciplinary in its outlook; the far-reaching importance of the subject matter meant that there were a number of different interested parties, and the complexity of the mind could only be successfully tackled by drawing upon ideas and expertise from a number of different disciplines. Linguistics has always been a core part of this enterprise (Miller, 2003). This chapter concerns two related issues within the field of cognitive science: conceptual frameworks and levels of analysis. By a conceptual framework, we denote abstract, theoretical frameworks that provide the means of conceptualising the phenomenon under study, illustrate important concepts and provide an overall architecture to which individual studies can be placed in reference (cf. Miles and Huberman, 1994, p. 18). Examples from linguistics include the Minimalist Framework of Chomsky (1995), as well as the theoretical architecture of Systemic Functional Linguistics (SFL), upon which this chapter focuses (cf. Matthiessen, 2007). Often, a conceptual framework will contain one or more sets of levels of analysis. In describing any phenomenon, there will be a plethora of information that could potentially be relevant. Levels of analysis enable researchers to structure this information according to hierarchical principles. Although a crucial notion within cognitive science, there exists ambiguity surrounding the term’s use, due to many different types of conceptual framework being available (Craver, 2007; Opie, 2010).


Jamie Williams et al.

Within cognitive science there exists a growing trend of embedded, embodied, extended and enacted cognition. This group of ideas, dubbed the “4E approaches” to cognition by Rowlands (2010, p. 3), emphasises that a theoretical understanding about the nature of cognition can only result from giving increased attention to the external environment, action and the body. The details and differences behind these four terms are outlined in Section 2.1. These ideas have grown out of dissatisfaction with the dominant view, which studies cognition in terms of a disembodied, decontextualised computational system. However, this 4E movement lacks an explicit conceptual framework that specifies the relevant levels of analysis needed to study cognition from this viewpoint. We believe that theoretical notions of stratification, instantiation, delicacy and rank can fill this role, and that cognition can be modelled as a stratified semiotic system, in a way that is true to the motivations behind these more externalist ideas within cognitive science. By discussing how SFL has modelled language as a stratified semiotic, and using this as an exemplar case, we aim to show that the use of the same theoretical ideas is warranted in cognitive science. This chapter aims to defend this position and outline some initial ways that these theoretical notions can be employed. It should be emphasised that this work is not a completed proposal of such a framework, but instead an argument that such an approach is viable, and one which will benefit from future development. The remainder of this chapter is organised as follows: Section 2 outlines the fundamentals of the 4E approaches to cognition, and identifies the gap in the literature that this work hopes to fill. Section 3 will motivate the application of systemic functional ideas in cognitive science in a wider sense, by arguing that cognition can properly be considered to be an open, dynamic, semiotic system, as has also been argued to be true for language by SFL-orientated scholars. Section 4 then discusses the applications of the dimensions mentioned above. Section 5 concludes with a discussion of future directions.

2. The Externalist Trend in Cognitive Science In the immediate aftermath of the cognitive revolution in the 1950s, the dominant idea of cognition was based upon the ‘computational metaphor’ (Eliasmith, 2003). In this tradition it was generally agreed that the basis of cognition was computation; the mind operated as a device that performed transformations upon meaningful representations (Edelman, 2008, Ch. 1). During the 1980s, the connectionist approach emerged, which took issue with the focus on serial computations of its predecessors, and instead advocated the study of cognitive processes through Artificial Neural Networks that could simulate a range of cognitive tasks (see Garson, 2015, for a review). Although there are a number of differences

Pushing the Boundaries of Appliability


between these two conceptualisations of cognition, they both share a common core: reliance upon the notions of computation and representation. These ideas have led to the dominant position in cognitive science that all of the important information for studying cognition is ‘internal’: aspects of the environment that organisms interact with, features of their body, and considerations about the actions that they actually do were essentially sidelined. However, the end of the last century saw a growing discontent with this position. For many, the computational perspective omitted too much potentially important information and failed to properly account for the fact that organisms actively do things cognitively within a world (Wilson and Galonka, 2013; Hatfield, 2014). Attention was instead focused, not on organisms as passive receivers of information, but instead on organisms as active doers, and this perspective required ideas of cognition to widen out from a solely internal, individualistic perspective. 2.1. The 4E Approaches: An Overview Within this growing trend, there are different opinions concerning how this shift towards active cognitive agents should be conceived. Rowlands (2010, p. 3) groups this range of ideas under the heading of the “4E approaches”:1 1. Embedded Cognition (e.g. van Gelder, 1995; Beer, 2000; Eliasmith, 2013) 2. Embodied Cognition (e.g. Shapiro, 2014) 3. Extended Cognition (e.g. Clark and Chalmers, 1998; Menary, 2010) 4. Enacted Cognition (e.g. O’Regan and Noë, 2001; Noë, 2004; Hutto and Myin, 2013) Embedded approaches emphasise two aspects of cognitive life: firstly, the environment in which a certain cognitive event occurs and, secondly, that cognitive processes are dynamic. In other words, “cognitive processes and their context unfold continuously and simultaneously over time” (van Gelder and Port, 1995, p. 2). Feedback is crucial for this view: any action that is performed by an individual has some effect on the external world, and in turn the environment has a clear effect upon the cognising organism (Beer, 1995, p. 182). To capture this dynamic interaction between an individual and its environment, a commonly used descriptive


Jamie Williams et al.

tool comes from dynamic systems theory, drawing upon the pioneering work of Thelen and Smith (1996). Explanations arise by understanding how the complex system of cognitive processes and their context evolve over time through their constant interaction. Embodied approaches also operate from an embedded perspective, but they place a greater emphasis on the nature of the body. Their thesis is not simply that cognitive processes necessitate a body in order to be implemented, but that the nature of the body plays a fundamental role in shaping the nature of cognition itself. As put by Leitan and Chaffey (2014, p. 3): “[embodied cognition] is held together by the key assumption that the body acts as a constituent of the mind rather than a passive perceiver and actor serving the mind.” Psycholinguistic and neurological evidence has also been put forward to strengthen the embodied position. For instance, physical movements made by respondents have been shown to affect word processing (Glenberg and Kaschak, 2002), and perception of motions has been shown to co-activate brain regions that typically correlate with the performance of the same action (Buccino et al., 2001). Similar neurological effects have been shown in the processing of words that denote motions (Hauk and Pulvermueller, 2004). Extended cognition takes the view that external factors can be constituents of processes that we consider to be mental, and therefore blur the boundary between the ‘inside’ and the ‘outside’. Proponents of this view argue that specific objects in the environment can be as much a part of certain mental processes as internal neurobiological signals or mental representations. The initial arguments for extended cognition were presented by Clark and Chalmers (1998), who argued that in certain situations external objects should properly count as constituents in a cognitive process if they played the same role as an internal process, such as memory recall. Enacted cognition is similar to both the embedded and embodied approaches in that it places an important role on both the body and the environment; cognition cannot be understood “without reference to interactions between organisms and environments” (Hutto and Myin, 2013, p. 4). It goes one stage further by prioritising the role that action plays in cognitive processes. For instance, O’Regan and Noë (2001) and Noë (2004) have developed the theory of sensorimotor enactivism, based upon the notion of sensorimotor contingencies: many facts about our cognitive acts are based upon expectations that our acts will have similar effects in similar situations. O’Regen and Noë (op. cit.) provide the example of haptic perception. If a blind person grasps a bottle, initially the perception could be of a number of different objects in her possession. However, as she moves her hand around the object, her movements become contingent with previous sensing experiences of bottles. This confirmation of past experiences allows her to finally perceive the object

Pushing the Boundaries of Appliability


as a bottle. In this way “[perception] isn’t something that happens in us, it is something we do” (Noë, 2004, p. 227).

2.2. Conceptual Frameworks in Externalist Theories of Cognition Although there is much to admire in the above claims, there is one gap in current work that this chapter aims to address: the lack of an explicit conceptual framework for an externalist approach to cognition. This gap has been indicated by Wilson and Clark (2008, p. 55), who opine that it does not necessarily pose a problem, since the initial task should be to argue on the basis of individual cases, before becoming concerned with the structure of any overarching theory. In principle, this is a sensible view in the initial stages of developing any new idea, and the 4E approaches are still considered to be an emerging trend within cognitive science. However, the increasing number of empirical and philosophical accounts necessitates the creation of a theoretical model upon which to position and assess the existing proposals. The scene seems prepared to begin the process of fleshing out a proposal for a fully articulated conceptual framework for an externalist cognitive science. Conceptual frameworks have played an important role in previous internal conceptualisations of the mind. The most famous example is Marr’s (1982) tri-level framework. This contains three levels of description that he argued were crucial for understanding cognition from a computational viewpoint: a description of the computation that the system has to complete, a description of the set of representations and algorithms the system requires to complete the computation, and a description of the physical realisation of the computational process. This approach has proved useful for framing debates within the discipline (Fodor and Pylyshyn, 1988; Neeleman, 2013). Similarly, Smolensky (1986) argued that within the connectionist paradigm, any Artificial Neural Network can be assessed according to two levels of interpretation: either it represents the neurological means of performing a particular action, or, alternatively, it could represent the conceptual strategy an organism may employ. Additionally, others have highlighted the importance of levels of mereology within neurological systems (Churchland and Sejnowski, 1992, p. 20; Craver, 2007, p. 166), which discuss the different structural levels at which neurological activity may take place. The overall situation in cognitive science that arises out of the brief overview just provided is as follows. Many of the theoretical positions in cognitive science can be divided into those which conceptualise cognition to be predominantly something that happens within the individual, and therefore the focus is more upon internal states, and the 4E approaches, which extend the view of cognition in the ways described above. Whereas


Jamie Williams et al.

we agree with proponents of the 4E movements that the former view suffers from conceptual limitations, this group of ideas has benefited from the presence of conceptual frameworks upon which to frame debates. Such frameworks have aided interpretation of results in studies undertaken. The lack of theoretical structure to underpin the development of the externalist ideas within cognitive science, although understandable at this stage in its development, hinders the full articulation of the underlying ideas. To reiterate, it is our belief that a conceptual framework based upon theoretical ideas developed in the context of SFL can provide a potential candidate to fill the above-identified gap. In the next section, we will provide an argument to support this view, that cognition and language (in the way conceptualised by scholars in SFL) both count as open, dynamic semiotic systems. If this holds, and one accepts that the theoretical structure provided by Halliday and others counts not just as a model of language, but as a model of semiotic systems of this type in a more general sense, then the plausibility of the above position follows.

3. Cognitive Systems as Open, Dynamic, Semiotic Systems Direct links between the terms ‘semiosis’ and ‘cognition’ are not readily forthcoming (pace Thibault, 2004), but discussions by cognitive scientists concerning the nature of cognition will be shown to be readily compatible with the view defended in this section. In order to do so, it will be necessary to show that cognitive systems are (1) semiotic, (2) open and (3) dynamic. What we denote by these terms will be discussed below. This characterisation is important, because although the position discussed here is found in the works of SFL-orientated researchers (esp. Thibault, 2004; Halliday and Matthiessen, 1999), our position is first strengthened by demonstrating supporting concepts from cognitive scientists working within broadly different traditions. 3.1. Cognitive Systems Are Semiotic The notion of meaning has long played an integral part in cognitive science (Dennett, 1968). Within the traditional, computational approach to cognition, the notion of meaning is fundamental to the notion of representation, which by definition carries meaningful content (cf. Papineau, 1997, and Eliasmith, 2013, for two examples). Mental representations themselves are deemed by Grush (2002, p. 349) as the factor that sets cognitive systems “apart from merely complex and interesting, but noncognitive systems, like elms, trees, and microwave ovens”. If mental representations are meaningful, and representations are a defining hallmark

Pushing the Boundaries of Appliability


of cognitive systems, then meaning becomes integral to understanding cognition itself. The crux behind many of these ideas is that our thoughts, feelings, fears, desires, etc. (and potentially those of other organisms) are intrinsically about things, or have intentionality. Thus, when someone wakes up in the morning and wants her first cup of coffee of the day, that thought or desire is about something outside of her. This ‘aboutness’, or intentionality, is inherently a semantic notion, and as a result the cognitive system itself has meaning as an integral part. Adams (2010, p. 325) succinctly puts it as follows: “Suppose you want to build a computer that can think for itself. To do so the symbols in the machine have to be meaningful to it.” Ignoring the myriad of details, arguments and hypotheses concerning how best to capture the precise nature of meaning within cognitive systems, the main point is that if we assume (as the majority of cognitive science theories do) that meaning plays an integral part in our theoretical perspective, then cognitive systems are by definition semiotic systems, or systems of meaning. It should be noted that this does not necessarily entail that they are semantic systems; the nature of meaning in this case does not have to be of the same type as meaning in natural language. The language of thought hypothesis (Fodor, 1975) is perhaps the best-known example of an argument for human thought being not just semiotic, but semantic in nature. 3.2. Cognitive Systems Are Open Systems Stemming from the idea of systems theory, open systems are defined as those that interact with some external source in the course of their operation (cf. Wilson, 2002, p. 630). Following such a definition, all living organisms can be described as open systems: sources of food, water and other crucial sustenance are gathered not from within the organism itself, but from outside. Therefore, considering any type of cognitive ability forces one to consider the existence of some external context with which that system is interacting. To perceive an apple in most circumstances requires an external source that causes that cognitive experience. To feel angry is determined in part by current or past experiences with the surrounding world, and so on. Further, there is evidence that points to the integral importance of external interaction for the proper functioning of cognitive systems. Merabet and Pascaul-Leone (2010) review a number of studies where sensory deprivation has led to loss of perceptual function and cognitive ability, or occasionally where the neurological pathways normally associated with such functions had been taken up by other functions entirely. In extreme cases, such as the infamous sensory deprivation experiments reported by Heron (1957), sensory deprivation can lead to a range of


Jamie Williams et al.

cognitive effects, including impaired thinking, hallucinations and abnormal neural oscillations. Cognitive systems thus operate in tandem with an external environment. As described by van Gelder and Port (1995, p. 2): “Cognitive processes and their context unfold continuously and simultaneously in real time.” In this way, then, they should be properly considered to be open systems. 3.3. Cognitive Systems Are Dynamic Systems Dynamic systems are those “whose behaviour changes over time, often in response to external stimulation or forcing” (Åstrom and Murray, 2008, p. 1). The idea that the cognitive systems change their behaviour over time should not come as a surprise. Indeed, the common cognitive ability of learning is tantamount to this dynamicism at a very basic level: the ability to change our future actions based upon past experiences is crucial to the survival of any organism with cognitive abilities. Although such an appeal to the dynamic nature of cognitive systems will likely not strike many in SFL as controversial, the group of researchers within the embedded cognition movement has placed this notion at the heart of their proposals. Beer (1995, p. 182) argues that “each of these two dynamical systems [cognition and context] is continuously deforming the flow of the other”. Within certain sections of psychology, dynamic accounts have been provided for a range of cognitive tasks: from decision and reaching tasks in young children (Thelen and Smith, 1996), analysis of a walking automaton (Beer, 1995) and the catching of moving objects (Beer, 2000). All these analyses and descriptions have been carried out under the view that cognitive systems should be properly considered to be dynamic systems. We have seen in this section accounts from non-SFL scholars that amount to considering cognition as (1) semiotic, (2) open and (3) dynamic. The first two of these are relatively uncontroversial: the idea that a characterisation of meaning is of fundamental theoretical importance, and that cognitive systems interact with an external environment. The third, perhaps, does not benefit from such universal acceptance, but it is interesting to note here that some work to improve the study of cognition as computation through the use of Artificial Neural Networks has incorporated dynamic ideas (Eliasmith, 2013). This idea, it seems, is gaining traction in cognitive science. If one accepts this view of cognitive systems as dynamic, open, semiotic systems, then the application of SFL for framing cognitive science seems ultimately reasonable, for it proposes that language is a semiotic system with similar properties. Much work within SFL, specifically when it comes to the architectural basis behind the descriptions, not

Pushing the Boundaries of Appliability


only has illuminated aspects of language use, but also provides one of the most complete proposals concerning the nature of semiotic systems and how they can be studied from an eco-social perspective. This can be shown by the application of SFL principles to areas not directly within linguistics. Hallidayian principles have been used in studies on visual grammar (Royce, 1999), as well as art and architecture (O’Toole, 1990). This has only been possible because the ideas contained within this linguistic theory pertain not only to language, but to semiotic systems more generally.

4. Architectural Applications With the current state of the field outlined, and a motivation provided for why SFL could have a meaningful role to play in the conceptualisation and study of cognitive systems, the rest of this chapter will provide a discussion of how architectural issues in SFL can be applied to cognitive science. By this we mean the application of the various theoretical dimensions that are important in treating language as a social semiotic: stratification, delicacy, rank and instantiation (Halliday and Matthiessen, 2014, p. 20). The one important omission from this work is that of metafunction, which is understood as the organisational patterns by which the functions of language may be patterned within most strata. A comprehensive argument for metafunctional ideas within cognitive science is provided in Thibault (2004), and interested readers should refer to this for proposals that are compatible with those presented here. Again, in a work of this length, only initial ideas can be presented. What follows, then, should be considered to be an outline of possible directions in which the project discussed here could be taken. 4.1. Stratification The complexity of language necessitates the need to “talk about language under different headings” (Halliday and Matthiessen, 2014, p. 24). SFL takes the view that language is best understood as a “system of systems” (Bache, 2013, p. 72). One way this meta-system can be organised is by division between varying strata, such that each one comprises a different way to systematise the ways that language meanings may be created or interpreted: the system of soundings (phonology), the system of wordings (lexicogrammar), the system of meanings (semantics) and the system of context (situational and cultural). This latter stratum is crucial, and its inclusion distinguishes SFL from many other theories of language: understanding of language and meaning can only be reached by investigating how these systems each operate with respect to themselves and each other, and within an overall eco-social system of context. The


Jamie Williams et al.

fundamental idea behind stratification is that the entire semiotic system can be considered to contain an expression plane and a content plane (see Taverniers, 2011, for discussion). The expression plane of language is considered to be either phonology or graphology, depending on the medium of the text, while the content plane is internally stratified into the systems of lexicogrammar and semantics. The highest stratum is that of context, which describes the overall meaning systems that language operates in, at least insofar as that context impinges on the language itself (distinct, for example, from the Material Situational Setting as per Hasan, 1981). Knowledge of these three systems: expression, content and context, provides SFL with one explanation of language. Proponents of SFL believe that the best way to conceive of language is as a resource for meaning-making activity, or in other words as a meaning-making potential (Halliday and Matthiessen, 1999, p. 1). This overall potential is studied by investigating how the phonological, lexicogrammatical and semantic resources are used in different contexts, as evidenced through analysing how humans use language to enact certain functions in social situations. The proposal here is to consider the semiotic system of cognition in a similar way: as a system of systems for meaning-making activity. In line with the three-way distinction of expression—content—context, we highlight three systems within a stratified framework for cognitive science: the neurobiological, behavioural and environmental. The first of these, at the lowest level of description, describes the overall neural resources that an organism has at its disposal for performing cognitive activity. This will include details concerning the vast array of electrical activity that is possible within a neurological system. The behavioural stratum outlines the potential number of cognitive actions that an organism can perform. It is important to note that at this stage there is no division made between processes that are traditionally considered to be ‘internal’, for instance perceiving, emotion, etc., and those which are ‘external’ in terms of strata. There may need to be some distinction made within the system between these types of process, but there seems little evidence to divide them according to different strata: if one takes the idea of embedded cognition seriously, then internal and external behaviours are equally dependent upon interaction with the external environment. The highest, the environmental stratum, provides details about the range of possible situations that cognitive behaviour can unfold in. There will be much to say about the precise nature of this system, and it is important to emphasise its explicit place in the framework. However, we will have to leave a detailed discussion of this for another time, in order to give it the space it deserves. It should immediately be apparent that there are many points of contact between the view of strata in semiotic systems as systems of meaningmaking, and the embedded view of cognition. Beer (2000) highlights

Pushing the Boundaries of Appliability


three important systems for this view: the system of neural events, the system of bodily actions and, finally, the contextual system. These are essentially the same three systems that are proposed above. The appeal to stratification in this proposal, then, is in line with much of the work in the above tradition. The view of cognition that it gives us is not necessarily one that looks for the structure of an internal set of representations, but instead one that investigates patterned activity as instances of various potentials for the creation of meaningful behaviour. This strikes us as a more naturalistic way of studying cognitive systems; instead of attempting to uncover an underlying set of rules that drive cognition, researchers prioritise how the system actually operates and allow regularities to be uncovered in this way. Additionally, there are potential benefits from employing the semiotic relation of realisation to link the systems mentioned above. The term already carries some currency within cognitive science, as it has formed an important part of the computational theory of cognition. Researchers there have claimed that any computation could be run using a much larger set of different representations and algorithms, which themselves could be implemented in a potentially infinite number of physical systems. In short, the thesis states that any description at a particular level of description can be linked to many more descriptions at the level below (see Neeleman, 2013 for discussion). As such, this inter-level relationship is unidirectional, and was drawn upon to argue for the supremacy of representational explanations within cognitive science (Wilson and Clark, 2008, p. 56). In SFL, however, realisation has a number of different properties, which indicate possible increased benefits for any multilevelled framework of cognition. Firstly, it is explicitly bidirectional: higher strata can be considered to activate certain features at the strata below, while lower strata construe or encode features in the immediately dominating level of description (Cloran, 1992, p. 58). Linked to this idea is the trinocular perspective possible in stratified semiotic systems: phenomena (e.g. language events) can best be understood by viewing them ‘from above’, ‘from below’ and ‘from around’ (Halliday, 1996; Halliday and Matthiessen, 1999, p. 504). In terms of the cognitive realm, the importance of any feature of future behavioural potential (i.e. a particular type of action) can be understood by looking at which environmental features it construes (from above), which aspects of the neurobiological potential it activates, and how it interacts with the vast range of other possible actions. In this way, the semiotic value of any feature within the system can be fully understood, and the systems imbued with meaning. Such an approach leads to a fully holistic view of cognition as a semiotic system: the meaning of the systems a whole can only be understood by looking at the interaction of the neurobiological, behavioural and environmental potentials.


Jamie Williams et al.

4.2. Delicacy One characteristic property of SFL is the highlighting of paradigmatic relationships between members of semiotic systems over syntagmatic relations that hold between linguistic units. The nature of such systems is usually described though system networks, which represent the choices available within a system (or parts of a system). System networks play a vital role within SFL theory because they allow the vast meaning-making potential of language to be made manageable. It is through such networks that specific features of this potential can be grouped as being members of the same higher level type. As such, linguists are able to demarcate a system of types that become more specific as further distinctions are made at a particular level of description. Therefore, the dimension of delicacy is interpreted here as operating over a number of levels of specificity. The fundamental assumption behind the description of language within SFL is that all strata can be described in this way, including that of the contextual system. There are largely agreed upon systems that operate at the stratum of lexicogrammar (see Halliday and Matthiessen, 2014) and a number of proposals for the level of semantics (Halliday and Matthiessen, 1999; Martin and Rose, 2007). The case is less clear once one turns to the stratum of context, though these have been explored and preliminary proposals for such system networks exist; recent proposals can be found in Bowcher (2014) and Hasan (2014). Turning to the notion of a stratified view of the cognitive system, the notion of delicacy has the potential to play a major role in bringing some structure to the notions of each of the potentials mentioned. In terms of the neurophysiological potential, it was proposed above that the best way to think of this resource for meaningful action was as a space of possible states in which the system can be. By looking at how this set of potential states is manifested in behaving events, how it acts as the realiser of behavioural and ultimately environmental features, it is possible to enquire about the operation of one aspect of the cognitive system. However, this set of possible states is initially unwieldy. Current estimates claim that there are around 86 billion neurons in the human brain, and if one considers the most delicate space of possible activation states by treating all of those activations independently, then the vast size of this potential becomes apparent. The same considerations apply to the higher behavioural stratum. In the case of humans, this would at least include all of the possible states that are currently described within SFL theory, for language is just one type of possible semiotic behaviour. Add to that all possible instances of memory, learning, perception, emotion, outward actions, etc., and once again the set of possible states is overwhelming. The dimension of delicacy has the potential to make these vast potentials more manageable for the purposes of theorising and investigation.

Pushing the Boundaries of Appliability


In terms of neurobiological potential, there are other types of activity than that of the individual neuron which can be considered. One such is the role of neural oscillations in meaningful behaviour. Oscillations are rhythmic or repeated brain activity, typically described mathematically as a wave. Oscillations are relevant to our discussion here because they permit us to treat a large amount of neural activity as a single event; we can generalise over many individual neurons firing as a unified whole if those neurons exhibit the firing patterns of the relevant type. In this way, we can divide up our neurobiological potential in a less delicate way, and there may be good reason to do so. The main means of differentiating types of oscillation is according to their frequency; the spectrum of possible frequencies is divided into bands. Examples include delta waves (1–4 Hz), theta waves (4–10 Hz) and beta waves (10–30 Hz) (cf. Buzsáki and Draguhn, 2004, p. 1926). Secondly, one can distinguish between oscillations of different amplitudes: high amplitude and low amplitude waves. Finally, the particular region of the brain from which these oscillations eminate seems to be important, in line with the ever-growing evidence that different areas of the brain are involved in different kinds of processing. As an example, theta oscillations have been seen to correlate with the formation and recall of spatial memories in both rats and humans (Skaggs, McNaughton, Wilson and Barnes, 1996; Kahana, Sekuler, Caplan, Kirschen and Madsen, 1999). In the case of humans, it was shown that the amplitude of the signal was stronger during the test trials, when presumably the memories were forming, and at a lower frequency during the experimental trial, when these memories were being recalled. Clearly, more experimental data on a wider range of cognitive functions is needed before we can begin to turn such observations into a fully fledged system. However, it does hint at one way to subdivide the overall neurobiological potential in terms of different types of oscillation activity at a higher neurological structural level. 4.3. Rank Language, evidently, is structured. As such, all linguistic theories must provide some means of accounting for the various mereological relationships that exist between linguistic units. Within the “Sydney School” of SFL, the dimension of rank provides the nature of the units that are held to be relevant for language at the different strata.2 Two rank scales are more or less uncontroversial, one operating at the phonological stratum, and the other at the stratum of lexicogrammar. These are given below (Halliday and Matthiessen, 2014, pp. 21–22): Phonological Rank Scale phoneme → syllable → foot → tonal group


Jamie Williams et al. Lexicogrammatical Rank Scale morpheme → word → phrase/group → clause

With regard to the higher stratum of semantics, there seems less consensus about the possibility of a unified rank scale, and indeed there have been some arguments against one existing at this level of description (Martin, 1992; Halliday and Matthiessen, 2014, p. 662). Regardless, it is clear that the dimension of rank operates over a series of levels of mereology. The fundamental principle is that of parts and whole, whereby units at the immediately lower level join together to form a unit at a higher rank, assuming more than one of them is necessary. Within neurobiological systems, it is clear that the brain is also formed of structured units, and that certain cognitive processes may be linked to signal processing at one of these levels. Below is a provisional rank scale for the neurobiological stratum, operating over the same principle as the rank scales for the linguistic strata above (cf. Churchland and Sejnowski, 1992, p. 20; Craver, 2007, p. 166): Neurobiological Rank Scale neuron → neural ensemble → brain region Neurons are the fundamental unit of neuronal processing. They are the individual brain cells whose purpose is to propagate electrical signals throughout a system. Neural ensembles are groups of neurons that have formed into an individual processing unit. Hypothesised by Hebb (1949), they are created according to one of Hebb’s postulates: neurons that consistently activate in sequence will have their connection strengthened. Moving up the scale, it is well known that the brain consists of identifiable regions. These can be defined in terms of similarity with regards to cytoarchitecture (e.g. Brodmann areas), or in terms of the gyri and sulci in the brain. Much work in current cognitive neuroscience that utilises neuroimaging technology is engaged in isolating particular psychological functions to a particular brain region (cf. Asp, 2013, for a proposal linking existing cognitive neuroscientific studies to ideas within SFL). The purpose of such a scale in this proposal is essentially the same as within SFL: firstly, it enables us to appreciate that such structural elements do exist, and, secondly, it enables researchers to postulate at which structural level meaningful activity takes place. Each system network starts with an entry condition that specifies the structural unit over which the stated oppositions are taken to operate. For instance, the lexicogrammatical category of negation does not apply to individual lexical items, but to clauses as a whole (Hasan, 2010). Within a neurobiological context, this would prove to be a useful tool. In describing the overall neural resources for meaning activity, it is important to appreciate that

Pushing the Boundaries of Appliability


behavioural activity may be realised by activation at different neuronal structures: it may be associated with the firings of individual neurons, neural ensembles or collective activity at the scale of the brain regions. The rank scale allows this added detail to be included within the overall framework. In terms of the proposed higher behavioural stratum, there is no obvious rank scale to advance at this stage. In some of the work that has been undertaken on the structure of human action, one structural unit seems to be that of the act (Pastra and Aloimonos, 2012). However, one difficulty arises from the fact that any physical act can potentially be subdivided into any number of sub-movements. Consider even a simple motion such as raising one’s arm: this could be subdivided into an enormous number of parts, such as each muscle movement, each slight deviation from the original position, etc. Importantly, each of these actions may have a neurobiological correlate, especially when it comes to muscle movements. The key would be to investigate not all possible sub-acts, but instead those which carry some meaning within the cognitive system. This approach would then have consequences on how the realising system is described. However, we leave this issue as an open question at this stage.

4.4. Instantiation The relationship between the system of language and the manifestation of this system through linguistic texts and language use is a crucial one in linguistic theorising. In contrast to other approaches in linguistics, SFL maintains that the two are intrinsically linked. The means of providing this link is provided by the cline of instantiation (see Halliday, 1992; Matthiessen, 2007, pp. 515–516, for discussion). The fundamental idea behind the cline of instantiation is that the system of language and an instance of language are two aspects of the same phenomenon. Any difference comes from a matter of perspective. The oft-used analogy is of that between climate and weather: they are both the same phenomenon but viewed at different time scales. In a sense, “climate is the system ‘behind’ the weather” (Halliday and Matthiessen, 1999, p. 328). Considered in this way, the following properties of the link between text and system arise: firstly, properties of the system will change much more slowly than those of the text, and the system is correctly viewed in one sense as a generalisation over texts. This second property highlights an important aspect for our purposes here: that there is constant interplay between the system and the instance. The nature of the system drives the properties of a particular text in a particular situational context, but these same statistical properties are formed by the output of the texts themselves.


Jamie Williams et al.

Additionally, the dimension of instantiation allows various midpoints between the extreme ends of text and system to be considered. These midpoints can be considered in one of two ways. On one view, they are text types: a number of texts produced in a similar situation are compared, and conclusions are drawn concerning the linguistic features that are likely or unlikely to appear. The second interpretation considers midpoints on the cline of instantiation as subsystems; they are variations of the overall system of language where the probability of certain features occurring has changed. The status of these theoretical constructs as a subsystem allows midpoints to act as a guide for the creation of new texts in similar situations. In this manner, the cline of instantiation is conceptualised here as operating over a series of levels of regularity. In considering the application of a dimension similar to instantiation in cognitive science, the motivation comes from psychological and neurological work that highlights the importance of past action driving future action. At the neurobiological stratum, this evidence comes from studies that have investigated the notion of neuroplasticity. At the behavioural stratum, the important role of learning and habitual action highlights the importance of this role. Neuroplasticity is the ability of neural systems to adapt their firing patterns based upon previous activity; certain activity is more or less likely to occur according to the output of the system (Malenka, 2002). Different types of neuroplasticity are thought to exist: Long Term Potentiation (LTP) occurs when repeated stimulation of a certain neuronal pathway makes this pathway more likely to activate in the future. Long Term Depression (LTD), on the other hand, has the opposite effect: continued activation results in the pathway being less likely to fire. There are two crucial aspects to note in this case. Firstly, the changes to the overall system in terms of future events being less or more likely to occur is activity driven. The output of the neurobiological system feeds back into the behaviour of the overall system. Additionally, these changes will then drive future events in similar situations: on presentation of the same stimuli certain outputs of the neurophysiological potential are more likely to fire in the future. This is in line with the way in which instantiation is conceived in SFL. Secondly, it is important for this proposal that these changes have some meaningful effect on the system as a whole. Neuroplasticity is hypothesised to be important for a number of cognitive processes. The most frequently cited of these are in the cases of memory formation (cf. Neves, Cooke and Bliss, 2008, for a review) and learning (cf. Caroni, Donato and Muller, 2012, for a review). At the higher stratum of the behavioural potential, similar patterns can be found. After all, organisms do not act in a random fashion; their behaviour in a certain situation is driven, although not entirely

Pushing the Boundaries of Appliability


dictated, by past behaviour in past situations. Regularity in human behaviour has been studied in a number of studies that have investigated habitual behaviour, and this has been considered to be contextually and purpose driven (Danner, Aarts and de Vries, 2008, p. 245). The authors proposed two main factors in the formation of habitual behaviour: (1) context stability, the presence of the same situation in the future, and (2) frequency of past behaviour. In line with the discussion above, these both have correlates with the fundamental idea behind instantiation. Additionally, the existence of midpoints can provide a useful modelling tool for cognitive science. By considering these points, it becomes possible to group similar events in terms of neurobiological or behavioural types: the nature of similar patterns of activity can be formed on the comparison of a number of events occurring in similar situations at these two levels of description. Also, on the alternative view, these midpoints can be considered to be subpotentials for their respective systems: they can provide the means of looking at how the probabilities in terms of system output for these systems differ in these situations from that of the system of a whole, in a similar way to what Halliday and Matthiessen (1999, §8) postulate for weather reports and cooking recipes in SFL. Having a cline also allows these comparisons to be ranked with respect to one another: instance types that are based upon more general properties can be placed closer to the system end of the cline than those that are narrower in their focus.

5. Looking Forward Although, in a chapter of this length and scope, only the bare beginnings of a new proposal can be outlined, the application of linguistic notions in cognitive science produces the following picture. Stratification allows the insight from embedded cognition to be maintained: Cognition is a system of systems consisting of a neurobiological, behavioural and environmental potential. Seeing these systems as part of the overall meaning-making potential of cognition, discoveries are made and organised by seeing how these systems are utilised in actual cases. Delicacy allows this vast potential to be divided into more or less specific types. Rank allows the various neurobiological and possible behavioural structures to be accounted for, and instantiation allows for the constant feedback between action and the underlying system. This highlights the role that action at various levels of description has in terms of understanding the whole system. We would hope that as more attention is given to give flesh to this initial skeleton, the overall benefit that cognitive science may gain from these ideas will be crystallised. This will be of benefit not only in that field, but in SFL itself, as it would help make further links between a social semiotic theory of language, and the operation of neurological and behavioural systems.


Jamie Williams et al.

Notes 1. It is important to note, however, that the labels given here are not always used consistently. For instance, the term embodied cognition is sometimes used to describe all four of the below ideas. 2. It should be noted that how to describe linguistic structures is not universally agreed upon within SFL. For example, in the Cardiff grammar, the notion of rank is presented differently as it is here. In this chapter, we are following SFL as presented particularly by Halliday, Hasan, Matthiessen and those of the socalled “Sydney School”. For our purposes, we need only accept the principle of rank, and will leave further complications for later discussion.

References Adams, F. (2010) Why we still need a mark of the cognitive. Cognitive Systems Research, 11(4), pp. 324–331. Asp, E. (2013) The twin paradoxes of unconscious choice and unintentional agents: What neurosciences say about choice and agency in action and language. In L. Fontaine, T. Bartlett and G. O’Grady (Eds.), Systemic Functional Linguistics: Exploring Choice. Cambridge: Cambridge University Press, pp. 161–178. Åstrom, K. J. and Murray, R. M. (2008) Feedback Systems: An Introduction for Scientists and Engineers. Princeton, NJ: Princeton University Press. Bache, C. (2013) Grammatical choice and communicative motivation: A radical systemic approach. In L. Fontaine, T. Bartlett and G. O’Grady (Eds.), Systemic Functional Linguistics: Exploring Choice. Cambridge: Cambridge University Press, pp. 72–94. Beer, R. D. (1995) A dynamical systems perspective on agent-environment interaction. Artificial Intelligence, 72, pp. 173–215. Beer, R. D. (2000) Dynamical approaches to cognitive science. TRENDS in Cognitive Science, 4(3), pp. 91–99. Bowcher, W. (2014) Issues in developing unified systems for contextual field and mode. Functions of Language, 12(2), pp. 176–209. Buccino, G., Binofski, F., Fink, G., Fadiga, L., Fogassi, L., Gallese, V., Seitz, R., Zilles, K., Rizzolati, G. and Freund, H. (2001) Action observation activates premotor and parietal areas in a somatic manner: An fMRI study. European Journal of Neuroscience, 13, pp. 400–404. Buzsáki, G. and Draguhn, A. (2004) Neuronal oscillations in cortical networks. Science, 304(5679), pp. 1926–1929. Caroni, P., Donato, F. and Muller, D. (2012) Structural plasticity upon learning regulation and functions. Nature Reviews Neuroscience, 13, pp. 478–490. Chomsky, N. (1995) The Minimalist Program. Cambridge, MA: MIT Press. Churchland, P. and Sejnowski, T. (1992) The Computational Brain. Cambridge, MA: MIT Press. Clark, A. and Chalmers, D. J. (1998) The extended mind. Analysis, 58(1), pp. 7–19. Cloran, C. (1992) Rhetorical Units and Decontextualization: An Enquiry Into Some Relations of Context, Meaning, and Grammar. Nottingham: School of English, University of Nottingham. Craver, C. (2007) Explaining the Brain: Mechanisms and the Mosaic Unity of Neuroscience. Oxford: Oxford University Press.

Pushing the Boundaries of Appliability


Danner, U., Aarts, H. and de Vries, N. (2008) Habit vs intention in the prediction of future behaviour: The role of frequency, context stability and mental accessibility of past behaviour. Journal of Cognitive Neuroscience, 15(2), pp. 294–313. Dennett, D. (1968) Content and Consciousness. London: Routledge. Edelman, S. (2008) Computing the Mind: How the Mind Really Works. Oxford: Oxford University Press. Eliasmith, C. (2003) Moving beyond metaphors: Understanding the mind for what it is. Journal of Philosophy, 100(10), pp. 493–520. Eliasmith, C. (2013) How to Build a Brain: A Neural Architecture for Biological Cognition. Oxford: Oxford University Press. Fodor, J. (1975) The Language of Thought. Cambridge, MA: Harvard University Press. Fodor, J. and Pylyshyn, Z. (1988) Connectionism and cognitive architecture: A critical analysis. Cognition, 28(1–2), pp. 3–71. Garson, J. (2015) Connectionism. In E. N. Zalta (Ed.), The Stanford Encyclopaedia of Philosophy. Available at: entries/connectionism (Accessed on 27 May 2016). Glenberg, A. and Kaschak, M. (2002) Grounding language in action. Psychonomic Bulletin and Review, 9(3), pp. 131–144. Grush, R. (2002) The architecture of representation. In W. Bechtel, P. Mandik, J. Mundale and R. S. Stufflebeam (Eds.), Philosophy and the Neurosciences. Malden, MA and Oxford: Blackwell, pp. 349–368. Halliday, M. A. K. (1992) Language as system and language as instance: The corpus as a theoretical concept. In J. Svartvik (Ed.), Directions in Corpus Linguistics: Proceedings of Nobel Symposium 82. Berlin: Mouton de Gruyter. Halliday, M. A. K. (1996) On grammar and grammatics. In R. Hasan, C. Cloran and D. Butt (Eds.), Functional Descriptions: Theory Into Practice. Amsterdam: John Benjamins, pp. 1–38. Halliday, M. A. K. and Matthiessen, C. M. I. M. (1999) Construing Experience Through Meaning: A Language-based Approach to Cognition. London: Continuum. Halliday, M. A. K. and Matthiessen, C. M. I. M. (2014) Halliday’s Introduction to Functional Grammar, 4th ed. Abingdon: Routledge. Hasan, R. (2010) The meaning of ‘not’ is not in ‘not’. In A. Mahboob and N. Knight (Eds.), Applicable Linguistics. London: Continuum, pp. 267–309. Hasan, R. (2014) Towards a paradigmatic description go context: Systems, metafunctions, and semantics. Functional Linguistics, 1(9). doi:10.1186/s40554014-0009-y (Accessed on 12 May 2015). Hatfield, G. (2014) Cognition. In L. Shapiro (Ed.), The Rutledge Companion to Embodied Cognition. London: Routledge, pp. 361–373. Hauk, O. I. J. and Pulvermueller, F. (2004) Somatotopic representation of action words in human motor and premotor cortex. Neutron, 41(2), pp. 301–307. Hebb, D. (1949) The Organisation of Behaviour. New York: Wiley and Sons. Heron, W. (1957) The pathology of boredom. Scientific American, 196, pp. 52–56. Hutto, D. D. and Myin, E. (2013) Radicalizing Exactivism: Basic Minds Without Content. Cambridge, MA: MIT Press. Kahana, M., Sekuler, R., Caplan, J., Kirschen, M. and Madsen, J. (1999) Human theta oscillations exhibit during task dependence during virtual water maze navigation. Nature, 399(24), pp. 781–784.


Jamie Williams et al.

Leitan, N. and Chaffey, L. (2014) Embodied cognition and its applications: A brief review. Sensoria: A Journal of Mind, Brain & Culture. Available at: http:// (Accessed on 22 April 2015). Malenka, R. (2002) Synaptic plasticity. In K. Davis, D. Charney, J. Coyle and C. Nemeroff (Eds.), Neuropsychopharmacology: 5th Generation of Progress. Philadelphia: Lippincott, Williams and Wilkins, pp. 147–157. Marr, D. (1982) Vision: A Computational Investigation Into the Human Representation and Processing of Visual Information. New York: W.H. Freeman and Company. Martin, J. R. (1992) English Text: System and Structure. Amsterdam: John Benjamins. Martin, J. R. and Rose, D. (2007) Working With Discourse: Meaning Beyond the Clause. London: Continuum. Matthiessen, C. M. I. M. (2007) The ‘architecture’ of language according to systemic functional theory: Developments since the 1970s. In R. Hasan, C. M. I. M. Matthiessen and J. Webster (Eds.), Continuing Discourse on Language: A Functional Perspective, vol. 2. London: Equinox, pp. 505–561. Menary, R. (2010) Introduction: The extended mind in focus. In R. Menary (Ed.), The Extended Mind. Cambridge, MA: MIT Press, pp. 1–26. Merabet, L. B. and Pascaul-Leone, A. (2010) Neural reorganization following sensory loss: The opportunity of change. Nature Reviews Neuroscience, 11(1), pp. 44–52. Miles, M. B. and Huberman, A. M. (1994) Qualitative Data Analysis: An Expanded Sourcebook, 2nd ed. Beverly Hills, CA: Sage Publishing. Miller, G. A. (2003) The cognitive revolution: A historical perspective. TRENDS in Cognitive Science, 7(3), pp. 141–144. Neeleman, A. (2013) Comments on Pullum. Mind and Language, 28(4), pp. 533–541. Neves, G., Cooke, S. and Bliss, T. (2008) Synaptic plasticity, memory and the hippocampus: A network approach to causality. Nature Reviews Neuroscience, 9, pp. 65–75. Noë, A. (2004) Action in Perception. Cambridge, MA: MIT Press. Opie, J. (2010) Levels and explanations. ASCS09: Proceedings of the 9th Conference of the Australian Society for Cognitive Science, pp. 270–276. O’Regan, J. K. and Noë, A. (2001) A sensorimotor account of vision and visual consciousness. Behavioural and Brain Sciences, 24, pp. 939–1031. O’Toole, M. (1990) A systemic-functional semiotics of art. Semiotic, 82(3/4), pp. 185–209. Papineau, D. (1997) Teleosemantics and indeterminacy. Australian Journal of Philosophy, 76, pp. 1–14. Pastra, K. and Aloimonos, Y. (2012) A minimalist grammar of action. Philosophical Transactions of the Royal Society of London, B: Biological Sciences, 367(1585), pp. 103–117. Rowlands, M. (2010) The New Science of the Mind: From Extended Mind to Embodied Phenomenology. Cambridge, MA: MIT Press. Royce, T. (1999) Visual-Verbal Intersemiotic Complementarity in the Economist Magazine. PhD thesis, University of Reading. Shapiro, L. A. (Ed.). (2014) The Routledge Handbook of Embodied Cognition. London: Routledge.

Pushing the Boundaries of Appliability


Skaggs, W., McNaughton, B., Wilson, M. and Barnes, C. (1996) Theta phase progression in hoippocampal neuronal populations and the compression of temporal sequences. Hippocampus, 6, pp. 149–172. Smolensky, P. (1986) Neural and conceptual interpretation of PDP models. In D. Rumelhart, J. McCelland and The PDP Research Group (Eds.), Parallel Distributed Processing: Explorations in the Microstructure of Cognition, vol. 2. Cambridge, MA: MIT Press, pp. 390–431. Taverniers, M. (2011) The syntax-semantics interface in systemic functional grammar: Halliday’s interpretation of the Hjelmslevian model of stratification. Journal of Pragmatics, 43, pp. 1100–1126. Thelen, E. and Smith, L. (1996) A Dynamic Systems Approach to the Development of Cognition and Action. Cambridge, MA: MIT Press. Thibault, P. J. (2004) Brain, Mind, and the Signifying Body. London: Continuum. van Gelder, T. (1995) What might cognition be, if not computation? Journal of Philosophy, 92(7), pp. 345–381. van Gelder, T. and Port, R. F. (1995) It’s about time: An overview of the dynamical approach to cognition. In T. van Gelder and R. F. Port (Eds.), Mind as Motion: Explorations in the Dynamics of Cognition. Cambridge, MA: MIT Press, pp. 1–43. Wilson, A. and Galonka, S. (2013) Embodied cognition is not what you think it is. Frontiers in Psychology, 4. Available at: fpsyg.2013.00058 (Accessed on 29 June 2016). Wilson, M. (2002) Six views of embodied cognition. Psychonomic Bulletin and Review, 9(4), pp. 625–636. Wilson, R. and Clark, A. (2008) How to situate cognition: Letting nature take its course. In P. Robbins and M. Aydede (Eds.), The Cambridge Handbook of Situated Cognition. Cambridge: Cambridge University Press, pp. 55–77.


The Interface of SFL and Pragmatics Reporting Strategies in Poe’s “William Wilson” Mounir Triki and Nesrine Triki

0. Introduction The aim of this chapter is to explore the complexity of reporting strategies in fiction drawing on the tools offered by Systemic Functional Linguistics (SFL), stylistics and pragmatics. Reporting in the context of this chapter takes a broader meaning than the one found in SFL, where it is considered as a type of projection. Outside SFL, reporting encompasses the various structural and semantic representations of direct and indirect speech and thought. We will first discuss how ideational and interpersonal projection can provide substantial material for the interpretation of a Gothic short story written by Poe and how such a functional grammatical analysis could constitute a solid base for the interpretation of the story from a pragmatic angle. The chapter emphasizes the value of blending linguistic and pragmatic frameworks in the analysis of literary texts. Because literature primarily lives within and through language, linguistic tools are thought to be mandatory in the understanding of literature (Hasan, 1989; Fowler, 1986). Hasan (1989, p. 104) maintains that “without linguistics, the study of literature must remain a series of personal preferences, no matter how much the posture of objectivity is adopted: being objective implies knowing the nature of that which one is being objective about”. Yet Hasan herself acknowledges that although linguistics is a fundamental step in approaching literary texts, it is not a “sufficient one” (ibid.). Because of their specificity, literary texts are thought to be problematic in their analysis as they often deviate from conventional writing styles used in non-literary genres. This is perhaps among the reasons for which literary critics often criticize linguists’ analysis of literary texts (Fowler, 1986). In particular, the use of reporting techniques in fiction seems to be more complex than in other writing genres. This chapter, therefore, seeks to show how it could be more enriching to blend SFL and pragmatic approaches to the analysis and interpretation of reporting. The underlying assumption is that the phenomenon of reporting is so complex that linguistic, stylistic and pragmatic frameworks need to be combined in order to optimize our understanding of reporting.

The Interface of SFL and Pragmatics


The chapter reviews the literature on reporting in those approaches and shows how it is possible to follow a two-step procedure in the analysis of “William Wilson”. The first step is the analysis of the ideational and interpersonal realizations of projection. The second is the interpretation of the findings using pragmatic tools.

1. The Grammar and Semantics of Reporting Because the approaches use different labels to refer to the different forms of speech and thought presentations, we will use ‘projection’ for the SFL sections, ‘speech and thought presentations’ for the stylistics sections and ‘reporting’ for the pragmatic ones. 1.1. Reporting in SFL 1.1.1. Clause as Representation In order to grasp the meaning of projection in SFL, it would be helpful to introduce it within a wider view of the theory, in particular, the three metafunctions of language, the different grammatical ranks and the logico-semantic systems. According to Halliday and Matthiessen (2014, p. 10), “[t]he clause is the central processing unit in the lexico grammar— in the specific sense that it is in the clause that meanings of different kinds are mapped into an integrated grammatical structure”. The different meanings that a clause can express are realized in the three metafunctions of language, namely the ideational (enacting of experiences), the interpersonal (speaker/hearer or writer/reader interaction) and the textual (information structure across clauses). At the ideational metafunction, clauses are analysed as representations of a certain experience enacted within different participants (Agent, Goal, Senser, Phenomenon, etc.), the types of processes these participants are involved in (material, relational, verbal, mental, behavioural and existential) and the various circumstances within which participants and processes interact (e.g. time, reason, angle etc.). When clauses are combined to form clause complexes, the logicosemantic meanings that could emerge are either Expansion (Elaboration, Extension and Enhancement) or Projection (Locution, Idea). Combined clauses could be of equal status (paratactic) or of unequal status (hypotactic), and sometimes clauses are rank-shifted to be participants in other clauses or embedded within nominal groups, in which case they are considered as embedded clauses. Projection is defined as “the logical-semantic relationship whereby a clause comes to function not as a direct representation of (nonlinguistic) experience but as a representation of a (linguistic) representation” (Halliday and Matthiessen, 2014, p. 508). Although such a definition seems to restrict the meaning of projection to the clause, it does so only


Mounir Triki and Nesrine Triki

because the clause in SFL is considered as the congruent realization of any meaning potential. In fact, as Halliday and Matthiessen point to, “a sequence of projection can thus be realized not only by the manifestation of projection in the clause nexus, but also by its manifestation in the clause or the group/phrase” (ibid., p. 720). Projection is, therefore, a “trans-phenomenal” (Halliday and Matthiessen, 1999, p. 223) category better thought of as a fractal motif (Halliday and Matthiessen, 1999; Arús, 2008) that could be realized at different clause and group ranks. Projection is analysed based on three systems: the level of projection, the mode of projection and the speech function of projection. The level of projection distinguishes between clauses projected in the form of meaning (idea), typically introduced via mental processes, and those projected in the form of wording (locution), delivered via verbal processes. Mental process clauses are classified into four types (Halliday and Matthiessen, 2014): emotion (hate, fear, love), cognition (think, remember, consider), desideration (want, wish, hope) and perception (feel, perceive, notice). These types of clauses involve two participants: a Senser, the one involved in a cognitive, perceptive or emotional experience, and a Phenomenon, the thing that is felt, thought or perceived. Verbal process clauses, on the other hand, are “clauses of saying, as in What did you say?—I said it’s noisy in here” (Halliday and Matthiessen, 2014, p. 302). The participants in a verbal process clause are the ‘Sayer’, the one issuing the saying activity; the ‘Receiver’, “the one to whom the saying is directed” (ibid., p. 306); and the ‘Verbiage’, the content of the saying. Circumstantially, projection could be signalled via a circumstance of ‘Matter’ marking the Verbiage as in about Paris in I don’t want to talk about Paris, or a circumstance of ‘Angle’ marking the Sayer as in according to the Prime Minister in According to the Prime Minister, taxes on revenues will be raised by 5%. 1.1.2. Mode of Projection The mode of projection sets the interdependency relations between clauses in clause complexes, in terms of parataxis (quotes) and hypotaxis (reports) “representing two degrees of remove from the original source” (Halliday and Matthiessen, 2014, p. 531) as well as metaphorically in rank-shifted clauses carrying the projected proposition functioning as Qualifier to the projecting head noun. Projection could also be used within clause simplexes in the form of comment adjuncts as in presumably in The suspect has presumably hidden the crime weapon in the baby diapers, or as participants as in the rumor in The rumor spread so quickly, and at group levels as a post-deictic as in alleged in The suspect refused to admit the alleged crimes. As for the speech functions of projection, these

The Interface of SFL and Pragmatics


could be either propositions (e.g. statements, claims, enquiries, etc. . . .) or proposals (e.g. offers, promises, demands, etc.). Another possible mode of projection, somehow controversial in its description in the literature, is the Free indirect style, which for Halliday and Matthiessen (2014, p. 532) is “not so much intermediate as a blend” of quoting and reporting features, but for Thompson is more like an intermediate category: “although I have implied that reports and quotes are identifiably separate types of projection, it would be truer to say that they are extremes on a cline, and that there are many intermediate cases” (2014, p. 205). Regardless of whether they are considered as blends or as intermediate categories, in SFL the ‘Free’ modes seem to be restricted to the ‘indirect’ categories and to the ‘speech’ representation even though Halliday and Matthiessen recognise that “free indirect speech can be projected both verbally and mentally” (2014, p. 532). More discussion about Free categories of projection will be dealt with from the stylistic perspective in section 1.2. 1.2. Reporting in Stylistics Stylistics, defined by Wynne (2005, p. 1) as “a field of empirical inquiry, in which the insights and techniques of linguistic theory are used to analyse literary texts”, offers detailed categorization of speech and thought presentations together with their semantic and pragmatic implications. Semino and Short (2004) refer to forms of reporting as Speech, Writing and Thought Presentations (SW&TP). Their model, which is a revised version of Leech and Short (1981), distinguishes between the three forms that an act of reporting could originally have. The three discourse presentations are presented along scales (see Table 5.1) that signal a move from the most involved form of reporting (left) to the most detached one (right). In addition to the traditional forms of direct (DS and DT) and indirect (IS and IT) speech and thought, Semino and Short’s model makes a clear distinction between Free Indirect Speech (FIS) and Free Indirect Thought (FIT) (1a1 and 1b), on the one hand, and between Free Direct Speech (FDS) and Free Direct Thought (FDT) (2a and 2b), on the other hand, thus allowing for a clear cut between the lexico-grammars of quotes and reports and between the nature of the represented material as either thought or speech (Semino and Short, 2004; Leech and Short, 2007). Table 5.1 Speech, writing and thought presentation scales (Semino and Short, 2004, p. 49) [N] [N] [N]








Mounir Triki and Nesrine Triki

(1) a.

He looked straight at her. He would definitely come back tomorrow! She was pleased. b. He looked straight at her. He would definitely come back tomorrow! She remained unaware of his plan until the following day. (2) a. He looked straight at her. ‘I’ll definitely come back tomorrow!’ She was pleased. b. He looked straight at her. I’ll definitely come back tomorrow! She remained unaware of his plan until the following day. Such a distinction is relevant to the present study as it allows readers of literature to differentiate between how a character’s mental and verbal representations are delivered to the reader and how such a difference could be deliberately misleading. More precisely, the non-presence of any reporting element makes it sometimes hard for the reader to set apart instances of narration (where no discourse presentation is being reported) and moments where the narrator is reporting a speech or thought event. A second relevant distinction in Semino and Short’s taxonomy lies within the new categories of Narrative Reports of Speech and Thought Act (NRSA and NRTA) (3a and 3b) and the categories of Narrator’s Representation of Voice (NV) and Internal Narration (NI) (4a and 4b). (3) a.

I again urged more flexibility, as I had learnt that some Ministers who would vote for her in the first ballot might not vote for her in the second. b. He looked straight at her and thought about his imminent return. She remained unaware of his plan until the following day. (4) a. We spoke to vice madam Michaela Hamilton from Bullwell, Notts, who arranged girls for a Hudson orgy at the Sanam curry house in Stoke. b. In the fleeting dizziness I had a nightmare vision. NRSA is different from IS in that it “prototypically has only one clause, with the ‘speech report’ verb often followed by a noun phrase or a prepositional phrase indicating the topic of the speech presented” Semino and Short (2004, p. 11). What is reported is, therefore, the “illocutionary force of a particular utterance” (ibid., p. 52). This category has a summarizing function even when the speech act verb or noun is followed by a lengthy description of the topic of the original speech event. NRTA, on the other hand, refers to cases where “someone is presented as performing mentally what would normally count as a speech act if it had been uttered” (Semino and Short, 2004, p. 130). As such, there is no clear reference to the propositional content of the thought and no use of a separate reported clause. The category of NV “captures minimal reports of speech, consisting either of simple references to the fact that someone spoke or of general

The Interface of SFL and Pragmatics


references to speech events” (Semino and Short, 2004, p. 69). There are several functions that this category could have, like introducing a speech event that will be reported in subsequent text or, in fiction, signalling that the speech event is (a) unimportant from the point of view that is adopted, or (b) too distant to be heard clearly from the point of view that is adopted, or (c) produced in such a way as to be inaudible from the point of view that is adopted. (ibid., p. 70) The category of NI refers to chunks where the narrator reports a character’s cognitive state or process, i.e. an experience that can be seen as involving some form of cognition, but without any indication of the occurrence of a specific thought act, let alone of any propositional content or wording that might have formed in the relevant person’s mind. (Semino and Short, 2004, pp. 132–133) This specific category might be controversial when looked at from an SFL perspective because it includes not only cognitive processes but also emotive ones as in (5). (5) A fortnight or so later Claudie invited Dr Wardener to have a glass of my father’s champagne and he called the nurses in to drink ‘my health’. The phrase made me a little sad. In examples like (5), Semino and Short argue that although an emotional state is being described, it has been caused by a cognitive process. Accordingly, whenever a cognitive process is involved, similar structures are tagged as NI. Such thought categories would therefore be identified via lexis that has to do “with cognition and emotion (e.g. verbs like ‘envied’, ‘amazed’, ‘sad’, ‘mind’, and nouns like ‘longing’, ‘curiosity’, ‘revulsion’)” (ibid., p. 133). 1.3. Pragmatic Approach to Reporting The third basis proposed for the analysis of reporting is the multifunctionality of reported speech in pragmatics. First, reporting is a confrontation of two ‘I’s. Reporting involves the person of the reporter and that or those of the reported people. Triki (1989 and 1991) has shown three main parameters to be constitutive of SELF in language. The first is the representation of the world in terms of person, space and time


Mounir Triki and Nesrine Triki

(the deictic parameter). The second is the representation of the world as perceived by one given centre (the perceptual parameter). The third is the representation of views about the world, feelings, attitudes, etc. (the affective/modal parameter). In all these centres, the mediation of the speaking subject is inexorably enshrined into and greatly influences the choice of linguistic forms. Fragmentation of SELF: Within a theory of Dialogical Self (DS) (e.g. St. Clair, Thomé-Williams and Su, 2005), there is no single centralized storyteller at work. DS theory retains James’ (1890) view of the self as consisting of different constituents: ‘I’ and ‘Me’. The ‘I’ refers to the reflexive parts of the self (the self as subject, knower, thinker, etc.), whereas the ‘Me’ is described as the sum of everything someone can be said to own (the self as object, as known, thought, etc.). Each ‘Me’, however, is endowed with a voice to tell its own story (St. Clair et al., 2005). We wish to argue that the dialogue between the narrator and his double in the short story at hand inscribes within the narrative of this inner conflict between competing aspects of SELF Heterogeneity of SELF: Hermans and Kempen (1993) develop the conception of the self as a multitude of different I-positions fluctuating within the imaginal landscape of the mind (Hermans, 2001), which is reminiscent of Bakhtin’s (2008) concept of ventriloquation to the effect that individual speakers always speak in the social languages of their time; thereby expressing the position of the group they belong to. Accordingly, we argue that the two voices competing for expression in “William Wilson” may be a form of ventriloquating. Culture and voice: Hermans, Kempen and Van Loon (1992) endorse the view that ‘Western’ culture, with its individualistic and rationalistic ideals of selfhood, influences the entire organization of dialogical selves, restricting its full potential and resulting in centralistically organized selves dominated by one or a few voices (St. Clair et al., 2005). The dominant conception of culture in DS theory today, however, stresses the existence of self and culture as a “multiplicity of positions among which dialogical relationships can develop” (Hermans, 2001, p. 243). In this conception, culture is a voice that may speak for itself or speak through another position (St. Clair et al., 2005). Thus, we argue that Poe could be said to be a precursor of this fragmentation of SELF into various social vantage points. Goffman’s (1986) Concept of the Self: Goffman, as reviewed by St. Clair et al. (2005), wants to show how even the most innocuous or apparently authentic of our social acts can be calculated to show the actor to his or her audience in a favourable light (Gouldner, 1970). Goffman’s is a social theory that dwells upon the episodic and sees life only as it is lived in a narrow interpersonal circumference, a-historical and noninstitutional, an existence beyond history and society, and one which comes alive only in the fluid, transient “encounter” (St. Clair et al.,

The Interface of SFL and Pragmatics


2005). Rather than conceiving of activities as a set of interlocking functions, Goffman’s dramaturgical model advances a view in which social life is systematically regarded as an elaborate form of drama in which— as in the theatre—men are all striving to project a convincing image of self to others.

2. Methodology To illustrate the complexity of reporting strategies and the need for combined approaches, a short story (Poe’s “William Wilson”) has been selected as a case study. The analysis of the text is to be considered as an illustration of the objectives previously mentioned and as a starting point for a more theoretically oriented discussion. Briefly, the story is about a protagonist depicted as wicked and evil (we will refer to him as Evil William) and a second character who bears his same first and family names, is born on the same day and comes from the same place, but who has the opposite moral values (he will be called Good William). Annoyed with Good William’s interference in his life, Evil William decides to get his revenge and stabs Good William only to discover that while he stabbed him he actually stabbed himself. This dual self of the narrator results in confusion about who is actually reporting and who is being reported, which is specifically the reason for which this short story has been chosen for investigation. The text (8068 words) was read manually, and all instances of reporting were coded using the Universidad Autónoma de Madrid Corpus Tool2 (O’Donnell, 2008) as either locutions or ideas and as either quotes or reports. We also investigated the use of modality in those reporting environments in order to identify the various enactments of evidentiality and attitudes. Various types of fuzzy structures were encountered. The first could be related to the blurry boundaries between narration and reporting; i.e. while some clauses seemed to be pure cases of narration, they turned out to be instances of reporting of what is referred to in the literature as the ‘Free’ type. The second difficulty is related to the classification of some verbal and mental processes and their potential projecting aspect as in: (6) While he spoke, so profound was the stillness that one might have heard a pin drop upon the floor. (7) [. . .] in which he was more than usually thrown off his guard, and spoke and acted with an openness of demeanor rather foreign to his nature. Indeterminacy in process-type categorization is a major difficulty faced by analysts while coding clauses. In particular, inconsistencies have been


Mounir Triki and Nesrine Triki

noted in the identification of verbal versus material processes (Gwilliams and Fontaine, 2015) and of verbal versus behavioural processes (O’Donnell, Zappvigna and Whitelaw, 2009). Gwilliams and Fontaine (2015) argue that a possible reason behind these coding differences is whether the analyst favours grammatical criteria for their coding decisions or whether they prioritize verb semantics. In line with this indeterminacy issue, the verb ‘speak’ in (6) and (7) could be thought of as a behavioural rather than verbal process (see O’Donnell et al., 2009, for a deeper discussion of coding differences in similar cases). Still, the two clauses are coded as cases of projection where the reader can understand that the Sayer ‘he’ was involved in some speaking activity but the content of this speaking was neither quoted nor reported in the clause. The content of the wording is sometimes recoverable from previous clauses. For example, clause (6) was preceded by a lengthy quote and could therefore be paraphrased as While he said “Gentlemen, I make no apology for this behavior . . . and the several little packages which may be found in the somewhat capacious pockets of his embroidered morning wrapper” so profound was the stillness . . . Such a coding decision is influenced by Semino and Short (2004) and by Hasan (1989). Hasan, for instance, argues that with such verbal processes in verbal art, what matters in these contexts are the circumstances and manners rather than the content itself (1989, pp. 86–87). Moreover, the identification of the source of projection, or in Martin and White’s (2005) terms the ‘Attribution’ of projected sequences, was not a straightforward task. It was almost impossible to make decisions about who was reporting who due to the dual self of the narrator (Evil William vs Good William) and the overlap between the two selves (which will be referred to as William Wilson).

3. Level and Mode of Projection in “William Wilson” Analysis has revealed the prevalence of reporting ideas at the expense of locutions (Table 5.2). Out of the 213 cases of projected speech and thought identified, locutions are projected in less than 32% of the total number of projection instances, while ideas amount to more than 68%. Locutions are related to several types of verbal and nominal processes as in (8) and (9).

Table 5.2 Level of projection in “William Wilson” Locution



68 32%

145 68%

213 100%

The Interface of SFL and Pragmatics


(8) He said that some person, apparently in great haste, demanded to speak with me in the hall. (9) “Mr. Wilson,” said our host, stooping to remove from beneath his feet an exceedingly luxurious cloak of rare furs, “Mr. Wilson, this is your property.” At the clause complex level, locutions are found to be realized paratactically and hypotactically as illustrated in (8) and (9). The verbal processes in said and demanded in (8) and said in (9) project the wordings of different characters in the text. Example (8) illustrates a common style in narratives, which is indirect speech or, in SFL terms, a ‘report’, whereas projection in example (9) takes the form of a ‘quote’ or direct speech. In terms of frequencies, quotes of locutions are not frequent in the short story, with only six cases identified. This alternation between quotes and reports seems to play a strategic role in the narrative process. Direct speech is used to quote very important messages in critical moments and scenes in the story (when Evil William’s cheating was discovered, when Evil William addresses Good William for the first time and when William Wilson stabbed himself) and gives therefore a theatrical, vivid and colorful aspect to the piece of fiction. Examples (10), (11) and (12), on the other hand, represent cases of projection where the reader has no access to the original verbiage that was uttered and where the verbiage itself comes in the form of “the name of saying” (Halliday and Matthiessen, 2014, p. 306) as in rumor, refusal and words or in the form of “the content of what is said” (ibid.) as in the day of the last conversation I there held with my singular namesake. (10) I was galled, too, by the rumor touching a relationship, which had grown current in the upper forms. (11) With a well-feigned show of reluctance, and not until after my repeated refusal had seduced him into some angry words which gave a color of pique to my compliance, did I finally comply. (12) I mention it at all but to define the day of the last conversation I there held with my singular namesake. Paratactic and hypotactic clause nexuses are also found in projected ‘ideas’. Ideas are mental representations of the narrator’s personal internal states of cognition or other characters’ as in (13), (14) and (15). (13) I would fain have them believe that I have been, in some measure, the slave of circumstances beyond human control. (14) I well remember it had no trees, nor benches, nor anything similar within it. (15) I perceived that we were even singularly alike in general contour of person and outline of feature.


Mounir Triki and Nesrine Triki

However, with only one exception of a paratactically quoted idea (example 16), all 19 direct quotes of ideas are used in the Free direct style (17, 18 and 19). (16) And again, and again, in secret communion with my own spirit, would I demand the questions “Who is he?—whence came he?— and what are his objects?” But no answer was there found. (17) This has been already too much an object for the scorn—for the horror—for the detestation of my race. To the uttermost regions of the globe have not the indignant winds bruited its unparalleled infamy? Oh, outcast of all outcasts most abandoned!—to the earth art thou not forever dead? to its honors, to its flowers, to its golden aspirations?—and a cloud, dense, dismal, and limitless, does it not hang eternally between thy hopes and heaven? I would not, if I could, here or to-day, embody a record of my later years of unspeakable misery, and unpardonable crime. (18) This reverend man, with countenance so demurely benign, with robes so glossy and so clerically flowing, with wig so minutely powdered, so rigid and so vast,—could this be he who, of late, with sour visage, and in snuffy habiliments, administered, ferule in hand, the Draconian laws of the academy? Oh, gigantic paradox, too utterly monstrous for solution! (19) Gasping for breath, I lowered the lamp in still nearer proximity to the face. Were these—these the lineaments of William Wilson? I saw, indeed, that they were his, but I shook as if with a fit of the ague in fancying they were not. What was there about them to confound me in this manner? I gazed;—while my brain reeled witha multitude of incoherent thoughts. The emboldened strings of words are to be understood as representations of the narrator’s inner thoughts. In (16), the italicized part stands as the projecting clause within which the circumstance in secret communion with my own spirit makes it clear that the process demand is not realized in the form of wording but rather in the form of thought. The use of opening and closing quotation marks explicitly signals the direct aspect of projection. Such explicit projecting elements and punctuation are not present in (17), (18) and (19) and in several other similar cases. Yet the emboldened strings could not be analysed as part of narration or as instances of Free Direct Speech because of two main arguments. The first is the sudden shift from the declarative, which is the unmarked style in narration (Semino and Short, 2004, p. 126), into the interrogative and exclamative; the second is the cues spotted in neighbouring co-texts as in (19), where the complex clause I gazed;—while my brain reeled with a multitude of incoherent thoughts offers hints that the questions asked in the previous clauses were not uttered but raised in the narrator’s mind.

The Interface of SFL and Pragmatics


Free direct ideas are construed as yes/no and WH-interrogative clauses and as exclamation clauses in interior monologues. Interrogative mental clauses, according to Halliday and Matthiessen (2014, p. 516), represent “an ‘undecided’ state of mind” and “are used to project indirect questions”. These Free direct categories are also construed in the form of vocatives,“Oh, outcast of all outcasts most abandoned!”, and signalled overtly via dashes and via punctuation marks (? and !) and juxtaposed in sequences of clauses. These vocatives, as suggested by Halliday and Matthiessen (2014, p. 510), cannot be reported and are construed as quotes. Interrogative clauses, however, could be paraphrased using a projecting clause of the type I thought to myself or If asked myself, followed by the projected sequences in the form of quotes. They are integrated within narration but could be understood as cases where the narrator shifts from narration of events or description of things and places into expressing inner thoughts and emotions related to unanswered questions, astonishment, admiration, anger or amusement. This deliberate blend of narration and projection contributes to the vivid aspect of the short story while, at the same time, it further complicates readers’ task in understanding the narrator’s flow of ideas. These ideas appear, therefore, to be as complicated and as mixed as the narrator’s identity. As is the case with locutions, ideas are also projected in the form of rank-shifted clauses as in (20) or prepositional phrases as in (21) or as a nominal group as in (22). (20) It was my intention, now, to put my scheme in operation, and I resolved to make him feel the whole extent of the malice with which I was imbued. (21) by bringing to mind dim visions of my earliest infancy—wild, confused and thronging memories of a time when memory herself was yet unborn. (22) I discovered, or fancied I discovered, in his accent, his air, and general appearance, a something which first startled, and then deeply interested me [. . .] Because reports are about reproductions of meanings rather than real wordings, with hypotactic combinations, the narrator has stronger influence on the speech or thought events being projected. Such an influence weakens when the narrator opts for quoting strategies, which have a role in “bringing characters to life” (Eggins, 2004, p. 274). Because the content of ideas cannot be interpreted as fulfilling speech functions of “giving and demanding information” (Halliday and Matthiessen, 2014, p. 516), we can only question and doubt whether the projected mental activities did or did not actually happen in the narrator and in the other characters’ minds; that is, the reader can raise questions about “the status of the validity of the information” (ibid.).


Mounir Triki and Nesrine Triki

The overwhelming projection of ideas (68% vs 32% for locution) reflects the type of narrative style used in the short story. It shows how the author, Poe, deliberately exposes his main character’s internal thoughts to the readers in order to enable them to build their own readings of his personality and his behaviour. Yet, as will be discussed in the next section, the sources of projection seem to hinder this reading and to mislead the reader.

4. Interpersonal Projection Two main aspects of interpersonal projection are identified in the short story. The first is the attribution of the projected locutions and ideas to a specific projector.3 The second relates to the various enactments of modalities accompanying projected chunks. Although the narrator is at the same time the projector in the short story, this same narrator turns out to have a dual self, one representing the evil, wicked and even criminal side of the character, while the second self, the doppelganger, is the direct opposite, featuring the epitome of good morals, kindness and respect. Throughout the text, it seems almost impossible to decide who of the two selves is narrating, quoting or reporting. In particular, the opening (example 23) and closing (example 24) sections of the short story are the most confusing. For this, the projector is considered as the source (even when not explicitly mentioned), who could be either Good William or Evil William. In these two sections, only 14 instances of projection were identified. (23) I long, in passing through the dim valley, for the sympathy—I had nearly said for the pity—of my fellow men. I would fain have them believe that I have been, in some measure, the slave of circumstances beyond human control. I would wish them [. . .] (24) It was Wilson; but he spoke no longer in a whisper, and I could have fancied that I myself was speaking while he said: “You have conquered, and I yield [. . .] In me didst thou exist—and, in my death, see by this image, which is thine own, how utterly thou hast murdered thyself.” In the opening scene, the ‘I’ is the source of all projected locutions and ideas, and although one could say that this ‘I’ refers to Evil William, and the ‘he’ in the closing scene refers to Good William, such a claim no longer suits when the reader reaches the last quote, whose source is at the same time ‘I’ and ‘he’ (the emboldened part in 24) and where the quote has two sayers, the two selves of the same character. The projector in the remaining part of the text is identified as Evil William (though a different interpretation could also fit), and he has projected his and others’ ideas and locutions 201 times. This shows that the

The Interface of SFL and Pragmatics


dominant voice readers can hear is that of the ‘Evil’, who is given much more space and time compared to the ‘Good’ one. Good William is given little space to express himself. All that the reader knows about him is delivered via the stories narrated by Evil William, who is the projector and source of all projected ideas and locutions. Only once was he quoted, which leaves room open for readers to doubt the truth behind Evil William’s descriptions and reports of his doppelganger. Interestingly, projected locutions and ideas in the short story are often blended with modality (119 instances in 213 occurrences of projection) (see Table 5.3). Modality in “William Wilson” contributes to the creation of figures where the narrator expresses wishes (example 25) about how he imagines others would think and behave, where he hesitates about his own words and thoughts (example 26) or where he refuses or denies actions and facts. In other words, modality helps in tracking the type of speech functions of ideas in particular, where “propositions are thought and proposals are hoped” (Halliday and Matthiessen, 2014, p. 527). Modality turns out to be a recurrent pattern in the short story, reaching therefore the status of ‘prominence’ that “contributes to the writer’s total meaning” (Halliday, 1971, p. 98). (25) I would fain have them believe that I have been, in some measure, the slave of circumstances beyond human control [. . .] I would have them allow—what they cannot refrain from [. . .] (26) It was about the same period, if I remember aright, that, in an altercation of violence with him (27) I can recall no occasion when the suggestions of my rival were on the side of those errors or follies so usual to his immature age [. . .] The juxtaposition of modal and projecting mental and verbal processes points to a subtle difference between real and hypothetical projection. Mental process clauses, as suggested by Halliday and Matthiessen (2014, p. 531), can be interpreted as cases of ‘hypothetical’, as opposed to ‘real’, projections and as implying ‘possible’ as opposed to ‘certain’occurrence of those mental activities. Hypothetical projections of ideas are found to be linked to the use of modal verbs like in (25), where the real alternative could be in the form of ‘I had them believe’ and ‘I had them allow’. These hypothetically projected ideas reflect William Wilson’s regrets and wishes

Table 5.3 Modality in projection Modal verbs Modal adjuncts Negative polarity Total

51 36 32 119


Mounir Triki and Nesrine Triki

about events and circumstances that are now impossible to change or to realize. Regrets and wishes, however, do not explicitly refer to one of the two selves. William Wilson explicitly hints at these hypothetical mental activities in the very beginning of the story where he wonders whether he could be living in a dream as illustrated in (28). (28) Have I not indeed been living in a dream? And am I not now dying a victim to the horror and the mystery of the wildest of all sublunary visions? It becomes even more difficult to trust the story of a dying narrator whose memories, to him, look more like a dream or a vision than real happenings. This, to some extent, could explain the mixed and distorted memories as in example (26). The presence of hypothetical projection helps in tracking and understanding the unstable state of mind of the narrator.

5. Pragmatic Analysis At this level of analysis, the focus shall be laid on deixis, modality and perception as intervening parameters in the construction and analysis of reporting strategies. Although these features have been integrated in the SFL section, it will be demonstrated that pragmatics could add more insights and could go beyond the lexico grammatical aspects. The ideational metafunction in SFL explains the various construals of experiences at the clause level and is analysed within the transitivity system. It has been shown (section 1.1.1) how reporting strategies are construed among the participants, the processes and the circumstances. We will explain how those participants, Sayers and Sensers in SFL, person deixis in Pragmatics, and circumstances of time and space (spatial-temporal deixis) could further intervene in the understanding of reporting events. 5.1. Deixis Within SFL, the study of deixis is linked to the various enactments of evidentiality, outlined in section 4, where the source of projection has been studied. The point has been made that there is a deliberate overlap between the narrator, who is at the same time the main character, and his conflicting selves, leading accordingly to confusion. Such confusion could be tracked along the different functions of person deixis in the short story, which are not necessarily used in the reporting instances but are rather located within a larger co-text. In the opening of the story, the narrator points to the fact that the name he is using to refer to himself is not his real name (29). This marks a first dichotomy between the ‘I’ of the real name of the narrator versus the ‘I’ of the fake name he chooses to use.

The Interface of SFL and Pragmatics


(29) In this narrative I have therefore designated myself as William Wilson,—a fictitious title not very dissimilar to the real. As the story progresses, the narrator alienates himself (the ‘I’ of Evil William) from his other self (‘he’ for Good William). Yet the ‘I’ and the ‘he’ refer to the same character, in which case the ‘I’ vs ‘Heh = ‘I’ (30) or the ‘I’ vs ‘he’ = ‘one’ as in (31). (30) This exception was found in the person of a scholar, who, although no relation, bore the same Christian and surname as myself;—a circumstance, in fact, little remarkable . . . Wilson’s rebellion was to me a source of the greatest embarrassment. Yet this superiority—even this equality—was in truth acknowledged by no one but myself. (31) It was difficult, at any given time, to say with certainty upon which of its two stories one happened to be. The deictic opposition ‘now’ versus ‘then’ expresses the opposition between the narrating self and the narrated self. However, either the present claim about the continuity of the past to include now is true, in which case the narrator is as outré as the protagonist, or it is not true, in which case the narrator is not reliable. Whatever the case might be, the reliability of the narrator is very much in doubt. This could be spotted in the ways place and time deixis are woven to create deliberate mystification of reference as shown in (32), where reference to the ‘here’ and ‘now’ of the reporter also equals the ‘there’ and ‘then’; in (33), where ‘these’ could be understood as either a deictic reference or an anaphoric reference; and in (34), where a clear distinction between the ‘now’ and ‘then’ is marked. (32) During the five years of my residence here, I was never able to ascertain with precision, in what remote locality lay the little sleeping apartment assigned to myself and some eighteen or twenty other scholars. (33) The extensive enclosure was irregular in form, having many capacious recesses. Of these, three or four of the largest constituted the play-ground. (34) Yet I must believe that my first mental development had in it much of the uncommon—even much of the outré. [. . .] In childhood I must have felt with the energy of a man what I now find stamped upon memory in lines as vivid, as deep, and as durable as the exergues of the Carthaginian medals. Reporting, therefore, inevitably involves reference to the grammatical category of person. The reporter has many options in so doing. One parameter would be whether the participants and the people referred to


Mounir Triki and Nesrine Triki

in the original speech event happen to be the same as or different from the participants and the people referred to in the reporting speech event. The determination of who is reporting what to whom about whom has a direct bearing on the choice of pronouns, which is taken here as a social demarcation strategy. The second parameter is the reporter’s perception of his/her social standing with respect to the addressee and the people referred to. This perception of power relations is reflected in the choice of referring expression. In addition to the grammatical category of person, reporting necessarily involves contextualization in the structures of space and time. The reporter has to locate the spatial deictic information with respect to a given deictic centre measured against his/her own underlying here (viewed as either proximal to it or distant from it). To the difficulties related to deixis is added one further complication overriding them, namely the speaker’s modal attitude to what is being reported (in terms of identification and belief vs distancing; involvement versus detachment). The above-mentioned possibilities provide the basis for the notion of narrative transformations (Triki, 1989). The argument is that despite the complex surface structures informing widely different narrative forms, every use of indices pertaining to person, space and time or to attitudes and modal judgements is amenable in essence to the binary opposition ‘I’ versus ‘not-I’. The complexity stems from the transformations which these basic structures undergo. The model proposed here is based on a basic opposition between one locus of consciousness as the axis of reference or anchorage and all the others being shifters (Triki, 1989); that is, they are defined with respect to this centre. The complexity stems from the confrontation of all these potential centres. The same linguistic devices could have different values depending on the identity of their defining centre. What is more, in case of displacement, they could be transformations of latent canonical devices. But sometimes transformations only operate half way through; that is, certain elements are transformed, and others are not; hence the ambivalence and potential confusion. 5.2. A Wider Context for Pragmatic Interpretation Because the case study here is an example of verbal art, literary, psychological, sociological and ideological perspectives are also needed to account for the subtle use of reporting strategies and the possible reasons behind the choices Poe made while depicting his protagonist and his plot. First, Kirsch’s (2011) functions of the unspeakable in American Gothic literature include the deliberate silences, irruptions, contradictions, omissions and other moments when the text either deliberately or perhaps unintentionally violates the conventions of written language—in order to explore their implications for socio-historical mindsets, for the

The Interface of SFL and Pragmatics


self-conscious construction of ‘American’ identity and for human psychology (Kirsch, 2011, p. 1). From the angle of reporting, the narrator’s insistence on the simultaneous existence of past and present (ibid., p. 2) and the ensuing simultaneity of history and present time suggests a profound anxiety over the impossibility of escaping history. This anxiety would be especially relevant and troubling for a region (the South) and a nation implicated in the immorality of slavery. Thus, in Poe’s “William Wilson,” linguistic excess (when this excess of words draws attention to its inability to contain definitive meaning) appears in moments of anxiety over the stability and uniqueness of the self, implicating a larger cultural insecurity of individual and national selfhood in a 19th century America which had only recently declared independence and had yet to fully distinguish itself culturally from Britain (ibid., pp. 1–2). Evidence from psychology (Scottoline, 2008, p. 91) shows that the perception of a threat to SELF is greater when it comes from within than from without. Scottoline (2008, p. 91) contends that Poe may not have invented the Evil Twin, but he certainly anticipated it, as well as the spookiness that comes from the fragmenting or doubling of the self and the splintering of identity. Poe must have known that no monster is half as scary as the evil within us, and it is tempting to wonder if he ‘wrote what he knew’, considering his own personal unhappiness and the fact that he assigned William Wilson his own birthday. Finally, the religious/biblical dimension is best developed by Scottoline (2008, p. 88), who compared the narrator to ‘Ishmael’, the Biblical envious brother of Isaac (Genesis XVI, XVII). She finds the twist/reversal of the doppelganger motif quite original (Scottoline, 2008, p. 89). Instead of the main character being the good one and the double being the bad one, in “William Wilson” the narrator is the bad one and the double is the good one. This reversal has intertextual allusions to John Milton’s Paradise Lost, where Satan was given a higher degree of audibility and credibility than Jesus (Scottoline, 2008, p. 90).

6. Prospects for the SFL-Pragmatics Interface The analysis of the short story introduced in sections 3, 4 and 5 constitutes an illustration of how texts could be analysed following a two-step procedure. The first step focuses on the lexico grammatical realizations of a specific language function such as reporting, leading thus to a deep understanding of the various choices made to report speech and thought. The power of the SFL framework in this first phase is to provide a clear understanding, both quantitatively and qualitatively, of the different meanings expressed, including metaphorical ones. A finer distinction between the various options is shown to be more prominent when other frameworks, such as the ones suggested by Leech and Short (1981) and Semino and Short (2004), are considered.


Mounir Triki and Nesrine Triki

When it comes to the overall understanding of the hidden, and perhaps the deliberate intentions of authors, pragmatic analysis comes into play as a second step and opens up the scope to look at the inferential meanings of reporting strategies by considering wider contexts. Certain questions raised in pragmatics could not be answered in SFL. These questions, although rooted in the social, cultural and ideological dimensions of language, an aspect of language use immensely recognized in SFL, but relatively not thoroughly followed in the analysis and interpretation of the grammatical analysis, could be handled by pragmatic analysis. The focus in pragmatics is on what is not explicitly stated; in other words, the focus is not on what a certain utterance means but rather on what a speaker means by a specific utterance (Leech, 1983). The notion of meaning beyond the clause could as such be extended to cover meanings beyond the surface level of instantiation to encompass pragmatic nuances related to ideological implications of the surface meaning as rightly suggested by Martin (1986, p. 227) in his extension of the scope of context to include ‘ideology’ as a level higher than genre and register. This complementarity between SFL and pragmatics is subject to a continuing debate in the literature about the extent to which SFL and pragmatics share common ground and common objectives. The two theories claim to provide analytical tools and methods that enable researchers to understand how language users deploy semiotic systems to interact with other language users in specific contexts. SFL suggests the investigation of language based on the three metafunctions of language which are thought to cover all aspects of human language properties, in the sense that any clause uttered would necessarily evoke a particular meaning that will be construed via the type of process used, the participants involved in the meaning generation and a particular relationship linking the speaker/ writer and hearer/reader. Yet, being a form of artistic writing whereby the author would transcend language and meanings created in the language, literary texts are often elusive in accepting linguistic analyses that stop at the level of description and interpretation based on the sole tools in the theories of grammar, hence the need for a pragmatic approach. This interface has been advocated by a number of researchers, pioneered by Butler (1988, p. 5), who acknowledges the “substantial overlap between Halliday’s concerns and those of pragmaticians”. Wylie (2009) advocates a functional account of the ideational metafunction as the seat of reference, being as it is the most pragmatic of the metafunctions rather than the least. He argues that Leech fails to make a convincing case for rejecting functionalism, while Halliday fails to make a convincing case for accepting it. Similarly, Song (2005) calls for a complementary relation between pragmatics and SFL. On the one hand, pragmatics and SFL have common interests in the function, context and meaning of language. On the other hand, the notions of presupposition and conversational implicature in pragmatics certainly benefit SFL, whereas the semiotic and functional conceptions of language in SFL favour pragmatics.

The Interface of SFL and Pragmatics


The pragmatic perspective (Mey, 2001) emanates from the importance attributed to the social dimension of language. As Pietarinen (2007, p. 127) has rightly argued, speech is not a personal possession, but a social one; for it belongs, not to the individual, but to the member of society. Moreover, Pietarinen (2007, pp. 129–131) notes the heterogeneous models of context: either being aware of the particulars of the conventions or normative grounds of language use may require a reconstruction of the socio-historical context within which the utterances must be understood, or the utterer’s and the interpreter’s propositional attitudes, including beliefs and intentions, delineate a different, cognitive type of context.

7. Conclusion In this chapter, we have examined the complexity of reporting in an example of fictional narrative to show how verbal and mental representations of characters can be so complex that a combination of linguistic and pragmatic tools could be suitable to better unveil their realizations and meanings. We have argued that such a combination of approaches is indeed helpful in the analysis and interpretation of verbal art where categories of narration and projection, on the one hand, and categories of lexico-grammar and logico-semantic meanings, on the other hand, could overlap and mislead the reader. We have demonstrated that SFL offers a linguistic framework of analysis which can constitute the basis of pragmatic and literary interpretations. Thus, understanding how locutions and ideas, for instance, are construed at different group and clause levels is a necessary step towards a solid interpretation of literary texts in terms of the socio-cultural, psychological and ideological mechanisms that govern narrators’ and characters’ use of reporting strategies.

Notes 1. All examples in this section are from Semino and Short (2004). 2. 3. The term ‘projector’ will include both Sayer and Senser.

References Arús, J. (2008) Ideational and interpersonal manifestations of projection in Spanish. In C. Wu, C. Matthiessen and M. Herke (Eds.), Proceedings of ISFC 35: Voices Around the World. Sydney: International Systemic Functional Congress, pp. 195–200. Bakhtin, M. M. (2008) The Dialogic Imagination: Four Essays, translated by C. Emerson and M. Holquist, edited by M. Holquist, 7th ed. Austin and London: University of Texas Press. Butler, C. S. (1988) Pragmatics and systemic linguists. Journal of Pragmatics, 12, pp. 83–102.


