CONDUCTING GENRE‑BASED

RESEARCH IN APPLIED
LINGUISTICS

This collection is a comprehensive resource on conducting research

in applied linguistics involving written genres that is distinctive in its
coverage of a multiplicity of interdisciplinary perspectives.
The volume explores the central approaches, methodologies, analyses,
and tools used in conducting genre-based research, extending the
traditional focus on a single framework for defining genres by explicating
the major approaches that have been invoked in applied linguistics.
Chapters address a mix of commonly used methodologies (e.g., case
studies, ethnographic approaches), types of analyses (e.g., metadiscourse,
rhetorical move-step analysis, multidimensional analysis, lexical bundles
and phrase frames, CALF measures, multimodal analysis), and studies that
focus on other areas of second language (L2) teaching and learning (e.g.,
multilingualism, the Teaching and Learning Cycle). Taken together, the
volume provides a theoretically and methodologically diverse introduction
to foundational topics in genre-related research, supported by detailed
discussions of the challenges and practical considerations to take into
account when conducting research involving written genres.
This book is a valuable resource for graduate students, faculty, and
researchers in applied linguistics, particularly those working in second
language acquisition, L2 writing, and genre theory and pedagogy.

Matt Kessler is Assistant Professor of Applied Linguistics at the University

of South Florida, where he teaches in the MA Applied Linguistics and
PhD Linguistics and Applied Language Studies programs. Matt’s research
focuses on issues pertaining to second language writing, genre-based
teaching and learning, and computer-assisted language learning. His work
has appeared in journals such as the Journal of Second Language Writing,
Language Learning, System, TESOL Quarterly, and more.

Charlene Polio is Professor in Linguistics, Languages, and Cultures at

Michigan State University, where she teaches in the MA TESOL and PhD
Second Language Studies programs. Her research focuses on methods
used in second language writing research, and she is the co-author of
Understanding, Evaluating, and Conducting Second Language Writing
Research (Routledge). She is also the co-editor of the journal TESOL
Quarterly.
 Second Language Acquisition Research Series
Susan M. Gass and Alison Mackey, Series Editors
Kimberly L. Geeslin, Associate Editor

The Second Language Acquisition Research Series presents and explores

issues bearing directly on theory construction and/or research methods
in the study of second language acquisition. Its titles (both authored and
edited volumes) provide thorough and timely overviews of high-interest
topics, and include key discussions of existing research findings and their
implications. A special emphasis of the series is reflected in the volumes
dealing with specific data collection methods or instruments. Each of these
volumes addresses the kinds of research questions for which the method/
instrument is best suited, offers extended description of its use, and outlines
the problems associated with its use. The volumes in this series will be
invaluable to students and scholars alike, and perfect for use in courses on
research methodology and in individual research.

Communicative Competence in a Second Language

Theory, Method, and Applications
Edited by Matthew Kanwit and Megan Solon

Second Language Acquisition and Lifelong Learning

Simone E. Pfenninger, Julia Festman, and David Singleton

The Role of the Learner in Task-Based Language Teaching

Theory and Research
Edited by Craig Lambert, Scott Aubrey, and Gavin Bui

Practice and Automatization in Second Language Research

Perspectives from Skill Acquisition Theory and Cognitive Psychology
Edited by Yuichi Suzuki

Conducting Genre-Based Research in Applied Linguistics

A Methodological Guide
Edited by Matt Kessler and Charlene Polio

For more information about this series, please visit: www​.routledge​.com​/Second-Langu​

age-A​cquis​ition​-Rese​a rch-​Serie​s/boo​k-ser​ies/L​E ASLA​R S
CONDUCTING GENRE-
BASED RESEARCH IN
APPLIED LINGUISTICS
A Methodological Guide

Edited by Matt Kessler and Charlene Polio

Designed cover image: © Getty Images | enjoynz
First published 2024
by Routledge
605 Third Avenue, New York, NY 10158
and by Routledge
4 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2024 selection and editorial matter, Matt Kessler and Charlene Polio; individual
chapters, the contributors
The right of Matt Kessler and Charlene Polio 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.
ISBN: 978-1-032-29284-7 (hbk)
ISBN: 978-1-032-29282-3 (pbk)
ISBN: 978-1-003-30084-7 (ebk)
DOI: 10.4324/9781003300847
Typeset in Galliard
by Deanta Global Publishing Services, Chennai, India
CONTENTS

About the Editors ix

About The Contributing Authors x

1 Introduction 1
Matt Kessler, Charlene Polio

SECTION I
Research methods 11

2 Case studies 13
Raffaella Negretti, Baraa Khuder

3 Ethnographic Research 35
Christine M. Tardy

SECTION II
Analyses, tools, and topics 57

4 Metadiscourse 59
Ken Hyland

5 Rhetorical move-step analysis 82

J. Elliott Casal, Matt Kessler


viii Contents

6 Lexical bundles and phrase frames 105

Viviana Cortes

7 Multidimensional analysis 127

Larissa Goulart, Shelley Staples

8 Complexity, accuracy, and fluency (CALF) measures 149

Charlene Polio, Hyung-Jo Yoon

9 Multimodal analysis 172

Larissa D’Angelo, Francesca Marino

SECTION III
Focus on the learner and learning 197

10 Multilingualism and multicompetence 199

Bruna Sommer-Farias

11 Systemic functional linguistics and the (expanded)

teaching and learning cycle 221
Kathryn Accurso, Jennifer Walsh Marr

Index 245
ABOUT THE EDITORS

Matt Kessler is Assistant Professor of Applied Linguistics at the University

of South Florida, where he teaches in the MA Applied Linguistics and
PhD Linguistics and Applied Language Studies programs. Matt’s research
focuses on issues pertaining to second language writing, genre-based teach-
ing and learning, and computer-assisted language learning. His work has
appeared in journals such as Journal of Second Language Writing, Language
Learning, System, TESOL Quarterly, and more.
OrcID: 0000-0001-5264-0059

Charlene Polio is Professor in Linguistics, Languages, and Cultures at

Michigan State University, where she teaches in the MA TESOL and
PhD Second Language Studies programs. Her research focuses on meth-
ods used in second language writing research, and she is the co-author
of Understanding, Evaluating, and Conducting Second Language Writing
Research (publisher: Routledge). She is also the co-editor of the journal
TESOL Quarterly.
OrcID: 0000-0003-2319-7478


ABOUT THE CONTRIBUTING
AUTHORS

Kathryn Accurso is Assistant Professor of Teaching Language and

Literacies at the University of British Columbia. She is an applied linguist
and former English teacher who uses SFL as part of a justice-oriented agenda
to support K-12 teachers in developing knowledge, beliefs, and practices for
reimagining and teaching disciplinary literacies in multilingual contexts.
OrcID: 0000-0002-1351-994X

J. Elliott Casal is Assistant Professor of Applied Linguistics at the University

of Memphis. His research interests include corpus linguistics, second lan-
guage writing, and English for Specific/Academic Purposes. His recent
work appears in Journal of Second Language Writing, Journal of English for
Academic Purposes, and Language Learning and Technology.
OrcID: 0000-0002-8920-9120

Viviana Cortes is Associate Professor in the Department of Applied

Linguistics and ESL at Georgia State University. She holds a PhD from
Northern Arizona University. Her research interests are corpus-based dis-
course analysis, formulaic language, English for Specific Purposes (ESP),
and descriptive grammar.
OrcID: 0000-0003-4704-779X

Larissa D’Angelo is Associate Professor of English Language and

Translation at the Department of Foreign Languages, Literatures and
Cultures of the University of Bergamo, and Head of the UNIBG Eye-
Tracking Lab. Her main research interests deal with biometric analyses,


 About The Contributing Authors xi

multimodality, Audio Visual Translation (AVT), and the translation of

children’s and young adult literature.
OrcID: 0000-0003-3565-6280

Larissa Goulart is Assistant Professor of Linguistics at Montclair State

University. Her research interest lies in corpus-based approaches to register
variation, academic discourse, and second language writing. She received
her PhD in Applied Linguistics from Northern Arizona University.
OrcID: 0000-0002-4190-4589

Ken Hyland is Honorary Professor at the University of East Anglia. He

has published 280 articles and 29 books on writing and academic discourse
with 77,000 citations on Google Scholar. The Stanford/Elsevier analysis
of the Scopus database shows he is the world’s most influential scholar in
language and linguistics. A collection of his work, The Essential Hyland,
was published in 2018 by Bloomsbury.
OrcID: 0000-0002-4727-8355

Baraa Khuder is Post-Doctoral Researcher at Chalmers University of

Technology in the Department of Learning and Communication. She has a
PhD in Applied Linguistics from Birkbeck, University of London, UK. Her
research focuses on students’ and academics’ interdisciplinary collaborative
writing and writing for publication.
OrcID: 0000-0003-2272-2189

Francesca Marino is PhD student in the Linguistics and Applied Language

Studies program at the University of South Florida. Her research interests
include multimodal discourse analysis and digital multimodal composing.
Francesca’s work has appeared in journals such as ELT Journal, Multimodal
Communication, and other venues.
OrcID: 0000-0003-2515-5642

Raffaella Negretti is Professor in Educational Psychology and Applied

Linguistics at Chalmers University of Technology in the Department of
Communication and Learning in Science. Her research spans academic
writing, metacognition, and genre pedagogy, appearing in many promi-
nent international journals. Raffaella currently serves as associate editor of
the Journal of Second Language Writing.
OrcID: 0000-0003-1948-1775