Mounir Triki and Nesrine Triki

Eggins, S. (2004) An Introduction to Systemic Functional Linguistics. New York and London: Continuum. Fowler, R. (1986) Studying literature as language. In T. D’haen (Ed.), Linguistics and the Study of Literature. Amsterdam: Rodopi. Goffman, E. (1986) Frame Analysis: An Essay on the Organization of Experience. Boston: Northeastern University Press. Gouldner, A. W. (1970) The Coming Crisis of Western Sociology. New York: Basic Books. Gwilliams, L. and Fontaine, L. (2015) Indeterminacy in process type classification. Functional Linguistics, 2(1). Halliday, M. A. K. (1971) Linguistic function and literary style: An enquiry into the language of William Golding’s ‘the inheritors’. In S. Chatman (Ed.), Literary Style: A Symposium. New York: Oxford University Press, pp. 330–368. Halliday, M. A. K. and Matthiessen, C. M. I. M. (1999) Construing Experience Through Meaning: A Language-based Approach to Cognition. London: Continuum. Halliday, M. A. K. and Matthiessen, C. M. I. M. (2014) An Introduction to Functional Grammar, 4th ed. London: Edward Arnold. Hasan, R. (1989) Linguistics, Language, and Verbal Art. Oxford: Oxford University Press. Hermans, H. J. M. (2001) The dialogical self: Toward a theory of personal and cultural positioning. Culture & Psychology, 7(3), pp. 243–281. Hermans, H. J. M. and Kempen, H. J. G. (1993) The Dialogical Self: Meaning as Movement. San Diego: Academic Press. Hermans, H. J. M., Kempen, H. J. G. and Van Loon, R. J. P. (1992) The dialogical self: Beyond individualism and rationalism. American Psychologist, 47, pp. 23–33. James, W. (1890) The Principles of Psychology, vol. 1. New York: Henry Holt. Kirsch, E. (2011) ‘Why? Why? Why? Why? Why?’: Functions of the ‘unspeakable’ in American gothic literature. AHSS Summer Research Report. Leech, G. (1983) Principles of Pragmatics. London and New York: Longman. Leech, G. and Short, M. (1981) Style in Fiction. London: Longman. Leech, G. and Short, M. (2007) Style in Fiction: A Linguistic Introduction to English Fictional Prose. New York: Pearson Longman. Martin, J. R. (1986) Grammaticalising ecology: The politics of baby seals and kangaroos. In T. Threadgold, E. A. Grosz, G. Kress and M. A. Halliday (Eds.), Language, Semiotics, Ideology,vol. 3. Sydney: Sydney Association for Studies in Society and Culture. Martin, J. R. and White, P. R. (2005) The Language of Evaluation: Appraisal in English. Basingstoke: Palgrave Macmillan. Mey, J. (2001) Pragmatics. Oxford: Blackwell. O’Donnell, M. (2008) The UAM CorpusTool: Software for corpus annotation and exploration. Proceedings of the XXVI Congreso de AESLA, Almeria, Spain. O’Donnell, M., Zappvigna, M. and Whitelaw, C. (2009) A survey of process type classification over difficult cases. In J. Carys and E. Ventola (Eds.), From Language to Multimodality: New Developments in the Study of Ideational Meaning. UK: Equinox eBooks Publishing, pp. 47–64. Pietarinen, A. V. (2007) On historical pragmatics and Peircean pragmatism. Linguistics and the Human Sciences, 2(1), pp. 123–143.

The Interface of SFL and Pragmatics


Scottoline, L. (2008) Identity crisis. In M. Connelly (Ed.), In the Shadow of the Master: Classic Tales by Edgar Allan Poe. Harper Collins e-books, pp. 87–94. Semino, E. and Short, M. (2004) Corpus Stylistics: Speech, Writing and Thought Presentation in a Corpus of English Writing. London: Routledge. Song, Y. (2005) The complementary relation between pragmatics and systemicfunctional grammar. US-China Foreign Language, 3(11). St. Clair, R. N., Thomé-Williams, A. C. and Su, L. (2005) The role of social script theory in cognitive blending. In M. F. Medina and L. Wagner (Eds.), Special Issue of Intercultural Communication Studies XI, vol. V. Thompson, G. (2014) Introducing Functional Grammar. London: Routledge. Triki, M. (1989) Linguistic and Perceptual Subjectivity: Towards a Typology of Narrative Voice. Unpublished PhD thesis, Essex University, UK. Triki, M. (1991) The representation of SELF in narrative. Journal of Literary Semantics, 20(2), pp. 78–96. Wylie, P. K. (2009) Leech and the limits of functionalism. In C. Wu, C. Matthiessen and M. Herke (Eds.), Proceedings of ISFC 35: Voices Around the World. Sydney: International Systemic Functional Congress, pp. 161–166. Wynne, M. (2005) Stylistics: Corpus approaches. In K. Brown (Ed.), Encyclopaedia of Language and Linguistics. Oxford: Elsevier. Available at: www.pala.


The Notion of a Multilingual Meaning Potential A Systemic Exploration Christian M.I.M. Matthiessen

1. Introduction: The Notion of a Multilingual Meaning Potential In this section, I will introduce the notion of a multilingual meaning potential and also a form of representation of it by means of multilingual system networks. I will begin by locating meaning as the key property of language in an ordered hierarchy of systems operating in different phenomenal realms, suggesting that these systems can be characterized as different kinds of potential—language being meaning potential. Then I will relate meaning potential to the phenomenon of multilingualism— emphasizing that it has been a default condition throughout most of human history, not an optional extra, and I will suggest the need to focus on the notion of a multilingual meaning potential. Preparing the ground for the rest of the discussion, I review our work on multilingual system networks as a key part of the representation of multilingual meaning potentials (see, e.g., Bateman, Matthiessen and Zeng, 1999). 1.1. Meaning Potential as a Fourth-Order Kind of Potential Language is a complex adaptive system—a very complex one, most likely the most complex of all human systems. This complexity is brought out if we locate language in an ordered typology of systems operating in different phenomenal realms, as proposed by Halliday (1996): see further Halliday and Matthiessen (1999/2006, Chapter 15), Halliday (2003a, 2005) and Matthiessen (2007). In this typology, there are four orders of system; they are, in increasing order of complexity (Figure 6.1): First order: Second order: Third order: Fourth order:

physical systems biological systems [physical + life] social systems [biological + value] semiotic systems [social + meaning]

Each higher order represents a new form of organization that is ‘overlaid’ on top of the lower order of organization. Thus, a system of a higher

A Multilingual Meaning Potential


order is also a system of the lower order(s), and in this respect it is more complex than systems of lower order(s). Language is a fourth-order, semiotic system, and it is thus also social, biological and physical—having the properties of each order of system. Thus, higher order systems inherit the properties of lower order systems. For example, language is embodied, having co-evolved with the human brain (biological);1 and it is collective as well as individual, being created, maintained and negotiated in interaction (social). Language can therefore be conceptualized and characterized in a number of different ways, from the vantage point of systems of different orders—the systems from which its properties derive. For example, language can be explored as an event potential through the analysis of soundwaves (physical), as a neuro-motor action potential through the analysis of neuro-motor actions (biological), as an interactive behaviour potential through the analysis of interactive behaviour (sociological) and as a meaning potential through the analysis of acts of meaning (semiotic). Since language is a fourth-order system, this last conception presupposes the other conceptions. Thus, to get at the defining characteristics of language, we study and model it as a meaning potential. However, in so doing, we note, as I put it above, that it inherits the properties of the other potentials: crucially, meaning potentials have the social properties of behaviour potentials—being involved in social interaction, negotiation and construction—and they have the biological properties of action potentials—being embodied together with other neuro-motor systems. The same poly-systemic point extends to the conception of individuals and their relationships to groups (cf. Halliday and Matthiessen, 1999, pp. 610–611): individual organisms (biological order) take on the many social roles that make up a person in interaction with others in social role networks (social order; cf. Firth, 1950; Halliday, 1978a; Butt, 1991), and persons take on the many semiotic roles that make up a meaning in the exchange of meanings with other meaners in communication networks (semiotic order; for some comments on the significance of this poly-systemic understanding of individuals in the conceptions of patients in healthcare, see Matthiessen, 2013b). This is of fundamental importance as we develop the notion of a meaner’s personalized meaning potential (see Matthiessen, 2009a), and it lies at the heart of accounts of identity and intersubjectivity, notions that have been foregrounded in the last couple of decades. Halliday introduced the conception of language as a meaning potential over four decades ago. He (1973, p. 55) characterized meaning potential as a form of behaviour potential, which makes perfect sense in terms of the ordered typology of systems represented in Figure 6.1. In this way, we can define language as a meaning potential, recognizing that in its internal organization, it is also a wording potential and a sounding potential (or a writing potential or a gesturing potential, depending on the nature of the expression plane).


Christian M.I.M. Matthiessen


4 : +meaning

meaning potential


3 : +value

behavioural potential


2 : +life

action potential


event potential


Figure 6.1 The ordered typology of systems operating in different phenomenal realms with associated potentials—increasing complexity, decreasing domain of operation

It is important to emphasize that characterizing language as a meaning potential is in no way a reification of language; continuing Firth’s understanding of language as activity, Halliday (1961/2002, pp. 42–43) has always theorized it as system-&-process, even half a century ago, when the general conception of language was still a very static one. This view contrasted with the “ ‘bricks and edifice’ view of language” (op cit., p. 65) that was the common default conception at the time and remained so for many years afterwards. Halliday continued to explore language as activity, and on one occasion he used the term languaging, taken from Peter Doughty, to bring this out (Halliday, 1985/2003, pp. 192–193): the noun language is of course a grammatical metaphor—Peter Doughty preferred to talk about languaging, to show that our object of study is more process than entity . . . Linguistics is about how people exchange meanings by ‘languaging’. This is very much in keeping with his (1961) characterization of language as patterned activity. ‘Potential’ is always potential to be actualized or instantiated as activity. The meaning potential is instantiated in acts of meaning (cf. Halliday, 1992a, but also, e.g., Halliday, 1977, on text as

A Multilingual Meaning Potential


semantic choice, i.e. as process). Potential and instance are related by the cline of instantiation, but not as separate phenomena—as Halliday has repeatedly emphasized (e.g. Halliday, 1991, 2002)—but rather as phases of the same phenomenon that can be observed from different vantage points. The conception of language as a meaning potential has had rather farreaching consequences in the theoretical, descriptive and analytical work on language in Systemic Functional Linguistics (SFL). For example, it made it possible for Halliday (e.g. 1991, 1992a, 2002) to extend meaning along the cline of instantiation from meaning potential to instantial acts of meaning, and to conceptualize language development as learning how to mean. Halliday (e.g. 1975) models ‘learning how to mean’ as the gradual development of a meaning potential—involving both continuity and transformation, with a growing number of options in meaning and thus increasing power to mean. 1.2. Multilingual Meaning Potential Against this background, I will now relate the notion of language as meaning potential to multilingualism;2 I will explore the notion of a multilingual meaning potential (cf. Ellis, 1987; Bateman, Matthiessen, Nanri and Zeng, 1991; Bateman et al., 1999). A multilingual meaning potential has all the physical, biological, social and semiotic properties of language as a meaning potential, for example being embodied in multilingual brains and enacted in multilingual groups; but in addition to its lower order properties, a multilingual meaning potential has a semiotic property: it is the potential to mean in more than one language. From the point of view of a speaker—a multilingual meaner—meaning in more than one language can manifest itself in a number of different ways, e.g. •

meaning in one language or another, depending on the meaning potential of the people the speaker is exchanging meanings with—as happens to travellers; meaning in one language and another, in the exchange of meanings with people who are also able to mean in more than one language— explored under the headings of code switching and code mixing; meaning in one language and another, but listening or reading in one and speaking or writing in another, recreating meanings and thus mediating between languages—that is, translating or interpreting.

These activities all presuppose that the speaker has a multilingual meaning potential,3 and as linguists we need to theorize and describe such multilingual meaning potentials—and we need to do it in such a way that we bring out the empowering nature of a multilingual meaning potential, showing very clearly what a multilingual meaner can mean in


Christian M.I.M. Matthiessen

a very wide range of contexts.4 In addition, we also need a model of the multilingual meaning potential to produce accounts of second language learning—that is, of the process whereby speakers learn how to mean in a new language (cf. Halliday, 1978b; Matthiessen, 2009a). Thus, after summarizing Halliday’s (1975) account of learning how to mean in mother tongue development, Saville-Troike (2012, Kindle location 1630) writes (my emphasis): One application of Halliday’s model to the study of SLA comes with seeing L2 learning as a process of adding multilingual meaning potential to what has already been achieved in L1. This is an approach that some of my colleagues and I have taken in our research. We have concluded that “Second language acquisition is largely a matter of learning new linguistic forms to fulfil the same functions [as already acquired and used in L1] within a different social milieu” (SavilleTroike, McClure and Fritz, 1984, p. 60). This needs to be conceptualized and investigated in light of Ortega’s (2013a, Kindle location 1301) vision: moving away from a “monolingual bias” to “a transformative bi/multilingual turn”. Thus, we need to use the theory and representation of the multilingual meaning potential to support extensive studies—preferably longitudinal— of how people learn how to mean in two or more languages (ranging from conditions of coordinate bilingualism to later bilingualism)—working towards the transformative multilingual turn. In addition, still in the area of second language learning and teaching, the model of multilingual meaning potential can also support what I would think of as a ‘helical’ return to contrastive analysis in the 21st century— contrastive analysis involving the notion of a multilingual meaning potential (cf. the contrastive account of Spanish in reference to English by Lavid, Arús and Zamorano-Mansilla, 2010). It might be thought that theorizing the multilingual meaning potential is essentially a task for certain specialist areas within linguistics concerned specifically with different manifestations of multilingualism (with bilingualism as a special case) and language contact. However, a very cogent case can be made that it is monolingualism that is exceptional and that multilingualism should therefore be a central concern in general linguistics. In other words, a general theory of language must be one that takes account of the multilingual meaning potential of speakers and of a speech community. Evans (2010) describes the situation in Arnhem Land (and other places around the world too), showing how multilingualism is the norm—and seen as a positive feature (in contrast with the Tower of Babel myth). Here multilingualism is both societal and individual. As noted above, within the field of ‘Second Language Acquisition’5 in applied linguistics, Lourdes Ortega (e.g. 2013a, 2013b) has argued very

A Multilingual Meaning Potential


cogently that the time has come for what she calls the bi/multilingual turn, building on but extending the social-interactive turn (which came after the cognitive turn). She points out that ‘Second Language Acquisition’ has been dominated for far too long by the conception of the monolingual speaker; i.e. it has taken the monolingual speaker as a point of reference even though taking the bilingual (or multilingual) speaker as the point of reference makes much more sense, especially in view of the fact that second language learners are striving to become bilingual speakers. She also problematizes the notion of a native speaker (whether ideal or not), bringing out the theoretically and ethically undesirable assumptions embodied in this notion (cf. Davies’ [2013], challenge of the distinction between native speakers and native users). There is, of course, a very extensive body of work on multilingualism (and bilingualism) as a social, cognitive and neurological phenomenon; and there is growing recognition for the need to take and develop the multilingual turn, as is brought out by May (2013). However, there do not appear to be a wealth of studies concerned with the representation of the kind of multilingual resources that must be characteristic of multilingual speakers (although there are certainly detailed explicit accounts in work on machine translation and multilingual text generation).6 And proposals that have been made have focused on syntactic structure; that is, they have been couched in terms of theories that are syntagmatically oriented rather than paradigmatically oriented as SFL is—and has been since the first half of the 1960s (see, e.g., Halliday, 1966a). Thus, the work by Sankoff (1998) and Sankoff and Poplack (1981) is concerned with the representation of syntactic structures needed to account for the possibility of code-switching but also for the constraints on it. The ‘formal linguistic theories successively in vogue’ have all been structure-based. The approach taken to grammar is thus syntagmatic in orientation rather than paradigmatic, but while we certainly need fully explicit representations of grammatical structure in functional terms, the paradigmatic orientation of SFL would make it easier to relate the representation of multilingual grammars to contextual factors influencing code switching (and code mixing) that have been investigated in sociolinguistics (see, e.g., Wodak, Johnstone and Kerswill, 2010, Part 5). This is essential if we are to develop a deeper understanding of code-switching as a multilingual meaning-making process. Further, researchers have tended to focus on one multilingual phenomenon such as code-switching rather than on the broader range of multilingual phenomena I mentioned above, but ultimately we will need to tackle all these phenomena. For example, a speaker who can ‘code switch’ is very likely also able to interpret and translate—not professionally, of course, but as a multilingual speaker. And, in addition, it makes sense to view multilingual processes together with multisemiotic, or ‘multimodal’, ones, for example to capture the complementarity of and division of semiotic labour between speaking


Christian M.I.M. Matthiessen

and gesturing in different language and gesture combinations (see, e.g., McNeill and Duncan, 2000; Chui, 2012). Thus, we can bring together the notions of a multimodal meaning potential (or better: a multisemiotic meaning potential; cf. Matthiessen, 2009b, on the degree of integration of meaning in different semiotic systems) and of a multilingual meaning potential to construct a multisemiotic and multilingual meaning potential. 1.3. The Development of a Representation of Multilingual System Networks To explore the notion of the multilingual meaning potential, I will draw on research John Bateman and I started about two decades ago in the research context of multilingual text generation informed by SFL. In the 1980s, in the context of text generation research at the Information Sciences Institute (ISI), University of Southern California and Kyoto University, a number of us worked on explicit computational representations of systemic functional grammar and on descriptions based on these representations of English and of Japanese. John Bateman and I worked on various problems in the representation of English and Japanese together—first while John was at the University of Kyoto and then when he joined us at ISI—and wrote up our experience of text generation based on SFL in Matthiessen and Bateman (1991). I documented the systemic functional description of English in Matthiessen (1995). This led naturally to the formulation of the task of multilingual text generation, i.e. the generation of texts in two or more languages based on the same ‘knowledge base’, or ‘meaning base’ as we conceived of it in Halliday and Matthiessen (1999). A central challenge was how to model and represent the resources of different languages as multilingual resources that could support multilingual text generation. There were various approaches that had been developed in machine translation—in particular, the interlingua and transfer approaches (both of which were advances over direct translation methods)—that might have been possible to draw on. However, they are syntagmatically oriented and not really informed by linguistic theory. So we built on the representation of system networks with realization statements, and extended their representational capacity so that it was possible to handle two or more languages—in principle, any number of languages; these are multilingual system networks: see Bateman et al. (1991) and Bateman et al. (1999). Our guiding principle was that the multilingual system network needed to integrate the resources of the different languages involved but that the integrity of each language needed to be maintained. Based on our framework, I directed a few projects concerned with multilingual text generation involving English, Chinese and Japanese, where Zeng Licheng made a central contribution to the computational modelling of multilingual system networks (e.g. Zeng,

A Multilingual Meaning Potential


1996; Matthiessen et al., 1998), and John developed the framework into a full-fledged platform for representing multilingual lexicogrammatical resources, the Komet-Penman MultiLingual (KPML) system (e.g. Bateman, 1997), which he has made freely available to the research community.7 It has supported a great deal of research, including the systemic functional description of the grammars of quite a range of languages. The multilingual system network representation that we developed originally in the early 1990s has been used in a number of examples comparing and contrasting a small number of languages: see, e.g., Teich (2003); Matthiessen (2004, p. 597); Teruya et al. (2007); and Matthiessen, Teruya and Wu (2008). 1.4. Organization of the Chapter 8 Drawing on the experience of working with multilingual system networks, I would like to take a step back, as it were, and view them as part of the account of the multilingual meaning potential of a speaker, and of a community. This will involve exploring the notion of the multilingual meaning potential in terms of some of the dimensions of systemic functional theory—very centrally, the hierarchy of stratification. My hope is that we will begin to see much more extensive work on the description of multilingual meaning potentials involving different sets of languages. Such descriptions are valuable in their own right and can support a range of applications—including a 21st century version of contrastive analysis. But at the same time, we need such descriptions to tackle hard multilingual phenomena theoretically. I will explore the modelling of multilingual meaning potentials step by step, starting with cases where differences between languages can be handled locally in the overall account of the multilingual meaning potential of two or more languages and moving towards cases where they cannot be accommodated locally but have to be dealt with non-locally. In Section 2, I will introduce examples of multilingual meaning potentials that can be represented as local to lexicogrammar, starting with differences that are confined to the syntagmatic axis and moving to differences that are systemic in nature and thus have to be represented paradigmatically. In Section 3, I will turn to examples of differences that are too great to handle within the lexicogrammatical stratum and thus have to be handled semantically. These examples involve complementary systems such as tense and aspect as complementary models of time. In Section 4, after having suggested in the previous section that we need to move up from lexicogrammar to semantics when it is not possible to bring out common meaning potential in the description of the lexicogrammatical systems of the languages constituting the meaning potential but also when it is necessary to show that similar lexicogrammatical


Christian M.I.M. Matthiessen

systems are deployed in semantically different ways, I will explore differences that arise in the semantics of registers operating in different contexts. In Section 5, I will summarize the account of the multilingual meaning potential introduced step by step in the previous section and suggest how this account can serve as a basis for a research programme concerned with multilingual meaning potentials.

2. Multilingual System Networks Within the Lexicogrammatical Stratum In the representation of multilingual meaning potentials involving two or more languages, certain types of difference—or incongruence—between the languages that make up the potential can be represented within the lexicogrammatical stratum against the background of common—or congruent—resources, and this is particularly likely to be the case when the different languages are quite close to one another, the limiting case being different dialects of the same language (as becomes clear in the context of translation in work on ‘dialect adaptation’; see, e.g., Weber and Mann, 1981; Mann and Weber, 1979). In the representation of the multilingual potential of lexicogrammar, I will start with the representation of structural differences—i.e. differences in the realization of systemic terms—in the environment of systemic similarities, and then move on to systemic differences. 2.1. Structural Realizational Partitions When we compare and contrast lexicogrammatical systems, we find many instances of systemic similarities and structural differences: languages have the same systemic contrast between two or more systemic terms, but the terms may be realized in different ways along the syntagmatic axis. This is variation ‘from below’ of the kind we would expect across the dialects of a language. For example, many—probably all—languages have a system of polarity with two primary terms, ‘positive’ vs ‘negative’; but while it is the ‘negative’ term that is marked in realization along the syntagmatic axis, realizations are fairly varied across languages although the variation is constrained and principled (see, e.g., Payne, 1985), as I illustrate in Matthiessen (2004, pp. 625–631): see Figure 6.2. The variation in the forms of realization does not affect the valeur of either term of the system; what matters is that the contrast between ‘positive’ and ‘negative’ is realized syntagmatically in some way so that the contrast is maintained syntagmatically. There are innumerable other examples of this kind, e.g. number (singular/plural, singular/dual/plural and a few other variants), tense (past/present/future, past/non-past), aspect (perfective/ imperfective).

A Multilingual Meaning Potential


Interpersonal POLARITY:

clause complex


negative verb projecting clause (e.g. Tongan ika) negative particle (e.g.


German nicht)



negative auxiliary (e.g. even ki )



verb: negative affix (e.g. Turkish ma; English- n’t ))

verb(al) form or particle closely associated with verb

Figure 6.2 Typological variation in realizations of the term ‘negative’ in the system of polarity (‘positive’/‘negative’)

The typological generalizations discussed above suggest that in a representation of a multilingual meaning potential, we must make it possible to show that languages share (parts of) the systemic organization and differ (to some extent) in how systemic options are realized syntagmatically. To achieve this, we use conditioned partitions within the multilingual system network. The partitioned parts of the system network are conditioned by languages. Any realization statement that does not appear within a partition is common to all the languages represented in the multilingual system network, but realization statements that appear within partitions are specific to the language or languages that condition the partition. 2.2. Systemic Partitions: Delicacy In our representation of the multilingual system network, it is possible to specify partitions for realization statements in different languages. Thus, we can ‘factor out’ syntagmatic differences while capturing systemic commonalities. However, the languages that make up a multilingual meaning potential may of course also differ systemically. Observations of a range of systems in many languages suggest that low-delicacy systemic contrasts are more likely to be congruent across languages than


Christian M.I.M. Matthiessen

higher-delicacy ones: as the delicacy of the description of systems is increased, we are likely to see more variation across languages—and this becomes very clear as we move from the grammatical zone of fairly low delicacy within lexicogrammar into the lexical zone of high-delicacy distinctions.9 As Teruya et al. (2007) show, this certainly seems to be the case for mood systems. The underlying principle is the one I already alluded to above: the grounding of the mood systems in the fundamentals of semantic systems of exchange. More delicate systems reflect variation across languages in the ‘fine-tuning’ of negotiation relating to force, key, bias (expectation), status and the like. There are clearly abundant examples of areas where languages differ with respect to the degree to which they differentiate the area in delicacy, so when we model multilingual meaning potentials we take account of the fact that they have to allow for variable delicacy across languages. I will just mention a couple of examples from systems in the ideational metafunction. In English, in the system of process type, one of the six primary process types is ‘existential’; but in Japanese, it is necessary to take one more step in delicacy to differentiate between aru (for inanimate Existents) and iru (for animate Existents). Similarly, in English, there is a general ‘intensive’ relational clause type with be as the Process; but in Spanish, it is necessary to take one more step in delicacy to differentiate between ser and estar (which is also used in ‘circumstantial’ clauses) as the Process. 2.3. Systemic Partitions: Delicacy and Simultaneity Multilingual system networks can accommodate differences in terms of delicacy among the languages that they cover: one language may embody further systemic differentiation than another. This is captured in the same way as syntagmatic differences by specifying a language-specific partition for the language that differentiates a given option further in delicacy. In addition, languages may also differ in terms of the number of dimensions of choice they involve in organizing some domain paradigmatically. In terms of the system network, this is the question of how many simultaneous systems are involved. It’s often the case that where some languages operate with a single systemic dimension in the organization of some semantic space, other languages operate with two or more. For example, within the interpersonal metafunction, languages may add more dimensionality by means of systems concerned with distinctions in tenor relating to power and solidarity (a classic example being Brown and Gilman’s 1960 study of the organization of pronouns in terms of power and solidarity in different languages).

3. Complementary Systems So far the multilingual lexicogrammatical system networks I have considered have involved both shared systems and realization statements and

A Multilingual Meaning Potential


non-shared ones represented within language-specific partitions. However, we also find areas within multilingual system networks at the level of lexicogrammar where two (or more) languages share no potential at all. This is likely to happen when languages operate with complementary models of some area of meaning (for the general notion of complementarities in language, see Halliday, 2008). Research is needed in order to produce a typological overview of sets of such complementary systems, but some complementarities have already been identified, including the transitive and ergative models of participation in the process of a figure (realized by a clause) and the tense and aspect models of time in the process of a figure (realized by a clause or a verbal group)—see Halliday (1987) for these two complementarities, and the modality and evidentiality ‘models’ for assessing information.10 To illustrate how complementary models are represented in a multilingual system network, I will turn to systems of tense and aspect as complementary ways of construing time in the process of a figure realized by a clause, using English and Chinese as examples.11 Halliday and McDonald (2004, p. 380) contrast tense and aspect as follows in their description of temporal categories in Chinese: As is characteristic of languages towards the eastern end of the Eurasian cultural continuum, time in Chinese is grammaticized as aspect rather than as tense: that is, the basic variable is not whether the process is construed as past, present or future relative to the time of speaking or other reference point, but rather whether the process is construed as imperfective or perfective relative to the context. The complementarity of the systems of tense and aspect as models of process time is shown diagrammatically in Figure 6.3. The complementary systems of tense and aspect must be handled at the level of semantics in the modelling of a multilingual meaning potential. The same would be the case in a description of a multilingual meaning potential involving a modality language such as English and an evidentiality language such as Quechua (see, e.g., Weber, 1989); in a system of modality, propositions are assessed in terms of probability—the degree of belief in the information, whereas in a system of evidentiality, propositions are assessed in terms of evidence—the source of the information. Describing the multilingual semantics of time in, e.g., English and Chinese or of epistemic assessment of propositions in, e.g., English and Quechua is a considerable challenge, and I have not attempted it here, leaving it on the agenda for the future. The complementarities just illustrated involve sets of systems like tense and aspect, modality and evidentiality. These complementary systems derive from the same metafunctions—ideational and interpersonal, respectively; and they are thus examples of complementary ways of


Christian M.I.M. Matthiessen

phenomena of experience

process unfolding through time

… construed on the model of location

construal in grammar systems


construed as TENSE: location: unfolding of process located relative to ‘now’ of speaking (or other reference time)



construed as ASPECT: bounding: unfolding of process as bounded / unbounded in time



present future


realizational resources in grammar

Figure 6.3 Tense and aspect as complementary systems of construing process time

construing experience and of enacting assessments. However, within the ideational metafunction, there are two complementary modes of construing experience, the logical mode and the experiential mode (for detailed discussion, see Halliday and Matthiessen, 1999). As we move around the languages of the world, we find that languages vary with respect to whether they construe a given domain of experience, e.g. time, space, the flow of events in terms of the logical model, or the experiential model, or a mixture of the two: languages vary in terms of the division of labour between these two complementary models for construing experience (see Matthiessen, 2004, pp. 574–580). For example, our experience of time in the unfolding of a process may be construed experientially or logically. Where time is construed experientially, it is, in a sense, modelled taxonomically, either as aspect (taxonomies of boundedness, most generally, ‘perfective’ vs ‘imperfective’, with further taxonomic distinctions such as ‘imperfective: durative/

A Multilingual Meaning Potential


progressive’) or as tense (taxonomies of locations relative to now or another reference time, e.g. ‘past’ vs ‘present’ vs ‘future’ [or ‘past’ vs ‘nonpast’], with further taxonomic distinctions such as ‘past: recent/remote’). As far as the tense model is concerned, this is linear time, differentiated into successively more delicate classes of time such as ‘recent’ vs ‘remote’ past. Where time is construed logically, it is modelled as serial time in a system of tense (I don’t think serial aspect models occur), e.g. ‘past’ vs ‘present’ vs ‘future’, and then, e.g., ‘past: past vs present vs future’—i.e. past in relation to now, and then past, present or future relative to that past, and then the possibility of another round of specifications, e.g. past in relation to now, then past in relation to that past, and then future in relation to that past, as in English (she) had been going to renew (the lease since mid March). When a multilingual meaning potential includes complementary experiential and logical models of some domain of experience, such as experiential and logical tense models of the domain of time, we are faced with the same type of challenge as discussed above for complementary systems: while representing the complementary systems within separate languagespecific partitions is straightforward, showing how they can be related to one another means transcending the complementarity so as to be able to bring out some kind of ‘translation equivalences’.

4. Context, Semantics, Lexicogrammar and Registerial Variation In the previous section, I suggested that in the representation of a multilingual meaning potential, commonalities that are not possible to capture in multilingual system networks at the level of lexicogrammar will have to be handled one level up, at the level of semantics, as shown schematically in Figure 6.4 (a). At the level of semantics, it is possible to abstract away from differences in wording to identify similarities in meaning; an early statement of this important insight is Martin’s (1983) text-based investigation of the semantics of ‘participant identification’ in three typologically quite different and genetically unrelated languages, English, Katê (Papuan) and Tagalog (Austronesian). 4.1. Registerial Differences Let me now take one step further and relate semantics to context. When we relate semantics to context, we have to take account of the contextual adaptation of meaning—i.e. of the registerial variation in meaning according to the context of use (see, e.g., Halliday, McIntosh and Strevens, 1964; Hasan, 1973; Halliday, 1978a, 1991; Matthiessen, 1993, 2013a, 2013b; Matthiessen and Teruya, 2016). In a sense, then, variation across languages is not only coming in ‘from below’; it is also coming in


Christian M.I.M. Matthiessen

‘from above’, as it were—variation originating in context. I’ll return to this point in the next section. Thus, even where two or more languages that are part of a multilingual meaning potential share lexicogrammatical systems such as mood, it may be the case that these systems are used by semantics in different ways depending on the nature of the register, as shown schematically in Figure 6.4 (b). This has to be investigated through the analysis (a) Diversified realization

semantics < language I, language II > feature c feature b

system feature d feature e

diversified realization lexicogrammar

feature a


feature m feature n feature o

feature a


feature p feature q feature r

(b) Diversified use of realization semantics

feature b


feature c feature d feature e

feature b


feature f feature g feature h

Diversified use of realization lexicogrammar < language I, language II > feature a


feature m feature n feature o

Figure 6.4 Diversification in the relations between semantics and lexicogrammar in the representation of a multilingual meaning potential

A Multilingual Meaning Potential


of naturally occurring comparable texts in the languages involved. Since the late 1990s, there have been a number of comparative studies in SFL based on comparable texts from comparable registers; they have been concerned with interpersonal systems, in particular speech function and mood—e.g. Teich (1999), Murcia-Bielsa (2000) and Lavid (2000), and, for an overview, see also Matthiessen et al. (2008), with notes on the Cross-Cultural Speech Act Realization Project outside SFL; and with textual ones, in particular theme—e.g. McCabe (1999) (for a recent review of translation studies concerned with textual choices, see Kim and Matthiessen, 2015). Here I will focus on one of the findings from the research referred to above: The interpersonal analysis of registerially comparable texts in different languages has shown that register plays a role in the realization of options within the semantic system of speech function by options within the grammatical system of mood. Teich (2003) draws attention to “cross-linguistic variation that is not predicted by divergence between language systems, but rather by differences in the mapping of semantics and lexico grammar according to register” (p. 58) and gives the following examples from instructional texts in English, French and German: (15) Set the switch to OFF when carrying the radio to prevent turning the power on accidentally. (16) Réglez-la sur OFF quand vous transportez la radio pour éviter de la set-it to OFF when you transport the radio for avoid to it mettre sous tension accidentellement. put under power accidentally (17) Beim Transportieren sollte der Schalter auf OFF gestellt werden, on transport should the switch on OFF put be damit das Gerät nicht versehentlich eingeschaltet wird. so-that the device not accidentally on-switched is And she comments on these examples: “Here, the same interpersonal meaning, a recommendation, is expressed by different lexico grammatical types: by an imperative in English, by a polite imperative in French and by a declarative (plus modal) in German” p. 59). In an instructional situation type—i.e. a situation type where a writer enables his or her readers to carry out some procedure by construing the method, the writer will use the semantic resources of the system of speech function to instruct them. Instructions, recommendations and other related speech functions such as warnings found in instructions are all types of ‘command’—demands for goods-&-services in the system of speech function as described by Halliday (1984). In this register, they are realized by ‘imperative’ clauses in English and French but by ‘modulated’ ‘declarative’ clauses in German


Christian M.I.M. Matthiessen

(modulation: obligation is one of the senses of the modal verb sollen in German). This description captures the important point that while many languages have a basic system of mood type where ‘imperative’ contrasts with ‘indicative’, the meaning of the category of ‘imperative’ varies somewhat from one language to another, and this variation can be illuminated when described in registerial terms (for the use of systemic partitions in the description of registers, see Matthiessen, 1993). Thus, in English, the ‘imperative’ is perfectly fine in the appeal segment of advertisements (realizing an ‘exhortative command’), as in For a free brochure, simply fill in the coupon or phone IBM Direct on 008 622 437; but in Chinese and Japanese, this would be a very unmarked option or even an impossible one: see Matthiessen et al. (2008, pp. 196–197). 4.2. Semantic and Lexicogrammatical Differences Let me now turn to an example of a somewhat more complex situation that can arise in the description of a multilingual meaning potential. Unlike the cases schematized in Figure 6.4 (a) and (b), this case involves both semantic and lexicogrammatical difference within a particular register. In an exploratory study of guidebook texts in a few languages, I found differences between English and German in the way that new places of interest are introduced in the course of walking tours, or, more generally, topographic procedures. The following pair of examples is quite representative: [German] In der Kathedrale selbst kann man die Kapelle Santa Llúcia (Heilige Luzie) bewundern, gegenüber dem Palast von l’Ardiaca (Erzdiakon). Heute noch kann man im Palast des Erzbischofs eine romanische Galerie mit Bogen sehen, die auf Säule des XII. und XIII. Jh. ruhen. “In the Cathedral itself, you can admire the Chapel of Santa Llúcia, opposite the Palace of l’Ardiaca (Archdeacon). Today you can still see, in the Archbishop’s Palace, a Romanesque gallery with arches, which rests on 12th- and 13th-century columns.” [English] In the Cathedral itself, there is the Chapel of Santa Llúcia, opposite the Archdeacon’s Palace, which itself contains a Romanesque gallery with arches supported by 12th- and 13th-century columns. In the German version, the places of interest are introduced as the Phenomenon of a ‘mental: perceptive’ clause with the tourist as Senser. This is a subjective way of introducing the place, from the point of view of the tourist. In contrast, in the English version, the places of interest are

A Multilingual Meaning Potential


introduced without reference to the tourist, first by an ‘existential’ clause and then by a ‘possessive relational’ one. This form of introduction can be characterized as objective precisely because it does not involve the point of view of the tourist walking around on his or her tour. The German analogue of the ‘existential’ clause, a structure with es gibt, cannot be used in a locative sense (see Matthiessen, 2001, and references therein). It is likely that a corpus-based study would bring out probabilistic tendencies, as Lavid’s (2000) study of the realization of requests and questions does. Our description of multilingual meaning potentials must thus include quantitative specifications as well as qualitative ones. This is precisely what one should expect in view of the probabilistic nature of language, theorized by Halliday (e.g. Halliday, 1959, 1993; Nesbitt and Plum, 1988; Matthiessen, 2006, 2015). Probabilistic information will be useful in all uses of multilingual meaning potentials. For example, referring to Halliday’s work on language as a probabilistic system, Toury (2004) draws attention to the importance of probabilities in translation.

5. Conclusion 5.1. The Notion of a Multilingual Meaning Potential Drawing on work on the representation of multilingual system networks we undertook originally in a computational linguistic research context (Bateman et al., 1991; Bateman et al., 1999), I have explored the notion of a multilingual meaning potential. The representation of a multilingual meaning potential includes the multilingual system networks that constitute the subsystems that make up the overall multilingual meaning potential but it also includes specifications of stratification under two main conditions (cf. Figure 6.4): (i) when commonality across the languages making up a multilingual meaning potential cannot be handled within a single stratum—that is, when it is necessary to shunt the task of representing commonality up from lexicogrammar to semantics; and (ii) when commonality across languages in the form of shared lexicogrammatical systems ‘hides’ different modes of deployment across languages—that is, when there are registerial differences in the lexicogrammatical realization of semantic options. The notion of a multilingual meaning potential is a natural development of Halliday’s (e.g. 1973) conception of language as meaning potential. It combines this conception of a language with multilingualism: a multilingual speaker operates with a multilingual meaning potential, and speakers in a multilingual speech community are, as it were, stake holders in a collective multilingual meaning potential, the share of each speaker depending on the nature of the multilingualism in the community and the functional complementarity of the languages involved. The notion of multilingual meaning potential is needed to describe how speakers


Christian M.I.M. Matthiessen

learn additional languages, how they switch between languages in the exchange of meanings or even mix them, and how they reconstrue meanings as they translate or interpret texts. And, as noted at the beginning, multilingualism must have been the unmarked condition for humans for most of our history—with multilingual regions such as Arnhem Land populated by multilingual speakers as a good model, as suggested by Evans (2010); so the notion of the multilingual meaning potential must therefore be part of the core of a general theory of language. 5.2. Stratification of the Multilingual Meaning Potential, and Its Context Like a monolingual meaning potential, a multilingual meaning potential is a system of systems. In stratal terms, it consists of multilingual semantic system networks, multilingual lexicogrammatical system networks, multilingual phonological system networks and multilingual phonetic system networks (or the written or gestural analogues in the expression planes of written languages and sign languages). Similarities and differences among the languages that make up the multilingual meaning potential are specified throughout this stratal hierarchy of subsystems. The multilingual meaning potential is a resource for making meaning in two or more languages. It operates in a context—in a semiotic environment— where the functional complementarity of the languages that make up the multilingual meaning potential is determined, both systemically and instantially, as speakers produce and interpret text unfolding through time. The nature of the functional complementarity is quite complex,12 as has been shown by the extensive literature on multilingualism concerned with external communal and personal factors; it depends on combinations of values within the contextual variables of field, tenor and mode: see Figure 6.5. For example, the different languages that constitute the multilingual meaning potential may have different registerial ranges, as was the case in Britain with English, Norman French and Latin before English morphed into a standard language and expanded to cover all the registers (see, e.g., Halliday, 2003b). 5.3. Descriptive Way in, and Source of Diversification In exploring the multilingual meaning potential, I started ‘from below’ in terms of axis and stratification:13 I started with cases where the languages that constitute a multilingual meaning potential share systems (organization paradigmatic along the paradigmatic axis) and only differ in the structural realization of options in these systems (organization along the syntagmatic axis). Let me return to the example of a multilingual mood system for a moment. The basic systems of ‘indicative’/‘imperative’,

A Multilingual Meaning Potential


Context: Functional Complementarity of languages constituting multilingual meaning potential

context field, tenor and mode factors determining functional complementarity of languages constituting multilingual meaning potential

Languages: multilingual meaning potential

multilingual semantic system network

multilingual lexicogrammatical system network

multilingual phonological system network

multilingual phonetic system network Languages: multilingual meaning

Figure 6.5 The stratal subsystems of a multilingual meaning potential consisting of two or more languages, and its context, where the functional complementarity of the languages constituting the multilingual meaning potential is specified

‘interrogative’/‘declarative’ and ‘elemental’/‘polar’ are found in many languages, but the patterns of realization vary to a certain extent. This is variation ‘from below’, from the point of view of the lower of the two axes—the syntagmatic axis, within the lower of the two strata within the content plane of language—the lexicogrammatical stratum. We associate such ‘low level’ variation with dialect variation (cf., e.g., Halliday et al., 1964, p. 88; Halliday, 1978a, pp. 35, 69, 185)—different ways of saying the same thing; and we can think of the geographical dispersal of languages as an extension of dialect variation14 (which it of course often actually is, as a result of speech communities splitting up as part of


Christian M.I.M. Matthiessen

migrations, as happened with the westward movement of certain IndoEuropean languages and the south-east movement of Bantu languages). The spread of realizational types illustrates the phenomenon of variation ‘from below’, and as I showed in Section 2.1, it is easy to accommodate such realizational differences within the representation of multilingual system networks. In the subsequent sub-sections of Section 2, I showed how our representation of multilingual system networks can also cope with systemic differences in both delicacy and simultaneity; and in Section 3, I went on to demonstrate that the representation can cope even with cases where there are no shared systems—a situation that I suggested arises with complementary systems such as tense and aspect. However, such cases present us with the challenge of specifying how multilingual speakers are able to mediate between systems in different languages that appear to be incommensurate. The general solution is stratal ascent—ascent from lexicogrammatical systems to semantic ones, the assumption being that by moving up to semantics in our multilingual description we are able to abstract away from differences in wording and postulate commonalities in meaning, as, e.g., translators must be able to do. But a great deal of research remains to be done in this area to find out how multilingual speakers operate with such complementary systems so that we can develop descriptions of semantic systems that can cover complementary lexicogrammatical systems such as tense and aspect and modality and evidentiality. Adopting the view ‘from below’, we can thus look at the diversification across languages that has to be accommodated in the modelling of multilingual meaning potentials, and we can ask how far it ascends upwards. There is an important generalization to be made here. In the comparison and representation of two or more languages, we can postulate a general fractal principle in the form of a cline from ‘most similar’ to ‘most different’. The principle is fractal in the sense that it is manifested along a range of semiotic dimensions in the overall ‘architecture’ of language (cf. Matthiessen, 2007). I have discussed and illustrated three of these explicitly—axis (Section 2.1), delicacy (Sections 2.2 and 2.3) and stratification (Section 3), leaving the discussion of rank for another occasion to save space (but see Halliday, 1966b; Catford, 1965; and Matthiessen, 2001, for the hierarchy of rank in translation). Differences among languages around the world—and among the languages that make up a multilingual meaning potential—can thus be interpreted, and even seen as originating, ‘from below’, in terms of the semiotic hierarchies of axis and stratification. At the same time, there is another force driving languages towards diversification—one that can be brought out by shifting the view to one ‘from above’. Seen from above, languages vary in terms of their registerial make-up; that

A Multilingual Meaning Potential


is, any language is an aggregate of registers, and the composition of that aggregate depends on what contexts the language has evolved to operate in. Languages have, of course, evolved as meaning-making resources for many recurrent tasks—contexts that we can recognize from one society to another; but there is also considerable variation in the registerial make-up of languages, as happens, for example, when ‘standard’ languages evolved the registers associated with the institutions of modern nation states, such as the registers of science and bureaucracy (cf. Halliday, 2003b). This type of variation is semantic in the first instance (cf. Hasan, 2009)—not variation in terms of different ways of saying the same thing but rather variation in what is being meant in different languages. It seems to me that we know much less about this kind of variation in the research context of multilingual studies (cf. Matthiessen et al., 2008) because semantic variation is much harder to investigate—this type of research presupposes fairly comprehensive descriptions of the semantic systems of languages embedded in their contexts of culture, and there are very few descriptions of this kind, if any. It is clearly easy enough to fall into the trap of making unsubstantiated sweeping generalizations about languages and the cultures they operate in. The explorations in SFL show very clearly that in order to begin to get at differences in modes of meaning or fashions of speaking across languages, we must work with syndromes of features—or ‘conspiracies’—and resist the temptation of cherry-picking single features (see, e.g., Hasan, 1984; Martin, 1988; Halliday, 1990). 5.4. Towards a Research Agenda My exploration in this chapter of the notion of the multilingual meaning potential is just that—an exploration. Parts of what I have presented have already been shown to work—the framework for representing multilingual system networks: thanks to Bateman’s (1997) KPML system, there is now a considerable body of descriptions of various languages based on multilingual system networks. But building on work of this kind, we can now sketch a research agenda concerned with the theory, description and application of multilingual meaning potentials—i.e. a framework where languages are conceived of and represented as resources for making meaning that are functionally complementary, complementary in the contexts in which persons make meaning. The research agenda includes the tasks of developing comprehensive descriptions of multilingual meaning potentials, of interpreting second language learning as the gradual ‘growth’ of a learner’s multilingual meaning potential (learning how to mean in a second language), of modelling code switching and code mixing as contextually influenced and


Christian M.I.M. Matthiessen

informed choices among options in a multilingual meaning potential, and of modelling translation and interpreting as the recreation of meaning in context by translators and interpreters based on the multilingual meaning potentials that they have mastered.

Notes 1. For a pioneering characterization of embodiment, see Varela, Thompson and Rosch (1991); cf. Thibault (2004). 2. The concern with multilingual issues has been part of the systemic functional traditional from the start, in fact going back to pre-systemic Firthian work: see, e.g., Teruya and Matthiessen (2015); and a number of us have worked towards a general notion of multilingual studies in order to bring together research strands and insights that have tended to be located within separate communities of researchers: see Matthiessen, Teruya and Wu (2008). 3. Or we could call it a polylingual meaning potential to echo Firth’s polysystemic conception of language and the everyday term ‘polyglot’. 4. Just as a meaner is also a person and a biological organism, so is a multilingual meaner: we need to seek to understand multilingual meaners semiotically, but also socially as persons and biologically as organisms (taking account of the body of work on bilingual and now also on multilingual brains—for a recent review of research, see Higby, Jungna and Obler, 2013). 5. I write this within quotes as a caution against the acquisition metaphor that ‘Second Language Acquisition’ is based on. Halliday (1978b/2007, p. 182) drew attention to the problem with this metaphor: “The prevailing metaphor for talking about the learning of the mother tongue in the 1960s was the metaphor of ‘acquisition’, suggesting that language is some type of commodity that the child has to acquire. One shouldn’t make too much of such metaphors; but it is noticeable how much of the work of this period is affected by the notion that language exists independently of people speaking and understanding; that there is an object called a set of rules which constitutes adult language, and it is the task of the child to acquire this ready-made object.” Since he made this observation around 35 years ago, the acquisition metaphor has continued to be widely used, enshrined also in the term ‘Second Language Acquisition’. In her 2011 plenary at the Association Internationale de Linguistique Appliquée (AILA) in Beijing, Diane Larsen-Freeman offered a very extensive and cogent critique of the acquisition model, showing how misleading the metaphor has been in the study and conceptualization of second language development. 6. However, there is a considerable body of work on the ‘bilingual lexicon’, including the question of the representation of lexical information in models of the bilingual lexicon; see, e.g., Pavlenko (2009). It would be interesting to work out the implications of such models in the very different conception of lexis in SFL—lexis as most delicate grammar (see, e.g., Halliday, 1961; Hasan, 1987; Tucker, 1998, 2007). 7. See 8. This is an abridged version of the chapter; the full version with illustrations can be consulted at 9. This point needs to be expanded a great deal particularly since typological studies of lexicalization patterns are not very common and have not been couched in systemic terms; but one can certainly read contributions such as Talmy (1985, 2007) and Viberg (1984) systemically.

A Multilingual Meaning Potential


10. I write ‘models’ in quotes simply because the notion of a model is more of an ideational one than an interpersonal one. 11. It is part of the nature of complementary systems that they are quite hard to sort out in the description of languages precisely because they are complementary: the linguistic complementarity is replayed as metalinguistic complementarity. Thus, as we interpret grammatical systems of time relating to the unfolding of processes, we are faced with the challenge of determining whether these systems are most insightfully described in terms of tense, aspect or a mixture of the two. Not surprisingly, there is a great deal of variation in the literature or—to be a bit more pointed—there is a great deal of confusion. Once aspect had been ‘discovered’ as a descriptive category, it was extended to the description of languages such as English where it really did not belong—introducing a confusion between aspects and relative tenses (cf. Comrie, 1976, 1985). This is not to say that there is no trace of aspect in English—there is; but the so-called perfect and progressive (or continuous) are not aspects but rather relative or secondary tenses. It would take another paper, or a monograph, to sort this out and provide the evidence; but the nature of the English system becomes clear once we confront accounts of it in terms of aspect with descriptions of languages with true aspect systems, as Halliday did based in part on his knowledge of Chinese and Russian— cf. Halliday and Ellis (1951). Other relevant systemic functional accounts include Halliday (1992b); Halliday and Matthiessen (2014), which provides a short account of Halliday’s description of tense in English as a logical system; Halliday and McDonald (2004) on Chinese (from which I have quoted here); and Matthiessen (1983, 1984, 1996). 12. And there are important educational implications, as is brought out in the work by Stephen Evans for the situation in Hong Kong (see, e.g., Evans, 2013). 13. ‘From below’ within the content plane; I have left the expression plane for another occasion. However, multilingual system networks can also be used to represent multilingual sounding potential, and the development of a multilingual phonological system network is an important part of the overall account of a multilingual meaning potential. 14. This is of course a simplification: dialect variation is not only geographical or regional but also relates to other aspects of the provenance of speakers, e.g. social class.

References Bateman, J. A. (1997) Enabling technology for multilingual natural language generation: The KPML development environment. Journal of Natural Language Engineering, 3(1), pp. 15–55. Bateman, J. A., Matthiessen, C. M. I. M., Nanri, K. and Zeng, L. (1991) The rapid prototyping of natural language generation components: An application of functional typology. In Proceedings of the 12th International Conference on Artificial Intelligence. Sydney and San Mateo, CA: Morgan Kaufman, pp. 966–971. Bateman, J. A., Matthiessen, C. M. I. M. and Zeng, L. (1999) Multilingual language generation for multilingual software: A functional linguistic approach. Applied Artificial Intelligence: An International Journal, 13(6), pp. 607–639. Brown, R. and Gilman, A. (1960) The pronouns of power and solidarity. In A. S. Thomas (Ed.), Style in Language. Cambridge, MA: MIT Press, pp. 253–276. Butt, D. G. (1991) Some basic tools in a linguistic approach to personality: A Firthian concept of social process. In F. Christie (Ed.), Literacy in Social Processes:


Christian M.I.M. Matthiessen

Papers From the Inaugural Australian Systemic Functional Linguistics Conference. Darwin: Centre for Studies of Language in Education, Northern Territory University, pp. 23–44. Catford, J. C. (1965) A Linguistic Theory of Translation. London: Oxford University Press. Chui, K. (2012) Cross-linguistic comparison of representations of motion in language and gesture. Gesture, 12(1), pp. 40–61. Comrie, B. (1976) Aspect. Cambridge: Cambridge University Press. Comrie, B. (1985) Tense. Cambridge: Cambridge University Press. Davies, A. (2013) Native Speakers and Native Users: Loss and Gain. Cambridge: Cambridge University Press. Ellis, J. (1987) Some ‘dia-categories’. In R. Steele and T. Threadgold (Eds.), Language Topics: Essays in Honour of Michael Halliday, vol. 2. Amsterdam: Benjamins, pp. 81–94. Evans, N. (2010) Dying Words: Endangered Languages and What They Have to Tell Us. Oxford: Wiley-Blackwell. Evans, S. (2013) The long march of biliteracy and trilingualism: Language policy in Hong Kong education since the handover. Annual Review of Applied Linguistics, 33, pp. 302–324. Firth, J. R. (1950) Personality and language in society. Sociological Review, 42(1), pp. 37–52. Reprinted in Firth, J. R. (1957) Papers in Linguistics 1934–1951. London: Oxford University Press, pp. 177–189. Halliday, M. A. K. (1959) The Language of the Chinese ‘Secret History of the Mongols’. Oxford: Blackwell. Reprinted in Halliday, M. A. K. (2006) Studies in the Chinese Language, vol. 8 in the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, pp. 3–171. Halliday, M. A. K. (1961) Categories of the theory of grammar. Word, 17(3), pp. 242–292. Reprinted in Halliday, M. A. K. (2002) On Grammar, vol. 1 of Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, Chapter 2, pp. 37–94. Halliday, M. A. K. (1966a) Some notes on ‘deep’ grammar. Journal of Linguistics, 2(1), pp. 57–67. Reprinted in Halliday, M. A. K. (2002) On Grammar, vol. 1 in the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, pp. 106–117. Halliday, M. A. K. (1966b) General linguistics and its application to language teaching. In M. A. K. Halliday and A. McIntosh (Eds.), Patterns of Language: Papers in General, Descriptive and Applied Linguistics. London: Longman, pp. 1–41. Halliday, M. A. K. (1973) Explorations in the Functions of Language. London: Edward Arnold. Halliday, M. A. K. (1975) Learning How to Mean: Explorations in the Development of Language. London: Edward Arnold. Halliday, M. A. K. (1977) Text as semantic choice in social contexts. In T. Van Dijk and J. Petöfi (Eds.), Grammars and Descriptions. Berlin: Walter de Gruyter, pp. 176–225. Reprinted in Halliday, M. A. K. (2002) Linguistic Studies of Text and Discourse, vol. 2 in the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, Chapter 2, pp. 23–81. Halliday, M. A. K. (1978a) Language as Social Semiotic: The Social Interpretation of Language and Meaning. London: Edward Arnold.

A Multilingual Meaning Potential


Halliday, M. A. K. (1978b) Is learning a second language like learning a first language all over again? In D. E. Ingram and T. J. Quinne (Eds.), Language Learning in Australian Society. Melbourne: International Press and Publications. Reprinted in Halliday, M. A. K. (2007) Language and Education, vol. 9 in the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, pp. 174–193. Halliday, M. A. K. (1984) Language as code and language as behaviour: A systemicfunctional interpretation of the nature and ontogenesis of dialogue. In M. A. K. Halliday, R. P. Fawcett, S. Lamb and A. Makkai (Eds.), The Semiotics of Language and Culture, vol. 1. London: Frances Pinter, pp. 3–35. Reprinted in Halliday, M. A. K. (2003) On Language and Linguistics, vol. 3 of the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, Chapter 10, pp. 226–250. Halliday, M. A. K. (1985) Systemic background. In J. D. Benson and W. S. Greaves (Eds.), Systemic Perspectives on Discourse. Norwood, NJ: Ablex Publishing Corporation, pp. 1–15. Reprinted in Halliday, M. A. K. (2003) On Language and Linguistics, edited by J. J. Webster. London and New York: Continuum, pp. 185–198. Halliday, M. A. K. (1987) Language and the order of nature. In N. Fabb, D. Attridge, A. Durant and C. MacCabe (Eds.), The Linguistics of Writing: Arguments Between Language and Literature. Manchester: Manchester University Press, pp. 135–154. Reprinted in Halliday, M. A. K. (2003) On Language and Linguistics, vol. 3 of the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, Chapter 5, pp. 116–138. Halliday, M. A. K. (1990) New ways of meaning: A challenge to applied linguistics. Greek Applied Linguistics Association, Journal of Applied Linguistics, 6 (Ninth World Congress of Applied Linguistics Special Issue), pp. 7–36. Reprinted in Halliday, M. A. K. (2003) On Language and Linguistics, vol. 3 of the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, Chapter 6, pp. 139–174. Halliday, M. A. K. (1991) The notion of ‘context’ in language education. In T. Le and M. McCausland (Eds.), Interaction and Development: Proceedings of the International Conference, Vietnam. University of Tasmania: Language Education, pp. 1–26. Reprinted in Halliday, M. A. K. (2007) Language and Education, vol. 9 in the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, pp. 269–290. Halliday, M. A. K. (1992a) The act of meaning. In J. E. Alatis (Ed.), Georgetown University Round Table on Languages and Linguistics 1992: Language, Communication and Social Meaning. Washington, DC: Georgetown University Press, pp. 7–21. Reprinted in Halliday, M. A. K. (2003) On Language and Linguistics, vol. 3 of the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, Chapter 17, pp. 375–389. Halliday, M. A. K. (1992b) The history of a sentence: An essay in social semiotics. In V. Fortunait (Ed.), La cultura italiana e le letterature straniere moderne. Bologna: Longo Editore [for University of Bologna], pp. 29–45. Reprinted in Halliday, M. A. K. (2003) On Language and Linguistics, vol. 3 of the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, Chapter 16, pp. 355–374.


Christian M.I.M. Matthiessen

Halliday, M. A. K. (1993) Quantitative studies and probabilities in grammar. In M. Hoey (Ed.), Data, Description, Discourse: Papers on the English Language in Honour of John McH. Sinclair. London: Harper Collins, pp. 1–25. Reprinted in Halliday, M. A. K. (2005) Computational and Quantitative Studies, vol. 6 in the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, Chapter 7, pp. 130–156. Halliday, M. A. K. (1996) On grammar and grammatics. In R. Hasan, C. Cloran and D. Butt (Eds.), Functional Descriptions: Theory Into Practice. Amsterdam: Benjamins, pp. 1–38. Reprinted in Halliday, M. A. K. (2002) On Grammar, vol. 1 of the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, Chapter 15, pp. 384–417. Halliday, M. A. K. (2002) Computing meanings: Some reflections on past experience and present prospects. In G. Huang and Z. Wang (Eds.), Discourse and Language Functions. Shanghai: Foreign Language Teaching and Research Press, pp. 3–25. Reprinted in Halliday, M. A. K. (2005) Computational and Quantitative Studies, vol. 6 in the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, pp. 239–267. Halliday, M. A. K. (2003a) Introduction: On the ‘architecture’ of human language. In M. A. K. Halliday (Ed.), On Language and Linguistics, vol. 3 of the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, pp. 1–29. Halliday, M. A. K. (2003b) Written language, standard language, global language. World Englishes, 22(4), pp. 405–418. Also in Kachru, B. B., Kachru, Y. and Nelson, C. L. (Eds.). (2006) The Handbook of World Englishes. Oxford: Blackwell, pp. 349–365. Halliday, M. A. K. (2005) On matter and meaning: The two realms of human experience. Linguistics and the Human Sciences, 1(1), pp. 59–82. Halliday, M. A. K. (2008) Complementarities in Language (Halliday Centre Series in Appliable Linguistics). Beijing: The Commercial Press. Halliday, M. A. K. and Ellis, J. O. (1951) Temporal categories in the Modern Chinese verb. Manuscript. Reprinted in Halliday, M. A. K. (2006) Studies in the Chinese Language, vol. 8 in the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, pp. 177–208. Halliday, M. A. K. and Matthiessen, C. M. I. M. (1999) Construing Experience Through Meaning: A Language-based Approach to Cognition. London: Cassell. Reprinted in 2006. London and New York: Continuum. Halliday, M. A. K. and Matthiessen, C. M. I. M. (2014) Halliday’s Introduction to Functional Grammar, 4th ed. London: Routledge. Halliday, M. A. K. and McDonald, E. (2004) Metafunctional profile of the grammar of Chinese. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (Eds.), Language Typology: A Functional Perspective. Amsterdam: Benjamins, pp. 305–396. Halliday, M. A. K., McIntosh, A. and Strevens, P. (1964) The Linguistic Sciences and Language Teaching. London: Longman. Hasan, R. (1973) Code, register and social dialect. In B. Bernstein (Ed.), Class, Codes and Control: Applied Studies Towards a Sociology of Language, vol. 2. London: Routledge and Kegan Paul, pp. 253–292. Hasan, R. (1984) Ways of saying: Ways of meaning. In M. A. K. Halliday, R. P. Fawcett, S. Lamb and A. Makkai (Eds.), The Semiotics of Language and Culture. London: Frances Pinter, pp. 105–162. Reprinted in Hasan, R. (1996)

A Multilingual Meaning Potential


Ways of Saying: Ways of Meaning: Selected Papers of Ruqaiya Hasan, edited by C. Cloran, D. Butt and G. Williams. London: Cassell, pp. 191–242. Hasan, R. (1987) The grammarian’s dream: Lexis as most delicate grammar. In M. A. K. Halliday and R. P. Fawcett (Eds.), New Developments in Systemic Linguistics: Theory and Description. London: Pinter, pp. 184–211. Hasan, R. (2009) Semantic Variation: Meaning in Society and Sociolinguistics, vol. 2 in the Collected Works of Ruqaiya Hasan, edited by Webster. J. London: Equinox. Higby, E., Jungna, K. and Obler, L. K. (2013) Multilingualism and the brain. Annual Review of Applied Linguistics, 33, pp. 68–101. Kim, M. and Matthiessen, C. M. I. M. (2015) Ways to move forward in translation studies: A textual perspective. In J. Munday and M. Zhang (Eds.), Discourse Analysis in Translation Studies. Special Issue of Target, 27(3), pp. 335–350. Lavid, J. (2000) Cross-cultural variation in multilingual instructions: A study of speech act realisation patterns. In E. Ventola (Ed.), Discourse and Community: Doing Functional Linguistics. Tübingen: Gunter Narr, pp. 71–85. Lavid, J., Arús, J. and Zamorano-Mansilla, J. R. (2010) Systemic Functional Grammar of Spanish: A Contrastive Study With English. London and New York: Continuum. Mann, W. C. and Weber, D. J. (1979) Prospectus for Computer-assisted Dialect Adaptation. Notes on Linguistics Special Publication 1. Dallas: Summer Institute of Linguistics. Martin, J. R. (1983) Participant identification in English, Tagalog and Kate. Australian Journal of Linguistics, 3(1), pp. 45–74. Martin, J. R. (1988) Grammatical conspiracies in Tagalog: Family, face and fate— with reference to Benjamin Lee Whorf. In M. J. Cummings, W. S. Greaves and J. D. Benson (Eds.), Linguistics in a Systemic Perspective. Amsterdam: Benjamins, pp. 243–300. Matthiessen, C. M. I. M. (1983) Choosing primary tense in English. Studies in Language, 7(3), pp. 369–430. Matthiessen, C. M. I. M. (1984) Choosing Tense in English. USC/ISI Report: ISI/ RR, pp. 84–143. Matthiessen, C. M. I. M. (1993) Register in the round: Diversity in a unified theory of register analysis. In M. Ghadessy (Ed.), Register Analysis: Theory and Practice. London: Pinter, pp. 221–292. Matthiessen, C. M. I. M. (1995) Lexicogrammatical Cartography: English Systems. Tokyo: International Language Sciences Publishers. Matthiessen, C. M. I. M. (1996) Tense in English seen through systemic-functional theory. In C. Butler, M. Berry, R. Fawcett and G. Huang (Eds.), Meaning and Form: Systemic Functional Interpretations. Norwood, NJ: Ablex Publishing Corporation, pp. 431–498. Matthiessen, C. M. I. M. (2001) The environments of translation. In E. Steiner and C. Yallop (Eds.), Beyond Content: Exploring Translation and Multilingual Text. Berlin: De Gruyter, pp. 41–124. Matthiessen, C. M. I. M. (2004) Descriptive motifs and generalizations. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (Eds.), Language Typology: A Functional Perspective. Amsterdam: Benjamins, pp. 537–673. Matthiessen, C. M. I. M. (2006) Frequency profiles of some basic grammatical systems: An interim report. In S. Hunston and G. Thompson (Eds.), System and Corpus: Exploring Connections. London: Equinox, pp. 103–142.


Christian M.I.M. Matthiessen

Matthiessen, C. M. I. M. (2007) The ‘architecture’ of language according to systemic functional theory: Developments since the 1970s. In R. Hasan, C. M. I. M. Matthiessen and J. Webster (Eds.), Continuing Discourse on Language, vol. 2. London: Equinox, pp. 505–561. Matthiessen, C. M. I. M. (2009a) Meaning in the making: Meaning potential emerging from acts of meaning. Anniversary Issue of Language Learning, 59(Supplement 1), pp. 211–235. Matthiessen, C. M. I. M. (2009b) Multisemiotic and context-based register typology: Registerial variation in the complementarity of semiotic systems. In E. Ventola and A. J. Moya-Guijarro (Eds.), The World Shown and the World Told. Basingstoke: Palgrave Macmillan, pp. 11–38. Matthiessen, C. M. I. M. (2013a) Modelling context and register: The long-term project of registerial cartography. Manuscript of book chapter submitted to L. Barbara and S. Cabral (Eds.), Teoria Sistêmico-Funcional para brasileiros [Systemic Functional Theory for Brazilians]. Santa Maria, Brazil: PPGL: Programa de Pós-Graduação em Letras. Universidade Federal de Santa Maria—UFSM. Matthiessen, C. M. I. M. (2013b) Applying systemic functional linguistics in healthcare contexts. Text and Talk, 33(4–5), pp. 437–466. Matthiessen, C. M. I. M. (2015) Halliday’s probabilistic theory of language. In J. J. Webster (Ed.), The Continuum Companion to M. A. K. Halliday. London and New York: Continuum, pp. 203–241. Matthiessen, C. M. I. M. and Bateman, J. A. (1991) Systemic Linguistics and Text Generation: Experiences From Japanese and English. London: Frances Pinter. Matthiessen, C. M. I. M. and Teruya, K. (2016) Registerial hybridity: Indeterminacy among fields of activity. In D. Miller and P. Bayley (Eds.), Permeable Contexts and Hybrid Discourses. London: Equinox. Matthiessen, C. M. I. M., Teruya, K. and Wu, C. (2008) Multilingual studies as a multi-dimensional space of interconnected language studies. In J. Webster (Ed.), Meaning in Context. London and New York: Continuum, pp. 146–221. Matthiessen, C. M. I. M., Zeng, L., Cross, M., Kobayashi, I, Teruya, K. and Wu, C. (1998) The multex generator and its environment: Application and development. Proceedings of the International Generation Workshop, Niagara-on-theLake, pp. 228–237. May, S. (Ed.). (2013) The Multilingual Turn: Implications for SLA, TESOL, and Bilingual Education. New York: Routledge. McCabe, A. (1999) Theme and Thematic Patterns in Spanish and English History Texts. PhD thesis, Aston University. McNeill, D. and Duncan, S. D. (2000) Growth points in thinking-for-speaking. In D. McNeill (Ed.), Language and Gesture. Cambridge: Cambridge University Press, pp. 141–161. Murcia-Bielsa, S. (2000) The choice of directives expressions in English and Spanish instructions: A semantic network. In E. Ventola (Ed.), Discourse and Community: Doing Functional Linguistics. Tübingen: Gunter Narr Verlag, pp. 117–146. Nesbitt, C. N. and Plum, G. (1988) Probabilities in a systemic grammar: The clause complex in English. In R. P. Fawcett and D. Young (Eds.), New Developments in Systemic Linguistics, Volume 2: Theory and Application. London: Frances Pinter, pp. 6–39.

A Multilingual Meaning Potential


Ortega, L. (2013a) Ways forward for a bi/multilingual turn in SLA. In S. May (Ed.), The Multilingual Turn: Implications for SLA, TESOL, and Bilingual Education. New York and London: Routledge, pp. 32–53. Ortega, L. (2013b) SLA for the 21st century: Disciplinary progress, transdisciplinary relevance, and the bi/multilingual turn. Language Learning, 63(Supplement 1), pp. 1–24. Payne, J. R. (1985) Negation. In T. Shopen (Ed.), Language Typology and Syntactic Description: Clause Structure. Cambridge: Cambridge University Press, pp. 197–243. Pavlenko, A. (2009) Conceptual representation in the bilingual lexicon and second language vocabulary learning. In A. Pavlenko (Ed.), The Bilingual Mental Lexicon: Interdisciplinary Approaches. Bristol, UK: Multilingual Matters, pp. 125–160. Sankoff, D. (1998) A formal production-based explanation of the facts of codeswitching. Bilingualism, Language and Cognition, 1, pp. 39–50. Sankoff, D. and Poplack, S. (1981) A formal grammar for code-switching. Research on Language & Social Interaction, 14(1), pp. 3–45. Saville-Troike, M. (2012) Introducing Second-language Acquisition, 2nd ed. Cambridge: Cambridge University Press. Saville-Troike, M., McClure, E. and Fritz, M. (1984) Communicative tactics in children’s second language acquisition. In F. R. Eckman, K. H. Bell and D. Nelson (Eds.), Universals of Second Language Acquisition. Rowley, MA: Newbury House, pp. 60–71. Talmy, L. (1985) Lexicalisation patterns. In T. Shopen (Ed.), Language Typology and Syntactic Description, Volume 3: Grammatical Categories and the Lexicon. Cambridge: Cambridge University Press, pp. 57–149. Talmy, L. (2007) Lexical typologies. In T. Shopen (Ed.), Language Typology and Syntactic Description, Volume 3: Grammatical Categories and the Lexicon. Cambridge: Cambridge University Press, pp. 66–168. Teich, E. (2003) Cross-linguistic variation in system and text: a methodology for the investigation of translations and comparable texts. Berlin: Walter de Gruyter. Teruya, K., Akerejola, E., Andersen, T. H., Caffarel, A., Lavid, J., Matthiessen, C. M. I. M., Petersen, U. H., Patpong, P. and Smedegaard, F. (2007) Typology of MOOD: A text-based and system-based functional view. In R. Hasan, C.  M.  I.  M. Matthiessen and J. J. Webster (Eds.), Continuing Discourse on Language: A Functional Perspective, vol. 2. London: Equinox, pp. 859–920. Teruya, K. and Matthiessen, C. M. I. M. (2015) Halliday in relation to language comparison and typology. In J. J. Webster (Ed.), The Continuum Companion to M. A. K. Halliday. London and New York: Continuum, pp. 427–452. Thibault, P. J. (2004) Brain, Mind and the Signifying Body: An Ecosocial Semiotic Theory. London and New York: Continuum. Toury, G. (2004) Probabilistic explanations in translation studies: Welcome as they are, would they qualify as universals? In A. Mauranen and P. Kujamäki (Eds.), Translation Universals: Do They Exist? Amsterdam and Philadelphia: Benjamins, pp. 15–32. Tucker, G. H. (1998) The Lexicogrammar of Adjectives: A Systemic Functional Approach to Lexis. London: Cassell. Tucker, G. H. (2007) Between grammar and lexis: Towards a systemic functional account of phraseology. In R. Hasan, C. M. I. M. Matthiessen and J. Webster


Christian M.I.M. Matthiessen

(Eds.), Continuing Discourse on Language: A Functional Perspective. London: Equinox, pp. 953–977. Varela, F. J., Thompson, E. and Rosch, E. (1991) The Embodied Mind: Cognitive Science and Human Experience. Cambridge, MA: The MIT Press. Viberg, Å. (1984) The verbs of perception: A typological study. In B. Butterworth, B. Comrie and Ö. Dahl (Eds.), Explanations for Language Universals. The Hague: Mouton, pp. 123–162. Weber, D. J. (1989) A Grammar of Huallaga (Huánuco) Quechua. Berkeley and Los Angeles: University of California Press. Weber, D. J. and Mann, W. C. (1981) Prospectus for computer-assisted dialect adaptation. American Journal of Computational Linguistics, 7(3), pp. 165–177. Wodak, R., Johnstone, B. and Kerswill, P. (Eds.). (2010) The SAGE Handbook of Sociolinguistics. Los Angeles, London, New Delhi, Singapore and Washington DC: SAGE. Zeng, L. (1996) Planning Text in an Integrated Multilingual Meaning Space. PhD thesis, Sydney University.


Maintenance versus Shift in Literary Translation SFL Perspectives on the Translation of User-Related Varieties in Socio-Cultural Contexts Akila Sellami-Baklouti

1. Introduction The concern of Systemic Functional Linguistics (SFL) with language use has made it suitable for application in translation (Kunz and Teich, 2017, p. 547). Indeed, the functional orientation of SFL has attracted translation scholars, who viewed it as a “new approach” that “provided translation studies with an alternative view which approached language as text” (Hatim and Mason, 1990, p. 36). This chapter aims to show how SFL perspectives can be suitable for application in the investigation of language variation in translation. Two types of perspectives are addressed in the present work. The first concerns the SFL notion of register and its centrality in some translation theories (Hatim and Mason, 1990; House, 2015; Steiner, 2004). This notion is useful in analysing language varieties in the Source Text (henceforth, ST) and correlating them with contextual factors before rendering them in the Target Text (henceforth, TT). The second perspective is concerned with the notion of choice, which is found in translation at two levels: first in the interpretation of the ST, and second in the recreation of these meanings in the TT. These choices can be situated on a maintenance-shift cline (Matthiessen, 2014), offering, thus, a framework for the analysis and evaluation of translation. Choice in translation may also be relevant to the strategies adopted by individual translators, which are, to some extent, determined by the socio-cultural context in which the translation will operate (Matthiessen, 2014). In order to show these theoretical constructs at work, this chapter investigates the translation of dialects, which represents one of the most important challenges to the translator, especially in the case of the literary text, where the quest for equivalence is complicated by the pressure to not lose the aesthetic effects of the ST (Matthiessen, 2014). More specifically, the chapter undertakes a comparative study of Mark Twain’s novel The Adventures of Tom Sawyer and four of its translations: one in Arabic (1963) and three in French (1884, 1904 and 2008). The choice of

122 Akila Sellami-Baklouti this novel is motivated by its purposeful richness in linguistic varieties; the choice of two language pairs (English-Arabic and English-French) is meant to test the impact of the target language status and ideology on the choice of translation strategies; and the choice of different historical periods in the case of the French translations is driven by the hypothesis that even in the case of the same target language, the socio-cultural context in which the translation operates exerts some pressure on the translator’s choices. The chapter starts with an overview of the SFL notions of register (section 2) and choice (section 3) and their appliability in Translation Studies, when the translation of dialects, i.e. user-related varieties, is at stake. The focus will then be narrowed down to the specific function of dialects in literary texts (section 4). After an overview of the four translations under study and the analysis procedure followed (section 5), a more practical discussion of these theoretical notions is based on a comparison of these translations in the way they rendered the dialects. First, dialect markers are linked to contextual features (section 6), and the different strategies adopted in their translation are overviewed (section 7); then, these strategies are discussed in relation to the socio-cultural context of the translations (section 8).

2. Register and Translation This section shows how the SFL notion of register has attracted translation scholars, who gave it an important position in their models of translation description and evaluation. Register, defined as “a variety of language determined by a particular set of values of the context” (Matthiessen, Teruya and Lam, 2010, p. 176), has engendered the recognition of language variation, neglected in structural accounts of language, and also of the importance of the context in this variation, as register is “determined by what the speaker is doing socially” (ibid., p. 176). The impact of contextual features on determining linguistic varieties was highlighted in Translation Studies (henceforth, TS) as early as the 1960s with Catford (1965, p. 84), who defined a language variety as “a sub-set of formal and/or substantial features which correlate with a particular type of socio-situational features”. So, in addition to a ‘common core’ of features that is shared among all the varieties of language, every variety has its features that are peculiar to it, which Catford (1965, p. 86) called “criteria or markers of the variety in question” (original italics). He added that these markers “may be at any level: phonetic, phonological, graphological, grammatical, lexical”. This implies that varieties may be established on the basis of a recurrence of markers which correlate with socio-situational factors; this procedure constitutes the methodological basis adopted in the analysis of linguistic variation in the novel under study. Also, an important implication of this recognition is that a

Maintenance versus Shift


translation, as a particular form of language use, is subject to the impact of context, as is any other form of language use (Kunz and Teich, 2017, p. 549). Language variation was addressed by Hatim and Mason (1990), who made a distinction between two dimensions of this variation. The first is user-related; such varieties are also called ‘dialects’, and they reflect differences from person to person. The second dimension, called ‘register’, is related to “the use to which a user puts language” (pp. 38–9). The focus of the present study is on the first type of variation, as correlations are to be established between the characters’ speech in the novel and their backgrounds before investigating how these varieties were rendered in the TTs. These user-related varieties, or dialects, are further subdivided into different types, based on the source of variation (Hatim and Mason, 1990). Geographical dialects are based on regional lines, temporal dialects reflect language change through time, social dialects reflect social stratification within a speech community, standard/non-standard variation is linked to functional variation, and idiolect is based on the individuality of a text user or idiosyncratic ways of using language. Register also occupies an important position in House’s translation model (1977, 1997, overviewed in House, 2015). It is considered “of prime importance for translation and translation criticism”, because it “captures the relationship between text and context” (House, 2015, p. 64). In the tenor dimension of register, House included “the text producer’s temporal, geographical and social provenance” (ibid., p. 64) as correlating with linguistic features in the ST and TTs. Accordingly, the first step in her model of translation consists of register analysis of the ST, whereby linguistic features (syntactic, lexical and textual) are linked to geographical, social and temporal situational factors. Then, a final Statement of Quality is based on the analysis of ‘mismatches’ identified in the comparison between the ST and the TT. House’s model is of much relevance to the present study both at the level of analysis of the corpus and in the discussion of the findings. Indeed, in the first part of the empirical study, linguistic markers are correlated with situational factors in the ST, and the comparative analysis is aimed at identifying any (mis-)matches between the ST and the TTs. Then, these (mis-)matches are interpreted in light of House’s distinction between an ‘overt’ and a ‘covert’ translation. In an overt translation, “the original is tied in a specific manner to the source language community and its culture”, but is at the same time “also of potential general human interest” (House, 2015, p. 54). House includes literary texts as examples of texts that may call for an ‘overt’ translation because their aesthetic value goes beyond the limits of the time in which they are produced. In a covert translation, however, a ‘cultural filter’ is used “to capture the socio-cultural differences between the ST and TT communities” (p. 66) with the purpose of achieving functional equivalence. This distinction is used to explain the strategies adopted by

124 Akila Sellami-Baklouti the different translators to meet the expectations of the target audiences to whom the translations studied are addressed. Register analysis was also given importance by Steiner (2004), who recognized textual variation as ‘the rule’ rather than an exception. For Steiner, for the practical analysis of a ST and the evaluation of a TT, a ‘Register theory’, “a theory of context of situation and of linguistic variation within such contexts” (Steiner, 2004, p. 44) is needed to connect text and context. According to Steiner, such register analysis would take into consideration “accents/dialects/socialects”, i.e. user-related variation, which is fundamental in translation when it is considered with respect to its function (p. 12). Steiner’s Register theory proves useful in the present study to analyse linguistic variation in the ST in relation to different characters’ social and ethnic backgrounds. Two more notions are borrowed from Steiner’s works for the purposes of this study; the first is the notion of a ‘proper translation’ (Steiner, 1998, p. 295), which requires that the register is kept constant across the process of the translation. This notion will be used to compare and evaluate the different translations of the novel under study. The second notion is that of ‘Goal orientation’, with which Steiner “is interested in participants’ goal, both the author’s and the translator’s, and, thus, on the function of both ST and TT” (Manfredi, 2008, p. 27). Indeed, the author of the ST and the translator act in different socio-cultural contexts and may have different intended readerships; so this difference in Goal orientation will be reflected in choices made by the translator when rendering linguistic variation. This notion of Goal orientation will be used to explain shifts identified in the corpus of TTs. It can be concluded from this overview that the notion of register has been “so successful in translation studies” (Kunz and Teich, 2017, p. 549) because it can be applied at both the theoretical and methodological levels. At the methodological level, Register theory provides a framework for the description of linguistic variation in the ST. Accordingly, the systematic study of dialects is based on establishing a correlation between linguistic features and contextual features. This framework is adopted in the first step of the analysis in the present study. First, dialect markers are identified through their deviations from standard forms and are classified into (a) phonological features, marked in the written text in the form of alternate spellings; (b) grammatical features, deviating from standard structural forms; or (c) lexical features, identified as dialectal lexis. Then, these markers are associated with the characters producing them with the aim of establishing correlations with geographical, temporal, social and idiolectal factors. At the theoretical level, the notion of context, closely related to register analysis, can be extended from the socio-cultural context in which linguistic variation in a text is observed, to the socio-cultural context in which a given translation is made. Accordingly, ‘mismatches’ between the ST and TTs can be explained in light of the difference between the ST author’s and the translator’s Goal orientations. Such a theoretical strand

Maintenance versus Shift


will be used to explain the translators’ opting for different strategies in translations. This leads the discussion to another important SFL notion that has proved useful in TS, which is choice.

3. Translation as Choice This section deals with choice, which constitutes another important SFL notion, in TS and shows how this notion will be applied in the analysis and comparison of the ST and its translations. For the purposes of the present study, ‘choice’ is to be understood, first, in its technical SFL reading as both “an option, a term in a system” and “an act of choice = selection (of an option in a system)” (Matthiessen et al., 2010, p. 69); and, second, in a more general sense of the common word as a choice among translation strategies to render in the TT the dialectal variation in a ST. Although in its technical sense, the meaning of ‘choice’ is “largely governed by the theoretical co-ordinates of the theory”, thus altering “its semantic identity” as a common word (Hasan, 2013, p. 269), in both dimensions choice is not independent of contextual factors. Choice constitutes a defining concept in SFL theory, which perceives a text as “the product of ongoing selection in a very large network of systems” (Halliday and Matthiessen, 2014, p. 23). Drawing a parallel between a text and a translation, as a translation is a particular form of text (cf. Catford, 1965), it can be concluded that choice also occupies an important position in the SFL approach to translation, defined by Matthiessen (2014, p. 272) “as the recreation of meaning in context through choice”. This choice is manifested at two levels: choice in the interpretation of the ST and choice in the creation of the translated text. At the first level, the translator interprets a text by showing the functional organization of its structure and showing what meaningful choices have been made (Halliday and Matthiessen, 2014, p. 24). At the second level, the translator tries to reproduce meanings in the TT through lexico grammatical choices. When the choices made by the translators are examined, various degrees of translation shifts may be found, and these degrees can be situated on a cline extending between translation equivalence and translation shift (Matthiessen, 2014, p. 279). Matthiessen argues that these shifts involve choices within the same metafunction or between different metafunctions. In the present study, Matthiessen’s maintenance-shift cline is used to identify the different degrees of shift at the level of realization of the dialect markers in the TT. In other words, the highest degree of maintenance is when the dialect marker in the ST is rendered in the TT by a dialect marker at the same stratum of realization; that is, phonological markers are rendered by phonological ones, and lexico grammatical markers by lexico grammatical ones, achieving thus formal and functional equivalence. A certain degree of shift is identified when the translator maintains

126 Akila Sellami-Baklouti the dialect marker, but renders it at a different stratum; for example phonological markers rendered by lexico grammatical markers and vice versa, achieving thus only functional equivalence. The notion of shift is also used in this study to describe instances of translation where the translator ‘ignores’ the dialect markers and renders the utterance displaying them in standard language. Maintenance and shift are to be understood as choices made by translators (i.e. the second reading of ‘choice’). Indeed, while dealing with texts displaying linguistic variation, translators may adopt different strategies. They may omit linguistic markers which signal dialectal variation through rendering the ST dialect into a standard variety, which may result in losing the ST effect (Hatim and Mason, 1990, p. 41). They may, alternatively, add linguistic markers to signal contextual meaning associated with dialects (Rosa, 2012, p. 85). In a third scenario, translators may opt for maintaining dialect markers in an attempt to render in the TT the meanings ‘evoked’ (Baker, 2011) by such variation in the ST. Apart from the first possibility, i.e. standardization of dialects, addition and maintenance can be meant to achieve ‘functional’ equivalence: adopting different means to reproduce the marked effect of the dialect use in the ST (Hatim and Mason, 1990). These strategies, however, cannot be fully explained unless the context in which they have been chosen is taken into consideration. In fact, like choices in a text being activated by context (Hasan, 2009), where contextual factors affect lexico grammatical choices, choices in translation are not independent of contextual factors, as they are made in two types of environment: “in the environment of other linguistic choices and also in the environment of the context in which the ST operates and in which the text being translated will operate” (Matthiessen, 2014, pp. 278–279). Putting aside the linguistic environment, this study focuses, first, on the environment of the context in which the ST operates; i.e. it interprets linguistic variation in light of the author’s characterization purposes. Then, the choices made by the different translators to recreate meanings in the TTs will be discussed in light of the context in which the TT is produced. Therefore, the contextual factors relevant to the situation in which the four translations studied have operated are used to explain the choices made by the translators to render linguistic variation in the ST. Such factors may include the acceptability to the target culture of the level of (in-) formality in the ST (Baker, 2011), and the translation’s Goal orientation (Steiner, 2004). This section has highlighted the relevance of the notion of choice in translation: choice of lexico grammar, choice of translation strategy and the contextual factors that may motivate such choices. These choices are investigated in the corpus of translations in how they render the meanings ‘evoked’ by dialectal variation in the ST. Given the literary nature of the

Maintenance versus Shift


ST under study, this dialectal variation can go beyond the spontaneity of casual speech to result in ‘intended’ meanings rather than ‘evoked’ ones.

4. User-Related Varieties in Literary Text When used in literary works, linguistic varieties are rather called ‘literary varieties’, defined as “pseudo-dialects and accents recreated in literature and film” (Rosa, 2012, p. 82). In using the two expressions ‘pseudodialects’ and ‘recreation’, Rosa highlights an important dimension that distinguishes linguistic variation from literary variation, that of spontaneity. While speakers in an authentic communicative situation will use different codes spontaneously (though, depending on the situation, speakers can be more or less careful about their linguistic choices), in a literary work authors carefully work on style, and thus the use of linguistic variation is a conscious authorial choice, as characters in the same novel may be made to speak different varieties to betray their social/ethnic/educational/ regional background (Lindqvist, 1993). When linguistic variation is used in fiction for characterization purposes, several categories may be distinguished (Rosa, 2012, p. 82): “characterising vs non-characterising, narrator vs character diction and groupal vs individual characterising discourse”. These three distinctions can be applied to the novel under study, as is shown in section 6. Discourse is characterizing when it bears markers of contextual features related to the speaker/ character. When characters are made to use a standard variety, which is also used by the narrator, discourse is considered non-characterizing. Within characterizing discourse, different groups may be identified based on a correlation between a systematic use of a linguistic marker (e.g. a given phonological form) and a common ethnic, age or social background. In addition to this groupal belonging, individual characterizing discourse may be used to identify an individual character to distinguish him/her from the rest of characters (e.g. Huck Finn in the novel under study).

5. Corpus and Procedure The empirical study is inspired by Munday’s (2002) systemic model for descriptive TS. The first step consists in the identification and description of dialect markers and their function in the ST. At the second level of analysis, a comparison is made between the ST on the one hand and the four TTs on the other. Table 7.1 provides publication details about the four translations. The choice of two language pairs (EnglishArabic and English-French) is meant to test the impact of the target language status and ideology on the choice of translation strategies; and the choice of different historical periods in the case of the French translations is driven by the hypothesis that the socio-cultural context

128 Akila Sellami-Baklouti Table 7.1 Target Text publication details Translation





Arabic translation


Maher Nassim (revised by Farid Abderrahman)


French translation 1

Fr. 1


French translation 2 French translation 3

Fr. 2

William-L. Hughes (translated with the author’s permission) François de Gaïl

Anglo-Egyptian Library, under the supervision of the Culture Department at the Higher Education Ministry, Cairo Bibliothèque Nouvelle de la Jeunesse, Paris

Fr. 3

Bernard Hoepffner



Mercure de France, Paris Editions Tristram

in which the translation operates exerts pressure on the translator’s choices. However, the choice of a single Arabic translation is due to the fact that this is the only full translation that could be found. The comparison between the ST and the TTs is based on 40 ST utterances that displayed an alternation between standard and non-standard varieties, illustrating, thus, phonological-graphological and lexico grammatical dialect markers. The final step of the analysis discusses translation strategies adopted by the four translators, linking their choices to their socio-cultural contexts.