Bruna Sommer-Farias is Assistant Professor in the Master of Arts in

Foreign Language Teaching program at Michigan State University. Her
xii About The Contributing Authors

research focuses on genre and corpus-informed approaches to curriculum

and instruction, language teacher development, and multilingualism.
OrcID: 0000-0002-2239-937X

Shelley Staples is Associate Professor at the University of Arizona. Her

research focuses on the use of corpus-based discourse analysis to investigate
language use across spoken and written contexts. In addition, her research
aims to inform language teaching and assessment, particularly in the areas
of English for Academic and Specific Purposes (EAP/ESP).
OrcID: 0000-0002-5927-4431

Christine M. Tardy is Professor of English Applied Linguistics at the

University of Arizona, where she teaches undergraduate and graduate stu-
dents in TESOL, applied linguistics, and writing. Her work appears in
numerous journals, edited books, and monographs, including Genre-Based
Writing: What Every ESL Teacher Should Know (University of Michigan
Press, 2019).
OrcID: 0000-0002-1692-0368

Jennifer Walsh Marr is an educational linguist and lecturer at University

of British Columbia Vantage College. She draws on critical pedagogy and
SFL-informed discourse analysis to help make implicit language and cul-
tural patterns more explicit and accessible to both instructors and students.
OrcID: 0000-0002-1688-4707

Hyung-Jo Yoon is Associate Professor in the Department of Linguistics/

TESL at California State University, Northridge. His research interests
include second language writing, task-based language teaching, and lan-
guage assessment, with a focus on building effective pedagogy for second
language development. He also conducts research related to computational
text analysis.
OrcID: 0000-0002-8553-3708
1
INTRODUCTION
Matt Kessler, Charlene Polio

People often think of genres as types of literature, such as mystery, fantasy,

or science fiction, or movie types such as comedies, dramas, or documen-
taries. In daily life, however, we can think of genres by the essential roles
they play, serving as the foundation for many people’s personal, academic,
and professional activities. Various written genres – the focus of this vol-
ume – range from street signs, medical bills, wedding invitations, research
papers, to conference abstracts, and more. They are oftentimes both dis-
tinct and easily identifiable, representing recurring forms of communica-
tion that convey specific messages to a target audience. As Martin (2009)
said, “You cannot not mean genres” (p. 13). In other words, any form of
meaningful communication is a genre and vice versa. We might contrast
genre-based writing with a classroom writing activity such as Write a para-
graph containing five past tense verbs. Due to the prominent role genres play
in communication, as a concept, genre has been remarkably influential in
first (L1) and second language (L2) composition.
Notably, despite the seemingly simple and straightforward nature of
the term genre, its conceptualization has been anything but. In fact, the
majority of genre-based scholarship and pedagogy to date have been under-
stood in accordance with three genre approaches, which have sometimes
been referred to synonymously as genre schools, theories, or traditions (e.g.,
Gebhard & Harman, 2011; Hyon, 2018; Paltridge, 2014; Rose, 2013).
These different approaches – the term we use throughout this introduc-
tory chapter – include: English for Specific Purposes, Systemic Functional
Linguistics, and the New Rhetoric/Rhetorical Genre Studies approach.
Crucially, each of these approaches conceives of genres and the activities

DOI: 10.4324/9781003300847-1
2 Matt Kessler, Charlene Polio

that surround them in distinct ways. Subsequently, their differences have

resulted in a diverse array of research aims, conceptual frameworks, and
pedagogical recommendations for classroom teachers. Because these three
approaches have served as the foundation for a sizable amount of genre-
based research and teaching (including many of the topics discussed in this
volume), we briefly describe them before further outlining the aims and
scope of this edited volume.