6. First Step: Dialects and Their Function in the ST The Adventures of Tom Sawyer, published in 1896, deals with the story of a typical mischievous boy who lives in a small Mississippi town, ‘Saint Petersburg’. The novel comprises, in addition to the main character, Tom, a large number of characters having different social and ethnic backgrounds. A large part of the novel is devoted to linguistic interaction between characters. Sellami-Baklouti’s (2005) quantitative study reveals that character diction represents about one third of the novel (23,225 words out of 71,733). The same study also revealed linguistic variation between the narrator’s diction, representing the standard variety, and the characters’ diction, classified into non-characterizing discourse, which is similar to the narrator’s diction, and characterizing discourse, bearing dialect markers. These markers, which are identified through their deviation from the standard variety, are of three types: phonological, grammatical and

Maintenance versus Shift


lexical. Phonological markers are identified through alternative spellings; therefore, the alternation between different orthographic forms of the same word can be seen as an alternation between a standard and a dialectal form; as, for example, in because vs becuz, fellow vs feller, believe vs bleeve and steady vs steedy. Phonological markers are also displayed by contractions (e.g. ’T would, ’t was) and assimilation processes (e.g. gimme instead of give me and lemme instead of let me). The second type of dialect markers is grammatical, and they are identified through their deviation from standard structural forms. Table 7.2 (adapted from Sellami-Baklouti, 2005) illustrates some of these markers. The third type is lexical, and they are identified on the basis of their dictionary entries as colloquial/slang lexis.

Table 7.2 Grammatical dialect markers in the Source Text Features



Absence of agreement between subject and verb

A third person singular subject is used with an noninflected verb A third person singular verb form is used with a non-third person singular subject

- It don’t hurt any more - He come along one day

Use of double negation in the same clause Use of two subjects in the same clause Absence of auxiliaries

Informal inflections of ‘to be’

Inflections of ‘to have’

The progressive auxiliary The perfective auxiliary The use of ‘ain’t’ as the present negative form of ‘to be’ The use of ‘warn’t’ as the past negative form of ‘to be’ The use of ‘hain’t’ as the present negative form of ‘to have’

- Why, you two was scuffling - when all the boys is having holiday - I reckon there ain’t no mistake ’bout where I’ll go - Well, I don’ know no kings, Tom - “Well, Tom Sawyer he licked me once.” - Then Mrs. Harper she began to cry - Oh, won’t it be fun! You going to have all the girls and boys? - I been to the circus three or four times - We ain’t doing any harm - Why, that ain’t a-going to do any good - You’d be always into that sugar if I warn’t watching you - He warn’t any more responsible than a colt but then he hain’t ever done anything to hurt anybody

130 Akila Sellami-Baklouti These dialect markers in the novel display a variation between the narrator and characters (example 1) and between different characters (example 2): (1) Narrator: The public are not slow in the matter. p. 120 Tom: They ain’t any such men now, I can tell you. p. 253 (2) Mr. Jones, the Welshman: I hope you’re good and hungry, because breakfast will be ready as soon as the sun’s up. p. 289 Huck Finn: I’ve come now becuz I wanted to know about it. p. 290 The study of this variation revealed a correlation between characters’ diction and contextual factors (Sellami-Baklouti, 2005). Social class constitutes a factor of variation, as the characters belonging to a higher social class (e.g. Judge Thatcher and Mr. Jones, the Welshman) are made to speak a standard variety that is similar to the narrator’s diction, i.e. non-characterizing discourse, while dialectal features are used by other characters, representing a lower social class. The age of characters represents another source of variation, as Tom and his friends are made to use peculiar dialect markers which are not found in the speech of the adult characters. Within this groupal variety, the utterances by Huck Finn, the town outcast, displayed some idiolectical markers, intended to single him out from the rest of the characters. Finally, ethnicity constitutes another source of variation in the novel, as Jim, the Black African slave, and Injun Joe are made to use dialect markers representative of the African American and the Native American varieties, respectively.

7. Second Step: Maintenance vs. Shift in the Corpus of Translations The analysis of the 40 utterances displaying dialect markers and their translations shows a variety of choices, implying different strategies adopted by the translators. While the following section will discuss the strategies preferred by each translator, the present section illustrates these strategies and discusses their position on the maintenance-shift cline. One of the strategies adopted by the translators is the omission of either an individual word bearing the dialect marker (example 3) or the whole utterance displaying the presence of such markers (example 4): (3) ST: Tom: ‘Well, I reckon it’s so, then. Becuz they say she’s a witch’. p. 71. Fr. 2: Je comprends tout, maintenant ! Ø On dit que c’est une sorcière ! p. 140.

Maintenance versus Shift


(4) ST: Huck: ‘Oh, Tom, I reckon we’re goners. I reckon there ain’t no mistake ’bout where I’ll go to. I been so wicked.’ p. 114. Fr. 1: The whole utterance has been skipped by the translator. In example 3, the translator has opted for omitting the word displaying the phonological dialect marker. As this word conveys a causal relation, this omission presents a shift at the logical metafunction, making implicit a causal link that is explicit in the ST. On the other hand, the omission of the whole content conveyed by the utterance in example 4 results in a shift at the ideational level. Shift in the corpus also results from the omission of the dialect marker through its neutralization and the choice of a standard form, as can be seen in examples 5 and 6: (5) ST: Tom: ‘Well, I reckon it’s so, then. Becuz they say she’s a witch’. p. 71. Fr. 1: Elle doit s’y connaitre; car on dit qu’elle est sorcière. p. 46. (6) ST: Jim shook his head and said: ‘Can’t, Mars Tom. Ole missis, she tole me I got to go an’ git dis water an’ not stop foolin’ roun’ wid anybody.’ p. 17. Ar. wa hazza jim ra?sa-hu salban wa qa:la: la: ?astaTi:Gu ?ayuha ?asayyidu and shook Jim head-his negatively and said: not can vocative master tom faqad Gahidat ?ilayya sayyidat-i ?alGaju:zu ?an ?a-dhhaba wa Tom as asked to-me mistress-my old that I-go and ?uHDHira ?al-ma:?a wa Hadharat-ni: mina ?atasakuGi ?aw ?attaHaduthi maGa Bring the-water and warned-me from hanging out or chatting with ?aHadin. p. 21. anybody

‫ ﻻ ﺃﺳﺘﻄﻴﻊ ﺃﻳﻬﺎ ﺍﻟﺴﻴﺪ ﺗﻮﻡ ﻓﻘﺪ ﻋﻬﺪﺕ ﺇﻟﻲ ﺳﻴﺪﺗﻲ ﺍﻟﻌﺠﻮﺯ ﺃﻥ ﺃﺫﻫﺐ‬:‫ﻭﻫﺰ ﺟﻴﻢ ﺭﺃﺳﻪ ﺳﻠﺒﺎ ﻭﻗﺎﻝ‬ ‫ﻭﺃﺣﻀﺮ ﺍﻟﻤﺎء ﻭﺣﺬﺭﺗﻨﻲ ﻣﻦ ﺍﻟﺘﺴﻜﻊ ﺃﻭ ﺍﻟﺘﺤﺪﺙ ﻣﻊ ﺃﺣﺪ‬ While, in example 5, neutralization is at the level of the word, this process of standardization is more clearly seen in example 6 from the Arabic text, where all the dialect markers representative of the Black American variety in the 19th century are rendered by Modern Standard Arabic, making Jim’s diction similar to the narrator’s diction and neutralizing any difference between this character, being representative of African Americans in the novel, and all the other characters.

132 Akila Sellami-Baklouti The third possible strategy marking a shift in the translation that is less frequently adopted consists in addition, as can be seen in example 7, where the translator adds an interjection, a characteristic feature of spoken language, to compensate for the phonological dialect markers present in the ST utterance: (7) ST: ‘I never thought of that. That’s so. Lemme go with you?’ p. 72. Fr. 1: Je ne pensais pas à ça. Laisse-moi aller avec toi, hein? p. 47. Moving on the cline towards maintenance, the translators may opt for rendering the dialect marker at a different stratum from the ST. The following shifts have been identified: (i) a phonological marker is rendered by a grammatical marker (example 8); (ii) a phonological marker is rendered by a lexical marker (example 9), (iii) a grammatical marker is rendered by a phonological marker in addition to the grammatical marker (example 10), and (iv) a grammatical marker is rendered by a lexical marker (example 11): (8) ST: ‘I never thought of that. That’s so. Lemme go with you?’ p. 72. (assimilation) Fr. 3: J’ai jamais pensé à ça. C’est vrai. Je peux venir avec toi? p. 67. (absence of inversion in an interrogative utterance) (9) ST: —’twould help a good deal. p. 291. (contraction) Fr. 3: ça serait sacrément utile. p. 250. (informal lexis) (10) ST: ‘Was you ever at a circus?’ said Tom. p. 83. (absence of agreement between subject and verb) Fr. 3: t’es déjà allée au cirque? p. 77. (contraction) (11) ST: ‘All right. But I bet you I ain’t going to throw off on di’monds. p. 243. Fr. 3: D’accord. Mais je te parie que tu ne vas pas bouder les diamants. p. 211. (informal lexis) At the maintenance end of the cline, translators remain at the same stratum while rendering a dialect marker in the TT; accordingly, the weak vowels in the ST are rendered by elision of a consonant (example 12), the absence of an auxiliary in the ST is rendered by the absence of inversion in an interrogative form (example 13), and lexical markers are rendered by lexical ones (example 14): (12) ST: ‘Well, I reckon it’s so, then. Becuz they say she’s a witch.’ p. 71.

Maintenance versus Shift


Fr. 3: ‘Eh bien, ça doit être vrai, alors, Pasqu’on dit que c’est une sorcière.’ p. 66. (13) ‘Say, Hucky, when you going to try the cat?’ p. 71. Fr. 3. Hucky, tu vas l’envoyer quand ton chat? p. 67. (14) ST: ‘I tell you better. What’s the name of the gal?’ p. 246. Fr. 3: ‘Je te dis qu’y vaudrait mieux. Comment elle s’appelle, la poule?’ p. 213. These five strategies adopted by the translators in rendering dialect markers have different positions on the shift-maintenance cline, as displayed by Figure 7.1. Omission is considered as the highest degree of shift because what is lost in the TT is not only the meaning evoked by the dialect marker, as is the case with standardization, but also the ideational dimension of the utterance. The range of meanings lost will depend on whether what is omitted is a single word, as in example 3, where the loss is not serious because the logical meaning can be reconstructed by the reader even if it is not explicitly expressed through the dialectal form of the conjunction used in the original text. However, we can speak about a more serious loss of meaning when the omission concerns a whole utterance as in example 4, where the omitted utterance offers insights into the psyche of the character. Addition, however, implies that the translator is aware of the effect of dialect markers in the ST, but for some reason—which will be discussed in the next section—cannot render them in the TT through dialect markers, so they resort to compensatory choices, as was shown in example 7. The highest degree of maintenance is depicted in the corpus when the translator renders the dialect marker at the same stratum, achieving thus both formal and functional equivalence. The choice of a dialect marker at a different stratum may be due to the fact that the translator thinks that it is more appropriate choice in the context of the TT language and culture. This takes the analysis a step further with a view to explaining the choice of these strategies in their socio-cultural contexts. While this section overviewed all the possible strategies, some systematic patterns have been depicted in the corpus, marking a preference for certain strategies by different translators. The following section discusses these patterns, trying to explain them in light of the socio-cultural context in which the translation was produced and the translator’s ideology.




Dialect marker at different stratum

Dialect marker at same stratum

Figure 7.1 Translation strategies on a shift-maintenance cline


134 Akila Sellami-Baklouti

8. Translators’ Choices and Contextual Factors This section mainly rests on the second SFL perspective on translation that is addressed in this chapter, namely that choices in translation are affected/shaped by the environment of the TT language and culture. We noticed preferred choices in the four translations under study, and the following sub-sections situate each of the studied translations in its sociocultural context.

8.1. The Arabic Translation The major strategy adopted in the Arabic translation under study is the neutralization of linguistic variation as the translator renders in Modern Standard Arabic all characters’ diction, neutralizing any difference between narrator and characters, on the one hand, and between different characters, on the other, removing thus from the TT the characterizing function of pseudo-dialects in the ST. Although this strategy leads to “losing the ST intended effect” (Hatim and Mason, 1990, p. 41), its choice may be better explained when taking into account contextual factors. One important factor is the institution behind this translation and its motivation for it. Indeed, the selection of the novel by the Culture Department at the Ministry of Education— as stated in the preface to the translation—implies that this translation was motivated by educational purposes, and this explains the choice of Modern Standard Arabic, which is the High Variety used as the medium of instruction in schools and opposed to Low Varieties, having mainly a spoken tradition and used for less prestigious functions (Ferguson, 1959). In addition, the use of Modern Standard Arabic guarantees a more widespread readership because the book can be understood by readers in the whole Arab world, which would not be the case if an Egyptian dialect or mixture of them had been selected to render the variation in the ST. The neutralization of pseudo-dialects can also be explained by the difference between the ST and TT in their Goal orientations (Steiner, 2004, p. 16). This difference is seen in the prefaces of the ST and the TT. In his preface to the novel, Twain started with the realistic aspect of the story and its characters and stressed the cultural dimension by stating that “the odd superstitions touched upon were all prevalent among children and slaves in the West at the period of this story” (Twain, 1876, p. 2), but he also hinted at a more general human interest by stating his plan “to try to pleasantly remind adults of what they once were themselves, and of how they felt and thought and talked”. It is this general human interest that was highlighted by the translator, who viewed the book as “the story of every boy beset by adolescence” (p. 7), and accordingly focused on the universality of the story, explicitly stating his view of the novel “as a big stage where characters may belong to any society/community

Maintenance versus Shift


be it eastern or western” (p. 8). Therefore, Nassim’s translation can be considered an ‘overt’ translation (House, 2015) because it focused more on the human ‘timeless’ nature of the literary text. Indeed, in translating characters’ diction, the translator neglected linguistic variation evoking characters’ regional and social belonging and rather focused on their human dimension because for him “the feelings depicted in this novel are the same that characterize life everywhere and every time” (p. 8). 8.2. Hugh’s Translation (1884) Apart from few instances where the translator rendered the dialect marker by adding a spoken discourse marker, as in example 7, the two major strategies adopted by Hugh are neutralization and omission of utterances containing dialect markers. The adoption of these strategies can mainly be explained by the motivation behind this translation and the audience to whom it was addressed. Like the Arabic one, this French translation was addressed to children, as it was published by Bibliothèque nouvelle de la jeunesse, and the editor, A. Hennuyer, collaborated with the Ministry of Education (Jenn, 2004, p. 238). According to Jenn (2004), the end of the 19th century and the beginning of the 20th century were characterized by a flourishing children’s book market that was enhanced by the government support of mass schooling. In this context, the goal of the translator was oriented more towards providing a book of adventures to children than depicting aspects of Western life through linguistic variation. 8.3. De Gail’s Translation (1904) De Gail’s translation displays more endeavour than Hugh’s in seeking functional equivalence through rendering the ST dialect markers. This process is characterized by two features: the first is the preference for lexical markers, even in cases where a phonological marker was used in the ST, displaying thus a shift in the language stratum; for example, the variation feller vs fellow was rendered by type vs individu. The second feature characterizing functional equivalence in this translation consists in keeping a certain degree of formality through the refinement of colloquialisms; for example, the variation gal vs girl was rendered by fille vs demoiselle. It can, therefore, be concluded that the attempt at seeking functional equivalence was limited by the fact that the book was addressed to children, as Jenn (2004) notes that at the beginning of the 20th century, The Adventures of Tom Sawyer used to be given as a prize to the best pupils. Hence, this preference for refined lexical markers may have been targeted to purify the text of vulgar words and ‘ungrammatical’ structures before making it accessible to children. This, in turn, can also be explained by a puristic linguistic ideology at the beginning of

136 Akila Sellami-Baklouti the 20th century, showing a lack of tolerance towards dialects in written literature, especially if we take into account that this translation was published by Mercure de France, which is “a prestigious and high-brow house catering to middle-class reader” (Jenn, 2006, p. 246).

8.4. Hoepffner’s Translation (2008) This translation displays more functional equivalence than the others, as the translator has tried to render in the TT all the dialect markers of the ST. Functional equivalence was in some instances achieved through a shift at the stratum level, where, for example, phonological dialect markers were rendered by lexical or grammatical markers, but in most examples, the translator maintained the dialect marker at the same stratum, achieving thus formal and functional equivalence; for example, Maît was used to translate Mars, and pasque to translate becuz. More noticeable in the translation is the fact that the translator used a dialect marker in the TT even in cases where the utterance in the ST has none, as shown in, for example, 15: (15) ST: ‘it sounds splendid. TT: ‘ça a de la gueule’. The effect of this endeavour to achieve functional equivalence is clear in the translation, as the translator managed to reproduce characterizing discourse, making a difference between the narrator and the characters, on the one hand, and between different characters, on the other, for example Huck Finn, the town’s outcast, is the only character who uses the form pasque, as opposed to all the other characters. The translator also managed to render the connotations conveyed by lexical distinctions in the ST; for example, poule was used to translate gal (both words are defined as having negative connotations in their respective languages). Hoepffner’s attempts at achieving functional equivalence may be explained by his awareness of the weaknesses of the previous translations in failing to achieve it. This awareness can be implied by the sub-title “New Translation” (“Nouvelle traduction”) on the title page. Also, the publisher’s note on the back cover of the translation, introducing Twain as one of the “pioneers of ‘spontaneous’ writing [who] introduced spoken language in writing”, implies an awareness of the importance of the linguistic variation in the ST and sets its rendering as an objective of the ‘new’ translation, which is the first according to the translation’s blurb “to render the incomparable taste and energy of the ST”. Accordingly, Hoepffner’s translation can be considered as the most ‘proper’ (Steiner, 1998) because the register has remained constant, with the same degree of (in-)formality. Thus, the distinctions between the narrator’s and the

Maintenance versus Shift


characters’ diction, and between characterizing and non-characterizing discourse, present in the ST have been appropriately rendered in the TT. This shows Hoepffner’s awareness of these evoked meanings, his endeavour to render them in the TT and his belief that “the burden of the translator” (Hoepffner, 2006) is to be faithful to the author’s language, even if this means they challenge the target language’s so-called purity.

9. Conclusion This chapter has argued for the relevance of the SFL notions of register and choice to the description and analysis of user-related linguistic variation in a literary text. This theoretical argument is supported by an empirical study investigating how the dialect markers in Twain’s The Adventures of Tom Sawyer are rendered by four translations representing different languages and different socio-cultural contexts. Combining ‘indicative’ and ‘imperative’ SFL perspectives on translation (Wang and Ma, this volume), the study started with a description of linguistic variation in the ST and its translation; and then shifted to the evaluation of the quality of the four translations, based on the way this variation is rendered. First, different degrees on the maintenance-shift cline (Matthiessen, 2014) were identified, and the strategies adopted by the different translators were explained in their respective contexts. Then, the comparative investigation of the four translations showed that a more ‘proper translation’ is produced when the translator is aware of the register variation of the ST and renders it in the TT, even if this challenges the target language ideology. The findings of the present study lead to the conclusion that the SFL constructs of register and choice—highlighted in this chapter—provide both a theoretical background and an analytical framework for the study of the translation of linguistic variation. Although the focus of the study is on SFL perspectives on TS, the findings also suggest that TS can have theoretical implications for SFL. Indeed, through showing the impact of the socio-cultural contexts on translators’ choices, the present study provides a further argument that variation between languages can also originate in the context, hence the need for a “contextual adaptation of meaning” (Matthiessen, this volume), which further highlights the ‘place’ of context as the highest stratum in linguistic investigation (Hasan, 2009).

Primary Works Source Text Twain, M. (1876) The Adventures of Tom Sawyer. Planet PDF. Available at: www. (Accessed on 29 September 2014).

138 Akila Sellami-Baklouti Translations De Gail, F. (1904) Les aventures de Tom Sawyer. Paris: Mercure de France. Hoepffner, B. (2008) Les aventures de Tom Sawyer. Auch: Tristam. Hugh, W. L. (1884) Les aventures de Tom Sawyer. Paris: Bibliothèque Nouvelle de la Jeunesse. Nassim, M. (1963) ‫ ﻣﻐﺎﻣﺮﺍﺕ ﺗﻮﻡ ﺳﻮﻳﺮ‬MuGHamaratu tu:m sawyer. Cairo: AngloEgyptian Library.

References Baker, M. (2011) In Other Words: A Coursebook on Translation. London and New York: Routledge. Catford, J. C. (1965) A Linguistic Theory of Translation. Oxford: Oxford University Press. Ferguson, C. A. (1959) Diaglossia. Word, 15, pp. 325–340. Halliday M. A. K. (1992) Language theory and translation practice, Rivista internazionale di tecnica della traduzione, n. 0, pp. 15–25. Halliday, M. and Matthiessen, C. M. I. M. (2014) Halliday’s Introduction to Functional Grammar. UK: Routledge. Hasan, R. (2009) The place of context in a systemic functional model. In M. A. K. Halliday and J. J. Webster (Eds.), Continuum Companion to Systemic Functional Linguistics. London and New York: Continuum, pp. 166–189. Hasan, R. (2013) Choice, system, realisation: Describing language as meaning potential. In L. Fontaine, T. Bartlett and G. O’Grady (Eds.), Systemic Functional Linguistics: Exploring Choice. Cambridge: Cambridge University Press, pp. 269–299. Hatim, B. and Mason, I. (1990) Discourse and the Translator. London and New York: Longman. Hoepffner, B. (2006) The burden of the translator. In M. Gonzalez and F. Tolron (Ed.), Translating Identity and the Identity of Translation. Buckinghamshire: Cambridge Scholars Press, pp. 193–198. House, J. (2015) Translation Quality Assessment: Past and Present. Oxon and New York: Routledge. Jenn, R. (2004) Les Aventures de Tom Sawyer: Traductions et adaptations. Palimpsestes: De la lettre à l’esprit: Traduction ou adaptation, 16, pp. 137–150. Jenn, R. (2006) From American frontier to European borders: publishing French franslations of Mark Twain’s novels “Tom Sawyer” and “Huckleberry Finn” (1884–1963). Book History, Vol. 9, pp. 235–260. Kunz, K. and Teich, E. (2017) Translation studies. In T. Bartlett and G. O’Grady (Eds.), The Routledge Handbook of Systemic Functional Linguistics. Oxon and New York: Routledge, pp. 547–560. Lindqvist, Y. (1993) Spoken language in literary prose—A translation problem. Translation and the Manipulation of Discourse, Selected papers of the CERA Research Seminars in Translation Studies 1992–1993, CETRA, Leuven. Manfredi, M. (2008) Translating Text and Context: Translation Studies and Systemic Functional Linguistics. Volume 2: From Theory to Practice, 2nd ed. Bologna: Cento di Studi Linguistico-Culturali.

Maintenance versus Shift


Matthiessen, C. M. I. M. (2014) Choice in translation: Metafunctional considerations. In K. Kunz, E. Teich, S. Hansen-Schirra, S. Neumann and P. Daut (Eds.), Caught in the Middle—Language Use and Translation: A Festschrift for Erich Steiner on the Occasion of His 60th Birthday. Saarland: Saarland University Press, pp. 271–333. Matthiessen, C. M. I. M., Teruya, K. and Lam, M. (2010) Key Terms in Systemic Functional Linguistics. London and New York: Continuum. Munday, J. (2002) Systems in translation: A systemic model for descriptive translation studies. In T. Hermans (Ed.), Crosscultural Transgressions. Research Models in Translation Studies II: Historical and Ideological Issues. Manchester: St Jerome, pp. 76–92. Rosa, A. A. (2012) Translating place: Linguistic variation in translation. Word and Text, II(2), pp. 75–97. Sellami-Baklouti, A. (2005) The ‘clue-bearing’ function of language: The politics of dialect in twain’s The Adventures of Tom Sawyer. Journal of Language and Literature, 3(1), pp. 40–67. Steiner, E. (1998) A register-based translation evaluation: An advertisement as a case in point. Target, 10(2), pp. 291–318. Steiner, E. (2004) Translated Texts: Properties, Variants, Evaluations. Frankfurt am Main: Peter Lang.


Textual and Logical Choices in the Dramatic Monologue of Teahouse and Its English Translations Wang Bo and Ma Yuanyi

1. Introduction As an appliable theory, the applicability of Systemic Functional Linguistics (hereafter SFL; e.g. Halliday, 1985) has been widely recognised in the field of translation studies (cf. Steiner, 2015). Within the SFL framework, a text can be described by the different metafunctional modes of meaning: the ideational (experiential and logical), the interpersonal and the textual. Thus, translation involves the recreation of all four modes of meaning (Matthiessen, 2001, 2014), and can be understood as “a process of simultaneous choices among options in logical, experiential, interpersonal and textual systems” (Matthiessen, 2014, p. 277). The application of SFL to translation studies can be traced back to Halliday’s (1956, 1962) early engagement with machine translation. Then the overall application continued with Catford (1965), which has been generally regarded as one of the first translation studies from an SFL perspective, and where Halliday’s (1961) scale and category grammar, an early model of SFL, is applied. In Catford’s (1965) monograph, translation equivalence and shift are examined from the perspectives of stratification and units in structures. In accordance with Halliday’s (2001) association of translation studies with the mood system, we can say that translation studies from the SFL perspective are indicative rather than imperative. In other words, a linguist’s translation theory is descriptive instead of prescriptive (cf. Toury, 1995), which concerns the nature of the translation process and the relationship between the source text and the target text. There have been a number of such descriptive studies, which draw on different aspects of SFL. For instance, Steiner and his group (e.g. Teich, 2003; Steiner, 2004; Hansen-Schirra, Neumann and Steiner, 2012) focus on register and corpus methodology. They provide methodologies for investigating comparable texts to identify systemic variations between languages. They also fill in gaps in studying the process of explicitation in translation, i.e. translated texts that are more explicit than the original (e.g. Steiner, 2008). Hatim and Mason (1990), on the other hand, contextualise translation from the perspectives of cohesion, coherence, register, genre, etc.

Textual and Logical Choices


There are also studies that combine indicative (descriptive) with imperative (prescriptive) approaches by taking translation evaluation into consideration. House (1977, 1997), for example, analyses the lexicogrammatical and registerial features of texts, and suggests two translation strategies, i.e. overt translation and covert translation, with regard to the cultural phenomena that she terms ‘the cultural filter’. In this chapter, we adopt a descriptive approach to the analysis of dramatic monologue in a playscript. The reason for focusing on drama in the present study is that the translation of drama has seldom been investigated compared to other literary genres, even in some of the monumental works by translation theorists. For example, two of the most influential studies from a linguistic perspective, Nida (1964) and Hatim and Mason (1990), completely neglected drama translation. The overall purpose of the study is to apply SFL to study the translations of Teahouse, which is an influential text that has not yet received enough attention from the SFL perspective, and to examine the changes and trade-offs made in the translations in systemic functional terms. In addition, the study also attempts to provide an alternative perspective on studying the textual and logical choices in translation by analysing choices in systems of theme, taxis and logico-semantic type. Following Matthiessen (2014), the different kinds of metafunctional translation shifts will be identified and discussed. The correlation between the lexicogrammatical choices and the context will be explored by way of analysing the contextual parameters.

2. Data: Dramatic Monologue in Teahouse Teahouse (the source text, hereafter ST) was written by a renowned Chinese writer, Lao She. There are two English versions of this play, which was translated by John Howard-Gibbon (target text 1, hereafter TT1) and Ying Ruocheng (target text 2, hereafter TT2). Teahouse is a monumental work in the history of Chinese literature. The two translators have adopted different strategies, thus providing two completely different translations. The ST and the two TTs have both been studied from various perspectives, including functional equivalence, the cooperative principle, skopos, etc. (see Ren, 2008), leading to different findings, but have never been investigated by way of an SFL analysis. Based on the registerial cartography developed by Matthiessen (e.g. Matthiessen, 2015; Matthiessen and Teruya, 2016), drama falls into the category of recreating, which is a semiotic process that recreates the other seven fields of activity by way of narration and/or dramatisation. As widely acknowledged, when translating different types of texts, translators need to adopt different strategies (cf. Reiss, 1971). Therefore, it is significant to situate drama within the map of a functional text typology, with which the discussion of the translators’ choices will be related. In

142 Wang Bo and Ma Yuanyi the coming sections, we will note that the field of activity being recreated in dramatic monologue is exploring, by which the character expresses his stances and positions through chanting. To recreate the field of exploring, differences of tenor and mode between the two translators have led to the variations in their lexicogrammatical choices. A description of the contextual parameters will be provided in Section 4.3. In the script of Teahouse, three kinds of text are identified, namely stage directions, dramatic monologue and dramatic dialogue. Dramatic monologue, different from the dialogues frequently found in drama, refers to a speech delivered by one speaker, with an extended length and internal coherence. In Teahouse, all dramatic monologues are addressed to the audience by a character named Shǎ Yáng (translated as Oddball Young and Silly Young in the two translations), who is a beggar who earns a living by chanting rhythmic storytelling. The complete dramatic monologue in Teahouse is selected for our analysis. All the lines by Silly Young are divided into clause complexes and ranking clauses. The number of ranking clauses in ST, TT1, and TT2, i.e. the volume of text to be analysed, is 155, 138 and 140 respectively.

3. Analytical Framework Based on a pilot study, we found that the two translators made more different choices in the textual and logical systems than in the interpersonal and experiential ones. Therefore, the following analytical framework is constructed to compare the textual and logical choices in the two TTs with those in the ST (see Figure 8.1). The systems involved in the lexicogrammatical analysis include theme, taxis and logico-semantictype. After the lexicogrammatical analysis, the three contextual parameters, i.e. field, tenor and mode, will be considered in the contextual analysis to give further accounts of the different

Figure 8.1 Theoretical framework of the present study

Textual and Logical Choices


choices of the two translators, because variations in contexts are likely to have influenced the two translators in making their choices, and will thus have led to the differences found in the lexicogrammatical analysis (cf. Thompson, 1999).

4. Analysis and Discussion 4.1. Analysis of Theme According to Halliday and Matthiessen (2014, p. 89), Theme “serves as the point of departure of message; it is that which locates and orients the clause within its context”. Based on previous typological descriptions of Chinese and English, we know that three types of Theme can be identified in both languages, viz. textual Theme, interpersonal Theme and topical Theme (cf. Halliday and Matthiessen, 2014; Halliday and McDonald, 2004; Li, 2007). Figure 8.2 sketches the frequencies of different types of Theme in the data. In terms of textual Themes, two are found in the ST, and forty-nine and forty-one are found in TT1 and TT2 respectively. Conjunctions, continuatives and conjunctive Adjuncts can all function as textual Themes. According to Table 8.1, most textual Themes in the TTs are conjunctions, such as “and”, “but”, “if”, “when”, which are added by the two translators, despite the fact that there are only two conjunctions in the ST, namely “还有” (PY: hái yǒu; IG: and)1 and “自从” (PY: zì cóng; IG: since). The addition of conjunctions in the TTs reveals the typological differences between Chinese and English, as it is a general feature for Chinese to use fewer conjunctions (cf. Li and Thompson, 1981). In Example 1,

Figure 8.2 Frequency of different types of Theme in the analysis

144 Wang Bo and Ma Yuanyi Table 8.1 Frequency of textual Themes

conjunction continuative conjunctive Adjunct Total




2 0 0 2

48 1 0 49

37 3 1 41

conjunctions, namely “But if” in TT1 and “But” in TT2, are found in the TTs but not in the ST. Further, an additional continuative—“Well”—is also added in TT1. Example 1 ST: 没 有 钱 的 只好 白 瞧 着。 PY: méi yǒu qián de zhǐ hǎo bái qiáo zhe IG: NEG have money SUB have to NEG look TT1: But if you’re broke. . . . TT2:


Well, watching’s free. But not a crumb for those who cannot pay.

Continuatives and conjunctive Adjuncts do not appear in the ST, and their occurrence in the TTs is quite low. These choices are rather the translators’ preferences. In Example 2, a continuative—“Ah”—is added only in TT2, and is used here to initiate a statement for Silly Young to start chanting and asking for money. Example 2 ST: 明天 好, PY: míng tiān hǎo IG: tomorrow good TT1: Tomorrow’s fine, TT2: Ah, tomorrow’ll be


For interpersonal Themes, only one interrogative element—“哪位” (PY: nǎ wèi; IG: which)—and two Vocatives—“小 姑娘” (PY: xiǎo gū niang; IG: little girl)—are found in the ST (see Table 8.2). However, more types of interpersonal Themes are added in the TTs, especially in TT2, for example Finite verbal operators like “would” and “don’t” and modal/comment Adjuncts like “never” and “no wonder”. These Finite verbal operators are added as a result of typological variation. As there is no Finite in Chinese, translators have to add them as interpersonal Themes in the TTs when they are translating imperatives with negative polarity or interrogatives from Chinese to English. In Example 3, while translating “哪 位” (PY: nǎ

Textual and Logical Choices


Table 8.2 Frequency of interpersonal Themes

Vocative Finite verbal operator interrogative element modal/comment Adjunct Total




2 0 1 0 3

3 1 0 0 4

3 4 0 2 9

wèi; IG: which MEAS) in an interrogative clause, both translators have added “would” that functions as an interpersonal Theme. Besides, an additional Vocative “Old timer there” is also added in TT1. Example 3 ST: 哪 位 爷,愿意 听, PY: nǎ wèi yé yuàn yì tīng IG: which MEAS gentlemen want to hear TT1: Old timer there, would you like to

hear the scene // About Yang Yanzhao and Mu Guiying? TT2: Would you like a story to cheer you up // Of heroes and heroines, For the modal/comment Adjuncts that are added as interpersonal Themes in TT2, they are largely a result of the translators’ personal choices. As illustrated in Example 4, the translator of TT2 adds “no wonder” to express his own assessment of the proposition that the empress dowager is angry. Example 4 ST: [Ø:这 件 事] 气 得 太后 PY: zhè jiàn shì qì de tài hòu IG: this MEAS event enrage VPART [Ø: 太后] 咬 牙 PY: tài hòu yǎo yá IG: empress dowager grind tooth [Ø: 太后] 切 齿 PY: tài hòu qiè chǐ IG: empress dowager grind tooth TT1: She raged,

empress dowager

she ground her teeth and cursed. TT2: No wonder the Empress Dowager was enraged. Table 8.3 and Table 8.4 summarise the frequency of topical Theme. We can see that fewer participants are found in TT1 and TT2. For TT2, it is

146 Wang Bo and Ma Yuanyi Table 8.3 Frequency of topical Themes ST participant process circumstance Total

147 4 3 154



121 5 8 134

107 5 22 134

Table 8.4 Frequency of marked and unmarked topical Themes ST marked topical Theme unmarked topical Theme Total

4 150 154



8 126 134

28 106 134

because the translator has deliberately selected circumstances as marked topical Themes. Therefore, a much larger number of marked choices are found here in TT2. For instance, in Example 5, the translator of TT2 purposely chooses “Tasty meat balls” as the marked topical Theme, so as to put “claim” in the culminative position, and to rhyme “claim” with “game” from the previous line. Example 5 ST: [Ø:您] 赌 一 卖 (碟) 干 炸 丸子 PY: nín dǔ yí mài dié gān zhá wán zi IG: you (HON) bet one MEAS plate dry fried meatball TT1: and [Ø:you] play // For a plate of meatballs— TT2:

losers pay. Tasty meat balls, the winners claim.

By the same token, a smaller number of participants are also found in TT1 compared to the ST. It is because TT1 tends to be briefer while translating some clauses that are similar in structure. Example 6 shows how four clauses in the ST are rendered as one clause in TT1, with the paralleled clauses in the ST being summarised and condensed. Example 6 [Ø:大 茶馆] 茶座 多, dà chá guǎn chá zuò duō big teahouse seat many 大 茶馆] 真 热闹, PY: dà chá guǎn zhēn rè nao ST: PY: IG: [Ø:

Textual and Logical Choices


IG: big teahouse really bustling [Ø: 大 茶馆] 也 有 老 来 PY: dà chá guǎn yě yǒu lǎo lái IG: big teahouse also exist old MOD [Ø: 大 茶馆] 也 有 少; PY: dà chá guǎn yě yǒu shào IG: big teahouse also exist young TT1: There’s lots of seats and lots of

fun, And lots of folk both old and young; TT2: Trade is brisk, lots of tea [Ø: are] sold, everyone [Ø: is] welcome, young and old. As for the metafunctional translation shifts suggested by Matthiessen (2014), we can find many Theme shifts in our data, i.e. translation shifts remaining within the textual metafunction. In Example 6, Theme shifts take place when the Themes in the ST, i.e. “茶座” (PY: chá zuò; IG: seat) and “大 茶馆” (PY: dà chá guǎn; IG: big teahouse), are changed to “There” in TT1, as well as “Trade”, “lots of tea” and “Everyone” in TT2. These shifts are largely motivated, as both translators have attached importance to the form of storytelling in the ST. Therefore, to make the TTs chantable, they choose words that rhyme with each other and put them at the end of the line, despite the various changes made to the Theme and Rheme position; take, for example, “sold” and “old” in TT2 that rhyme with each other. In other words, the shifts here at the lexicogrammatical strata are to maintain the equivalence elsewhere (cf. Halliday, 2009, 2012). Some typological similarities and differences between English and Chinese can be seen based on the lexicogrammatical analysis of Theme: Firstly, the basic choices in the system of theme are similar in English and Chinese. Choices of textual Theme and interpersonal Theme are both optional in major clauses, and choices of topical Theme are both compulsory. In Examples 4, 5 and 6, all major clauses have a topical Theme, such as “茶座” (PY: chá zuò; IG: seat) in Example 6, while textual and interpersonal Themes are not found. Secondly, the topical Theme is more often omitted in the Chinese ST. These omitted topical Themes can be added back in the analysis, such as “大 茶馆” (PY: dà chá guǎn; IG: big teahouse) in Example 6. In the English TTs, the omission of topical Theme is also possible; for instance, in “And fill the place with cultured talk”, the topical Theme is “college students”, which is added back in our analysis. Thirdly, in both languages, the typical order of Theme is textual ^ interpersonal ^ topical. Though textual and interpersonal Themes are rarely found in the Chinese ST, they always occur before the topical Theme. For instance, in “自从 那 日本 兵, 八 年 占据 老 北京。” (PY: zì cóng nà rì běn

148 Wang Bo and Ma Yuanyi bīng bā nián zhàn jù lǎo běi jīng; IG: since that Japanese soldier eight year occupy old Beijing), the textual Theme “自从” (PY: zì cóng; IG: since) is found before the topical Theme “那 日本 兵” (PY: nà rì běn bīng; IG: that Japanese soldier). In “小 姑娘, 别 这样’ ” (PY: xiǎo gū niang bié zhè yàng; IG: little girl don’t this), the interpersonal Theme realised by the Vocative “小 姑娘” (PY: xiǎo gū niang; IG: little girl) is before the topical Theme “别” (PY: bié; IG: don’t). Similar examples can also be found in the English TTs, as in “Now, little girl, don’t be so forlorn” from TT2; the textual Theme “Now” realised by a continuative, and the interpersonal Themes— “little girl” and “don’t” realised by a Vocative and a Finite verbal operator respectively, are all before the topical Theme “be” realised by a process. Fourthly, different from English, there is no system of finite in Chinese. Therefore, in clauses of imperative and polar interrogative mood, there is no interpersonal Theme realised by a Finite verbal operator in Chinese. For example, in an imperative like “这些 事,别 多 说” (PY: zhè xiē shì bié duō shuō; IG: these matter don’t much say), there is no Finite, and “这些 事” (PY: zhè xiē shì; IG: these matter) functions as the topical Theme. In the English TTs, Finites can function as interpersonal Themes in yes/no interrogatives and imperatives, such as “Would” in “Would you like a story to cheer you up” and “Don’t” in “Don’t be hard on poor old me”. Fifthly, in the English TTs, circumstances that function as topical Themes are analysed as marked choices, such as “into the teahouse” in “into the teahouse I go”. However, in the Chinese ST, circumstances in clauses whose process types are analysed as existential or relational: existential are considered as the unmarked topical Themes (cf. Halliday and McDonald, 2004). For instance, in “[Ø:朝 中] 还 有 那 康有为 和 梁启超” (PY: cháo zhōng hái yǒu nà kāng yǒu wéi hé liáng qǐ chāo; IG: court in also exist that Kang Youwei and Liang Qichao), “朝中” (PY: cháo zhōng; IG: court in) is here considered as the unmarked topical Theme. In terms of the textual choices, translators have to choose to interpret the messages and the sequences of messages in the Chinese source text; they also have to choose to re-present the textual meanings among the options in Chinese (Matthiessen, 2014). Based on the lexicogrammatical analysis of Theme, we can summarise some of their trade-offs as follows: Firstly, the translators need to add a large amount of textual Themes based on their interpretation of the ST, to recreate the flow of information in the ST, and to reproduce a cohesive TT in English. Most of these textual Themes are realised by conjunctions like “and” and “but” in the TTs. In Example 7, both translators add “but” to show the adversative relation and link the clause to the previous discourse. Example 7 ST: PY: IG:

[Ø:他] 动 脑筋, tā dòng nǎo jīn he use brain

Textual and Logical Choices


[Ø: 他] 白 费 力, PY: tā bái fèi lì IG: he NEG waste strength TT1: But all his effort is in vain; TT2: But all his efforts, alas, are looking pretty thin, Secondly, while choosing certain mood types in the English system of mood, translators have to be aware of the systemic differences between the two languages. For instance, interpersonal Themes like “would” have to be added in “Old timer there, would you like to hear the scene about Yang Yanzhao and Mu Guiying?” from TT1. Thirdly, textual choices sometimes have to give way to the need for rhyming and performability. Translators thus resort to translation strategies like dynamic equivalence, etc. (cf. Nida, 1964; Wang and Ma, 2016). In Example 8, changes are made to Theme choices in the ST. No Vocative is found in the ST, in which “王掌柜” (PY: wáng zhǎng guì; IG: Manager Wang) functions as the topical Theme. In the two TTs, “王掌柜” (PY: wáng zhǎng guì; IG: Manager Wang) is translated as a Vocative. The translator of TT1 puts “Proprietor Wang” in the culminative position, in order to rhyme “Wang” with “tongue” in the next line. However, in TT2, “Manager Wang” occupies the thematic position, and “times” is used at the end, so as to rhyme with “rhymes” in the next line. These shifts are motivated because they reflect the translators’ concern for recreating the rhymed way of storytelling in the translations. Example 8 王掌柜, 大 发 财 wáng zhǎng guì dà fā cái Manager Wang big make fortune 金 银 元宝 一齐 来。 PY: jīn yín yuán bǎo yì qí lái IG: gold silver ingot together come TT1: You’ve made your pile, // Of silver and gold, Proprietor Wang. TT2: Manager Wang, for you these seem profitable times, ST: PY: IG:

4.2. Analysis of Taxis and Logico-Semantic Type After the analysis of taxis and logico-semantic type, some quantitative differences can be observed (see Figure 8.3). In the system of taxis, among the choices of parataxis and hypotaxis, parataxis is the preferred choice in the ST, TT1 and TT2. As for hypotaxis, its frequency in TT2 is much higher than those in the ST and TT1. Example 9 (Table 8.5) shows how choices of taxis vary among the ST, TT1 and TT2. We can see that the tactic structures in the ST, TT1 and TT2

150 Wang Bo and Ma Yuanyi

Figure 8.3 Frequency of taxis and logico-semantic type

are all different. Instead of choosing parataxis, hypotaxis is here selected by the translator of TT2, marked explicitly by “till”. A tactic shift has taken place, which is a shift from parataxis to hypotaxis, and is still within the logical metafunction. The motivation for this shift is the consideration of the lyrical form. In this way, the clause with hypotactic structure that ends with “stop” can rhyme with “shop” in the previous line. For logico-semantic type, as seen in Figure 8.3, the overall frequency for projection is similar among the ST, TT1 and TT2. As for expansion, the choices are different to some extent. In terms of elaboration and enhancement, compared to the choices made in the ST, a lower frequency is found in TT1, while a higher frequency is found in TT2. Extension, which is the preferred choice in the ST and TT1, turns out to be the least favourite choice in TT2, with the lowest occurrence being found compared to those of elaboration and enhancement. A detailed comparison between the ST and TT2 has revealed that eight shifts of logico-semantic type from extension to enhancement are in total found in TT2. As illustrated in Example 10 (Table 8.6), the extending relation in the ST combines the two paratactic clauses, whereas in TT2, an extra clause—“Life’s hard for a rhymester”—is added, and the logico-semantic relation is changed to that of enhancement, by using a paratactic link of “so”. The reason is the translators’ preferences in the reconstrual of the logico-semantic types. The translator of TT2 chooses a free strategy, which allows him to make various changes to the logico-semantic types in his translation. Thus, the enhancing relations (×) not found in the ST are twice used in TT2. The translator of TT1, on the other hand, chooses to translate the extending relations (+) equivalently. As revealed in Example 11 (Table 8.7), different choices are made in the ST, TT1 and TT2, and shifts of logico-semantic type are found. We

(我) 大傻杨, PY: wǒ dà shǎ yáng IG: I Silly Young [Ø:我] 打 竹板儿, PY: wǒ dǎ zhú bǎ n er IG: I beat bamboo clapper [Ø:我] 一来PY: wǒ yì lái IG: I come [Ø:我] 来 到 大 茶 馆儿。 PY: wǒ lái dào dà chá guǎ n er IG: I come PV big teahouse



Tongue a’wagging

[Ø:I’m] still dirt-poor.


My clapper [Ø:is] clacking, I’m here once more.


I’m always here.

I’m Oddball Yang, a balladeer. This is Yutai Teahouse;


+2 1



[Ø:我] 打 竹板 PY: wǒ dǎ zhú bǎ n IG: I beat bamboo clapper 我 又 来, PY: wǒ yòu lái IG: I again come 数来宝 的 还是 没 发财。 PY: shǔ lái bǎ o de hái shì méi fā cái IG: rhythmic storytelling SUB still NEG get rich



Tactic structure




Tactic structure

Tactic structure

Table 8.6 Example 10

+3 =4



Tactic structure

Table 8.5 Example 9


+2 1


Tactic structure




Tactic structure

so a beggar I remain.

Life’s hard for a rhymester,

Beating my clappers, here I am again,


and from shop to shop, I make my rounds till here I stop.

I’m Silly Yang,



“Reform’s rebellion” —so she said.

=2 2

She screamed for heads;

She screamed for blood,


=2 “1

1 +2


她 要 杀, PY: tā yào shā IG: she want kill 她 要 砍, PY: tā yào kǎn IG: She want decapitate 讲 维新 的 都 是 要 造反。 PY: jiǎng wéi xīn de dōu shì yào zào fǎn IG: suggest reform SUB all be want rebel.



Tactic structure


Tactic structure

Table 8.7 Example 11

So the movement was crushed, nipped in the bud.

×2 α ×2 +β

wanting blood,

she screamed,

11α 1 1×β




Tactic structure

Textual and Logical Choices


can see that projection of locution is added in both TTs, and the relationship of enhancement is omitted in TT1. Additional logico-semantic types are also added in TT2, such as enhancement and extension. The choice of rhyming and the concern for performability have played their part in the translators’ choices. As the ST rhymes “砍” (PY: kǎn; IG: decapitate) with “反” (PY: fǎn; IG: rebel), a similar rhyming pattern is recreated in TT2, i.e. “blood” and “bud”, whereas “heads” and “said” in TT1 is a less complete rhyming pattern, though the translator has also intended to rhyme in these two lines. Both translators are trying to produce a kind of rhymed couplets in the TTs, which is suitable for performance. The translator of TT2 takes performability more seriously (Ying, 1999), and thus sacrifices the choices in other modes of meaning such as logical relations. The translator of TT1 also expects his translation to be read by readers, so he tries to produce a formally equivalent translation (cf. Nida, 1964) by way of reconstruing the extending relation in the ST (see also Section 4.3). Besides translation shifts of taxis and logico-semantic type, which are within the logical metafunction, there are also translation shifts from textual to logical metafunction and vice versa. A shift from textual to logical metafunction takes place in Example 12 (Table 8.8), as no textual Theme is found in the ST, while several of them are found in the TTs, namely “But” in TT1 and “But”, “For” and “nor” in TT2. As held by Matthiessen (2014), textual transitions can be either implicit or explicit. In the ST, all textual transitions are implicit, left for readers or the audience to infer. However, in both TTs, such transitions are made explicit by way of adding cohesive conjunctions, thus involving translation shifts from textual meaning to logical meaning. The reason behind this is that the Chinese monologue is written in the form of doggerel, and the textual connectors are seldom used. In the couplet style of storytelling, the numbers of the Chinese characters in two adjacent lines are normally the same. Conjunctions in Chinese, such as “然后” (PY: rán hòu; IG: and), “可是” (PY: kě shì; IG: but) and “因为” (PY: yīn wéi; IG: because), not only would be regarded as redundant, but would also destroy the rhythmic pattern in Chinese (see also Section 4.3 for the analysis of mode). Conversely, there are also translation shifts from logical to textual metafunction. In Example 13 (Table 8.9), the translator of TT1 has broken the tactic relations between the clauses in the ST, and has translated the tactically related clauses as structurally unrelated clauses that form a cohesive sequence. In this way, TT1 has become less explicit than the ST. By the same token, in Example 9 (Table 8.5), TT1 also disregards the tactic structure in the ST, resulting in a similar translation shift. Based on several examples, we can see that the translator of TT1 prefers translating in such a way, which results in shifts from logical to textual metafunction.

[Ø:他] 动 脑筋, PY: tā dòng nǎo jīn IG: he use brain [Ø:他] 白 费 力, PY: tā bái fèi lì IG: he NEG waste strength 胳膊 拧 不 过 大腿 去。 PY: gē bo nǐng bú guò dà tuǐ qù IG: arm wring NEG PV leg PV

这些 事, 别 多 说, PY: zhè xiē shì bié duō shuō IG: these matters NEG much say 说着 PY: shuō zhe IG: say VPART 说着 PY: shuō zhe IG: say VPART 就 许 掉 脑壳。 PY: jiù xǔ diào nǎo ké IG: VADV maybe drop head






Tactic structure

Table 8.9 Example 13




Tactic structure

Table 8.8 Example 12

Tactic structure



Tactic structure

The longer you’ll keep your head from harm.

So the less you say about reform.


There are certain things you can’t attain.

But all his effort is in vain;





Tactic structure

Talking too freely will surely risk my neck!

and hold myself in check,

But I’d better stop


nor with tails did win.

For with heads he lost,

×β α ×β+β

But all his efforts, alas, are looking pretty thin,



Tactic structure

Textual and Logical Choices


For logical choices, more similarities than differences between Chinese and English are found according to the analysis in this study. In both languages, the basic choices in the systems of clause complexing include taxis, logico-semantic type and recursion (cf. Halliday and Matthiessen, 2014). For the system of taxis, the choices are hypotaxis and parataxis. For the system of logico-semantic type, the fundamental categories include expansion and projection. More delicate choices for expansion and projection are also identified, with those for expansion being elaborating, extending and enhancing, and those for projection being locution and idea. One of the primary differences between the two languages can be found in the lexical choices of conjunctions that mark out the tactic structure explicitly. However, most such conjunctions in our data are omitted, and the only example is “自从” (PY: zì cóng; IG: since), which suggests an enhancing logico-semantic relation. When translators interpret the relations or when researchers analyse the text as data, these omitted conjunctions can be added back. For instance, in Example 14 (Table 8.10), conjunctions, i.e. “所以” (PY: suǒ yǐ; IG: so), can be added to the Chinese ST, thus revealing the logico-semantic relations of enhancing. The two translators have both interpreted the relations as enhancing, and the conjunction “so” is thus added to both TTs to equivalently reconstrue the relations. In terms of the logical choices, the difficulties for the translators can be summarised as follows: Firstly, the translators need to choose how to interpret the choices of taxis and logico-semantic type in the ST. Since the textual transitions in the ST are all omitted, the translators have to make their own inferences about the logico-semantic relations between the clauses. Secondly, based on their interpretations, the translators have to reconstrue the logico-semantic relations in the TTs they are producing. As seen in our analysis, the choices are made in systems like taxis, logico-semantic type and recursion (whether to stop or continue a clause complex). In addition, similar concerns for rhyming and performability also play a part in the translators’ reconstrual of logical meaning (see Section 4.1 and Examples 1, 5, 12 and 13). Based on the translators’ personal preferences, they are free to make completely different logical choices from those in the ST, by adding or omitting clauses, reorganising the number of clauses in a clause complex, etc. For instance, tactic structures can be omitted when translators reorganise the number of clauses in a clause complex. In Example 15 (Table 8.11), the translator of TT2 translates one clause complex as two clause complexes, and has omitted the tactic structure of “1 + 2” in the ST. Besides, the alteration of taxis is also found when parataxis in the ST is changed to hypotaxis in TT1. In the ST, the two clauses are of equal status and are in

[Ø: who are] Outdoing the Japs in everything.

Now the Kuomintang are in Beijing,


国民党, 进 北京, PY: guó mín dǎng jìn běi jīng IG: Kuomintang enter Beijing 横行霸道 一点 不 让 日本 兵。 PY: héng xíng bà dào yì diǎn bú ràng rì běn bīng IG: tyrannical little NEG lose out Japanese soldier




Tactic structure


So they waste our wealth on foreign ones.

Tactic structure

Table 8.11 Example 15

×2 ×2

they’ve got to have guns,

But in order to fight


[Ø:军阀] 为 打仗, PY: jūn fá wèi dǎ zhàng IG: warlord to fight [Ø:所以] [Ø:军阀] 要枪炮, PY: suǒ yǐ jūn fá yào qiāng pào IG: so warlord want gun [Ø:所以] [Ø:军阀] 一 堆 一 堆 给 洋人 老 爷 送 钞票。 PY: suǒ yǐ jūn fá yì duī yì duī gěi yáng rén lǎo ye sòng chāo piào IG: so warlord one pile one pile CV foreigner lord send money



Tactic structure


Tactic structure

Table 8.10 Example 14

Tactic structure



Tactic structure

[Ø: They are] As cruel a tyrant as the Japs [[could ever be]].

Then to old Beijing came the KMT!


So to foreign countries went silver by the tons.

one must buy guns,

In order to fight,


Textual and Logical Choices


a paratactic relation. In TT1, the two clauses are linked by way of a hypotactic relation, as the first clause is playing a dominant role, and the second clause, realised by a relative clause, is dependent upon the first one. 4.3. Analysis of Context In this part, we will move from lexicogrammar to context, with the purpose of interpreting the pattern of the dramatic monologue we have just analysed. As noted by Butt, Fahey, Feez, Spinks and Yallop (1994), the relationship between text and context is dynamic and reversible, in that texts can reveal context, and context is realised in texts. On the one hand, knowledge of the context helps us predict the lexicogrammar of the text. On the other hand, lexicogrammatical analysis is helpful for an understanding of the context, as the total meaning encoded in the lexicogrammar will become the sign of context (cf. Halliday, 1978, 1991). Our analysis of the context of situation is characterised by the three contextual parameters, namely field (what is going on), tenor (what is the social relationship between the people who take part) and mode (what role language plays in the text). In terms of field, as previously discussed, the field of activity being recreated in ST is that of exploring (cf. Matthiessen, 2015; Matthiessen and Teruya, 2016). Before each act of the play, the storyteller of the dramatic monologue will express his stances and positions by way of chanting rhythmic storytelling. That is also related to the purpose of creating this character, i.e. to introduce the situations going on in the teahouse. In the TTs, the field of activity remains unchanged. If we carry out a transitivity analysis of the data, we will find that despite the omissions and additions found in the data of dramatic monologues, most of the process types in the ST that are used to introduce the teahouse and the historical background, and to express opinions, have been translated equivalently in both TTs. The description of tenor will focus on the relations between the writer, the translators and the readers. Firstly, in terms of institutional role, Lao She—the writer of the ST—is a famous Chinese playwright. The translator of TT1, John Howard-Gibbon, used to be a journalist and a college teacher of English language and literature in China. The motivations for him to translate Teahouse include his interest in Chinese language, his love of Chinese literature and his desire to improve his Chinese (HowardGibbon, 2004). As for the translator of TT2, Ying Ruocheng was a professional actor at Beijing Peoples’ Art Theatre, and he also performed in Teahouse. By the time he translated this play, Howard-Gibbon’s translation of the play had already been published. Ying believed that HowardGibbon’s translation was error-ridden, with many cultural connotations being translated in a wrong way (Howard-Gibbon, 2004), and also was

158 Wang Bo and Ma Yuanyi not suitable for performance (Ying, 1999). Thus, he wanted to provide a colloquial translation characterised by concise language. Secondly, with regard to power and status roles, the roles of Lao She, Howard-Gibbon, Ying and the audience are all unequal. Lao She and the two translators all adopt the role of recounter and are responsible for the characterisation, i.e. depiction, of the various characters in the play. The complementary role is meanwhile assigned to the audience or readers. Thirdly, in terms of distance, the playwright and the translators are authorities, as persons holding authority and specialists in the area of drama, whereas the readers and the audience are unseen and unknown. Ying, the translator of TT2, attempts to bridge this distance and puts great emphasis on performability. However, no similar attempt is found in Howard-Gibbon, as his focus mainly lies in the accuracy of his translation. In his own words, his limited fluency in the language has forced him to “dwell over the text” and has made him very “text sensitive” (HowardGibbon, 2004). As Ying concentrates on the form of the rhythmic storytelling and performability, he sacrifices the textual, logical and even experiential modes of meaning in TT2. Based on the discussions in the previous sections, the shifts found in TT2 include addition of textual Themes (Examples 1 and 2), addition or substitution of topical Themes (Examples 1, 3 and 6), addition of interpersonal Themes (Examples 3 and 4), alteration of topical Themes in the ST into Rhemes (Example 5), changes in tactic structures (Examples 11, 12 and 13), etc. Compared to Ying, Howard-Gibbon focuses largely on experiential meaning and the form of the rhythmic storytelling (cf. Howard-Gibbon, 2004). Some translation shifts are also found in TT1, such as addition of textual Themes (Examples 1 and 5), substitution of topical Themes (Example 6), changes in tactic structures (Example 11, 12 and 13), etc. However, such shifts occur with a lower frequency compared to TT2 and are mainly used to guarantee the equivalence of experiential meaning and the translator’s rhyme choices. In terms of mode, the ST is written to be spoken and to be performed on stage. It is monologic and written in the form of Chinese rhythmic storytelling. The lines are written in couplets, with end rhymes being used at the end of each line. For the two TTs, TT1 is rather written to be read by readers, as revealed in various studies (cf. Ren, 2008). TT2, however, is translated into colloquial and simple language, with the purpose of being performed (Ying, 1999). It has been used as English sub-titles for performances in other countries since the 1980s. To sum up, the contextual analysis has given proof of the reasons for the previously discussed lexicogrammatical choices made by the two translators, because their choices are to some extent influenced by the context. Quantitatively, TT1 is much closer to the ST than TT2 in many respects, such as in the use of the interpersonal Theme, marked and

Textual and Logical Choices


unmarked topical Theme, parataxis, elaboration and extension as well as projection. As mentioned above, TT1 is translated to be read rather than performed, and thus has a lower requirement for the rhyming scheme or the use of colloquial language. However, to realise the performability of the translation, the translator of TT2 has to make various changes to the ST, such as increasing the use of the marked topical Theme and textual Theme, changing the use of parataxis and extension, etc.

5. Conclusion This chapter attempts to reveal the textual and logical choices made in the original and the two translations of the dramatic monologue of Teahouse. We have reported our quantitative findings in the analysis of Theme, taxis and logico-semantic type. Some different choices made by the two translators have been pointed out, and the reasons are explained. Further, we have illustrated four different kinds of metafunctional translation shift in relation to our analysis, such as shift remaining in the textual metafunction, shift remaining in the logical metafunction, shift from textual to logical metafunction and shift from logical to textual metafunction. We have also related the translators’ concern for performability to the choices they made, and discussed the translation difficulties as revealed in the analysis. In the present study, Chinese rhythmic storytelling has for one of the first times been analysed in light of SFL. We have applied SFL to the study of drama translation, and the detailed lexicogrammatical analysis has reflected the translators’ preferences in terms of textual and logical meaning. Based on the lexicogrammatical analysis of the data, we have compared Chinese and English and discussed some systemic similarities and differences between the two languages, thus contributing to various areas of multilingual studies in SFL, which include studies on language typology (cf. Mwinlaaru and Xuan, 2016). This study also presents us with the options translators have and the challenges they are expected to meet in translation, especially in the translation of Chinese rhythmic storytelling or limericks, in the case of our data.

Note 1. We have provided Pinyin (PY) and interlinear glossing (IG) for the examples in Chinese, and the following abbreviations are used in referring to interlinear glossing (cf. Halliday and McDonald, 2004; Li, 2007): CV: coverb HON: honorific MEAS: measurer MOD: verbal particle: modal NEG: verbal particle: negative PV: postverb

160 Wang Bo and Ma Yuanyi SUB: subordinating VADV: verbal adverb VPART: verbal particle

References Butt, D., Fahey, R., Feez, S., Spinks, S. and Yallop, C. (1994) Using Functional Grammar: An Explorer’s Guide. Sydney: National Centre for English Language Teaching and Research. Catford, J. C. (1965) A Linguistic Theory of Translation. London: Oxford University Press. Halliday, M. A. K. (1956) The linguistic basis of a mechanical thesaurus, and its application to English preposition classification. Mechanical Translation, 3(3), pp. 81–88. Halliday, M. A. K. (1961) Categories of the theory of grammar. Word, 17(3), pp. 241–292. Halliday, M. A. K. (1962) Linguistics and machine translation. Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung, 15(1–2), pp. 145–158. Halliday, M. A. K. (1978) Language as Social Semiotic: The Social Interpretation of Language and Meaning. London: Edward Arnold. Halliday, M. A. K. (1985) An Introduction to Functional Grammar. London: Edward Arnold. Halliday, M. A. K. (1991) The notion of ‘context’ in language education. In T. Lê and M. McCausland (Eds.), Language Education: Interaction and Development, Proceedings of the International Conference, Vietnam, April, 1991. Launceston: University of Tasmania, pp. 1–26. Halliday, M. A. K. (2001) Towards a theory of good translation. In E. Steiner and C. Yallop (Eds.), Exploring Translation and Multilingual Text Production: Beyond Content. Berlin: Mouton de Gruyter, pp. 13–18. Halliday, M. A. K. (2009) The gloosy ganoderm: Systemic functional linguistics and translation. Chinese Translators Journal, 1, pp. 17–26. Halliday, M. A. K. (2012) Pinpointing the choice: Meaning and the search for equivalents in a translated text. In A. Mahboob and N. K. Knight (Eds.), Appliable Linguistics. London: Continuum, pp. 13–24. Halliday, M. A. K. and Matthiessen, C. M. I. M. (2014) Halliday’s Introduction to Functional Grammar, 4th ed. London and New York: Routledge. Halliday, M. A. K. and McDonald, E. (2004) Metafunctional profile of the grammar of Chinese. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (Eds.), Language Typology: A Functional Perspective. Amsterdam: John Benjamins, pp. 253–305. Hansen-Schirra, S., Neumann, S. and Steiner, E. (Eds.). (2012) Cross-linguistic Corpora for the Study of Translations: Insights From the Language Pair EnglishGerman. München: Walter de Gruyter. Hatim, B. and Mason, I. (1990) Discourse and the Translator. London: Longman. House, J. (1977) A Model for Translation Quality Assessment. Tübingen: Gunter Narr. House, J. (1997) Translation Quality Assessment: A Model Revisited. Tübingen: Gunter Narr.

Textual and Logical Choices


Howard-Gibbon, J. (2004) Counting Your Treasures. Available at: www.chinadaily. (Accessed on 13 February 2004). Li, C. and Thompson, S. (1981) Mandarin Chinese: A Functional Reference Grammar. Berkeley, Los Angeles and London: University of California Press. Li, E. S. (2007) A Systemic Functional Grammar of Chinese. London and New York: Continuum. Matthiessen, C. M. I. M. (2001) The environments of translation. In E. Steiner and C. Yallop (Eds.), Exploring Translation and Multilingual Text Production: Beyond Content. Berlin: Mouton de Gruyter, pp. 41–124. Matthiessen, C. M. I. M. (2014) Choice in translation: Metafunctional considerations. In K. Kunz, E. Teich, S. Hansen-Schirra, S. Neumann and P. Daut (Eds.), Caught in the Middle: Language Use and Translation: A Festschrift for Erich Steiner on the Occasion of His 60th Birthday. Saarbrücken: Saarland University Press, pp. 271–333. Matthiessen, C. M. I. M. (2015) Register in the round: Registerial cartography. Functional Linguistics, 2(9), pp. 1–48. Matthiessen, C. M. I. M. and Teruya, K. (2016) Registerial hybridity: Indeterminacy among fields of activity. In D. R. Miller and P. Bayley (Eds.), Hybridity in Systemic Functional Linguistics: Grammar, Text and Discursive Context. Sheffield: Equinox, pp. 205–239. Mwinlaaru, I. N. and Xuan, W. W. (2016) A survey of studies in systemic functional language description and typology. Functional Linguistics, 3(8), pp. 1–41. Nida, E. A. (1964) Towards a Science of Translation: With Special Reference to Principles and Procedures Involved in Bible Translating. Leiden: Brill. Reiss, K. (1971) Translation Criticism: The Potential and Limitations: Categories and Criteria for Translation Quality Assessment, translated by E. F. Rhodes. Manchester: St. Jerome. Ren, X. (2008) Translator on Stage: A Systematic Study on Ying Ruocheng’s Drama Translation. Beijing: China Social Sciences Press. Steiner, E. (2004) Translated Texts: Properties, Variants, Evaluations. Frankfurt am Main: Peter Lang. Steiner, E. (2008) Explicitation: Towards an empirical and corpus-based methodology. In J. J. Webster (Ed.), Meaning in Context: Implementing Intelligent Applications of Language Studies. London and New York: Continuum, pp. 234–277. Steiner, E. (2015) Halliday’s contribution to a theory of translation. In J. J. Webster (Ed.), The Bloomsbury Companion to M. A. K. Halliday. London and New York: Bloomsbury, pp. 412–426. Teich, E. (2003) Cross-linguistic Variation in System and Text: A Methodology for the Investigation of Translations and Comparable Texts. Berlin and New York: Mouton de Gruyter. Thompson, G. (1999) Acting the part: Lexico-grammatical choices and contextual factors. In M. Ghadessy (Ed.), Text and Context in Functional Linguistics. Amsterdam: John Benjamins, pp. 101–124. Toury, G. (1995) Descriptive Translation Studies and Beyond. Amsterdam and Philadelphia: John Benjamins. Wang, B. [王博] and Ma, Y. [马园艺] (2016) ‘译者的选择— 《茶馆》粤语译本的 翻译策略’ [Choice of the translator: Translation strategies for the Cantonese

162 Wang Bo and Ma Yuanyi version of Teahouse], in 杨连瑞 [L Yang] (Ed.), 中国外语研究:2015 年卷 [Foreign Language Research in China: 2015]. Qingdao: China Ocean University Press, pp. 82–92. Ying, R. [英若诚] (1999) ‘序言’ [Preface]. In S. Lao (Ed.), 茶馆 [Teahouse], translated by R. Ying [英若诚]. Beijing: China Translation and Publishing Corporation, pp. 1–11.


A Contrastive Description of Projection in English and Spanish Across Ranks From the Clause Nexus to the Group Jorge Arús Hita

1. Introduction1 Systemic Functional Linguistics (SFL) has throughout the years proven to be a reliable tool for the general description of languages as well as for the description of specific areas within languages. One of those areas is that of projection, i.e. “the logical-semantic relationship whereby a clause comes to function not as a direct representation of (non-linguistic) experience but as a representation of a (linguistic) representation” (Halliday and Matthiessen, 2014, p. 508). As Halliday and Matthiessen point out, projection serves a number of important discursive uses, such as “to attribute to sources in news reporting, to represent views in scientific discourse, to construct dialogue in narrative . . ., to frame questions in conversation” (2014, p. 509). Studies on projection have for the most part focused on the description of quoting and reporting, as in examples (1) and (2), respectively, i.e. at the clause nexus level.2 Yet, as shown by Matthiessen and Teruya (in preparation), projection is a fractal pattern; i.e. it is manifested in different environments in the lexicogrammar. We find projection not only within the clause nexus, as in (1) and (2), but also within the simple clause (3) and the group (4). In (3), the comment Adjunct reportedly, a constituent at the simple clause level, makes the rest of the clause appear as a projection. This could have been expressed as they say that plans are afoot for . . . The same effect is achieved by the post-determiner alleged in (4); acting at the group level—within the nominal group the alleged conspiracy—it manages to provide a meaning of projection which most directly concerns the Head of the group but also affects the rest of the clause, as the statement is not about a first-order reality—a conspiracy— but a second-order one—an alleged conspiracy, i.e. what someone has claimed to be a conspiracy. (1) “That was probably the turning point,” said the former champion (K5J W_newsp_other_social)3


Jorge Arús Hita

(2) the DRA Foreign Minister claimed that Afghanistan was ready to reach a settlement (GVK W_ac_polit_law_edu) (3) Plans are reportedly afoot for a new soap called ‘Ministry’ (A3D W_newsp_brdsht_nat_report) (4) The overt acts of the alleged conspiracy are said to be set out in the statement of claim (FDY W_ac_polit_law_edu) The fractal nature of some semantic categories—one of which is projection—was first suggested by Halliday and Matthiessen (2009) in their meaning-based description of the construal of experience. That projection is found also in environments other than the clause nexus is already present in the third edition of Halliday’s introduction to Systemic Functional Grammar (Halliday and Matthiessen, 2004, pp. 603–613; also in 2014, pp. 676–686). The fractality of projection is then fully expounded in Matthiessen and Teruya (in preparation). The fractality of projection is a multilingual phenomenon, as shown by several researchers in Teruya (2015). Arús Hita (2015) offers an account of projection in Spanish, where this resource is also found at different ranks, as illustrated by examples (5–8). As in the English examples above, (5, 6) illustrate projection within a clause nexus, quoting and reporting respectively, whereas in (7) the Comment Adjunct supuestamente (‘supposedly’) plays a similar role to allegedly in (3), and the post-determiner presunto is comparable in effect and scope to alleged in (4). (5) ¡Cuánta fe!—exclamó la mujer con renovada ansiedad (19-F la víspera y el día [‘So much faith!—exclaimed the woman with renewed anxiety’]) (6) La gente piensa que cuando hacemos publicidad es para vender (19OR España Oral EEDU018A [‘People think that we advertise in order to sell’]) (7) Supuestamente el límite para algunos cursos son quince estudiantes (19-OR Habla Culta San Juan (PR): M2 [‘supposedly the limit for some courses is fifteen students’]) (8) El presunto abuso le provocó crisis depresivas (19N Hon: Prensa: 98May27 [‘The alleged abuse brought about bouts of depression’]) The present chapter draws upon Arús Hita’s (2008, 2015) description of Spanish projection as well as on Matthiessen and Teruya’s (in preparation) work for English to provide a detailed contrast of projection across ranks in English and Spanish. We discuss the similarities and dissimilarities in terms of the semiotic potential and realisation of projection in both languages. The chapter is structured as follows: sections 2 and 3 look at projection in English and Spanish within the clause nexus and other lexicogrammatical domains, respectively—the latter focusing on the clause and

A Contrastive Description of Projection


the group. Both sections include contrastive discussion. This discussion is summarised in the conclusion before pointing out some venues for future research.

2. Projection Within the Clause Nexus As said above, the clause nexus is the environment with which projection is typically identified. Indicative of that is the fact that Halliday’s first two editions of his introduction to Systemic Functional Grammar (1985, 1994) only cover this kind of projection. It is not until the third edition (Halliday and Matthiessen, 2004) that projection is more explicitly described in other environments within the lexicogrammar. When speaking about projection within the clause nexus, it is important to recall the projection systems identified by Halliday and Matthiessen (2014, pp. 509–511): (i) level of projection (idea [mental] vs locution [verbal]); (ii) mode of projection (reporting [hypotactic] vs quoting [paratactic]; and (iii) speech function (proposition [statements and questions] vs proposal [offers and commands]). Concerning the level of projection, both English and Spanish coincide in their potential for projecting ideas and locutions; i.e. in both languages verbal and mental processes may project, as illustrated by the contrastive pairs (9a, b) and (10a, b), the former verbal, the latter mental. (9) (a) One policeman said he was planning to leave the force after 21 years (FBL, W_pop_lore) (b) La policía dijo que Pastor Argüelles, de 50 años y odontóloga de profesión, tenía una profunda herida de cuchillo en el cuello (19-N Hon:Prensa:98May16 [‘The police said that Pastor Argüelles, aged 50 and a professional odontologist, had a deep knife cut in her neck’]) (10) (a) he thought that perhaps eighteen months was a reasonable guess before he would actually manage to get somebody (JJX, S_courtroom) (b) y pensó que era muy dura para él esta carrera de médico (18 En la carrera [‘and he thought that the medical career was too hard for him’]) As for mode of projection, both languages behave similarly in terms of their potential for quoting and reporting and the relationship between projecting and projected clauses. Thus, quotes in English and Spanish are paratactically related to the projecting clause, whereas reports are hypotactically so (see Halliday and Matthiessen, 2014, p. 509; Lavid, Arús Hita and Zamorano-Mansilla, 2010, pp. 66–69).4


Jorge Arús Hita

The cross-linguistic similarity also extends to the fact that quotes may precede or follow the projecting clause, whereas reports are always after the main clause. Only in very colloquial spoken Spanish is it possible to come across a report preceding the projecting clause (e.g. que vengas, dice mamá [lit. ‘that you should come, says mum’]). Yet this is so rare and highly marked that our corpus scrutiny yielded no results. The English and Spanish pairs in (11–13) illustrate the points just made. (11) Quote before projecting clause (paratactic relation) (a) ‘Why didn’t you say so before?” exclaimed the Bookman (AMB, W_fict_prose) (b) ¡Me han robado el reloj!—exclamó el hombre, muy excitado (19 F, Hijo de ladrón [‘They’ve stolen my watch!—exclaimed the man, very upset’]) (12) Quote after projecting clause (paratactic relation) (a) But he then added: ‘I hope the Government will not delay further its clear commitment to join the full EMS’ (A4K, W_newsp_brdsht_nat_report) (b) Entonces añadió Chisquín, alentado por los aplausos de sus compañeros,—¿qué derecho es ese que se nos da? (18, Don Gonzalo Gonzáles de la Gonzalera [‘And then Chisquín added, encouraged by his workmates’ applause,—what’s that right they’re giving us?’]) (13) Report, always after projecting clause (hypotactic relation) (a) They decided that Victoria would accompany them to the funeral (C98, W_fict_prose) (b) pensó que era necesario plasmar en el papel aquellas vivencias de juventud (19F, Ayer soñé con Valparaíso [‘he thought it was necessary to put those experiences from his youth in black and white’]) Still within mode of projection, both languages display the same potential for mixed projections, i.e. those with some characteristics of reports and some of quotes, and for projecting clauses embedded within a projection. Mixed projection may involve reports containing some quoted parts, as in (14), or the well-known free indirect speech, with paratactic structure but hypotactic shifting of time and person reference (see Halliday and Matthiessen, 2014, pp. 531–533 for English; Lavid et al., 2010, p. 81, Maldonado González, 1999, p. 3552 for Spanish). Thus, (15a) is halfway between a quote such as She wondered: is it true? and a

A Contrastive Description of Projection


report such as She wondered if it was true. Likewise, Spanish (15b) uses the structure of the quote ¿O seré yo? Se preguntó (‘or will that [i.e. the problem] be me? She wondered’) with the time and person reference corresponding to the report Se preguntó si sería ella (‘She wondered whether it [i.e. the problem] would be herself’). Projection surrounding a projecting clause in English and Spanish is illustrated by the examples in (16). Notice, incidentally, that the Spanish example (16b) has no inverted commas to signal the quote. This happens quite often both in Spanish and in English. (14) (a) in the example above officers had reported that the water was ‘clear and satisfactory’ (FA1, W_ac_polit_law_edu) (b) Asimismo, el abogado Daniel Olivas especificó que su representada “acudirá a los tribunales internacionales para denunciar a Ortega” (19N, Hon: Prensa: 98Jun22 [‘Likewise, the lawyer Danier Olivas specified that his client “will go to the international courts of justice to sue Ortega” ’] (15) (a) Was it true, she wondered (BMU, W_fict_prose) (b) ¿O sería ella?, se preguntó. (19F, Algo terrible y poderoso [‘Or, would it (the problem) be her(self)? She wondered.’] (16) (a) ‘I’m not sure’, remarked the Captain, ‘that I care to be coupled with mere pickpockets and footpads . . . but that Huge Reward sounds interesting!’ (ALS, W_fict_prose) (b) Buena memoria tienes, observó el villano, para acordarte de todo eso, porque debías ser muy joven entonces (18, Don Felipe el Prudente: novela histórica [‘you have a fine memory, remarked the villain, if you remember all that, because you must have been very young at that time’]) It is when looking at speech function that we can find the largest contrast between English and Spanish, in particular within the projection of reports, as quotes in principle simply reproduce the words as they were previously uttered (or, more rarely, thought) by someone. Each language uses different resources for capturing the fact that, as Halliday and Matthiessen point out, in contrast with reported propositions, the mode in reported proposals is “ ‘irrealis’, or non-actualized” (2014, p. 525). To a large extent, Spanish marks this distinction by means of the indicative and the subjunctive moods, respectively, whereas in English the distinction is manifested differently, as seen below.5 Reported propositions in Spanish are for the most part projected with the verb in the indicative (e.g. estaba in 17a), which is not different from what happens in English, as illustrated by the two reported propositions in (17b).


Jorge Arús Hita

(17) (a) Le preguntaron si en verdad estaba decidida a casarse (19-F, Cien años de soledad [‘they asked her whether she was truly resolved to get married’]) (b) He was asked if he’d been drinking and he said he’d a glass of wine (K1E, W_news_script) When the projecting clause and the reported statement refer to the same grammatical Subject, both Spanish and English may opt for a nonfinite realisation of the report. For instance, in (18a) el cónsul is the Subject of the projecting clause and would have been the referent of the (presupposed) Subject in the projection if this had been realised finitely (e.g. el cónsul reconoció que (él) había hablado con la doctora Trujillo un día [‘the consul admitted that he had spoken with Doctor Trujillo one day’]). English, in contrast, resorts to gerundival clauses in these cases (a finite realisation of [18b] would have been ‘They admitted they had had Ecstasy which had been bought at rave parties’). (18) (a) el cónsul reconoció haber hablado con la doctora Trujillo un día (19-N, US: Herald: 98Jul23 [‘the consul admitted speaking with Doctor Trujillo one day’]) (b) They admitted having Ecstasy which had been bought at rave parties (CEN, W_newsp_other_report) Examples such as (18a, b) illustrate that, within the realm of projection, “there will always be ‘borderline cases’, instances where the line is hard to draw” (Halliday and Matthiessen, 2014, p. 545). Non-finite realisations are typically associated with act clauses (see Halliday and Matthiessen, 2014, pp. 503–508), which typically find their environment in relational and mental: perceptive processes (Halliday and Matthiessen, 2014, p. 504), functioning in the latter as macrophenomena (see Halliday and Matthiessen, 2014, p. 251; Matthiessen, 1995, p. 273 for English; Lavid et al., 2010, p. 125; Arús, 2007, pp. 289–290 for Spanish). Examples (18a, b), however, are clearly verbal, and, to make matters more complicated, projections with admit, like the above-suggested rephrasing of (18b)—They admitted they had had Ecstasy which had been bought at rave parties—are considered embedded facts in Halliday and Matthiessen (2014, p. 545), more precisely of the kind happening in some cognitive and verbal processes where “a projected element may occur which is not projected by that process” (2014, p. 544, emphasis in the original). In any case, regardless of the exact nature of this kind of projections, we have seen that they stand in clear realisational contrast with the more conventional finite realisation of reported statements such as those in (17a, b), while at the same time exhibiting an infinitival vs gerundival contrast in Spanish and English which makes it relevant for the contrastive purposes of this chapter.

A Contrastive Description of Projection


Still within the realm of reported propositions, there is a cross-linguistic contrast worth pointing out and which accounts for a rather recurrent error made by Spanish learners of English as a foreign language and vice versa. It concerns the arrangement of the modal structure in reported questions, notably of the wh- type: whereas English realises Mood in reported questions as Subject ^ Finite (e.g. 19a), Spanish, where there really is no Mood system comparable to that of English (see Lavid et al., 2010, pp. 8–28), resorts to what could be analysed as Finite ^ Subject. Example (19b) incidentally shows something which is proper to this kind of structures in Spanish, i.e. the possible presence of the nexus que before the wh- element, in contrast with English, where the nexus that is never realised in that position (no possibility, then, to have something like *Paul asked that where he might find a workshop). (19) (a) Paul asked where [Subject:] he [Finite:] might find a bookshop (CD2, W_fict_prose) (b) y él les preguntó que dónde [Finite:] estaba [Subject:] la señora (18, Colección de documentos para la historia de México: versión actualizada [‘and he asked them where the lady was’ (lit. ‘where was the lady’)]) Turning now to reported proposals, in Spanish these have their verb in the subjunctive in the case of commands (20a) or suggestions (21a), unless these are expressed by means of a modal verb, in which case the modal verb is in the indicative (22a). As we can see in the b examples in (20,  21), English also realises reported proposals differently from reported propositions. In fact, English makes a further grammatical distinction between the reporting of commands, which use the full infinitive (20b), and of suggestions, which may use the gerund (e.g. 21b) or a that-clause with a residual form of the present subjunctive that formally coincides with the bare infinitive, as in example (21c), where the lack of third person singular inflection in the verb indicates that this is in the subjunctive form. In this, reported suggestions in English and Spanish are therefore not that distant, grammatically speaking. This proximity is actually more manifest in the realisation of reported suggestions by means of a modal verb, where both languages show a clear parallelism, as seen in (22a, b). (20) (a) nos pidió que siguiéramos hablando en la calle (19-F, El nombre prestado [‘he asked us to keep talking in the street’]) (b) Yeah, they told me to phone you up (KD3, S_conv) (21) (a) Sugiere que alguien suba un poco el volumen del CD (19-F, Aquel cuyos ojos se colman de lágrimas [‘he suggests that someone turn up the CD volume a little’])


Jorge Arús Hita

(b) One amendment suggested scrapping the council’s 6,250 donation to the cathedral organ restoration fund and spending the money on retraining the unemployed (CFC, W_newsp_other_report) (c) I . . . would suggest that he too withdraw a minor claim; then add a new demand (A0R, W_fict_prose) (22) (a) dicen que debemos amarnos los unos a los otros (19-F, Sobre héroes y tumbas [‘they say we must love one another’]) (b) I suggest that we should delay submitting this to Number Ten until you have had a chance to draft the actual terms of reference, as you propose (J17, W_fict_prose) Among reported proposals we also find non-finite clauses where, as seen in the English example (21b), and in contrast with non-finite reported propositions (see 18a, b), the implied Subject does not need to be the same as that of the projecting verbal clause. In fact, (21b) illustrates what seems to be a recurrent pattern in both English and Spanish, i.e. the non-finite realisation of reported proposals not overtly referring to anyone in particular. Thus, the actual Agent of the scrapping mentioned in (21b) is left undefined. One could certainly understand who is supposed to perform that action, but what matters here is that no specific referent is expressed. The same can be observed in the Spanish and English examples (23a, b), where the advice in (23a) is targeted to people in general and in (23b) to, presumably, some music lovers, but in both cases such referents are left unspecified and the report realised non-finitely, with an infinitive in Spanish, with a gerund in English. (23) (a) Por ejemplo, los médicos recomiendan mantener una dieta balanceada, no fumar y hacer ejercicio para gozar de buena salud (19-N, Guat:Gerencia:98JUN9 [‘For instance, doctors recommend (lit. ‘to keep’) a balanced diet, not smoking and doing exercise in order to enjoy good health’]) (b) I would advise listening to Disc 2, track 1 (BMC, W_pop_lore) The above does not mean that non-finite clauses are the only possible realisation of reported proposals with unspecific agency; passive thatclauses and, above all, ranged clauses also abound. Examples of the former can be seen in (24a, b), whereas the latter are illustrated by (25a, b). Notice that the Ranges in the (25a, b) processes are nominalisations of congruent realisations such as ordenó que se suspendieran/ordenó suspender (‘ordered that . . . should be cancelled’) and ‘ordered that all red stars should be removed’.6 Because processes (25a, b) are ranged, we are then no longer dealing with a clause nexus but a simple clause.