1.1 The three genre-based approaches (ESP, SFL, and RGS)

For teachers and researchers in applied linguistics and L2 writing, the first
and perhaps most well-known genre-based approach is English for Specific
Purposes (ESP). As an approach, ESP emerged from the works of John
Swales (1990, 2004). In fact, Swales is typically credited with first introduc-
ing the concept of genre to applied linguistics through his transformative
work, which analyzed the rhetoric of academic research article introduc-
tions (see Swales, 1981 and a later reprint in 2011). Since that time, both
Swales and others (e.g., Vijay Bhatia, Ken Hyland, and Ann Johns) have
been instrumental in developing ESP into a fully formed approach to genre.
In ESP, genre has often been understood according to a definition put
forward by Swales (1990), in which he defined genres as “a class of com-
municative events, the members of which share some set of communicative
purposes. These purposes are recognized by the expert members of the
parent discourse community, and thereby constitute the rationale for the
genre” (p. 58). Typically, an ESP approach to genre focuses on both the
rhetorical moves and the language used to realize those moves. For exam-
ple, a simple move in an email to a professor might be a salutation, but
depending on the relationship one has with the professor, the move could
be expressed as Dear Prof. Polio or Hi Charlene. Of course, these moves are
more complex in an academic publication, for example. So typically, genre
analysis is accomplished by analyzing a text’s rhetorical moves, lexico-gram-
matical features, and attempting to understand these features in connec-
tion with their discourse community or rhetorical context (Hyland, 2008;
Paltridge, 2014). The ESP approach has been highly influential in the field
of language teaching with both studies of target texts, (which both L1 and/
or L2 writers hope to produce), and studies of learner texts to highlight
areas that might need instruction.
The second influential genre-based approach is Systemic Functional
Linguistics (SFL). Notably, much of SFL’s early use and development
occurred in the context of K-12 classrooms in Australia, so it has some-
times been referred to as the Sydney School of genre (Hyon, 1996). As
an approach, SFL emerged from the work of M. A. K. Halliday (1978),
 Introduction 3

although more wide-ranging and divergent discussions appeared later as

well (e.g., Halliday & Matthiessen, 2004; Martin, 1984; Martin & Rose,
2008). In SFL, the social context in which the genre is produced is of the
utmost importance. This is not to say, however, that context is not impor-
tant in ESP scholarship, but ESP researchers traditionally have tended to
focus more so on the text-based features of the genre being studied. In
terms of SFL, Halliday (1978) understood human communication/mean-
ing and social contexts as being intimately intertwined, and he conceived
of meaning as being structured along three dimensions, including what he
called ideational, interpersonal, and textual meanings. When it comes to
the production of genres, Halliday and Matthiessen (2004) explain that
these dimensions account for the linguistic choices that writers make when
expressing different constructs or ideas (ideational), how writers interact
with their readers by taking on different roles and establishing relationships
(interpersonal), and how the text is ultimately constructed using cohesive
devices and specific discourse sequences (textual). Returning to the email
example from earlier, the email writer’s opening/salutation (i.e., whether
it is more formal or informal) might establish the writer’s relationship with
the professor (interpersonal); the first full sentence in the body paragraph
might convey different meanings depending on the intent of the email,
such as making a request (ideational); and, specific types of vocabulary
or grammatical constructions might also be used depending on whether
the writer is making a request, submitting a document for the professor’s
review, and so on (textual).
It is important to note here that within SFL scholarship, there are vari-
ous sub-communities who see genre somewhat differently, and these differ-
ences primarily pertain to understanding the relationship between genres
and their contexts. For instance, in Halliday’s account of SFL, genre is
part of a broader meaning-making resource system, in which text is only
one of many modes or resources at a writer’s disposal (e.g., aural, visual,
and gestural modes can also be leveraged). Conversely, other SFL scholars
place more of an emphasis on aspects pertaining to the linguistic (i.e., writ-
ten or oral) register patterns used by people when communicating (e.g.,
Martin, 1992). For readers interested in learning more about the complexi-
ties of SFL, we refer them to Chapter 11 (see Accurso and Walsh Marr,
this volume). Despite these subtle differences, as an approach, SFL is some-
what similar to ESP in that its proponents often adopt a form of genre
analysis as a means for describing the structure and language of written
genres. Specifically, SFL practitioners have typically adopted a pedagogical
approach to teaching genres called the Teaching and Learning Cycle (TLC).
This cycle includes three (or more) stages, including the teacher mod-
eling/deconstructing the target genre; the teacher and his/her students
4 Matt Kessler, Charlene Polio

collaboratively constructing the genre; and finally, the students producing

the genre independently (e.g., Caplan & Farling, 2016; Gebhard, 2019). Of
course, ESP-oriented practitioners also integrate aspects of the TLC, but
in SFL approaches, it is forefronted. Importantly, in both research and in
practice, SFL tends to place more of an emphasis on understanding genres
in tandem with the cultural context in which they are produced. In other
words, writers produce genres in ways that reflect their specific cultural
context and rhetorical situation. Thus, SFL views genres less so as fixed
rhetorical patterns, (thus making it distinct from ESP), and instead, as a
series of dynamic choices that are made by people in response to a particu-
lar situation.
The third and final genre-based approach that has garnered much atten-
tion is Rhetorical Genre Studies (RGS). Until more recently, RGS has fre-
quently been referred to as New Rhetoric and/or North American New
Rhetoric (e.g., Hyon, 1996; Paltridge, 2004; Wang, 2019). Primarily
responsible for RGS’s emergence is Carolyn Miller (1984), although other
scholars such as Bazerman (1988, 1994) and Freedman and Medway
(1994) have also been influential in its development. Somewhat similar to
SFL, RGS places significant emphasis on the social situations and contexts
in which genres are produced. In particular, Miller (1984) argued that gen-
res “serve as keys to understanding how to participate in the actions of a
community” (p. 165). For instance, when learning how to write an email
request, write a recommendation letter, or how to give an oral academic
presentation, we are not simply learning a generic form that exists in iso-
lation. Instead, we are learning how to take on various roles in different
social situations, and subsequently, this enables us to better understand
the different situations in which we may find ourselves within our society.
Thus, RGS sees genres as playing an important role by fostering communi-
cation among humans, and even helping to create a shared reality.
Because of the overt focus on the context and social actions of the tar-
get genres, RGS proponents like Miller and others (e.g., Aviva Freedman)
have often been interested in investigating aspects of the text in relation
to the locale or place, how a text may function in a different context, and
how genres may change over time (see Paltridge, 2004). In addition, the
focus is often on how a text is constructed. As such, researchers may use
ethnographic methods to study how texts are created (Wang, 2019). For
example, Wang (2016) explored the intertextual practices (writing from
sources) of Chinese ESL students both through text analysis, and surveys
and interviews.
Unlike ESP and SFL, when it comes to classroom pedagogy, RGS has
tended to be much less prescriptive in terms of its recommendations. That
is, RGS does not have anything similar to ESP’s focus on rhetorical moves
 Introduction 5

or akin to SFL’s TLC. Perhaps most notably, among some RGS propo-
nents, there have even been conflicting views about how explicitly genres
should be taught in a classroom since the classroom context often does not
accurately represent the intended social context in which the genre will be
used (see Bawarshi & Reiff, 2010). That being said, when genres are explic-
itly taught in the classroom, RGS practitioners may focus on helping learn-
ers become more aware of the context and rhetorical situation the genre is
responding to, along with how features of the text may be shaped by the
context itself (e.g., Coe, 1994; Devitt, 2004). For example, when using
RGS, Freedman (1999) notes that teachers and their students may act as
researchers to investigate the complexities of a given genre’s context. When
doing so, students will be prompted to pay particular attention to features
of the context in which it is produced (e.g., where they might find the
genre, who the audience is, and why the audience would be interested in
it). Students might also engage in what Johns (2013) calls “textual analysis
through a contextual lens” (p. 3), analyzing why certain linguistic choices
were made depending on the aforementioned contextual factors.
As shown, while there is some overlap among the three genre-based
approaches, there are also clear differences in the ways that ESP, SFL,
and RGS each define and conceptualize genre. These differences may be
summed up in reference to three main areas, which pertain to: (1) the vary-
ing degrees of stress that each approach places on understanding text exter-
nal factors such as the social context and the construction of texts; (2) the
different frameworks each approach uses for making sense of text-related
features (e.g., rhetorical moves versus ideational, interpersonal, and textual
meanings); and finally, (3) the varying degrees to which each approach sug-
gests specific pedagogical activities in the classroom, including activities
that prompt students to focus on different elements throughout the com-
posing process.

1.2 Volume aims, audience, and chapter previews

As outlined in the previous section, the field of applied linguistics has seen
a substantial amount of genre-based research over the past few decades,
and a large portion of this scholarship has been motivated by ESP, SFL,
or RGS. Crucially, Tardy (2016) has noted that the goal of much genre-
based research is to demystify the practices of academic and professional
communities, thereby providing novices with an entry point for engaging
in unfamiliar practices. However, for novice researchers such as graduate
students and early-career scholars who may be new to genre research, there
are currently no existing texts that provide an introduction to methods for
studying genre. As such, the current volume attempts to fill this gap in the
6 Matt Kessler, Charlene Polio