A Contrastive Description of Projection


(24) (a) los científicos no recomiendan que se consuma esa cantidad (19-N, España: ABC [‘scientists don’t recommend that such an amount be ingested’]) (b) They recommend that the sealed sections of the hull containing the reactor compartments should be buried in an open trench in an arid or semi-arid environment (B7G, W_non_ac_nat_science) (25) (a) La administración municipal ordenó la suspensión de toda actividad recreativa (19-N, Col:Semana:840 [‘the municipal administration ordered the cancellation of all leisure activities’]) (b) Two months ago the government ordered the removal of all red stars from public buildings (A7V, W_newsp_brdsht_nat_report) We have just seen that non-finite reported propositions in Spanish and English contrast in that the former are infinitival while the latter are gerundival. On the other hand, both languages coincide in choosing the infinitive when the non-finite report is introduced by an interrogative adverb, as is the case in (26a, b). Notice that these wh- projections are in fact hard to categorise concerning the proposition vs proposal divide. Although they are included here among proposals, interpretable as suggestions about how to make fishing baits and rods (26a), and how to plant seeds (26b), there could be grounds to interpret them as statements of how these activities are performed. In the latter case they would be propositions rather than proposals. The report as suggestion, or even command, is more foregrounded in those cases when, even if there is a sense of applicability to anyone in general, it is clearly targeted to someone in particular. Such is the case in (27a), where the advice is clearly geared to the Argonauts thanks to the presence of the deictic clitic les referring to them as Addressees of the message. Likewise, in the English example (27b), the reference to the first person singular, both as the Addressee in the projecting clause and as the Subject of the conditional clause if I am in pain, counterbalances the otherwise generally applicable suggestion what to do. (26) (a) El libro contaba cómo hacer cebos y cañas para pescar (19-AC Enc: Pesca deportiva [‘the book told (about) how to make fishing baits and rods’]) (b) He explained how to plant the seeds (FS2, W-fict-prose) (27) (a) En agradecimiento, Fineo les dijo cómo pasar a través de las Simplégades, las rocas que guardaban la entrada al mar Euxino (19-AC, Enc: Argonautas [‘in gratitude to them, Phineas told them how to sail though the Symplegades, the rocks guarding the entrance to the Euxine sea’] (b) They’ve told me what to do if I am in pain (K1C, W_news_script)


Jorge Arús Hita

Although projection at the clause nexus level concerns mental and verbal processes, the discussion so far has mainly relied on verbal examples, as many of the nuances associated with quoting and reporting are more easily observed in this kind of processes. Thus, whereas both quoting and reporting abound in verbal processes, mental processes are likelier to project a report than a quote (see Halliday and Matthiessen, 2014, p. 523), and mental reports tend to be propositions rather than proposals, since offers and commands are typical of conversational exchange. As Halliday and Matthiessen (2014, p. 527) state, mental proposals are restricted to things that can be hoped. It is possible, however, to observe some of the general cross-linguistic contrasts that have been so far discussed in mental processes, too. For instance, examples (28a, b) are mental, and they serve to illustrate the subjunctive vs full infinitive Spanish/ English contrast in reported proposals already discussed and exemplified by means of verbal (21a, b). There is a caveat, however, and it is that both the Spanish and the English processes in (28), although mental: desiderative, are not really being used to refer to the Senser’s wishes but, as is customary with gustaría and would like, to perform a polite request. (28) (a) También me gustaría que me comentara algo sobre sus principales logros o sus principales satisfacciones en . . . en el ejercicio de su profesión (19-OR, Habla Culta: Bogotá: M14 [‘I would also like you to tell me about your main achievements or your satisfactions in your profession’] (b) I would like you to let me go down to the fair (KE3, S_conv) One kind of contrast that is specific to mental processes and again shows how hard it is sometimes to apply the proposition vs proposal distinction to mental processes is the one illustrated in (29a, b). Even though Halliday and Matthiessen claim that “while propositions are thought, proposals are hoped” (2014, p. 527), it is hard to see the projections in either (29a) or (29b) as clear proposals. They share with proposals the fact they refer to non-actualised actions, but they are neither offers nor commands proper. In any case, what seems to be a general pattern with mental: desiderative processes is that reports are projected in the subjunctive in Spanish and in the indicative in English (differently from the indicative/indicative for propositions and subjunctive/infinitive [or thatclause] similarity/contrast seen in verbal processes above). (29) (a) Espero que dentro de poco pueda tocarse regularmente en Navidad o Reyes, especialmente para el público infantil (19-OR, Entrevista (ABC) [‘I hope in the near future it can be regularly played on Christmas Day or Epiphany’] (b) We hope they will resume talks, but we cannot waste much more time (AAE, W_newsp_brdsht_nat_sports)

A Contrastive Description of Projection


Spanish and English show a similar predisposition to include, among projecting (mainly quoting) processes, behavioural ones, which, as is well known, can be quite close to verbal meaning. Thus, the same as a behavioural verb such as moan in (30a), proposed by Halliday and Matthiessen (2014, p. 524), can easily project a quote, a similar process in Spanish may act likewise, as seen in (30b). (30) (a) ‘It’s too dangerous,’ moaned Mcduff (AMB W_fict_prose) (b) “Eran los versos de mi abuelo”, se quejó mi padre sin energía (19-F, Los hombres de a caballo [‘“Those were my Grandfather’s verses”, moaned my father without energy’] Up to this point we have dealt with projection in its most canonical, logical sense. It is, however, possible to find cases of projection within the clause nexus which are not entirely logical but also, to some extent, interpersonal. As pointed out by Matthiessen and Teruya, “In interpersonal manifestations [of projection, only the projected status is represented, related to the process of projecting” (in preparation, p. 11). As will be seen in section 3, interpersonal projection really comes into its own at the clause and group levels. It is, however, possible at the clause nexus level to leave part of the projecting element—more precisely the Speaker or Senser (usually a ‘thinker’)—out of the picture in such a way that the ‘hearsay’ effect proper to interpersonal projection (Matthiessen and Teruya, in preparation, p. 26) is strongly felt. This is typically achieved through a passive verbal or mental process such as those in (31a, b). As we can see, in both of them the writers detach themselves from the responsibility of the claims by presenting such claims as made by someone else yet without making the sources explicit (in these particular cases, because readers can easily make themselves a rough picture of those sources, i.e. scientists). This is the reason, then, why the projections in (31a, b) are logical as well as interpersonal: the claims are explicitly presented as reports of what others expect or think, which is the logical part of the projection, while at the same time introducing a hearsay effect, the interpersonal side of the projection, by not specifying who expects (31a) or thinks (31b) those projections. Notice that the grammar of the projected clause is quite different in each language: whereas Spanish projects a that-clause—in the subjunctive in this case but it could be in the indicative if projected by a different verb—reporting passives in English, such as are thought in (31b), are characterised by projecting a full-infinitive clause. (31) (a) Se espera que los robots reemplacen a los cirujanos en aquellas tareas que exigen mayor precisión y estabilidad que las que la mano y el ojo del hombre pueden garantizar (19-N, España: ABC [‘robots are expected to replace surgeons at those tasks


Jorge Arús Hita demanding a larger precision and stability than the hand and the eye can provide’] (b) The gems are thought to belong to a female member of an Arab royal family (CH2, W_newsp_tabloid)

As a way to conclude the contrast of projection at the clause nexus rank, let us point out two issues which have incidentally appeared in some of the examples analysed so far. One of them has to do with the possibility, in both languages, of having a non-human Sayer which is in some way ‘personified’ to report speech. Such was the case in English (21b) and Spanish (26a), reproduced now as (32a) and (32b), respectively. In both (32a, b) we can see that, as Halliday and Matthiessen state, “when the Sayer is realized by a nominal group denoting a symbol source other than a human speaker . . . the ‘present in’ [i.e. a progressive tense] is unlikely” (2014, p. 305). Thus, both suggested and contaba are in the simple past. It must be noticed, however, that progressive tenses are less used in Spanish than in English, as the aspectual meaning captured by progressive forms in English is often expressed in Spanish by means of the present simple and the imperfect (Zamorano-Mansilla, 2006; Rojo, 1990). This is also reflected in verbal processes, as seen in (33a, b), which illustrate this contrast in the present. (32) (a) One amendment suggested scrapping the council’s 6,250 donation to the cathedral organ restoration fund and spending the money on retraining the unemployed (CFC, W_newsp_other_ report) (b) El libro contaba cómo hacer cebos y cañas para pescar (19-AC Enc: Pesca deportiva [‘the book told (about) how to make fishing baits and rods’]) (33) (a) People are saying that Sydney Burrows has been caught. Is that true? (C8D, W_fict_prose) (b) Dicen por ahí que Cerezo tiene encinta a su mujer (18, Tradiciones peruanas [‘They are saying (lit. ‘they say’) around there that Cerezo’s wife is expecting’] The second issue worth pointing out is that, for once, both languages behave similarly concerning word order in that, when a quote precedes the quoting clause, the latter tends to favour Process ^ Sayer/Subject over Sayer/Subject ^ Process, as seen in (11a, b), reproduced as (34a, b). This relative order or the Subject and the verb—motivated in this case by endweight, i.e. longer elements at the end—is quite normal in Spanish but rather unusual in English.

A Contrastive Description of Projection


(34) (a) “Why didn’t you say so before?” exclaimed the Bookman (AMB, W_fict_prose) (b) ¡Me han robado el reloj!—exclamó el hombre, muy excitado (19 F, Hijo de ladrón [‘They’ve stolen my watch!—exclaimed the man, very upset’])

3. Projection Within Other Lexicogrammatical Domains The discussion of projection within the clause nexus in the previous section has allowed us to take a close look at the workings of logical projection, with a brief reference to the combination of this with interpersonal projection. Moving out of the clause nexus also takes us out of the logical metafunction and into the experiential and interpersonal metafunctions as modes of organisation of projection. As advanced in section 2, the main difference between logical projection and the other modes of projection has to do with the explicitness of the projection. As Matthiessen and Teruya put it, “experiential and interpersonal manifestations constitute reduced versions of the projecting part” (in preparation, p. 11). Thus, whereas in interpersonal projection only the projecting tends to be fully represented, we will see that in experiential projection both the projecting and the projected parts are represented but their relationship is not explicitly manifest. 3.1. Projection at Clause Level At the clause level, projection can be manifested in three major ways: (a) experientially, in an intensive identifying relational clause (Matthiessen and Teruya, in preparation, pp. 18–25; Halliday and Matthiessen, 2014, pp. 540–542); (b) also experientially, as a circumstance of Angle (Matthiessen and Teruya, in preparation, pp. 25–26; Halliday and Matthiessen, 2014, pp. 676–677); and (c) interpersonally, as a modal Adjunct (Matthiessen and Teruya, in preparation, pp. 26–28; Halliday and Matthiessen, 2014, pp. 677–686). Among intensive identifying relational processes, there are, in both English and Spanish, a number of possible realisations involving experiential projection, of which we will focus on two here. One kind concerns those clauses, typically identifying, with the otherwise verbal Process nominalised as the Token of the relational process. Such is the case in (35a, b), where the logically projecting he argues that . . . and respondió que . . . (‘he answered that . . .’] are realised as the Tokens his argument and su respuesta, respectively. The other kind concerns those processes where the Token and the Value are related by means of a verb with a clear verbal (sometimes mental) meaning, thus resulting in a sort of hybrid relational/verbal process. This is illustrated by English (36a), with Token


Jorge Arús Hita

the survey and projected Value that the debate . . . and Spanish (36b), with Token Los estudios and projected Value que quienes comienzan . . . Precisely because of the hybrid nature of these processes, the projection involved looks both logical—from the verbal interpretation—and experiential—from the relational interpretation, with the projecting element embedded as a simple clause participant. However, if we consider that, as Matthiessen and Teruya point out, the apparent projection is ‘pre-projected’ (in preparation, p. 21) rather than actually projected, the relational interpretation should prevail over the verbal one. (35) (a) His argument is that medicine rests on the discovery of natural laws while legal norms derive from decisions, influenced by the actions of lawyers themselves (CMS, W_ac_soc_science) (b) su única respuesta fue que ya iría otro año (19-N, España: ABC [‘his only answer was that he would go some other year’] (36) (a) The survey demonstrates that the debate over membership of the ERM has become clouded by issues of personalities and politics (A9D, W_newsp_brdsht_nat_commerce) (b) Los estudios demuestran que quienes comienzan a fumar desde la adolescencia tienen 50% de posibilidades de morir a causa del tabaco (19-N, Guat:Gerencia:98MAY28 [‘studies show that those starting to smoke in adolescence have 50% chances of dying because of tobacco’] As seen in (35) and (36), English and Spanish behave similarly concerning the construction of experiential projection in relational processes. The examples in (37a, b) in turn show that they also make a similar exploitation of circumstances of Angle to project at the clause level. Both the president of the Can Manufacturing Institute and la portavoz Kerry Fennelly would have been the Sayers had the projections been construed logically by means of verbal clause complexes. (37) (a) According to the president of the Can Manufacturing Institute, Michael Dunn, the country’s waste stream will soon contain almost no aluminium cans if present trends continue (J2U, W_misc) (b) Normalmente se reciben unos 200 al día, según la portavoz Kerry Fennelly (19-N, US: Herald: 98 Sep 19 [‘Around 200 are usually received a day, according to spokesperson Kerry Fennelly’] Turning now to interpersonal projection, (38a, b) include typical comment Adjuncts, i.e. reportedly in English, aparentemente in Spanish, used in each language for speakers to disengage themselves from responsibility for what is said (see also examples [3] and [7] in the introduction), in a

A Contrastive Description of Projection


similar way to what can be achieved by means of logical projection in verbal processes of the type They/people say . . . or It is said . . . Notice that this has been studied from different approaches, notably pragmatics, under a number of different labels, such as ‘hedging’ (e.g. Kaltenböck, Mihatsch and Schneider, 2010), ‘evidentiality’ (e.g. Marín-Arrese, Hassler and Carretero, 2017), ‘stance’ (e.g. Hyland and Feng, 2016; Jaffe, 2012), ‘engagement’ (e.g. Hyland, 2005) and ‘voice’ (e.g. Hyland and Sancho Guinda, 2012). From a perspective closer to SFL, appraisal theory also speaks of ‘engagement’; in particular, the attribution sub-type fulfils a role very similar to that of interpersonal projection (see, e.g., Martin and White, 2005). (38) (a) He is allegedly being held in a secret location in Riyadh where it is feared he is at risk of torture and execution (CJS, W_pop_lore) (b) Hasta unos 35 esqueletos humanos del tipo neandertalense fueron aparentemente depositados en una fosa en este lugar (19 A-C, Enc: Edad de piedra [‘As many as 35 human Neanderthaltype skeletons were apparently laid in a pit in this place’]) Whereas this kind of interpersonal projection is, as seen in (38a, b), possible in both languages, where English resorts to comment Adjuncts Spanish often prefers alternative realisations such as interpersonal metaphors with Parece (ser) que (39) or the use of the conditional (40)—in a similar way to the use of the subjunctive in German (see Holsting, 2008). Notice that in the case of the use of the conditional, the projection does not take place at the clause level, but rather at the group level by the use of the corresponding verbal inflection. (39) Parece ser que ocurrían unos ruidos muy extraños en el edificio7 (19-OR, España Oral: CDOC012B [‘Apparently (lit. ‘it seems that/ looks like’) there were some very strange noises in the building’]) (40) Entretanto, se conocieron versiones de que Pinochet habría sufrido un retroceso en su convalecencia de una operación a una hernia lumbar (19-N, US: Herald: 98Oct22 [‘In the meantime, versions were known that Pinochet had apparently suffered (lit. ‘would have suffered’) a relapse in his convalescence . . .’])

3.2. Projection Within the Group Projection happening at the group level mostly concerns the use of post-determiners in nominal groups, very often related to the comment Adjuncts used at clause level, e.g. allegedly/alleged; reportedly/reported in English or presunto/presuntamente; supuesto/supuestamente in Spanish (see Halliday and Matthiessen, 2014, pp. 680–685 for a list of both


Jorge Arús Hita

modal Adjuncts and post-determiners in English). This kind of projection is therefore also interpersonal. Examples (41a, b) illustrate the use of this resource in English and Spanish. (41) (a) Government sources yesterday denied the reported rift between the Prime Minister and the Chancellor (A4C, W_newsp_brdsht_ nat_misc) (b) Una Comisión de Amnistía Internacional verificará hoy la presunta violación a los derechos humanos en el Chapare de Cochabamba (19-N, Bolivia: ERBOL:04/16/96 [‘an Amnesty International commission will verify today the alleged violation of human rights in Chapare de Cochabamba’]) As seen above, Spanish can realise interpersonal projection at the verbal group level by means of the conditional. Since it has already been discussed a propos its contrast with clause-level interpersonal projection in English, we will not illustrate it again in this section. However, it must be pointed out that whereas Spanish can project interpersonally within both nominal and verbal groups, this is circumscribed in English to the nominal group, as seen.

4. Conclusion This chapter has looked at the similarities and dissimilarities in the construal of projection in English and Spanish at different ranks. The contrastive discussion of projection within the clause nexus has been the richest in this respect due to the large number of variables resulting from the combination of the three major systems of level of projection, mode of projection and speech function. In particular, speech function has been found to account for most of the contrast seen in these pages. Table 9.1 summarises the main contrasts discussed, with a parenthetical reference to the numbers of the examples used to illustrate the discussion. Outside the clause nexus, the overall picture is simpler yet arguably subtler because, as seen, projection is not as explicitly expressed as in the more canonical, logical realisation. Below the clause nexus, the crosslinguistic contrasts found concern the use in Spanish of interpersonal metaphors of the parece (ser) que type in addition to comment Adjuncts, the favourite resource in English, and the use, also in Spanish, of the conditional tense as a mode of projection at the verbal group level. Looking at projection as a fractal pattern has allowed us to ascertain the difference between projection, which is a semantic category, and quoting and reporting, which are lexicogrammatical realisations of projection

A Contrastive Description of Projection


Table 9.1 Main contrasts between English and Spanish projection in the clause nexus English Reported proposition vs reported proposal Non-finite realisation of reported propositions and proposals Arrangement of modal structure in reported questions Nexus in reported wh- questions Reports in mental: desiderative processes Interpersonal projection in clause nexus


Proposition: finite (20b) Proposition: indicative (17a) Proposal: non-finite (17b) Proposal: subjunctive (20a) Gerund: (18b) (23b) Infinitive: (18a) (23a)

Subject ^ Finite (19a)

Finite ^ Subject (19b)

No that (19a) Indicative (29b)

Possibility of que before wh- (19b) subjunctive (29a)

Non-finite (31b)

Finite (31a)

at the clause complex level. Because projection is usually considered at the clause nexus rank, there is the risk of equating projection with quoting and reporting, when in fact these are only possible realisations of projection. As we have seen, other possible realisations are by means of comment Adjuncts or circumstances of Angle in a simple clause, postdeterminers in the nominal group, tense (only in Spanish) at the verbal group level, and specific relational constructions. The journey through English and Spanish projection has proven quite fruitful not only to compare the two languages in this respect but also, hopefully, to gain further understanding of the multiple nuances of this semantic concept and its lexicogrammatical realisations. For future work remain the examination of projection at yet other levels than the ones covered here, for instance, how projection is marked phonetically; or the body language associated with projection (let us just think of how we use our fingers to indicate quote marks); not to mention the ways in which projection is signalled beyond the clause nexus, at discourse level. All this certainly suggests that there still is a lot to be learned about this semantic category and the way it manifests itself in different languages.

Notes 1. The research leading to this chapter has been possible thanks to funding by the Hong Kong Polytechnic University through the research project Projection around the World: Clause Complex and Rhetorical Relation.


Jorge Arús Hita

2. See, for instance, the description of projection systems in different languages in Caffarel, Martin and Matthiessen (2004). 3. All examples in this chapter are from Brigham Young University-Brithish National Corpus (Davies, 2004–), for English, and Corpus del Español (Davies, 2002–), for Spanish. 4. Note that there is no reason to assume the same parallelism in all languages. In Dagaare, for instance, reported clauses are free clauses, and quotes may be hypotactically related to the projecting clause (Mwinlaaru, 2015). Arabic may also quote hypotactically and report paratactically (Bardi, 2015). 5. This kind of contrast with English has also been pointed out for other languages; see, for instance, Thoma (2008) about projection in English and Greek. 6. Nominalisation plays an important role in projection, not only in verbal and mental processes but also in relational processes (see section 3). 7. Parece (ser) que constructions express evidential meaning in a very similar fashion to objective explicit metaphorical expressions of probability in English, such as it’s likely Mary knows (in Halliday and Matthiessen, 2014, p. 689). That is why they are considered interpersonal metaphors here.

References Arús Hita, J. (2007) Hacia una especificación computacional de la transitividad en el español: Estudio contrastivo con el inglés (Towards a Computational Specification of Transitivity in Spanish: A Contrastive Study With English). PhD thesis, Universidad Complutense de Madrid. Arús Hita, J. (2008) Ideational and interpersonal manifestations of projection in Spanish. In C. Wu, C. M. I. M. Matthiessen and M. Herke (Eds.), Proceedings of ISFC 35: Voices Around the World, pp. 195–200. Arús Hita, J. (2015) Projection as a Fractal Motif in Spanish. [Presentation within Teruya (2015)], Hammamet, Tunisia, 27th March. Bardi, M. A. (2015) Projection as a Fractal Motif in Arabic. [Presentation within Teruya (2015)], Hammamet, Tunisia, 27th March. Caffarel, A., Martin, J. R. and Matthiessen, C. M. I. M. (2004) Language Typology: A Functional Perspective. Amsterdam: Benjamins. Davies, M. (2002–) Corpus del español: 100 Million Words, 1200s-1900s. Available at: (Accessed on 22 May 2017). Davies, M. (2004–). BYU-BNC (Based on the British National Corpus from Oxford University Press). Available at: (Accessed on 22 May 2017). Halliday, M. A. K. (1985) An Introduction to Functional Grammar, 2nd ed. London: Routledge. Halliday, M. A. K. (1994) An Introduction to Functional Grammar. London: Routledge. Halliday, M. A. K. and Matthiessen, C. M. I. M. (2004) Halliday’s Introduction to Functional Grammar, 2nd ed. London: Routledge. Halliday, M. A. K. and Matthiessen, C. M. I. M. (2009) Construing Experience Through Meaning: A Language-based Approach to Cognition. London: Continuum. Halliday, M. A. K. and Matthiessen, C. M. I. M. (2014) Halliday’s Introduction to Functional Grammar, 4th ed. London: Routledge. Holsting, A. (2008) Projecting clause complexes and the subjunctive mood as means of projection in German. In N. Nørgaard (Ed.), Systemic Functional

A Contrastive Description of Projection


Linguistics in Use. Odense Working Papers in Language and Communication, vol. 29, pp. 381–399. Hyland, K. (2005) Stance and engagement: A model of interaction in academic discourse. Discourse Studies, 7(2), pp. 173–192. Hyland, K. and Feng, J. (2016) Change of attitude? A diachronic study of stance. Written Communication, 33(3), pp. 251–274. Hyland, K. and Sancho Guinda, C. (2012) Stance and Voice in Written Academic Genres. London: Palgrave Macmillan. Jaffe, A. (2012) Stance: Sociolinguistic Perspectives. New York: Oxford University Press. Kaltenböck, G., Mihatsch, W. and Schneider, S. (2010) New Approaches to Hedging. Bingley: Emerald Publishing. Lavid, J., Arús Hita, J. and Zamorano-Mansilla, J. R. (2010) Systemic Functional Grammar of Spanish: A Contrastive Study With English. London: Continuum. Maldonado González, C. (1999) Discurso directo y discurso indirecto. In I. Bosque and V. Demonte (Eds.), Gramática descriptiva de la lengua española. Madrid: Espasa-Calpe, pp. 3551–3595. Marín-Arrese, J. I., Hassler, G. and Carretero, M. (Eds.). (2017) Evidentiality Revisited: Cognitive Grammar, Functional and Discourse-Pragmatic Perspectives. Amsterdam: Benjamins. Martin, J. R. and White, P. R. R. (2005) The Language of Evaluation: Appraisal in English. London: Palgrave Macmillan. Matthiessen, C. M. I. M. (1995) Lexicogrammatical Cartography: English Systems. Tokyo: International Language Science Publishers. Matthiessen, C. M. I. M. and Teruya, K. (in preparation) Projection as a fractal motif: Semantic and lexicogrammatical manifestations. Mwinlaaru, I. (2015) Projection as a Fractal Motif in Dagaare. [Presentation within Teruya (2015)], Hammamet, Tunisia, 27th March. Rojo, G. (1990) Relaciones entre temporalidad y aspecto en el verbo español. In I. Bosque (Ed.), Tiempo y aspecto en español. Madrid: Cátedra, pp. 17–43. Teruya, K. (2015) Projection as a fractal motif: Multilingual studies of Arabic, Chinese, Dagaare, English, Japanese, Spanish and Thai. Colloquium held at the First Tunisian International Systemic Functional Linguistics Conference and Workshop, Hammamet, Tunisia, 26–28 March. Thoma, C. (2008) Projection in modern Greek (In Spanish). In C. Wu, C. M. I. M. Matthiessen and M. Herke (Eds.), Proceedings of ISFC 35: Voices Around the World, pp. 225–228. Zamorano-Mansilla, J. R. (2006) Stative situations and the progressive construction (be/estar + gerund) in English and Spanish: A contrastive corpus-based analysis. In M. Carretero, L. Hidalgo Downing, J. Lavid, E. Martínez Caro, J. Neff, S. Pérez de Ayala and E. Sánchez Pardo (Eds.), A Pleasure of Life in Words: A Festschrift for Angela Downing. Madrid: Universidad Complutense de Madrid, pp. 377–399.

10 Towards a Systemic Functional Contrastive Analysis of Japanese and English Corporate Legal Discourse Sonya Chik 1. Introduction Language comparison, whether in the form of linguistic typology that is concerned with “parts of the linguistic systems in many languages” or translation studies that focuses on “the recreation of instances of meaning identified in one language in another language” (Teruya and Matthiessen, 2015: 428), has received considerable attention, especially in the research areas of second/foreign language (SL/FL) learning, multilingual text generation, multilingual typology and translation studies. The present chapter is concerned with contrastive analysis, a form of linguistic comparison in between translation studies and linguistic typology, which focuses on the variation among systems in different languages based on the analysis of texts within one or a small set of registers (Teruya and Matthiessen, 2015). Originally developed as a method to support SL/ FL learning (Lado, 1957), contrastive linguistics has expanded from the early works of contrastive rhetoric in the late 1960s (Kaplan, 1966), which focused on the analysis of paragraph organisation in English as a Second Language (ESL) student essay writing, to an interdisciplinary area of applied linguistics in other genres such as business reports, journal articles, editorials, etc. (Connor, 2002). Of immediate relevance to this study are the various contrastive works on the analysis of different linguistic aspects of Japanese and English. They include extensive contrastive studies by Hinds (1983, 1990), who analysed the rhetoric of Japanese and English expository writing and found their organising principles to be dramatically different, referring to Japanese text as ‘reader-responsible’ as opposed to ‘writer-responsible’ in English. Focusing on the different grammatical features of Japanese and English, Ikegami (1981, 1991) contends that Japanese is a ‘becomelanguage’, which contrasts sharply with English as a ‘do-language’. This contrastive view suggests that given the same event, Japanese tends to describe the world as a ‘change in state’ conceived as a whole, whereas English tends to describe the event as an action with someone doing something to someone else. More recently, contrastive analysis between

Corporate Legal Discourse


Japanese and English has been conducted from systemic functional (SF) perspectives. For example, Thomson, Fukui and White (2008) analysed Japanese news reports using the appraisal framework and found that White’s (2000) definition of ‘reporter’ voice in English news report also applies to Japanese news stories. With a focus on the systems of MOOD and MODALITY, Mizusawa (2008) compared directives in Japanese and Australian workplace administrative discourse and found that Japanese favours incongruent realisation of command as compared to English. Findings from these previous contrastive studies have shed light on the understanding of the different linguistic features of Japanese in comparison with other languages in different contexts of situation. However, despite growing interest in cross-linguistic research in professional discourse, especially in the areas of multilingual text production and translation studies (e.g. Kranich, 2016), contrastive analysis in Japanese and English in language for specific purposes (LSP) has remained under-researched. Adopting a holistic approach to text in context, this study draws on the SF perspective to explore cross-linguistic variation between Japanese and English in corporate legal discourse. Privacy policies, written corporate legal documents that are publicly available on major organisations’ websites, have been selected as the object of research for (1) their value as a professional discourse and applicability in multilingual text production and translation studies in LSP, and (2) their representativeness of the ‘enabling: regulating’ text type/register. Using a combination of qualitative and quantitative analysis, this study explores the variation in linguistic choices that construe meaning in Japanese and English from the perspectives of context, semantics and lexicogrammar. Specifically, it aims to investigate three research questions: (1) What are the commonalities and divergences in the construal of ideational meaning in Japanese and English written corporate legal discourse?; (2) to what extent are the divergences in linguistic features structurally and registerially motivated?; and (3) what are some of the benefits of adopting an SF approach to contrastive analysis? The answers to these questions will not only reveal salient features of corporate legal discourse instantiated in privacy policies but also shed light on typological and registerial characteristics of Japanese as compared to English in the given text type/ register. The rest of the chapter is organised as follows. First, I will briefly review the key features of Systemic Functional Linguistics (SFL) theory and its appliability to multilingual and translation studies. Second, I will present a case study to demonstrate how contrastive analysis using the SF approach can benefit from a holistic perspective on language description that operates at a high degree of semiotic abstraction. Third, I will discuss the cross-linguistic variation between Japanese and English in the construal of experiential meaning under the given social context. Finally,


Sonya Chik

I will conclude the chapter by summarising the findings and proposing possible directions for future research.

2. Theoretical Framework—SFL Theory SFL theory is a comprehensive theoretical framework for analysing text in context. It is designed in such a way that it is appliable to describe any language. Halliday himself started applying SFL theory to the study of Mandarin Chinese in the 1950s (see Halliday, 1956, 1959). However, it was not until the 1990s and 2000s that a multitude of SF research on different languages started to sprout, e.g. German (Teich, 1999; Steiner and Teich, 2004), Tagalog (Martin, 2004), Thai (Patpong, 2006), French (Caffarel, 2004), Japanese (Teruya, 2004; 2007) and Spanish (Lavid, Arús and Zamorano-Mansilla, 2009). These important works on SF description of different languages laid the foundation on which new ways of approaching contrastive linguistics and translation studies are based. Under SFL theory, language is theorised as a semiotic system for making meaning. It is comprehensive in the sense that “it is concerned with language in its entirety”. In other words, any one aspect of the language is always interpreted with reference to the total picture and essentially plays a part that contributes to the whole (Halliday and Matthiessen, 2014, p. 20). In contrastive analysis, it is important to ground the analysis on a theoretical framework that is able to cover the known parameters of cross-linguistic variation, including function and form, as well as system and text (Teich, 2003, p. 29). The SF framework also enables a multidimensional view of text from the context, semantics and lexicogrammar under the notion of stratification, where language is considered as a stratified resource organised in strata that are differentiated according to level of abstraction (Halliday and Matthiessen, 2014, pp. 24–27). These strata are related by means of realisation. At the top level is the context stratum, which consists of contextual parameters of field, tenor and mode, which correspond with the respective metafunctions, or functionalities that are intrinsic to language (see Halliday and Hasan, 1989, p. 26). Field, or what’s going on in the situation, corresponds with the ideational metafunction that construes human experience; tenor, or who is taking part in the situation, corresponds with the interpersonal metafunction that enacts interpersonal relations; and mode, or what role is being played by language in the situation, corresponds with the textual metafunction that facilitates text organisation and creates cohesion in the text (Halliday and Hasan, 1989, pp. 13–14). In theory, if the text is grounded in contextual considerations, semantic and lexicogrammatical considerations will align themselves with the contextual ones (Halliday and Matthiessen, 2014, p. 33). In terms of stratification and metafunction, the present study focuses on the field variable of corporate legal discourse and aims to investigate the variation in linguistic resources selected from the

Corporate Legal Discourse


ideational metafunction to construe meaning in Japanese and English in the strata of context, semantics and lexicogrammar. Another dimension of language that interacts with stratification is instantiation, or system as instantiated in the form of text. Closely related to the notion of instantiation is the notion of register, or a functional variety of language (Halliday, McIntosh and Strevens, 1964; Halliday, 1978)—the patterns of instantiation of the overall system associated with a given type of context (a situation type). In principle, every situation type has an associated register based on evidence gathered from the various contexts of situation operating within that institution (cf. Hasan, 2009, pp. 172–173). Within the broad category of register, there are eight primary types of socio-semiotic processes that can be further grouped under semiotic processes where language constitutes the context (e.g. expounding in a history textbook, reporting in news reports, recreating in novels, sharing in casual conversation and exploring in editorials), semiotic processes that lead to social processes (e.g. recommending in advertisements and enabling in regulations and guidelines) and social processes where language facilitates the doing process (e.g. doing in service encounters; also see Matthiessen, 2015, for a full description of the eight types of socio-semiotic processes). In this study, I considered privacy policies as an instance of corporate legal discourse, which operates under the sociosemiotic process of ‘enabling’. Its primary field of activity is to regulate social behaviour relating to the management and handling of privacy. By analysing instances of privacy policies, this study attempts to bring to light the commonalities and divergences between Japanese and English in realising the situational meaning potential in regulating social behaviour under corporate communication. The findings from this study will provide grammatical evidence that characterises this text type as such. Examples of text-based comparative studies of languages within particular registers in European languages include Teich (2003) on English and German scientific writing, where she identifies and confirms various typological and registerial features in both languages using a combination of corpus-based and computational analysis rooted in SFL, and Lavid (2000) on English, German and Italian administrative forms, which offers intriguing findings on some variation in the three languages in terms of the relations between requestive strategies realised in Mood and the socio-cultural contexts in where the forms are produced. Focusing on units of meaning and their relation to the context of situation in which the text operates rather than on isolated syntactic forms, the SF approach to contrastive analysis takes a socio-semiotic orientation that aims at contributing to the overall linguistic research with application to areas such as translation studies and multilingual text production, providing tools that not only identify similarities and differences across languages based on registers, but also offer an SF explanation of why a text means what it does in different languages. The current study adopts


Sonya Chik

the SF approach to contrastive analysis by analysing a small sample of texts of one register/text type to draw inferences about variation between the language systems in Japanese and English. It aims to contribute to the current pool of SF contrastive studies by extending the comparison between English and European languages to Asian languages in the LSP genre.

3. Data and Method 3.1. The Corpus A specialised corpus that comprises thirty privacy policies (fifteen English and fifteen Japanese) extracted from the social media websites and corporate websites of top multinational companies and Japanese corporations was compiled for this study (Interbrand, 2014; Interbrand Japan, 2014). The corpus is further divided into two sub-corpora: (1) parallel texts from social media websites consisting of Japanese and English source texts (STs) and their respective target texts (TTs),1 and (2) comparable texts, which are STs for each language from top global brands and Japanese local brands in different industry sectors including telecommunications, the automobile industry, retail, finance, internet services, food and beverages, and diversified and personal care (see Table 10.9 in the Appendix). The full corpus consists of 121,021 words (41,406 words in English and 79,615 words in Japanese).2 3.2. Methodology The analysis is divided into two parts. In part one (section 4), the data is examined ‘from above’ by systematically examining the contextual environment in which the privacy policy operates using the contextual parameters of field, tenor and mode. Drawing on Hasan’s Generic Structure Potential, or GSP, (Hasan, 1984/1996), the generic structure of the privacy policy is explored at the semantic level to identify the total range of textual structures available within the privacy policy as a text type, including obligatory and optional elements that make up the structure of any instance of a privacy policy. The GSP is then mapped against the corresponding rhetorical structure using the semantic system of Rhetorical Relations developed by Matthiessen (1995, 2002, forthcoming), a modified version of the classical Rhetorical Structure Theory (RST; e.g. Mann and Thompson, 1988; Mann, Matthiessen and Thompson, 1992) that is systematised and integrated into the SFL theory as a part of the holistic account. In part two (section 5), divergences between Japanese and English in the construal of ideational meaning are explored ‘from below’ at the lexicogrammatical level, using a pair of English-Japanese privacy

Corporate Legal Discourse


policies from Translations are used as the object of study as they are “a valuable source for investigating cross-linguistic differences in the usage conditions of comparable grammatical means in a source and a target language” (Teich, 2003, p. 17). First, the experiential configuration of both the ST and the TT is examined and compared from transitive and ergative perspectives. Second, variation in the lexicogrammatical realisation of the semantic strategies pertaining to the major communicative functions of privacy policies are identified and categorised. Finally, salient features are compared and verified with those drawn from both sub-corpora of parallel and comparable texts. The aim is (1) to identify variation in semantic choices in fulfilling the discourse function in Japanese and English as reflected in the lexicogrammatical realisation of the semantic patterns, and (2) to identify the motivating factors such as registerial constraints and communicative intention (cf. Teich, 1995) behind such variation as reflected in the translation shifts.

4. Contextual Analysis and the Generic Structure of Privacy Policies 4.1. Contextual Analysis Under SFL, any situation type can be characterised in terms of the three contextual variables of field, tenor and mode. The experiential domain under the field of discourse in a privacy policy is primarily ‘regulating’, supported by ‘explaining’. It is ‘regulating’ in the sense that both speaker and addressee are expected to abide by the rules and regulations as stated in the privacy policy, and ‘explaining’ in the sense that the scope and boundary of the right to privacy are described and supported by the reasons, purposes and conditions under which privacy is handled. In terms of tenor, it is characterised by different roles for speaker and addressee, including institutional roles (service provider and customer), status roles (expert and layperson) and speech roles (regulator and regulated), etc. In terms of mode, the privacy policy is a formal written monologic text distributed through the online channel with language playing the central role in constituting the situation as well as facilitating actions to be performed by the addressee. The combination of these three contextual variables determines the different uses of language, in this case the language used in explaining and regulating privacy in the privacy policy. The contextual variables resonate with the semantic system in the first instance; they also penetrate the lexicogrammar (Halliday and Matthiessen, 2014, pp. 34–35). In section 4.2, I will discuss how the rhetorical organisation of the privacy policy correlates with the field of activities in the contextual environment of the privacy policy.


Sonya Chik

4.2. The Generic Structure and Rhetorical Relations of Privacy Policies In this study, the full corpus is analysed for the GSP of the privacy policy. Corresponding to the field of activities in the privacy policy, which are mainly ‘explaining’ and ‘regulating’, a generic structure is identified following Hasan’s GSP (1984/1996). It is observed that the most common generic organising pattern of privacy policies is typified by a generic structure of Background ^ General Statement ^ Description with a corresponding nucleus-satellite rhetorical structure. Figure 10.1 presents a schematic illustration of the global organisation of a representative example of an English privacy policy ( in complementary views from GSP and the semantic system of Rhetorical Relations. First, a Background is presented to set the scene for the General Statement: This Privacy Policy describes how and when Twitter collects, uses and shares your information when you use our Services, which introduces the subject matter of the privacy policy and is realised by a Global Nucleus expanding through the rhetorical relation of ‘elaboration’ to the Description stage. The Description stage is further expanded through a Global Satellite that comprises the remainder of the text, organised in multi-nuclear ‘additive’ series of text segments.

Figure 10.1 Global organisation of a privacy policy under the GSP and the semantic system of RHETORICAL RELATIONS (Twitter privacy policy)

Corporate Legal Discourse


Although the GSP/RST is typical in most privacy policies (twenty-two out of a total of thirty in the full corpus), there are five instances that do not have the Background, and three instances that only have the Description stage. As a result, the GSP of the privacy policy is further refined to [Background] ^ [General Statement] ^ Description, where both [Background] and [General Statement] become optional. Moreover, under the Description stage of the GSP, four obligatory components that are concerned with using, collecting, sharing and managing personal information are identified. It is not surprising that the GSP of [Background] ^ [General Statement] ^ Description is consistent across languages between English and Japanese as the privacy policy is becoming standardised as a global legal document under which conventions are to be observed across countries and cultures (cf. Bhatia, Candlin, Engberg and Trosborg, 2003). However, divergences across languages between English and Japanese are identified in the lexicogrammatical realisation of the generic global structure of the discourse. For example, English tends to organise the General Statement and Description in a taxonomical development with explicit lexical cohesion; e.g. verbs in the General Statement such as collect and use correspond to nominal phrases such as collection and use as section heads in the Description stage. By contrast, Japanese favours a General Statement that highlights the service provider’s commitment to legal compliance, which contains lexical items such as kojin joohoo hogo ‘privacy protection’, sengen-suru ‘declare’ and hooreijunshu ‘compliance’, with little direct lexical reference to the lexical components that make up the Description stage.

5. Contrastive Analysis—A Case Study A case study of a pair of privacy policies consisting of an English ST and a Japanese TT from Twitter is conducted to investigate divergences and commonalities in ideational meaning between the two languages, which, as a general rule, carries the highest value in terms of ‘translation equivalence’ (Halliday, 2001, p. 16). The focus of the ideational analysis will be on the experiential configuration of events that contribute to the explanation of the four obligatory components: using, collecting, sharing and managing of personal information under the Description stage of the GSP identified in the previous section. Variation in terms of shifts in translation is analysed and interpreted systematically, with an attempt to identify the motivating factors behind the shifts. The results are first compared with another pair of parallel texts consisting of a Japanese original and its English translation from Mixi, a Japanese social media website, and is further verified with sample texts from the comparable sub-corpus.


Sonya Chik

5.1. Lexicogrammatical Analysis—Experiential Configuration At the lexicogrammatical level, the sample text pair from is examined under the system of TRANSITIVITY, the grammatical system for construing (interpreting and representing) human experience, under which each clause represents a quantum of change modelled as a figure. From a transitive perspective, a figure consists of a nuclear configuration that is particularised for each category of process (doing, being, sensing and saying) that unfolds through time and participants that are directly involved in the process in some way, and optional circumstances of time, space, cause, manner, etc. (see Halliday and Matthiessen, 1999). If examined under the complementary ergative model, which is generalised across process types, a figure is interpreted from a nuclear perspective with a focus on causation, i.e. Agency. The nuclear configuration consists of Process + Medium, the participant through which the process comes into existence. A figure is middle when it is self-engendered—that is, the process is actualised by the Medium without any external Agent—and is effective when the actualisation of the process is caused by an Agent. Using a pair of privacy policies from, I first examined the experiential configuration of the English ST and compared it with the Japanese TT from an ergative perspective. It is observed that experience construed in the sample text pair is primarily effective (78% for the English ST and 68% for the Japanese TT); that is, the majority of the events are activated by external Agent(s). However, the Japanese TT has a higher proportion of middle clauses than does the English ST (32% vs 22%), indicating a divergence in experiential configuration between the two versions, a detailed explanation of which can be postulated through a complementary interpretation from a more specific transitive perspective focusing on process types. Under the transitive model, the figure of doing is further categorised into material and behavioural processes, sensing and saying into mental and verbal processes, and being into relational and existential processes. The frequency distribution of process type in both the English ST and the Japanese TT of the Twitter privacy policy is presented in Table 10.1. From a transitive perspective, the sample privacy policy is dominated by figures of doing in both the English ST (80%) and the Japanese TT (71.7%). However, it is worth noting that the Japanese TT has a significantly higher percentage of relational clauses (18.9% vs 8.2% in the English ST), which aligns with the high percentage of middle clauses (32% in the Japanese TT vs 22% in the English ST) from an ergative perspective, indicating that some figures of doing in English may be shifted to figures of being in Japanese during the translation process. A similar frequency distribution is observed in the pair of parallel texts from the Mixi privacy policy, where Japanese is the original (66% effective) and

Corporate Legal Discourse


Table 10.1 Frequency distribution of PROCESS TYPE in the Twitter privacy policy, by number and percentage (Japanese n = 180 clauses and English n = 246 clauses) Figure


English ST

Japanese TT

Doing Sensing/saying

material mental verbal relational existential

197 26 2 20 1

129 15 0 34 2


80.0% 10.6% 0.8% 8.2% 0.4%

71.7% 8.3% 0% 18.9% 1.1%

the English a translation (68% effective), suggesting that privacy policies in general are dominated by effective clauses, with Japanese favouring relational processes (14%) compared to English (4%). To investigate the experiential divergence further, I examine the participant roles in relation to process type and agency from both transitive and ergative perspectives. The privacy policy is a written document that operates in a social exchange between the service provider (speaker) and the individual user (addressee). Its purpose is to construe knowledge about privacy and regulate social behaviour pertaining to the right to privacy under the context in which personal information is given in exchange for products and services. During the exchange, four main participants are involved: (1) ‘privacy’ in the form of personal information realised by nominal groups such as ‘personal information’ or ‘contact information’, (2) the ‘service provider’, or the speaker realised by the pronoun ‘we’ or the name of the service provider, (3) the ‘user’ or addressee, realised by the pronoun ‘you’ or the noun ‘user’ or ‘customer’, and (4) a ‘third party’, realised as a nominal group. Privacy is usually construed as an abstract thing in the form of personal information and data, assuming the nucleus participant role of Medium from the ergative perspective, which corresponds with the Goal that is being impacted by the Actor’s performance of the Process from a transitive perspective. Examples 1–3 (Tables 10.2–10.7) are presented from both transitive and ergative perspectives to give a complementary view of the experiential configuration of events related to using, collecting, sharing and managing personal information. If viewed from a general ergative perspective, the emphasis is on whether the external Agent is present. In contrast, if viewed from a specific transitive perspective, the focus of attention is brought towards the semantic type of process involved, and the participant roles such as Actor and Goal in the material process. Using translation shifts as a point of departure to investigate variation in language use in Japanese and English privacy policies, I compare the findings with data from the respective STs and identify salient features that are representative of both the language system and the register of the language under


Sonya Chik

scrutiny. Three types of translation shift are identified and explained in detail with illustrating examples in the next section. 5.2. Translation Shifts From the SF Perspective 5.2.1. Same AGENCY: Shift in PROCESS TYPE Consider the English-Japanese parallel text pair from the Twitter privacy policy in Examples 1a and 1b. From an ergative perspective, causation is highlighted in both the English ST and the Japanese TT as both are effective clauses with the speaker assuming the Agent role that brings about some change to the Medium your contact information. However, in the English ST, the speaker is the external Agent that brings about change to the Medium by directly acting on it in a material process. By contrast, in the Japanese TT, although the speaker also assumes the Agent role, the clause is realised in a relational process of the ‘assigned’ type. In relational clauses of the ‘assigned’ type, “the relationship of ascription or identity is represented as being brought into existence by an external Agent”

Table 10.2 Example 1a—English ST (effective: material) We

may use

your contact information









“We may use your contact information”

Table 10.3 Example 1b—Japanese TT (effective: ‘attributing’ relational) Ø Twitter wa

gorenraku saki joohoo o

shiyoo suru mono to


Ø Twitter WA

p:contact information (honorific: respect) O

use-inf NOM TO


nom group

noun phrase


Ø Senser



Ø Attributor



Process: relational


Ø Agent




verbal gp Process: mental

“(Twitter) considers (your) respectful contact information something that we use.” (author’s translation)

Corporate Legal Discourse


and can be achieved through mental projection, in which case the Attributor is like the Senser of a mental clause, as illustrated in Example 1b, allowing interpretation from complementary perspectives of mental and relational clauses (Teruya, 2007, pp. 274–275). In this case, an external Agent or Attributor Twitter, although implicit, is present to attribute some quality, shiyoo suru mono ‘something that we use’, to the Carrier, gorenraku saki joohoo ‘contact information’. In this experiential construction, the material process use is nominalised to become a quality through the nominaliser mono, which turns the verb process shiyoo suru ‘use’ into a noun phrase shiyoo suru mono ‘thing that we use’ that can be attributed as a quality to the Carrier gorenraku saki joohoo ‘contact information’. The speaker thus becomes the external Agent that changes the value or attribute of the Medium your contact information. By selecting this type of relational process, the experiential configuration is shifted from a figure of doing in English to a figure of being in Japanese. This shift is significant in the sense that (1) it is selected repeatedly in the sample Japanese TT (fifteen instances), (2) it sets a distance between the actual ‘performing’ of actions that are a potential threat to the addressee’s privacy right and gives power to the speaker to legitimise those actions through attributing values to the nominalised action that functions as an Attribute in a relational process, and (3) it serves as a salient feature that characterises the privacy policy as a legal text type. Other material processes that are nominalised through the lexical item mono to construe the attributing relational process include hozonsuru ‘store’, shuushuusuru ‘collect’ and shiyoosuru ‘use’. The complementary transitive and ergative perspectives are particularly useful in allowing an accurate specification in contrasting how semantically related processes behave differently in the two languages (Lavid and Arús Hita, 2002). This specific feature of a “relational process of the assigned type” also exists in the Japanese ST of the Mixi privacy policy. It is further verified in the rest of the Japanese texts in both the parallel and comparable sub-corpora and is found to be a salient feature (thirty-five instances) that characterises the privacy policy as a sub-type of the legal register. 5.2.2. Same PROCESS TYPE: Shift in Orientation Another common type of shift identified in the sample pair of parallel texts concerns the shift in orientation from the addressee in the English ST to the speaker in the Japanese TT. In the privacy policy, events that are related to the collection of personal information can be construed either from the speaker’s perspective through assigning the Agent role to the speaker, e.g. We collect/receive, etc., or from the addressee’s perspective through assigning the Agent role to the addressee, e.g. you provide/upload, etc. (Example 2a).


Sonya Chik

Table 10.4 Example 2a You


some personal information, such as your name, username, password, and email address.









“You provide some personal information, such as your name, username, password, and email address.”

Table 10.5 Example 2b Ø User ni

Ø Twitter WA

Ø User NI personal information such as name, username, password, and email address O

Transitive Ø Actor


go teikyoo shimei, yuuzaa-mei, pasuwaado, meeruadoresu nado no kojin joohoo o

Ø Twitter wa

Ø Cir: source


p.provide fml (honorific- (honorificrespect) humble)

nom group




Ø receiver Ø giver



Ø Agent



auxiliary: receive

“(Twitter) receives (your) providing some personal information, such as name, username, password, and email address.” (author’s translation: Japanese ‘speaker’ perspective in receiving orientation)

By assigning the Agent role to the addressee in a material clause, as in Example 2a, the addressee is directly involved as the ‘doer’ in the events and is assigned the responsibility for initiating the ‘giving’ of personal information. Two important observations are made with regard to assigning the Actor/Agent role to the addressee. First, assigning the Agent role to the addressee as a nuclear participant in a material clause is only common in English. In privacy policies in Japanese, regardless of whether Japanese is the source or the target language (as verified by the data from the Mixi privacy policy, where Japanese is the original, with an English translation), the addressee is left implicit and seldom assigned the Actor/

Corporate Legal Discourse


Agent role; that is, most events are construed with the speaker as the Actor/Agent. Second, unlike in the English ST, where the addressee is directly assigned the responsibility of performing the material process of ‘giving’ information through the use of second person pronoun you as the Actor (as in Example 2a), in the Japanese TT the orientation of ‘receiving’ is favoured, allowing the construal of the experience from the speaker’s perspective even in events where the addressee is the ‘doer’ who actualises the process (see Example 2b). In Japanese, the general verbs of ‘giving’ and ‘receiving’ that serve as verb honorifics in Japanese construe social status and distance between interlocutors (Hiraga, 1999). In Example 2b, the verb itadaku, an honorific form of morau ‘receive’ is attached to the verb teikyoo-suru ‘provide’, construing the experience from the speaker’s perspective in a receiving orientation. Moreover, the social hierarchical status between the speaker/ service provider (inferior) and addressee/customer (superior) is grammaticalized through the verb marked with the honorific prefix go in goteikyoo “provide” (honorific-‘respect’) itadakimasu “receive” (honorific auxiliary denoting ‘humble’), allowing both speaker and addressee to be elided in the clause. This is an example of the construction of social roles and hierarchy of interpersonal meaning through experiential grammar, and this is frequently selected in business contexts to mark the social hierarchical status of corporations and customers (cf. Mizusawa, 2008). Verbs that are frequently represented in the receiving orientation marked with honorification and politeness include goriyoo-itadaku ‘use’, ookuri-itadaku ‘send’ and gokoonyuu-itadaku ‘buy’, which are common selections in Japanese privacy policies in both parallel and comparable corpora. It is observed that the preference for speaker orientation in the construal of experience in Japanese privacy policies is consistent across the parallel and comparable corpora, meaning that it is a salient feature rather than a mere translation shift. The motivation is both typological and registerial. It is typological in the sense that the verbs of giving and receiving in their honorific forms in Japanese constrain the speaker to take the recipient’s point of view while signalling social standing and distance between interlocutors (Shibatani, 1990). It is registerial in the sense that the tenor relations that enact the social roles (service provider vs customer) in a commercial situation determine the simultaneous selection of the ‘receiving’ type of material process together with honorification to enact the social hierarchical roles of the interlocutors. By contrast, construing experience that is oriented towards the addressee using a direct way of writing that resembles the spoken style is preferred in the English ST. The motivation can be attributed to the “Plain Language Movement” that started in the 1970s with the aim of making legal writing intelligible to ordinary people by adopting a writing


Sonya Chik

style that is clear and direct, and free from technical jargon and intricate syntactic structure (Trosborg, 1997, p. 141). Another motivation could be attributed to different preferences in managing tenor relations for institution and customer between the English-speaking and Japanese-speaking communities. Unlike the Japanese-speaking community, where social distance and hierarchy are made explicit through grammaticalisation, the English-speaking community prefers a friendly and direct approach by involving both the speaker ‘we’ and addressee ‘you’ as Actors/Agents who participate in events where the social distance is reduced and the hierarchy neutralised to increase audience acceptance (cf. Lavid, 2000). 5.2.3. Shift in AGENCY and PROCESS TYPE Another semantic strategy specific to the English ST in explaining the use, collection and sharing of personal information is the assigning of an additional Agent role using the ‘three-participant causative’ experiential configuration (cf. Halliday and Matthiessen, 2014, p. 579) to legitimise or justify the action. In a ‘three-participant causative’ configuration, a second-order Agent from the ergative perspective that corresponds to the Initiator from the transitive perspective is present, extending the causation to the Actor/first-order Agent through a causative verbal complex such as ‘help . . . (to) do’, ‘enable . . . to do’, ‘allow . . . to do’, etc. Example 3a shows the ‘three-participant causative’ configuration in a material clause in the English ST, which is juxtaposed with the relational clause in the Japanese TT in Example 3b, with contrasting AGENCY between effective in the English ST and middle in the Japanese TT. In Example 3a, the demonstrative pronoun this (the user’s personal information) is assigned the Initiator role that brings about the action performed by the Actor (speaker) to perform an event display ads about things you may have already shown interest in that is presented as beneficial to the addressee. The aim is to increase the addressee’s positive regard towards the idea of ‘providing’ personal information. The Initiator assumes a ‘causing or agentive role’ (Martin, Matthiessen, Painter and Webster, 2010, pp. 108–110) that activates a causative verbal group complex such as help + verb in the material process. This experiential Table 10.6 Example 3a this



to display

ads about things [[you may have already shown interest in]].











Corporate Legal Discourse


Table 10.7 Example 3b kore ni yori,

tsuittaa [[[[yuuzaa ga kanshin wa o shimeshita kotogara ni kansuru]] kookoku o hyooji suru]] koto ga

this NI Twitter YORI WA Transitive Cir: means Ergative

Carrier domain

kanoo ni


display ads that the user has shown interest in NOM GA Carrier

possible NI




Attribute Process: inceptive Process

“Through this, the act of Twitter displaying ads about things the user has shown interest in will become possible” (author’s translation)

configuration is a salient feature in English privacy policies (fourteen instances in the sample English ST and fifty instances in the full corpus) as it serves to legitimise or justify actions of disclosing, sharing and providing personal information. Verbs that are frequently co-selected to form the verbal complex such as help + verb include provide, improve, prevent, ensure, customise, manage, etc. By contrast, the demonstrative pronoun kore ‘this’ in the Japanese TT (Example 4b) is construed as a Circumstance of means marked by the particle ni yori, through which the nominalised act of ‘displaying ads’ (Carrier/Medium) will become ‘possible’ (Attribute/Range). In this case, the PROCESS TYPE is shifted from material to relational, also resulting in a shift in AGENCY from effective to middle. In the Japanese ST, instead of representing the quantum of change being brought about by an additional external Agent in an effective clause, the preference is to construe change in a middle clause through attributing some quality over time in a relational process. The event [. . .] surukookoku o hyooji suru ‘display ads that . . .’ is nominalised through koto, or what is referred to as ‘a rank-shifted nominal clause marker’ in Japanese (see Teruya, 2004, p. 215) to become a ‘thing’ that assumes the Carrier/Medium role. This kind of shift also explains the higher proportion of relational clauses in the Japanese TT. In Examples 1–3, I have discussed three different kinds of translation shifts in Japanese TTs that are identified within the ideational metafunction. These shifts are representative of the variation identified in both the parallel and comparable corpora. While some of the shifts are structurally motivated (e.g. grammaticalisation and nominalisation), others are registerially constrained (e.g. ‘give-&-receive’ speaker orientation in the business register and ‘suru mono to suru’ relational


Sonya Chik

process of the attributing type in the legal register in Japanese, vs the ‘you’ or addressee orientation in English). It is also interesting to learn that cross-linguistic variation between Japanese and English in the ideational meaning of privacy policies is largely attributable to the tenor relations that determine the different speech and social roles of the speaker and addressee, which are affected by the culture and contextual norms. In the English-speaking community, the preference is to construct an equal status between the service provider (speaker) and the user (addressee) by involving the addressee in events as Actor/Agent, while in the Japanese-speaking community, the preference is to keep the social distance between speaker and addressee through experiential construction of hierarchical roles and representing events that are oriented towards the speaker.

6. Concluding Remarks In this chapter, I have presented some findings on a contrastive analysis of corporate legal discourse in Japanese and English from an SF perspective. While the cross-linguistic variation identified in this study is illustrative rather than exhaustive, this chapter has demonstrated how a holistic approach to text in context can enrich the findings on variation across languages using the privacy policy as a representative text type in the socio-semiotic process of enabling in corporate communication. With a focus on the variation in registerial rather than universal features in Japanese and English, I have examined both parallel and comparable texts instantiated in English and Japanese privacy policies in relation to their respective language systems, as well as the relations between function and form under the ideational domain. The results of the analysis reveal that the privacy policy has a GSP of [Background] ^ [General Statement] ^ Description rhetorically organised in a nucleus-satellite structure, which is consistent across languages in English and Japanese. Under the Description stage of the GSP, cross-linguistic variation between Japanese and English is identified as translation shifts within the ideational metafunction as exemplified in a case study using a pair of parallel texts from Three types of shift are observed from complementary ergative and transitive perspectives. First, a shift in PROCESS TYPE from material in the English ST to relational attributing type in the Japanese TT reveals that Japanese tends to represent processes that relate to the collection, use and sharing of personal information as nominalised attributes that are assigned a value through mental projection suru mono to suru,3 a feature that is typical in legal discourse (Tajima, 2005, pp. 49–53). Second, when describing events activated by the addressee, unlike in English, where the Actor/Agent role is assigned to the addressee ‘you’, a speaker orientation is preferred

Corporate Legal Discourse


in Japanese. The speaker is foregrounded as the Actor/Agent who is socially inferior to the addressee through the verb honorifics itadaku, constructing a receiving type of material process that signals the social status and distance of the interlocutors. Finally, in relation to the semantic strategy of ‘legitimising’ events that might otherwise be perceived as posing threats to the addressee’s privacy, English tends to employ a ‘three-participant causative’ verbal group complex, e.g. allow . . . to do, to highlight things such as personal information or the act of collecting personal information as the second-order Agent/Initiator that brings about the action performed by the speaker/addressee that is either legitimate or beneficial to the addressee. By contrast, Japanese tends to represent the causation as Circumstance of means in a relational clause, shifting the AGENCY from effective in the English ST to middle in the Japanese TT. The cross-linguistic variation in semantic strategies and lexicogrammatical features identified in this study provides grammatical evidence for some of the registerial characteristics of the languages in contrast that are specific to privacy policies, a sub-type of LSP under the corporate legal domain. The findings also help to explain the choice of certain linguistic features to fulfil specific discourse functions in the given context of situation, thereby enabling the prediction of lexicogrammatical choices for similar text types, which can be useful in the areas of multilingual text production and translation studies. Table 10.8 presents a summary of the findings of the contrastive analysis of English and Japanese privacy policies from a cross-stratal perspective between semantics and lexicogrammar. This study has demonstrated how a contrastive analysis that takes a ‘trinocular perspective’, to use Halliday’s term (see Matthiessen, Teruya and Lam, 2010, p. 201), from context (field, tenor and mode), semantics and lexicogrammar using a holistic theoretical framework such as the SFL theory enables a rich and deep interpretation of the data by identifying where and why the variation occurs in relation to the respective context and culture. Given the limited space here, I was only able to explore the cross-linguistic variation in the experiential meaning of privacy policies under the ideational domain. The current study may serve as a point of departure towards a fully integrated SF contrastive analysis that includes other metafunctions simultaneously at work in Japanese and English corporate legal discourse, which may shed light on the linguistic challenges facing multilingual corporate communications in the globalised business environment (cf. Böttger, 2004, 2007). The challenge for the analyst remains, however, deciding on which aspects of the SF framework to focus on in order to provide useful results that best characterise both the register and the different languages under study.

Table 10.8 Summary of findings—contrastive analysis of English and Japanese privacy policies Strata



Semantics (full corpus) Generic GSP: [Background] ^ [General GSP: [Background] ^ [General Structure Statement] ^ Description Statement] ^ Description Potential Rhetorical RST: nucleus-satellite RST: nucleus-satellite Relations (global) Lexicogrammar (case study—one pair of parallel texts) Experiential configuration

Translation shifts

ergative: effective 78% > ergative: effective 68% > middle 22% middle 32% transitive: material 80% > transitive: material 71.7% > mental 10.6% > relational relational 18.9% > mental 8.2% > verbal 0.8% > 8.3% > existential 1.1% existential 0.4% Same AGENCY: shift in PROCESS TYPE material relational of ‘assigned’ type e.g. ‘we may use your (mental projection) information’ realised in nominalised embedded clause suru mono to suru (feature of legal register) e.g. ‘we consider your information something that we use’ Same PROCESS TYPE: shift in orientation speaker-oriented addressee-oriented e.g. ‘we receive your providing e.g. ‘you provide some some information’ information’ speaker: Actor (receiver) addressee: Circumstance: source realised in verb honorifics itadaku to construct a ‘receiving type of material process that signals social status and distance (feature of business register) Shift in AGENCY and PROCESS TYPE relational attributing/ ‘three-participant causative’ inceptive—create distance material process—semantic Causation is realised in strategy to legitimise events Circumstance of means and actions that pose threats e.g. ‘through this, [[the display to addressee’s privacy of ads about things that the Causation is assigned to the user has shown interest in]] second-order Agent/Initiator will become possible’ e.g. ‘this allows us to display ads about things that you may have already shown interest in’

Corporate Legal Discourse


Table 10.9 Corpus of the study (n = 30 texts and 121,021 words: 41,406 words in English and 79,615 words in Japanese) Sub-corpus 1: Parallel texts Industry


Social media

LINE Mixi Mobage GREE Facebook Twitter Google WhatsApp

English ST

2,720 2,803 2,420 2,640

Japanese TT

Japanese ST

English TT

4,708 3,205 1,790 10,697

4,100 960 4,707 2,445

7,186 8,397 8,470 N/A

Sub-corpus 2: Comparable texts Industry


Telecommunications Automobile

SoftBank Toyota Mercedes-Benz Mizuho American Express Calbee Coca-Cola Rakuten Amazon Lawson IKEA Toshiba GE Unicharm Gillete

Finance Food and beverages Internet services Retail Diversified Personal care

English ST

Japanese ST 10,343 4,089

2,900 1,900 940 4,870 4,550 5,130 2,780 2,380 2,020 2,610 1,900 3,840 3,521

Appendix Notations and symbols used in this chapter are given in the following. 1. Morphological notations fml NOM p.

formal: desu, shimasu nominaliser: no, mono, koto politeness/honorification: o-, go-; itadaku (honorific) “receive”

2. Systemic notations / ^ [[]]

conflation ordering rank-shifted (embedded) clause


Sonya Chik Senser and other terms with initial capitals: names of functions MOOD and other terms in all upper case: names of systems

3. Transcription In transcribing Japanese, a modified Hepburn style is used.

Notes 1. Parallel texts contain a pair consisting of the ST and the TT, which is usually a translation of the source text into the target language. However, in the parallel texts (sub-corpus 1) of this corpus, while all Japanese TTs are translations of the English STs, only the English TT of Mixi (a Japanese social media website) is a translation of a Japanese ST. The English TTs of LINE, Mobage and GREE are adaptations of the Japanese ST with content adapted and localised for the English-speaking community of the target markets in the United States and Europe. 2. As a general rule, the ratio for converting a character count in Japanese to word count in English is around 2.5 Japanese characters to one English word. In the sample data, the variation in word counts between Japanese and English texts is further attributed to the divergence in content in terms of coverage and depth of the individual privacy policy in both parallel texts (sub-corpus 1) and comparable texts (sub-corpus 2). 3. Suru mono to suru is usually translated as shall that signals the modality of obligation in English legal discourse.

References Bhatia, V., Candlin, C., Engberg, J. and Trosborg, A. (2003) Multilingual and Multicultural Contexts of Legislation. Germany: Peter Lang. Böttger, C. (2004) Genre mixing in business communication. In J. House and J. Rehbein (Eds.), Multilingual Communication, vol. 3. Amsterdam and Philadelphia: John Benjamins, pp. 115–129. Böttger, C. (2007) Lost in Translation?: An Analysis of the Role of English as the “Lingua Franca” of Multilingual Business Communication. Hamburg: Kovač. Caffarel, A. (2004) Metafunctional profile of the grammar of French. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (Eds.), Language Typology: A Functional Perspective. Amsterdam: Benjamins, pp. 77–137. Connor, U. (2002) New directions in contrastive rhetoric. TESOL Quarterly, pp. 493–510. Halliday, M. A. K. (1956) Grammatical categories in modern Chinese. Transactions of the Philological Society, 55(1), pp. 177–224. Reprinted in M. A.  K. Halliday (2006) Studies in the Chinese Language, vol. 8 in the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, pp. 209–248. Halliday, M. A. K. (1959) The Language of the Chinese ‘Secret History of the Mongols’, vol. 17. Published for the Society by B. Blackwell (Publications of the Philological Society 17.) Reprinted in M. A. K. Halliday (2006) Studies in the Chinese Language, vol. 8 in the Collected Works of M. A. K. Halliday, edited by J. J. Webster. London and New York: Continuum, pp. 3–171.

Corporate Legal Discourse


Halliday, M. A. K. (1978) Language as Social Semiotic: The Social Interpretation of Language and Meaning. London: Edward Arnold. Halliday, M. A. K. (2001) Towards a theory of good translation. In E. S. C. Yallop (Ed.), Exploring Translation and Multilingual Text Production: Beyond Content. Berlin, Germany: Mouton de Gruyter, pp. 13–18. Halliday, M. A. K. and Hasan, R. (1989) Language, Context, and Text: Aspects of Language in a Social-Semiotic Perspective. Oxford: Oxford University Press. Halliday, M. A. K. and Matthiessen, C. M. I. M. (1999) Construing Experience Through Meaning: A Language-based Approach to Cognition. London: Cassell. Halliday, M. A. K. and Matthiessen, C. M. I. M. (2014) Halliday’s Introduction to Functional Grammar, 4th ed. London: Routledge. Halliday, M. A. K., McIntosh, A. and Strevens, P. (1964) The Linguistic Sciences and Language Teaching. London: Longman. Hasan, R. (1984/1996) The nursery tale as a genre. Nottingham Linguistic Circular, 13, pp. 71–102. Reprinted in Hasan, R. (1996) Ways of Saying: Ways of Meaning: Selected Papers of Ruqaiya Hasan, edited by C. Carmel, D. Butt and G. Williams. London: Cassell, pp. 1951–1972. Hasan, R. (2009) The place of context in a systemic functional model. In Continuum Companion to Systemic Functional Linguistics. London and New York: Continuum, pp. 166–189. Hinds, J. (1983) Linguistics and written discourse in particular languages: Contrastive studies: English and Japanese. Annual Review of Applied Linguistics, 3, pp. 78–84. Hinds, J. (1990) Inductive, deductive, quasi-inductive: Expository writing in Japanese, Korean, Chinese, and Thai. Coherence in Writing: Research and Pedagogical Perspectives, pp. 87–110. Hiraga, M. K. (1999). Difference as Distance: Metaphorical Base of Honorific Verb Construction in Japanese. Paper presented at the Cultural, Psychological, and Typological Issues in Cognitive Linguistics: Selected Papers of the Bi-annual ICLA Meeting in Albuquerque, July 1995. Edited by Masako K. Hiraga, and Sherman Wilcox, John Benjamins Publishing Company, 1999. Ikegami, Y. (1981) Suru to Naru no Gengogaku. Tokyo: Taishuukan. Ikegami, Y. (1991) The Empire of Signs: Semiotic Essays on Japanese Culture. Foundations of Semiotics, vol. 8. Amsterdam and Philadelphia: John Benjamins. Interbrand. (2014) Best Global Brands 2014. Available at: www.bestglobalbrands. com/2014/ranking/ (Accessed on 24 April 2015). Interbrand Japan. (2014) Best Japan’s Global Brands 2014. Available at: http:// (Accessed on 24 April 2015). Kaplan, R. B. (1966) Cultural thought patterns in inter-cultural education. Language Learning, 16(1–2), pp. 1–20. Kranich, S. (2016) Contrastive Pragmatics and Translation: Evaluation, Epistemic Modality and Communicative Styles in English and German. Pragmatics & Beyond New Series, vol. 261. John Benjamins. Lado, R. (1957) Linguistics Across Cultures: Applied Linguistics for Language Teachers. Ann Arbor: University of Michigan Press. Lavid, J. (2000) Cross-cultural variation in multilingual instructions: A study of speech act realisation patterns. Language in Performance, pp. 71–86.


Sonya Chik

Lavid, J. and Arús Hita, J. (2002) Nuclear transitivity in English and Spanish: A contrastive functional study. Languages in Contrast, 4(1), pp. 75–103. Lavid, J., Arús Hita, J. and Zamorano-Mansilla, J. R. (2009) Systemic Functional Grammar of Spanish: A Contrastive Study With English, vol. 1. London and New York: Continuum. Mann, W. C., Matthiessen, C. M. I. M. and Thompson, S. A. (1992) Rhetorical structure theory and text analysis. In Discourse Description: Diverse Linguistic Analyses of a Fund-raising Text, pp. 39–78. Mann, W. C. and Thompson, S. A. (1988) Rhetorical structure theory: Toward a functional theory of text organization. Text, 8(3), pp. 243–281. Martin, J. R. (2004) Metafunctional profile of the grammar of Tagalog. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (Eds.), Language Typology: A Functional Perspective. Amsterdam: Benjamins, pp. 255–304. Martin, J. R., Matthiessen, C. M., Painter, C. and Webster, J. J. (2010) Deploying Functional Grammar. Beijing: The Commercial Press. Matthiessen, C. M. I. M. (1995) Lexicogrammatical Cartography: English Systems. Tokyo: International Language Sciences Publishers. Matthiessen, C. M. I. M. (2002) Combining clauses into clause complexes: A multifaceted view. In J. B. M. Noonan (Ed.), Complex Sentences in Grammar and Discourse: Essays in Honor of Sandra A. Thompson. Amsterdam: John Benjamins, pp. 237–322. Matthiessen, C. M. I. M. (2015) Register in the round: Registerial cartography. Functional Linguistics, 2(1), pp. 1–48. Matthiessen, C. M. I. M. (forthcoming) The Semantic System of Rhetorical Relations: Rhetorical Structure Theory Revised. Book manuscript. Matthiessen, C. M. I. M., Teruya, K. and Lam, M. (2010) Key Terms in Systemic Functional Linguistics. London: Continuum. Mizusawa, Y. (2008) Investigating the Directive Genre in the Japanese and Australian Workplace: A Systemic Functional Approach. PhD thesis, University of Wollongong, Australia. Patpong, P. (2006) A Systemic Functional Interpretation of Thai Grammar: An Exploration of Thai Narrative Discourse. PhD thesis, Macquarie University. Steiner, E. and Teich, E. (2004) Metafunctional profile of the grammar of German. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (Eds.), Language Typology: A Functional Perspective. Amsterdam: John Benjamins, pp. 139–184. Shibatani, M. (1990). The languages of Japan. Cambridge University Press. Tajima, M. (2005) Hooreiyoogo Handobukku (kaiteihan) [A Handbook of Legal Terminology (revised edition)]. Japan: Gyoosei. Teich, E. (1995) Towards a methodology for the construction of multilingual resources for multilingual generation. Paper presented at the Proceedings of the IJCAI’95 Workshop on Multilingual Text Generation, pp. 136–148. Teich, E. (1999) System-oriented and text-oriented comparative linguistic research: Cross-linguistic variation in translation. Languages in Contrast, 2(2), pp. 187–210. Teich, E. (2003) Cross-Linguistic Variation in System and Text: A Methodology for the Investigation of Translations and Comparable Texts, vol. 5. Berlin: Walter de Gruyter. Teruya, K. (2004) Metafunctional profile of the grammar of Japanese. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (Eds.), Language Typology:

Corporate Legal Discourse


A Functional Perspective. Amsterdam and Philadelphia: John Benjamins, pp. 185–254. Teruya, K. (2007) A Systemic Functional Grammar of Japanese, 2 vols. London: Continuum. Teruya, K. and Matthiessen, C. M. I. M. (2015) Halliday in relation to language comparison and typology. In J. Webster (Ed.), The Bloomsbury Companion to M. A. K. Halliday. London and New York: Bloomsbury Publishing, pp. 427–452. Thomson, E. A., Fukui, N. and White, P. R. (2008) Evaluating ‘reporter’ voice in two Japanese front-page lead stories. In E. A. Thomson and P. R. R. White (Eds.), (Ed.), Communicating Conflict: Multilingual Case Studies of the News Media. New York: Continuum. Trosborg, A. (1997) Rhetorical Strategies in Legal Language: Discourse Analysis of Statutes and Contracts. Tübingen, Germany: Gunter Narr. White, P. R. (2000) Media objectivity and the rhetoric of news story structure. Language in Performance, pp. 379–397.

11 Grammaticalising Attitude Clause Juncture Particles and negotiation in Dagaare Isaac N. Mwinlaaru

1. Introduction Over the past half century, linguists working in language typology, language description and related fields, such as conversation analysis and interactional sociolinguistics, have increasingly given attention to spoken discourse, in general, and dialogic interactions, in particular.1 The consequence is a continual enrichment of our knowledge of the different ways by which speakers encode their attitude and other interpersonal meanings in interactions. Key grammatical systems that have been widely investigated in this regard include evidentiality (see contributions in Chafe and Nichols, 1986; Aikhenvald and Dixon, 2003), modality (e.g. Bybee, Perkins and Pagliuca, 1994, Chapter 6; Palmer, 2001), and the latest addition of mirativity, the grammatical means by which speakers indicate surprise or unexpected information (DeLancey, 1997, 2012). One linguistic phenomenon that has come under focus is modal particles that are placed in clause juncture positions to host speakers’ prosodic cues and encode their (inter)subjective stance. These particles are noted to realise meanings like those encoded by intonation in languages such as English and other European languages (cf. Halliday and Greaves, 2008, pp. 123–128). Many studies on these modal particles have focused on Asian languages (e.g. Morita, 2005, 2015; Kim and Sohn, 2015), with some studies exploring linguistic resources with similar behaviour in European languages, such as clause final conjunctive and adverbial items (cf. Hancil, Haselow and Post, 2015). There is, however, an apparent neglect of African languages in the literature, as is reflected in a recent review by Hancil, Post and Haselow (2015), although clause juncture particles are pervasive in many African languages. Clause juncture particles have also often been treated in isolation, without recognising them as forming a grammatical system, and many studies have focused on only final particles although some of these modal particles also occur clause initially and even clause medially. On the other hand, systemic accounts of Asian languages (e.g. Halliday and McDonald, 2004; Teruya, 2004, 2007; see also Teruya et al., 2007) have recognised clause juncture particles as forming a grammatical system, the system of

Grammaticalising Attitude


negotiation.2 However, only sketches of this system have been provided in these languages, normally in relation to the discussion of mood, and there is a great need to investigate this system comprehensively and more systematically within and across languages. The present study is an attempt to fill these gaps and envisages a systematic typology of negotiation across languages. It is when we treat clause juncture particles systemically that we can begin to map them meaningfully across languages. The study examines negotiation in the Lobr dialect of Dagaare (Niger-Congo: Gur), spoken in Ghana and Burkina Faso. The account is guided by Halliday’s (1996) notion of a trinocular vision, an approach that allows us to observe the system from three points of view: (i) The first view is ‘from above’ lexicogrammar, that is, in the semantics stratum of language. From this point of view, negotiation is defined as a semiotic resource that enacts the clause interactively as a negotiable unit of discourse by encoding the speaker’s (inter) subjectivity—their attitude towards the clause or the listener. (ii) The second view is ‘from below’ the clause, profiling particles that are placed in clause juncture positions to enact attitudinal meanings in the clause. (iii) The third view is ‘from roundabout’, within the clause itself. Here, the study will explicitly identify the various attitudinal meanings that are enacted within the system of negotiation and their systemic relationship with one another. It also considers the interaction between negotiation and mood—i.e. how the same particle realises different meanings across indicative and imperative clauses, extending the primary mood contrast in Dagaare between indicative and imperative in delicacy. This also concerns the placement of the particles as juncture prosodies either at clause initial or clause final position and how these positions interact. The unit of analysis is the clause, which may be major (e.g. The thief ran away) or minor (e.g. Oh yes!), simplex (e.g. I saw him) or complex (e.g. The thief ran away because I saw him), full (e.g. Where did you see him?) or elliptical (e.g. . . . in the kitchen) (cf. Halliday and Matthiessen, 2014, pp. 125–127, pp. 193–196; see also Mwinlaaru, 2017, pp. 98–106, on Dagaare). The data set for the study consists of naturally occurring spoken texts, including casual conversations, recreational texts, media discourse and workshop reports. Constructed examples are used for illustration where convenient. The rest of the chapter is organised as follows: §2 will proceed to examine the interpersonal structure of the Dagaare clause to provide a context for the discussion on negotiation. The system of negotiation itself is discussed in §3 based on the three views outlined above. §4 concludes the study.


Isaac N. Mwinlaaru

2. The Interpersonal Structure of the Dagaare Clause The interpersonal structure of the Dagaare clause is represented as follows: (Negotiator) (^ Subject) (^ Adjunct) ^ Predicator (^ Complement) (^ Adjunct) (^ Negotiator). The parentheses indicate optional elements, and the caret shows that the elements are ordered. An illustration is given in Figure 11.1 (the equal sign (=) indicates clitics). The Subject is realised by a nominal group, and it is defined as the modal responsibility of the clause and/or the nub of the argument in propositions. In Figure 11.1, Ʋ (‘S/he’) is the subject of the interrogation. The Predicator is typically realised by the verbal group, consisting of the verb, obligatorily marked for aspect, and particles realising systems such as polarity, modality, tense and directionality (see Mwinlaaru, 2017, pp. 120–127, for details; see also Bodomo, 1997, pp. 80–92, on Central Dagaare). The Complement is typically realised by nominal groups, while the Adjunct is typically realised by adverbial groups. The Negotiator is realised by modal particles that occur as juncture prosodies in clause initial and clause final position, and it enacts the clause as a negotiable unit of discourse. The placement of the Negotiator at clause initial and clause final positions has interactive significance. The beginning of the clause is where the speaker is taking off in the exchange, and the end of the clause is where s/he is potentially handing over the turn. These juncture positions are therefore at ‘risk’ for hosting the speaker’s prosodic cues in the exchange. The following dialogue sheds light on the functions of the clause elements in the flow of interaction (all personal names in conversations are pseudonyms): (1)




Filipi ɓaw nι a Phillip pick.pfv foc def Subject Predicator Complement ‘Phillip picked the matchet from here.’ Ʋ̃ ʋ̃, Ɩ ̃ bɛ ɓaw no 1sg neg.ind.nfut pick.pfv Subject Predicator ‘No, I haven’t picked (it).’ Fʋ ɓaw na. 2sg pick.pfv affr Subject Predicator Negotiator ‘You picked (it).’

sʋɔ a ka. matchet def here Adjunct ɛ. naffr Negotiator

Figure 11.1 Illustration of the interpersonal structure of the Dagaare clause

Grammaticalising Attitude


In the first clause, Speaker A initiates the exchange by making a claim in respect of B (Phillip). The Subject, ‘Phillip’, carries the modal responsibility, and the Complement a sʋɔ (‘the matchet’) is introduced as the focus of new information, signalled by the focus particle nι. In the next two clauses, we see a shift in the person deixis realising the Subject in A’s and B’s utterances, corresponding to each speaker’s subjective encoding of modal responsibility in the clause. The Predicator marks the polarity of the clause, and the polarity values shift between positive and negative to carry the argument in the exchange forward. Again, once the Complement and Adjunct are established as shared information, they are discarded, making way in subsequent turns for the Negotiator to enact the speaker’s assessment of the clause as the turn is being handed over (see Mwinlaaru, 2017, pp. 113–136, 229–233, for details).

3. The System of NEGOTIATION As mentioned in the preceding sections, the system of negotiation is realised by juncture prosodies, particles placed at the initial or final position of the clause as Negotiator elements in the clause. These particles occur only in free clauses as opposed to bound clauses, and, indeed, free clauses serve as the domain for the realisation of negotiation in Dagaare. The particles ground the clause within the semantic space open to speaker and listener as something that can be negotiated. There are two kinds of negotiation particles, namely those that contribute to making mood distinctions, particularly in indicative clauses, and those that are optional in the clause and only indicate the speaker’s attitude towards the proposition or proposal realised by the clause (cf. Teruya et al., 2007; Mwinlaaru, Matthiessen and Akerejola, 2018). In this study, all negotiation particles, including those that contribute to mood distinction, will be considered together as realising the system of negotiation (see Table 11.1; cf. Halliday and McDonald, 2004, pp. 341–342, on Chinese). The meanings of negotiation particles mostly vary across moods, but also based on the ‘affective loading’ provided by different interactional contexts (Halliday and McDonald, 2004, p. 342). Their categorization and interpretation are therefore fuzzy. Table 11.1 glosses their general meanings. The categories of value in the table indicate their relative degree of force. As the table also shows, a few of the particles are restricted to indicative clauses. Also, except for the hesitative particle mɛ́ ̃ and the exhortative particle na, which are clause initial particles respectively associated with indicative and imperative clauses, the rest of the particles are clause final. Despite the fuzziness in the meanings of negotiation particles, it can be generalised that negotiation in the indicative clause is mostly concerned with epistemic stance, the degree to which the speaker is committed to the knowledge claims of propositions, and, in the imperative, negotiation typically modulates the proposal realised by the clause. Although these two


Isaac N. Mwinlaaru

Table 11.1 Negotiation particles and their general meanings across moods Value


Meaning Indicative




dɛ ka kaka kɛ̃ / wɛ1 ka Kaka wɛ

definitive assertive strongly assertive admonitive

na (clitic forms: n=, =a) ɩ, e, ɛ

affirmative (‘factual’) non-affirmative (‘non-factual’) opinative

bɩɩ na2 bɩ wɛ/kpo1 ʋ ya mɛ́2 mɔ̀ yaa

1 2


neutral interrogative biased interrogative biased interrogative biased interrogative; mirative hesitative counter-expectation empathic

Imperative mild insistence strong insistence admonitive mildly imploring strongly imploring exclamative; requestive prohibitive suggestive exhortative


Particles that are sub-dialectal variants for realising the same meaning Clause initial negotiation particle

categories of negotiation are like categories of modality, there is a systemic distinction between negotiation and modality in Dagaare. modality is realised by particles placed within the verbal group and consists of two subsystems: probability and desirability (cf. Mwinlaaru, 2017, 172–179). The section will proceed to discuss the different uses of negotiation particles in the indicative (§3.1) and imperative (§3.2) moods. 3.1. Negotiation in the Indicative Mood: Epistemic and Affective Stance Attitudinal negotiation in the indicative mood can be divided into two main types, based on the orientation of the negotiation, whether it is (1) proposition-oriented or (2) interactant-oriented. These two orientations evoke the notions of subjectivity and intersubjectivity. However, most of the categories discussed below as proposition-oriented do not only encode the speaker’s subjectivity but simultaneously invite the listener to respond to it. Negotiation meanings are therefore better treated as varying degrees on a cline from subjectivity to intersubjectivity rather than as discrete categories of subjectivity and intersubjectivity.

Grammaticalising Attitude


(1) Proposition-oriented negotiation: Proposition-oriented negotiation indicates the speaker’s attitude towards the propositional content of a clause. This type of negotiation represents different degrees of assertiveness. For convenience, they will be grouped based on their semantic relatedness. The categories here comprise (i) the affirmative and nonaffirmative, (ii) the various types of interrogative negotiation, (iii) the definitive, assertive and strongly assertive, (v) the exclamative, (vi) the opinative and (vii) the hesitative and counter-expectation (see Table 11.1). (i) Affirmative and non-affirmative: These two relates to factuality—the speaker ends the clause by clearly assessing its polarity value, which is signalled earlier in the verbal group by polarity particles, as s/he is potentially about to hand over the turn to the addressee. The Negotiator element for the affirmative is realised by the clause final particle na (example 2).3 For the non-affirmative, the Negotiator is realised by one of the phonologically variant final particles, e, ɛ or ι, depending on its phonetic environment in terms of tongue root vowel harmony (example 3): (2)






Fʋ wõ =n a lɛ [[ʋ na yel a]]? 2sg hear.pfv=foc def dem she rel say.pfv junc ‘You heard what she said?’ Ɩ̃ wõ a na. 1sg hear.pfv 3pl.nhm affr ‘I heard it.’ Fʋ baw nι bom kaw? 2sg know.pfv foc thing some ‘You know something?’ Ɩ̃ bɛ baw ɛ. 1sg neg.ind.nfut know.pfv naffr ‘I don’t know.’

As (2) and (3) show (both from the play St Maria), the affirmative particle asserts and negotiates the positive value of the clause while the nonaffirmative particle asserts and negotiates the negative value of the clause. In (3), for instance, polarity is realised by the particle bɛ in the verbal group realising the Predicator. At the end of the clause, however, the speaker reiterates the polarity value of the clause as an interpersonal punch as s/he is potentially about to hand over the turn to the listener. In (2), the polarity of the clause is zero-marked, and the affirmative particle na serves as a juncture prosody reflecting the positive value of the proposition. (ii) Interrogatives: The various interrogative particles reduce the epistemic force of the proposition and at the same time indicate the speaker’s bias towards the propositions for which they are seeking confirmation from the addressee. The categories here comprise the neutral interrogative (4), initiative biased interrogative (5) and responsive biased interrogative (6). The relevant particles are highlighted in bold in the dialogues


Isaac N. Mwinlaaru

(all from the play St Maria). Other particles realising biased interrogation are identified in Table 11.1. (4)





B: (6)



A dɔɔ nɔw ya-rɛ na bι? def man adv be:mad-ipfv affr int ‘Is the man possibly mad? / The man is possibly mad, is he?’ Ɩ̃ bɛ baw ɛ wɛ! 1sg neg know.pfv naffr excl ‘I don’t know!’ Nι dι na wɛ? 2pl eat.pfv affr int ‘You have eaten, right?’ Ʋ̃ ʋ. ‘Yes.’ Ɩ̃ zã̀a wa na wa nyɛ̃ 1sg yesterday come.pfv affr prox see.pfv a ɩ̃ pɔw-yaa ʋ bιɛrɛ. def 1sg daughter 3sg be:sick.ipfv ‘Yesterday I came home to see that my daughter was sick.’ Ʋ bιɛrɛ ya? 3sg be:sick.ipfv int ‘She was sick, you say?’

As the examples show, the interrogative clause is realised by simply adding an interrogative particle to a declarative clause. The implication is that the interrogative clause is a non-assertive proposition, a proposition where the speaker relatively defers the epistemic claim to the listener. The biased interrogative (examples 5 and 6) signals some degree of commitment to the epistemic claim of the proposition, while the neutral interrogative (example 4) completely defers commitment to the listener. The two biased interrogatives, initiative and responsive, show different orientations in the exchange. The initiative flows from the speaker and thus indicates more commitment to the claim of the proposition, while the responsive realises an echo-question, often assessing the proposition affectively as a surprise. Thus, in addition to its default interrogative uses, the particle ya has a mirative meaning, and both meanings are often enacted simultaneously (cf. DeLancey, 1997; Aikhenvald, 2012, on mirativity). Further, the interrogative particle ya is also used in ‘self-talk’, where the speaker signals to the listener that s/he is finding it difficult to cognitively retrieve information from his/her consciousness. The following dialogue from a casual conversation illustrates this phenomenon: (7)


Zan nʋ waar. [. . .] Tɩ dɛ a dɛ Zan come.ipfv 1pl adv affr adv fɩɩrɩ [laughter] [. . .]. be:in trouble ‘Zan is the one coming . . . We are just in trouble . . .’

Grammaticalising Attitude B:




Ʋ dɛ bɛ pawr a mɩ̀ ɛ. Ɛcɛ 3sg adv neg.ind.nfut get.ipfv 3pl.nhm also naffr but a a a . . . ãa lɛ nɩ a na ya? def def def who cop foc def dem int ‘S/he is just not getting some of it. And the the the . . . who is this even?’ A Gɔbr. def Gobr ‘Gobr.’ Ʋʋ. ‘Yes.’

The material situation surrounding this exchange is a family get-together, where a group of relatives are sharing a bottle of wine under a tree. A neighbour sees them and walks towards the scene to join the celebration. Speaker A alerts the group of the prospective guest, indicating the inadequacy of the wine. Speaker B, at this point, inquires of the whereabouts of his younger brother and, for a moment, cannot remember his name. The particle ya added to the clause, which is already a ‘wh’-interrogative, simply indicates that the proposition is not a genuine question but rather a self-reflection. Such clauses are also commonly used as rhetorical questions showing surprise or unexpectedness at seeing, for instance, a long ‘lost’ friend. Thus, ya clearly lies between the boundary of an interrogative marker, obligatory as it were, and an optional attitudinal marker. (iii) Definitive, assertive and strongly assertive: First, definitive negotiation gives a high degree of force to a proposition. It adds an emphatic punch to the proposition realized by the clause, showing a high epistemic commitment on the part of the speaker. It is realised by the particle dɛ. An example is the elliptical clause Ziri dɛ! (‘Lie!’) in Speaker C’s turn in example (8). The extract is an episode in a biblical drama (The Story of Jesus) where the Jewish chief priest cautions his elders about Jesus’s popularity: (8)




Nɩbɛ yaga zie de ʋ na ʋ ɩ people many place take.pfv 3sg affr 3sg be.pfv bɛ nàa baarɩ kɛ̃. 3pl king finish.pfv adm ‘Many people have accepted him to be their king already, how careless you are!’ ||| Nàa ya? Nɩ-baalbɛ, pɔw-ƴɛmɛ, nyanyuuru nàa? ||| king int thieves king Bɛ mɩ́ saa nɩ dɩã , || daw-bio 3pl hab show.pfv foc today tomorrow bɔr   || bɛ yiiri bɛ bɛr. ||| disappear.pfv 3pl forget.pfv 3pl leave.pfv ‘King? King of sick people, barren women and thieves? They appear today and disappear tomorrow (and) people forget about them.’ ||| Ziri dɛ! ||| A ʋ yuor yire na def 3sg name come:out.ipfv affr lie       emp


Isaac N. Mwinlaaru


dɔwlɛ bibie za || ɛ a nɩbɛ increase.ipfv days all conj def people mʋʋrɛ ʋ mɩ̀. ||| praise.ipfv 3sg adv ‘Lie! He is getting popular every day, and the people are praising him too’. ||| Siza na.   ||| Nàa nʋ.   ||| Ɛcɛ damnʋ dɛ truth king conj trouble adv wa lɛ bere a tew pʋɔ a,    || cond be:again exist.ipfv def town inside junc nyιmɛ so a sawna. ||| 2pl.emp own.pfv def blame ‘It is true. He is king. But if any trouble just occurs in the town again, you will have the blame.’