literature, serving as a guidebook to researching written genres in applied

linguistics.
The goal of this collection is to provide practical advice, with a particu-
lar focus on the methodological how to aspect of conducting genre-based
research. In terms of audience, this book is intended for researchers in the
field of applied linguistics, but it also may be of particular interest to those
in fields such as education, L1 composition studies, or related areas. When
we conceived of this volume, we specifically had in mind graduate stu-
dents (at both the master’s and doctoral levels), in addition to early-career
faculty who might be interested in expanding their research repertoire.
Because genre-based research has oftentimes been highly interdisciplinary
and varied in its approach with genre being understood differently based
on the approach/school/theory adopted, this book does not take a single
approach to operationalizing and defining genre. Instead, each chapter cov-
ers a different topic, namely an approach or a tool, and the authors of each
chapter discuss how different approaches such as ESP, SFL, and/or RGS
have informed research on that topic.
Following this chapter (i.e., Chapter 1, the Introduction), this volume is
divided into three main sections. These sections include: Section I (Research
Methods), Section II (Analyses, Tools, and Topics), and lastly, Section III
(Focus on the Learner and Learning). In Section I titled Research Methods,
there are two chapters, which cover established qualitative research meth-
odologies that have been prominent in explorations of written genres. The
two chapters in this section include Chapter 2 on the use of Case Studies
(authored by Raffaella Negretti and Baraa Khuder), and Chapter 3 on
Ethnographic Research (Christine Tardy). Notably, each of these chapters
touches upon the specific methods and tools that have been used by genre
researchers when adopting case studies (e.g., interviews) and ethnographic
methods (e.g., longitudinal observations, the collection of artifacts).
Although Section I covers research methodologies that are purely quali-
tative in nature, in Section II, titled Analyses, Tools, and Topics, there are six
chapters that adopt a greater diversity of methods, including quantitative,
qualitative, and mixed methods. Unlike Section I, which covers holistic
methodologies used in genre-based scholarship, Section II shifts the focus
to some of the most prominent topics and types of analyses that researchers
have used when investigating written genres. Notably, some of the analy-
ses in this section have been used in keeping with a specific genre-based
approach (e.g., rhetorical move-step analysis with ESP), while others have
been combined with multiple approaches for different purposes. In the
chapters for Section II, there is also a focus on the different automated or
digital tools that researchers have made use of when conducting various
analyses. The chapters in Section II include: Chapter 4 on investigations
 Introduction 7

of Metadiscourse (Ken Hyland), Chapter 5 on the use of Rhetorical Move-

Step Analysis (J. Elliott Casal and Matt Kessler), Chapter 6 on stud-
ies that analyze Lexical Bundles and Phrase Frames (Viviana Cortes),
Chapter 7 on Multidimensional Analysis (Larissa Goulart and Shelley
Staples), Chapter 8 on the adoption of Complexity, Accuracy, and Fluency
Measures (Charlene Polio and Hyung-Jo Yoon), and Chapter 9, which cov-
ers Multimodal Analysis (Larissa D’Angelo and Francesca Marino).
The final section of this volume, Section III, addresses scholarship with
a Focus on the Learner and Learning. As the title of this section suggests,
there has been a large body of scholarship that has tended to focus on
aspects of the learners themselves, particularly, by focusing on L2/mul-
tilingual learners as they engage with different written genres. Many of
these studies have broadly explored issues related to the types of instruction
learners have received, and subsequently, how learning progresses and is
affected by a host of factors (e.g., learners’ L1s, existing genre knowledge,
and more). The final two chapters of this volume address these areas, with
Chapter 10 covering the topic of Multilingualism and Multicompetence
(Bruna Sommer-Farias), and Chapter 11 covering the topic of Systemic
Functional Linguistics and the (Expanded) Teaching and Learning Cycle
(Kathryn Accurso and Jennifer Walsh Marr).

1.3 Organizational structure of the chapters

Although each of the ten content chapters (i.e., Chapters 2–11) addresses
different research methods, tools, or topics pertaining to genre-based
research, each chapter is organized similarly. In particular, each chapter
follows a similar internal structure that consists of seven sections, which are
titled: (1) Introduction to the approach and definition of genre, (2) Goals, (3)
Common research methods, (4) Example studies, (5) Issues and challenges, (6)
Study-in-focus, and (7) Future research directions.
Each chapter begins with a section called Introduction to the approach
and definition of genre. In this opening section, chapter authors provide
important information about the history and origins of the method, analy-
sis, or topic. When doing so, the authors explain how genre has typically
been defined by those who have adopted the method/analysis, including
whether the method/analysis is related to a particular genre-based approach
(i.e., ESP, SFL, or RGS), or if it cuts across multiple genre theories. In
this section, information is also provided about the extent to which the
method/analysis has been applied to different text types (e.g., academic
writing, multimodal genres), and the characteristics of the writers them-
selves (e.g., whether the texts were produced by experts or novices, L1 writ-
ers or L2/multilingual writers).
8 Matt Kessler, Charlene Polio