This dialogue is rich in negotiation. The chief priest (A) admonishes his elders for their inability to control the activities of Jesus, leading to his popularity among the people. In the first clause, he uses the admonitive particle kɛ̃ to signal his disappointment and sound caution to the elders. In essence, the particle kɛ̃ prosodically enacts the whole statement as an admonition rather than a bare statement of fact. In defence, one of the elders (B) tries to mitigate the seriousness of the situation. Note his use of the biased interrogative particle ya not only to query the proposition made by the chief priest but also to show surprise. Of particular interest here is the underlined clause, where another elder (C) challenges the position of his colleague with a definitive clause, signalled by the particle dɛ. This particle colours the proposition it attaches to as conclusive. In other words, the speaker lays a strong epistemic claim to the proposition. In general, the extract clearly exemplifies negotiation in action in the flow of discourse. The next proposition-oriented negotiation to discuss consists of those that have been labelled assertive and strongly assertive in Table 11.1. On a scale of degree of force, they are less strong than the definitive negotiation although they are also of high epistemic value. A mildly assertive negotiation is realised by the particle ka, while a strongly assertive negotiation is realised by its reduplication counterpart, kaka. Examples are given in the following constructed examples: Ʋ wa na ka. 3sg come.pfv affr m.ins ‘He has come, I insist.’ (10) Ʋ bɛ wa ɩ kaka. 3sg neg.ind.nfut come.pfv naffr s.ins ‘He hasn’t come, I strongly insist.’ (9)

The interactional contexts where a speaker typically uses these two assertive forms are where the speaker is reacting to or challenging a

Grammaticalising Attitude


proposition made by the addressee by insisting on the truth value of the proposition the particle attaches to. (iv) Exclamative: Another proposition-oriented negotiation is the exclamative. Exclamation is realised in Dagaare only as a form of attitudinal stance, unlike in many Indo-European languages, where it can be realised by a sub-type of the declarative mood (e.g. How gracefully she walks!) (cf. Halliday and Matthiessen, 2014, Chapter 4, on English). Except for the interrogative clause, any Dagaare clause can be turned into an exclamative by adding the particle wɛ to it. An example is the utterance by Speaker B in example (4). The dialogue gives other illustrations (from The Story of Jesus): (11)

Pɔw ɩ, a fʋ baalʋ sanι na. woman voc def 2sg sickness heal.pfv affr ‘woman, your sickness is healed.’ Voice 1: Wi! Alɛ! Ʋ sanι na wɛ! intj intj 3sg heal.pfv affr excl ‘What! She is healed!’ Voice 2: Ʋʋ    wɛ! Nɔ-ɓaa yel =ʋ. Nɩ wa yes   excl shocking 2pl come.pfv nyɛ̃ a! see.pfv prt ‘Yes! It is a marvellous incident. You all come and see!’ Jesus:

Example (11) is an extract from a movie episode on a healing miracle of Jesus. Here, the exclamative particle occurs in two clauses, the first being an affirmative clause (Voice 1) and the second being a minor clause (Voice 2). The exclamative particle in these clauses shows the surprise and feeling of awe of the speakers at the miraculous healing. As Moutaouakil (1999, p. 7) observes of exclamation, it “signals the speaker’s evaluation of [their] attitude towards the content of the linguistic expression with the peculiarity that the source of the evaluation is the impression made on the speaker by this content.” The realisation of exclamation as an assessment category in Dagaare rather than a mood type resonates with Moutaouakil’s (1999) cross-linguistic characterisation of exclamation as a kind of modal assessment. He notes that while declarative, interrogative and imperative clauses have structural properties that distinguish them across languages, exclamations can take variant forms even within the same language. As (11) shows, exclamation in Dagaare has no unique structural realisation, only being indicated optionally as the speaker’s attitude encoded in the clause (also see §3.2 for instances of exclamation in the imperative clause). (v) Opinative: The opinative is realised by the particle bɩɩ and enacts the proposition realised by the clause as the speaker’s opinion on an issue or as his/her personal conviction. An example is given in the following dialogue (from the play St Maria):


Isaac N. Mwinlaaru



Mãa lɩɛbɛ nɩ faara o! [. . .] 1sg.emp turn.ipfv foc priest prt [. . .] ‘I am becoming a priest! . . .’ Father: Bʋnʋ ya? A sukuul ɩ ̃ na yaw fʋ, fʋ What int def school 1sg rel put.pfv 2sg 2sg bɛ zawrɩ ɛ? [. . .] neg.ind.nfut refuse.pfv naffr ‘What? The school I put you in, didn’t you stop? . . .’ Fʋ manɛ a faara mɩnɛ na mɩ́ zɩ 2sg assume.pfv def priests pl rel hab sit.pfv a lɛ a, zu gbɩlɩ lɛ bɛ ɛ? [. . .] def dem junc head.pfv knots cop 3pl naffr ‘You assume the priests who sit like that, they are uneducated?’ (= ‘You assume the priests you see sitting in church are uneducated?’) Son: Ɩ̃ na lɩɛb =a. Bɛ na 1sg pos.ind.fut turn.pfv=affr 3pl pos.ind.fut de =m =a bɩɩ. take.pfv=1sg.acc=affr op ‘I will become (one). They will admit me, in my opinion.’

In this extract, the father’s discouraging response to his son’s declaration of becoming a priest implies that the son will not be admitted to the priesthood since he is a school dropout. Note again the use of the biased interrogative particle ya by the father to show surprise at his son’s statement or unexpectedness. In reaction to the father’s comments, the son explicitly enacts his conviction in the final clause with the opinative particle bɩɩ. It is also worth pointing out that the particle o in the first clause in (12) has negotiatory value although, unlike the other particles, this particle is very general (or even para-linguistic) and has no specific meaning. It is a juncture prosody that amplifiers the volume of the utterance. In fact, its use is not limited to Dagaare and would be familiar to speakers of West African languages in general. (vi) Hesitative and counter-expectation: Next, hesitative and counterexpectation particles are associated with interrogative clauses. The hesitative is indicated by the clause initial particle mɛ́ ̃. Examples are given in (13) and (14): (13)

Political opinion interview Mɛ́ ̃ tɩ be =n a ka a mʋtɔw hst 1pl exist.pfv=foc def here def afternoon za wɛ? all int ‘I believe all of us have been here since this afternoon, right?’ (14) St Maria play A: A libie nɩ ana. def foc.cop dem ‘These are the money.’

Grammaticalising Attitude B:




Na dɔm! Na dɔm =ɩ tɩ nyɛ̃ exh squat.pfv exh squat.pfv=com 1pl see.pfv a! [. . .] prt ‘Please squat! Please squat with (it) and let’s see!’ Mɛ́ ̃ a ta na wɛ? hst 3pl reach.pfv affr int ‘I believe it has reached (the agreed amount), right? Ɩ̃ bʋɔlɩ ɩ̃ pɔw-yaa ʋ wa sɔr 1sg call.pfv 1sg daughter 3sg prox count.pfv kaa nyɛ̃. check.pfv see.pfv ‘Let me call my daughter to come and count and see.’

As the examples show, the hesitative particle is an epistemic downtoner. The speaker hedges the proposition by hesitating in claiming full knowledge of it, inviting the addressee to confirm the proposition. Thus, it often co-occurs with the biased interrogative particle wɛ (and also ʋ or kpo) (see Table 11.1), as the examples indicate. Semantically, the hesitative particle simultaneously indicates both the speaker’s bias and uncertainty towards the proposition. Because of this, even if it occurs in a clause without an interrogative particle, the clause mostly still carries the sense of a question (e.g. Mɛ́ ̃ tɩ be =n a ka a mʋtɔw za, ‘I believe all of us have been here since this afternoon’). The following extracts, especially the dialogue in (16), illustrate this phenomenon further: (15) The story of Jesus ||| Mɛ́ ̃ nyɩ taa na lɔb zʋkpar na hst 2pl mod pos.ind.fut throw.pfv proverb dem kʋ̀ =m: || “dɔkta a, sanɩ fʋ tʋɔra!” ||| give.pfv=1sg.acc doctor voc heal.pfv 2sg self ‘I believe you will tell me this proverb: “doctor, heal yourself.” ’ (16) Workshop interview Agric Officer: [. . .] A tɩ kɔb [[tɩ na mɩ́ [. . .] def 1pl farming 1pl rel hab kʋɔr a]], mɛ́ ̃ tɩntɛr =ʋ tɩ farm.ipfv junc hst 1pl mɩ́ kɔ yaga. hab farm.pfv plenty ‘. . . Our farming we do, I believe it is plains we cultivate more.’ Host: mmm intj ‘Yeah.’ Agric Officer: Soo, bɛ kɔ =n tɩntɛr. so 3pl farm.pfv=foc plains ‘So they cultivated plains.’


Isaac N. Mwinlaaru

Generally, as Yang and Yap (2015, p. 54) characterised a Mandarin stance marker, the hesitative particle is used when the speaker is quite certain about the epistemic claims of the proposition, yet, at the same time, is reluctant to come across as being assertive. Counter-expectation, on the other hand, is the grammatical encoding of irony, as it were. What is negotiated in the clause could be either a situation the speaker is reacting to or the propositional content of the clause itself. In other words, it either indicates (i) whether the situation engendering a question is opposite to the expectation of the speaker or (ii) that the speaker believes the opposite of the propositional content of a question. It is realised by the particle mɔ̀, illustrated in (17) (from The Story of Jesus): (17)

Blind man: Alɛ ka! Bʋʋ nʋ? Bʋʋ nιɛ nʋ wait m.ins what what adv

mɔ̀? ce Man:

‘Wait, I implore you! What is it? What is even (happening)?’ Nazaratι tew Yeezu nʋ tɔlɛ a ka. Nazareth town Jesus pass.ipfv def here ‘It is Jesus of Nazareth (who) is passing here.’

The use of mɔ̀ in (17) is an instance of an assessment of the situation engendering the question as counter to the speaker’s expectation. The extract is from the episode of the healing of the blind Bartolomeo by Jesus. The blind man hears an unusual commotion and movement of a large crowd along the road where he sits in quest of alms. He approaches a fellow closer to him and queries about the unusual drama unfolding before him. The particle mɔ̀, which punctuates the interrogative clause, signals the oddity of the situation. Another example is given in (18) from the same text, but this time showing counter-expectation to the propositional content of a clause (complex): (18)

||| Yel tɩ nyɛ̃! ||| A zuru yab [[tɩ na say.pfv 1pl see.pfv def heads payment 1pl rel ya-rɛ kʋ̀-rɛ a nàa Sɩzaar a]], a pay-ipfv give-ipfv def King Caesar junc 3pl.nhm sɛw na   || bɩɩ a bɛ be:acceptable.pfv affr conj 3pl.nhm neg.ind.nfut sɛw mɔ̀? ||| be:acceptable.pfv ce ‘Tell us! The taxes we pay to Caesar, is it even acceptable or is it not acceptable?’

In this extract, the Jewish elders approach Jesus to inquire about the appropriateness of the payment of taxes to Caesar. The use of the counterexpectation particle in the alternative interrogative clause marks the proposition as counter to the expectations of the speaker (see Mwinlaaru

Grammaticalising Attitude


2017, pp. 150–153, on alternative interrogative clauses in Dagaare). It must be noted that the counter-expectation particle scopes over the whole clause (complex). While the particle prompts the addressee that the speaker disprefers one of the alternatives, it does not make the preference explicit. (2) Interactant-oriented negotiation: Interactant-oriented negotiation, on the other hand, clearly indicates the speaker’s intersubjective attitude towards the addressee. It comprises admonitive and empathic negotiation. The admonitive is realised by the particle kɛ̃ or wɛ, depending on the sub-dialect. An example is the first clause in example (8), which has already been discussed: Nɩbɛ yaga zie de ʋ na ʋ ɩ bɛ nàa baarɩ kɛ̃ (‘Many people have accepted him to be their king already, how careless you are!’). Here, the admonitive particle kɛ̃ prosodically enacts the whole statement as an admonition rather than a bare statement of fact. Empathic negotiation, as the name suggests, signals that the speaker empathises with the addressee or is emotionally involved in the proposition realised by the clause in some way. It is realised by the particle yaa, illustrated here by constructed examples: (19) A mama na wa na yaa? def mum pos.ind.fut come.pfv affr em ‘Mum will come, alright?’ (20) A: Ɛ Zan wa? cont Zan be:where ‘And where is Zan?’ B: Wi! Ʋ kpi na yaa. intj 3sg die.pfv affr em ‘Oh! S/he is dead, sorry for the shock.’

Example (19) is a typical consoling statement to a child crying for his/ her absent mother, while example (20) gives a scenario where a bereaved relative responds to a question on the whereabouts of the deceased, whose demise the questioner is ignorant of. In both clauses, the empathic marker indicates that the speaker is emotionally involved in the proposition by empathising with the addressee. While in (19) the speaker signals endearment towards the worried child, in (20) the speaker anticipates the shocking effect of the proposition and the listener’s embarrassment and indicates his/her empathy. Generally, the empathic particle features prominently in care-taker talk, where is it used as an endearment marker (as in example 20). 3.2. Negotiation in the Imperative Mood: Modulating Proposals This section proceeds to discuss negotiation in the imperative mood. As mentioned earlier, negotiation here modulates the force of the proposal by either toning it down, enforcing it or indicating disinterest on the part of the speaker. This modulation strategy maps out a semiotic space


Isaac N. Mwinlaaru

consisting of more delicate sub-types of the imperative. They can broadly be classified into (1) hortative and (2) non-hortative types. (1) Hortative negotiation: Hortative negotiation generally encourages or exhorts the addressee to execute the proposal. In a sense, it modulates the force of the proposal by negotiating obligations. Two hortative types are identified: (i) adhortative and (ii) exhortative. These are semantically distinguished based on the degree to which the speaker can enforce the proposal. With the adhortative, the speaker positions her/himself as relatively less able to enforce the proposal. The addressee is empathised with and encouraged or urged to execute the proposal. The Negotiator is realised by the particle yaa: (21) Ta kõne ɩ yaa! neg.imp cry.ipfv naffr em ‘Don’t cry, okay!’

The exhortative, on the other hand, strongly encourages or exhorts the addressee to execute the proposal. It is marked by the clause initial particle na, immediately preceding the Predicator. Examples are given in the following (from the play St Maria): (22) Husband: Na bɔ bʋndɩrɩ kʋ̀ mãa! exh search.pfv food give.pfv 1sg.emp ‘Please, get me food! (and let’s stop talking about this issue)’ Wife: Ãnʋ bɔ bʋndɩrɩ kʋ̀ fʋ? [. . .] who find.pfv food give.pfv 2sg [. . .] ‘Who (should) get you food?’ Husband: Bʋʋ so fʋ yele a lɛ? what own.pfv 2sg say.ipfv def dem ‘Why are you saying so?’ Wife: Fʋ kʋ̀ =m =ɩ libir? 2sg give.pfv=1sg.acc=foc ‘Have you given me money?’ (23) Ɩ ̃ pɔw-yaa na ƴɛrɛ ɩ̃ zie! 1sg daughter exh speak.ipfv 1sg place ‘My daughter, please speak to me! (and ignore your father)’

As the English translations in (22) and (23) show, the exhortative particle is used in contexts where the proposal the speaker is making is competing with other topics for the attention of the listener (example 22) or where the speaker is competing with other voices for the attention of the listener (example 23). The particle na is thus used to exhort the listener to align with the speaker’s interest. An observation of interactions in casual conversation reveals that the use of this particle can be offensive to the nonaddressee participant whose interest the listener is exhorted to ignore. A case in point is (23), where a mother exhorts the daughter to continue their conversation and ignore an interruption from the father.

Grammaticalising Attitude


(2) Non-hortative negotiation type: The non-hortative types comprise (i) prohibitive, (ii) requestive/exclamative, (iii) admonitive, (iv) insistent/ imploring and (v) suggestive negotiation. (i) The prohibitive negotiation corresponds to the non-affirmative in the indicative mood and is realised by the same set of phonologically variant particles (see Table 11.1). For want of space, it will not be discussed further here. (ii) Requestive/exclamative: The requestive entreats or requests the addressee to ensure the success of the proposal realised by the clause. The Negotiator is realised by the particle wɛ, as exemplified by the underlined clauses in the following extract (from The Story of Jesus): (24)




Fʋ bɔbr kɛ tι ι ŋmιn? 2sg want.ipfv proj 1pl do.pfv what ‘What do you want us to do?’ Faa tι wɛ! save.pfv 1pl excl ‘Save us, please!’ Yel wɛ! Tι ι ŋmιn? say.pfv excl 1pl do.PFV what ‘Tell (us), please! What (should) we do?’

These utterances are made by a restless crowd following John the Baptist in a biblical drama. The requestive clauses that realise Speaker B’s and C’s proposals entreat him to answer their questions. In the Bartolomeo episode introduced earlier (example 17), the blind man (A in example 25) employs the requestive to entreat Jesus for healing. The extract is repeated here in an extended form (the relevant clause is underlined): (25)




Alɛ ka! Bʋʋ nʋ? Bʋʋ nιɛ wait m.ins what what adv nʋ mɔ̀? ce ‘Wait! What is it? What is even (happening)?’ Nazaratι tew Yeezu nʋ tɔlɛ a ka. Nazareth town Jesus pass.ipfv def here ‘It is Jesus of Nazareth (who) is passing here.’ Oyi! Yeezu u! Oyi! Yeezu u! intj Jesus voc intj Jesus voc Daviir bi-dɛb ι, zɔ a ɩ̃ nι-baalʋ wɛ! David child-male voc run.pfv def 1sg pity excl ‘Oh, Jesus! Oh, Jesus! Son of David, have pity on me, please!’

Supplicatory contexts such as these are typical environments for a requestive Negotiator. The speaker defers to the listener to motivate her/him to bring about the proposal. In addition to its requestive use, the particle wɛ is also used in the imperative mood to realise exclamation (cf. §3.1). However, the use of


Isaac N. Mwinlaaru

wɛ as an exclamation marker in the imperative is limited. It is only used with the perceptive verb nyɛ̃ (‘see’), as in the following extract from a casual conversation: (26) Nyɩnɛ na fʋ dɛ ŋmʋrɛ a lɛ cere where 2sg adv rush.ipfv def dem.dist go.ipfv nɩ? Nyɛ̃ wɛ! com see.pfv excl ‘Where is it that you are just dashing like that to? Look at that!’

Here, Nyɛ̃ wɛ! (‘See!’) is not an invitation for the listener to engage in a perceptive process, but rather an indication of the speaker’s disapproval of the listener’s careless behaviour, indicated in the preceding clause. The perceptive verb nyɛ̃ (‘see’) has been recruited for the realisation of exclamations, and it has almost lost its sense of perception in this environment. As the English translation shows, this incongruent use of perceptive verbs is not limited to Dagaare but is also possible in English and tend to be a cross-linguistic phenomenon. Below are other exclamation clauses where nyɛ̃ (‘see’) is used other than those with the exclamative particle we (from The Story of Jesus): (27) Nɩ nyɛ̃ a nɩɩ! 2pl see.pfv 3pl.nhm prt ‘You look at that (= Look at what you have caused)!’ (28) Fʋ bɛ nyɛ̃ yɛrʋ! 2sg neg.ind.nfut see.pfv speech ‘What a speech!’

Examples (27) and (28) are instances of recurrent formulaic expressions. Example (27) is a depreciative exclamation, while (28) is an appreciative exclamation (cf. Moutaouakil, 1999, pp. 8–10). Except for yɛrʋ (‘speech’) in (28), all the individual words in these two clauses are semantically empty; they are non-referential. The pronouns are impersonal and have no specific reference, and the verb nyɛ̃ does not necessarily indicate perception by sight in clauses such as (28). A more congruent realisation of the process here would be wõ (‘hear’) rather than nyɛ̃ (‘see’) since what is perceived is sound rather than a physical object. Thus, while (27) is used to indicate the speaker’s frustration towards the listener (i.e. depreciative exclamation), (28) shows the speaker’s admiration of the speech alluded to (i.e. appreciative exclamation). (iii) Admonitive: The admonitive negotiation in the imperative mood has a similar meaning as in the indicative mood. But here it either cautions or urges the addressee against the event or action represented by the Predicator/Process (example 29) or indicates that the speaker is indifferent to whether the proposal is enforced or not. That is, it is the addressee’s own business (example 30). There is sub-dialectal variation in

Grammaticalising Attitude


the realisation of the Negotiator. Mostly, speakers in Burkina Faso use a distinctive particle kɛ̃, while speakers in Ghana use the particle wɛ, which has the same form as the exclamative/requestive particle, and the two can only be distinguished in context. Illustrations of the admonitive are given in the constructed examples below: (29) Zɩ wɛ/kɛ̃! Bɛ dɩrɛ na. sit.pfv adm 3pl eat.ipfv affr ‘You sit! They are eating. (You won’t have any food to eat.)’ (30) A: Ɩ wa bɩ? 1sg come.pfv int ‘Should I come?’ B: Wa wɛ/kɛ̃! come.pfv adm ‘Come! (It’s your business; I don’t care).’

(iv) Insistent/imploring: The particle ka (mild) or kaka (strong) in the imperative clause can be interpreted as realising insistence or imploring, depending on the tenor and/or the nature of the demand imposed on the listener. In the following clauses, ka is interpreted as imploring the listener to carry out the command since, in each case, the listener is in a more privileged or powerful position than the speaker. (31) The Story of Jesus Alɛ ka! Wait.pfv m.ins ‘Wait, I implore you!’ (32) St Maria play Pɛw mɛ ka! Ɩ̃ dɩ =n san. lend.pfv 1sg.acc m.ins 1sg.nom owe.pfv=foc debt. Mãa cen, ʋ kʋ̃ kʋ̀ mɛ̃ ɩ. 1sg.emp go.pfv 3sg fut.neg give.pfv 1sg.acc naffr ‘Lend me, I implore you! I owe (her) a debt. (If) I go, s/he will not give (it) to me.’

Example (31) is from the blind Bartolomeo to a passer-by, and (32) is from an insolvent husband to his apparently well-to-do wife. A similar utterance from father to son, such as Wa kaka! (‘Come, I insist!’), will be interpreted as insistence. Thus, while in the declarative mood, ka and kaka boost the speaker’s commitment to the knowledge claims of the proposition, in the imperative they modulate the speaker’s demand for goods and services. (v) Suggestive: The final negotiation type to consider in the imperative mood is the suggestive, realised by the clause final particle bɩɩ. It is illustrated by the following constructed dialogue: (33)


Ɩ̃ sɛ̃ a nɛn? 1sg roast.pfv def meat ‘I (should) roast the meat?’


Isaac N. Mwinlaaru B:

Dʋw bɩɩ. Boil.pfv op ‘Boil it, I suggest / I think!’

As the exchange indicates, this particle is always used in reaction to a preceding utterance, and it enacts the imperative clause as an alternative course of action. 3.3. Negotiation Concord To recapitulate, the particles realising negotiation relatively have different meanings in the indicative and imperative moods. In the indicative mood, they mostly enact the speaker’s assessment of the epistemic claim of propositions. In the imperative, they indicate the speaker’s assessment of the tenor and the force of proposals. These are, however, relative characterisations. For instance, interactant-oriented negotiation in indicative clauses is also motivated by tenor. Generally, the particles realising negotiation are juncture prosodies that extend the indicative and imperative mood distinctions in delicacy. However, more than one attitudinal marker can be found in a single clause, at most up to three. In the indicative mood, this often happens in polar interrogative clauses, which normally combine the affirmative or non-affirmative particle with the selected interrogative particle (see Table 11.1). Also, a clause which already has an obligatory Negotiator for mood contrast can take an optional Negotiator element (examples 19 and 20). This section is particularly concerned with a situation where more than one optional Negotiator is used in a clause to amplify negotiation. This normally involves a clause initial particle and a clause final particle. I refer to this phenomenon as negotiation concord to highlight the semantic resonance between the co-occurring particles. In the indicative mood, such co-occurrence consists of the initial hesitative particle mɛ́ ̃ and a variant of the initiative biased interrogative particle (see Table 11.1). An example of this has already been introduced in examples (13) and (14) in the preceding section. Both the initial and final particles indicate the speaker’s uncertainty towards the epistemic claim of the proposition (cf. §3.1). The following combinatory possibilities are identified in the imperative mood: (i) (ii) (iii) (iv) (v)

Na . . . . . . . . . . . . . . . . . . . . . . . . . . . ka (the imploring sense)! Na . . . . . . . . . . . . . . . . . . . . . . . . . . . kaka (the imploring sense)! Na . . . . . . . . . . . . . . . . . . . . . . . . . . . wɛ (the requestive sense)! Na . . . . . . . . . . . . . . . . . . . . . . . . . . . yaa! Na . . . . . . . . . . . . . . . . . . . . . . . . . . . bιι!

As these combinations show, negotiation concord partly contributes to classifying particles into the ‘value’ categories in Table 11.1. Clause initial

Grammaticalising Attitude


na can only combine with categories of median or low value, and not high value categories. On the other hand, mɛ́ ̃, which is of low value, can only occur with categories within the range of its value. An illustration of negotiation concord involving na is given below by a dialogic musical interlude in a traditional folktale about a wild duck and her paralytic duckling (Gbɛr-be-yeni Wʋba, ‘One-Leg Paralytic’; repeated lines are omitted): (34) Duckling: Ɩ ̃ ma a, ɩ̃ ma a, bɛ 'wɔbr 1sg mother voc 1sg mother voc 3pl chew.ipfv fʋ na wɛ! 2sg affr excl ‘My mother, my mother, they are eating you!’ Duck: Ɩ̃ bie e, ɩ̃ bie e, na ta 1sg child voc 1sg child voc exh neg.imp kõne ι wɛ! cry.ipfv naffr excl ‘My child, my child, please don’t cry, please!’ Mãa nιɛ pιrɛ yɛryɛryɛr ƴawnɛ 1sg.emp adv struggle.ipfv ideo give.ipfv bɛlɛ nɛ. 3pl.emp dem.dist ‘For I have earnestly been feeding those others.’ Bɛlɛ nιɛ bɛ kõne ι. Ɛ 3pl.emp adv neg.ind.nfut cry.ipfv naffr conj fʋʋ pɑ̃a nιɛ kõne. 2sg.emp adv adv cry.ipfv ‘They are not even crying and now you are crying.’ Gbɛr-be-yeni wʋba a, na ta kõne leg-one paralytic voc exh neg.imp cry.ipfv ι wɛ! naffr excl ‘One-legged paralytic, please don’t cry, please!’

The story relates that Mother Duck refused to feed her paralytic duckling due to its deformity but took very good care of her healthy ducklings. When she was attacked and killed by a group of hunting boys, all the ducklings scattered, except the paralytic one, who followed the boys wherever the mother was taken. In this exchange, the duckling employs the exclamative marker to indicate her desperation at her mother being eaten. Of particular interest here is the combination of the exhortative particle na and the requestive marker wɛ in Mother Duck’s response. With this negotiation concord, she negotiates the proposal at both ends of the clause to strongly show her remorse and empathise with and entreat the duckling to stop crying. This contributes to creating an emotionally charged semiotic context, the social function of which is to evoke the right emotions in the young audience and inject in them the moral import of the story.


Isaac N. Mwinlaaru

4. Conclusion This chapter examined one modal assessment system in Lobr Dagaare, namely negotiation. It first discussed the interpersonal structure of the Dagaare clause and identified the Negotiator, realised by clause final and initial particles, as the negotiatory element in the clause. It has been shown that many of the particles have different meanings across the indicative and imperative moods. In the indicative they mostly indicate epistemic stance and, in a few instances, affective stance. In the imperative, on the other hand, they modulate proposals and, again, show the speaker’s affective attitude. Thus, these juncture prosodies expand the meaning potential of the grammar of mood by realising delicate or finer indicative and imperative meanings. As a characteristic of highly delicate items, the meanings of the particles are relatively fuzzy, and one challenge that results from this is how to gloss them. Where the meanings are closely related, one core meaning has been chosen as a general gloss. One key contribution of this study to the extant literature on clause juncture particles is the holistic approach adopted. The particles are defined and categorised based on their discourse functions and in relation to their systemic environment—how they make meaning in relation to other possible choices and how they differ across moods. This approach is proposed as a model for future research.

Abbreviations Not in Leipzig Glossing Rules ADM—admonitive, AFFR—affirmative, CE—counter-expectation, CONJ—conjunction, CONT—continuative, EM—empathy marker, EMP—emphatic, EVT—eventuality, EXCL—exclamative, EXH— exhortative, EXIST—existential, HAB—habitual, HM—human, HST—hesitative, IDENT—identifying pronoun, IDEO—ideophone, INT—interrogative, INTJ—interjection, INS—insistence, JUNC— juncture, M—mild, MOD—modal particle, OP—opinative, POS— positive, PROJ—projection marker, PRT—particle, S—strong.

Texts Used for Illustrations 1. St Maria play—An unscripted play by students of St Maria Junior High School, Nandom, Ghana, August 5, 2013. Author’s transcription. 2. The Story of Jesus—A film on the Life and Times of Jesus Christ according to the Gospel of Luke. Burkina Faso, August 20, 2012. Available at Author’s transcription.

Notes 1. I acknowledge funding from the University Grants Council, Hong Kong (Award no. PF12-10332). I thank Christian Matthiessen for discussions on this study.

Grammaticalising Attitude


2. Halliday and McDonald (2004, pp. 341–342) name this system as ASSESSMENT. I use the term NEGOTIATION from Teruya (2004, pp. 193–194). Following Matthiessen (2004, pp. 631–635), I reserve the term (MODAL) ASSESSMENT as a general label for all systems enacting the speaker’s evaluation in the clause (e.g. evidentiality, modality, negotiation, etc.). 3. When the affirmative particle na occurs in serial verb constructions, it comes after the first verb.

References Aikhenvald, A. Y. (2012) The essence of mirativity. Linguistic Typology, 16, pp. 435–485. Aikhenvald, A. Y. and Dixon, R. M. W. (Eds.). (2003) Studies in Evidentiality. Amsterdam and Philadelphia: Benjamins. Bodomo, A. (1997) The Structure of Dagaare. Stanford: Stanford University, CSLI Publications. Bybee, J., Perkins, R. and Pagliuca, W. (1994) The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago and London: University of California Press. Chafe, W. L. and Nichols, J. (Eds.). (1986) Evidentiality: The Linguistic Encoding of Epistemology. Norwood, NJ: Ablex Publishing Corporation. DeLancey, S. (1997) Mirativity: The grammatical marking of unexpected information. Linguistic Typology, 1, pp. 33–52. DeLancey, S. (2012) Still mirative after all these years. Linguistic Typology, 16, pp. 529–564. Halliday, M. A. K. (1996) On grammar and grammatics. In R. Hasan, C. Cloran and D. G. Butt (Eds.), Functional Descriptions: Theory Into Practice. Amsterdam and Philadelphia: Benjamins, pp. 1–38. Halliday, M. A. K. and Greaves, W. (2008) Intonation in the Grammar of English. London and Oakville: Equinox. Halliday, M. A. K. and Matthiessen, C. M. I. M. (2014) Halliday’s Introduction to Functional Grammar, 4th ed. London and New York: Routledge. Halliday, M. A. K. and McDonald, E. (2004) Metafunctional profile of Chinese. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (Eds.), Language Typology: A Functional Perspective. Amsterdam and Philadelphia: Benjamins, pp. 305–396. Hancil, S., Haselow, A. and Post, M. (Eds.). (2015) Final Particles. Berlin and Boston: Mouton. Hancil, S., Post, M. and Haselow, A. (2015) Introduction: Final particles from a typological Perspective. In S. Hancil, M. Post and A. Haselow (Eds.), Final Particles. Berlin and Boston: Mouton, pp. 3–35. Kim, S. H. and Sohn, S.-O. (2015) Grammar as an emergent response to interactional needs: A study of final kuntey ‘but’ in Korean conversation. Journal of Pragmatics, 83, pp. 73–90. Matthiessen, C. M. I. M. (2004) Descriptive motifs and generalisations. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (Eds.), Language Typology: A Functional Perspective. Amsterdam and Philadelphia: Benjamins, pp. 537–674. Morita, E. (2005) Negotiation of Contingent Talk: The Japanese Interactional Particles ne and sa. Amsterdam and Philadelphia: Benjamins.


Isaac N. Mwinlaaru

Morita, E. (2015) Japanese interactional particles as a resource for stance building. Journal of Pragmatics, 83, pp. 91–103. Moutaouakil, A. (1999) Exclamation in Functional Grammar: Sentence Type, Illocution or Modality? Working Papers in Functional Grammar 69. Amsterdam: University of Amsterdam. Mwinlaaru, I. N. (2017) A Systemic Functional Description of the Grammar of Dagaare. PhD thesis, The Hong Kong Polytechnic University. Mwinlaaru, I. N., Matthiessen, C. M. I. M. and Akerejola, E. S. (2018) A systembased typology of mood in Niger-Congo languages. In A. Agwuele and A. Bodomo (Eds.), Routledge Handbook of African Linguistics. London and New York: Routledge, pp. 93–117. Palmer, F. R. (2001) Mood and Modality, 2nd ed. Cambridge: Cambridge University Press. Teruya, K. (2004) Metafunctional profile of the grammar of Japanese. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (Eds.), Language Typology: A Functional Perspective. Amsterdam and Philadelphia: Benjamins, pp. 185–254. Teruya, K. (2007) A Systemic Functional Grammar of Japanese, vols. 1 and 2. London and New York: Continuum. Teruya, K., Akerejola, E., Andersen, T. H., Caffarel, A., Lavid, J. Matthiessen, C., Petersen, U. H., Patpong, P. and Smedegaard, F. (2007) Typology of mood: A text-based and system-based functional view. In R. Hasan, C. M. I. M. Matthiessen and J. J. Webster (Eds.), Continuing Discourse on Language: A Functional Perspective, vol. 2. London: Equinox, pp. 859–920. Yang, Y. and Yap, F. H. (2015) ‘I am sure but I hedge’: Fear expression kǒngpà as an interactive rhetorical strategy in Mandarin broadcast talk. Journal of Pragmatics, 83, pp. 41–56.

12 Systemic Socio-Semantic Stylistics (SSS) as Appliable Linguistics The Cases of Literary Criticism and Language Teaching/Learning Donna R. Miller and Antonella Luporini* 1. Introduction As Halliday puts it, “the value of a theory lies in the use that can be made of it” (1985b, p. 7). Indeed, the aim of an appliable linguistics is to challenge the boundaries between theory and practice (e.g. Halliday, 2002 [2009], p. 3; Mahboob and Knight, 2010, p. 4; Matthiessen, 2012, p. 436). This chapter would do just that, focusing on the practice of doing stylistics in a Systemic Functional Linguistics (SFL) perspective (Halliday, 1985a and subsequent editions), i.e. on the Systemic Socio-Semantic Stylistics (SSS) framework (Hasan, 1985 [1989], 2007, pc),1 and on how it might achieve its ‘social accountability’ (Matthiessen, 2012, p. 436) in higher academic contexts. These contexts for us are literary criticism and the teaching and learning of English, which are conflated in our teaching of stylistics. The second section pays too-brief tribute to the work of scholars who were among the first to avow the value of a linguistic approach to the study of literature. Here we also bring together language, literature and English Language Teaching (ELT), under the label of ‘Pedagogical Stylistics’ (see, e.g., Burke, Csábi, Week and Zerkowitz, 2012). Our next section succinctly describes the stylistics model we have long worked with: Hasan’s SSS framework and the theory behind it. Section 4 then discusses the methodology adopted in our class activities. Select concrete illustrations of doing SSS are offered in section 5. The two-fold aim is to show how SSS is perfectly instrumental to both doing linguistic criticism and further developing already appreciable L2 skills, through stylistics, in the Italian academic setting.2 In closing, we cursorily summarise the case we’ve aimed to make.

2. Linguistics, Literature and ELT What follows offers a fleeting look at twentieth-century arguments for a linguistic approach to the study of literature (but see Miller, forthcoming). The dénouement of the tradition of purely author-centred literary theory came about with the work of the Moscow Linguistic Circle, and


Donna R. Miller and Antonella Luporini

specifically with the most famous of the so-called Russian Formalists, Roman Jakobson, well-known for his ‘poetic function’, whose focus is on the message for its own sake (1960, p. 356 ff.). Soon afterwards, in 1964, coincidently the same year as Hasan’s PhD thesis, Halliday wrote, “Literature is language for its own sake: the only use of language, perhaps, where the aim is to use language” (in Halliday, MacIntosh and Strevens, 1964, p. 245, our emphasis). The correspondence between Jakobson’s and Halliday’s views is clear. In her PhD thesis (1964), Hasan first put forth an analogous proposition: It is not that there is art [somewhere ‘out there’, so to speak], and the job of language is simply to express it; rather it is that, if there is art, it is because of how language functions in the text. . . . [I]n verbal art the role of language is central. Here language is not as clothing to the body; it is the body. (reiterated in 1985 [1989], p. 91) But back in the 1930s Jakobson was collaborating with the Prague Structuralists, notably Mukařovský, aiming at identifying the formal and functional linguistic mechanisms responsible for articulating the aesthetic motivation they both saw as the impetus of literature, and working together towards the theorisation of what would become seminal and durable stylistic concepts. With time Jakobson proposed grammatical parallelism (henceforth GP)—the structural patterning in texts that he theorised as the empirical linguistic evidence of his ‘poetic function’—and the notion of pervasive parallelism (PP; 1966), while Mukařovský moved on to further hypothesising the role of foregrounding (1964, 1977, 1978). But the theoretical intersection endured, despite the undoubtedly greater ‘success’ of the latter—not only in SSS, but in stylistics tout court. Foregrounding’s mileage is indeed impressive, but Jakobson’s reach was just as significant, if not as wide-ranging. In the 1960s, Roger Fowler began his own struggle to reinstate literature as a legitimate object of linguistic study—as the renowned acerbic, if also at times amusing, Fowler-Bateson debates (1967, 1968) testify. He was also the first to draw proper attention to the ‘Mukařovský-Jakobson Theory’, in astutely—and quite rightly—noting the theoretical convergence of Mukařovský’s foregrounding and Jakobson’s GP (1986, p. 73). But Fowler explicitly and energetically distances himself from their ‘aesthetic’ positions on the functions and motives of foregrounding and parallelism. For him, literature, in short, is not special. Although Hasan was sceptical of the central role of GP in her model to the last (pc in an ongoing discussion with Miller, 10 January 2011), she would strongly disagree on this point, for reasons we’ll make clearer below. All the scholars mentioned in this condensed account have made a seminal contribution to the inception and the recognition of stylistics as a discipline. And it is also thanks to their insights if today we can seek “to

Systemic Socio-Semantic Stylistics (SSS)


foreground pedagogical stylistics within the wider framework of literary linguistics, putting teaching and teaching-driven research at the top of our professional agendas” (Burke et al., 2012, p. 1). Pedagogical stylistics, defined as “the application of stylistic techniques in teaching” (McIntyre, 2011, p. 10), has a by now quite long tradition in English as L2 contexts (Clark and Zyngier, 2003), where it has been shown to improve learners’ linguistic sensitivity and awareness (Clark and Zyngier, 2003; Zyngier, Fialho and Rios, 2007; cf. also Fogal, 2015).3 Of course, the question of “how to teach literature to enable the learner to produce an independent reasoned ‘reading’ of some existing work in literature” was, for Hasan, a vital one, “socially, morally and pedagogically” (Hasan, 2011, p. xv). And one of the undeniable advantages of her functional approach is that it allows students to go beyond their personal, subjective responses and to engage hands on with the linguistic mechanisms of the text (Lukin, 2008, p. 103; cf. forthcoming). We will show how they do just that in our English as L2 classrooms at the University level. But first we will describe the theoretical model we work with.

3. Language, Linguistics and Verbal Art4 Our analytically and pedagogically tested (Miller, 1998, 2000) conviction is that the SSS model is unique, but also uniquely valid, its main—and most essential—strong point being its association of text to the linguistic system, i.e. to meaning potential (Halliday, 1982 [2002]: 128). It is also unique because of both its holistic, coherent and systematic nature, and its premise that literature is ‘special’. This belief goes against the claim of contemporary mainstream stylisticians (e.g. Fowler, but also Simpson, 2004 [2014]) that literature is merely another text type, to be investigated with the same analytical tool kit applied to any other register. We dissent, as Hasan would, and espouse her contextually rooted ‘double-articulation’ model—comprising firstly a semiotic system of language, but also a distinct semiotic system of verbal art—what makes it special—as in Figure 12.1. 3.1. ‘Double-Articulation’ Figure 12.1 represents how verbal art, indisputably a kind of language use in a particular social context, is not simply a register like any other. For Hasan, this is because the context-language connection in verbal art is fraught with complexities which other registers are not heir to (Hasan, 2007, pp. 22–23); thus, it requires a different methodological take. The First Level: The Semiotic System of Language This is the multiple coding system which is the valid entry point for the analysis of any register (Hasan, 1985 [1989], p. 92). Each stratum is realised, i.e. becomes accessible to us, through the one below it. In


Donna R. Miller and Antonella Luporini

Figure 12.1 The double-articulation framework (based on Hasan, 1985 [1989], p. 99)

short, semantics, or meanings, are realised in lexicogrammar, or wordings, which become accessible to us in phonology, or soundings, or, in the case of a written text, in graphology, in written symbols. The Second Level: The Semiotic System of Verbal Art The first stratum of this second system is called ‘verbalisation’. As is evident, it is the meeting point of the two systems, where the system of verbal art concretely links up to the system of the language of the text. Thus, verbalisation contains the whole of the first system. It equals it; hence the broken line in the figure. The highest stratum of the second order of semiosis is called ‘theme’ in reference to the ‘deepest’ meaning of any verbal artefact (Hasan, 1985 [1989], pp. 97–98). For Hasan, such deep-ness is very close to a generalisation which can be viewed as a hypothesis about some aspect of the human condition—what Aristotle may have had “in mind when he declared that art is truer than history” (Hasan, 1985 [1989], p. 97). Without this theme, or reflection on humanity, and its ‘symbolic articulation’, there is, for Hasan, no verbal art (1985 [1989], p. 100). The middle stratum is that of symbolic articulation. Looking at it from ‘below’, this is the place where the basic first-order meanings are expanded upon, or enriched, and are made ‘art’. How this happens is through foregrounding, or the patterning of patterns. This is the process by which the meanings emerging through analysis of the semiotic system of language are symbolically turned into signs, for the aesthetic motive,

Systemic Socio-Semantic Stylistics (SSS)


or ‘artistic intention’ (Mukařovský, 1977) of expressing a theme— that deeper meaning. The synergy between symbolic articulation and Jakobsonian PP is argued in Miller (2016) and treated here as given. 3.2. More on Foregrounding For Mukařovský, the most basic notion of foregrounding is that of ‘contrast’, a divergence between what the text establishes as a ‘norm’ and what emerges as being different, as standing out against such norms, which are what makes up the ‘background’. As Hasan rightly points out, however, the analyst can only be concerned with contrast which is significant, i.e. significant enough to be called—because functioning as— foregrounding. But how can the analyst tell when it is significant? And just what does ‘significant’ mean? Although there is no simple map that unproblematically plots the road to certainty, there are modus operandi that can be applied. ‘Significant foregrounding’ is foregrounding that ‘counts’. And what counts, for Hasan, is not occurrence or non-occurrence of linguistic mechanisms—at least not in and of themselves. What matters is ‘consistency’ and Mukařovskian ‘motivation’. Motivation means that the features are articulated in a way that is working towards the construction of theme. And if foregrounding is noticeably meaningful, then it will also prove to be ‘consistent’, and in two ways: the stability of its semantic direction, and the stability of its textual location. The former signifies that “the meanings which are being highlighted by the foregrounded patterns converge toward the same direction” (1985 [1989], p. 95). In practice we find it is often the very semantic tension between fore- and background that is responsible for construing this kind of consistency. The stability of textual location does not refer to predictable, regular, material locations in the text, but simply means that significant patterns of foregrounding tend to take place in textually significant places (1985 [1989], p. 95). 3.3. The Context of Creation Space limits preclude adequate treatment of this essential aspect of the approach, which is seen as operating: ‘over’, i.e. presiding over the functions, in toto, of the entire dual framework. Figure 12.2 depicts the three ‘ingredients’ to be probed, vis-à-vis the verbal artist in relation to those of his or her discourse community, whose language, artistic conventions and world view the artist cannot help engaging with, either to basically repropose or to some degree distance him/herself from them. Probing language means investigating how a writer exploits his/her contemporary resources for making (‘standard’ and non-) meanings


Donna R. Miller and Antonella Luporini

Figure 12.2 The context of creation (our representation)

available; it may mean discovering how these have been modified over time. A whole history of artistic conventions are there to be adopted, or flouted, though ‘originality’ is not rampant. No artist creates in a vacuum—as Hasan puts it: “It is not an accident that the theme, the symbolic articulation, and the patterns of language display a great deal of similarity within any one period of verbal art” (1985 [1989], p. 102). Probing the verbal artist’s world view means examining the taken-forgranted ideological assumptions of a discourse community, which cannot help but flow into its verbal art. But research into even quasi-obsolete belief and value systems may at times be needed, since “the greater the distance between the context of creation and reception, the more inaccessible the meanings of the text become” (Hasan (1985 [1989], p. 101). In our view, scrutiny of the context of creation should follow, rather than anticipate, analysis at both semiotic levels. Following, it may shed light on the theme provisionally formulated; anticipating, it would tend to condition findings.

4. Applying SSS: Methodology Our SSS teaching takes place within courses in English language and linguistics for undergraduate and graduate students at the School of Foreign Languages, Literatures, Interpreting and Translation of the University of Bologna. English is an L2 for practically all of our students; the level of proficiency for incoming undergraduates is at least B1 of the Common European Framework of Reference (but for 70 per cent, higher), tested upon admittance through the Oxford Online Placement Test; the level in the main is C1 for postgraduates. So our students come with considerable competence in English, but negligible language awareness, which they need guidance to. The English-teaching

Systemic Socio-Semantic Stylistics (SSS)


staff opted for SFL as a theoretical framework over fifteen years ago, as it “offers a rich interpretation of meaning through Halliday’s theory of metafunctions” (Macken-Horarik, Sandiford, Love and Unsworth, 2015, p. 148). They were convinced that (1) to learn how to mean in L2 one must understand the mechanisms of a language, (2) to do that one needs to be able to identify their functions, and (3) to talk about these functions, a metalanguage is essential. Thus, we progressively provide students with SFL metalanguage—from scratch, as SFL is not ordinarily taught in Italian secondary schools. With that metalanguage we can better tackle some of the main challenges faced by L2 English teaching/ learning in the twenty-first century, e.g. enhancing students’ knowledge of language as a multi-functional resource to produce meaning, and so also improving their awareness of the effects of linguistic choices and ultimately their own competence to actively exploit them (cf. MackenHorarik, Love and Unsworth, 2011). In teaching SSS, we gradually apply the framework to language in literature (Hasan, 1985 [1989], p. 94), opting for texts that are stylistically uncomplicated and/or that have “themes with which the students can identify” (McKay, 1986, p. 194). But our choice of texts would also confirm that the model is “maximally applicable to the genre [i.e. to literature], irrespective of variations in time, sub-genre, and the critic’s response” (Hasan, 1985 [1989], p. 90). Among these figure Shakespeare and nineteenth-, twentieth- and twenty-first-century prose and poetry,5 both canonical and popular. The text is, firstly, viewed as an instrument, a window onto the semiotic system of language itself, of which it is a concrete instantiation (Halliday, 1982 [2002], pp. 130–132). A key notion in our workshop activities, and indeed at the core of Hasan’s SSS framework, is that of stratification (Halliday and Matthiessen, 2004, p. 24 ff.; Macken-Horarik et al., 2015, p. 149), considered from the two-fold perspective of double-articulation: firstly, intra-systemically, at the level of the semiotic system of language; and secondly, inter-systemically, between the two interrelated semiotic systems of language and verbal art. SSS analysis starts in the second undergraduate year, when basic metafunctions have been grasped. It takes the form of guided interactive workshops. To date, we concentrate on the experiential and interpersonal metafunctions, focusing on foregrounded patterns of transitivity and appraisal mechanisms. We begin with texts having prominent transitivity patterns, which are more readily handled by students approaching SSS for the first time. This experiential lens enables them to reflect on what the text is ‘about’: doing, feeling, being, saying or behaving. Interpersonal semantics is brought into the picture gradually, firstly, where relevant, through mood and modality systems and afterwards appraisal systems (Martin and White, 2005), as these are introduced only in the second year and are in any case more slippery.


Donna R. Miller and Antonella Luporini

The step upward, to the system of verbal art, where we systematically trace the relevance of foregrounding for the symbolic articulation of the theme, is initially daunting for students, but ever more readily negotiated through continual application. In the case of longer texts such as novels, we also steer them to replicate evidence from our corpus-assisted investigations (cf., e.g., Miller and Luporini, 2015, forthcoming), which allows us to draw attention to quantifiable patternings supporting qualitative findings. A preliminary theme is formulated with all the evidence in and its reconsideration performed only after the text’s context of creation is systematically examined. This is also the stage where the potential of SSS as a tool for cultural teaching/learning is more fully exploited, though opportunities for cultural observations are not lacking at any stage. We believe that our practice leads to improved language awareness and L2 performance but fear we still merit Fogal’s (2015) censure of the under-collecting and underreporting of data properly demonstrating the belief, as it is not yet possible for us to offer robust, retrievable and replicable proof of the effectiveness of our practices. As this research project is ‘younger’ than those practices and still ongoing, we are only now beginning to marshal comparative diachronic statistics specifically related to the number of participants in SSS workshops; their biodata, including individual level of pre- and post-practice language awareness and proficiency; and questionnaires on student satisfaction with SSS analysis workshops: overwhelmingly positive to date (June 2017). However, techniques to better monitor these aspects are planned, e.g. collecting samples of written assignments and assessments to track our attenders’ L2 and SSS development (cf. also Moore and Schleppegrell, 2014).6

5. Applying SSS: Sample Practice The following illustrations are presented in terms of Jakobson’s PP, whose correspondence with symbolic articulation has been asserted above. The workshops described below are illustrative, in no way exhaustive either of what we do globally, or locally, with these texts themselves. The average time dedicated to a single text at the undergraduate level is four hours, and at graduate level, having 25 per cent more time to work with, from six to eight hours, excluding home assignments. 5.1. A Second Year Undergraduate Example Second year students are generally familiar with Harry Potter (HP), the seven-volume saga of a young orphan whose life radically changes when he finds out about his magical powers and starts attending the

Systemic Socio-Semantic Stylistics (SSS)


Hogwarts School of Witchcraft and Wizardry. Indeed, the books have had a profound cultural impact on children and adults worldwide, as also confirmed by their sales figures, their copious translations (sixtyeight languages according to the American publisher’s website) and the success of the film series, which is still periodically screened on TV in Italy. We focus on one 655-word segment from the end of the first book (US edition), relating the first battle between Harry and his main antagonist, the dark wizard, Lord Voldemort (Rowling, 2001, pp. 364–367). Analysis probes the power imbalance between them as construed by processes and -er roles, or—in PP terms—Doer ^ process structures. Students work on the segment highlighted for clauses having Harry, Voldemort and their body-parts as Doers, identifying the general process types and related participant roles of the main characters: the first instances are examined together, and the remainder in small groups of four to five. Students are then guided to calculate the raw and relative frequency of each process category for Harry and Voldemort. The results at this stage, summarised in Table 12.1, show that Harry is mainly a Senser, although frequently also an Actor or Behaver, while Voldemort primarily acts as Sayer. At this point, quantitative data need to be complemented by qualitative considerations at a slightly more delicate level. This we do by (1) assigning Harry’s mental processes to their sub-class (perception, cognition, desire, emotion); (2) looking more closely at the material clauses having him as Actor; and (3) considering the ‘coupling’ between polarity/ modality and transitivity structures. Students are again asked to work autonomously in groups after initial guided examples. Table 12.1 Process types with Harry Potter and Lord Voldemort as Doers Harry Potter

Lord Voldemort

Process type

Raw frequency %

Process type

Raw frequency %

Mental Perceptive Cognitive Desiderative Emotive Material Behavioural Relational

15 12 2 1 0 13 10 1

32.5% 25.0% 2.5%

Mental Perceptive Cognitive Desiderative Emotive Material Behavioural Relational

3 1 1 0 1 3 1 4

15.8% 5.3% 21.0%

Verbal Existential Total: 40

1 —

2.5% —

Verbal Existential Total: 19

7 1

36.8% 5.3%




Donna R. Miller and Antonella Luporini

With reference to activity 1, students notice perception predominating, co-textually construing Harry as a mainly ‘passive’ participant; even his behaviours are mostly reactive (e.g. ‘scream’, ‘stumble’). Moreover, analysis reveals that five of the thirteen material processes with Harry as Doer are merely actions predicted by Voldemort (so not experientially relevant). Polarity and modality, however, offer the clearest insights. Harry’s actions repeatedly couple with negative polarity, or are modulated in terms of inability, e.g.: (1) He couldn’t move a muscle. (2) Harry would have screamed, but he couldn’t make a sound. (p. 364) The negativity/inability patterns in significant textual locations for Harry contrast with findings for Voldemort, whose actions always have positive polarity, as the students promptly observe. Space constraints preclude looking at Voldemort’s Sayer patterns in detail, but students are guided to note the bestial, aggressive quality of his verbal processes (e.g. ‘hiss’, ‘snarl’) and cruel Verbiage. His weapons are his words. We then have students focus on body-parts as Doers: what Simpson calls meronymic agency, noting that such a “(literal) disembodiment of a character often makes what they do, say or think appear involuntary, cut adrift from conscious intervention” (2004 [2014], p. 79). The following examples, in which a part of Harry’s own body is acting against the rest, reveal once again his weakness. (3) Harry tried to take a step backward, but his legs wouldn’t move. (p. 364) (4) Harry’s scar was almost blinding him with pain. (p. 366) Finally, to spotlight the asymmetrical power hierarchy construed in the passage, we introduce Hasan’s cline of dynamism (1985 [1989], p. 46), with its distinction between more active and more passive -er roles. With Table 12.2 we show how most of Harry’s processes fall into the passive end of the cline. At this point, we tentatively formulate the theme emerging from analysis of this brief text segment. Students suggest the timeless struggle between good and evil, but the need for love and friendship in such struggle is also put forward, since Harry finally wins only thanks to his friends’ help. To arrive at a final formulation, the students are first invited to read an open-access paper providing the full transitivity analysis for the segment examined and another passage, from the seventh book, relating the last

Systemic Socio-Semantic Stylistics (SSS)


Table 12.2 Harry and his body-parts (excluding actions predicted by Voldemort) on a cline of dynamism (based on Hasan, 1985 [1989], p. 46) Harry Potter + DYNAMIC Actor + animate Goal Actor + inanimate Goal Sayer + Recipient Sayer + Target Sayer Senser Actor NO Goal Behaver Carrier

12.9% — — — 3.2% 38.7% 12.9% 32.3% —


fight between Harry and Voldemort (Luporini, 2016). The comparison between the transitivity patterns in the two excerpts highlights Harry’s growth, and the totally overturned power hierarchy in the sequel. Then, they research the context of creation. Regarding language and artistic conventions, the students readily find mention of Rowling’s creative and evocative use of language, and the multiple intertextual strands in HP. Regarding world view, they report the author’s commitment to edifying literature, and her Britishness, permeating the saga (most notably, perhaps, in the organisation of Hogwarts, echoing the austere British public school culture; cf. Goatly, 2004). The deepest theme is worked out in the conclusive workshop: “Life presents increasingly difficult challenges and we must face the dark side; those trials, the support of our loved ones, and even the rules we are subject to, can help us grow ‘taking the right side’”, as Harry does in this ‘extended’ bildungsroman. 5.2. A Third Year Undergraduate Example The history of World War I (1914–1918) is taught in Italian secondary schools, and war literature in English may also be. In their undergraduate degree course, students can choose contemporary history, which covers it. Additionally, copious centennial commemorations have recently spotlighted the conflict. Thus, for our students the backdrop of Siegfried Sassoon’s acerbic social poem, “Does it Matter?”, is a familiar one. Siegfried Sassoon’s “Does it Matter?” (1918) 1. Does it matter?—losing your legs? . . . 2. For people will always be kind,

240 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

Donna R. Miller and Antonella Luporini And you need not show that you mind When the others come in after hunting To gobble their muffins and eggs. Does it matter?—losing your sight? . . . There’s such splendid work for the blind; And people will always be kind, As you sit on the terrace remembering And turning your face to the light. Do they matter?—those dreams from the pit? . . . You can drink and forget and be glad, And people won’t say that you’re mad; For they’ll know you’ve fought for your country And no one will worry a bit.

The most consistent significant foregrounding in this poem is between its globally ironic surface wordings/meanings, enhanced by PP, and what the text is actually saying/meaning—a contrast students grasp straightaway. Unearthing the textual patternings construing it, however, needs guidance. In the second undergraduate year, appraisal systems are taught; thus, students are prepared to take on the challenge of exploring irony in terms of ‘discordant couplings’, either between appraisal selections and what is being appraised, or even among the appraisal variables themselves (Martin, 2000, pp. 163–164). Space constraints preclude citing the further theoretical references we consider. Starting with the title’s—and then thrice reiterated (in the first line of each stanza)—yes-no rhetorical question to the addressee ‘you’, students invariably exclaim ‘Yes!’ They also perceive, however, that the dominant semantic direction of the poem apparently construes reasons why ‘it’ does NOT ‘matter’ at all. They are then asked to identify just who/what is being appraised in these same significant semantic locations: i.e. the almost casually appended post-posed subjects which are typically held to ‘matter’ very much indeed. Together we investigate the GP of Act + possessive deictic + Thing in the first lines of the first two stanzas, also lexically identical except for the lost physical possessions, ‘legs’ and ‘sight’, and then the structurally different third stanza’s psychological aftereffect of combat in the trenches: ‘those dreams from the pit’. Transitivity points raised at this juncture include the semantics of ‘losing’ as depriving of/causing to not have, where the Depriver—warfare itself—is only implied. Students play with co-representational wordings for ‘matter’ too: a relational attributive reading, i.e. ‘it’ as ‘important’, or a mental (worry you) one, which they astutely tend to prefer for its focus on Senser ‘you’. This reflection prompts analysis of the key participant chains: ‘people’ vs ‘you’, in a constant PP-enacted tension throughout the poem. In

Systemic Socio-Semantic Stylistics (SSS)


groups, students list the wordings denoting/connoting both ‘you’ and the ‘others’. NGs reported for ‘you’ include the (ex-)possessions of the protagonist, possessive deictic + Thing (‘your sight’ and ‘your legs’), as already noted, but also possessive deictic + a now-spoiled ‘face’, and the abstraction ‘country’—none of which, we note, inscribes evaluation in itself, but alerts us to it. Rightly, students find ‘people’ less tangible, uniform in what they will/ won’t do/be/say: we note the fully parallel structures in lines 2 and 8 where ‘people will be’ (facilely) ‘kind’—typically enacting the appraisal category of +ve (positive) judgement (cf. Martin and White, 2005, p. 52 ff.)—and that they are negated Sayers (line 13) and Sensers (line 15) and the predicted Sensers of a then-still-accepted-as ‘noble’ behaviour: line 14’s Fact of having ‘fought for your country’. Possible enacted judgement is investigated by exploring motivations for their reticence: pity and/or politeness?; aversion to being made uncomfortable? ‘You’’s range of action is also circumscribed: to being a potential Actor, Senser and Carrier, though never a Sayer. We examine how, in stanza 1, line 3, ‘you’ is actually advised not to complain (be a bore); how, in stanza 2, line 7, he is flippantly told of there being ‘splendid work for the blind’, and in stanza 3, line 12, of the possibility/ability, in the face of ‘those dreams from the pit’, to ‘drink and forget and be glad’—poignantly contrasting with all ‘you’ effectively does in the poem: to ‘sit . . . remembering’ and ‘turning [his] face to the light’ (lines 9 and 10)—‘light’ contrasting with blindness and the darkness of the pit. Also foregrounded are the parallel contrasting Acts: of you’s ‘losing’, ‘remembering’ and ‘turning’, vs the only vigorous material action in the poem and, as such, something only ‘the others’ can do: ‘hunting’, prior to (rather distastefully) ‘gobbl[ing]’. None of this evaluative reading is clearly inscribed in the poem. On the surface, ‘it’ is simply not a problem. Yet exaggeration of this kind may strongly provoke -ve (negative) judgement. Also examined is how the slippage between surface and ‘real’ meaning is also enacted, and compacted, by parallelism in the domain of verse. The provisional theme students propose is something like ‘The results of war do matter’. They are then assigned at-home research into the poem’s context of creation. They report, with regard to language and literary conventions, that Sassoon’s war poetry experimented with new modes of expression demanded by the atrocities of war and thus is seen as contributing to the modernist rejection of conventional ‘poetic diction’; and, with regard to his world view, that he strove to promote social attention to the shell-shocked and mutilated victims of the trenches. This extra-textual evidence corroborates the results of analysis of the consistently motivated PP in the text: that its deepest theme is akin to “Society’s view that war is glorious must be overturned by exposing its lasting physical and psychological effects on those who fight.”


Donna R. Miller and Antonella Luporini

5.3. A Graduate Course Example Coetzee’s Foe (1986 [1987]) is not a ‘stylistically uncomplicated’ text of the kind we typically choose (cf. the “Applying SSS: Methodology” section), but it complies with postcolonial interests at the MA level. Moreover, recent socio-cultural events regarding the author’s native country, South Africa, include the film Invictus on the life of Mandela (2009) and his death in 2013, but also the 2010 FIFA World Cup. Coetzee also won the Nobel Prize for literature in 2003. The novel is a rewriting of Defoe’s Robinson Crusoe (RC)—now read as dealing with globalisation by the superior white population, and linked to waves of asylum-seeking ‘Fridays’. RC is also still (2017) required reading in Italian secondary schools, and taught in second year undergraduate literature courses in our school. We guide students through a corpus-assisted study of the text in electronic format (corpus linguistics [CL] is also taught at the graduate level), using WordSmith Tools (Scott, 2012). Students are made to experiment with the research stages themselves. Starting with the lemmatised wordlists compiled from the carefully cleaned texts of Foe and RC, used as a reference corpus, lets them identify ‘truth’ as a keyword in the text, a finding allied to what the novel is very schematically ‘about’: The female protagonist, Susan, is shipwrecked while returning from a long unsuccessful search for her daughter. She is cast up on Cruso’s island, where he lives with ‘his man’ Friday, now mute and unrecognisable as his namesake in RC. All three are rescued, but Cruso dies before they reach England, and Susan becomes obsessed with telling their ‘true’ story. Mr. Foe, the writer to whom she turns for help, has other ideas. Then, using the British National Corpus (BNC) as a large reference corpus, they demonstrate the PP of the lemma ‘truth’ in Foe. On the basis of our CL study (Miller and Luporini, 2015, forthcoming), we pose a research question: whether the cumulative instances of ‘truth’ can be seen to construe the notion’s relativity in the novel. This brings us to appraisal analysis. Firstly, we have the students automatically retrieve and save all thirtyfive instances of ‘truth’ in the text in sentence-length concordances. We then have them retrieve collocates and discuss what they find: in particular (see Table 12.3), negative polarity, adversative ‘but’—a possible signal of relativity in itself—the NG ‘story’ and verbal process ‘tell’, reminding them that these last two are among the top-ranking lemmas in the wordlist for Foe. The first step of appraisal analysis is the identification of appraisers and appraisees/eds in the environment of ‘truth’. Before sending the students back to the saved concordances to do this, we provide a breakdown of the four clearly marked ‘parts’ of the novel and a synopsis of their contents. With such background knowledge, they are surprisingly capable of distinguishing the voices.

Systemic Socio-Semantic Stylistics (SSS)


Table 12.3 Collocates for ‘truth’ in Foe Truth (35) T-score


2.91783 2.17457 2.14231 2.10594 1.96179 1.86994 1.83516 1.76560 1.75017 1.71169 1.63702

not I It no tell Would You We but Give story

Divided into groups, students draw up a table of the results of their analyses and report their findings. Results are typically similar, unsurprisingly, as the categories of targets of appraisal are largely limited to ‘story’/‘history’/‘tale’, ‘Friday’ and ‘Cruso’, while the key appraiser is ‘Susan’. After a brief review of the systems and the categories of invoked and inscribed appraisal, analysis of ‘truth’ concordances then begins. Again the students work in groups, each analysing circa seven/eight extended concordances; subsequently, each group reports their findings, which the class discusses. Guided, students note how evaluation is predominantly of abstractions—and irrealis. Groups basically agree that +ve judgement is the category almost exclusively enacted, mostly on veracity (i.e. truthtelling), and predominantly by invoking. However, we point out how examples may challenge a clear-cut distinction between +ve and -ve appraisal, setting up a continuum across which especially Susan the truth-seeker moves, uncertainly. There is space for but one example of such fuzziness, and of the relativity of truth: (5) Return to me the substance I have lost, Mr Foe: that is my entreaty. For though my story gives the truth, it does not give the substance of the truth (I see that clearly, we need not pretend it is otherwise.) (p. 51) On one hand, Susan evaluates her veracity as positive but, on the other, as negative, somehow wanting. And here we also see enacted the novel’s— and unequivocally Susan’s—major overreaching evaluative thrust: desperately wanting truth to be known, and told (+ve veracity of ‘truth’ as the ‘ideal’), but at the same time despairing of ever being able to know and tell it (-ve veracity of what is said and/or the capacity to say it).


Donna R. Miller and Antonella Luporini

Much more is done: e.g. examining ‘truth’ with regard to Friday’s silence vs Foe’s and Susan’s words and their power, or impotence, to (re) shape (hi)stories, and so, veracity, none of which can be illustrated here. In formulating the theme, students are made aware that even more extensive analysis would be needed. For now we stop at the indeterminacy of truth. Atypically, we examine the context of creation with the students, sharing our contrastive corpus work showing that the use of tru* in Foe was unmistakably sui generis. With reference to artistic conventions, we note how Foe links up to RC contratextually, deconstructing it, by, among other things, creating a female narrator who despairs of the truth of her narration—thus enacting the angst of contemporaneous postcolonial literature. Concerning the writer’s world view, we read Coetzee’s enigmatic Noble Lecture, which again forefronts his fascination with the volatile nature of truth-telling. With this work, the indeterminacy of truth as theme is fully corroborated—now as inherent to the essential human condition.

6. Final Thoughts From the above testimonies we believe that the permeable distinction between theory and practice (Halliday, 2002 [2009], p. 3) emerges forcefully, showing how practice, in a dynamic cyclical process, feeds back into theory and methodology, strengthening the capacity of both to respond and be appliable to our academic context (cf. Halliday, 1985b, p. 1). In this chapter, despite our current over-reliance on intuitive teaching results—soon to be remedied—we hope to have made a convincing case for our multiple-strand thesis: i.e. that worthwhile ‘lit crit’ needs a linguistic approach; that SSS—rooted in SFL—is arguably the most satisfactory descriptive and analytical framework for the task; that its value extends to English language and literature teaching, and learning; and that, in the final analysis, SSS can be seen to be a thoroughly rewarding appliable linguistics indeed.

Notes * This chapter is the fruit of continuous collaboration; nonetheless, sections 1, 2 and the Foe sub-section (5.3.) were written up jointly; Luporini is responsible for section 4 (methodology) and the HP sub-section (5.1.); and Miller for the third and final section and the sub-section on Sassoon’s poem (5.2.). 1. The label Systemic Socio-Semantic Stylistics, or SSS, is Ruqaiya Hasan’s—her last formulation before her untimely passing (pc to Miller: 1 January 2015). 2. Here and in what follows, we use ‘L2’ to refer collectively to both second and foreign languages. 3. This is what we aimed to do with a recent cross-linguistic symposium we organised in Bologna entitled “The Literature Text in Language Learning:

Systemic Socio-Semantic Stylistics (SSS)


Comparing Experiences”, which met with an overwhelmingly positive reaction from the students—all from the School of Foreign Languages, Literatures, Interpreting and Translation. For more on the event, the first of a planned series, see (last accessed 7/06/2017). The initiative also prompted us to set up an ongoing project to continue the research reported in this chapter, entitled ‘SSS, the Corpus and the Consumer’ (at www., last accessed 7/06/2017). 4. The title of Hasan (1985 [1989]). 5. Despite common belief that prose is ‘easier’ than poetry, it clearly depends on what prose and poetry we’re talking about. An attempt to grade levels of difficulty is of course made. 6. Concerning demographics, much information on all enrolled students for the academic year 2016–2017, including their geographical origin, age and gender, can be found at for the most numerous undergraduate Degree Course, “Foreign Languages and Literatures”; at Report-0980-2016.pdf for “Asian Languages, Markets and Cultures”; and at for the language-oriented graduate Degree Course “Language, Society and Communication” (all last accessed 5/06/2017). A preliminary report on the initial implementation of these monitoring practices is provided in our presentation to the European Systemic Functional Linguistics Conference, 2017, in Salamanca: “A Fine-Tuned SSS as Appliable Pedagogical Stylistics: A Case Study”.

References Burke, M., Csábi, S., Week, L. and Zerkowitz, J. (Eds.). (2012) Pedagogical Stylistics. London: Bloomsbury. Clark, U. and Zyngier, S. (2003) Towards a pedagogical stylistics. Language and Literature, 12(4), pp. 339–351. Coetzee, J. M. (1986 [1987]) Foe. London: Penguin Books. Fogal, G. G. (2015) Pedagogical stylistics in multiple foreign language and second language contexts: A synthesis of empirical research. Language and Literature, 24(1), pp. 54–72. Fowler, R. (1986) Linguistic Criticism. Oxford: Oxford University Press. Fowler, R. and Bateson, F. W. (1967) Argument II. Literature and linguistics. Essays in Criticism, 17, pp. 322–347. Fowler, R. and Bateson, F. W. (1968) Argument II (continued): Language and literature. Essays in Criticism, 18, pp. 164–182. Goatly, A. (2004) Corpus linguistics, systemic functional grammar and literary meaning: A critical analysis of Harry Potter and the Philosopher’s Stone. Ilha do Desterro, 46, pp. 115–154. Halliday, M. A. K. (1982 [2002]) The de-automatization of grammar: From Priestley’s ‘An Inspector Calls’. In J. J. Webster (Ed.), Linguistic Studies of Text and Discourse, Volume 2 of The Collected Works of M. A. K. Halliday. London: Continuum, pp. 126–148. Halliday, M. A. K. (1985a) An Introduction to Functional Grammar. London: Edward Arnold (revised 2nd edition 1994; revised 3rd edition, with C. M. I. M. Matthiessen, 2004).


Donna R. Miller and Antonella Luporini

Halliday, M. A. K. (1985b) Systemic background. In J. D. Benson and W. S. Greaves (Eds.), Systemic Perspectives on Discourse, Volume 1 of Selected Theoretical Papers From the 9th International Systemic Workshop. Norwood, NJ: Ablex Publishing Corporation, pp. 1–15. Halliday, M. A. K. (2002 [2009]) Applied linguistics as an evolving theme. In J. J. Webster (Ed.), Language and Education, Volume 9 of The Collected Works of M. A. K. Halliday. London: Continuum, pp. 1–19. Halliday, M. A. K., MacIntosh, A. and Strevens, P. (1964) The Linguistic Sciences and Language Teaching. London: Longman. Hasan, R. (1964) A Linguistic Study of Contrasting Linguistic Features in the Style of Two Contemporary English Prose Writers. Unpublished PhD thesis, University of Edinburgh. Hasan, R. (1985 [1989]) Language, Linguistics and Verbal Art. Oxford: Oxford University Press. Hasan, R. (2007) Private pleasure, public discourse: Reflections on engaging with literature. In D. R. Miller and M. Turci (Eds.), Language and Verbal Art Revisited: Linguistic Approaches to the Study of Literature. London: Equinox, pp. 41–67. Hasan, R. (2011) A timeless journey: On the past and future of present knowledge. In Selected Works of Ruqaiya Hasan on Applied Linguistics. Beijing: Foreign Language Teaching and Research Press, pp. xiv–xliii. Jakobson, R. (1960) Closing statement: Linguistics and poetics. In T. A. Sebeok (Ed.), Style in Language. Cambridge, MA: MIT Press, pp. 350–377. Jakobson, R. (1966) Grammatical parallelism and its Russian facet. Language, 42(2), pp. 399–429. Lukin, A. (2008) Reading literary texts: Beyond personal responses. In Z. Fang and M. J. Schleppegrell (Eds.), Reading in Secondary Content Areas: A Language-based Pedagogy. Ann Arbor, MI: University of Michigan Press, pp. 84–103. Lukin, A. (forthcoming) Language, linguistics and verbal art: The contribution of Ruqaiya Hasan to the study of literature. In R. Wegener, A. Oesterle and S. Neumann (Eds.), On Verbal Art: Essays in Honour of Ruqaiya Hasan. Sheffield: Equinox. Luporini, A. (2016) Spotlighting fantasy literature with the tools of frame semantics and systemic functional linguistics: A case study. In D. R. Miller (Ed.), Quaderni del CeSLiC: Occasional Papers. Bologna: CeSLiC and ALMADL. doi:10.6092/unibo/amsacta/5162, pp. 1–31. Macken-Horarik, M., Love, K. and Unsworth, L. (2011) A grammatics ‘good enough’ for school English in the 21st century: Four challenges in realising the potential. Australian Journal of Language and Literacy, 34(1), pp. 9–23. Macken-Horarik, M., Sandiford, C., Love, K. and Unsworth, L. (2015) New ways of working ‘with grammar in mind’ in School English: Insights from systemic functional grammatics. Linguistics and Education, 31, pp. 145–158. Mahboob, A. and Knight, N. (2010) Appliable linguistics: An introduction. In A. Mahboob and N. Knight (Eds.), Appliable Linguistics. London: Continuum, pp. 1–12. Martin, J. R. (2000) Beyond exchange: Appraisal systems in English. In S. Hunston and G. Thompson (Eds.), Evaluation in Text: Authorial Stance and the Construction of Discourse. Oxford: Oxford University Press, pp. 142–175.

Systemic Socio-Semantic Stylistics (SSS)


Martin, J. R. and White, P. R. R. (2005) The Language of Evaluation: Appraisal in English. Basingstoke: Palgrave Macmillan. Matthiessen, C. M. I. M. (2012) Systemic functional linguistics as appliable linguistics: Social accountability and critical approaches. D.E.L.T.A., 28(Especial), pp. 435–471. McIntyre, D. (2011) The place of stylistics in the English curriculum. In L. Jeffries and D. McIntyre (Eds.), Teaching Stylistics. Basingstoke: Palgrave Macmillan, pp. 9–29. McKay, S. (1986) Literature in the ESL classroom. In C. Brumfit and R. Carter (Eds.), Literature and Language Teaching. Oxford: Oxford University Press, pp. 191–198. Miller, D. R. (1998) Insegnare la lingua speciale del testo letterario: L’approccio sociosemiotico. In M. Pavese and G. Bernini (Eds.), L’apprendimento linguistico all’Università: le lingue speciali. Roma: Bulzoni, pp. 271–293. Miller, D. R. (2000) A linguistic approach to the teaching of Lawrence’s ‘verbal art’ to non-native speakers. DHLR (D.H. Lawrence Review), 29(3), pp. 66–69. Miller, D. R. (2016) Jakobson’s place in Hasan’s social semiotic stylistics: ‘pervasive parallelism’ as symbolic articulation of theme. In W. Boucher and J. Liang (Eds.), Society in Language, Language in Society: Essays in Honour of Ruqaiya Hasan. Basingstoke: Palgrave Macmillan, pp. 59–80. Miller, D. R. (forthcoming) Language and literature. In G. Thompson, W. L. Bowcher, L. Fontaine and J. Y. Liang (Eds.), Cambridge Handbook of Systemic Functional Linguistics. Cambridge: Cambridge University Press, Chapter 27. Miller, D. R. and Luporini, A. (2015) Social semiotic stylistics and the corpus: How do-able is an automated analysis of verbal art? In A. Duguid, A. Marchi, A. Partington and C. Taylor (Eds.), Gentle Obsessions: Literature, Linguistics and Learning: In Honour of John Morley. Rome: Artemide, pp. 235–250. Miller, D. R. and Luporini, A. (forthcoming) Software-assisted systemic sociosemantic stylistics—appraising tru* in J. M. Coetzee’s Foe. In R. Wegener, A. Oesterle and S. Neumann (Eds.), On Verbal Art: Essays in Honour of Ruqaiya Hasan. Sheffield: Equinox. Moore, J. and Schleppegrell, M. (2014) Using a functional linguistics metalanguage to support academic language development in the English Language Arts. Linguistics and Education, 26, pp. 92–105. Mukařovský, J. (1964) Standard language and poetic language. In P. Garvin (Ed. and Trans.), A Prague School Reader on Aesthetics, Literary Structure and Style. Washington, DC: Georgetown University Press, pp. 17–30. Mukařovský, J. (1977) The Word and Verbal Art: Selected Essays by Jan Mukařovský, edited and translated by J. Burbank and P. Steiner. New Haven and London: Yale University Press. Mukařovský, J. (1978) Structure, Sign and Function: Selected Essays by Jan Mukařovský, edited and translated by J. Burbank and P. Steiner. New Haven and London: Yale University Press. Rowling, J. K. (2001) Harry Potter and the Sorcerer’s Stone. New York: Scholastic. Sassoon, S. (1918) Does it Matter?, The First World War Poetry Digital Archive,, last accessed 20/12/2017.


Donna R. Miller and Antonella Luporini

Scott, M. (2012) WordSmith Tools Version 6. Liverpool: Lexical Analysis Software. Simpson, P. (2004 [2014]) Stylistics, 2nd ed. London: Routledge. Zyngier, S., Fialho, O. and Rios, P. A. D. P. (2007) Revisiting literary awareness. In G. Watson and S. Zyngier (Eds.), Literature and Stylistics for Language Learners. Basingstoke: Palgrave Macmillan, pp. 194–209.

13 Reading and Writing Across the Curriculum Teresa Benítez, Norma Barletta, Diana Chamorro, Jorge Mizuno and Gillian Moss

1. Introduction Reading and writing are exceptionally important processes at all educational levels, especially in higher education, considering their prominent role in the acquisition and communication of knowledge in the different areas. Future professionals permanently apply these skills to learn disciplinary content, keep updated about the latest advances in their fields, and participate in the academic and professional communities sharing their practices and theoretical constructions. However, not all students bring to the university the necessary skills to comprehend and produce the genres of their disciplines, which are more complex and have different characteristics from the genres they used to manage at school. There is thus a gap between the preparation students receive in high school and the academic challenges of the university; a gap which constitutes an obstacle to optimal performance in the different disciplines. This gap has attracted researchers’ attention in different parts of the world (Martin and Rose, 2003/2007, 2008; Rose and Martin, 2012; Carlino, 2005; Fernández and Carlino, 2010), giving rise to a series of initiatives and interventions intended to close the breach. In some Latin American countries such as Chile, Argentina and Colombia, interventions are implemented through courses intended to enhance students’ reading and writing competences at some point during the first college year (Narvaja de Arnoux, Di Stefano and Pereira, 2011; Moyano, 2007; 2010; Moyano and Natale, 2012; Peña, 2011). Nevertheless, the efficiency of these courses is questioned because the process is not continued across the curriculum. Although recent initiatives emphasize the work with the specific genres of the disciplines, such interventions are scarce, and information about the outcomes is still limited. Researchers from the Language Department of Universidad del Norte in Barranquilla, Colombia, also identified low reading and writing performance among new students, and academic problems associated with that deficiency. Studies focusing on first year students’ reading and writing, and the way content-area teachers fostered the development of these competences in their classes, were conducted between 1999 and 2005

250 Teresa Benítez et al. (Colectivo Comunicaciones, 2002). Results confirmed the weaknesses in high school preparation for the academic processes at the university. Likewise, a marked emphasis on content teaching in the disciplines also corroborated that content-area instructors usually take for granted students’ ability to read and write and do very little—or nothing—to help them approach successfully the texts of their disciplines. These results led to a series of initiatives, including workshops to provide contentarea teachers with theoretical and methodological tools to guide their students’ reading and writing in their subjects. However, they were not entirely successful due to a lack of institutional policy and a clear methodology for accompanying teachers. In 2006, the institution welcomed two general Communicative Competence courses—rooted in genre theory and genre pedagogy as developed by the Sydney School (Rose and Martin, 2012)—that first year students needed to take as basic subjects during their first year of college studies. These courses lay the groundwork, but they have been insufficient to familiarize students with the specific genres of the disciplines. Therefore, it was decided to implement a programme named Communicative Efficacy, whose main purpose is to strengthen students’ communicative competences across the curriculum, in order to enhance efficient learning and communication in their academic and professional lives. This initiative consists of a teacher development programme intended to provide content-area teachers with theoretical and methodological tools to continue enhancing students’ reading and writing competences across the curriculum. Teacher-researchers from the Language Department piloted the programme between the years 2010 and 2012. Adjustments were made as the results of the pilot were analysed. In 2014, Communicative Efficacy became an institutional programme. In this chapter we will describe the characteristics of the programme: theoretical foundations, methodology, teacher development workshops, genres identified in the corpus of texts used by the content-area teachers, achievements and challenges.

2. Theoretical Foundations of the Programme Since its inception, Systemic Functional Linguistics (SFL) has been characterized by its conviction of the necessity for a socially accountable linguistics, in contrast with the more formalist tradition in linguistics largely prevalent in the twentieth century. This has led Halliday and his colleagues and disciples to think of SFL as an appliable linguistics; that is to say, a type of linguistics which has the potential to be applied to the solution of real-world problems in communities (Halliday, 1985/2003). The creation and implementation of linguistically based programmes of community-oriented action requires the empowerment of practitioners through the democratization of access to semiotic resources (Matthiessen,

Reading and Writing Across the Curriculum


2012), hence the enormous importance of SFL-based programmes in educational settings. Genre theory (Martin, 1992; Martin and Rose, 2003/2007, 2008) and genre pedagogy (Rose and Martin, 2012), in particular, imply rejection of the educational model in which knowledge and understanding of text structures and their relation to social communicative purposes are the exclusive domain of expert teachers and professors. Instead, the aim is to make these structures and purposes explicit and accessible to teachers and professors in all fields as well as to students. Such programmes necessarily require the integration of theory and practice and a transdisciplinary approach permitting linguists to work together with experts in other disciplines (Matthiessen, 2012). The theoretical strength offered by SFL in such contexts is that it is a holistic theory of language in context which permits the elaboration of comprehensive descriptions of languages, including the genres and registers used in central social institutions such as schools and universities (Christie and Martin, 1997) and also those which are specific to each discipline. Such descriptions also include the registerial relations between contexts of situation and features of lexicogrammar used in texts. One of the practical strengths offered by SFL in this regard is the teaching and learning cycle proposed in genre pedagogy (Rose and Martin, 2012), which involves a Vygotskian socio-cultural approach to the use of scaffolding and accompaniment in classroom settings in order to make the theory accessible to teachers and students. This cycle will be described in more detail below. Inspired by this tradition of appliability and social accountability, the Communicative Efficacy programme at Universidad del Norte adopted SFL and, in particular, genre theory and pedagogy as the theoretical and practical underpinning which would enable us to work successfully with professors from other disciplines in order to facilitate students’ access to the semiotic resources of their chosen discipline.

3. Methodology For the implementation of this programme we adapted the stages of Participatory Action Research (PAR) design, which involves participants in a cycle of planning, action and observing, and reflection in order to analyse their own practice in terms of weaknesses and strengths, and implement improvements as a result of this reflection (Kemmis and McTaggart, 1990; McGregor and Cartwright, 2011). In the planning stage, the first step was the selection of the genre to be taught and the design of a series of activities based on the teaching and learning cycle of the genre pedagogy proposed by Martín and Rose (2005). This cycle includes three main stages: deconstruction, joint construction and independent construction (Martin, Christie and Rothery, 1987; Martin and Rothery, 1990; Martin, 1997; Rose and Martin, 2012).

252 Teresa Benítez et al. There is also a preliminary stage, ‘building the context’ (Feez and Joyce, 1998), which Martín and Rose (2005) consider a permanent feature of the cycle. The purpose of this is to familiarize the students with the social purpose and the context of production and circulation of the chosen text. In the deconstruction stage the teacher demonstrates how the structural pattern of the text (its phases and stages) unfolds in the text and how the lexical items, grammar forms and cohesive devices contribute to the construction of the meanings of the text. During the joint construction stage, students and teacher collaborate to produce a text similar to the one deconstructed. Students may also work in small groups to produce the text while the teacher supervises and offers support as needed (Moyano, 2010, 2012) In the independent construction stage students are asked to produce individually a text of the chosen genre. Students are encouraged to assume increasing responsibility for the writing process. Feez and Joyce (1998) propose a fifth stage they call ‘linking to related texts’ in which the teacher and students compare their texts with other texts of the same or a different genre or register. Martín and Rose (2005) expect students to control and reach a critical approach towards the text. One of the strengths of the genre pedagogy is that by providing enough scaffolding to the students, it enables them to attain more control over the genre than they would have if they had been left alone (Moss, 2016). In the second stage of the PAR cycle (action), we defined four steps for the implementation of the programme: awareness, planning, implementation/ evaluation and consolidation. The first step, awareness, consisted of a series of workshops on topics such as SFL principles, genre theory and some worldwide successful cases of the genre-based pedagogical approach. Thirty teachers from different disciplines attended these workshops. After that, twenty-two of them agreed to continue in the programme and work towards the transformation of their class methodology. Each of these content-area teachers was assigned as a partner a specialist in reading and writing from the Spanish Language Department. At this point, an expert in genre pedagogy was brought in as an advisor to ensure unification of methodological and analytical criteria. The second step involved collaborative work between the Spanish language teachers and their partners from the other disciplines. The latter selected a text to be read by their students in their classes and another text to be used as a model for written production. Both texts were analysed from the SFL perspective by the language partner and the programme advisor. Once the texts were analysed and discussed by the language partner and the advisor, the analyses were presented and explained to the content-area teacher. For the sake of the preparation of the reading class, special attention was paid to possible difficult passages in the text, for instance, a high lexical density, a great number of

Reading and Writing Across the Curriculum


nominalizations, technical lexis, incomplete information and so on. Taking this into account, teachers planned the lessons in which they would teach the text, following the teaching and learning cycle (Martín and Rose, 2005) of genre pedagogy. For the preparation of the writing class, the focus was on the stages of the chosen genre, the use of cohesion devices and lexicogrammatical choices appropriate to the academic register. These processes were accompanied by both the language partner and the advisor. During the third step, teachers implemented the planned reading and writing classes. The texts written by the students were assessed by the language teachers and their partners from the other disciplines. Then the results were compared, and feedback was provided to the students. The lessons were video-recorded and analysed independently by both partners based on a reflection guide. These analyses were then shared, and the class development, difficulties, achievements and challenges were discussed. These meetings were audio-recorded and served as data for the assessment of the programme. Besides, teachers wrote a reflection about the process of teaching and assessing reading and writing in their classes. The fourth step (consolidation) consisted of a new implementation of the planned lessons by the teacher, but in a more autonomous way. The partners met less frequently in order to share opinions about the development of the lessons. These meetings were also audio-recorded. The programme lasted three semesters: stages 1 and 2, one semester; stage 3, one semester; and stage 4, one semester. Table 13.1 summarizes the process described above.

Table 13.1 Participatory Action Research (PAR) cycle and the implementation of the programme PAR cycle stages Programme Planning

Acting and observing


Awareness (workshops on SFL, genre pedagogy, successful cases of the pedagogy). Planning: Spanish teachers analysed the genres and linguistic features of the disciplinary texts; Spanish and content-area teachers jointly prepared a reading and a writing cycle. Content-area teachers implemented their planned lessons in two semesters. In the first semester there was close guidance by the Spanish teachers. In the second semester, contentarea teachers received less guidance and developed more autonomy. Spanish partners collected data through videorecording, teacher reflection journals and students’ survey. Data were jointly analysed, and implications about achievements and challenges in the implementation of the programme were derived.

254 Teresa Benítez et al.

4. Teacher Development Two workshops were carried out prior to the planning stage. The first was for those Spanish teachers who were unfamiliar with genre pedagogy at the outset of the programme. This workshop presented some of the basic principles of SFL, emphasizing the importance of (a) the concept of language as a meaning-making resource and (b) the aim of democratization of access to semiotic resources. It then presented the basic principles of genre theory as developed by the Sydney School (Martin, 1992; Christie and Martin, 1997; Martin and Rose, 2003/7) and an introduction to genre pedagogy (Martín and Rose, 2005; Rose and Martin, 2012). The workshop concluded by modelling the application of the teaching and learning cycle with an academic text. Teachers were provided with a bibliography and encouraged to read further on the topic. The second of the workshops held prior to the planning stage of the programme was for both the Spanish teachers and their partners from the other disciplines.1 This second workshop provided a rationale for reading and writing programmes at the tertiary level in terms of the need to teach students to deal with both a new context of culture and new genres and registers. Such programmes require knowledge of contexts and texts, a theory of learning and a pedagogical approach. The workshop went on to present basic concepts of genre theory and an adaptation of the teaching and learning cycle for the tertiary level. The pedagogical design draws on the Sydney School genre pedagogy and the socio-cultural pedagogy of Vygotsky (1934/1986), Bruner (1987) and Wertsch (1985). The programme implemented at the General Sarmiento National University in Buenos Aires was described, emphasizing the importance of institutional policy and close collaboration between Spanish teachers and their partners from other disciplines. The third workshop in the series was held after the planning stage of the programme and before the implementation stage and was aimed at the partner teachers from other disciplines. By this stage, they had worked in detail, with their Spanish partners, on the analysis of the generic and registerial characteristics of the texts chosen for the reading activities and were planning the pedagogical process to be applied. The workshop consisted mainly of the modelling of a reading class, using a linguistics text in order to reproduce the situation in which the teacher is an expert in the topic of study while the students are novices. By way of introduction to the activities, the basic principles of the Vygotskian socio-cultural pedagogical approach were reiterated: the need for initial scaffolding in order to maximize experiences of success, followed by strategy development and the gradual removal of scaffolding in order to promote autonomous learning and make it unnecessary to cover all content in class. Six stages were suggested for implementation of the reading class and modelled during the workshop:

Reading and Writing Across the Curriculum 1. 2. 3. 4. 5. 6.


Building context Purpose and structure of the text Detailed reading of difficult sections of the text Re-representation (in another semiotic mode) of ideas from the text Reaction to the text Self-assessment

These six stages are an elaboration on the deconstruction phase of the Sydney School teaching and learning cycle, providing detailed suggestions for specific activities as support for teachers unused to working explicitly with reading. The final workshop in the series was carried out after the implementation stage of the programme and before the consolidation stage and was again aimed at the partners from other disciplines. At this point, all the partners had implemented their reading lessons, and about half had also implemented writing lessons. They had also discussed these classes with their Spanish teacher partners. The leader of the workshop had viewed and analysed a large proportion of the (video-recorded) lessons, both reading and writing. One of the characteristics observed was that partners tended to concentrate exclusively on the text in hand, without making explicit to students the idea that strategies used for this text could be used in tackling other texts. It was therefore decided to centre the workshop activities on reading strategies. The workshop began by reviewing the basic pedagogical principles being applied in the programme and commenting on activities observed in the implementation classes. A number of reading strategies were recommended and related to each of the six stages in the reading lesson. Examples were modelled, with the teachers in the workshop using texts which had been worked on in classes during the implementation stage. The reading strategies recommended were derived from three principal sources: •

SFL: unpacking of grammatical metaphors, use of macroTheme and hyperThemes, identification of voices in the text (Halliday and Martin, 1993; Martin and Rose, 2003/7; Martin and White, 2005). Academic reading strategies (Davies and Greene, 1984), e.g. DARTs (Directed Activities Related to Texts), and identification of textual units (Davies, 1995). Work on reading strategies in the EFL context (Oxford, 1990), e.g. cognitive and metacognitive strategies.

The workshop included microteaching activities in which partner teachers selected appropriate strategies for working with their own texts and for a specific stage in the reading lesson and modelled them with their colleagues.

256 Teresa Benítez et al. A review of activities observed in implementation classes was presented with a view to sharing ideas and suggestions among partners. The final session of the workshop was devoted to the ‘re-representation of ideas’ stage of the reading lesson. It was decided to give prominence to this stage, not only because of its importance in developing reading skills but also because very few of the partner teachers had included work on this stage in their lessons, probably because they were uncertain of how to approach it. The importance of this stage was made more explicit, in terms of cognitive development, favouring of diverse learning styles, encouragement of collaborative work and interpretative skills. Some of the activities discussed were labelling diagrams and constructing synoptic tables, flowcharts, tree diagrams, Venn diagrams, concept maps and model building (as, for example, in architecture). There was no workshop addressing specifically the teaching of writing in the joint construction or independent construction stages. The teachers that taught a specific genre for writing received one-to-one orientation regarding how to apply the reading for writing cycle.

5. Genres The genres found in the texts analysed largely corresponded to the elemental genres of school science identified by Rose and Martin (2012) and Dreyfus, Humphrey, Mahboob and Martin (2015). We identified instances of stories (narrative), informing genres such as histories (biographical recount, historical recount and historical account), explanations (sequential, factorial, consequential), reports (description and classification), procedures (procedural recount) and one interview. Finally, we found evaluating genres in the form of arguments (exposition and discussion). The most common genres were Arguments (eight expositions, two discussions) and Informing genres. We also found two factual stories and one narrative. In Table 13.2 we synthesize the distribution of elemental genres found in the texts used for reading by the teachers involved in the project. In some of the texts we identified only one elemental genre; such was the case of the biographical recount on Ethel Fenwick, the interview with Mies van der Rohe and the narrative text about Ann Mulcahy at the Xerox Company. Though the text about Ann Mulcahy has the stages and the engaging purpose of a narrative, it is different from the typical literary narrative in that it also has an informative and didactic purpose. Table 13.3 shows a description of the stages in the text “The Challenge of the Manager: How Anne Mulcahy Created a New Xerox” (Gareth and George, 2010) and how they were clearly signalled through the sequence of hyperThemes. Most of the time, however, elemental genres were embedded in “longer texts . . . stringing elemental genres together, one after another, until the

Background Explanation of events

Historical account



Orientation Sequence of events

Biographical recount

Factual stories

Classification Description: types Classification Description Phenomenon Factors Phenomenon Explanation Phenomenon Consequences

Classifying report



Descriptive report Factorial

Definition Exemplification


Orientation Complication Resolution Evaluation





Genre group

Mechanical Engineering





Business Administration


Table 13.2 Elemental genres identified across the texts1 analysed

“Foundations of Metal Smelting” (Groover, 2007)

“The Concept of Education” (Fermoso, 1982)

“The Challenge of the Manager: How Anne Mulcahy Created a New Xerox” (Gareth and George, 2010) “Crowning the Edifice: Ethel Fenwick and State Registration” (Griffon, 1995) “Care in the History of Nursing” (Gonçalves, de Sena and Gómez Serrano, 2000) “Metapsychology: Psychoanalytical Concept of Personality” (Brainsky, 1986)



Manufacturing Processes Fifth semester

Adult Critical Adult Care Seventh semester Psychodynamic Tendencies and Authors Fourth semester Nursing in Education Fifth semester

Introduction to Nursing First semester

Evolution of Administrative Thought Second semester


Issue Sides Resolution Questions Answers



The titles of texts have been translated from Spanish.


Thesis Arguments Reiteration



Purpose Method Results

Procedural recount




Genre group

Table 13.2 (Continued)

“Understanding the Concepts of Play and Its Pathologies” (Bolívar, 1998) Administrative Law. Definition, characteristics, nature (Santofimio, 2003) “Conversations with Mies van der Rohe—American Certainties” (Puente, 2006)

Child Education




“Use of Metakaolin, Waste Glass and Optical Fibre in the Development of Translucent Concrete” (Franco, Pérez and Çruz, 2013) “Existential Foundations of the Human Mind” (De Castro and García, 2011)


Civil Engineering


History of Architecture Fifth semester

Administrative Law Sixth semester

Play and Fun Third semester

Existential Tendencies and Authors Fifth semester

Construction Materials Fifth semester


Reading and Writing Across the Curriculum


Table 13.3 Stages in the narrative text Stage Orientation




In 2001, Xerox, the well-known photocopier company, was about to go bankrupt . . . The person chosen to plan the transformation of the company was Ann Mulcahy. . . . The main administrative challenge faced by Mulcahy was to decide how to modify the structure and the control systems of the company to reduce Xerox’ high operation costs and save millions of dollars. . . . she had to decide how to better invest the remaining money in research to innovate new types of digital copying machines. To find a solution to this problem, Mulcahy used her leadership abilities By carefully listening both to employees and to clients, Mulcahy . . . obtained new elements leading to new strategies that have transformed the line of products of the company. . . . working closely with clients, Mulcahy and her administrators have formulated new strategies for Xerox, based on innovating with improved products and services. The result was that the profits increased dramatically.

work that needs to be done is accomplished” (Dreyfus et al., 2016, p. 39). These longer texts constitute macro-genres (Martin and Rose, 2008). These comprise a series of linked segments, each of which has its own generic structure. Some macro-genres have already been described in the existing literature. Such is the case of the research report (Swales, 1990), an example of which we found in “Use of Metakaolin, Waste Glass and Optical Fibre in the Development of Translucent Concrete”. Table 13.4 shows how it unfolds. The text “Foundations of Metal Smelting” (Groover, 2007) can be considered a macro-report (Dreyfus et al., 2016). It includes a number of reports as needed to construct the field. Table 13.5 contains a description of the genres involved in the initial sections. An interesting characteristic found across several genres in the academic texts analysed was a stage we called Preamble. It is a contextualization or orientation, which makes reference to the importance of the topic in the history of the discipline or for humankind in general. Its function is to introduce the topic to a reader, a newcomer to the field or the topic, thus softening the task of learning it. Regarding the linguistic features of the genres studied, the ones that are likely to represent difficulties for the students are technical terminology, the use of grammatical metaphors especially in complex nominal groups, syntactic intricacy, abstract and decontextualized language, and in some cases texts with faulty syntax, spelling, punctuation or translation.

Table 13.4 Stages of the research paper Abstract

MacroTheme for the whole article


A contextualizing descriptive report on the general topic of construction materials and on the materials which are the object of study (concrete and glass) and a research warrant for the worthiness of the study (Hood, 2010). A formal statement of the objective of the research as a macroNew at the end of the introduction and at the beginning of the description of the experiment. This consists of two phases: Materials and Methods. The former entails a number of descriptive and compositional reports describing characteristics and amounts of the materials used in the experiment. Compositional reports were instantiated through non-verbal texts in tables. The Methods phase describes a series of processes carried out in the experiment. The purpose of this stage is achieved mainly through consequential explanations. It also includes Descriptive Reports. This part works as macroNew. It includes a summary of the experiment and its results, as well as an evaluation thereof.

Purpose Method

Results Conclusions

Table 13.5 Macro-report components Sections




Definition Descriptive report

Certain terms are defined: smelting, ingot. It describes advantages and disadvantages of smelting

Section 1: Overview of smelting technology Sub-section 1: Smelting process



Sequential explanation Classification report

Sub-section 2: Moulds for sand smelting

Descriptive report

How to carry out a smelting operation Smelting processes are classified into two categories It describes the basic characteristics of a mould Explanation of the process of sand smelting

Sequential explanation Section 2: Heating and pouring Sub-section 1 Sub-section 2

Genres Descriptive report Sequential explanation Factorial explanation

Purpose The formula of calorific energy is explained It explains factors affecting the pouring of metal

Reading and Writing Across the Curriculum


6. Achievements and Challenges In this section we explain the achievements of the content-area teachers and their challenges in the adoption of a new pedagogy for teaching reading and writing in their subjects. Some teachers followed the key phases of the cycle (Feez and Joyce, 1998; Rose and Martin, 2012): explaining in detail the genre purpose and the settings in which it can be used; analysing a representative genre sample and identifying its phases; guiding the genre writing; involving students in independent writing; and giving feedback to the students. In the contextualization stage, they used a variety of strategies such as identifying the text type or genre to read and write, explaining their social purpose and importance, and clarifying the purpose of producing a text of this kind. For example,2 in an engineering class the teacher explains: The goal of this activity is [. . .] to write a text that reports research results derived from some laboratory tests that you have carried out throughout the semester based on the text you have in your hands, and which we are going to analyse [. . .] and I tell you who the authors of the article are. Why is it important for me, as a student, to write such texts? Who is the target audience? [. . .] Is it a professional or a technician related to particular subjects? In the deconstruction phase, teachers made explicit to their students the generic structure of the text to be read and written. That means they explained in detail each section or stage of the text, the key features and possible variations, the purpose of each section and the type of information in each one. Teachers find this phase useful because they have the opportunity not only to explain the importance, relevance and appropriateness of including each type of information and lexicogrammatical choices, but also to discuss the consequences of making such decisions with their students: (Reading from projected example) “(Full Name) legal age . . . identified as shown below my signature”. I like this kind of writing. Those people who put the number of the identification card above and then at the bottom . . . it’s possible to make a mistake writing the number of the identification card several times, then, if somebody says he/ she is going to identify themselves at the bottom, okay, go to the bottom . . ., there is less chance of making a mistake. Another very important achievement of teachers in this phase was the interaction. They inquired about students’ previous knowledge about the subject, which provided the appropriate time and space for building up a

262 Teresa Benítez et al. common starting point. Let’s look at an example. An engineering teacher asked for information that students could find in the conclusion section of a research article: S: T:

S: T: S: T:

That . . . metakaolin improves the properties . . . That means that you would make a conclusion that points out the advantages with respect to other component resistance, you would stress the percentage limit, which is an important conclusion. Tell me another advantage. It affects the durability in a positive way. It has a positive feature of durability because of a pore network that helps material permeability. It mentions the use of additives and the reasons. Take into account that they may be used and it’s necessary, because I cannot show all the positive aspects of the material either. The author has been clear in presenting the constraints, drawing attention not only to the advantages but to the constraints as well, and comparing against other studies. What does this mean? If I read one article, it doesn’t mean it’s the panacea of all materials.

In the joint construction stage, content-area teachers guide the writing through strategies, stressing the central phases or functions of the texts in order to practice structuring a whole text. In a Law class, when the teacher finishes modelling the Right of Petition genre, she organizes the students in small groups in order to write a sample. After twenty minutes, the teacher shows on the board one of the texts written by some of the students: This is the petition (enclosing it in a circle). Then, this petition should be here (she draws an arrow pointing to a section below) [. . .] I think that this could be improved in the writing (she encloses in a circle the last section). We are not told the working hours, we don’t know whether the contract is verbal or written . . . [. . .] I would say norms in general, to use a more technical language: “in the following norms and jurisprudence”. As we can see, the teacher makes some comments; she gives feedback about text organization, lack of information, wordings and text structure. She also suggests the use of technical language. The ability to recognize the features of a particular genre allows contentarea teachers to select the text to be written in class. In her reflection journal, one of the teachers expressed: A detailed analysis of the example used for modelling the writing helped me realize that the text I had selected wasn’t an ideal sample,

Reading and Writing Across the Curriculum


so I changed the activity: I asked the students to evaluate the way in which the text was written, and I decided to rewrite it. The above teacher reflection confirms Hyland’s premise that knowledge about a genre, the research article in this case, increases teachers’ awareness of texts and ability to confidently advise students on writing (Hyland, 2013). In addition, teachers’ statements evaluating the experience reflected that they had become aware of the following: a.

genres follow a recognizable, predictable pattern, and consequently genres can be taught; b. genre teaching is inherent to their content area, so it is their responsibility; c. time used in explicit teaching of genre is essential to reach an understanding of what is read and autonomy in writing; d. the explicit deconstruction of the genre and the modelling of reading strategies constitute scaffolding, which ensures students understand the topic taught, and guides them in the use of such strategies when dealing with other texts. Despite the number of achievements evident in the teachers’ journals and recorded dialogues with partners, the teachers and the programme in general are faced with a number of challenges: 1. In order to devote time to the implementation of their new syllabi, the content-area teachers should reduce the scope of their courses in terms of content so as to be able to enhance the students’ reading and writing abilities. Deciding on the balance between content and language teaching is not an easy task for teachers. 2. It is difficult for non-linguist teachers to identify the generic characteristics of the texts they ask their students to read and write. This requires dedication and initial close guidance by more expert colleagues. 3. Since the teachers are familiar with the texts they discuss in their classes, they are usually unaware of the complexities and challenges these texts pose to their students: abstract language, unknown terminology, poor translations, assumed knowledge, information gaps and ideological biases. 4. In the teaching to write cycle, teachers are willing to do the deconstruction of the text, but some did not implement the joint construction stage. Instead, they explained the characteristics of the texts students had to write and assigned the writing for evaluation. It seems that helping students to write in the classroom at the university level is not a usual activity.

264 Teresa Benítez et al. 5. For the Spanish teachers, it is a challenge to get familiar with the texts that are read in the disciplines. These texts are usually macro-genres, and they have mixed characteristics. It takes time to gather a corpus, analyse it, make a generic description and untangle the complexities they entail in order to devise proper methodological guidance for the content-area teachers to deal with them in their classes.

7. Conclusions The Communicative Efficacy programme at Universidad del Norte has achieved considerable success in a number of ways and for a number of reasons. The achievements were summarized in the previous section. Here, in the conclusion, we will set out what we believe to be the reasons for this success: an appropriate theoretical framework, combining SFL with Vygotskian socio-cultural pedagogical theory; commitment on the part of the teachers involved and their willingness to take up the challenge of learning and implementing new theoretical and methodological concepts; and clear institutional support. The choice of an SFL theoretical framework, in particular the Sydney School genre pedagogy, as the basis for designing and implementing a proposed new methodology across the curriculum, has given the programme a degree of coherence and conviction which had been lacking in previous initiatives. As regards the teachers, the programme has posed challenges for all, challenges which they have met, in the main, with enthusiasm and willingness to adapt and learn. The Spanish teachers have taken on the double task of learning about genre theory and genre pedagogy and of guiding the implementation of the pedagogy using texts from disciplines in which they are not experts. Their partners from other disciplines have faced new experiences in working together with a teacher from another discipline; in learning, accepting and implementing an unfamiliar methodology; and, perhaps most importantly, in assuming responsibility for supporting their students’ reading and writing processes. This last was particularly difficult for those—the majority—who had always assumed that students arrived at university knowing how to read and write academic texts and that their sole responsibility was to teach content. Furthermore, for many teachers, accustomed to working within a more constructivist, implicit pedagogical framework, adopting a more teacher-directed and explicit pedagogy meant a move out of their comfort zone and into uncharted territory. In this respect, the insistence on the Vygotskian socio-cultural approach, with its emphasis on initial scaffolding leading to ever-greater student autonomy, has been crucial for convincing teachers of the appropriateness of the pedagogy. Finally, the level of institutional support for the programme has been another decisive factor. The programme was presented to all faculties as

Reading and Writing Across the Curriculum


an institutional initiative that was to be taken seriously by all and would be taken into account for teacher evaluation purposes. Support has been provided for bringing in external advisers and for video-recording the class observations. Looking towards the future, in order to ensure the sustainability of the programme, it will be essential to count on continued institutional support. As well as involving an ever-higher number of teachers in the programme, it is of the utmost importance to ensure that those teachers who have already completed the three stages of the programme are offered continuing support in their use of genre pedagogy. If these conditions are met, we believe that Communicative Efficacy, in the long term, has the potential to transform the Universidad del Norte students’ performance in academic reading and writing tasks and, therefore, their general academic success.

Notes 1. The second workshop was run by Dr Estela Moyano of the Universidad Nacional General Sarmiento, Buenos Aires, who was invited as a guest lecturer and advisor to the programme. The other workshops were all led by Dr Gillian Moss, the Universidad del Norte team advisor. 2. The examples in this section have been translated from Spanish.

References Bruner, J. S. (1987) Actual Minds Possible Worlds. Cambridge, MA: Harvard University Press. Carlino, P. (2005) Concepciones y formas de enseñar escritura académica. Un estudio contrastivo. Signo & Seña, 16, pp. 71–117. Christie, F. and Martin, J. R. (Eds.). (1997) Genre and Institutions: Social Processes in the Workplace and School. London: Cassell. Colectivo Comunicaciones. (2002) Comprensión y competencias lectoras en estudiantes universitarios: resultados y recomendaciones de una investigación. Barranquilla: Ediciones Uninorte. Davies, F. (1995) Introducing Reading. Harmondsworth: Penguin. Davies, F. and Greene, T. (1984) Reading for Learning in the Sciences. Edinburgh: Oliver and Boyd. Dreyfus, S., Humphrey, S., Mahboob, A. and Martin, J. (2016) Genre Pedagogy in Higher Education. New York: Palgrave Macmillan. Feez, S. and Joyce, H. (1998) Text-based Syllabus Design. Sydney: National Centre for English Language Teaching and Research. Fernández, M. E. and Carlino, P. (2010) ¿En qué se diferencian las prácticas de la universidad y las de la escuela secundaria? Perspectivas de alumnos y profesores en Ciencias Humanas y Veterinarias. Lectura y Vida, 31(13), pp. 6–19. Gareth, J. and George, J. (2010) Administración contemporánea, 6th ed. México: McGraw Hill. Groover, M. (2007) Fundamentos de manufactura moderna. México: McGrawHill.

266 Teresa Benítez et al. Halliday, M. A. K. (1985/2003) Systemic background. In J. D. Benson and W. S. Greaves (Eds.), Systemic Perspectives on Discourse, Volume 1 (ADPS15), pp. 1–15. Reprinted in M. A. K. Halliday and J. Webster (2003) On Language and Linguistics, vol. 3 in the Collected Works of M. A. K. Halliday. London: Continuum. Halliday, M. A. K. and Martin, J. R. (1993) Writing Science: Literacy and Discursive Power. London: Falmer Press. Hood, S. (2010) Appraising research: Evaluation in academic writing. London: Palgrave Macmillan. Hyland, K. (2013) Genre and discourse analysis in language for specific purposes. In C. A. Chapelle (Ed.), The Encyclopedia of Applied Linguistics. Oxford: Wiley-Blackwell, pp. 2527–2534. Kemmis, S. and McTaggart, R. (1990) The Action Research Planner, 3rd ed. Victoria: Deakin University Press. Martin, J. R. (1992) English Text: System and Structure. Philadelphia: John Benjamins. Martin, J. R. (1997) Analysing genre: Functional parameters. In F. Christie and J. R. Martin (Eds.), Genres and Institutions. London: Cassell, pp. 3–39. Martin, J. R., Christie, F. and Rothery, J. (1987) Social processes in education: A reply to Sawyer and Watson (and others). In I. Reid (Ed.), The Place of Genre in Learning: Current Debates. Geelong, Victoria: Centre for Studies in Literacy Education, Deakin University, pp. 58–82. Martin, J. R. and Rose, D. (2003/2007) Working With Discourse: Meaning Beyond the Clause. New York: Continuum. Martín, J. R. and Rose, D. (2005) Designing literacy pedagogy: Scaffolding democracy in the classroom. In R. Hasan, C. Matthiessen and J. Webster (Eds.), Continuing Discourse on Language: A Functional Perspective. London: Equinox, pp. 252–280. Martin, J. R. and Rose, D. (2008) Genre Relations: Mapping Culture. Sheffield: Equinox. Martin, J. R. and Rothery, J. (1990) Literacy for a Lifetime: Teachers’ Notes. Sydney: Film Australia. Martin, J. R. and White, P. R. R. (2005) The Language of Evaluation: Appraisal in English. Hampshire: Palgrave Macmillan. Matthiessen, C. M. I. M. (2012) Systemic functional linguistics as appliable linguistics: Social accountability and critical approaches. D.E.L.T.A., 28, pp. 435–471. McGregor, D. and Cartwright, L. (2011) Developing Reflective Practice: A Guide for Beginning Teachers. New York: McGraw-Hill. Moss, M. G. (2016) La pedagogía de géneros. In M. G. Moss, T. Benítez Velásquez and J. Mizuno Haydar (Eds.), Textos que se leen en la universidad. Una mirada desde los géneros discursivos en la Universidad del Norte. Barranquilla: Editorial Universidad del Norte, pp. 19–26. Moyano, E. I. (2007) Enseñanza de habilidades discursivas en español en contextos pre-universitarios: Una aproximación desde la LSF. Revista Signos, 40(65), pp. 573–608. Moyano, E. I. (2010) Escritura académica a lo largo de la carrera: Un programa institucional. Revista Signos, 43(74), pp. 465–488.

Reading and Writing Across the Curriculum


Moyano, E. I. (2012) Hacia la caracterización de géneros profesionales: Algunas reflexiones teórico-metodológicas. In S. Nothstein, M. C. Pereira and E. Valente (Eds.), Libro de Actas del Congreso Regional de la Cátedra UNESCO en Lectura y Escritura: Cultura Escrita y Políticas Pedagógicas en las Sociedades Latinoamericanas Actuales. Los Polvorines: Universidad Nacional de General Sarmiento, pp. 1567–1586. Moyano, E. I. and Natale, L. (2012) Teaching academic literacy across the university curriculum as institutional policy. The case of the Universidad Nacional de General Sarmiento (Argentina). In C. Thaiss, G. Bräuer, P. Carlino, L. Ganobcsik-Williams and A. Sinha (Eds.), Writing Programmes Worldwide: Profiles of Academic Writing in Many Places: Perspectives on Writing. Fort Collins, CO: The WAC Clearinghouse and Parlor Press, pp. 23–34. Narvaja de Arnoux, E., Di Stefano, M. and Pereira, C. (2011) La lectura y la escritura en la universidad, 2nd ed. Buenos Aires: Eudeba. Oxford, R. (1990) Language Learning Strategies: What Every Teacher Should Know. Boston: Heinle and Heinle. Peña, F. (2011) Leer y escribir. Prácticas necesarias en la universidad. Educere, 15(52), pp. 711–719. Rose, D. and Martin, J. R. (2012) Learning to Write, Reading to Learn: Genre, Knowledge and Pedagogy in the Sydney School. Sheffield: Equinox. Swales, J. (1990) Genre Analysis: English in Academic and Research Settings. Cambridge: Cambridge University Press. Vygotsky, L. (1934) Thought and Language, translated by Eugenia Hanfmann 1986, Cambridge, MA: MIT Press. Wertsch, J. V. (1985) Culture, Communication, and Cognition: Vygotskian Perspectives. New York: Cambridge University Press.

Analysed texts Bolívar, C. (1998) Aproximación a los conceptos de lúdica y ludopatía. Proceedings of the VII Congreso Nacional de Educación Física y Recreación, Caldas, 3–8 November 1998. Manizales: Funlibre. Available at: www.redc (Accessed on 5 September 2014). Brainsky, S. (1986) Manual de psicología y psicopatología dinámicas: fundamentos de psicoanálisis. Bogotá: Carlos Valencia Editores. De Castro, A. and García, G. (2011) Psicología clínica: Fundamentos existenciales, 2nd ed. Barranquilla: Editorial Universidad del Norte. Fermoso, P. (1982) Teoría de la educación. Una interpretación antropológica. Barcelona: CEAC. Franco, D., Pérez, E. and Çruz, R. (2013) Uso del metacaolín, vidrio reciclado y fibra óptica en la elaboración de un concreto translúcido. ITECKNE, 10(2), pp. 158–166. Gareth, J. and George, J. (2010) Administración contemporánea, 6th ed. México: McGraw Hill. Gonçalves, A., de Sena, R. and Gómez Serrano, C. (2000) El cuidado en la historia de la enfermería. In R. de Sena (Ed.), Educación de la enfermería en América Latina. Bogotá: Universidad Nacional del Colombia R.E.A.L., pp. 3–31.

268 Teresa Benítez et al. Griffon, D. P. (1995) Crowning the edifice. Ethel Fenwick and state registration. Nursing History Review, 3, pp. 201–212. Groover, M. (2007) Fundamentos de manufactura moderna. México: McGrawHill. Puente, M. (2006) Conversaciones con Mies Van der Rohe: Certezas americanas. Barcelona: GG Editores. Santofimio, O. (2003) Tratado de derecho administrativo. Introducción a los conceptos de la administración pública y el derecho administrativo, 3rd ed. Bogotá: Universidad Externado de Colombia.

14 On the Relevance of the Textual Metafunction for Spanish Learners/Teachers of English1 Ana Elina Martínez Insua

1. Introduction: Initial Assumptions and Hypotheses Taking as a point of departure the assumption that the organisation of constituents is crucial if the act of communication is to be successful, the Theme position (i.e. the clause initial position from a Hallidayan point of view) in languages such as English becomes most important. Studies on Theme are abundant in the literature (Fries, 1981, 1992, 1994, 1995a,b; Martin, 1992, 1995; Berry, 1995, 2013; Matthiessen, 1995; Whittaker, 1995, among many others), and there are also more specific studies on the need for teachers of English as a foreign language to consider Theme, and especially thematic patterning, as a tool in their teaching and evaluation of their students’ writing (Alonso-Belmonte and McCabe-Hidalgo, 1998). The present piece of research is a modest attempt to contribute to this latter group of studies. The general aim is to give evidence of the appliability of Systemic Functional Linguistics to aid non-linguists, more specifically, those involved in the process of learning and teaching English. Alonso-Belmonte and McCabe-Hidalgo (1998, p. 15) analyse Theme/ Rheme as a very helpful “tool of instruction for the teacher to evaluate L2 writing at the level of discourse”, and claim (Alonso-Belmonte and McCabe, 2003) that “information progression is achieved through thematic patterning, and . . . can be used to help improve” learners’ production in English. Encouraged by such observations, the present study approaches Theme and hypothesises that gaining awareness of the features of the textual metafunction and the thematic structure of native English may help teachers and students of English as a foreign language. The assumption is that the exploration of Theme in native and learner English spoken texts can be exploited in order to help our English language learners achieve greater awareness of the need to pay attention to native organisation of constituents, as well as the need to seek resemblance to native texts in the organisation of their own production. We assume not only that Theme/Rheme is the “basic form of the organization of the clause as message” (Halliday, 1985, p. 53), or is relevant as a very helpful tool of instruction for teachers of English as

270 Ana Elina Martínez Insua L2 (Alonso-Belmonte and McCabe-Hidalgo, 1998), but also that paying attention to Theme/Rheme during the process of learning and teaching may help learners to gain awareness of its relevance in the construction and production of cohesive discourse, as well as to gain proficiency and efficiency in their foreign language production. To our knowledge, while Theme/Rheme in L2 writing has been analysed previously (e.g. Alonso-Belmonte and McCabe-Hidalgo, 1998), not much research has been conducted on the distribution of Theme/Rheme in learners’ spoken discourse; hence our present attempt to contribute to this area of study by focusing on oral production. In this line, detecting and analysing possible differences in the thematic components used by learners and natives may shed light on which features of English native discourse are critical to learners of the language when trying to achieve proficiency. As regards the methodology necessary for this type of analysis, we share O’Donnell’s (2013) belief that [a] learner corpus with texts allocated into proficiency levels is a useful resource when designing a curriculum for EFL grammar education, as it can provide insights into which grammatical features are most critical to the learner at each stage of their progress. (2013, p. 571) O’Donnell (2013, p. 574) posits that “[i]t is relatively easy to use a learner corpus to see what learners need to be taught”, especially in terms of lexis or syntactic features. As a matter of fact, as O’Donnell explains, comparisons between levels of usage of vocabulary or syntactic features have been made with the aim to reveal which lexis or syntactic features are under-used by a group of learners, and thus where more teaching would be valuable (e.g. Granger, 1997; Dagneaux, 1995, among others). Having Theme/Rheme as the main interest, the present study seems to be, however, beyond lexis and syntactic features. Notions such as cohesion were somehow neglected in language teaching up to the mid-1970s. Nowadays, by contrast, it is true that most general coursebooks of English as a foreign language and most reading and writing courses incorporate work designed to help learners grasp the cohesive devices of written English: discourse connectors, ellipsis, conjunctions, and so forth. However, L2 writing instructors still come across ESL student compositions in which sentences, assessed in isolation, are grammatically correct and yet the overall effect is one of incoherence. (Alonso-Belmonte and McCabe, 2003) Something similar seems to happen when one focuses on L2 oral discourse: grammatical well-formedness and incoherence sometimes co-occur in

Relevance of the Textual Metafunction


the learners’ production, while instruction does not always answer the question of why this happens, and does not reduce the possible effects of L1 transfer in discourse. As explained by Chocano et al. (2007, p. 1), English and Spanish, having fixed word order and free word order, respectively, differ in the devices they employ to order constituents in the sentence, and an “in-depth investigation into word order in advanced learners of L2 English . . . will thus offer answers to questions regarding . . . general issues related to L1 transfer” (Chocano et al., 2007, p. 1). The present study means to shed light on this circumstance, calling our attention to the relevance that the study of Theme and thematic organisation have in L2 learning and teaching. Further inspiration for the present study comes from Kaplan’s (1995, p. 21) claim that the logic expressed through the organisation of written text is culture-specific, such that “speakers of two different languages will organize the same reality in different ways” (McCabe and AlonsoBelmonte, 2000, p. 78). Given the impossibility to collect data in such an ideal situation, with speakers of the two languages at issue (namely English and Spanish) observing and organising the same reality, and bearing in mind McCabe and Alonso-Belmonte’s (2000, p. 77) claim that the way in which text producers view the field of discourse motivates their Theme selection, the data analysed in this study have been retrieved from two corpora of spoken English compiled following the same design of tasks, topics and transcription conventions, as Section 3.1 explains in detail. In a nutshell, one may say that the aim of the present study is to use both a learner corpus and a native control corpus so as to try and see if Spanish learners of English need to be taught anything about the thematic structure of native English. If so, an attempt will be made to see what learners need to be taught, and also to point towards where more teaching would be valuable or helpful. In this line, the analysis of the data will aim to test two main hypotheses. On the one hand, it is hypothesised that (i) the thematic areas produced by natives and learners when they speak English differ from each other. On the other hand, taking as a basis preliminary findings that indicate that marked Themes are more frequently present in the spoken production of Spanish learners than in native production, it is hypothesised that (ii) the use of marked Themes is more frequent in the English discourse produced by Spanish learners. Before going on with the study itself, it is necessary to say that the space constraints necessarily imposed on the study, linked to the small number of texts that can be covered in studies based on manual analyses, imply that it will not be possible to draw firm conclusions, and will prevent us from trying to draw hard and fast conclusions. The ultimate aim of this study is, rather, to find pointers and hypotheses for future work (Berry, Thompson and Hillier, 2014, p. 108), which will hopefully be developed in future stages of our project.

272 Ana Elina Martínez Insua Section 2 approaches the notion of Theme and clarifies which of its many characterisations is adopted in this study. Section 3 presents the study itself, describing first of all the samples of language used, and trying to provide a plausible discussion of the findings afterwards. Finally, Section 4 contains some concluding remarks and points towards possible lines of further research with which we would like to continue.

2. The Notion of Theme in This Study As pointed out, this piece of research pays attention to the features of Themes produced in English oral discourse by both Spanish university learners of English and university native speakers of English. Assuming that the meanings contained in Themes relate ultimately to the broader socio-cultural context (Forey and Thompson, 2008, p. 3), an attempt was made to work with samples that had been compiled on similar bases. Thus, an attempt was made to ensure that the contexts of production were as similar as possible. Independently from the samples analysed, however, one cannot deny the existence of an ongoing debate concerning the boundaries and interpretation of Theme. Taking as a point of departure the Hallidayan consideration of Theme, many scholars have worked on identifying it in texts, on determining what factors should be considered in deciding what to include in Theme, and on proposing specific guidelines for establishing boundaries between Theme and Rheme. For our purposes, we adopt Berry’s (2013, among others) notion of Theme, according to which “the Theme of a clause is everything up to the main verb of the clause” (2013, p. 248).2 Berry shares Brown and Yule’s (1983, p. 133) notion of Theme as an orienting element in that Theme is regarded not only as the starting point of the message, but also as an element with the role of connecting the message to what has already been said, and the discourse function of orienting “the listener/reader to the message that is about to be perceived and provid[ing] a framework for the interpretation of that message” (Fries, 1995b, p. 318). McCabe and Alonso-Belmonte (2000, p. 89)3 clearly explain the cognitive and communicative motivations of Theme selection, and describe it as a process influenced both by the way in which the speaker perceives reality and by the way in which that speaker tries to transmit such perception to addressees. In their own words, there are grounds to support the notion that thematic selection has a cognitive motivation, at least in part, in that Theme selection allows us a reflection of the speaker’s perception of reality. . . . Thus, the Theme of the clause, which formally is the initial element of the clause, functionally combines the expression of the speaker’s perception of reality and the concerns of the speaker to communicate

Relevance of the Textual Metafunction


that perception of reality to the listener. It is, thus, both cognitive, in the sense that it refers to the world of experience, and communicative, in the sense that it has a discoursal role. This combined function of Theme goes some way in explaining some of the problems involved with pinning down the function of Theme, and separates it from notions of givenness and aboutness. (McCabe and Alonso-Belmonte, 2000, p. 80) Following Berry, we concentrate only on one part of the Theme: the syntactic Subject or Subject Theme (SubjTh).4 In our aim to replicate Berry’s (2013) analysis of contentlight and contentful Subject Themes, only Themes of independent clauses are considered for inclusion in our database, as the main contribution comes from the thematic structure of independent clauses (see also Halliday, 1985; Brown and Yule, 1983). More specifically, only declarative main clauses are explored. Also following Berry, we shall be discussing “content weight in terms of meanings, in terms of what the Subject Themes refer to” (Berry, 2013, p. 249). Weight is regarded, then, as the amount of meaning or reference. The lesser the meaning of a Subject (for example of a pronominal form), the lighter it is regarded to be in terms of content. Conversely, the greater the amount of semantic content of the Subject (a noun or a nominal group), the more contentful it is taken to be. Besides content weight, the informational reference of SubjTh is taken into consideration in this study. Thus, as in Berry (2013, p. 252), (i) SubjTh consisting of personal pronouns with antecedents are assumed to be referring to ‘Given Topics’ (GivTops); (ii) SubjTh consisting of a noun/nominal group that refers to an aspect of a discourse topic which has been mentioned before are assumed to be referring to ‘Resumed Topics’ (ResTops), a label used for references to topics that were already introduced some time ago and, therefore, need to be reactivated; and (iii) SubjTh referring to aspects of the discourse topic which have not been introduced before are assumed to refer to ‘New Topics’ (NewTops). As explained by Berry, this classification draws on Dik’s (1997) distinction between New Topic (NewTop), Given Topic (GivTop) and Resumed Topic (ResTop),5 and is reinforced by the adoption of labels related to content weight: while SubjTh referring to GivTops can be regarded as contentlight, those referring to ResTops and NewTops are labelled as contentful (Berry, 2013, p. 258). In addition to these three informational categories, a fourth one, called ‘Qualified Resumed Topic’ (QResTop), is used in this analysis of the corpus.6 The label QResTop is applied to SubjTh that refer to aspects of the discourse topics which have been mentioned before and contain some kind of qualification bringing semantic nuances to the resumption of the topic (Berry, pers. comm.). These four categories are illustrated by some of the SubjTh in (1)–(3), drawn from the two samples under analysis, which were subsequently

274 Ana Elina Martínez Insua drawn from the corpora VICOLSE (Vigo Corpus of Learner Spoken English) and LOCNEC (Louvain Corpus of Native English Conversation). Section 3 will further characterise these two corpora: (1) VICOLSE Riding HoodResTop continued her way to her grandma’s house when she reach it he kno—sheGivTop knocked the door who is calling *answer—asked a voice from inside (. .) IGivTop am (.) grandma I’mGivTop your *niece I’mGivTop your granddaughter IGivTop bring you some food that mum that mamma gave me for you oh come in come in *answered a little voice from inside so sheGivTop came in and sheGivTop found her grandma in bed but was a surprise that she wasn’t her grandma it was the wolf who was *wearing her grandma’s nightclothes ha ha ha I’mGivTop going to eat you said the wolf but Little Red Riding HoodQResTop escaped running to call the woodman and who came with her to the house and killed the the wolf (. .) after that theyGivTop found the grandma *hided in the wardrobe7 (2) VICOLSE this roomResTop shows a very big mess as begin the morning as in the noon (. .) the owner of this roomNewTop was John Burnett must be a very messy person (. .) nothingNewTop is in his place (. .) or where it must be (. .) a sandwichNewTop is there in the morning and itGivTop remains there in the noon only with a bit a bit given to it (3) LOCNEC [unorthodox English teacher . erm form a group called Dead Poets Society8 mhm where they erm . . read poetry and and treat it as a . a living thing . and erm . . one of the boys isNewTop particularly inspired by this erm . besides he wanted to become an actor (. . .) erm but he decides he wants to become an actor . and erm he takes part in a play against his erm . father’s wishes and er . his father’s furious with him and eventually the boyResTop commits suicide The SubjTh the owner of this room (2) and one of the boys (3) are NewTops in that they refer to aspects of the discourse topic which are mentioned for the first time. Notice that Riding Hood (1) and this room (2) are analysed as ResTops even though these extracts occur at the beginnings of their respective discourses. The reason why these two SubjTh cannot be said to refer to NewTops is that, even if they have not been previously

Relevance of the Textual Metafunction


mentioned in the discourse (they are discourse-initial elements), Riding Hood and the room were present in the prompts given to the speakers (see Section 3.1.1 on the tasks carried out by the speakers). The SubjTh The boy (3), by contrast, refers to a ResTop that has been previously mentioned in the discourse before other entities had been introduced. It is true that although there is not a great distance between the first and second mentions of this topic (the boy), distance is not the only relevant matter here, but rather the existence or not of other entities in between the two mentions. Brown and Yule’s (1983, p. 173) alternative term for (Berry’s) ResTop, ‘displaced given entity’, is helpful here, as it makes reference to the fact that an entity can only be regarded as a GivTop if it has not been displaced by another entity. In and erm . . one of the boysNewTop is particularly inspired by this erm . besides heGivTop wanted to become an actor, the boy is alluded to twice, by the nominal group one of the boys and by the pronoun he. This second mention, the third person pronoun he, refers to a GivTop. This can be regarded as a GivTop because no other entity has occurred between it and the first reference that would displace it from the GivTop status. By contrast, when the boy is mentioned again in the following message, when the speaker says erm but he decides he wants to become an actor . and erm he takes part in a play against his erm . father’s wishes and er . his father’s furious with him and eventually the boyResTop commits suicide, the SubjTh the boy cannot be regarded as a GivTop, because although the boy has been mentioned before in the immediately preceding discourse, other entities have been introduced to displace the boy from the GivTop status (e.g. his father). This explains why it has to be referred to by a noun/nominal group and cannot be referred to by a pronoun, which would be ambiguous. The case of the SubjTh Little Red Riding Hood (1) is a slightly different one. Because it is qualified (by means of the modifiers Little Red), it is giving us a more complete, modified view of Riding Hood (even if in this case we are dealing with a name, rather than with a qualified common noun) and so has to be regarded as a QResTop. As pointed out elsewhere (Martínez-Insua, 2017), taking this classification into account, the distinction between contentful and contentlight does not seem to be clear-cut, but rather scalar. It seems to be plausible to speak of a cline of contentfulness that goes from the contentlight GivTop towards the increasingly more contentful ResTop, QResTop and NewTop, as Figure 14.1 represents. Considering that the usual and unmarked pattern of information flow in text tends to be for the Theme to coincide with given information, and the Rheme to provide new information (Prince, 1981; Alonso-Belmonte and McCabe, 2003, among others), we may expect the frequency of contentlight GivTops to be higher than for more contentful ones. Importantly, according to Berry’s (2013, p. 264) findings in her study of discourse produced by school children, there seems to be a strong tendency

276 Ana Elina Martínez Insua

Figure 14.1 Contentfulness scale

for SubjTh of informal spoken English to realise contentlight options and for those of formal written English to realise contentful options. Bearing this in mind, Section 3 details our study of spoken English discourse produced by learners and native speakers of English.

3. The Study The study is based on the analysis of 600 SubjTh produced by native speakers of English and 600 SubjTh produced by Spanish learners of English in their spoken discourse. The data were drawn from the two mentioned corpora that Section 3.1 describes in detail, namely VICOLSE (Vigo Corpus of Learner Spoken English, Universidade de Vigo) and LOCNEC (Louvain Corpus of Native English Conversation, Université Catholique de Louvain—Centre for English Corpus Linguistics). As will be described in Section 3.2, in our attempt to test the working hypotheses, with Theme being our variable, the aspects considered in the study are (i) the (un)markedness of the SubjTh, (ii) their complexity, (iii) their weight and (iv) the reference they convey. 3.1. The Corpora 3.1.1. VICOLSE (Vigo Corpus of Learner Spoken English) VICOLSE is a spoken corpus of 100,000 words compiled by TizónCouto (2014), following the design of tasks, topics and transcription conventions used in LOCNEC (the native control corpus in this study), which are inherited in turn from the LINDSEI (Louvain International Database of Spoken English Interlanguage) project. The data in VICOLSE consist of the oral production of 86 university students (all of them native speakers of Spanish or Galician) who had to retell a story, describe a real-world scene, give their personal opinion and comment on a familiar/current topic. In the first task the students were asked to read the first part of a tale and then tell the remaining part of the story in

Relevance of the Textual Metafunction


their own words and with their own ideas. In the second task, they chose among several current topics (education, a famous TV show, a film, etc.) and told as much as they could, expressing their own personal opinions. Finally, students took part in a picture-based storytelling activity (TizónCouto, 2014, p. 167). 3.1.2. LOCNEC (Louvain Corpus of Native English Conversation) LOCNEC, containing 162,000 words, is the native counterpart of the Louvain International Database of Spoken English Interlanguage (LINDSEI), created at the Université Catholique de Louvain within the framework of De Cock (2003). LOCNEC contains the spoken production of 50 British students at Lancaster University. As explained by Tizón-Couto (2014, p. 185), the oral production contained in LOCNEC was elicited through the same tasks as those used in the interviews with non-native students in LINDSEI, and the texts have been transcribed following the same conventions. It represents the best option as a native control corpus, being comparable to VICOLSE to a large extent. The comparability of VICOLSE and LOCNEC is, therefore, high in that they share the same design of tasks and topics proposed for their compilation, and practically identical transcription conventions (inherited from the LINDSEI project). Apart from the differences in the native language of their participants (something that enables us to use one of the corpora as the control), their main difference is in their modality. Whereas the tasks in LOCNEC are carried out in the form of an interview, VICOLSE mainly consists of monologic recordings. Undeniably, this modality difference constitutes a relevant downside in the study, since theme choice might be affected by turn-taking. The findings in Section 3.2 will be accordingly treated with caution, and further research will be suggested as necessary to further test the hypotheses. 3.2. Findings and Discussion As pointed out above, once the 1,200 Themes were selected and introduced into a database, four different aspects were considered: (i) their (un)markedness, (ii) their being simple or multiple Themes, (iii) their content weight and (iv) the reference they contain. As hypothesised, the thematic areas of learner and native discourse were attested to differ from each other as regards the complexity of the Themes, their weight and their reference. Such findings help us confirm, at least partially, the initial hypothesis that the thematic areas of the two samples differ from each other. Nevertheless, our findings refute our second hypotheses, as the distribution of marked and unmarked Themes was found to be similar in both samples. Figures 14.2 to 14.5 show these findings graphically.

278 Ana Elina Martínez Insua 3.2.1. Markedness As regards the frequency of marked and unmarked Themes, no statistically significant differences were found between our learner sample and the native control one. The initial position in unmarked Themes is occupied by the subject itself, whereas in marked Themes other constituents precede the subject. Expectedly enough, unmarked Themes (e.g. [6] or [7]) are the predominant option in both samples, and no statistically significant differences were attested in their distribution across the two samples. (6) I’ve I’ve been . . I’ve only been abroad once and that was to Paris (LOCNEC) (7) my mum said to me to take you some food (VICOLSE) As mentioned in previous sections, the preliminary findings suggested higher frequencies of marked Themes in the spoken production of Spanish learners of English than in the native production. However, the analysis of these larger samples refutes this initial finding. As a matter of fact, the chi-square test confirms that the result is not significant at p < .05 (the p-value is 0.4503). Further research, using larger samples from both corpora, will be required before discarding the initial hypothesis definitely, but it is true that the differences do not seem to be remarkable. The initially higher frequency of marked Themes in the learners’ production may be due to the fact that more descriptive extracts (corresponding to picture-based tasks) were considered in the pilot sample. In such learners’ descriptive discourse, the presence of circumstantial information before SubjTh seemed to be remarkably abundant. In the samples under analysis, both learners and native speakers of English seem to produce messages with circumstances preceding subjects,

Figure 14.2 Theme markedness Chi-square statistic (χ2 (1) = 0.57, p-value = 0.4503). The distribution is not significant.

Relevance of the Textual Metafunction


such as (8) or (9), with relative frequency, but seem to opt similarly for unmarked Themes, such as those in (6) or (7), where the clause initial element is the subject, as their default option. (8) half an hour later Little Red Riding Hood reached the house (VICOLSE) (9) I think it’s three years here here (. . .) for a degree (. . .) only three years, in our country it‘s four years (LOCNEC) Except for (un)markedness, the rest of the aspects tested corroborate, however, our initial hypothesis that the thematic areas under analysis differ from each other, and statistically significant differences were observed between learner and native spoken English as regards the complexity, the content weight and the reference of the Themes speakers use. 3.2.2. Complexity When considering their complexity, the Themes were classified as either simple or multiple. Simple Themes are formed by one single constituent (in this case, the subject), whereas multiple Themes contain other elements in addition to the subject. The SubjTh used by the Spanish learners turned out to be simple with a significantly higher frequency than those in the native control sample. Even though examples such as (10) and (11) (where the Theme is multiple in that the subject is preceded by the textual markers and and but, respectively) are remarkably frequent in both samples (82.6% of the Themes in VICOLSE and 87.2% of those in LOCNEC), they are significantly more frequent in LOCNEC. By contrast, simple Themes like

Figure 14.3 Theme complexity Chi-square statistic (χ2 (1) = 4.4, p-value = 0.0359). The distribution is significant.

280 Ana Elina Martínez Insua my mum in (7) are significantly more frequent in VICOLSE (17.3% of the Themes in VICOLSE vs 12.8% in LOCNEC). (10) and that is quite an usual thing (LOCNEC) (11) she knocks at the door but nobody opens it (VICOLSE) As regards the textual and interpersonal elements in such multiple Themes, we consider it worth mentioning that cases of SubjTh containing interpersonal elements were only attested in VICOLSE. Even if their frequency is very low (only 11 cases), these interpersonal elements included vocatives and modal or comment adjuncts. It would be interesting to go into a deeper analysis of these cases, and also try to confirm whether this tendency persists in larger samples. This is, however, beyond the scope of the present contribution. 3.2.3. Contentfulness and Reference When focusing on the contentfulness (content weight) of the SubjTh under analysis, we found that the use of contentful Themes as clause initiators is remarkably more frequent in the learner sample. While contentlight SubjTh such as those in (12) are predominant in the native control sample, there is a tendency for the Spanish learners of English in the study to initiate their spoken messages with contentful SubjTh, that is, references to NewTops or ResTops. A NewTop introduces new material into the discourse, and even though a ResTop does not introduce new material, it does “give indications of a new attention to old material” (Berry, 2013, p. 258). Examples like (13), which are significantly more frequent among Spanish learners, illustrate this new attention granted to a piece of old material. (12) it’sGivTop a film which I . I had never seen before but everyone else seemed to have seen and so IGivTop was interested to see it . and it’sGivTop about erm . life in a[ei] . eh public school in America IGivTop think it’s in about the nineteen fifties (LOCNEC) (13) and they started crying inside the stomach of the *wolf (. .) because of the *noise eh (tch) Hugh the woodmanNewTop eh came to the house and saw the wolf and she ehm eh Little Red Riding HoodResTop started crying and screaming and HughResTop eh knew that the wolf have eaten them and he killed the wolf em (tch) opened the *opened the stomach of the wolf and he *caught Little Red Riding Hood and her grandmother (VICOLSE) Bearing in mind Berry`s (2013, p. 258) claim “that a Subject Theme is contentful if it refers to a NewTop or a ResTop [/QResTop], but contentlight if it does not so refer”, the findings represented in Figure 14.4

Relevance of the Textual Metafunction


Figure 14.4 Theme weight Chi-square statistic (χ2 (1) = 125.93, p-value < 0.0001). The distribution is highly significant.

below evince highly significant differences in the use of contentlight and contentful SubjTh by natives and learners. SubjTh containing light given references are remarkably less frequent in the learners’ spoken discourse. Being a matter of the relative newness or givenness of the entities that are introduced into the discourse, “references to NewTops would take more processing than references to GivTops. . . . GivTops can be assumed to be already in the hearer’s mind as they have just been mentioned in the text” (Berry, 2013, p. 260), while NewTops have not been so, and ResTops and QResTops have to be retrieved from longer ago. The statistical significance of the differences between the two samples is illustrated by the data in Figure 14.5. In addition to the overwhelming majority of contentlight SubjTh in native speech (ca. 85%), what Figure 14.5 suggests is that the different types of references are distributed more evenly in the learners’ SubjTh. In native discourse, more than 96% of the SubjTh refer to either GivTops (ca. 85%) or NewTops (ca. 12%), and only ca. 3% of them refer to ResTops or QResTops. By contrast, in the learners’ discourse, ResTops and QResTops represent ca. 20% of all the SubjTh, and the frequencies of NewTops (ca. 25%) and GivTops (ca. 55%) are remarkably lower than those in the native sample. It is our belief that these statistically highly significant differences may be behind some of the difficulties encountered by Spanish learners of English in achieving proficiency in their speech. Accepting that contentlight SubjTh imply less processing difficulty than contentful ones, demand less processing time and also grant “the conversationalists regular rests which enables them to recoup their energies ready for the more challenging references likely to occur after the verb” (Berry, 2013, p. 260), we may detect some extra degree of processing difficulty in the learners’ discourse.

282 Ana Elina Martínez Insua

Figure 14.5 Theme reference Chi-square statistic (χ2 (3) = 143.17, p-value < 0.0001). The distribution is highly significant.

The introduction of notions such as Theme, reference or contentfulness in the process of teaching and learning English as L2 might help to make learners aware of certain features of the English language that may be otherwise difficult to clarify. If learners of English know what SubjTh are, it may be easier for them to understand that SubjTh in native spoken English tend to be contentlight, and this helps decrease the processing difficulty. Similarly, it may be easier for teachers to explain that this is not necessarily so in written discourse, as when dealing with written English, “readers can take as long as they need to assimilate” the references of SubjTh (Berry, 2013, p. 260). We cannot but admit that our findings should necessarily be taken with caution, given the small size of the samples analysed. Nonetheless, they seem to support the advisability of calling learners’ attention to the Given-before-New information pattern that predominates in English. Figure 14.5 clearly illustrates the natives’ preference for given, contentlight SubjTh, as opposed to the much higher frequencies of new and resumed SubjTh attested in the learner’s discourse. Considering Berry’s (2013, p. 264) study of British school children’s written discourse, where “there does seem to be a strong tendency for the Subject Themes of informal spoken English to realize contentlight options and for the Subject Themes of formal written English to realize contentful options”, the spoken discourse of the learners in VICOLSE seems to

Relevance of the Textual Metafunction


be less genuinely ‘speechy’ than the spoken discourse of the natives. As a matter of fact, while the discourse in LOCNEC coincides with Berry’s findings on spoken English, the learners’ discourse in VICOLSE appears to be closer to the written-like end of the cline suggested. Should we try to establish a connection with Berry’s cline of content weight (2013, p. 258) and our contentfulness scale (Figure 14.1, repeated here for convenience) at this point, we could say that while the SubjTh from the native sample clearly stick to the contentlight/GivTop end of the scale, those from the learners’ sample linger around the centre of the scale. Thus, inspired by Berry (2013, p. 265), we may say that if the samples discussed here turn out to be representative, it seems that Spanish learners of English, when practising their spoken discourse, need to learn not to suppress “what seems to be the most natural way of communicating— foregrounding themselves” and their addressees in their oral production. Let us insist that some words of caution are mandatory here. We should not forget that even if VICOLSE and LOCNEC display a considerable degree of comparability, the samples contained in the two corpora are not identical in their modality. While LOCNEC mostly contains dialogic samples, VICOLSE does not in all the cases, and this implies that our samples are not completely free from bias. If one assumes a cline from informal spoken English to formal written English, one might expect dialogue to be slightly closer to the informal spoken English end (Berry, pers. comm.). Consequently, one might expect our VICOLSE sample to be a little further from the informal spoken English end than our LOCNEC sample.

4. Concluding Remarks The present study is an attempt to contribute to the debate on the appliability of Systemic Functional Linguistics to help solve problems in the world. More specifically, the aim of the study is to evince the usefulness of the framework and its approach to language as aids for Spanish people involved in the process of learning and teaching English as a foreign language. The findings obtained in our testing of the thematic area of the spoken messages produced by English natives and learners of the language hint at the potential benefits that introducing the notion of Theme in the teaching-learning process may bring about. Highly significant differences were attested in the complexity and content weight of the SubjTh produced by natives and learners. Similarly, the types of references they include in their SubjTh were found to be significantly different. Without forgetting that the corpora used are comparable but not completely identical, we may conclude that Spanish learners of English might benefit from knowing about the thematic structure of English. To return here our initial attempt to see what Spanish learners need to be taught, or where more teaching would be valuable, it seems

284 Ana Elina Martínez Insua to be the case that becoming familiar with notions such as Theme, contentfulness, given-before-new information patterning, (end-)weight and reference might give Spanish learners of English stronger tools to become proficient speakers of English. This might help them make their spoken discourse approach that of natives. Teachers, in turn, might find an ally in the notion of (Subject) Theme when trying to explain to Spanish learners of English how to construct fully effective messages in the target language. Once the learner is familiar with the notion of Theme, and knows its main features and tendencies in the discourse of native English speakers, explanations and corrections may become simpler, clearer and more effective. No doubt, further research is necessary to confirm or refute these tentative conclusions. We aim to enlarge the samples analysed and refine the classification of the SubjTh by elaborating a fine-grained contentfulness scale, taking into consideration Prince’s (1981, p. 237) Assumed Familiarity Scale.9 Prince’s classification of discourse entities as New, Inferrable or Evoked, with their subsequent sub-types, may well help us refine our contentfulness scale, and render it more accurate and able to account better for the SubjTh in our samples.

Notes 1. We are grateful to the following institutions for generous financial support: the Spanish Ministry of Economy and Competitiveness and the European Regional Development Fund (grants no. FFI2013-44065-P and FFI2016-77018-P), and the Autonomous Government of Galicia (grant no. GPC2014/060). On a personal level, we are gratefully indebted to Margaret Berry for extensive discussion, insightful suggestions and most generous help on many aspects of this study. We would also like to thank the editors and reviewers of an earlier version of the chapter for their detailed and helpful comments. Any errors remaining in the chapter are the sole responsibility of the author. 2. For the controversy on the notion of Theme in Systemic Functional Grammar, see also Berry (1995) or Downing (1991), among others. Other possible views on the extent of the Theme may be found in the literature (for example Berry, 1996, pp. 29–31), but their detailed examination would go beyond the space constraints of this contribution. 3. See also McCabe and Alonso-Belmonte (2000) for a detailed review of the literature on different approaches to Theme (Theme as a starting point, Theme as an orienting element, Theme as idea, cognitive approaches to Theme, etc.). 4. The label ‘Subject Theme’ is a term that Berry (2013, p. 248) takes from Fawcett (2008, p. 182) so as to refer to what the former had previously labelled as ‘Basic Theme’. Berry’s label ‘Subject Theme’ will be adopted in this study. 5. Initially, Berry (1995) used Brown and Yule’s terms; later she adopted Dik’s terms as more transparent (pers. comm.). 6. Berry (pers. comm.) called our attention to the possibility to include this fourth category. Brown and Yule (1983, p. 174) refer to expressions of the form the + property + noun and the fact that they are “used almost exclusively in identifying displaced entities [that is, those labelled as resumed here]”.

Relevance of the Textual Metafunction


7. The asterisks in VICOLSE mark mispronounced words. The dots in between brackets signal the speaker’s pauses. 8. In LOCNEC ... and ... delimit the speakers’ turns. Whilst the letter “A” enclosed between angle brackets always signifies the interviewer’s turn, the letter “B” between angle brackets indicates the interviewee’s (learner’s) turn. The end of each turn is indicated by either or (https://www.uclouvain. be/en-307849.html) 9. We would like to thank Margaret Berry for suggesting the elaboration of this fine-grained contentfulness scale.

References Alonso-Belmonte, I. and McCabe-Hidalgo, A. (1998) Theme-Rheme patterns in L2 writing. Didáctica, 10, pp. 13–31. Alonso-Belmonte, I. and McCabe-Hidalgo, A. (2003) Improving text flow in ESL learner compositions. The Internet TESL Journal, IX(2). Available at: http:// (Accessed on 31 July 2016). Berry, M. (1995) Thematic options and success in writing. In M. Ghadessy (Ed.), Thematic Development in English Texts. London and New York: Pinter, pp. 55–84. Berry, M. (1996) What is theme? A(nother) personal view. In M. Berry, C. Butler, R. Fawcett and G. Huang (Eds.), Meaning and Form: Systemic Functional Interpretations. Meaning and Choice in Language Studies for Michael Halliday, Advances in Discourse Processes, vol. LVII, Norwood, NJ: Ablex Publishing Corporation, pp. 1–64. Berry, M. (2013) Contentful and contentlight subject themes in informal spoken English and formal written English. In G. O’Grady, T. Barlett and L. Fontaine (Eds.), Choice in Language: Applications in Text Analysis. Sheffield: Equinox, pp. 243–268. Berry, M., Thompson, G. and Hillier, H. (2014) Theme and variations. In M. A. Gómez-González, F. Ruiz de Mendoza and F. Gonzálvez-García (Eds.), Theory and Practice in Functional-Cognitive Space. Amsterdam: Benjamins, pp. 107–126. Brown, G. and Yule, G. (1983) Discourse Analysis. Cambridge: Cambridge University Press. Chocano, G., Jiménez, R., Lozano, C., Mendikoetxea, A., Murcia, S., O’Donnell, M., Rollinson, P. and Teomiro, I. (2007) An exploration into word-order in learner corpora: The WOSLAC project. In P. Rayson Davies, S. Hunston and P. Danielsson (Eds.), Proceedings of the Corpus Linguistics Conference 2007. Birmingham: University of Birmingham. Dagneaux, E. (1995) Expressions of Epistemic Modality in Native and NonNative Essay Writing. MA dissertation, Université catholique de Louvain. De Cock, S. (2003) Recurrent Sequences of Words in Native Speaker and Advanced Learner Spoken and Written English. Unpublished PhD thesis, Université Catholique de Louvain. Dik, S. C. (1997) The Theory of Functional Grammar. Part I: The Structure of the Clause, 2nd ed., edited by K. Hengeveld. Berlin and New York: Mouton de Gruyter. Downing, A. (1991) An alternative approach to theme: A systemic functional perspective. Word, 42(2), pp. 119–143.

286 Ana Elina Martínez Insua Fawcett, R. P. (2008) Invitation to Systemic Functional Linguistics Through the Cardiff Grammar: An Extension and Simplification of Halliday’s Systemic Functional Grammar, 3rd ed. London: Equinox. Forey, G. and Thompson, G. (2008) Introduction. In G. Forey and G. Thompson (Eds.), Text Type and Texture. London: Equinox, pp. 1–7. Fries, P. H. (1981) On the status of theme in English: Arguments from discourse. Forum Linguisticum, 6, pp. 1–38. Fries, P. H. (1992) Information flow in written advertising. In J. E. Alatis (Ed.), Georgetown University Round Table on Languages and Linguistics: Language, Communication and Social Meaning. Washington: Georgetown University Press, pp. 336–352. Fries, P. H. (1994) On theme, rheme and discourse goals. In M. Coulthard (Ed.), Advances in Written Text Analysis. London: Routledge, pp. 229–249. Fries, P. H. (1995a) A personal view of theme. In M. Ghadessy (Ed.), Thematic Development in English Texts. London and New York: Pinter, pp. 1–19. Fries, P. H. (1995b) Themes, methods of development, and texts. In R. Hasan and P. H. Fries (Eds.), On Subject and Theme. Amsterdam and Philadelphia: John Benjamins, pp. 317–359. Granger, S. (1997) On identifying the syntactic and discourse features of participle clauses in academic English: Native and non-native writers compared. In J. Aarts, I. de Mönnink and H. Wekker (Eds.), Studies in English Language and Teaching. Amsterdam and Atlanta: Rodopi, pp. 191–202. Halliday, M. A. K. (1985) Introduction to Functional Grammar, 1st ed. London: Edward Arnold. Kaplan, R. B. (1995) Contrastive rhetoric. In T. Miller (Ed.), Functional Approaches to Written Text: Classroom Applications. Paris: The Journal of TESOLFrance in Association With U.S. Information Service, pp. 21–39. LINDSEI = Louvain International Database of Spoken English Language. (2010) Edited by G. Gilquin, S. De Cock and S. Granger. Louvain-la-Neuve: UCL Presses Universitaires de Louvain. LOCNEC = Louvain Corpus of Native English Conversation. S. De Cock (compiler). See: Martin, J. R. (1992) English Text: System and Structure. Amsterdam and Philadelphia: John Benjamins. Martin, J. R. (1995) More than what the message is about: English theme. In M. Ghadessy (Ed.), Thematic Development in English Texts. London and New York: Pinter, pp. 223–258. Martínez-Insua, A. E. (2017) On the contentfulness of theme in English historical medical texts. In S. Neumann, R. Wegener, J. Fest, P. Nietmietz and N. Hützen (Eds.), Challenging Boundaries in Linguistics. Bern: Peter Lang, pp. 111–131. Matthiessen, C. (1995) THEME as an enabling resource in ideational ‘knowledge’ construction. In M. Ghadessy (Ed.), Thematic Development in English Texts. London and New York: Pinter, pp. 20–54. McCabe, A. and Alonso-Belmonte, I. (2000) Theme, transitivity and cognitive representation in Spanish and English written texts. Revista Canaria de Estudios Ingleses, 40, pp. 77–94. O’Donnell, M. (2013) From learner corpora to curriculum design: An empirical approach to staging the teaching of grammatical concepts. Procedia: Social and Behavioral Sciences, 95, pp. 571–580.

Relevance of the Textual Metafunction


Prince, E. F. (1981) Toward a taxonomy of given-new information. In P. Cole (Ed.), Radical Pragmatics. New York: Academic Press, pp. 223–256. Tizón-Couto, B. (2014) Clausal Complements in Native and Learner Spoken English: A Corpus-based Study With LINDSEI and VICOLSE. Bern: Peter Lang. Whittaker, R. (1995) Theme, processes and the realization of meanings in academic articles. In M. Ghadessy (Ed.), Thematic Development in English Texts. London and New York: Pinter, pp. 105–128.

15 Nominal Groups in Arabic and English Experiential Differences and Effect on Tunisian EFL Learners’ Translations Dorra Moalla 1. Introduction The translation of nominal groups (NGs)1 from English to Arabic can be worth studying for grammatical and pedagogical reasons. From a grammatical perspective, in both languages, the NG is an important carrier of meaning that can serve the role of subject or complement in clause modal structure and different participant roles in experiential structure (Halliday and Matthiessen, 2004). Fontaine (2017, p. 267) stresses the centrality of this unit in the clause, arguing that the NG is “the only unit that can simultaneously express the main functional elements for all three main metafunctions of the clause”. Given the importance of this unit in discourse, the study participants, who are Tunisian learners of English as a Foreign Language (EFL) enrolled in second year to obtain a bachelor in English, are frequently required to translate this structure. In this learning situation, translation is an obligatory subject in the curriculum for the three years of undergraduate education. The Tunisian context is a multilingual context in which Arabic is the native language while English is a foreign language growing in importance both in the academic context and in the professional one. In this context, the students are quite aware of the need to promote their translation skills to increase their chances to get better jobs in the future. Considering the semantic importance of NGs in discourse and the concentration of translation in the curriculum, there is a need to identify the sources of problems that interfere with their effective translation. Studies on Arabic NGs and issues in their translation are, however, very scarce. Bardi (2008), for instance, focused on the functional role of NGs in Arabic across the three metafunctions but did not offer an experiential division of this structure based on Systemic Functional Linguistics (SFL). However, such a division is needed to enable translation researchers to identify the sources of translation problems. Applied linguists have also shown genuine interest in translation through developing techniques and materials in the curriculum, listing error taxonomies and identifying their

Nominal Groups in Arabic and English


causes and through dealing with strategies as means to deal with translation problems (Hatim, 2013; Hatim and Munday, 2004). They remain, however, quite focused on ways to investigate language in translation rather than language itself in translation. This study argues for a close approximation between SFL and applied linguistics to provide new perspectives for exploring issues in translating NGs in the context of EFL. SFL provides a framework that divides the NG into functional components. This framework can be used by applied linguists as a parameter to compare NGs and to identify the areas of equivalences and shifts across the languages. Previous studies relied on speculation about potential areas of difficulty, which may be attributed to the degree of translation shift in the systems of languages (Catford, 1965; Connor, 2002; Nida, 1964). In the same line, this chapter postulates that the translation of NGs may be potentially challenging for Tunisian EFL learners because of the divergence in Deictic elements, the potential for expansion and position of experiential categories across the languages. By administering authentic translation tasks to Tunisian EFL learners and comparing their performance with the degree of shift and equivalence across the languages, this chapter will probe the effect of these variables on the translation of NGs instead of relying on speculation about their potential effects. The next section describes the theoretical framework of this study, focusing on the NG within the dimensions of language organization, describing the experiential classification of the NG in SFL and Arabic grammar and defining translation shifts and equivalences. Section 3 introduces the research problem and hypotheses. Sections 4 and 5 describe the research methodology and results respectively. The chapter ends with a summary of the main results and suggestions for future research.

2. Key Concepts and Theoretical Framework The purpose of this section is to describe the levels of language organization in SFL and the grammar of constructs in Arabic in an attempt to situate the NG within the hierarchy of language. It also aims to describe the experiential classification of this structure in both languages, drawing on SFL and Arabic grammar. This section ends with a review of relevant research on translation shifts and equivalences. 2.1. Dimensions of Language Organization Munday (2008, p. 5) defines translation as a process involving “the translator changing an original written text in the original verbal language (the source language) into a written text in a different verbal language (the target language)”. In order to develop a clear picture of the mechanisms of language transfer in translation, I consider it important to divide language


Dorra Moalla

into its hierarchical units. This identification of dimensions of language organization aims to situate the NG within language hierarchy. According to SFL, language is organized according to local and global paradigms that often overlap. At the level of global organization, language in context is ordered into a number of strata including phonetics, phonology, lexicogrammar, semantics and context (Halliday and Matthiessen, 2004, p. 25). Stratification is a hierarchical system in which the different constituents interact. Human beings construe experience and enact social relationships, which are within the stratum of semantics. These processes are materialized through wording, which is the stratum of lexicogrammar (Halliday, 2001). The local dimension of language organization includes rank, which organizes the resources of the lexicogrammar stratum into a hierarchy of ranks from clause to group and from group to word. Within the clause and group, resources are organized into system networks. In this chapter, the focus is laid on the NG unit in both English and Arabic. In English, the NG is a unit at the group rank, situated between the clause and the word. In Arabic, there is a lack of functional description of NGs. Though there is a controversy in Arabic grammar relative to the concepts used to distinguish the different ranks (Bardi, 2008), Tunisian university textbooks situate the NG as a rank unit between Aljomla-‫ﺍﻟﺠﻤﻠﺔ‬ ‘the sentence’ and Alkalima-‫‘ ﺍﻟﻜﻠﻤﺔ‬the word’. This rank unit is labelled Almorakab-‫ﺍﻟﻤﺮﻛﺐ‬, which can be translated as ‘compound’ or ‘construct’. In this chapter, I have opted for the construct concept because it can have broader structural realizations (noun-noun combination, adjectivenoun . . . etc.) as opposed to the compound, which is more closely related to noun- noun combinations. In my attempt to explore the issue of NG translation from English to Arabic, I will try to demonstrate throughout this chapter that though the NG is operating at the group rank within the stratum of lexicogrammar, semantic factors are relevant to explain NG translation issues and choices. This stands in harmony with SFL as a meaningdriven theory. After situating the NG within language dimensions, I will focus on the experiential description of this unit in SFL and its description in Arabic grammar. In SFL, the NG is a meaning-loaded grammatical unit that serves to characterize and represent the Thing element (Halliday and Matthiessen, 2004, p. 312). Fontaine stressed the meaning-driven nature of this structure, arguing that it is “the main grammatical resource speakers have for referring to a referent” (2013, p. 46). This referring function is also stressed in the NG in Arabic. A noun construct is considered as a grammatical and semantic entity used “to refer to the existence of things in their generality or specificity” (Kabani, Bannour, Missaoui and Chaouech, 2000, p. 101).

Nominal Groups in Arabic and English


In addition to the semantic description of this structure, Systemic Functional Grammar (SFG) provides a structural description. There is agreement among researchers that the NG is a sequence of constituents including a Head/Thing plus Modifiers (Fontaine, 2013; Halliday and Matthiessen, 2004). Both SFG and Arabic grammar argue for a semantic relationship between Thing and Modifiers, but while SFG argues that Modifiers are experientially classified and are clustered around the Thing element, Arabic grammar analyses the NG as a hierarchical system that can include embedded NGs. The nature of each NG is defined according to the functional and structural relationships between head and modifiers. Example (1) from Jobrane (1918, p. 487) is used to illustrate the similarities and differences between SFG and the noun construct framework in their analysis of the NG in italics. (1)

‫ﻭ ﺗﻤﻸ ﺃﺣﻼﻣﻲ ﺑﺄﺷﺒﺎﺡ ﺃﺭﺽ ﻗﺼﻴﺔ ﻣﺎ ﺭﺃﺗﻬﺎ ﻋﻴﻨﻲ‬ ‘and my dreams were filled with the ghosts of a remote land my eyes have never seen’

If we consider the NG represented and glossed as example (2),2 it is worth mentioning that the main similarity between SFG and the noun construct framework is the functional, meaning-driven classification of constituents. For instance, ‘remote’ is an Epithet that indicates some quality of the Thing (Halliday and Matthiessen, 2004, p. 318). The same element quasiyatin-‫‘ ﻗﺼﻴﺔ‬remote’ is a qualification na‘at-‫ ﻧﻌﺖ‬for the head noun ārdhin-‫‘ ﺃﺭﺽ‬land’. (2)

āchbah-i ārdhin quasiyatin me raātha ‘ayni āchbaḥ-i ārdh-in quasiy-at-in me raāt-ha ‘ayn-i ghosts- landremotenever see.past.3rd. eye-poss case case fem-case sing-it det ‘the ghosts of a remote land my eyes have never seen’

The main difference between the two grammars is their analysis of the way Modifiers relate to the Thing element. In SFG, the Modifiers are clustered around the Thing. For instance, in the NG ‘a remote land my eyes have never seen’, the experiential categories of Unspecific Deictic, Epithet and Qualifier, respectively realized by the elements ‘a’, ‘remote’ and ‘my eyes have never seen’, occur around the Thing element ‘land’ (see Table 15.1). The parallel NG in Arabic, the relative clause ‘my eyes have never seen’, modifies the qualifying construct ārdhin quasiyatin-‫‘ ﺃﺭﺽ ﻗﺼﻴﺔ‬a remote land’ (see Table 15.2). Another main difference between the two grammars is that SFG classifies the NG into experiential categories, as I have already shown. The noun construct framework3 introduces an analysis of the NG as a hierarchical system including a higher level NG āchba hi ardhin quasiyatin


Dorra Moalla

Table 15.1 Experiential analysis of the NG in English in example (1) using SFG The


of a remote land my eyes have never seen



Qualifier a



my eyes have never seen





Table 15.2 Analysis of the NG in Arabic in example (1) using the grammar of constructs ‫ﻣﺎ ﺭﺍﺗﻬﺎ ﻋﻴﻨﻲ‬ me raātha ‘ayni my eyes have never seen










adjective/ noun/ qualification- qualified‫ﻧﻌﺖ‬/‫ﻣﻨﻌﻮﺕ ﺻﻔﺔ ﻣﺸﺒﻬﺔ‬/‫ﺍﺳﻢ‬ relative clause/qualification‫ﻧﻌﺖ‬/‫ﻣﺮﻛﺐ ﻣﻮﺻﻮﻟﻲ‬

qualifying construct/ qualified-‫ﻣﻨﻌﻮﺕ‬/‫ﻣﺮﻛﺐ ﻧﻌﺘﻲ‬

qualifying construct/annexed to- ‫ ﻣﻀﺎﻑ ﺇﻟﻴﻪ‬/‫ﻣﺮﻛﺐ ﻧﻌﺘﻲ‬

noun/annexed‫ﻣﻀﺎﻑ‬/ ‫ﺍﺳﻢ‬

annexation construct-‫ﻣﺮﻛﺐ ﺇﺿﺎﻓﻲ‬

me raātha ayni-‫( ﺃﺷﺒﺎﺡ ﺃﺭﺽ ﻗﺼﻴﺔ ﻣﺎ ﺭﺃﺗﻬﺎ ﻋﻴﻨﻲ‬see Table 15.2). Embedded in this NG, there is a lower level one ārdhin quasiyatin me raātha ayni‫ﺃﺭﺽ ﻗﺼﻴﺔ ﻣﺎ ﺭﺃﺗﻬﺎ ﻋﻴﻨﻲ‬. In noun construct grammar, the nature of the NG shifts as the functional and structural relationships between head and modifier. The higher level construct is an annexation construct, in which ārdhin quasiyatin me raātha ayni-‫ ﺃﺭﺽ ﻗﺼﻴﺔ ﻣﺎ ﺭﺃﺗﻬﺎ ﻋﻴﻨﻲ‬is annexed to the head noun ashbahi-‫‘ ﺃﺷﺒﺎﺡ‬ghosts’ (see Table 15.1). This annexation has structural implications manifested in the genitive case of the modifier. The embedded NG is a qualifying construct, which includes two qualifications: the adjective quasitatin-‫‘ ﻗﺼﻴﺔ‬remote’ and the relative clause me raātha ayni-‫‘ ﻣﺎ ﺭﺃﺗﻬﺎ ﻋﻴﻨﻲ‬my eyes have never seen’ postmodifying the head noun ārdhin-‫‘ ﺃﺭﺽ‬land’. This qualifying relationship has structural implications of agreement (this will be further explained in the following section). 2.2. Experiential Classification of NGs in English As developed by Halliday and Matthiessen (2004, 2014), in experiential terms, the Thing, which is “the semantic core of the nominal group” (Halliday and Matthiessen, 2004, p. 325), is preceded and followed by elements that serve the two basic functions of “specifying a class

Nominal Groups in Arabic and English


of things” and indicating “some category of membership within this class” (Halliday and Matthiessen, 2004, p. 312). These elements are realized by the functional elements of Deictic, Numerative, Epithet and Classifier. •

• • •

Deictic: The Deictic element indicates “whether or not some specific subset of the Thing is intended” (Halliday and Matthiessen, 2014, p. 365). Deictic is realized by the system of determination. Numerative: It indicates a quality or order of the Thing. It is realized by a numeral. Epithet: It indicates some quality of the Thing or the speaker’s attitude towards it. It is typically expressed by an adjective. Classifier: It refers to a particular sub-class of the Thing.

The element that occurs after the Thing is the Qualifier. It is a rankshifted element with a rank equivalent to the NG or higher than it, i.e. a NG, prepositional phrase (PP) or clause. 2.3. Description and Pattern of Modification of NGs in Arabic In view of the lack of a functional framework for analysing NGs in Arabic, the noun construct framework al-morakabet al-ismia-‫ﺍﻟﻤﺮﻛﺒﺎﺕ ﺍﻹﺳﻤﻴﺔ‬ was adopted to describe NGs in Arabic. I consider it the most adequate framework for NG comparison between Arabic and English. This view is shared by Bardi (2008, p. 2), who stressed the compatibility between SFG and Arabic grammar in general, suggesting that “the Arab tradition is closer to Halliday’s approach than to other western structural approaches”. He argues that the functional meaning-driven orientation of the Arabic tradition is supported by two major facts. The first is that this grammar evolved with the linguistic sciences of rhetoric and semantics. The second fact is that Arabic grammar was text-driven. This grammar, along with these sciences, was originally trying to study the Quran and ancient Arab poetry to show their distinctiveness and determine why these texts were more highly appreciated by readers and listeners. In this endeavour, grammatical structures and meaning were viewed as inextricably related. This view of language as a meaning-conveyor is also echoed in the description of NG in the grammar of constructs. It provides a functional description of this structure through specifying the functional relationship between head and modifiers (for instance, a qualifying construct is composed of a head that is being qualified and a postmodifier that qualifies this head). This compatibility between SFG and the grammar of constructs is highlighted numerically in the corpus of the study (for instance, 78% of NGs including Epithets are translated into qualifying constructs a definition of qualifying constructs with examples will be provided in this section).


Dorra Moalla

In the grammar of constructs, the noun/thing can be followed by one or more than one grammatical unit that specifies it. They together form a noun construct morakab Ismi-‫ﻣﺮﻛﺐ ﺇﺳﻤﻲ‬. The noun constructs are divided into several categories depending on their structural realizations, the features of the noun head and postmodifier, and the functional relationship between thing and modifier. In the context of the present study, focus is laid on four types of constructs: The annexation construct al-morakab al-iḍaafi-‫ﺍﻟﻤﺮﻛﺐ ﺍﻹﺿﺎﻓﻲ‬: the concept of annexation construct was adopted from Ryding (2005), as opposed to other translations like the ‘possessive compound’, because the latter reduces the semantic values of this NG. As defined by Kabani et al. (2000, pp. 48–58), in this construct, the head noun is annexed to another noun to specify it, as in example (3). It is composed of the head noun labelled the annexed almoḍaaf-‫ﺍﻟﻤﻀﺎﻑ‬, which is ghorfatu-‫‘ ﻏﺮﻓﺔ‬room’, and the element following it, labelled the annexed to almoḍaaf ilayhi-‫ﺍﻟﻤﻀﺎﻑ ﺇﻟﻴﻪ‬, which is a ṭaami-‫‘ ﺍﻟﻄﻌﺎﻡ‬dining’. The head is an unspecified one-word unit, which is free of number, gender and nunation.4 The postmodifier can be a specified noun and marked by the genitive case almajrūr-‫ﺍﻟﻤﺠﺮﻭﺭ‬, as in example (3). (3)

‫ﻏﺮﻓﺔ ﺍﻟﻄﻌﺎﻡ‬ ghorfatu a- ṭa‘am-i room spec-dining-gen case ‘the dining room’

The qualifying construct almorakab ana a͑ ti-‫ﺍﻟﻤﺮﻛﺐ ﺍﻟﻨﻌﺘﻲ‬: The head noun can be followed by a qualification to qualify it (Kabani et al., 2000, p. 101). This construct is composed of the head noun, the qualified alman o ͑ ut-‫ﺍﻟﻤﻨﻌﻮﺕ‬ and a postmodifier, the qualification ana a͑ tu-‫ﺍﻟﻨﻌﺖ‬. The qualification can take different grammatical forms like a single word (an adjective), a PP morakab bil Jarr-‫ﻣﺮﻛﺐ ﺑﺎﻟﺠﺮ‬, a relative clause morakab mawṣouli-‫ﻣﺮﻛﺐ ﻣﻮﺻﻮﻟﻲ‬ or a clause. The qualification agrees with the qualified in gender, number, specification and case. This is illustrated in examples (4), (5) and (6). In example (5), the qualified alfaṭoura-‫‘ ﺍﻟﻔﻄﻮﺭ‬the meal’ is a singular, specified masculine noun. This noun is in the accusative case almanṣub‫ ﺍﻟﻤﻨﺼﻮﺏ‬since the construct is the object of the sentence. The qualification alladhidha-‫‘ ﺍﻟﻠﺬﻳﺬ‬the delicious’ is an adjective that agrees with the qualified in number, gender, specification and case. (4)

“֞‫ﺃﻛﻠﺖ ﻛﻌﻜﺔ֞ ﻟﺬﻳﺬﺓ‬ ākaltu ka‘ka-tan ladhidha-tan eat. 1st sing. cake-fem.sing. delicious- fem.sing. past unspec.nun unspec.nun ‘I ate a delicious cake’

Nominal Groups in Arabic and English (5)

‫ﺃﻛﻠﺖ ﺍﻟﻔﻄﻮ َﺭ ﺍﻟﻠﺬﻳ َﺬ‬ ākaltu al-faṭour- a eat. 3rd sing. spec-meal-masc.sing. past case ‘I ate the delicious meal’



al-ladhidh-a spec-delicious-masc.

‫ﺃﻛﻠﺖ ﺍﻟﻜﻌﻜﺔ ﺍﻟﺘﻲ ﻁﺒﺨﺘﻬﺎ ﺃﻣﻲ‬ ākaltu al-ka‘kata allati eat. 3rd spec-cake rel pron sing.past ‘I ate the cake that my mother baked’

ṭabakhat-ha bake. 3rd sing.past -it

om-i motherposs det

The distinction construct almorakab atamyiizi-‫ﺍﻟﻤﺮﻛﺐ ﺍﻟﺘﻤﻴﻴﺰﻱ‬: This construct is divided into two categories. In the first category, the distinguished almomayaz-‫ﺍﻟﻤﻤﻴﺰ‬, which is the head, is a one-word element, and the distinction atamyiiz-‫ ﺍﻟﺘﻤﻴﻴﺰ‬is an unspecified noun or a PP starting with the preposition min-‫‘ ﻣﻦ‬of’, used to avoid ambiguity about this noun (Zied, 2009). In example (7) from the Quran in Yusuf Surah (Quran: 12–5, Islam International Publications Limited), the numeral āḥada‘achara-‫ﺃﺣﺪ ﻋﺸﺮ‬ ‘eleven’ is distinguished as kawkaban-‫‘ ﻛﻮﻛﺒﺎ‬stars’. In example (8), the unit of measurement mitran-‫‘ ﻣﺘﺮﺍ‬one meter’ is distinguished by the PP mina aṣoufi-‫‘ ﻣﻦ ﺍﻟﺼﻮﻑ‬of wool’. (7)

‫ﺇﻧﻲ ﺭﺃﻳﺖ ﺍﺣﺪ ﻋﺸﺮ ﻛﻮﻛﺒﺎ‬ Inni ra āytu āḥada‘achara kawkab-an see.1st sing.past eleven star-unspec ‘I saw (in my dreams) eleven stars’


‫ﺍﺑﺘﻌﺖ ﻣﺘﺮﺍ ﻣﻦ ﺍﻟﺼﻮﻑ‬ ibta‘tu mitran mina a-ṣouf-i buy.1st sing.past one meter prep spec-wool- gen case ‘I bought one meter of wool’

In the second category of distinction constructs, the distinction specifies the whole clause preceding it. In example (9) from the Quran in Maryam Surah (Quran: 19–5, Islam International Publications Limited), the distinction chayban-‫‘ ﺷﻴﺒﺎ‬hoariness’ specifies the reason why the head glistens In example (10), the distinction does not relate to a single word but clarifies the relationship between the elements of the whole previous clause. The distinction kholoqan-‫‘ ﺧﻠﻘﺎ‬behaviour’ is used to specify the level at which I found him the best person. (9)

ً ‫ﻭﺍﺷﺘﻌﻞ ﺍﻟﺮﺃﺱ ﺷﻴﺒﺎ‬ wa ichta‘ala a-raāsu and glisten. 3rd. sing.past spec-head ‘and the head glistens with hoariness’

chayb-an hoariness-unspec


Dorra Moalla


‫ﻭﺟﺪﺗﻪ ﺃﺣﺴﻦ ﺍﻟﻨﺎﺱ ﺧﻠﻘﺎ‬ wajadtu-hu āḥsana a-nasi kholoq-an find.1st sing. best spec- people behaviour-unspec past- him ‘I found him the best behaving person’

The pseudo-verbal construct almorakah chibh alisnadi-‫ﺍﻟﻤﺮﻛﺐ‬ ‫ﺷﺒﻪ ﺍﻹﺳﻨﺎﺩﻱ‬: It is composed of a derived structure like an adjective, an active participle ism fa ͑el-‫ ﺇﺳﻢ ﻓﺎﻋﻞ‬or a passive participle ism maf ʿoul-‫ﺇﺳﻢ ﻣﻔﻌﻮﻝ‬, superlative . . . etc. These structures, which constitute the head of this construct, are viewed as nominal because they are nominal structures in Arabic morphology as̠ arf-‫ﺍﻟﺼﺮﻑ‬. These structures can be paraphrased into a verb and are followed by a structure related to it, as illustrated in Table 15.3 and example (11). (11)

‫ﺭﺃﺕ ﺍﻣﺮﺃﺓ ﻻﺑﺴﺔ ﻟﺒﺎﺳﺎ ﺗﻘﻠﻴﺪﻳﺎ‬ raāt imraāt-an labisa-t-an libas-an see. 3rd. womanwearing-fem- clothessing.past unspec unspec unspec ‘She saw a woman wearing traditional clothes’

taqlidi-an traditionalunspec

Table 15.3 Example of a pseudo-verbal construct in Arabic ‫ﻟﺒﺎﺳﺎ ﺗﻘﻠﻴﺪﻳﺎ‬ traditional clothes A qualifying construct/direct Object**- ‫ﻣﻔﻌﻮﻝ ﺑﻪ‬/‫ﻣﺮﻛﺐ ﻧﻌﺘﻲ‬

‫ﻻﺑﺴﺔ‬ wearing

‫ﺍﻣﺮﺃﺓ‬ a woman

‫ﺭﺃﺕ‬ she saw

active participle* ‫ﺍﺳﻢ ﻓﺎﻋﻞ‬-

pseudo-verbal construct/qualification- ‫ﻧﻌﺖ‬/‫ﻣﺮﻛﺐ ﺷﺒﻪ‬ ‫ﺍﺳﻨﺎﺩﻱ‬


qualifying construct/direct object- ‫ﻣﻔﻌﻮﻝ ﺑﻪ‬/‫ﻣﺮﻛﺐ ﻧﻌﺘﻲ‬


*The active participle Ism al-Fa ͑el is a derived form that can be paraphrased into the verb ‘‫‘ ’ﺗﻠﺒﺲ‬to wear’. **This structure is related to the derived form and plays the role of its Direct Object because the construct can be paraphrased as ‘‫( ’ﻛﺎﻧﺖ ﺍﻟﻤﺮﺃﺓ ﺗﻠﺒﺲ ﻟﺒﺎﺳﺎ ﺗﻘﻠﻴﺪﻳﺎ‬the woman was wearing traditional clothes).