The second section in each chapter is called Goals. This section builds
on section one by further explicating the typical aims or objectives of the
chapter’s method, analysis, or topic. When describing the goals, the authors
also provide readers with a series of sample (but general) research questions
that previous researchers have asked in past studies.
Sections III and IV are perhaps the most significant and largest sections
within each chapter. Section III is titled Common research methods, and
this is where the chapter’s authors describe how researchers have gone about
adopting the specific method or analysis. When doing so, the authors share
the different types of tools researchers typically use when engaging in the
method/approach, along with any specific steps that are often taken when
conducting the analyses. Importantly, in this section, the chapter’s authors
explain how (or the extent to which) the research is related to other methods
or analyses covered in this volume. Immediately following this is Section
IV, which provides readers with Example studies. As the name suggests, in
this section, the authors showcase a series of five to seven exemplary studies
involving the chosen method, analysis, or topic. These example studies are
described in the text, and they also appear in a formatted table that breaks
down key information about each study’s context, research questions, data
collected, and comments on the study’s design.
Following the example studies section, Section V covers Issues and chal-
lenges. Since this edited volume is primarily intended for graduate students
and early-career researchers, this section provides an account of the various
challenges researchers might encounter or need to consider when conduct-
ing research on the topic or when adopting the method/analysis. For each
chapter, this consists of authors discussing three to four key issues and chal-
lenges that are important to be aware of.
The sixth section is called Study-in-focus. Although multiple example
studies are provided for readers in Section IV, in this section, the authors
highlight one typical and well-designed study that is not included in the
earlier table of example studies. When showcasing the study-in-focus, the
study appears in a callout text box, which includes information about the
study’s research questions, context and population, design and procedure,
and findings. After readers review a summary of the study-in-focus, the
chapter’s authors take readers through the study and explain the meth-
odological choices that the researchers had to make. This also includes
explanations of why/what makes it an exemplary study, along with how
the researchers dealt with or attempted to address any common issues and
challenges.
The seventh and final section of each chapter covers Future research
directions. As the name suggests, in this section, the chapter’s authors
make a series of suggestions for future scholars who may be interested in
 Introduction 9

conducting genre-based research adopting the method, analysis, or on the

topic covered in the chapter. When doing so, the authors provide read-
ers with descriptions of three to four specific directions, and they include
example research questions when able.

1.4 Conclusion
In closing, we hope that this edited volume will serve as a useful guide for
those graduate students and early career scholars who are interested in con-
ducting genre-based research in the future. Importantly, we note here that
there are undoubtedly different methods, analyses, and topics that have
been omitted from this volume. However, this text is specifically meant
to showcase the most prominent methods, analyses, and topics that have
been investigated in the field of applied linguistics. If an analysis or topic
has been omitted, it was not done so on ideological grounds. Rather, it was
done so because we feel that the topics covered in this book are reflective of
the major publication trends in applied linguistics scholarship over the past
few decades. As such, we hope that this volume will serve as a useful start-
ing point for readers who wish to gain a foundational understanding of key
areas involved in genre-based scholarship. We close by encouraging readers
to further investigate the various academic journals, books, and resources
discussed by the contributing authors in their respective chapters.

References
Bawarshi, A. S., & Reiff, M. J. (2010). Genre: An introduction to history, theory,
research, and pedagogy. Parlor Press.
Bazerman, C. (1988). Shaping written knowledge: The genre and activity of the
experimental article in science. University of Wisconsin Press.
Bazerman, C. (1994). Systems of genres and the enactment of social intentions.
In A. Freedman & P. Medway (Eds.), Genre and the new rhetoric (pp. 79–101).
Taylor & Francis.
Caplan, N. A., & Farling, M. (2016). A dozen heads are better than one:
Collaborative writing in genre-based pedagogy. TESOL Journal, 8(3), 564–
581. https://doi​.org​/10​.1002​/tesj​.287
Coe, R. M. (1994). Teaching genre as process. In A. Freedman & P. Medway
(Eds.), Learning and teaching genre (pp. 157–169). Boynton/Cook.
Devitt, A. J. (2004). Writing genres. Southern Illinois University Press.
Freedman, A. (1999). Beyond the text: Towards understanding the teaching and
learning of genres. TESOL Quarterly, 33(4), 764–767. https://doi​.org​/10​.2307​
/3587890
Freedman, A., & Medway, P. (Eds.). (1994). Genre in the new rhetoric. Routledge.
Gebhard, M. (2019). Teaching and researching ELLs’ disciplinary literacies: System
functional linguistics in action in the context of U.S. school reform. Routledge.
10 Matt Kessler, Charlene Polio