Having now described the experiential classification of NGs in both languages, I will concentrate on translation shifts and equivalences because this chapter focuses on the effect of these variables on the translation of the different experiential categories.

Nominal Groups in Arabic and English


2.4. Translation Equivalences and Shifts Translators and translation researchers agree that equivalence is a central and organizing principle for evaluating translation. This concept is, however, not consistently used in translation studies. Early studies associated equivalence with translation quality. They attributed differential values to equivalence at the different strata and ranks. Nida (1964) was influenced by semantic and pragmatic theories and valued dynamic equivalence that prioritizes the naturalness of the message through adaptation to the linguistic and cultural expectations of the receptor. Catford (1965) focuses on equivalence at the lexicogrammar stratum, introducing formal correspondence between two languages as a system-based approach. He describes a formal correspondent as “any target language category (unit, class, element of structure . . . etc), which can be said to occupy as nearly as possible, the ‘same’ place in the ‘economy of the target language’ as the given source language category occupies in the source language” (p. 27). Halliday (2001) acknowledges that equivalence carries differential values at the different ranks, strata and metafunctions and affects translation quality. For instance, at the rank level, clause complex equivalence is more highly valued than phrasal equivalence. He, however, contends that equivalence can be highly valued at lower levels. In the same spirit, this study tries to demonstrate that equivalence at the NG rank can be used as a criterion for evaluating translation quality. As indicated in the introduction, the NG is important in realizing functional elements of the clause. This structure has an interesting potential for meaning building and expansion. At the same time, the translation of this structure may be challenging for Tunisian EFL learners because of the potential effect of translation shifts imposed by the systems of the two languages. While early research on translation associated equivalence with translation quality, in a more recent study, Matthiessen (2014) viewed translation equivalences not as criteria to judge translation quality but as the outcome of choices in a meaning-building process. The discrepancy across studies in the conceptualization of equivalence was also echoed in the conceptualization of translation shift. Catford (1965) argues that whenever a parallel structure is not possible, shift, a departure from formal correspondence, occurs. He draws a distinction between level shift and category shift. Level shift occurs when an element is expressed grammatically in one language (for example aspect) and lexically in another language. Category shift is subdivided into four categories (Catford, 1965, pp. 75–82): • •

Structural shift is a variation in grammatical structure or pattern. For example, Arabic is an VS language while English is an SV one. Class shift occurs when two languages use different structural categories. For example, an adjective in the source language is shifted into a noun in the target language.

298 •

Dorra Moalla Rank shift occurs when the translation equivalent in the target language is different in rank from the source language. For example, a phrase is translated into a clause. Intra-system shift occurs when the two languages have equivalent systems but the translator opts for a non-corresponding term in the target language system. For example, the English he has a broken leg is translated into French into il a la jambe cassée (Catford, 1965, p. 80).

Catford’s taxonomy of shifts was a valuable contribution to the field of translation, but it was heavily criticized for its focus on formal linguistic criteria and scientific probability rather than communicative considerations (Munday, 2008). It also simplified the translation process through conceiving shift as a deviation from equivalence dictated by failure to find equivalent structures. More recent research associated shift with choice in meaning. Through examining shifts in the translation of NGs modified by embedded clauses from English to Chinese, Fang and Wu (2009) argue that translators opt for translation shifts not only because of incompatibilities in the systems of the languages but also due to translator preferences and the functional role of the embedded clause. Within the same vein, Matthiessen (2014, p. 322) argues that translation is a process “of choice in meaning” among options in the logical, experiential, interpersonal and textual modes of meaning. The outcomes of these choices are degrees of translation shift that can be located on a cline extending from translation equivalence to translation shift. Despite the doubt cast upon the validity of equivalence and shifts in translation, they remain relevant concepts for the present study because it is applicable to the NG rank, in contrast to other studies (Bardi, 2008; Matthiessen, 2014), which are clause-oriented. By means of administering translation tasks to Tunisian EFL learners, this chapter re-examines the effect of this variable on translation performance from English to Arabic across different experiential categories. In view of the fact that translation is a complicated human activity, this chapter postulates that translation performance can equally be attributed to reasons related to the features of the NG (experiential features and structural realizations) and to translator-related factors, such as his/her awareness of the naturalness of the translation in the target language and translator preferences.

3. Research Problem and Hypotheses The present study draws on SFL to identify the difficulties that Tunisian EFL learners encounter in the translation of NGs from English to Arabic. It proposes to test the hypothesis that equivalences and shifts between Arabic and English account for translation success or failure. This hypothesis has two aspects. The first aspect is that equivalence

Nominal Groups in Arabic and English


in the pattern of modification and experiential classification between the two languages is a strong predictive factor of translation success. The second aspect is that the divergence in these areas accounts for the problems in EFL learners’ translations. To explain this, let us consider example (12), taken from Halliday and Matthiessen (2004, p. 429), where the NG in question is identified by italics (I focused on the NG in apposition to highlight the areas of difference in its realization across the languages): (12) Haffaz Assad, Syria’s autocratic president who dreamed of Arab unity but watched his neighbours sign peace deals with Israel, died Saturday.

In Table 15.4, the experiential structure of this NG is presented following the analysis given by Halliday and Matthiessen (2004, pp. 312–329). This nominal group is translated into Arabic in example (13). (13) rais-u

Souria al- mostabid-u

president-case Syria


alladhi haluma bi wihdati al‘arabi relative clause

Both NGs specify the Thing element ‘president’ raisu-‫ﺭﺋﻴﺲ‬. They nonetheless display many differences in terms of experiential features. In the experiential structure in English, the characterization of ‘president’ is expressed by the functional elements of Specific Deictic, Epithet and Qualifier. In Arabic, the categorization is realized by a set of embedded NGs, as shown in Table 15.5. In both languages, the adjective ‘autocratic’ almostabidu-‫ ﺍﻟﻤﺴﺘﺒﺪ‬is used to realize the Epithet function in SFG and qualification in Arabic grammar. While in English it is a modifier of the Thing ‘president’, in Arabic it is used as a qualification for an embedded NG called morakab iḍaafi-‫ﻣﺮﻛﺐ ﺇﺿﺎﻓﻲ‬, which is raisu souria-‫ﺭﺋﻴﺲ ﺳﻮﺭﻳﺎ‬ ‘Syria’s president’. This function has implications for the specification of this adjective in Arabic. Therefore, the adjective agrees in number, gender and specification with this construct. In English, adjectives are not subject to agreement rules. In English, the second Specific Deictic element ‘Syria’s’ is a modifier for the Thing ‘president’. The equivalent element in Arabic is a postmodifier for this head. The main similarity between the two languages is the Qualifier in SFG and qualification in Arabic Table 15.4 Experiential description of example (12) Syria’s



who dreamed of Arab unity

Specific Deictic possessive determinative

Epithet adjective

Thing noun

Qualifier relative clause


Dorra Moalla

Table 15.5 Analysis of the NG in Arabic following the grammar of constructs ‫ﺍﻟﺬﻱ ﺣﻠﻢ ﺑﻮﺣﺪﺓ ﺍﻟﻌﺮﺏ‬ who dreamed of Arab unity

‫ﺍﻟﻤﺴﺘﺒﺪ‬ the autocratic

‫ﺳﻮﺭﻳﺎ‬ Syria

‫ﺭﺋﻴﺲ‬ President

annexed to-‫ ﻣﻀﺎﻑ ﺇﻟﻴﻪ‬head-‫ﻣﻀﺎﻑ‬ annexation construct-‫ﻣﺮﻛﺐ ﺇﺿﺎﻓﻲ‬ qualification- ‫ ﻧﻌﺖ‬qualified- ‫ﻣﻨﻌﻮﺕ‬ qualifying construct-‫ﻣﺮﻛﺐ ﻧﻌﺘﻲ‬ qualification-‫ ﻧﻌﺖ‬qualified-‫ﻣﻨﻌﻮﺕ‬ qualifying construct-‫ﻣﺮﻛﺐ ﻧﻌﺘﻲ‬

grammar, which is realized by a restrictive relative clause labelled morakab mawsouli-‫ ﻣﺮﻛﺐ ﻣﻮﺻﻮﻟﻲ‬in Arabic. English relies on premodification and postmodification (Quirk et al., 2005). Modification is realized by possessive determinative elements and an adjective. Postmodification is realized by a relative clause. The NG in Arabic displays a hierarchical structure relying solely on postmodification. Table 15.5 illustrates the modification pattern of the NG in Arabic following the analysis of Kabani et al. (2000). At the higher level of hierarchy, the NG constitutes one construct with a qualifying function morakab na a͑ ti-‫ﻣﺮﻛﺐ ﻧﻌﺘﻲ‬. The qualified Rāisu Souria almostabidu-‫ﺭﺋﻴﺲ ﺳﻮﺭﻳﺎ ﺍﻟﻤﺴﺘﺒﺪ‬ ‘Syria’s autocratic president’ is postmodified by a relative clause morakab mawsouli-‫ﻣﺮﻛﺐ ﻣﻮﺻﻮﻟﻲ‬. As shown in Table 15.5, this NG contains embedded NGs. For instance, rāisu souria-‫‘ ﺭﺋﻴﺲ ﺳﻮﺭﻳﺎ‬Syria’s president’ is an annexation construct morakab Iḍaafi-‫ ﻣﺮﻛﺐ ﺇﺿﺎﻓﻲ‬composed of the head noun rāisu-‫‘ ﺭﺋﻴﺲ‬president’ postmodified by souria-‫‘ ﺳﻮﺭﻳﺎ‬Syria’s’. Given the differences in NG realization, the question worth raising is whether these differences account for translation quality. To answer this question, the chapter adopts a methodology that will be described in the next section.

4. Methodology 4.1. Test Design and Data Collection A set of thirteen NGs in English were written by the researcher and then inserted into sentences. The selection of authentic NGs was discarded because they often occur within clause complexes, which may constitute intervening variables that may constitute additional challenges for translators and can divert their attention from NG translation. The NGs displayed different patterns of modification (they included both Modifiers and Qualifiers), different experiential categories (Deictics,

Nominal Groups in Arabic and English


Numeratives, Epithets, Classifiers, Thing elements and Qualifiers) and different structural realizations (for example PPs, finite and non-finite Qualifiers . . . etc.). They varied in word number (from four to thirteen words). They also displayed different degrees of complexity (the number of embedded NGs ranged from one to four). Thirty students were asked to translate these NGs into Arabic. This task was piloted to check the task validity and to decide about the time allocated to it. The researcher wrote the translation of the NGs into Arabic. These translations were used as the standards against which the students’ translations were evaluated. Two specialists in Arabic checked and revised them to verify the two following conditions: • •

The equivalence between the structures across the languages (formal equivalence) The acceptability of the translation in the target language (dynamic equivalence)

4.2. Methods of Data Analysis Different steps of data analysis were followed: 1. The NGs were divided into their experiential categories (Deictic, Numerative, Epithet, Classifier . . . etc.) in order to evaluate the participants’ mastery of the different experiential functions. 2. The different NGs were divided into their component structural realizations (for instance noun as Classifier, PP as Qualifier . . . etc.). Then participants’ performance was calculated. 3. After undertaking a quantitative analysis of participants’ performance, a qualitative analysis was undertaken. It offered an in-depth examination of particular NGs in order to identify the main translation problems and strategies.

5. Results Based on the quantitative analysis of the translations; participants generally did not provide an exact equivalent structure in the target language, with only 24% of translations being equivalent structures. Analysis of participants’ performance in experiential categories indicated that English Qualifiers were the most difficult to translate. As shown in Figure 15.1, to-infinitives and ing-clauses Qualifiers were challenging. In contrast, participants were quite successful in translating English PPs and relative clauses Qualifiers into equivalent structures in Arabic. Analysis of participants’ performance according to structural realizations for the different experiential categories indicated that the participants found it difficult to match the ed- and ing-participles as Modifiers with the equivalent structure in Arabic. However, they were more successful in finding the Arabic


Dorra Moalla



34 17

PP as Qualifier



Ed-participle as Qualifier

To-infinitive as Qualifier

Ing-participle as Qualifier

Appositive clause as Qualifier

Relative clause as Qualifier

Figure 15.1 Percentage of Arabic equivalents for English Qualifiers in participants’ performance

equivalent structures of English nouns used as Classifiers and adjectives used as Epithets and Numeratives. However, within a single category, there was a lack of consistency in participants’ performance. When translating NGs with an Epithet, the percentage of translations that matched the structure ranged from 3.5% to 83.5% across the examples. In the same spirit, the translation of ingclause as Qualifiers into an equivalent structure in Arabic ranged between 7% and 96% across the examples. This fact has probably reduced the reliability of the quantitative analysis. Because the discrepancies between elements within a category were obscured by quantitative analysis, I have opted for a qualitative analysis that undertakes an in-depth examination of specific experiential categories and structures in order to account for reasons for translation problems and choices. The qualitative analysis will highlight the following two findings. The first finding is that translation equivalences and shifts are strong predictive factors of translation performance for several experiential categories. The second finding is that shifts and equivalence effects remain relative, and translator-related factors account for translation performance. 5.1. Structural Shift Effect 5.1.1. Genitive: The Impact of Translation Shifts The effect of the language shift hypothesis was manifested in the translation of the Possessive Deictic (genitive). As illustrated in the research

Nominal Groups in Arabic and English


hypothesis, the Possessive Deictic element is realized in different ways in the two languages in both experiential features and pattern of modification. I will illustrate this shift through two examples and will highlight the translation strategies adopted by the participants to cope with this shift. Consider this example (14): (14)

I heard about our competitor’s launch of a new model of fax machines.

The parallel translation of this NG can be realized through the use of the pseudo-verbal construct almorakab chibh alisnadi-‫ﺍﻟﻤﺮﻛﺐ ﺷﺒﻪ ﺍﻹﺳﻨﺎﺩﻱ‬, which is illustrated in example (15). This construct is based on the derived noun labelled verbal noun almaṣdar-‫ﺍﻟﻤﺼﺪﺭ‬, which is iṣdar‫‘ ﺇﺻﺪﺍﺭ‬launch’. This form is derived from the verb ās̠ dara-‫‘ ﺃﺻﺪﺭ‬to launch’. This head is followed by the annexation construct morakab iḍaafi-‫ ﻣﺮﻛﺐ ﺇﺿﺎﻓﻲ‬monafisuna-‫‘ ﻣﻨﺎﻓﺴﻨﺎ‬our competitor’ and the PP linamoudhagin jadidin-‫‘ ﻟﻨﻤﻮﺫﺝ ﺟﺪﻳﺪ‬of a new model’ taking the object role, as shown in Table 15.6. This construct is pseudo-verbal because it can be paraphrased as ‘Our competitor launched a new model of fax machines’. (15)

‫ﺍﺻﺪﺍﺭ ﻣﻨﺎﻓﺴﻨﺎ ﻟﻨﻤﻮﺫﺝ ﺟﺪﻳﺪ ﻣﻦ ﺁﻻﺕ ﺍﻟﻔﺎﻛﺲ‬ is̠dar

monafisi-na li-namodhaj- jadid-in min eleti al-fax in launch competitor- prep-model- new-nun prep machines specour nun fax ‘the launch of our competitor’s new model of fax machines’

Table 15.6 Analysis of example (15) using the grammar of constructs ‫ﺍﻟﻔﺎﻛﺲ‬



‫ﻟﻨﻤﻮﺫﺝ ﺟﺪﻳﺪ‬


‫ﺟﺪﻳﺪ ﻣﻦ‬




the machines of new PP/distinction-‫ﺍﻟﺘﻤﻴﻴﺰ‬



model of







competitor launch

qualifying construct/ distinguished‫ﻣﺮﻛﺐ ﻧﻌﺘﻲ‬/‫ﺍﻟﻤﻤﻴﺰ‬

distinction construct-‫ﻣﺮﻛﺐ ﺗﻤﻴﻴﺰﻱ‬

PP/direct object-‫ﻣﻔﻌﻮﻝ ﺑﻪ‬/‫ﻣﺮﻛﺐ ﺑﺎﻟﺠﺮ‬

prep pronoun/ noun/ annexed annexed— to‫ﻣﻀﺎﻑ‬ ‫ﺿﻤﻴﺮ‬/‫ﻣﻀﺎﻑ‬ ‫ﺇﻟﻴﻪ‬ annexation construct‫ﻣﺮﻛﺐ ﺇﺿﺎﻓﻲ‬

pseudo-verbal construct based on the derived noun ‘launch’‫ﻣﺮﻛﺐ ﺷﺒﻪ ﺇﺳﻨﺎﺩﻱ ﻗﺎﺋﻢ ﻋﻠﻰ ﺍﻟﻤﺼﺪﺭ‬

head noun


Dorra Moalla

The analysis indicated that the NG was not among the translation choices in translation. For example, 40% of participants substituted the NG with a clause complex (projection), as shown in example (16): (16)

‫ﺳﻤﻌﺖ ﺃﻥ ﻣﻨﺎﻓﺴﻨﺎ ﺍﺻﺪﺭ ﻧﻤﻮﺫﺟﺎ ﺟﺪﻳﺪﺍ ﻣﻦ ﺁﻻﺕ ﺍﻟﻔﺎﻛﺲ‬ ‘I heard that our competitor launched a new model of fax machines’

The use of a clause reflects a subconscious awareness of the congruent representational value of this NG (Halliday and Matthiessen, 2004, p. 637). The genitive has an actor role, while the noun has a process one. In the same vein, reflecting the participants’ awareness of the participant role of the Possessive Deictic, 25% of participants opted for substituting the original NG with a NG having a relative clause qualifier ‘‫‘ ’ﺳﻤﻌﺖ ﻋﻦ ﻧﻤﻮﺫﺝ ﺟﺪﻳﺪ ﻣﻦ ﺁﻻﺕ ﺍﻟﻔﺎﻛﺲ ﺍﻟﺬﻱ ﺃﺻﺪﺭﻩ ﻣﻨﺎﻓﺴﻨﺎ‬I heard about a new model of fax machines that our competitor launched’. In example (17), I consider the following NG in italics that includes a Possessive Deictic: (17)

This is the country’s largest trading company.

As opposed to the NG in English, the genitive occurs in final position in the parallel NG in Arabic in the form of a PP lilbiladi-‫‘ ﻟﻠﺒﻼﺩ‬of the country’. This shift was a challenging difficulty in the translation, and, accordingly, 68% of participants changed the possessive meaning into a locative meaning fil biladi-‫‘ ﻓﻲ ﺍﻟﺒﻼﺩ‬in the country’. Opting for a non-equivalent structure in the translation of the possessive can be explained by the translators’ awareness of the choices in meaning that the system offers rather than by the failure to find an equivalent structure. This explanation is consolidated by Brinton and Brinton’s (2010, p. 319) argument that possession can be treated as a special type of location that indicates “the location of an entity coincidental with a person” (here the country). 5.1.2. Noun Classifier: The Impact of Translation Shifts Another experiential shift from English to Arabic that constituted a major challenge for participants was the translation of noun Classifiers, as illustrated in the following example (18): (18)

He attended optional professional development workshops.

In the NG in italics, the Head noun is modified by two Classifiers: ‘optional’ and ‘professional development’. Arabic offers two possible translations of this NG. In the first translation, the adjective occurs after the head, followed by a PP ‘workshops for professional development’. This structure, introduced as example (19), is increasing in Arabic and is dictated more by the rules of usage than by the rules of grammar.

Nominal Groups in Arabic and English (19)


‫ﺍﻟﻮﺭﺷﺎﺕ ﺍﻻﺧﺘﻴﺎﺭﻳﺔ ﻟﻠﺘﻄﻮﻳﺮ ﺍﻟﻤﻬﻨﻲ‬ al-warach-at-u al-ikhtiari-at-u li-ataṭwiri almihnii spec-workshop- spec-optional- prep-development professional plu.fem-case fem-case ‘the optional workshops for professional development’

A more grammatically accepted translation is illustrated in example (20) and Table 15.7. (20)

‫ﻭﺭﺷﺎﺕ ﺍﻟﺘﻄﻮﻳﺮ ﺍﻟﻤﻬﻨﻲ ﺍﻻﺧﺘﻴﺎﺭﻳﺔ‬ warachat-u a-taṭwir-i al-mihni-i workshops- spec-development- spec-professionalcase case case ‘optional professional development workshops’

al-ikhtiari-yat-u spec-optionalfem-case

This translation differs from the English NG in the pattern of modification, experiential classification and specification. This construct is based on postmodification only. ‘Professional development workshops’ are qualified as optional. The whole construct is a qualifying construct that includes an annexation construct, which is ‘professional development workshops’. In experiential classification, the NG is a qualifying construct, unlike the NG in English, which is based on Classifiers. As opposed to the unspecified NG in English, the one in Arabic is specified through the affixation of al-‫‘ ﺍﻝ‬the’ in the embedded qualifying construct and through aliḍaafa-‫ﺍﻹﺿﺎﻓﺔ‬, (see Table 15.7). This discrepancy is echoed in participants’ performance, with 60% of them opting for the PP in NG final position ‘the optional workshops for professional development’, which is the less grammatically acceptable structure.

Table 15.7 Analysis of example (20) using the grammar of constructs ‫ﺍﻻﺧﺘﻴﺎﺭﻳﺔ‬ the optional



the professional

the development



‫ﻭﺭﺷﺎﺕ‬ workshops

qualifying construct-‫ﻣﺮﻛﺐ ﻧﻌﺘﻲ‬ annexed to-‫ﺍﻟﻤﻀﺎﻑ ﺇﻟﻴﻪ‬


annexation construct-‫ﻣﺮﻛﺐ ﺇﺿﺎﻓﻲ‬ qualification-‫ﺍﻟﻨﻌﺖ‬

qualified-‫ﻣﻨﻌﻮﺕ‬ qualifying construct-‫ﻣﺮﻛﺐ ﻧﻌﺘﻲ‬


Dorra Moalla

5.2. Translation Shift: A Weak Variable The previous section has demonstrated that experiential and structural variation in NGs between Arabic and English may be major sources of translation problems. Through analysing participants’ translation of specific experiential categories, this section will show that the translation shift/ equivalence hypothesis may be a weak predictor of participants’ performance. This finding has two aspects. Many translation shifts were quite easy to translate, and equivalent structures were difficult to translate. This section tries to show that, bearing in mind that translation is a complicated human activity, translation performance cannot be predicted purely based on structural features. Translator-related factors may override or limit the structural effect. These factors may include translator preferences, his/her awareness of the naturalness of the message in the target language, and the negotiation process between translator knowledge of the system and the semantic value of NGs. The effect of translator-related factors will be highlighted through the translation of specific experiential categories.

5.2.1. Agreement in Epithet/Qualifying Constructs Agreement in qualifying constructs stands as an example of the weak effect of structural and experiential shifts. The sentence considered as example (21) includes an embedded NG in italics. Agreement in English is gender free. Number is marked by the pluralization of the Thing/Head, and the NG is unspecified, denoting ‘angry or frustrated customers’ as a general category. (21) She has the ability to calm angry or frustrated customers

The Arabic translation of this NG, represented as example (22), is marked for number, gender and specification. There is agreement in number and gender between the qualification ana͑ at-‫ﺍﻟﻨﻌﺖ‬, which is alghaḍibina aw almoḥbaṭina‫‘ ﺍﻟﻐﺎﺿﺒﻴﻦ ﺃﻭ ﺍﻟﻤﺤﺒﻄﻴﻦ‬angry or frustrated’, and the qualified alman͑ out-‫ﺍﻟﻤﻨﻌﻮﺕ‬ ‘customers’ through the plural and masculine inflection of the different elements. Specification is realized for both the qualification and the qualified through the addition of the specification prefix al-‫‘ ﺍﻝ‬the’ for the different components of this NG. This variation did not have an important impact on participants’ performance. The majority of the participants (63%) managed to translate it into a specified qualifying construct. This may be attributed to the participants’ awareness of naturalness in the target language. (22)

‫ﺍﻟﺰﺑﺎﺋﻦ ﺍﻟﻐﺎﺿﺒﻴﻦ ﺍﻭ ﺍﻟﻤﺤﺒﻄﻴﻦ‬ a-zaba-in-i al-ghaḍib-in-a spec-customer-plu. spec-angry-plu. masc-case masc-case ‘angry or frustrated customers’

aw al-moḥbaṭ-in-a or spec-frustratedplu.masc-case

Nominal Groups in Arabic and English


5.2.2. The Specific Deictic/Possessive Determinative The Possessive Deictic is not realized according to the same experiential and structural pattern in Arabic and English. In English, the possessive determinative element is a separate word that occurs before the Head/ Thing element. In example (23), mounafisuna-‫‘ ﻣﻨﺎﻓﺴﻨﺎ‬our competitor’, the possessive determinative element is a pronoun suffixed to the head/thing element, together forming an annexation construct. This variation did not constitute a challenging difficulty (85% of participants managed the translation of this construct). This ease in translating this experiential category, despite the structural and morphological difference, may be attributed to the similarity in the semantic value of the structures as well as to the participants’ knowledge of the target language. (23) ‫ﻣﻨﺎﻓﺴﻨﺎ‬ mounafisu-na competitor-poss

5.2.3. Epithet/‘Morakab Tamyiizi-‫’ﻣﺮﻛﺐ ﺗﻤﻴﻴﺰﻱ‬ Another example challenging the structural variation hypothesis was the translation of the NG in italics introduced as example (24): (24)

This is one of the fastest growing global brands.

The translation of this NG into Arabic, shown in example (25), is different from the NG in English in word class classification and experiential structure. As shown in Table 15.8, modifiers in English include a compound adjective realized by a superlative and an ing-participle ‘fastest growing’ and an adjective ‘global’. The NG is specified through the use of the Specific Deictic in NG initial position. In Arabic, ‘global’ and ‘fastest’ are realized by the same word class. The translation of the ing-participle, on the contrary, resulted in a class and structural shift as it is realized by an unspecified noun marked by case inflection and occupying an end position in this NG (see Table 15.8). Specification is realized at two levels. At the highest level, the whole NG is specified through annexation, and at a lower level, the qualifying construct almarketi al‘alamiyati-‫‘ ﺍﻟﻤﺎﺭﻛﺎﺕ ﺍﻟﻌﺎﻟﻤﻴﺔ‬the global brands’ is specified through Table 15.8 Experiential division and structural realization of the NG in English in example (24) The








definite article

compound adjective






Dorra Moalla

the prefixation of ‘al-‫ ’ﺍﻝ‬for both the qualified and the qualification (see Table 15.9). (25) ‫ﺃﺳﺮﻉ ﺍﻟﻤﺎﺭﻛﺎﺕ ﺍﻟﻌﺎﻟﻤﻴﺔ ﺗﻄﻮﺭﺍ‬ āsra‘i al-mark-et-i al-‘alamiy-at-i fastest spec-brand-plu.femspec-global-femcase case ‘the fastest global brands in growth’

taṭawor-an growth-unspec

In experiential structure, the NG in English is modified by an Epithet and a Classifier. In Arabic, this NG includes embedded NGs, as illustrated in Table 15.9. The NG comprises an annexation construct morakab iḍaafi‫ ﻣﺮﻛﺐ ﺇﺿﺎﻓﻲ‬headed by the superlative asra‘i- ‫‘ ﺃﺳﺮﻉ‬fastest’, which is labelled almoḍaaf-‫ﺍﻟﻤﻀﺎﻑ‬. This head noun is modified by the qualifying construct almarketi al‘alamiyati ‘the global brands’. A distinction Tamyiiz-‫ﺗﻤﻴﻴﺰ‬ occurs in construct end position to specify the level at which this global brand was fast (growth). This construct can be literally translated as ‘the fastest global brands in growth’. Despite the structural and experiential shifts between the NG in English and the one in Arabic, 80% of participants were quite successful in its translation into Arabic. This success may be attributed to the participants’ association process between the semantic value of the NG in English and structural options offered by the system to express this meaning in Arabic rather than to equivalence/shift criteria.

Table 15.9 Analysis and structural realization of the NG in Arabic in example (25) ‫ﺗﻄﻮﺭﺍ‬





the global

the brands








qualifying construct-‫ﻣﺮﻛﺐ ﻧﻌﺘﻲ‬ annexed to-‫ﻣﻀﺎﻑ ﺇﻟﻴﻪ‬ distinction-‫ﺗﻤﻴﻴﺰ‬


annexation construct-‫ﻣﺮﻛﺐ ﺇﺿﺎﻓﻲ‬

pseudo-verbal construct based on the superlative-‫ﻣﺮﻛﺐ ﺷﺒﻪ ﺇﺳﻨﺎﺩﻱ ﻗﺎﺋﻢ ﻋﻠﻰ ﺇﺳﻢ ﺍﻟﺘﻔﻀﻴﻞ‬

5.3. Structural Equivalence and Translation Problems In this section, I will try to show that semantic, structural and experiential parallelism between NGs is not a facilitating factor in translation.

Nominal Groups in Arabic and English


5.3.1. The Translation of the Ing-Participle as Epithet Example (26) (the NG is in italics) stands as an example supporting the relatively weak effect of structural equivalence: (26)

An employee who is well aware of changing trends in the field

As will be shown, this NG is quite similar to its Arabic translation in semantic value and structural features. This similarity, however, did not eliminate translation failure, as only 13.5% of participants were successful in its translation. The English NG is composed of a Head ‘trends’ preceded by an ingparticiple as a Modifier and followed by a PP Qualifier. The translation of Qualifiers with a locative meaning ‘in the field’ did not raise real problems. In contrast, the translation of the ing-participle was challenging. In English, this structure can be classified as an Epithet because it closely relates to the finite tense in the present continuous ‘trends which are changing’. As shown in Example (27), the Arabic translation of the ing-participle almotaghayirati-‫ ﺍﻟﻤﺘﻐﻴﺮﺓ‬is a derived form of the active participle ism faa ͑el‫ ﺇﺳﻢ ﻓﺎﻋﻞ‬that represents the doer of the action. This form is derived through two processes of prefixation and infixation from the verb taghayara-‫ﺗﻐﻴﺮ‬ ‘to change in state’. Semantically, this active participle qualifies the head noun ‘trends’ as continuously changing in state. The construct anaza‘ati almotaghayirati-‫‘ ﺍﻟﻨﺰﻋﺎﺕ ﺍﻟﻤﺘﻐﻴﺮﺓ‬changing trends’ is a qualifying construct morakab na ͑ ati-‫ ﻣﺮﻛﺐ ﻧﻌﺘﻲ‬composed of the head anaza‘ati-‫‘ ﺍﻟﻨﺰﻋﺎﺕ‬trends’, which is the qualified, and the postmodifier ‘changing’, which is the qualification. Therefore, the two NGs exhibit structural and semantic similarities. They can be paraphrased using a finite verb in the continuous, in which the head plays the role of actor/Faa e͑ l-‫‘ ﻓﺎﻋﻞ‬trends that are changing’. Semantically, the modifier in both languages indicates the quality of the Thing as ‘changing’, a role typically assumed by an Epithet in SFG and a qualification in Arabic grammar (Halliday and Matthiessen, 2014, p. 379). (27)

‫ﺍﻟﻨﺰﻋﺎﺕ ﺍﻟﻤﺘﻐﻴﺮﺓ‬ a-naza‘-at-i spec-trend-plu.fem-case ‘changing trends’

al--taghay--r-at-i spec- change-fem-case

As shown at the beginning of this section, this similarity was not predictive of participants’ translation success. This may be explained by the preferences for particular structures. Thus, a majority (66.75% of participants) substituted the original qualifying construct with an annexation construct taghayoru anaza‘ati-‫‘ ﺗﻐﻴﺮ ﺍﻟﻨﺰﻋﺎﺕ‬trend change’, which resulted in structural, experiential and semantic shifts. The head of the NG has become ‘change’, rather than ‘trends’ as in the original NG. Semantically


Dorra Moalla

Table 15.10 Original NG versus translation preference in example (27) ‫ﺍﻟﻨﺰﻋﺎﺕ ﺍﻟﻤﺘﻐﻴﺮﺓ‬ (Changing trends)

Original nominal group changing




active participle


‫ﺗﻐﻴﺮ ﺍﻟﻨﺰﻋﺎﺕ‬ (Change in trends)

Participants’ translation preference the trends


annexed to




speaking, the continuous meaning associated with ism faa ͑el-‫ﺇﺳﻢ ﻓﺎﻋﻞ‬ ‘active participle’ has not been faithfully translated in this translation (see Table 15.10). 5.3.2. Translation of Non-Finite Qualifiers The weak effect of structural and experiential similarity on participants’ translations was particularly clear in the translation of non-finite clauses used as Qualifiers. If we consider example (28), the NG (in italics) contains a non-finite ed-participle clause Qualifier. (28)

The advertising agency organised a successful trial run in Mexico City.

The structurally equivalent NG in Arabic is illustrated in example (29). The two NGs display similar structural features and experiential structures. Both contain a clausal postmodifier. In the NG in Arabic, the edparticiple clause is paralleled with a clausal postmodifier labelled a verbal construct morakab isnadi-‫ﻣﺮﻛﺐ ﺇﺳﻨﺎﺩﻱ‬, which functions as a qualification of the head noun moḥawalatan-‫‘ ﻣﺤﺎﻭﻟﺔ‬a trial’. The verb form is obtained through the inflection of the verb āqama-‫‘ ﺃﻗﺎﻡ‬to run’. The inflected structure is a passive verb without an actor almabni lilmajhoul-‫ﺍﻟﻤﺒﻨﻲ ﻟﻠﻤﺠﻬﻮﻝ‬, a parallel structure to the ed-participle clause. (29)

‫ﻣﺤﺎﻭﻟﺔ ﻧﺎﺟﺤﺔ ﺍﻗﻴﻤﺖ ﺑﻤﺪﻳﻨﺔ ﻣﻜﺴﻴﻜﻮ‬ moḥawalatan nejiḥatan -q--mat trial successful -run ‘a successful trial run in Mexico City’

bi-madinat meksiko in-city Mexico

Despite the correspondence between the two NGs in the Qualifier structure and semantic value, this parallelism was not a predictive factor

Nominal Groups in Arabic and English


of translation success. Only one participant managed to translate the qualifier into a passive verb. The participants opted for more frequently used structures in Arabic, with 55% of participants turning the edparticiple clause into an annexation construct morakab iḍaafi-‫ﻣﺮﻛﺐ ﺇﺿﺎﻓﻲ‬, which is moḥawalutu houroubin-‫‘ ﻣﺤﺎﻭﻟﺔ ﻫﺮﻭﺏ‬a running trial’. This may indicate that the verb ‘to run’ ‘to organize’ in the past participle form was mistaken for the noun ‘run’, resulting in a deviation from the original meaning. Many participants (36%) opted for ignoring the ed-participle Qualifier, turning this NG into ‘‫‘ ’ﻣﺤﺎﻭﻟﺔ ﻧﺎﺟﺤﺔ ﻓﻲ ﻣﺪﻳﻨﺔ ﻣﻜﺴﻴﻜﻮ‬a successful trial in Mexico City’. These structural preferences indicate that a negotiation process occurred between the translator, who is aware of what the system offers, and the semantic value of the NG. This negotiation has, however, induced a Short Circuit Hypothesis. This concept was adopted from reading research (Clarke, 1988). According to this hypothesis, the reader draws on his/her background knowledge to build limited and erroneous deductions about lexical or grammatical cues. In the same vein, the translators made erroneous deductions induced by the presence of the PP ‘in Mexico City’, and accordingly the Qualifier was reduced to the locative meaning. The presence of the word ‘run’ has also induced wrong deductions about its semantic value.

6. Conclusion This chapter has endeavoured to seek appliabilities of the field of SFL to translation and language learning. On the grounds that the NG is a central structure from grammatical and pedagogical perspectives, the disciplines of translation and language learning have cross-fertilized with the field of SFL to provide a new understanding of NG translation problems. SFL has offered a framework for the classification of NGs into experiential categories and structural realizations. This framework has served to compare and contrast the experiential classification and structural realizations of NGs in order to highlight the areas of equivalence and shift in this unit across the languages. Afterwards, through administering translation tasks to Tunisian EFL learners, this study attempted to provide a real probe of the effect of shifts and equivalences on translation based on authentic tasks rather than on predictions. The results indicate that although the translation shift and equivalence hypothesis is a relatively strong predictor of participants’ performance for several experiential categories (for instance genitive and compound nouns as Classifiers), the effect of this variable remains relatively weak and even counter-productive. Many equivalent structures (for instance non-finite Qualifiers) are difficult to translate, and many translation shifts are easily translated (for example ing-participles as Epithets, and Possessive Deictics). In view of the fact that translation is a complicated human endeavour, the present study has posited that translation performance cannot be


Dorra Moalla

attributed to purely structural reasons. Translator-related factors remain important factors for determining translation performance. They include the translator preferences and the concern with naturalness in the target language. This chapter has shown that translation success and failure are not attributable to one-to-one structural correspondences between Arabic and English or to incompatibilities between the systems of the two languages. It is a much more complex process involving negotiations between the translator, who is aware of the options the system offers, and the semantic value of the NG, which leads to opting for more natural resources in the target language or more frequently used ones. Though the role of structural factors has been relatively undermined in this chapter, a functional approach to this issue remains valid because both linguistic and non-linguistic factors are embodied by structural realizations in the different experiential categories. This chapter has offered a reconsideration of translation equivalence/ shift effects through providing a real and reliable probe of this variable. This diagnosis can be exploited by translation specialists and teachers to plan remedial work to overcome these problems. The results of the study remain, however, exploratory in nature and should be consolidated by future research undertaking an in-depth examination of the effect of equivalences and shifts within individual experiential categories (for example PP Qualifiers, Classifier, Epithet . . . etc.). Future research developing an SFL-based experiential classification of NGs in Arabic is needed to undertake a more reliable comparison of NG translation from English to Arabic.

Notes 1. The NG is a group unit within Systemic Functional Grammar, in which the Head is a noun. It is similar to the noun phrase (NP) in descriptive grammar (see Quirk, Greenbaum, Leech and Svartvik, 2005). 2. In the glosses, the following abbreviations are used: ACT PART: active participle; FEM: feminine; GEN CASE: genitive case; MASC: masculine; NUN: nunation; PASS INFLEC: passive inflection; PLU: plural; POSS: possessive; POSS DET: possessive determinative; PREP: preposition; REL PRON: relative pronoun; SING: singular; SPEC: specified; UNSPEC: unspecified; 3rd SING. PAST: the verb is in the third person singular in the past; 1st SING.PAST: the verb is in the first person singular in the past. 3. The noun construct framework al-morakabet al-ismia-‫ ﺍﻟﻤﺮﻛﺒﺎﺕ ﺍﻹﺳﻤﻴﺔ‬is part of a school of grammar adopted in the study of language in Tunisian universities, which is the grammar of constructs. It views language as a hierarchical system composed of the word, the group and the clause. Within the group level, there are, for instance, almorakab bil jar-‫( ﺍﻟﻤﺮﻛﺐ ﺑﺎﻟﺠﺮ‬PP) and almorakab alismi-‫ﺍﻟﻤﺮﻛﺐ ﺍﻻﺳﻤﻲ‬. This group is equivalent to the NP in English grammar. It is composed of a head, which is a noun, and postmodifier(s). In Arabic grammar, the relationship between head and modifiers is governed by functional and structural rules (this will be explained within this section).

Nominal Groups in Arabic and English


4. Nunation is a marker of indefiniteness in Arabic for nouns and adjectives through the addition of a final nun-‫ ﻧﻮﻥ‬to the final consonant. It has three different pronunciations depending on the case of the noun or adjective: ‘nun’ in the nominative case, almarfou ͑u-‫‘ ;ﺍﻟﻤﺮﻓﻮﻉ‬nin’ in the genitive case, almajrur‫ ;ﺍﻟﻤﺠﺮﻭﺭ‬and ‘nan’ in the accusative case, almans̠ ub-‫ﺍﻟﻤﻨﺼﻮﺏ‬

References Bardi, M. A. (2008) Towards a Systemic Functional Description of Arabic. Unpublished PhD thesis, University of Sydney, Australia. Brinton, J. L. and Brinton, M. D. (2010) The Linguistic Structure of Modern English. Philadelphia: John Benjamins Publishing Company. Catford, J. C. (1965) A Linguistic Theory of Translation. London: Oxford University Press. Clarke, M. A. (1988) The Short circuit hypothesis of ESL reading-or when language competence interferes with reading performance. In P. L. Carrell, J. Devine and J. E. Eskey (Eds.), Interactive Approaches to Second Language Reading. New York: Cambridge University Press, pp. 114–124. Connor, U. (2002) New directions in contrastive rhetoric. TESOL Quarterly, 36, pp. 493–510. Fang, J. and Wu, C. (2009) Exploring shifts in translating English nominal groups modified by embedded clauses: A corpus-based approach. A paper presented at ASFLA Conference: Practising Theory: Expanding Understanding of Language, Literature and Literacy, held on the Queensland University of Technology, Brisbane, Wednesday 30 September–Friday 2 October. Fontaine, L. (2013) Analysing English Grammar: A Systemic Functional Introduction. Cambridge: Cambridge University Press. Fontaine, L. (2017) The English nominal group: The Centrality of the thing element. In T. Bartlett and G. O’Grady (Eds.), The Routledge Handbook of Systemic Functional Linguistics. England: Routledge, pp. 267–283. Halliday, M. (2001) Towards a theory of good translation. In E. Steiner and C. Yallop (Eds.), Beyond Content: Exploring Translation and Multilingual Text. Berlin: De Gruyter, pp. 13–18. Halliday, M. and Matthiessen, C. (2004) An Introduction to Functional Grammar, 3rd ed. London: Arnold. Halliday, M. and Matthiessen, C. (2014) Halliday’s Introduction to Functional Grammar. UK: Routledge. Hatim, B. (2013) Teaching and Researching Translation. London: Routledge. Hatim, B. and Munday, J. (2004) Translation: An Advanced Resource Book. London: Routledge. Jobrane, K. J. (1918) Almajmou‘a alkamila liā‘mel Jobrane Khalil Jobrane: Al‘awa s̠if. Lebanon: Naoufel. Kabani, S., Bannour, O., Missaoui, S. and Chaouech, M. (2000) Annahw Alaarabi. Tunis: Centre National Pédagogique. Matthiessen, C. (2014) Choice in translation: Metafunctional considerations. In K. Kunz, E. Teich, S. Hansen-Schirra, S. Neumann and P. Daut (Eds.), Caught in the Middle-Language Use and Translation: A Festschrift for Erich Steiner on the Occasion of His 60th Birthday. Saarland: Saarland University Press, pp. 271–333.


Dorra Moalla

Munday, J. (2008) Introducing Translation Studies: Theories and Applications. London: Routledge. Nida, E. A. (1964) Toward a Science of Translating. Leiden: E. J Brill. Quirk, R., Greenbaum, S., Leech, G. and Svartvik, J. (2005) A Comprehensive Grammar of the English Language. London: Longman. Ryding, K. C. (2005) A Reference Grammar of Modern Standard Arabic. Cambridge: Cambridge University Press. Zied, M. (2009) Alwajizu fi annahwi. Karthoum: Assahwa Publications.

16 Cancer Care as an Integrated Practice Consultations Between an Oncologist and Patients With Advanced, Incurable Cancer Neda Karimi, Alison Moore and Annabelle Lukin 1. Introduction Oncological care is going through a transformation from a practice in which the patient is the object of the medical practitioner’s implementation of anti-tumour therapies to a multifaceted practice where the patient is an active participant in a dialogic relationship with a medical practitioner. The oncologist’s role is far from simple. Surbone, Zwitter, Rajer and Stiefel (2012) suggest diagnostics, treatment and communication as the three pillars of oncology. Cherny and Catane (2011) add palliation to this list. The development of oncology practice management, psychooncology and palliative oncology as interdisciplinary fields of study and the introduction of palliative medicine into oncological care show the complexity and the diversity of the competing demands on the oncologist. This cultural shift entails a diversification in the range of business considered as part of the purview of the oncologist. A recent statement of these diverse responsibilities (Cherny and Catane, 2011), published in the Oxford Textbook of Palliative Care, adds to preventative oncology, diagnostic evaluation and anti-tumour therapies— arguably core to oncology as a professional specialization—the further tasks of ‘communication’, ‘symptom control’, the ‘optimization of the social support’ and ‘care of the dying patient’. None of these tasks are specific to the job of an oncologist. However cultural pressures that have their roots in Carl Rogers’s concept of ‘client-centred therapy’ (1942), Michael Balint’s notion of ‘patient-centred medicine’ (Balint, Ball and Hare, 1969), Barney Glaser and Anselm Strauss’s concept of ‘open awareness’ (1965) and Cicely Saunders’s ‘hospice movement’ (1967) have created the conditions under which these tasks have come to be considered an essential part of medical oncology practice, even a ‘moral and clinical imperative’ (Cherny, Catane and Kosmidis, 2003). What is the effect of these cultural pressures on the context of oncological consultations? What do they mean for the tenor relations


Neda Karimi et al.

between oncologist and patient? Do they change the nature of the social process—the field in Halliday’s terms—of the clinical dialogue? How does this changing context of culture translate into actual instances of oncology consultations? Despite the substantial influence of concepts such as patient-centred care, there is very little research that describes or evaluates patient-centred care in terms of specific communicative and interactive patterns, leaving the question still open as to what such practice really looks like in terms of its linguistic profile—or. in other words, as a register. We consider these questions, drawing on a case study of one oncologist going about the business of ‘doing oncology’ in consultations with ten different patients, at different stages in their cancer treatment. The case study reported here is selected from a larger sample of data collected as part of a randomized controlled trial (RCT) of an intervention called a ‘communication support programme’ (CSP), designed and conducted by researchers at the Centre for Medical Psychology and Evidence-Based Decision-Making (CeMPED, University of Sydney) at medical oncology clinics based at or affiliated with major hospitals in Sydney, Australia (Walczak et al., 2014).1 All patients in the study were diagnosed with a terminal cancer. Thus, the oncologists were treating patients who could not—and did not—survive their cancers. While the CSP targeted the patients and their care givers, the focus of this chapter is on the oncologist as the “prime cancer specialist” who “coordinates cancer care in all of its phases” (Cherny et al., 2003, p. 1335). The data used here is a corpus of ten consultations conducted by the same oncologist, but most of the examples are from consultations with three patients (and their care givers) at three different stages of advanced cancer care: early stages (text I), mid-way (text II) and final stage/symptom management only (text III). The oncologist’s discourse, as we will discuss in more detail throughout the chapter, suggests a philosophical stance that approaches the patient as a social human being. The oncologist perceives herself not as the absolute knower but as an informed professional who assumes and encourages partnership and facilitates a dialogic relationship with her patients. We chose this oncologist among fourteen oncologists in the corpus not only because she had patients at all three stages of the trajectory of illness but also because she seemed to be aware of all the clinical roles of the medical oncologist outlined by Cherny and Catane (2011) and interwove them seamlessly with the tasks which define her specialization. The next section presents a brief account of register and its variables (field, tenor and mode) as the theoretical tools that were applied in this case study and explains their relevance to the study of oncology consultations. This is followed by a more elaborated account of the three contextual variables and the analyses of the oncology consultations in terms of each variable in the next three sections. Finally, a summary of the results is presented in the last section.

Cancer Care as an Integrated Practice


2. Register Analysis Halliday, McIntosh and Strevens (1964) divide the situational features that realize text into the three categories of field, tenor and mode. Field refers to “the nature of the social action that is taking place”; tenor refers to the “the nature of the participants, their statuses and roles”; and mode refers to “what part the language is playing” (Halliday and Hasan, 1985/89, p. 12). A particular set of values for the field, tenor and mode determines a functional variety of language or a ‘register’, for example the ‘medical consultation’ register or, more delicately, the ‘oncology consultation’ register (although ‘register’ settings need not always generate recognizable existing practices—see Moore [2017]). While traditionally field, tenor and mode have been described intuitively and based on experience (Hasan, 2009), recent scholarly attempts have been made to model the parameters of context in a systemic fashion similar to the descriptions we see at the level of grammar and semantics (Hasan, 2014; Bowcher, 2014; Butt, 2003; Hasan, 2016 [2014]), the applications of which we see in media and medical communication research (Moore, 2015; Lukin, Moore, Herke, Wegener and Wu, 2011; Lukin, 2008; Scott, 2009; Wegener, 2011; Moore, 2004; Moore, 2016). Describing context using system networks enables the analyst to contrast the contextual possibilities of human interaction in delicacy and order (from most general to most delicate). Building and using context networks enhances order and precision in the study of context (Hasan, 2014), something that is much needed when looking at complex situations such as oncology at the end of life. In the perspective of the above theoretical explanations, an oncology consultation is an instance of the social practice of oncological care conducted by a specialist medical practitioner and in a broader sense an instance of the current cultural institutions of healthcare and palliative care. At the same time, the oncology consultation is the realization of meaning and grammar as text. Therefore, text and culture are related through the principles of realization and instantiation (for a more detailed account of the relation between texts, aspects of healthcare, and healthcare as an institution, refer to Matthiessen [2013]). Delineating the complex choices that the oncologist makes at the level of context as well as their textual realizations when consulting patients at different stages of their cancer journey can help us understand how the expert oncologist manages to actualize this multifaceted role, constructing a new culture of medical practice while at the same time being constructed by it.

3. The Field of Oncology Consultations Field of discourse has been paradigmatically modelled by Butt (2003) and Hasan (1999, 2009). In addition, Bowcher (2014) combines the


Neda Karimi et al.

FIELD networks of Hasan and Butt to propose a revised and unified network. The network used here is Hasan’s 2009 network of FIELD.2 Hasan’s networks, in Bowcher’s words, “are global in their application and are situated close to the system end of the cline of instantiation: context as system” (Bowcher, 2014, p. 185) and therefore seem suitable for the analysis of the context of oncology consultation, a context that has not been analysed paradigmatically to date. Hasan’s (2009) FIELD network consists of three primary m-systems (member systems) of VERBAL ACTION, SPHERE OF ACTION and PERFORMANCE OF ACTION to describe the contextual parameters of FIELD. Figure 16.1 shows the vectors and their primary systems in Hasan’s (2009) FIELD network. The system of VERBAL ACTION concerns the options [ancillary] and [constitutive] where [constitutive] itself selects [practical] or [conceptual]. A verbal action is practical if it is sensible and entails “the future occurrence of some material action” (Hasan, 1999, p. 283). It is conceptual if it

Figure 16.1 FIELD network (Hasan, 2009, p. 183)

Cancer Care as an Integrated Practice


is intellig-ible and “do[es] not call for any physical action” (Hasan, 1999, p. 288). The option of [practical] can be further divided into two options: [instruct] or [plan]. The option of [conceptual] opens to three simultaneous sets of option: [relation-based] or [reflection-based], [narrating] or [informing], and [first order] or [second order]. To illustrate, applying this m-system to our data showed that in the oncology consultations VERBAL ACTION is [constitutive] and that oncology consultation was not only about [practical] activities, but [conceptual] actions constituted a significant part of the oncological practice in every stage of the patient’s illness. The practical actions in the oncologist’s data not only involved planning/ instructing regarding the different treatments and tests (instructing regarding pain and other symptom/side-effect alleviation, and planning the next consultation) but also extended to incorporate instructions on using family and community-based services such as in-home care and psychological care, instruction on using special equipment for comfort and support at the end of life, instruction regarding access to palliative care, planning of non-medical activities such as vacations or family visits and coaching and navigating the patient across the healthcare system. Table 16.1 shows a list of the practical actions that we witnessed in the corpus along with instances of how these actions were realized textually by the oncologist. The conceptual actions included prognosis, reiteration of the goal of the care, medical and clinical reasoning and the promotion of evidencebased research, psychological counselling and casual talk. The data contains instances where the oncologist reminded the patient and the care giver of the untreatable nature of the cancer and the goal of the care; engaged in a pedagogic discourse to inform the patient and the care giver about the illness, treatment options, clinical research, the patient’s body and what to expect in the future and how to deal with it; provided psychological support and invited the patients to share their feelings and thoughts; and engaged with the patient and the companion in conversations about everyday matters related to the patient’s ‘lifeworld’ (Mishler, 1984) (Table 16.2). In all these instances, both practical and conceptual activities (including reflection-based conceptual activities), the oncologist seemed to be constructing a new clinician-patient relationship and activating and facilitating a new ‘patient’ role that is different from what the patients may have perceived before or at the beginning of their journey. Even in the activities that projected a future material action, there seemed to be a relationbased dimension. They conveyed an oncologist-patient relationship that is centred around the patient as a social human being and is based on empathy and support and continuity of care. This shows another layer of complexity in ‘doing oncology’. It also suggests that the system network of field can be further developed in this area to model this relation-based dimension of the practical and reflection-based conceptual actions.

Table 16.1 Actions in the corpus instantiating the Field parameter [practical] and their textual realizations Giving instruction to do/ Planning the treatment, blood test, blood transfusion, scan . . . Instructing regarding pain and other symptom/ side-effect alleviation

Planning the next consultation Instructing on using family and communitybased services such as in-home care and psychological care

Instructing on using special equipment for comfort and support Instructing regarding access to palliative care Planning non-medical activities such as vacations, family visits or daily activities

Gatekeeping (paperwork related to other aspects of the patient’s life being affected by the cancer) Coaching patients and navigating the health system

• Now you are on that tough every two week schedule. That was proven to be superior for paediatric and adolescent sarcoma and because we wanted to get control of this and because you were handling it fine we said we’d try it. We can go to every three weeks. • . . . and then you’ll give the Aloxi to the nurses. On the night of the chemo you can either have nothing or two Maxolon or half as often, or a whole as often if you’ve got nausea. The next morning you take two Dexamethasone and the small Emend . . . • Now I’m away that week so I’ll get XXXX to cover me but we might be on the week after anyway, in which case I’m here. • What about getting some regular planned extra help for a couple of hours in the afternoon so you know you’ll be able to get some rest. We’ve still got a long road ahead. You’ve got a little baby. • But this is an emotional disease. We’re going to have good news at times and not so good—so if you have a good rapport with someone— . . . And you’ve got to make that rapport now while you’re well. • Because of the back pain—the other thing is a sheepskin and you can buy those or rent those. • Have you seen the palliative care . . . Yes, I think it is [worthwhile] and, you know, it’s not terminal care it doesn’t mean that because you see them that, you know . . . • Look usually I say to people we know what’s going to happen next week or next month, we don’t know what’s going to happen in six months’ time—so go sooner rather than later. Now sometimes you stay well for a long time but you’re more sure of something in the shorter term than the longer term. So I think about Easter is a better time in terms of I would be more sure that you’ll be in a good condition. • Look I’m very happy to write that I think you need two bedrooms and you need to be in the area. • Oh do you want to go and tell the NUM, the Nurse Unit Manager, Lindsay, because you’re not the first person today and the system really needs to operate better than that . . .

Cancer Care as an Integrated Practice


Table 16.2 Actions in the corpus instantiating the Field parameter [conceptual] and their textual realizations Prognosis and reiteration of the goal of the care [conceptual: informing] Medical and clinical reasoning and the promotion of evidencebased research [conceptual: informing] Psychological counselling [conceptual: narrating] Casual talk [conceptual: narrating]

• Well remember what I said in the first place, unless we can cut it all out, which is a big, big ask, we can’t get rid of it . . . You need to be on continuing treatment and even with that at some point it stops working, okay. • We know that if—so they did do some studies saying what about if we stop and we watch you and if it starts to grow again we start again—and there’s a bit of split in oncologists about how they interpret those results. If you stop you need to be watched like a hawk and your average time before you start again is three or four months. • Are you happier where you are now? • Patient’s mother: I’m going next week. Oncologist: Are you? You’ve been evicted? Patient: No, she’s more than welcome to stay.

Moving to the second primary m-system of Hasan’s FIELD network, that of SPHERE OF ACTION, there are options of [specialized] or [quotidian] and [official (institutional)] or [private (individuated)]. Oncology consultation is a [specialized] and [institutional] social practice. The data showed that even in instances where the topics seemed to be quotidian topics, such as organizing a vacation or a family-related matter, the topic development and mapping between topics and turns was not what can be found in a casual conversational context or other quotidian situation. For example, in such a context ‘holiday’ correlates with ‘certainty’ and ‘health status’ in a way that it would not do in everyday chat. What explains this paradox is the oncologist’s patient-centred approach through which the commonplace and ordinary aspects of the patient’s life also become the business of care, producing what Candlin (1995) calls ‘comprehensive coherence’. We can test [institutional] versus [individuated] by considering Hasan’s (1996) notion of a context as ‘multiply coded’. Accordingly, institutionalized social processes would logically be multiply coded semiotically; that is to say the fact that they are institutionalized would be indicated by the fact that the different modes of meaning would single them out, and the boundaries set for a process by one mode would be commensurate with those set by another mode of meaning. (Hasan, 1996, p. 46)


Neda Karimi et al.

To test whether oncology consultation is an institutionalized activity, we can say that it is multiply coded through ways of behaving, the necessity of performing it in a specific setting and time and conducting it in a more or less defined way that the oncologists master throughout their long-term training, and is, therefore, an institutionalized practice. These options, however, seemed to vary in degree depending on the time or the patient’s stage of illness. If we consider them not as binary options but as two ends of a continuum, as Hasan (1996) suggests, then we observed from the data that as the patient gets closer to the end and when the treatment is stopped the linguistic behaviour changes in such a way that the selections in these systems seemed to move closer to the [quotidian] and [individuated] end. The last primary m-system in the FIELD network is PERFORMANCE OF ACTION, which covers the spatio-temporal nature of the activity. PERFORMANCE OF ACTION consists of two options, [bounded] and [continuing]. The first option refers to action that is carried on within one spatio-temporal location, and the second option refers to continuing action conducted across several different but connected events at different spatio-temporal locations. Going further in delicacy, [continuing] action can be either [sequenced] or [conditional]. The work of oncology consultation is a continuing activity, spanning from the day that a patient is diagnosed with cancer (or sometimes earlier) until the last consultation. This, according to Cherny and Catane (2011, p. 113), requires a varied range of responsibilities for the oncologist as the coordinator of cancer care in all of its phases: ‘the diagnostic: ambulatory or inpatient’, ‘curative primary therapy’, ‘ambulatory palliative therapy’, ‘sedentary palliative therapy— interactional’ and ‘sedentary palliative therapy—non-interactional’.

4. The Tenor of Oncology Consultations We adopted Hasan’s system network of INTERACTANT RELATIONS (Hasan, 2015, 2014, 2016 [2014]) to describe the tenor of oncology consultation (Figure 16.2). Hasan’s tenor network consists of the three simultaneous primary m-systems of AGENTIVE ROLES, TEXTUAL ROLES and SOCIAL ROLES. The AGENTIVE ROLES system relates to “the part an interactant is playing in the achievement of the goal implicit in the social practice” (Hasan, 2015, p. 34); the TEXTUAL ROLES are based on “the considerations around which their meaning-wording roles are calibrated” (Hasan, 2014, p. 33); and the SOCIAL ROLES refer to the interactants’ “interactive biography and their social positioning” (Hasan, 2014, p. 33). A central topic in the literature on end-of-life and palliative care is the topic of the goals of care and the scope of roles that oncologists take not only in achieving these goals but also in guiding the patient with advanced cancer to be a protagonist in this social practice and to

Cancer Care as an Integrated Practice


Figure 16.2 A fragment of INTERACTANT RELATIONS network (Hasan, 2014, p. 33; Hasan, 2013)

actively work towards reaching these goals, as we briefly discussed in the previous section. Tools such as Hasan’s tenor network, and its capacity for a close mapping of how such roles configure and vary, are therefore expected to be of use in the understanding of this aspect of ‘doing


Neda Karimi et al.

oncology’. In particular, we expect the INTERACTANT RELATIONS system network to tell us more about the relation-based dimension of the oncology consultation practice and its connection to a patient-centred care for advanced cancer patients. As was mentioned above, AGENTIVE ROLES refers to the roles the oncologist and the patient (and the companion) take in order to achieve the goals of the oncology practice. In the instances analysed here, patients appeared to have the responsibility for paying careful attention to and reporting symptoms, side-effects and quality of life information; conducting self-care; and participating in the process of decision-making in an informed and active manner. The oncologist, on the other hand, paired the patient’s self-reported information and other information that she elicited from the patient with biomedical results, i.e. blood tests and scans, and assessed the overall wellbeing of the patient. The data shows how the expert oncologist and the patient worked collaboratively to reach a decision about the patient’s care plan based on the information provided by the patient and the biomedical results, but also included the patient’s personal preferences. To pursue the long-term goal of keeping a balance between more treatment and quality of life, the oncologist’s role, however, was not confined to the assessment of the patient’s state and treatment/symptom management. Assisting the patient to navigate the healthcare system, giving advice to the patient regarding a vacation, and writing letters and reports and filling out forms in relation to other aspects of the patient’s life that were affected by the cancer, such as public housing and public home care service, were among the practical activities that the oncologist performed, as outlined above. Patient education through sharing of medical and clinical reasoning was another role the oncologist was engaged in. The oncologist helped the patient understand the nature of the process, explained the existing treatment options, talked about the most recent clinical research findings and their limitations, informed the patients and the care givers of the social and medical services available to them and taught them self-care in terms of pain and symptom management. She established a therapeutic relationship with her patients and provided care beyond what might be the expected order of being a cancer patient, including checking up on the patient’s care giver and family members. A diverse range of roles, and their textual realizations, have been documented through the analysis of tenor in this section, corresponding to the characteristics of oncology as a field seen in the previous section. Implicit to all these agentive roles is a facilitating role observed in the consultations we analysed, which exemplifies the relation-based dimension of ‘doing oncology’ referred to in the previous section. This facilitating role, which is a key aspect of the tenor of the oncology consultations observed, can arguably be explained as the oncologist’s ongoing work of teaching the patient what it means to be in the role of a cancer patient from the

Cancer Care as an Integrated Practice


philosophical point of view held by the oncologist in question, regarding the rights and responsibilities such patients have. That is, the data shows the oncologist not only facilitating greater patient agency by making the basis of her recommendations thoroughly explicit and inviting the patient to be involved in resolving the best way forward in terms of treatment/ non-treatment decisions, but also actively coaching the patient to be agentive. These features of the clinical dialogue allow us to infer a certain philosophical/ideological orientation of the physician. In the remainder of this section, using Hasan’s system network of interactant relations, we will elucidate the details of how the oncologist performs this facilitating role. Moving one step further in delicacy within the system networks, AGENTIVE ROLES consists of three simultaneous systems of GOALORIENTATION, COMMENCEMENT and ACTUAL OPENING. The GOAL-ORIENTATION system is about the awareness of the interactants of the goal of the social practice and their disposition to that goal. The network opens to a simultaneous set consisting of m-systems with the options [aware] or [unaware] and [one] or [both]. Before going through these options, a few words on the goals of the oncology consultations we analysed seem necessary. To do this, we make use of Butt’s (2003) network of GOAL-ORIENTATION (Figure 16.3). In the data we analysed, there were both immediate and longitudinal goals to the consultation. The immediate goals were variable depending on time and the patient’s condition and were overt most of the time. The longitudinal goal of the oncology practice for advanced cancer patients, or at least

Figure 16.3 Butt’s (2003) network of GOAL-ORIENTATION


Neda Karimi et al.

the oncologist’s goal, was to help the patient live longer with anticancer treatments as long as the side-effects from these treatments did not significantly reduce the quality of the time left. This required an ongoing process of calibrating quantity and quality of time, based on which the care plan was decided on or modified along the course of illness, as the corpus showed. This goal was a constant goal that over time was made overt by the oncologist if she found out that the patients or the care givers had different expectations. In fact, the reiteration of the goal of the care and the promotion of ‘open awareness’ of dying (Glaser and Strauss, 1965) formed one of the primary activities of the oncologist. While the oncologist was obviously aware of this goal, the patients and the care givers could at times misunderstand the aim of the treatment, as other researchers have also observed (Weeks et al., 2012), or simply had different goals that were not visible. As Hasan (2016 [2014], p. 447) puts it, “When goals are not ‘in’ the action, there is no linguistic means of knowing which of the invisible goals is being pursued” by the patients and care givers. The mismatch between goals in the oncology consultation practice can be due to the relation-based nature of this social activity, based on Hasan’s (1999, p. 234) argument that the interactants’ awareness of the goal of the social activity is at its lowest when that activity is ‘relation based’. In fact, the oncologist-patient role relationship was not simple. There were various dimensions to being a patient or an oncologist, as we will see in the coming paragraphs. The next m-systems of AGENTIVE ROLES are COMMENCEMENT and ACTUAL OPENING. Commencement is related to the control over how the social activity commences. When commencement is not free, it can be governed internally (control is built into conventions of social practice) or externally (control is external, i.e., some feature in material situational setting) (Hasan, 2013). Information about which interactant, the initiator or the respondent, actually opens the social activity is specified by ACTUAL OPENING. In most of the consultations we observed, the oncologist was both the initiator and the actual opener of the consultation. In text III the patient was the actual opener of the consultation, where she initiated a greeting move to ask about the oncologist’s vacation. This is perhaps because the SPHERE OF ACTION becomes more individualized as the patient gets closer to the end of life; possibly it construes the decreasing social distance that comes with many repeat consultations over time and the patient’s familiarity with the setting and ‘the rules of engagement’ in that setting (see below); either way the patient has more control over the commencement of the consultation. In general, the oncology consultation sessions we analysed began with a run down on side-effects and symptoms initiated by the oncologist, which seems to show the oncologist role as centred in anti-tumour therapy and symptom control, as Cherny and Catane outline (see previous section). One point worth considering is the influence of the broader culture of medical

Cancer Care as an Integrated Practice


practice within which the oncologist has been trained and has become acculturated on how the consultations commence. In this light, oncology consultation is an instance of healthcare as a cultural institution. It is the healthcare system within which the clinicians have been trained and are working that governs the actual opening of the consultation, and the oncologist plays the role of the ‘mediator’ from this perspective. In terms of the contextual dimensions of TEXTUAL ROLES in INTERACTANT RELATIONS an account of attitude as a psychological notion seems useful. Attitude is defined as “a psychological tendency that is expressed by evaluating a particular entity with some degree of favor or disfavor” (Eagly and Chaiken, 1993, p. 1). Although attitude is mostly communicated through non-verbal resources (Mehrabian, 1972), our analysis of the attitude of the oncologist was only based on the written transcripts as we did not have access to the audio records of the transcripts and there was no video-recording in this project. Accordingly, our analysis showed that the oncologist held a [biased: positive] attitude towards the patient, and the focus was on [addressee]. One manifestation of the oncologist’s positive attitude is her attempt at the construction of a joint project through using first person plural pronouns instead of separating herself from the patient. Of course, the institutional ‘we’ can also be a controlling, patronizing mechanism (Pennycook, 1994; Kinsman et al., 2010). What makes us interpret it as positive attitude is the bigger picture of the oncologist’s commitment to the endorsement and reinforcement of shared decision-making, as we explain in this chapter. Another manifestation of the oncologist’s positive attitude and her focus on her addressee was her respect for the topics that were initiated by the patient and her willingness and interest to get involved in the details of those patient-initiated moves despite the time pressure (note that the average consultation time length for the oncologist under study is not different from the average time across the whole data set, which is around twenty-one minutes). An example is the oncologist’s engagement in a nineteen-turn conversation about the patient’s visit to Dubbo in text III that seems to have no bearing on the treatment and symptom control business of the consultation. We suggest that showing a positive attitude (in this sense) is one contextual aspect of the facilitating role of the oncologist. Another point that needs to be mentioned here before we move on to the last m-system is the positive attitude towards research and evidence-based practice that we witnessed in the texts, especially when discussing treatment options and their effects, which is also stressed by Cherny and Catane (2011). The last primary m-system in INTERACTANT RELATIONS is the system of SOCIAL ROLES, which consists of two simultaneous sets of SOCIAL DISTANCE and STATUS. As a result of the oncologist’s knowledge and expertise and her semiotically managed influence in the social practice of consultation, the oncologist appeared to be in a higher


Neda Karimi et al.

position than the patients or care givers. This hierarchized relationship, though, did not seem to have an effect on the quality of the social distance between them. This can be partly due to the long history that the interactants shared together as oncologist and patient; the patients seemed to be seeing the oncologist for a long time. However, it seemed independently to be the case that the oncologist consistently chose ways of speaking that construed the patient as an active human being and as a ‘meaner’, an approach that realizes what Bernstein calls ‘weak insulation’ between the oncologist and the patient and the care giver. What we inferred from the oncologist’s practice of ‘doing oncology’ could arguably be described as a ‘socio-semiotic’ philosophy of care. We dub it a ‘socio-semiotic’ philosophy of care, whereas many of the contextual features of oncology consultation that we have described here have been associated with cultural movements or moments such as postmodernism and consumerism, in both celebratory and critical modes (see, e.g., Morris, 2000; Henderson and Petersen, 2002; Yeatman, Dowsett, Fine and Gursansky, 2009), or the medical ‘gaze’ of modernism (Foucault, 1973), as displayed in the evidence-based medicine movement and its commitment to innovation, progress and authority. One of the advantages of Hasan’s contextual parameters approach is that it allows us to capture configurations of contextual features without having to declare their provenance, which can then be explored separately. This then arguably helps show how different philosophies and zeitgeists may co-occur and vary within a coherent community of practice such as oncology. This socio-semiotic philosophy of care is transformed into the oncologist’s discourse through what the sociologist Basil Bernstein terms weak ‘insulation’ between the social categories. This weak degree of insulation between the expert oncologist and the patient and his/her care giver creates a different division of labour and generates weak degrees of classification in their interpersonal relations. Classification refers to the degree of boundary maintenance between the social categories: agents (here the oncologist, patients and care givers) and discourses or contexts. Classification is affected by the degree of insulation between the categories. A strong insulation between categories creates a principle of strong classification, and a weak insulation between categories gives rise to a principle of weak classification (Bernstein, 1990). In our data the social division of labour, as our earlier discussions in the chapter suggest, was composed of several categories: ‘advisor’, ‘therapist’ and ‘transmitters’ on the part of oncologists, and ‘advisee’, ‘therapee’, ‘acquirer’ and ‘reviewer’ on the part of patients and care givers. In this configuration the oncologist acted on the principle that her roles permeate each other and constantly reclassified her roles. Not only did she provide medical advice to the patient, but she also felt the need to convince the patient (and the companion) through reasoning and providing educational information on why such advice is suggested. Arguably this does not detract from patient agency

Cancer Care as an Integrated Practice


but in fact enables the patient to be more agentive (cf. Moore, 2004, 2017). In addition, offering advice and education to the patient was not confined to the area of biomedicine, but extended to other areas, including the healthcare system, where the oncologist navigated the patient across the healthcare system, as well as the patient’s personal life, where the oncologist attended to the life of the patient as affected by the illness. The permeable nature of the oncologist’s roles suggests that weak classification is another aspect of this facilitating role of the oncologist. Consistent with this argument, our transitivity analysis of over 5,000 clauses suggests that the patients did not construe themselves merely as the site of the disease and the goal of the treatment but as conscious and active participants who were semiotically agentive, as they used more mental processes and fewer relational semiotic processes to convey inner world experiences, and who exerted ‘agency despite cancer’ through identifying themselves as the Actors of material actions related to the realm of the mundane activities of a social human being or material actions targeted towards palliating symptoms and side-effects. Now, if we consider a discourse spectrum that ranges from a completely specialized medical voice to a completely non-medical lay voice, our analysis suggests that, since the classification seems weaker, the discourse of both oncologist and patient was not polarized at each end of the spectrum but is close to a discourse that is a mix of specialized and non-specialized voices for both the oncologist and the patient. An instance of weak classification in oncology consultation is the occurrence of a ‘pedagogic’ move that is usually followed by treatment advice whereby the oncologist explains the rationale behind such advice. Another instance is the doctor’s attention to the daily life of the patient and the effect of cancer on that. Another aspect of social distance between the interactants is the concept of framing. Framing is defined as “the principle regulating the communicative practices of the social relations within the reproduction of discursive resources” (Bernstein, 1990, p. 36). Accordingly, when framing is strong, the oncologist explicitly regulates the distinguishing features of talk and the kind of roles she takes during the consultation. By contrast, when framing is weak, this control is not unilateral; rather, the shift between the different roles is controlled by the oncologist and the patient in partnership. Our analysis shows that the oncologist does not exercise strong control over the unfolding of the discourse, and her framing is weak. Even in those consultations in which the patient is receiving active anticancer treatment, where her role as someone who offers advice and planning regarding treatment spans a large part of the interaction, the classification between interactants was weak, and the control over this weak classification was distributed between the oncologist, the patient and the companion: the weaving in and out of the different roles was jointly construed. Of course, this does not imply that the oncologist does not have a fixed agenda,


Neda Karimi et al.

nor does it imply an equal status for the oncologist and the patient, as mentioned earlier. The situation can be described as what Hasan refers to as “showing respect for the other’s individuality” (2001, p. 65). An example of the oncologist’s weak framing is her readiness to entertain a shift in the activity initiated by the patient or the companion. Another example is her disposition to provide detailed and elaborated answers to the questions raised by the patient or the companion, as well as her endorsement of question asking. At the beginning of this section we suggested a facilitating role on the part of the oncologist and argued that it is a key aspect of the tenor of the analysed consultations. Throughout this section, we have explained the different dimensions of this facilitating role. Table 16.3 summarizes these dimensions and presents an example for each aspect.

5. The Mode of Oncology Consultations Let us finish the discussion of the contextual parameters of oncology consultation with a review of mode as partially modelled by Hasan (2014, 2016 [2014]) (Figure 16.4). MODE in Hasan’s words “is all together concerned with contact” (2014, p. 28). The two dimensions of MODE or the two primary m-systems are MATERIAL CONTACT and SEMANTIC CONTACT. Hasan’s paradigmatic account of MODE is in her own words “a tentative fragment of MODE” (Hasan, 2014, p. 28) and needs further exploration. However, it is still the most comprehensive network as it brings into account the parameter of SEMANTIC CONTACT. MATERIAL CONTACT “is an attempt to describe the possible means of expression that normal human beings can use in producing and receiving meaning-wording in many parts of the world today” (Hasan, 2016 [2014], p. 429). The material contact of oncology consultation is then phonic at the PRODUCTION POINT and direct at the RECEPTION POINT. SEMANTIC CONTACT “concern[s] the description of intelligible aspects of language, which would primarily facilitate access to those meanings whose choice activates continuity, coherence and textual organisation” (Hasan, 2016 [2014], p. 429). It consists of four simultaneous systems: ELOCUTION, TURN MANAGEMENT, READINESS and RELEVANCE. Of these systems we will just look at TURN MANAGEMENT and READINESS only at the point of origin level as the networks currently stop here. Our analysis of the number of turns suggests turn-taking is roughly equal between the oncologist and the patient. However, if we look at the length of turns, this equality no longer holds true. While in text III the turn length distribution between the oncologist and the patient is nearly equal, in texts I and II the oncologist’s share in terms of clause count is higher than the patients’ and the care givers’ overall share (67%

Facilitating role

Weak framing


• The construction of a joint project But what that will do over time is start giving us some numbness in your fingers and toes and that is a cumulative thing. • Respect for patient-initiated topics • Promotion of evidence-based practice and research • Pedagogical moves and clinical reasoning Yes, that’s right. So the Oxynorms are great for extra pain relief but if you need lots of them and regularly that means you need to up the—so the idea is that you’re only on regular medication and the occasional one otherwise you’re taking tablets all the time. So I would go up to Oxynorm—I would go up for the Oxycontins to 30’s which you can do by two 15’s if you’ve got those leftovers. • Attention to the life of the patient What about getting some regular planned extra help for a couple of hours in the afternoon so you know you’ll be able to get some rest? We’ve still got a long road ahead. You’ve got a little baby. • Readiness to entertain a shift in the activity initiated by the patient or the companion (During the move in which the oncologist instructs the patient to take pain relief medication) Patient: Yes, okay, and the other thing is I wanted to sort of, I’m going to Dubbo for a week on Monday, so I didn’t want to sort of start the patch and get sort of, because I’ve got about a 5½ to 6 hour train trip, so I thought I might wait until I come back. Oncologist: Are you going up with someone? Patient: No, no. Oncologist: What are you going to do up there, go to the zoo?

Positive attitude

Weak classification

As discussed in Tables 16.1 and 16.2

Agentive roles

Table 16.3 Aspects of the facilitating role of the oncologist

Facilitating role

Agentive roles

Table 16.3 (Continued)

Patient: I’ve got a friend up there—no, I am going to the zoo, yes. Oncologist: Yeah, good, fantastic. Patient: One of my students is up there and we’ve kept in touch over the—yes. Oncologist: Okay, terrific, oh it’s a fantastic zoo. Patient: Yes, yes, so I am, so. Oncologist: Alright, good, excellent. I did say it as a joke, but that’s good, I’m happy, yeah. • Giving detailed and elaborated answers to the questions Patient: [Do I reduce Dex] Down to one and then down to a half? Oncologist: Or down to one and a half, and then one. It doesn’t matter. Cutting down on the Dex too quickly will make you feel lousy. Stopping it altogether quickly is bad. Your body doesn’t have it but there’s no right answer. As long as you go slowly you can titrate it to how you feel. So you can go to one and a half for two days, then one for two days, then half. • Endorsing question asking Have you got any questions for me at this stage?

As discussed in Tables 16.1 and 16.2

Cancer Care as an Integrated Practice


Figure 16.4 A fragment of the MODE network (Hasan, 2014, p. 27)

in II and 76% in I). That is, although the oncologist had fewer turns in consultations I and II, her turns were considerably longer. Based on the information provided in Table 16.4, we can form the hypothesis that there is a trend in the length of the oncologist’s turns depending on where the patient is in the trajectory of illness. Accordingly, as the patient gets closer to the end of life, the semantic work on the part of the oncologist decreases, and the patient contributes more to the interaction. As we saw in the section on field, as this change in mode occurred, the sphere of


Neda Karimi et al.

Table 16.4 Interactants’ share of discourse in texts I, II and III Text


Text III Patient Oncologist Text II Patient, Partner, Mum Oncologist Text I Patient, Wife Oncologist

Turn count Word count Clause count [% participation] [% participation] [% participation] 137 [50%] 137 [50%] 109 [56%]

1,939 [51%] 1,827 [49%] 1,324 [32%]

338 [51%] 330 [49%] 240 [33%]

85 [44%] 133 [54%] 113 [46%]

2,757 [68%] 708 [19%] 3,109 [81%]

484 [67%] 146 [24%] 459 [76%]

action in field moved towards being more [quotidian] and [individuated], and the role relations became more of therapist-therapee. This reciprocal relation, which was also witnessed between the field and the tenor, supports Hasan’s argument that the configuration that results from the choice of symbolic mode, social process, and social relation is not a simple combination; its meaning is not additive, not just the sum of the meanings of the three; rather, contextual configuration is like a chemical solution, where each factor affects the meaning of the others. (Hasan, 1995, p. 231) Moving to the m-system of READINESS, Hasan (2014) does not define this notion, but if we assume that READINESS is related to the participants’ acquaintance with the function of the meaning/wording that is used in a specific social practice, then both the oncologist and the patients generally seem to be ready for what they receive. An example is the familiarity of the patients with their various roles that they had mastered throughout their cancer journey, arguably as a result of the enabling role of the oncologist. The oncologist also seemed to be ready for the topics that the patient raised and never failed to address them comprehensively, which shows her ability to juggle all these different roles that we tried to outline and that Cherny and Catane (2011) suggest.

6. Conclusion Our application of the contextual system networks in the study of palliative oncology consultations suggests that oncology consultation practice as a ‘contextual concern’ (Matthiessen, 2013, p. 444) can be examined in detail through the analysis of recurrent patterns in a number of oncology consultation texts which are the textual realization of instances of oncology consultation practice for advanced cancer patients. It revealed the different aspects of what today is considered a good oncology practice for advanced

Cancer Care as an Integrated Practice


cancer patients along with their textual realizations. Using paradigmatic models of the contextual parameters of field, tenor and mode allowed us to analyse these different aspects not as detached and self-enclosed elements but as a gestalt with all the interrelations and interdependencies discussed above. The medical oncologist’s role in advanced cancer care, as we tried to show in this chapter, is multifaceted and complex and is the function of choices from various interwoven contextual elements and their textual realizations. Our contextual analysis arguably demonstrates the actualization of the roles that Cherny and Catane (2011) outline for the medical oncologist in the practice of the expert oncologist under study. In addition, we found two further roles, namely ‘facilitator’ and ‘mediator’, in her practice.

Acknowledgements We thank our colleagues Professor Phyllis Butow and Dr Adam Walczak from the Centre for Medical Psychology and Evidence-Based DecisionMaking (CeMPED), who provided the data used in this project and commented on an earlier draft. This chapter was written while the first author, Neda Karimi, was receiving an Australian Postgraduate Award (APA) scholarship for PhD research.

Notes 1. The trial protocol and all study forms and materials have received the approval of the Lead Human Ethics Review Board at Royal Prince Alfred Hospital (Protocol Number: X10-0032 and Approval Number HREC/10/RPAH/51) and from the governance officers at each of the participating recruitment sites. 2. ‘Point of origin’, which refers to the system network’s object of enquiry, is always shown in upper case, the m-systems are written in lower case (except for the point of origin m-systems which are shown in upper case), and the ‘options’ are displayed in square brackets. For some further information on system networks, see Hasan (2014). 3. The sensible/intelligible distinction is after Russell (e.g. 2004); the hyphenation is Hasan’s.

References Balint, M., Ball, D. H. and Hare, M. L. (1969) Training medical students in patient-centered medicine. Comprehensive Psychiatry, 10, pp. 249–258. Bernstein, B. (1990) Class, Codes and Control. London and New York: Routledge. Bowcher, W. L. (2014) Issues in developing unified systems for contextual field and mode. Functions of Language, 21(2), pp. 176–209. Butt, D. G. (2003) Parameters of Context: On Establishing the Similarities and Differences Between Social Processes. Unpublished monograph, Macquarie University. Candlin, S. (1995) Towards Excellence in Nursing: An Analysis of the Discourse of Nurse and Patients in the Context of Health Assessments. PhD thesis, University of Lancaster.


Neda Karimi et al.

Cherny, N. I. and Catane, R. (2011) Palliative medicine and modern cancer care. In G. Hanks, N. I. Cherny, N. Christakis, M. Fallon, S. Kaasa and R. K. Portenoy (Eds.), Oxford Textbook of Palliative Medicine, 4th ed. Oxford: Oxford University Press, pp. 111–124. Cherny, N. I., Catane, R. and Kosmidis, P. (2003) ESMO takes a stand on supportive and palliative care. Annals of Oncology, 14, pp. 1335–1337. Eagly, A. H. and Chaiken, S. (1993) The Psychology of Attitudes. Fort Worth, TX: Harcourt Brace Jovanovich College Publishers. Foucault, M. (1973) The Birth of the Clinic: An Archaeology of the Human Sciences. New York: Routledge. Glaser, B. G. and Strauss, A. L. (1965) Awareness of Dying. NJ: Aldine Transaction. Halliday, M. A. K. and Hasan, R. (1985/89) Language, Context and Text: Aspects of Language in a Social-Semiotic Perspective. Victoria: Deakin University Press. Halliday, M. A. K., Mcintosh, A. and Strevens, P. (1964) The Linguistic Sciences and Language Teaching. London: Longmans. Hasan, R. (1995) The conception of context in text. In P. H. Fries and M. Gregory (Eds.), Discourse in Society: Systemic Functional Perspectives. Norwood, NJ: Ablex Publishing Corporation, pp. 183–284. Hasan, R. (1996) What’s going on: A dynamic view of context in language. In C. Cloran, D. G. Butt and G. Williams (Eds.), Ways of Saying: Ways of Meaning. London: Cassell, pp. 37–50. Hasan, R. (1999) Speaking with reference to context. In M. Ghadessy (Ed.), Text and Context in Functional Linguistics. Amsterdam and Philadelphia: John Benjamins Publishing Company, pp. 219–329. Hasan, R. (2001) The ontogenesis of decontextualised language: Some achievements of classification and framing. In A. Morais, I. Neves, B. Davies and H. Daniels (Eds.), Towards a Sociology of Pedagogy: The Contribution of Basil Bernstein to Research. New York: Peter Lang Publishing, pp. 47–79. Hasan, R. (2009) The place of context in a systemic functional model. In M. A. K. Halliday and J. J. Webster (Eds.), Continuum Companion to Systemic Functional Linguistics. London: Continuum, pp. 166–189. Hasan, R. (2013) Tenor: Three primary Systems in Interactant Relations. Unpublished work. Hasan, R. (2014) Towards a paradigmatic description of context: Systems, metafunctions, and semantics. Functional Linguistics, 1(9), pp. 1–54. Hasan, R. (2015) Tenor: Rethinking Interactant Relations. Unpublished work. Hasan, R. (2016 [2014]) Towards a paradigmatic description of context: Systems, metafunctions, and semantics. In J. Webster (Ed.), Context in the System and Process of Language: The Collected Works of Ruqaiya Hasan, vol. 4. UK and USA: Equinox, pp. 389–469. Henderson, S. and Petersen, A. (2002) Consuming Medicine: The Commodification of Health Care. London: Routledge. Kinsman, H., Roter, D., Berkenblit, G., Saha, S., Korthuis, P. T., Wilson, I., Eggly, S., Sankar, A., Sharp, V., Cohn, J., Moore, R. and Beach, M. (2010) ‘We’ll do this together’: The role of the first person plural in fostering partnership in patientphysician relationships. Journal of General Internal Medicine, 25, pp. 186–193. Lukin, A. (2008) Journalistic voice, register and contextual configuration: A case study from the Spanish and Argentinian press. In E. Thomson and P. R. R. White (Eds.), Communicating Conflict: Multilingual Case Studies of the News Media. London and New York: Continuum, pp. 143–172.

Cancer Care as an Integrated Practice


Lukin, A., Moore, A. R., Herke, M., Wegener, R. and Wu, C. (2011) Halliday’s model of register revisited and explored. Linguistics and the Human Sciences, 4, pp. 187–213. Matthiessen, C. M. I. M. (2013) Applying systemic functional linguistics in healthcare contexts. Text & Talk, 33(4–5), pp. 437–467. Mehrabian, A. (1972) Nonverbal Communication. New Brunswick and London: Aldine Transaction. Mishler, E. G. (1984) The Discourse of Medicine: Dialectics of Medical Interviews. Norwood, NJ: Ablex Publishing Corporation. Moore, A. R. (2004) The Discursive Construction of Treatment Decisions in the Management of HIV Disease. PhD thesis, Macquarie University. Moore, A. R. (2015) Can semantic networks capture intra-and inter-registerial variation? Palliative care discourse interrogates Hasan’s message semantics. In W. L. Bowcher and J. Y. Liang (Eds.), Society in Language, Language in Society: Essays in Honour of Ruqaiya Hasan. UK: Palgrave Macmillan, pp. 83–114. Moore, A. R. (2016) Surgical teams in action: A contextually sensitive approach to modelling body alignment and interpersonal engagement. In A. Baldry and E. Montagna (Eds.), Interdisciplinary Perspectives on Multimodality: Theory and Practice. Campobasso, Italy: Palladino. Moore, A. R. (2017) Register analysis and message semantics. In T. Bartlett and G. O’Grady (Eds.), Routledge Handbook of Systemic Functional Linguistics. London: Routledge, pp. 418–437. Morris, D. B. (2000) How to speak postmodern: Medicine, illness, and cultural change. Hastings Center Report, 30, pp. 7–16. Pennycook, A. (1994) The politics of pronouns. ELT Journal, 48, pp. 173–178. Rogers, C. R. (1942) Counseling and Psychotherapy: Newer Concepts in Practice. Boston: Houghton Mifflin Company. Russell, B. (2004) The Analysis of Mind. New York: Cosimo Classics. Saunders, C. M. (1967) The care of the terminal stages of cancer. Annals of the Royal College of Surgeons of England, 41, pp. 162–169. Scott, C. E. (2009) Reporting Armistice: A Diachronic, Functional Perspective. PhD thesis, Macquarie University. Surbone, A., Zwitter, M., Rajer, M. and Stiefel, R. (2012) New Challenges in Communication With Cancer Patients. New York: Springer. Walczak, A., Butow, P. N., Clayton, J. M., Tattersall, M. H., Davidson, P. M., Young, J. and Epstein, R. M. (2014) Discussing prognosis and end-of-life care in the final year of life: A randomised controlled trial of a nurse-led communication support programme for patients and caregivers. BMJ Open, 4, e005745. Weeks, J. C., Catalano, P. J., Cronin, A., Finkelman, M. D., Mack, J. W., Keating, N. L. and Schrag, D. (2012) Patients’ expectations about effects of chemotherapy for advanced cancer. New England Journal of Medicine, 367, pp. 1616–1625. Wegener, R. (2011) Parameters of Context: From Theory to Model and Application. PhD thesis, Macquarie University. Yeatman, A., Dowsett, G., Fine, M. and Gursansky, D. (2009) Individualization and the Delivery of Welfare Services. London: Palgrave.

17 The Process of Perception in the Early Academic Article David Banks

Preliminary Remarks The first ever academic periodical appeared on 5 January 1665 in Paris. It was the Journal des Sçavans. It was followed two months later on 6 March by the Philosophical Transactions, which appeared in London. Both of these journals still exist, thus providing an extensive corpus of academic texts over a period of three and a half centuries (Kronick, 1962, 1991). The Journal des Sçavans was edited by Denis de Sallo, who founded the journal at the instigation of Colbert, Louis XIV’s first minister. The Philosophical Transactions was the brainchild of Henry Oldenburg, one of the secretaries of the Royal Society. The recent history of France had led to that country becoming powerful and stable. In 1643 Louis XIV had come to the throne at the age of five. There was a period of regency which lasted until the death of Mazarin, who had been coregent with the Queen, in 1661. Louis then took over personal control of the affairs of state, and chose as his first minister Colbert. The political system was that of absolute monarchy, and Colbert did everything possible to enhance the glory of the ‘Sun King’, Louis XIV. In order to do this he felt it necessary to control everything; indeed, he was something of a control freak! But under his control France became the economic and cultural centre of Europe (Gignoux, 1941). The situation across the Channel in England was quite the reverse. In 1605 the Gunpowder Plot failed, but was symptomatic of the divisions which would tear the country apart. In 1625 Charles I came to the throne, but his reign was marred by constant disputes with Parliament, mainly over questions of finance and religion. This came to a head in 1642 with the outbreak of civil war, which lasted seven years until the capture and execution of Charles I by beheading in 1649. This was followed by a period of republican government under Cromwell, but his regime was tyrannical and puritanic. Cromwell died in 1658, but his son, who had been supposed to take over, resigned after a very short period, thus leaving the way open for the Restoration of the monarchy. This took place in 1660 with the return of Charles II. Thus, in the 1660s England had gone

The Process of Perception


through a half century of total chaos. In fact, although there was great hope for the future, the chaos did not end; the disputes with Parliament continued throughout the reign of Charles II and into the abortive reign of James II, until, in 1688, in the Glorious (bloodless) Revolution, William and Mary were brought in as co-monarchs, while James II went into exile (Kishlansky, 1996; Hill, 1969 [1961]; Clark, 1956). In getting de Sallo to create a periodical, control was Colbert’s objective, and that included controlling new knowledge. It was thought that if people had revolutionary ideas in the arts and sciences, perhaps they would get revolutionary ideas in politics too! The new knowledge of the period was to be found mainly in books, so de Sallo’s journal turned out to consist basically of book reviews; and since, if control is one’s purpose, there is no point in having exclusions, the journal covered all disciplines, including theology, law and history (Morgan, 1928). Most of the members of the Royal Society were virtuosi, or gentlemen scientists of private means. This, however, was not the case for Oldenburg, who had to earn his living. Since he had become the centre of a network of scientific correspondence, he conceived the idea of using his position to create extra income. His idea was to use his correspondence to make a bulletin of scientific news which he could then sell, to help make ends meet. Hence, his bulletin, the Philosophical Transactions, was based on the letters and other items he received in the post (Gotti, 2006, 2011). And since his potential readership was the members of the Royal Society and like-minded people, whose main interest was science and technology, or ‘natural philosophy’ as it was then called, it was fairly natural that the contents of the Philosophical Transactions should be restricted to that area (Bluhm, 1960; Hall, 2002).

Corpus and Sub-Corpus In my recent work (Banks, 2017), I have used a corpus of articles from the Journal des Sçavans and the Philosophical Transactions representing the period 1665–1700. The articles are taken from the issues of 1665, 1675 and 1685, for both journals in each case, and 1695 for the Journal des Sçavans and 1694 for the Philosophical Transactions. This corpus has approximately 77,000 words for the Philosophical Transactions component, and about 66,000 for the Journal des Sçavans. The year 1700 provides a useful cut-off point for our period of study since the Académie Royale des Sciences, which had been founded in 1666, did not publish its findings in a widely distributed publication. Its publications were luxurious limited editions which were considered the personal property of the King, and Louis XIV used them as gifts for illustrious visitors. Hence, they were not widely available. It was only in 1699 that the Académie made the decision to start publishing in a more


David Banks

open fashion, and the first publication, containing the papers for the year 1699, actually appeared in 1702 (Licoppe, 1994, 1996). For the purposes of my analyses, I use a system of process types, based on those usually used in Systemic Functional Linguistics (Halliday and Matthiessesn, 2014). My variant of this system comprises five types: material (physical actions and events), mental (events of a cerebral nature), relational (relationships between two entities or between an entity and one of its characteristics), verbal (acts of communication) and existential (statements of existence) (Banks, 2005), where mental process can be subdivided into cognitive, affective and perception processes. This is shown is Figure 17.1. It will be noted that this excludes the category of behavioural process used by many within the systemic linguistics field. I feel that this category is hybrid and to a certain degree incoherent, and so cannot be considered to be a process type on par with the five major types mentioned in Figure 17.1. I think the phenomena that behavioural process is intended to deal with can be better treated at a more delicate level, with, for example, the distinction between processes of communication which can or cannot project, being treated as a subdivision of verbal process into projecting and non-projecting types. Similarly, the distinction between processes of perception which are voluntary and those which are not can be dealt with as a subdivision of perception process into voluntary and involuntary types (Banks, 2016). O’Donnell, Zappavigna and Whitelaw (2008) have shown that within Systemic Functional Linguistics, there is a wide range of views on the analysis of process types, from a semantic, conceptual view at one end of the cline, to a structural grammatical view at the other end. My position is very much towards the conceptual end of that scale. In this study1 I shall be specifically considering processes of perception. This includes all forms of perception by the eyes or ears, and since, as I said above, my conception of perception process includes both voluntary and involuntary processes, perception process in this study includes all forms of seeing and looking, hearing and listening.

Figure 17.1 Types of process

The Process of Perception


Of the finite verbs in the Journal des Sçavans component of the corpus, 15% encode mental processes, of which 9%, 1.4% of the total, are perception processes. In the Philosophical Transactions component, 19% of the finite verbs encode mental processes, of which 29%, or 5.5% of the total, are processes of perception. Hence, it seems evident that perception processes are more frequent in the Philosophical Transactions than in the Journal des Sçavans. In order to look at this in greater detail I have taken a subsample, based on the issues for the year 1675. This subsample contains five issues of the Journal des Sçavans, with approximately 14,000 words, and three issues of the Philosophical Transactions, with about 23,000 words. The disciplines represented in the two journals are shown in Table 17.1. It can be seen that of the twenty-seven items that appear in the Journal des Sçavans sample, the commonest discipline to appear is theology, with five items, followed by history, with three. Of the twenty-seven items, no more than ten could be classed as scientific. The Philosophical Transactions sample contains nineteen items, the commonest discipline being agriculture, with five items, followed by astronomy, with four. Of the nineteen items, at least sixteen could be regarded as scientific.

Perception Processes in the Journal des Sçavans Fifteen cases of finite verbs encoding perception process were found in the Journal des Sçavans sample. The numbers found in individual disciplines are shown in Table 17.2.

Table 17.1 Disciplines represented (1675) Journal des Sçavans

Philosophical Transactions


No. of articles


No. of articles

Theology History Astronomy Biography Geography Classics Mathematics Medicine Technology Agriculture Bibliography Chemistry Editorial Literature

5 3 2 2 2 2 2 2 2 1 1 1 1 1

Agriculture Astronomy Physics Geography Geology Economy General Mathematics

5 4 3 2 2 1 1 1


David Banks Table 17.2 Perception processes by discipline (Journal des Sçavans) Astronomy Biography Chemistry Geography History Mathematics Technology Theology

3 3 2 2 2 1 1 1

The disciplines with the highest number are astronomy and biography, with three examples each. The verb concerned is usually voir, which accounts for eleven of the examples, with one example each of apercevoir, faire l’observation, regarder and faire regarder. (1) . . . & ce sont ces nuages que nous voyons quelquefois estre agitez par les vens, & qui nous cachent bien souvent tout le Ciel. [astronomy]2 ‘. . . and these are the clouds that we sometimes see being shaken by the wind, and which frequently hide all of the sky from us.’3 (2) De Chancelier on le voit devenir Archevesque de Cantorbery & Primat de l’Angleterre . . . [biography] ‘From Chancellor we see him become Archbishop of Canterbury and Primate of England.’ Example (1) is from the field of astronomy and example (2) from the field of biography. However, it may be felt that example (2) is, at best, a marginal example of a perception process. It is a moot point as to whether seeing that a historical figure had become Archbishop of Canterbury can be assimilated to processes of perception (rather than, say, cognition). Nevertheless, these marginal examples have been included for the sake of argument, since there are relatively few in the Journal des Sçavans sample, and excluding them would simply reinforce the argument I shall be making. (3) Et ce fut alors que sur la surface de ce sel on vit s’élever comme des fougeres en grand nombre. [chemistry] [And it was then that a great quantity of something like fern was seen to arise on the surface of this salt.] (4) On y voit souvent deux Soleils avec trois Arc en Ciel qui passent entre les deux images du Soleil, & le Soleil veritable. [geography] [You sometimes see two suns with three rainbows which pass between the two images of the sun and the real sun.]

The Process of Perception


(5) . . . qui n’a pas seulement donné de nos jours de quoy le faire, mais qui a encore laissé un revenu considerable pour en faire les reparations, & pour y entretenir une fameuse Imprimerie ; doù nous voyons sortir tous les jours de grands Ouvrages. [history] [. . . who has not only given us the means of doing it today, but moreover has left a considerable income for repairs and the maintenance of a famous printing house; from it we see great books emerging every day.] Examples (3), (4) and (5) come from the fields of chemistry, geography and history respectively. Once again, example (5) from the field of history might be considered marginal. To what extent is seeing that a printer has produced great books a question of perception? The borderline cases which occur are, as here, in the humanities, rather than the sciences.

Perception Processes in the Philosophical Transactions If perception processes occurred with the same frequency in the Philosophical Transactions as they do in the Journal des Sçavans, then, given that there are fifteen examples in the Journal des Sçavans for 14,000 words, and the Philosophical Transactions sample has 23,000 words, we would expect to find about twenty-five. In fact, there are eighty-one. The distribution of these by discipline is given in Table 17.3. The field with the highest number of examples is physics, with twentysix, followed by geology, with twenty-three, and geography, with fourteen. The commonest verb is see, with twenty-eight occurrences, followed by hear, with sixteen, and observe, with fifteen. It will be noted that almost 20% of the occurrences (that is, the examples of hear) are cases of aural, rather than visual, perception. Table 17.4 gives a full list of the verbs which encode perception processes in the Philosophical Transactions sample. The verb see occurs in almost all the disciplines. Examples (6), (7) and (8) are from the fields of physics, geology and geography respectively.

Table 17.3 Perception processes by discipline (Philosophical Transactions) Physics Geology Geography Agriculture Astronomy General Biology

26 23 14 7 4 4 3


David Banks Table 17.4 Perception verbs (Philosophical Transactions) See Hear Observe Find Perceive Make (observations) Appear Discern Glance Look Take notice Taste

28 16 15 9 4 2 1 1 1 1 1 1

(6) After that both these liquors were well purged of Air, the upper Glass was sunk into the lower, so as that the Spirit of Wine was mingled with the Aqua fortis; at which instant there was yet seen a very considerable Ebullition. [physics] (7) That which I saw was little, because a great part of the Pit was filled up . . . [geology] (8) When extraordinary Snow falls, and Shepheards are not present to drive their Sheep under shelters, the Sheep gather themselves close together ; and the Snow so covering them, that they cannot be seen for a while . . . [geography] The verb hear occurs mainly in items on physics and geology. (9) I have heard, that Mr. Joshua Childray, whose ingenious disquisitions you have published in several of your Volumes, made a Journal of the quantity of Rain that fell at any time where he abode, in England, these many years. [physics] (10) That those who were in the Bink whilst it was fired, never heard any more noise than that which is usually made by a flash of Gunpowder in the open air, although those in the other Binks, and out, heard a very great one. [geology] Example (9) from the field of physics is taken from an article about climate differences between England and Scotland. Example (10) from the field of geology deals with the problem of coal gas in mines. The presence here of the verb hear as the second most common verb encoding perception processes is significant. For the members of the Royal Society the notion of witness was important: it was not only important to state what one had seen and heard, but also to report what others had seen and heard. The verb observe (fifteen occurrences, plus two of make observations) occurs in all disciplines.

The Process of Perception


(11) Only this we observed one day, that having mingled together, without the Recipient, some of that Oyl and vinegar and Spirit of Wine, and put this mixture in vacuo, it did not boyl up so soon as when there was no Oyl . . . [physics] (12) The shaft of the Coal-pit is about fifteen yards deep; the Soyl a stiff mire, shaly about the middle of the shaft, dry at the bottom (as they say, though I observed some moisture about the middle) and without any quarry of stone . . . [geology] (13) And further, that these things might also be diligently observ’d about the Islands over the whole VVorld, or as far as is possible. [geography] Examples (11), (12) and (13) are from the fields of physics, geology and geography respectively.

Some Tentative Discussion We have seen that cases of perception processes are much more common in the Philosophical Transactions than in the Journal des Sçavans, even though marginal examples in the latter were included in the count. This would seem to indicate that scientific subjects attract the use of verbs encoding perception processes more than the humanities, since the Philosophical Transactions is mainly scientific, which is not the case of the Journal des Sçavans. This does seem to be brought out by the evidence. Of the fifteen examples which occur in the Journal des Sçavans, nine are in scientific subjects: astronomy, chemistry, geography, mathematics and technology; the other six are in the humanities, in the fields of biography, history and theology. There are no examples of perception processes in the fields of classics, literature or editorial matter, nor, perhaps surprisingly, in medicine. If it is accepted that the sciences attract the use of perception processes more than the humanities, then the relative paucity of perception processes in the Journal des Sçavans can be explained by de Sallo’s decision to extend his scope to all the disciplines of his time, so that the humanities takes a fair proportion of the space in the Journal des Sçavans. Thus, this linguistic feature is the result of de Sallo’s editorial decision, and this editorial decision can itself only be understood in terms of the historical context in which he found himself, that is, in the position of editing a clandestinely state-sponsored journal whose ultimate objective was to control all the new knowledge that was then appearing. This is a striking illustration of the way in which not only the general features of a genre, but even the detailed linguistic features can be traced back to the context from which they emanate, and of which they form a part. However, if this were the only explanation of the difference in perception process use between the two journals, taking into account the


David Banks

number of words per article, the different disciplines, as well as factors taken into account earlier, my rough estimate would be that we might expect up to sixty examples in the Philosophical Transactions sample, but not eighty or more. Among the factors that might contribute to this difference, I feel that the differing ideological stances of the intellectuals of the Royal Society and those in France is of great importance. The Royal Society had taken as its maître à penser Francis Bacon. Bacon had promoted the idea that true knowledge should be based on observation and experimentation, as opposed to theory and logic, which had been the classically based way of thinking throughout the Middle Ages and up till then. He also believed that new knowledge, particularly of a scientific or technical nature, should have practical application; it was important that it should be used for the improvement of the human condition (Farrington, 1951 [1949]). This probably accounts for the importance of agricultural items in the Philosophical Transactions sample. It is probably also significant that this fits in well with the Protestant work ethic then prevalent in England (Merton, 1938; Weber, 1976 [1930]). And, perhaps most importantly, he espoused a system of inductive argumentation; that is, argumentation should be based on data, hence the importance of observation and experimentation to discover the facts which are the data on the basis of which arguments can be built up (Bacon, 1905; Henry, 2002). This was enshrined in Sprat’s History of the Royal Society (Sprat, 2003 [1667]), published in 1667, just seven years after the Royal Society had been founded. This was an account of the Royal Society rather than a history in our modern sense, and was often considered as a defence or manifesto of the Royal Society. This contains several fairly extensive eulogious passages on Bacon, of which the following is an example: And of these, I shall onely mention one great Man, who had the true Imagination of the whole extent of this Enterprize, as it is now set on foot; and that is, the Lord Bacon. In whose Books there are every where scattered the best arguments, that can be produc’d for the defence of Experimental Philosophy; and the best directions, that are needful to promote it. All which he has already adorn’d with so much Art; that if my desires could have prevail’d with some excellent Friends of mine, who engag’d me to this Work: there should have been, no other Preface to the History of the Royal Society, but some of his Writings. . . . He was a Man of strong, cleer, and powerful Imaginations: his Genius was searching, and inimitable: and of this I need give no other proof, than his Style itself; . . . The course of it is vigorous, and majestical: The Wit Bold, and Familiar: the comparisons fetch’d out of the way, and yet the most easie: in all, expressing a soul, equally skill’d in Men, and Nature. (Sprat, 2003 [1667], pp. 35–36)

The Process of Perception


French intellectuals, on the other hand, were still under the influence of Descartes. His approach was quite the opposite of that of the English empiricists. He argued for a deductive system of argumentation, which starts with reason and builds up an argument logically. Thus, theory is built up first; the only value of observation and experimentation is in the subsequent attempt to test or validate the theory. It is significant that at the very beginning of his book Opticks, of which the first edition appeared in 1704, but which was probably drafted as early as the late 1660s, Newton says, “My design in this Book is not to explain the Properties of Light by Hypotheses, but to propose and prove them by Reason and Experiments” (1952 [1730], p. 1). In the terms of the time, this was a real Cartesian putdown. Newton is nailing his empiricist colours to the mast, and stating that he is light-years away from the Cartesian approach. The Cartesians started from hypotheses, that is, theory, which might then be tested (or not). This is what Newton refuses: his argumentation is on the basis of the data which his experiments provide (Fara, 2002; Hakfoort, 1988). French scientists were to adopt the empiricist position only slowly, and it was only in the course of the eighteenth century, and after the support for this position from Voltaire and his mistress, Emilie du Châtelet, that the process was complete (Le Ru, 2005). And I would claim that the influence of Descartes can still be felt in French academic writing in the humanities today. If the influence of Bacon on English thought, and that of Descartes on French thought, is taken into account, then it seems clear that the empiricist emphasis on the importance of observation and experimentation to provide scientific data will inevitably lead to the use of processes of perception, and that this can explain the increased number of perception processes in the Philosophical Transactions sample.

By Way of Conclusion In this short chapter we have seen that finite verbs encoding perception processes are more frequent in the Philosophical Transactions than in the Journal des Sçavans. It has been shown that the exact sciences seem to attract perception processes rather more than the humanities do. I have suggested that the difference in the use of finite verbs encoding perception processes is partly due to the editorial decisions of de Sallo and Oldenburg, and that these decisions themselves are a function of the historical situation which generates them. I have also suggested that the influence of Bacon on English thought, producing an empirical approach, and the influence of Descartes on French thought, is an important factor to be taken into account. All of this is of interest and importance for the history of science in general, and in particular for the development of academic writing in French and English.


David Banks

Notes 1. Earlier versions of this chapter were presented (in French) at the Journée d’étude international “Vision, observation, autopsie . . . Le regard savant vu par ses mots”, Brest, France, 20 May 2014, and (in English) at the First Tunisian International Systemic Functional Linguistics Conference and Workshop, Hammamet, Tunisia, 26–28 March 2015. 2. In the examples, the original spelling and typography have been retained, with the exception of ‘long-s’, which has been replaced by a modern ‘s’. In addition, relevant segments are printed in bold. 3. Glosses of French examples are mine. They are intended to give the sense, but not to be polished translations.

References Bacon, F. (ed. John M. Robertson). (1905) The Philosophical Works of Francis Bacon. London: George Routledge and Sons. Banks, D. (2005) Introduction à la linguistique systémique fonctionnelle de l’anglais. Paris: L’Harmattan. Banks, D. (2016) On the (non)necessity of the hybrid category behavioural process. In D. R. Miller and P. Bayley (Eds.), Hybridity in Systemic Functional Linguistics: Grammar, Text and Discursive Context. Sheffield: Equinox, pp. 21–40. Banks, D. (2017) The Birth of the Academic Article. Le Journal des Sçavans and the Philosophical Transactions, 1665–1700. Sheffield: Equinox. Bluhm, R. K. (1960) Henry Oldenburg, F.R.S. (c.1615–1677). In H. Hartley (Ed.), The Royal Society: Its Origins and Founders. London: The Royal Society, pp. 47–56. Clark, G. (1956) The Later Stuarts, 1660–1714, 2nd ed. Oxford: Oxford University Press. Fara, P. (2002) Newton: The Making of Genius. London: Picador. Farrington, B. (1951 [1949]) Francis Bacon: Philosopher of Industrial Science. London: Lawrence and Wishart. Gignoux, C.-J. (1941) Monsieur Colbert. Paris: Grasset. Gotti, M. (2006) Disseminating early modern science: Specialized news discourse in the philosophical transactions. In N. Brownlees (Ed.), News Discourse in Early Modern Britain. Bern: Peter Lang, pp. 41–70. Gotti, M. (2011) The development of specialized discourse in the philosophical transactions. In I. Taavitsainen and P. Pahta (Eds.), Medical Writing in Early Modern English. Cambridge: Cambridge University Press, pp. 204–220. Hakfoort, C. (1988) Newton’s optics: The changing spectrum of science. In J. Fauvel, F. Raymond, M. Shortland and R. Wilson (Eds.), Let Newton Be! Oxford: Oxford University Press, pp. 81–99. Hall, M. B. (2002) Henry Oldenburg, Shaping the Royal Society. Oxford: Oxford University Press. Halliday, M. A. K. (2014) Text Linguistics, the How and Why of Meaning, edited by J. J. Webster. Sheffield: Equinox. Halliday, M. A. K. and Matthiessen, C. (2014) Halliday’s Introduction to Functional Grammar, 4th ed. London: Routledge. Henry, J. (2002) Knowledge Is Power: Francis Bacon and the Method of Science. Cambridge: Icon Books.

The Process of Perception


Hill, C. (1969 [1961]) The Century of Revolution, 1603–1714. London: Sphere Books. Kishlansky, M. (1996) A Monarchy Transformed: Britain 1603–1714. London: Penguin. Kronick, D. A. (1962) A History of Scientific and Technical Periodicals: The Origins and Development of the Scientific and Technological Press 1665–1790. New York: Scarecrow Press. Kronick, D. A. (1991) Scientific and Technical Periodicals of the Seventeenth and Eighteenth Centuries: A Guide. Metuchen, NJ: Scarecrow Press. Le Ru, V. (2005) Voltaire Newtonien, Le combat d’un philosophe pour la science. Paris: Vuibert/ADAPT. Licoppe, C. (1994) The crystallization of a new narrative form in experimental reports (1660–1690). The experimental evidence as a transaction between philosophical knowledge and aristocratic power. Science in Context, 7(2), pp. 205–244. Licoppe, C. (1996) La formation de la pratique scientifique. Le discours de l’expérience en France et en Angleterre (1630–1820). Paris: La Découverte. Merton, R. K. (1938) Science, technology and society in seventeenth century England. Osiris, 4, pp. 360–632. Morgan, B. T. (1928) Histoire du Journal des Sçavans depuis 1665 jusqu’en 1701. Paris: Presses Universitaires de France. Newton, I. (1952 [1730]) Opticks, or a Treatise of the Reflections, Refractions, Inflections & Colours of Light. New York: Dover. O’Donnell, M., Zappavigna, M. and Whitelaw, C. (2008) A survey of process type classification over difficult cases. In C. Jones and E. Ventola (Eds.), New Developments in the Study of Ideational Meaning: From Language to Multimodality. London: Continuum, pp. 47–64. Sprat, T. (2003 [1667]) History of the Royal Society of London for the Improving of Natural Knowledge. Whitefish, MT: Kessinger [facsimile of the 1667 edition. London: J. Martyn]. Weber, M. (1976 [1930]) The Protestant Ethic and the Spirit of Capitalism. London: George Allen & Unwin.

18 Construing Space in Academic Writing Fatma Benelhadj

1. Introduction The notion of ‘space’ has been the concern of intellectuals for a long time, yet the way humans use language to communicate about space is as important as space itself. As Landau (2002, p. 395) explains, “[s]patial representations support our capacity to perceive, act upon, and think about space”. These representations are important in the sense that they shape, and are shaped by, the way space is perceived, thought about and acted upon, or otherwise experienced. Previous linguistic studies have provided a range of perspectives on space including typologies for motion expressions (Beavers, Levin and Wei Tham, 2010), categorizations of spatial entities (Aurnague, Hickmann and Vieu, 2007) and spatial language and dialogue (Coventry, Tenbrink and Bateman, 2009). However, a study by Matthiessen and Kashyap pointed out that space can be construed differently “for different purposes in different registers” (Matthiessen and Kashyap, 2014, p. 1). In fact, experiences with space may not be universal, as they may be affected by the context. Construing space differently is possible only due to the linguistic resources offered to the speaker. From the perspective of Systemic Functional Linguistics (SFL), language has an experiential dimension that represents “our experience of what goes on” (Halliday and Matthiessen, 2014, p. 220), and is reflected in a “configuration of a process, participants involved in it, and any attendant circumstances” (Halliday and Matthiessen, 2014, p. 212). Circumstantials are “optional augmentations of the clause”, and they tend to augment the “experiential centre . . . temporally, spatially, causally, and so on” (Halliday and Matthiessen, 2014, p. 221; emphasis added). Circumstantials are very important in pointing to, and moving through, space; “they are typically expressed . . . as either adverbial groups or prepositional phrases—mostly the latter, since adverbial groups are largely confined to one type, those of Manner” (Halliday and Matthiessen, 2014, p. 311). The present chapter seeks to investigate how space is construed through focusing on prepositional phrases (henceforth PPs) of space.1

Construing Space in Academic Writing


In the following section the notion of space in language is reviewed. Section 3 discusses the treatment of context in SFL. Following this, the corpus and the methodology adopted in the study are explained in section 4. The analysis and key findings are then discussed in section 5. The chapter concludes in section 6, showing that while certain frequencies are ‘global’ and reflect general features of academic writing, others are ‘local’ as they reveal different construals of space across the registers.

2. Space in Language English is classified among the “satellite-framed languages” (Beavers et al., 2010, p. 333), where motion events are encoded in satellites (particles). In fact, Biber, Johansson, Leech, Conrad and Finegan (2007, p. 787) studied the syntactic realisation of adverbials of place, and concluded that “place is realized overwhelmingly by prepositional phrases”. Prepositions bear some similarities to verbs in the sense that both typically require complements. Quirk, Greenbaum, Leech and Svartvik (2005) propose that the preposition establishes a relation between two entities, one of them being the prepositional complement. Within the framework of SFL, this linking role of the preposition is further stressed by the claim that “the preposition serves as a ‘minor process’” and that a prepositional phrase is “a clause-like structure in which the Process/Predicator function is performed by a preposition and not by a verb” (Halliday and Matthiessen, 2014, p. 329/423). The similarity to verbs also extends to the semantic relation that a preposition expresses between its complement, and other previous element(s). Prepositions express different meanings that situate one element with respect to others, and “when we use a preposition to indicate space, we do so in relation to the dimensional properties, whether subjectively or objectively conceived, of the location concerned” (Quirk et al., 2005, p. 673). This means that the different meanings expressed by prepositions impose a particular dimension on the location they describe. Space is, therefore, classified into different meanings. From an SFL perspective, space is one of the circumstantial meanings, which can be extent (walk for seven miles) or location (work in the kitchen). Each of these can be definite (at home) or indefinite (near). Locations can be either absolute (in Australia) or relative (near). Finally, location can show either rest (in Sydney) or motion (to/from Sydney) (Halliday and Matthiessen, 2004, 2014). Table 18.1, adapted from Halliday and Matthiessen (2014, p. 317), summarises the classification of spatial meanings. In addition to circumstantial elements which reflect the experiential dimension of the clause, spatial meanings can be textual, functioning as a ‘conjunctive adjunct’ (at this point). Furthermore, Halliday and Matthiessen (2014, p. 317) add that “space includes not only concrete space, but also abstract space”. Abstractness may result from an abstract location


Fatma Benelhadj

Table 18.1 Semantic classification of space circumstantials Extent (distance)

Location (place)



five miles


a long way


at home




in Australia




here, nearby


there, a long way away

Rest Motion

in Sydney, at the airport Towards

to Sydney

Away from

from Sydney

(this brings us to the purpose of the symposium) or an abstract participant (a great sadness came over him). In addition to their different meanings, prepositions can be ranked on a scale of explicitness since each expression “carries a different degree of explicitness in the encodings of referents in the world” (Svorou, 1994, p. 6), and a degree of specificity, defined as “the amount of detail with which spatial relations are described” (Svorou, 1994, p. 6). To illustrate both concepts, Svorou gives the example of “on the left hand side of Main Street as you go towards downtown, next to the Utica subway station”, which shows a high degree of explicitness because “it makes use of surrounding landmarks” and a high degree of specificity since “it not only tells us about the surrounding objects, but also the direction of travel” (1994, p. 6). The linguistic resources expressing space offer to the speaker different moments of choice, and are thought to be motivated by the context.

3. Context From an SFL perspective, “language operates in context” (Halliday and Matthiessen, 2014, p. 32). Hasan (2009, p. 170) holds context to be “the highest stratum in the theory” as it encompasses the SEMANTICS and LEXICOGRAMMAR strata, which means that the semantic and lexicogrammatical choices are explained with reference to the context where the text operates. Context is defined as a configuration of field, tenor and mode representing the register of the text, which corresponds, in academic writing, to disciplines (Gardner, 2012). According to Hyland (2004, p. 8), a discipline has “its particular norms, nomenclature, bodies of knowledge, sets of conventions and modes of inquiry”. As Gardner explains (2012, p. 54), students are “socialised into the language and ways of thinking of

Construing Space in Academic Writing


academic disciplines through participation in lectures, reading textbooks and writing assignments”. In this study, two disciplines are compared: the medical and social sciences. These two disciplines are known for having three main differences (cf. Ernest, 1994; Becher, 1994): having different foci (the human body versus the human being in society), following different research paradigms (scientific and interpretative) and being classified as ‘hard’ and ‘soft’ sciences respectively. In addition to the register, or the context of situation, a text can be defined according to the culture where it was created, represented in the genre of the text. This chapter focuses on the genre of ‘research articles’ (henceforth RAs), as they are “the main means employed by the hard sciences for the introduction of new findings and claims in the community” (Koutsantoni, 2006, p. 20). RAs are organised following the IMRD structure (Introduction, Methods, Results and Discussion), which is described as coming “into almost universal use in research journals” (Day and Gastel, 2006, p. 9). These features can be viewed as relating to the highest stratum of context. Within the theory of SFL, the three strata of context, meaning and wording are related since, as Hasan (2009, p. 171) explains, “lexicogrammatical choices CONSTRUE semantic choices CONSTRUE contextual ones”. The focus of this chapter is on PPs expressing space, as lexicogrammatical choices, and its aim is to identify the meanings of space they construe in relation to the contexts of academic writing in the medical and social sciences.

4. The Corpus A corpus of RAs from both disciplines was collected from online journals (ScienceDirect), as represented in Table 18.2. The prepositions in the corpus were identified using AntConc2 (Version 3.3.5), a free concordance program. A sample was chosen by applying a stratified random sampling technique because the concordance revealed that certain prepositions are more frequent than others and because the corpus contains two disciplines: medical and social sciences, and four sections: Introduction, Methods, Results and Discussion. The stratified random sampling technique “is applied so as to obtain a representative sample. In this technique, the population is stratified into a number of non-overlapping subpopulations or strata and sample items are selected from each stratum” (Kothari, 2004, p. 16). Among the sampled PPs, only Table 18.2 The corpus

Number of words Number of items

Medical RAs

Social sciences RAs

245,000 50

246,000 30


Fatma Benelhadj

Table 18.3 Spatial references Preposition Number of items

For 1

To 13

In 48

Between 15

On 1

From 15

At 9

Figure 18.1 The coding scheme

those expressing space are studied in this chapter; they form 102 PPs, and they are classified as shown in Table 18.3. The sample was then annotated using the UAM CorpusTool (Version  2.8.14), software for corpus annotation, using a coding scheme (Figure 18.1) based on systems of spatial meanings. The coding scheme contains additional coding schemes to identify the discipline (medical or social sciences) and section (Introduction, Methods, Results or Discussion) for each PP. The analysis, being both quantitative and qualitative, seeks to compare the way space is construed across the registers and sections of the RA.

5. Analysis and Findings In the construction of space, different moments of choice are available, as suggested in Figure 18.1, the first being whether the spatial references have an experiential or textual meaning. The analysis has revealed that experiential spatial references (95 cases) outnumber textual ones (7 cases), as shown in Table 18.4. Spatial textual references are made to the text itself, or to particular sections, while spatial experiential references construe the circumstances of the actions in the text, as revealed in examples (1) and (2): (1) MA INT 3. In the present study, we demonstrated that TCR gene rearrangement in fetal thymocytes was perturbed by the E47 deficiency in a dosage-sensitive fashion. (textual)

Construing Space in Academic Writing


Table 18.4 Textual and experiential references

Experiential Textual

Medical sciences

Social sciences


50 3

45 4

95 7

(2) MA INT 2. The development of genetically engineered mouse (GEM) PCa models can help dissect the role of the FGF axis in disease progression in a physiological setting. (experiential) Space is construed differently in these examples; in the first example, it designates the study presented in the article, while in the second example, space represents the circumstances around the experiments reported in the text. These different spaces reveal two different personas for the speaker: the author of the text and the researcher, respectively. Space can be revealing of identity; within each space, a particular role emerges. However, the use of experiential references reveals the focus on the researcher’s role, rather than the author’s. The cohesion of the text emanates from its content rather than from the use of conjunctive devices, a result similar to that of Sellami Baklouti (2011, p. 514), who claims that “hard disciplines abstracts . . . display . . . a higher probability of absence of overt conjunctive adjuncts”. In fact, in science, cohesion results from the rigour of the argument, rather than the use of conjunctions and transitions, and this makes the scientific text difficult to read for the non-expert. 5.1. Textual References Textual references are made either to the whole study or to precise sections or tables. The latter are found in the social sciences texts in the Results and Discussion sections. Tables are frequent in the Results because this section presents “the main data that support (or reject[ing]) the hypotheses in the form of tables and graphs” (Hartly, 2008, p. 47), as shown in example (3): (3) SA RES 5. In Table 18.2, we present the effects of the broad and restricted White working class variables with no controls. In this corpus, social sciences texts seem to rely on tables, while graphs are more frequent in the medical texts. Such differences can be related to the nature of the study in each discipline. In the medical sciences, the data is about the human body, which needs graphs to be reproduced (photographs, electrograms, etc.). In the social sciences, researchers try to translate the qualitative data into numerical data in the form of tables.


Fatma Benelhadj

Tables can be recalled in the Discussion sections since while Results sections are about describing results, the Discussion is about re-describing them, according to Swales (1990), as example (4) shows: (4) SA DIS 1. There may, in fact, be other relevant variables that are not included in the models presented in Table 18.1 and Table 18.2. A prominent feature of a Discussion section is that it should emphasize “how to . . . limit and qualify your claims so that they are in conformity with the evidence” (Harmon and Gross, 2010, p. 42; my emphasis). Tables are not only used to present the data but also to limit the scope of the findings. The other textual references are made to the whole contribution. They are PPs headed by the preposition ‘in’ and meant to describe what is ‘contained’ in the ‘study’ or ‘paper’. (5) MA DIS 7. In the present study, alcohol use was prospectively examined at 6-month intervals using a comprehensive, validated instrument. (6) SA MET 5. In this paper, we describe patterns of change in segregation for the 100 most populous metropolitan areas (as of the 2000 census) in the U.S. A contrast is implied between what is proposed in the RA and what already exists in the literature. These references are not purely textual as they emphasize the originality of the research work. They are tools used for important announcements about the methodology, mainly in the Methods and Discussion sections. 5.2. Experiential References Experiential references to space are those representing the experience with space that is reported in the text. Table 18.4 reveals a difference in the distribution of experiential references across the two disciplines (50 in the medical texts and 45 in the social sciences texts) that can be explained by the other meanings expressed in the social sciences, namely ‘respect’. PPs of respect are used “to identify a relevant point of reference in respect of which the clause concerned derives its truth value”, and these PPs bear a similarity to other meanings, mainly locative ones (Quirk et al., 2005, p. 483). The study of all the meanings expressed by the PPs revealed that PPs of ‘respect’ are more frequent in the social sciences. In addition to the importance of spatial references, authors in the social sciences do need to express the respect to which they are expressing their messages. These PPs are very similar to those expressing place, yet the place they are referring to is not a real and concrete one, as revealed in the following example. (7) SA INT 5. Moreover, the scale at which segregation is measured will determine the extent of segregation observed—a measure of the

Construing Space in Academic Writing


variation in racial composition among street blocks will not reveal the extent to which racial composition varies among regions of the country; likewise, a measure of the variation in racial composition among regions will not reveal the extent to which the average block differs from the next. The variation is not situated inside ‘racial composition’, but is rather presented with respect to the notion of ‘racial composition’. This shows the importance of situating one’s variables with respect to precise values in the social sciences, as possible interpretations may exist. The focus in social sciences is not on real places, as it is in the medical sciences, but rather on defining one’s terminology and variables with respect to preestablished ones. Experiential references are studied across the sections of RAs. The results are presented in Table 18.5. Table 18.5 shows that experiential references of space are not equally distributed across the sections of the RA of each discipline. The Introduction and the Methods are similar with respect to the number of references. However, there are more spatial references in the Results of medical articles, which is not the case in the social sciences. Generic features are not universal; they can be subject to disciplinary differences. The difference between the two disciplines is also explicit in the second moment of choice: concrete and abstract. 5.3. Concrete and Abstract References Spatial references can be either concrete or abstract. The results reveal that this distinction seems to be influenced by the context, as its distribution differs with respect to the medical and social sciences, as revealed in Table 18.6. Table 18.6 shows that most of the concrete references are made in the medical sciences (29 cases), while in the social sciences, most references are abstract (35 cases). The chi-square test shows a high significance for this result, which is further stressed by the correlation coefficient Q, being beyond the rejection interval: Q = A – B / (A + B) = 0.657

Table 18.5 Experiential references of space across the sections and the disciplines

Introduction Methods Results Discussion

Medical sciences

Social sciences


11 11 18 10

11 11 10 13

22 22 28 23


Fatma Benelhadj

Table 18.6 Concrete and abstract spatial references

Concrete Abstract

Medical sciences

Social sciences


29 21

10 35

12.528 12.528

Table 18.7 PPs expressing concrete and abstract places

Concrete Abstract








0 1

2 11

28 13

2 13

0 1

7 8

0 9

This correlation implies that a systematic relation exists between the type of reference (abstract or concrete) and the register. A register is distinguished by its combination of field, tenor and mode, where the field presents the topic dealt with in the text. Biber and Conrad (2009, p. 46) claim that “the words used in a text are to a large extent determined by the topic of the text”, which actually reflects the field of the texts. Thus, the presence of concrete references reflects the field of medical sciences, whereas the abstract references are a reflection of the social sciences. The two disciplines seem to have two distinct fields, which would make them distinct registers. The spatial references can be concrete or abstract based on the place itself; however, the analysis revealed that certain prepositions are associated with one meaning over the other, while others are equally used to convey both meanings, as shown in Table 18.7. Table 18.7 shows that not all the prepositions express the meaning of space, although it is thought to be the most basic of all the meanings. This shows that the central meaning of prepositions is not necessarily spatial but rather conceptual (Tyler and Evans, 2003, p. 15). Each preposition is linked to a mental concept which is translated into a particular meaning whenever used in context. The direct linguistic context plays an important role in activating the meaning of the preposition. Prepositions such as with, as, by and about do not express spatial meanings in the studied PPs. Other prepositions, such as for, of and on, do not convey this meaning frequently, according to Table 18.7. This can be related to the degree of specificity of these prepositions or “the amount of detail with which spatial relations are described” (Svorou, 1994, p. 6). For instance, the prepositions with and by express a conceptual meaning of closeness and proximity, which is not specific compared to other meanings of path (to), position (at), origin (from), etc. Spatial prepositions are those which have a more specific conceptual meaning.

Construing Space in Academic Writing


Furthermore, Table 18.7 shows that the prepositions to, in, between, from and at express most of the spatial meanings, and that to, between and at are rather associated with abstract meanings, while in and from express both abstract and concrete meanings. The meanings of the PPs are classified into concrete or abstract based on the types of complements they have, as shown in examples (8) and (9): (8) MA RES 28. Consistent with the previous data, the Ile161Met mutation did not markedly impact on the interaction between CD94NKG2C and HLA-E. (concrete) (9) SA MET 27. Frame analysis is particularly applicable to the media due to its appreciation of the link between the production and consumption of the news story. (abstract) The same preposition can express concrete or abstract spatial meanings, the difference being the complement of the preposition. The preposition to originally expresses a movement to a precise point. Metaphorically, the meaning of this preposition can extend and cover any movement with a target, or transition: (10) MA MET 11. In brief, one round of amplification was performed, RNA samples were labeled with Cy3 or Cy5 dyes, and samples were hybridized to the Mouse Operon oligo set 4.0 Chip. (concrete) (11) MA DIS 11. In such cases, one reliable indicator of disease is that the pain is out of proportion to the physical examination or the cutaneous findings [13]. (abstract) For further analysis of these meanings, the functions of those PPs have been studied, yielding the results in Table 18.8. Table 18.8 summarises the distribution of the grammatical functions of postmodifier, adverbial and complement with respect to abstract and concrete spatial meanings. The analysis reveals that most abstract meanings are associated with the functions of postmodifiers and complements, while most concrete senses are adverbials. Unlike adverbials, which occur freely in the clause, postmodifiers and complements are syntactically and semantically bound to other linguistic units. This result is proved by the

Table 18.8 Functions of abstract and concrete spatial PPs

Postmodifier Adverbial Complement




22 20 14

13 23 3

0.350 5.020 4.687


Fatma Benelhadj

chi-square test, which is significant, thus showing that the distribution of the grammatical functions according to the meanings is not even. The analysis has revealed that to functions as part of a complement to certain verbs by indicating the direction of their movements. This is also the case for the function of the postmodifier, since most of the nouns are nominalised verbs. Conceptually, to expresses the positive meaning of a movement to “the endpoint of a path” (Lindstromberg, 2010, p. 30), and this meaning seems to be shaped not only by the prepositional complement but mainly by the preceding verb or noun. The PP, therefore, expresses a metaphorical meaning of direction, convergence, pointing, etc. More precisely, it shows the direction of the effect of the preceding verb or noun. (12) MA MET 12. Nonidet P-40 detergent (NP-40) was added to a final concentration of 1% before PNGase F addition. In addition to to, preposition between is used in the corpus to express abstract locations, as shown in Table 18.7. This preposition functions as a postmodifier to other nouns: (13) SA MET 27. Frame analysis is particularly applicable to the media due to its appreciation of the link between the production and consumption of the news story. (14) MD DIS 28. The aim of this chapter was to identify individual residues of core protein required for the interaction between core and DDX3. This preposition expresses a meaning of centrality of the heads link and interaction towards the prepositional complements. Furthermore, Table 18.7 shows that spatial meanings, whether abstract or concrete, are expressed by the prepositions in, from and at. The first preposition expresses containment in space, while the second has the meaning of negative source, and the third expresses the meaning of a positive position. Containment can be in real or abstract places: (1) MA MET 9. The nitrocellulose membrane was blocked for 1 h with 5% skimmed milk (Oxoid, UK) at 37 °C, incubated with the hyperimmune serum diluted in phosphate buffered saline (PBS) for 1 h and washed with PBS-0.1% Tween-20 (PBST). (2) SA DIS 2. Differentiating between employees in terms of their bargaining power in the employment relationship has also been a wellknown theme in labor market sociology. (3) MA MET 23. Values were obtained by phlebotomy if not available from clinical records within 4 months of the interview.

Construing Space in Academic Writing


(4) SA INT 25. Even Crothers (2002), a critic of Putnam, advocates “street-level” civic revitalization at the neighborhood and community levels. Whether concrete or abstract, the meaning of containment of the preposition in seems to be expressed in the same way, and this is also the case for the meanings of origin and position. The difference between the references can be related to the register. Concrete and abstract spatial references seem to be register-related, as far as the prepositions in, from and at are concerned. However, certain prepositions, to and between, are associated with abstract rather than concrete PPs because of the grammatical function they have; they complement verbs and nouns, and this shapes their meanings to drive them away from pure concrete references. The degree of specificity of a preposition seems to be in a way responsible for its meaning and its function in the clause. More specific prepositions seem to have more concrete spatial meanings, while those which are less specific are rather used to express abstract meaning as they are associated with other grammatical categories (N, V, etc.). Less specific prepositions allow the creation of new and abstract meanings from the interaction with the words that they complement. The classification of concrete and abstract references differs also across the sections in the medical sciences, as shown in Table 18.9. Table 18.9 reveals the reliance on concrete references in the Results section and the abstract ones in the Discussion. The concrete references are mostly expressed by the preposition in: (5) MA RES 4. On the other hand, the titers of primary IgG1 antibodies in miR-155 deficient mice were reduced to 20% of WT controls when NP17-BSA was used as the capture antigen (Figure 1B, middle). (concrete/Results) (6) MA DIS 5. As in all observational studies, we cannot rule out bias due to imbalance in an unmeasured covariate. (abstract/Discussion) Concrete references in the Results section present the results. Those in the Discussion section are rather abstract as they present more general Table 18.9 Concrete and abstract references in medical RAs Medical




Introduction Methods Results Discussion

6 7 14 2

5 4 4 8

0.069 0.184 4.516 7.410


Fatma Benelhadj

references, which is basically the difference between the sections: more generalisations in the Discussion and presentation of data in the Results. Another moment of choice in the construal of spatial meanings is that of extent versus location. 5.4. Extent and Location Spatial references can be made to a particular point in place, called location, or to a whole interval, that is, a distance, called extent. In the studied sample, the semantic distribution of PPs shows that space PPs are all classified as locations rather than extents. Both registers seem to locate actions in space, but do not opt for defining distances, which can again be linked to the focus of these registers. The sciences studied here do not opt for measuring distances as part of their methodologies. Besides, all of the locations are definite, depicting precise positions in space. This highlights the ability of prepositions to express precise meanings. In fact, most of the PPs have been annotated as definite since indefinite references are typically conveyed via adverbs. These definite locations can be either rest or motion. The result of the annotation of this system is presented in Table 18.10. Table 18.10 shows that most spatial references express rest locations (68 percent), while 32 percent of all the spatial references express motion. The difference between the two categories seems to follow from the preposition itself, as these two meanings differ with respect to the prepositions that introduce them. For further insights, Table 18.11 summarises the distribution of prepositions. Table 18.11 shows that certain prepositions like in, between, on, after and at are associated with rest locations, while other prepositions such as to and from are linked to motion. As for the preposition in, although associated with rest positions, one occurrence expresses motion: (7) MA INT 8. The flow of information through the amygdala starts with cortical and thalamic input into the lateral and basolateral nuclei, then proceeds from there through efferent projections to the central nucleus of the amygdala, bed nucleus of the stria terminalis, and the nucleus accumbens. This example has been classified as motion because the meaning of into is different from in, as the preposition to adds to the complex preposition Table 18.10 Rest and motion references across the disciplines

Rest Motion

Medical sciences

Social sciences


35 15

30 15

65 (68%) 30 (32%)

Construing Space in Academic Writing


Table 18.11 Prepositions expressing rest and motion

Rest Motion








0 1

0 13

40 1

15 0

1 0

0 15

9 0

Table 18.12 Grammatical functions of motion PPs

Postmodifier Adverbial Complement

Medical sciences

Social sciences


3 7 5

7 1 7

2.40 6.10 0.55

the meaning of path. This shows how complex prepositions depart from the original meanings expressed by their prepositions. Among the prepositions expressing motion, the most basic, to and from, are relied on in the corpus, to the exclusion of other possibilities like out of, off, onto, etc. (Quirk et al., 2005). This emphasises the inclination towards choice of simple linguistic items in the domain of academic writing. The prepositions expressing motion are to and from, classified by Quirk et al. (2005) as expressing positive and negative meanings, respectively. These two prepositions bear some similarities in that they express symmetrical movements, with the first converging towards a point, and the second diverging from another. Motion PPs seem to be explored differently across the two registers, as they seem to have different functions, as shown in Table 18.12. Table 18.12 shows that PPs are associated with the function of adverbials in the medical register, while they are postmodifiers or complements in the social sciences register. The chi-square test is highly significant, showing a correlation between motion PPs and the function of complement, on the one hand, and rest PPs and adverbials, on the other. The preposition seems to be responsible not only for the meaning expressed by the whole PP but also for the function of the PP in grammar, and this can be associated with the link between certain verbs or adjectives and prepositions: (8) SA RES 23. The effect of cohabitation on fertility is not significantly different from single status in any country sampled. (9) MA MET 23. Values were obtained by phlebotomy if not available from clinical records within 4 months of the interview. The same preposition is explored differently in the different contexts. It expresses a static meaning of source and functions as an adverbial


Fatma Benelhadj

in the medical context, while in the social sciences texts, it rather expresses a departure from a status, which is highlighted by the adjective different. This shows how realities are different across the two disciplines: in the medical texts, descriptions are made of real-life phenomena, which would require their location in rest positions, while in the social sciences, interpretations are made, and relations are sought, and these can be rather construed by relational adjectives and verbs, which are followed by PPs.

6. Conclusion This chapter has examined the way space is construed in academic writing. It has focused on PPs expressing space in RAs from the medical and social sciences. The main claim defended is that space, expressed via PPs, is construed differently in different contexts, which means that lexicogrammatical choices of PPs are activated by features of the context, namely the register and the sections of the RA. The analysis has revealed certain similarities across the corpus, thus revealing ‘global frequencies’ about the field of academic writing (Halliday and Webster, 2009), while other choices reveal differences in the way space is perceived in the two disciplines. For instance, certain textual references are used across the registers to highlight the contribution(s) in the RAs, while experiential meanings reveal differences among registers, as concrete references are in the medical texts, while abstract ones are in the social sciences texts. Moreover, the genre of RAs, although showing global features, reveals differences among the disciplines, thus highlighting a dialogic relation between register and genre as the two main components of the context. Experiential references are equally distributed in the Introduction and Methods sections, while they are different in the Results and Discussion sections. Generic features may not be universal; they can be influenced by the register. Furthermore, the meaning of space expressed by PPs seems to be influenced by the prepositional complements (abstract and concrete) and the preposition itself. Spatial references are those where the preposition expresses a more specific conceptual meaning (Svorou, 1994, p. 6). Furthermore, a link can be established between the meaning of the PP and its grammatical function, as abstract references are complements or postmodifiers, and therefore syntactically and semantically bound. Register differences are also present in the way linguistic features are explored. In fact, the same preposition (from) expresses different meanings across the registers: it is static in the medical texts and relational in the social sciences texts.

Construing Space in Academic Writing


Notes 1. This chapter is based on a wider research project on prepositional phrases available at 2. This can be downloaded from html.

References Aurnague, M., Hickmann, M. and Vieu, L. (2007) The Categorization of Spatial Entities in Language and Cognition. London: John Benjamins Publishing Company. Beavers, J., Levin, B. and Wei Tham, S. (2010) The typologies of motion expressions revisited. Journal of Linguistics, 46(2), pp. 331–377. Becher, T. (1994) The significance of disciplinary differences. Studies in Higher Education, 19(2), pp. 151–161. Biber, D. and Conrad, S. (2009) Register, Genre and Style. Cambridge: Cambridge University Press. Biber, D., Johansson, S., Leech, G., Conrad, S. and Finegan, E. (2007) Longman Grammar of Spoken and Written English, 6th impression. UK: Pearson Education Limited. Coventry, K. R., Tenbrink, T. and Bateman, J. (Eds.). (2009) Spatial Language and Dialogue. Oxford: Oxford University Press. Day, R. and Gastel, B. (2006) How to Write and Publish a Scientific Paper, 6th ed. Cambridge: Cambridge University Press. Ernest, P. (1994) An Introduction to Research Methodology and Paradigms. Exeter, UK: School of Education, University of Exeter Research Support Unit. Gardner, S. (2012) Genres and registers of student report writing: An SFL perspective on texts and practices. Journal of English for Academic Purposes, 11, pp. 52–63. Halliday, M. A. K. and Matthiessen, C. M. I. M. (2004) An Introduction to Functional Grammar, 3rd ed. London: Edward Arnold. Halliday, M. A. K. and Matthiessen, C. M. I. M. (2014) An Introduction to Functional Grammar. 4th ed. London: Edward Arnold. Halliday, M. A. K. and Webster, J. J. (2009) Continuum Companion to Systemic Functional Linguistics. London and New York: Continuum. Harmon, J. E. and Gross, A. G. (2010) The Craft of Scientific Communication. Chicago and London: The University of Chicago Press. Hartly, J. (2008) Academic Writing and Publishing: A Practical Handbook. London: Rutledge, Taylor and Francis Group. Hasan, R. (2009) The place of context in a Systemic Functional Model. In M. A. K. Halliday and J. Webster (Eds.), Continuum Companion to Systemic Functional Linguistics. London: Continuum, pp. 166–189. Hyland, K. (2004) Disciplinary Discourses: Social Interactions in Academic Writing. Ann Arbor: University of Michigan Press. Kothari, C. R. (2004) Research Methodology: Methods and Techniques. New Delhi: New Age International Ltd. Publishers. Koutsantoni, D. (2006) Rhetorical strategies in engineering research articles and research theses: Advanced academic literacy and relations of power. Journal of English for Academic Purposes, 5, pp. 19–36.


Fatma Benelhadj

Landau, B. (2002) Spatial cognition. In V. Ramachandran (Ed.), Encyclopedia of the Human Brain, vol. 4, pp. 395–418. San Diego: Academic Press. Lindstromberg, S. (2010) English Prepositions Explained. Amsterdam and Philadelphia: John Benjamins Publishing Company. Matthiessen, C. M. I. M. and Kashyap, A. K. (2014) The construal of space in different registers: An exploratory study. Language Sciences, 45, pp. 1–27. Quirk, R., Greenbaum, S., Leech, G. and Svartvik, J. (2005) A Comprehensive Grammar of the English Language. London: Longman. Sellami Baklouti, A. (2011) The impact of genre and disciplinary differences on structural choice: Taxis in research article abstracts. Text & Talk, 31(5), pp. 503–523. Svorou, S. (1994) The Grammar of Space. Amsterdam and Philadelphia: John Benjamins Publishing Company. Swales, J. (1990) Genre Analysis: English in Academic and Research Settings. Cambridge, UK: Prentice-Hall International. Tyler, A. and Evans, V. (2003) The Semantics of English Prepositions: Spatial Scenes, Embodied Meaning and Cognition. Cambridge: Cambridge University Press.

19 Elaboration Postmodifiers in Academic and Popular Medical Articles Imen Ktari

1. Introduction Systemic Functional Linguistics (SFL) has been often considered as a theory of language, not only grammar, since it seeks “to account for how language enables human beings to communicate with one another in the ways they do” (Butler, 2009, pp. 63–64). One of the means of communication is the text. Indeed, one of the advances of SFL is its prioritization of text rather than the clause. Unlike many other theories, SFL has given paramount importance to text and genre analysis, considering the text as an ‘instantiation’ of the system (Halliday and Matthiessen, 2014). As such, interpreting the systemic choices in a text leads to an interpretation of the text itself. Text is, nonetheless, not a homogeneous concept as different texts belong to different genres having different communicative purposes and formal features. This chapter focuses on the logico-semantic relationship and more precisely on elaboration postmodifiers within academic articles (AAs) and popular articles (PAs) as an example of a particular genre and seeks to show how the lexico grammatical choice of elaboration postmodifiers is genre-driven. The analysis adopts a qualitative and quantitative approach to a onemillion-word-corpus equally divided between AAs and PAs belonging to medical discourse. In what follows, Section 2 reviews the key areas of concern in relation to postmodifiers in English, with a focus on the nominal group as well as the logico-semantic relationship within postmodifiers. Section 3 argues that these choices could be quantified and that their probabilities of occurrence in the text could be context-sensitive, highlighting the notions of choice, system network, text and context, and genre. Section 4 sets out the methodology, describing the corpus under study and presenting the analytical and methodological tools. The main findings of this investigation are presented in Section 5, drawing on the frequency distribution of elaboration postmodifiers in the whole corpus and, most importantly, across the two genres. Finally, Section 6 summarizes the chapter and offers some pointers for future research.


Imen Ktari

2. Postmodifiers in SFL This part reviews the nominal group in English with a focus on postmodifiers. It then examines the logico-semantic relationship within postmodifiers. 2.1. Nominal Groups Nouns are used to represent entities and classify things in the world. As a result, they “tend not to change very much over time”, and they are “referred to repeatedly in discourse as the same thing” (Verspoor and Sauter, 2000, p. 71). Syntactically speaking, a noun is part of a larger unit called the noun phrase (NP) or, in SFL terms, the nominal group (ngp). In fact, from the SFL perspective, these two terms are not equivalent since “words expand to form groups . . . clauses contract to form prepositional phrases . . . clauses and phrases get embedded inside (nominal groups)”. The preposition acts as a reduced process and the embedded nominal group is the complement of this process (Halliday, 2003, p. 19). With its “infinite potential for pre-and post-modification with embedded and subembedded structures”, the ngp is “one of the most loaded groups” (Pop, 2007, p. 56). In fact, it can be constituted of a head, a determinative (predeterminers, central determiners and postdeterminers), premodification and postmodification (Quirk, Greenbaum, Leech and Svartvik, 2005, p. 1239). Fawcett (2000), on the other hand, provides the following classification: determiner (d), modifier (m), thing (th) and postmodifier (q). In the same context, Halliday and Matthiessen (2014) argue that the ngp consists of Deictic, Numerative, Epithet, Classifier, Thing and Postmodifier. A presentation of these views is found in Table 19.1. The example proposed, a little house with a yard, is taken from Fontaine (2017, p. 268). Throughout this chapter, I will use the terms head and postmodifier, thereby focusing on the head-modifier relationship. According to Fawcett (2000, p. 217), the function of the head of the nominal group is (assuming it is a noun) to state the ‘cultural classification’ of the referent. . . . [T]he head realizes one type of meaning that relates to the referent, while the modifier realizes another. Postmodifiers, on the other hand, have been defined as the elements that occur after the Thing element (Halliday and Matthiessen, 2004, p. 323). They are different from other elements of the ngp because “they are generally expressed by prepositional phrases and clauses, which means that they are not operating at the same level or rank as the rest

Elaboration Postmodifiers


Table 19.1 Basic analysis of the nominal group by various grammars (Fontaine, 2017)

Fries (1970)



determiner 2

looseknit head modifier Epithet Thing Head Modifier head

Halliday (1985 onwards) Deictic Premodifier Fawcett (1980, 2000) Quantifying determiner Martin (1992) Deictic Epithet Fontaine (2012) deictic determiner Modifier

house with a yard

thing thing

restrictive modifier Qualifier Postmodifier Qualifier Qualifier Qualifier

of the nominal group” (Fontaine, 2017, p. 277). The units expressing the postmodifier are “embedded or rank-shifted in that they operate at a rank that is either equivalent to or higher than the nominal group” (ibid.). 2.2. Logico-Semantic Relationship Within Postmodifiers Functionally speaking, postmodifiers are connected to the system of the logico-semantic relationship describing “the semantic relations, the ways in which clauses that are either independent or dependent build on the . . . meanings of the clauses they relate to” (Eggins, 2004, p. 270). The logico-semantic relationship distinguishes between expansion, the “logico-semantics of developing on previous meanings”, and projection, “the logico-semantics of quoting and reporting speech or thoughts” (p. 271). The main function of expansion is to “define, delimit or specify” (Halliday and Matthiessen, 2004, p. 428). Three distinctions are made within expansion: elaboration, extension and enhancement, as illustrated in the scheme in Figure 19.2. Elaboration is achieved through “the typical defining relative clause, introduced by who, which, that, or in so-called ‘contact clause’ form without any relative marker” (p. 428). It provides information that restates, clarifies, refines or adds a descriptive attribute, as illustrated in example (1). (All illustrations are taken from Halliday and Matthiessen, 2004). (1) The man who came to dinner stayed for a month. Extension is realized through the sense of possession, introduced by whose, of which/which . . . of or a ‘contact’ relative ending with of (ibid., p. 432), as in example (2). (2) I recently read an incredibly well-written story about a couple whose thirty-something-year-old son dies of an illness.


Imen Ktari

Enhancement, on the other hand, implies a circumstantial (time, place, manner or cause) relationship between the postmodifier and the head. (3) Some may precipitate directly from sea water in areas where volcanism releases abundant silica. The second type of postmodifiers falls within projection, where “the secondary clause is projected through the primary clause” (ibid., p. 427), either through expressing an idea (what is thought) (4) or locution (what is said) (5). (4) To what extent do you buy into the belief that if the individual becomes enlightened, that adds to the betterment of the universe in and of itself. (5) I was very intrigued by your take on Huck Finn in that piece, and your argument that the great American novel of the century was Uncle Tom’s Cabin. Space does not permit a detailed review of all types of logico-semantic relationship since the focus here is on elaboration postmodifiers only. See Halliday and Matthiessen (2014) for a full account.

3. SFL Approach In this section, the central SFL concepts of choice, system, text and context, and genre are discussed. 3.1. Choice in SFL SFL is “a variety of system-structure theory” (Halliday, 2009, p. 63, emphasis in original), where structure is “the syntagmatic ordering in language: patterns, or regularities, in what goes together with what” (Halliday and Matthiessen, 2004, p. 22, emphasis in original), and system is the paradigmatic ordering in language, or “what could go instead of what” (ibid., emphasis in original). However within SFL, priority is given to system and to systemic choices, hence the name systemic in SFL (p. 23). Halliday (2013, p. 21) argues that “a system network is the theoretical representation of a potential, the potential that is inherent in some particular set of circumstances; since a language is a resource for making meaning, I referred to it as a meaning potential” (emphasis in original). Such a system network, may be read either procedurally, with meaning as choosing (in circumstance a, choose either xor yor z), or declaratively, with meaning as choice (in circumstance a, either xor yor

Elaboration Postmodifiers


zis on the cards); ‘system’ is really short for ‘system-&-process’ (Halliday, 2013, p. 21). In the same respect, Fontaine (2013, p. 95) argues that choice might concern either “the system network where choice is seen as the available semantic options (i.e. a term in a system), which involves classification”, i.e. a product, or “the language system where choice is seen as the process of selecting options”. Winograd (1983, p. 289), on the other hand, claims that choice involves classification that “is being imposed by an observer”. Choice, for Halliday, is “more appropriately considered as selection (in a Darwinian sense) rather than as decision” (Fontaine, 2013, p. 97). Tucker (1998, p. 41), however, argues that “a system network displays the choices in meaning which can be made, and which lead to a possible realization at the level of form. What they do not do, and cannot do, is provide the basis for these choices”. This makes us think that choice is “an active process; i.e. choices are made by the speaker—and this almost sounds like ‘deciding’” (Fontaine, 2013, p. 97). Fawcett (2013) points out that “ ‘choice’ implies ‘choosing’ and ‘choosing’ entails ‘deciding’: choosing (like many lexical verbs) is a complex process, involving not only ‘deciding’ but also ‘wanting’ and ‘coming to have’” (in Fontaine, 2013, p. 98). In the same respect, Berry (2013) raises many questions as to whether the choices simply comprise an inventory of options available in a language, or whether humans actually make these choices; whether they make the choices consciously or unconsciously; whether the choices are free or constrained; whether they are collective or individual; whether they are choices of meaning or choices of form or choices of context; and whether choices of meaning are contextually determined or whether they determine the context. She concludes saying that “it is not really a matter of either/or, but a matter of both/and” (366). This chapter presents choice as a kind of process, i.e. the process of opting for some options and not others. 3.2. System Network and Probabilities in SFL According to SFL, each network of system is “potentially being a candidate for quantitative analysis” (Halliday, 2005, p. 95), and the frequency of the entries of this system is translated into probabilities within the system. Any systemic choice is probabilistic in nature where the likelihood of an option being selected is either equiprobable or skewed. These probabilities may be global, i.e. “pertaining to the language as a whole, in all contexts and registers” (ibid.), or local, that is, “particular to one subsystem or text type or even one body of text” (Halliday and Webster, 2009, p. 252). In the text, probabilities are local and may be conditioned by two types of factors: intrastratal conditioning, involving


Imen Ktari

probabilities conditioned by factors internal to the system, and interstratal conditioning, concerned with one stratum affecting probabilities in another. Any system cannot, however, function in isolation but rather in relation to text and context, as will be explained in the next section. 3.3. Text and Context Considering the text as a “window on the system” (Halliday and Matthiessen, 2004, p. 3), SFL puts corpus-based studies at the heart of linguistic investigation since they allow for authentic data, can include spoken language and, more importantly, make it possible “to study grammar in quantitative terms” (Halliday and Matthiessen, 2014, p. 35). The concept of text, however, “is not absolute” (Sellami-Baklouti, 2011, p. 504) because different texts belong to different contexts, defined as “the extralinguistic environment in which language operates, as spoken or written discourse” (Halliday and Webster, 2009, p. 240). Halliday (1985, p. 7) refers to context of situation (“environment of text”; Halliday and Webster, 2009, p. 240) and context of culture (“environment of linguistic system”; ibid.), both being “necessary for the adequate understanding of the text” (Halliday, 1985, p. 7). The relation between text and context is tightly connected, in the sense that “the parameters of the relevant context ‘activate’ or ‘motivate’ the uptake of particular semantic and grammatical choices, and simultaneously we can take the grammatical selections in the text as evidence of the ‘motivational relevancies’ in the context” (Hasan, 1999), in what is referred to as “activation-construal dialectic” (Hasan, 2009, p. 17). This chapter takes the scientific discipline as a context. According to Halliday (2004), the scientific discipline is characterized by certain features such as interlocking definitions (defining a scientific entity through providing a series of definitions that are related) and technical taxonomies or “highly ordered constructions in which every term has a definite functional value” (Halliday, 2004, p. 164). Other characteristics involve special expressions which make up technical grammar, lexical density (the number of lexical words per clause) and syntactic ambiguity, where some expressions are equivocal. Another feature is grammatical metaphor, known as “the transference of the linguistic representation of the semantic components of a situation between different lexicogrammatical categories” (Galve, 1998, p. 369), thus creating tension “between the wording (literal) and meaning (transferred)” (Halliday and Martin, 1993, p. 540). A final feature of the scientific discourse, as argued by Halliday (2004), is semantic discontinuity, “when the writers sometimes make semantic leaps, across which the reader is expected to follow them in order to reach a required conclusion” (p. 177).

Elaboration Postmodifiers


3.4. Genre The scientific discipline, however, is not homogeneous as it consists of AAs and PAs. In fact, although they both “report on the same subjectmatter and in the same field, one should not be misled into assuming that they belong to the same genre or that one or the other is the sub-genre of another” (Nwogu, 1990, p. 354). They are rather “acknowledged as two distinct genres” (De Oliveira and Pagano, 2006, p. 628) that display differences at the level of their communicative purposes and formal features. In this chapter, I will use the term genre and not register for the following reasons. A register is defined as “a variety associated with a particular situation of use (including particular communicative purposes)” (Biber, Conrad and Reppen, 2009, p.6), and is identified “often on the basis of a specific configuration of three contextual factors: field of discourse, mode of discourse and tenor of discourse" (Halliday, McIntosh and Strevens, 1964, p.9), which makes them “restricted to typical configurations of contextual features as realized through different aspects of the grammatical system” (Bhatia, 2004, p. 32). Genres, on the other hand, are said to be “much more action-oriented than registers” (ibid.). Biber et al. (2009) argue that a register would include “three major components: the situational context, the linguistic features, and the functional relationships between the first two components” (p. 6). As such, a register analysis tries to relate these three components in the sense that the linguistic analysis of structures is linked to the situational context, which in turn involves functional features (communicative purposes). The interrelationship between registers, genres and disciplines can be represented visually as shown in Figure 19.1, from Bhatia (2004).





Research Article Text Book


Academic Essay Business




Figure 19.1 The relationship between genre, register and discipline (Bhatia, 2004)


Imen Ktari

AAs and PAs are considered as two distinct genres for while AAs are “specialist-to-specialist” (Varttala, 1999, p. 1), i.e. written by experts researching on a particular field and listed with their institutional affiliations and targeting experts, or in Parkinson and Adendorff’s (2004) terms “the few intelligent”, PAs are often written by a journalist or non-expert who is ‘generic’ rather than ‘specific’ or a ‘real scientist’ (Parkinson and Adendorff, 2004, p. 10). As far as their communicative purposes are concerned, AAs aim at communicating research and scientific ideas to “persuade reader of their knowledge claims” (Myers, 1990, p. 9), making use of a more specialized jargon and a more complicated technical style (Parkinson and Adendorff, 2004). The purpose of PAs, on the other hand, is to “inform and entertain the general public by reporting on new claims not yet endorsed as fact by the community” (Myers, 1990, p. 10), hence the focus on people’s thoughts and claims rather on scientific theories and methods as in the case of AAs (ibid.). As such, PAs “present a sequential narrative of nature” involving the participants rather than the scientific activity (ibid, p. 2). In such a way, they help “maintain a vital relationship between researchers and the general public” (Giannoni, 2008, p. 1) by “correspond[ing] to the needs of less informed groups of readers such as students and laymen” (Varttala, 1999, p. 2). The selection of these two genres is motivated by the fact that such contextual factors affect the choice of elaboration postmodifiers. The empirical part of this study will try to provide some quantitative support for the claim that genre affects such a choice. To this end, some methodological tools are needed, as explained in the following section.

4. Corpus and Methodology The corpus under study is a one million-word-corpus, equally divided between AAs and PAs and taken from different journals and magazines, as shown in Table 19.2. The corpus was annotated using the UAM CorpusTool Software1 (O’Donnell, 2008). This software allows the design of a scheme taking into account the logico-semantic relationship within postmodifiers and genre. The scheme shown in Figure 19.2 has been designed for corpus annotation, considering the variables of the analysis, namely elaboration postmodifiers and genre (AAs and PAs). Probabilities are also considered on the basis of the frequency distributions of the variables concerned. The empirical part relies on significance tests such as the chi-square test, which confirms the correlation between the frequency distribution of the logico-semantic relationship (elaboration mainly) within postmodifiers and genre (AAs and PAs).

5. Corpus Analysis and Findings Accounting for 90 percent, elaboration postmodifiers are much more frequent than extension (0 percent) and enhancement (10 percent) postmodifiers

Elaboration Postmodifiers


Table 19.2 Corpus description




Number of words

Academic Radiology, Acute Pain, Alzheimer’s and Dementia, Cancer Cell, Cardiovascular Pathology, Clinical and Applied Immunology Reviews, Genetics and Evolution, Heart & Lung, Heart Rhythm, Human Pathology, Infection, International Journal of Cardiology, International Journal for Parasitology, Fertility and Sterility, Magnetic Resonance, Metabolism, Microbes and Infection, Microbial Pathogenesis, Molecular and Cellular Endocrinology, NeuroImage, The American Journal of Cardiology, The Journal of Acute and Critical Care, The Lancet New Scientist, New York Times, Scientific American, Time, London Times



Figure 19.2 Coding scheme designed for corpus annotation

within the whole corpus. This pattern is the same across the two genres under investigation, as revealed in Table 19.3. Nonetheless, these frequencies are genre-related as the Chi-square test reveals high significance between the frequency of occurrence of elaboration and the two genres. As already discussed in Section 3.4, AAs display a high frequency of taxonomies and technical words. Some definitions and explanations are thus needed, as illustrated in examples (6), (7) and (8). (6) CBP/p300 are general mediators which bridge the basal transcriptional machinery to the ER-α/co-activator complex, irrespective to which specific protein is presentSRC1, SRC2 or SRC3(AA 8037). (7) In line with previous studies ([Shang and Brown, 2002], [Shang et al., 2000] and [Stossi et al., 2006]), activated ER-αbinds to gene promoters, recruits co-activators or co-repressors which determine the subsequent transcriptional up- or down-regulation, respectively(AA 5811).


Imen Ktari

Table 19.3 Distribution of expansion qualifiers across genres Elaboration

AAs Pas









1,770 1,002

91.6% 88.2%

0 7

0% 0.6%

224 309

8.3% 11.2%

(8) We performed western blot analysis of PFSK-miR-520g cells using pan-PKC- antibodies which identify most Ca2+-dependent phosphoPKC isoforms (PKCα, βI, βII), anti-phospho-CAMKII, as well total and phosphospecific JNK1/2 antibodies(AA 4820). Postmodifiers in examples (6), (7), and (8) are associated with their heads mediators, co-expression and (pan-PKC-) antibodies, respectively, through an identifying relation. Indeed, such words belong to the category of technical words or ‘ultra-specialized’ (Lowe, 2009) terms pertaining to the medical discourse only. The postmodifiers related to these heads act as definitions of them. As such, postmodifiers in examples (6) and (7) could be reworded as general mediators can be defined as those which bridge the basal transcriptional machinery to the ER-α/co-activator complex, irrespective to which specific protein is present SRC1, SRC2andSRC and co-activators or co-repressors are by definition those which determine the subsequent transcriptional up- or down-regulation, respectively. The same applies to example (8), where the head (pan-PKC-) antibodies and its corresponding postmodifier, which identify most Ca2+-dependent phospho-PKC isoforms (PKCα, βI, βII), anti-phospho-CAMKII, as well total and phosphospecific JNK1/2 antibodies, are related through a defining relationship. This is in line with the communicative purposes of AAs, which seek to present and explain scientific issues and where definitions constitute “a feature of scientific and other academic forms of writing” (Wilkins, 1986, p. 53) that “boost argumentation and guarantee understandability in academic discourse” (Triki, 2014, p. 211). Thus, these postmodifiers are “not cases of language embellishment or superfluous items within texts” (ibid.); they rather “guarantee that the antecedent is free of all ambiguity, uncertainty, or obscurity”, and they “guide the reader while trying to understand claims” (ibid.). PAs, on the other hand, display fewer technical words as they are often written by journalists targeting lay people. Examples (9) and (10) of elaboration postmodifiers from PAs follow. (9) This kind of magnetic stimulation has nothing to do with the little bar magnets that arthritis sufferers sometimes wrap around their wrists or attach to their backs. (PA 6027) (10) For women who took up smoking later lung cancer mortality is less rising. Another clearly established risk factor is exposure to

Elaboration Postmodifiers


ionising radiation. This may be responsible for cancers in people living around Chernobyl in Ukraine, Toikamura in Japan and for people working in nuclear power plants. (PA 7359) Postmodifiers in examples (9) and (10) are not used to mainly define the heads to which they are attached as these heads (magnets, women and people) have clear common meanings. These postmodifiers rather specify their referent. In fact, the postmodifier in example (9), that arthritis sufferers sometimes wrap around their wrists or attach to their backs, is used to refine and specify the head magnets in the sense that the writer of this PA is talking about one specific kind of bar magnets which is the one that arthritis sufferers sometimes wrap around their wrists or attach to their backs. Similarly, not all smoking women present less risk for lung cancer, but only those who took up smoking later. Likewise, the writer in example (10) restricts the category of people with high risk for cancer to two main categories that are revealed in the two postmodifiers: the first refers to people living around Chernobyl in Ukraine, Toikamura in Japan and the second to those working in nuclear power plants. This is in line with the generic features of PAs, where a generic and simpler language is often used, as opposed to AAs, where more technical and specialized terms are found, corresponding to the degree of scientific knowledge of their writers and target audience are found. This corresponds to the degree of scientific knowledge of the writers and target audience of PAs and the needs of the latter, being lay people. As such, elaboration postmodifiers in AAs are related to their heads through a defining relationship, while those in PAs through a specifying relationship, restricting the categorization of their heads, which highlights the impact of genre on the choice of elaboration postmodifiers.

6. Concluding Remarks In summary, this chapter has presented a discussion of how the lexico grammatical structure is genre-driven, through an analysis conducted on elaboration postmodifiers in AAs and PAs. The analysis of the corpus of AAs and PAs has revealed two main conclusions. First, based on statistical evidence, it is proved that genre does not really affect the choice of elaboration postmodifiers in terms of frequency (in both AAs and PAs, elaboration postmodifiers are the most frequent), but it does in terms of their implications. This is a result of the disciplinary conventions of the scientific medical writing to which both genres belong. In fact, the main purpose of the scientific discipline is to explain and define scientific issues, which is possible mainly through elaboration. More importantly, the analysis has revealed that elaboration postmodifiers in AAs are linked to the meaning of defining, while in PAs these are associated with the meaning of specifying, which is tightly related to generic variations.


Imen Ktari

An analysis of the other types of logico-semantic relationship (expansion: extension and enhancement as well as projection: locution and idea) will be conducted in a future work to give a broader picture of the SFL perspectives on genre analysis and more precisely on the interaction of lexico grammar and the context. I hope to have shown the impact of genre on the semiotic choice of elaboration postmodifiers in AAs and PAs in the sense that the lexico grammatical choice might mean different things in different genres (Halliday, 2005, p. 51). This confirms that the system could not be studied independently of the text, which could not be studied independently of genre, which is at the heart of SFL. This is very important as it sheds light on the interaction of SFL with and its application to other disciplines such as genre studies, computational linguistics, etc.

Note 1. Version 2.3 beta, downloaded from

References Berry, M. (2013) Towards a study of the differences between formal written English and informal spoken English. In L. Fontaine, T. Barlett and G. O’Grady (Eds.), Systemic Functional Linguistics: Exploring Choice. Cambridge: Cambridge University Press, pp. 365–383. Bhatia, V. K. (2004) Interdiscursivity in critical genre analysis. English for Specific Purposes, 16(3), pp. 181–196. Biber, D., Conrad, S. and Reppen, R. (2009) Corpus Linguistics: Investigating LanguageStructure and Use, 5th ed. New York: Cambridge University Press. Butler, C. (2009) Lexical phenomenon in Functional Discourse Grammar and Systemic Functional Linguistics. In S. Slembrouck, M. Taverniers and M. Van Herreweghe (Eds.), From Will to Well: Studies in Linguistics Offered to AnneMarie Simon-Vanderbergen. Gent: Academia Press, pp. 55–67. De Oliveira, J. M. and Pagano, A. S. (2006) The research article and the science popularization article: A probalistic functional grammar perspective on direct discourse representation. Discourse Studies, 8(5), pp. 627–646. Eggins, S. (2004) An Introduction to Systemic Functional Linguistics, 2nd ed. New York and London: Continuum. Fawcett, R. (1980). Cognitive Linguistics and Social Interaction: Towards an Integrated Model of a Systemic Functional Grammar and the Other Components of an Interacting Mind. Heidelberg: Julius Groos and Exeter University. Fawcett, R. (2000) A Theory of Syntax for Systemic Functional Linguistics. Amsterdam: Benjamins. Fawcett, R. (2013) Choice and choosing in Systemic Functional Grammar: What is it and how is it done? In L. Fontaine, T. Barlett and G. O’Grady (Eds.), Systemic Functional Linguistics: Exploring Choice. Cambridge: Cambridge University Press, pp. 115–134.

Elaboration Postmodifiers


Fontaine, L. (2012). Analysing English Grammar: A Systemic Functional Introduction. Cambridge: Cambridge University Press. Fontaine, L. (2013) Semantic options and complex functions: A recursive view of choice. In L. Fontaine, T. Bartlett and G. O’Grady (Eds.), Systemic Functional Linguistics: Exploring Choice. Cambridge: Cambridge University Press, pp. 95–114. Fontaine, L. (2017) The English nominal group: The centrality of the thing element. In T. Bartlett and G. O’Grady (Eds.), Routledge Systemic Functional Linguistic Handbook. London: Routledge, pp. 267–283. Fries, P. (1970). Tagmeme Sequences in the English noun phrase. In Collins. P and Lee. D (eds). The Clause in English: In Honour of Rodney Huddleston. Amesterdam: John Benjamins, pp. 93–111. Galve, I. G. (1998) The textual interplay of grammatical metaphor on the nominalizations occurring in written medical English. Journal of Pragmatics, 30, pp. 363–385. Giannoni, D. S. (2008) Popularizing features in English journal editorials. English for Specific Purposes, 27, pp. 212–232. Halliday, M. A. K. (1985) Spoken and Written Language. Victoria: Deakin University Press. Halliday, M. A. K. (2003) Introduction: On the ‘architecture’ of human language. In J. Webster (Ed.), On Grammar. London: Continuum, pp. 1–32. Halliday, M. A. K. (2004) The Language of Science. London: Continuum. Halliday, M. A. K. (2005) Towards probabilistic interpretations. In J. Webster (Ed.), Computational and Quantitative Studies. London: Continuum, pp. 42–62. Halliday, M. A. K. (2009) Methods—techniques—problems. In M. A. K. Halliday and J. Webster (Eds.), Continuum Companion to Systemic Functional Linguistics. London: Continuum, pp. 59–86. Halliday, M. A. K. (2013) Meaning as choice. In L. Fontaine, T. Barlett and G. O’Grady (Eds.), Systemic Functional Linguistics: Exploring Choice. Cambridge: Cambridge University Press, pp. 365–383. Halliday, M. A. K. and Martin, J. R. (1993) Writing Science: Literacy and Discursive Power. London: Falmer Press. Halliday, M. A. K. and Matthiessen, M. I. M. (2004) AnIntroduction to Functional Grammar, 3rd ed. Great Britain: Oxford University Press. Halliday, M. A. K. and Matthiessen, M. I. M. (2014) An Introduction to Functional Grammar, 4th ed. London: Arnold. Halliday, M. A. K., McIntosh, A. and Strevens, P. (1964) The Linguistic Sciences andLanguage Teaching. London: Longman. Halliday, M. A. K. and Webster, J. (Eds.). (2009) Continuum Companion to Systemic Functional Linguistics. London and New York: Continuum. Hasan, R. (1999) Speaking with reference to context. In M. Ghadessy (Ed.), Text and Context in Functional Linguistics. Amsterdam: Benjamins, pp. 219–328. Hasan, R. (2009) The place of context in a systemic functional model. In M. A. K. Halliday and J. J. Webster (Eds.), Continuum Companion to Systemic Functional Linguistics. London and New York: Continuum, pp. 166–189. Lowe, I. (2009) Characteristics of the Language of Science. Available at: www. (Accessed on 12 November 2009).


Imen Ktari

Martin, J.R. (1992). English Text: System and Structure. Philadelphia, PA and Amesterdam: John Benjamins. Myers, G. (1990) Writing Biology: The Social Construction of Popular Science. Madison: University of Wisconsin Press. Nwogu, K. N. (1990) Discourse Variation in Medical Texts: Schema, Theme and Cohesion in Professional and Journalistic Accounts. UK: Department of English Studies, University of Nottingham. O’Donnell, M. (2008) The UAM CorpusTool: Software for corpus annotation and exploration. In Carmen M. Bretones Callejas et al. (Eds.), Applied Linguistics Now: Understanding Language and Mind/La Lingüística Aplicada Hoy: Comprendiendo el Lenguaje y la Mente. Almería: Universidad de Almería, pp. 1433–1447. Parkinson, J. and Adendorff, R. (2004) The use of popular science articles in teaching scientific literacy. English for Specific Purposes, 23, pp. 379–396. Pop, A. (2007) Simplicity Versus Complexity of Noun Group Structure in Print Commercial Advertising. Available at: Viewer.aspx?Year=2007&ID=pop (Accessed on 10 October 2009). Quirk, R., Greenbaum, S., Leech, G. and Svartvik, J. (2005) A Comprehensive Grammar ofthe English Language. London: Longman. Sellami-Baklouti, A. (2011) The impact of genre and disciplinary differences on structural choice: Taxis in research article abstracts. Text & Talk, 31(5), pp. 503–523. Triki, N. (2014) Elaboration paradigms in PhD theses introductions. In M. Guirat and M. Triki (Eds.), Deviation (s). Tunisia: Laboratory on Approaches to Discourse (LAD), pp. 202–225. Tucker, G. (1998) The Lexicogrammar of Adjectives: A Systemic Functional Approach to Lexis. London: Cassell Academic. Varttala, T. (1999) Remarks on the communicative functions of hedging in popular scientific and specialist research articles on medicine. English for Specific Purposes, 18, pp. 177–200. Verspoor, M. and Sauter, K. (2000) English Sentence Analysis: An Introductory Course. Amesterdam John Benjamins. Wilkins, D. A. (1986) Notional Syllabuses. Oxford: Oxford University Press. Winograd, T. (1983) Language as a Cognitive Process, Volume 1: Syntax. Reading: Addison-Wesley.

20 A Systemic Functional Analysis of the Use of Personal Reference in the Medrano Burglary Court Hearings Ameni Hlioui 1. Introduction This chapter examines the relationship between genre and the choices of different linguistic features in a text. As Halliday and Matthiessen (2014, p. 23) advocate, in Systemic Functional Linguistics (SFL), language is seen as “a resource for making meaning, and meaning resides in systemic patterns of choice”. It follows from this that identifying these patterns through identifiable linguistic features will provide some understanding of the meaning of a text. The primary concern of this study is to examine the choice of personal reference in the legal genre. This investigation will be based on the court hearings of the Medrano burglary trial at two levels: the use of personal reference and the use of different transitivity configurations of processes and participants to construct different competing narratives of the same version of truth. These linguistic features will be seen as evidence of choices which are affected by the social dimension of personal pronouns. As Pennebaker (2011, p. 90) states, personal pronouns can be thought of as “the most social of all word categories”. After many years of research on courtroom discourse, Pennebaker (2011) found that personal reference can reflect the social status of the different participants taking part in a court setting as well as the credibility of their versions of truth. The social status of these participants will affect the personal reference with which they will address each other, and the different versions of the same story each speaker is trying to put forth will result in different choices of processes and different participant roles assigned to this personal reference. Within SFL, Hasan (2013, p. 298) argues that “choice with system and realisation is the main instrument in achieving the SFL objective of describing language as a meaning potential”. The analysis of the corpus in this study will involve both the experiential metafunction, through which people negotiate their own configuration of what happened in the case setting, and the interpersonal metafunction, through which “users of language establish, negotiate and assume their positions in social

382 Ameni Hlioui relationships” (Halliday, 1994, p. 68). Linking these two metafunctions is the context variable of tenor. This variable is related to the various kinds of relationships holding between participants in a situation. These different relationships can be explained by the different institutional backgrounds which give participants different status roles.

2. Background 2.1. The Rationale Behind the Choice of Personal Reference Quirk, Greenbaum, Leech and Svartvik (2005, pp. 335–336) note that personal pronouns represent the most important central class of pronouns because of their frequency and grammatical characteristics. These pronouns are used for specific reference. Indeed, first and second person pronouns are used to refer to those directly involved in the discourse situation, i.e. the speaker/writer and the addressee (p. 347), while third person pronouns “refer to third parties not directly involved in the origination or reception of the utterance in which they occur” (p. 340). Finally, the non-personal pronoun it is typically used to refer to inanimate objects, abstractions and whole clauses and sentences (p. 348). This study will investigate this referential property to identify who are the participants directly involved in the discourse events and those who are absent or absented from them. Moreover, according to Quirk et al. (2005, pp. 341–354), personal pronouns have many other specific uses like ‘prop a