Gebhard, M., & Harman, R. (2011). Reconsidering genre theory in K-12 schools:
A response to school reforms in the United States. Journal of Second Language
Writing, 20(1), 45–55. https://doi​.org​/10​.1016​/j​.jslw​.2010​.12​.007
Halliday, M. A. K. (1978). Language as social semiotic: The social interpretation of
language and meaning. Edward Arnold.
Halliday, M. A. K., & Matthiessen, C. M. (2004). Halliday’s introduction to
functional grammar (3rd ed.). Routledge.
Hyland, K. (2008). Genre and academic writing in the disciplines. Language
Teaching, 41(4), 543–562. https://doi​.org​/10​.1017​/S0261444808005235
Hyon, S. (1996). Genre in three traditions: Implications for ESL. TESOL Quarterly,
30(4), 693–722. https://doi​.org​/10​.2307​/3587930
Hyon, S. (2018). Introducing genre and English for specific purposes. Routledge.
Johns, A. M. (2013). Rhetorical genre studies (RGS) and the teaching of EFL
reading and writing-or genre is much more than text structure. In CELC
symposium bridging research and pedagogy (pp. 1–7).
Martin, J. R. (1984). Language, register, and genre. In F. Christie (Ed.), Language
studies: Children’s writing: Reader (pp. 21–30). Deakin University Press.
Martin, J. R. (1992). English text: System and structure. John Benjamins.
Martin, J. R. (2009). Genre and language learning: A social semiotic perspective.
Linguistics and Education, 20(1), 10–21. https://doi​.org​/10​.1016​/j​.linged​
.2009​.01​.003
Martin, J. R., & Rose, D. (2008). Genre relations: Mapping culture. Equinox.
Miller, C. R. (1984). Genre as social action. Quarterly Journal of Speech, 70(2),
151–167. https://doi​.org​/10​.1080​/00335638409383686
Paltridge, B. (2004). Academic writing. Language Teaching, 37(2), 87–105.
https://doi​.org​/10​.1017​/S0261444804002216
Paltridge, B. (2014). Genre and second-language academic writing. Language
Teaching, 47(3), 303–318. https://doi​.org​/10​.1017​/S0261444814000068
Rose, D. (2013). Genre in the Sydney school. In J. Gee & M. Hanford (Eds.), The
Routledge handbook of discourse analysis (pp. 209–225). Routledge.
Swales, J. (1981 [2011]). Aspects of article introductions. University of Michigan
Press.
Swales, J. (1990). Genre analysis: English in academic and research settings.
Cambridge University Press.
Swales, J. M. (2004). Research genres: Exploration and applications. Cambridge
University Press.
Tardy, C. M. (2016). Beyond convention: Genre innovation in academic writing.
University of Michigan Press.
Wang, W. (2016). Intertextual practices in academic writing by Chinese ESL
students. Applied Linguistics Review, 7(1), 53–72.
Wang, W. (2019). Text analysis. In J. McKinley & H. Rose (Eds.), The Routledge
handbook of research methods in applied linguistics (pp. 453–463). Routledge.
SECTION I

Research methods


2
CASE STUDIES
Raffaella Negretti, Baraa Khuder

2.1 Introduction to the approach and definition of genre

The use of case studies in the social sciences and specifically in second
language teaching has reflected the social turn in the field and the shift in
the theoretical framework from focusing on cognition to identity theories
and socialization. This acknowledgment of individual differences and reluc-
tance to view learners as masses encouraged the use of case studies among
multilingual writers (Polio & Friedman, 2016). The purpose of a case study
is to investigate in all its complexity a specific singular case or multiple cases
under a certain set of typical or atypical circumstances. In genre research,
case studies have been applied primarily to study how single students or
groups of students learn to produce a genre under specific circumstances
and/or over time (e.g., Casanave, 2010; Kessler, 2021; Tardy, 2005), or to
study the learning generated by genre-based approaches to teaching, such
as explicit instruction (e.g., Huang, 2014) and use of genre-based materials
in subject teaching (e.g., De Oliveira & Lan, 2014). More rarely, the case
has been the genre itself, with the aim of documenting the process that
leads to its production (e.g., McGrath, 2016).
What characterizes then a case study approach? Because case studies are
a well-established methodology beyond genre research, we start first with
some general considerations about this methodological approach. The key
characteristic defining an inquiry as a case study is the focus on a case,
delineated by clear boundaries – its “casing” (Ragin & Becker, 1992, p.
217). The underlying assumption is that the case, and the specific circum-
stances in which it is examined, can provide in-depth insights that have the
potential to apply to similar sets of cases and circumstances, or alternatively
DOI: 10.4324/9781003300847-3
14 Raffaella Negretti, Baraa Khuder

provide an answer to a specific question. In simple terms, we could say that

a case study is an example of some sort, from which we can learn something.
Therefore, case studies can illustrate how a theory may apply to real situ-
ations, and conversely how lived experiences and authentic situations may
problematize and add complexity to theoretical assumptions.
Therefore, the delimitation of the case is an important starting point in
case studies: a clear description of the unit of study and its specific charac-
teristics, as well as a clear description of the circumstances under which it
is studied which make it interesting to investigate (Yin, 2009). That said,
there are two main approaches to case studies in research which correspond
to different views about research paradigms. In the interpretive/social con-
structivist paradigm (Merriam, 2009; Stake, 1995), there is often a per-
sonal interaction (transaction) between the researcher, often the teacher,
and the case (e.g., Huang, 2014); whereas in a post-positivist paradigm
(e.g., Eisenhardt, 1989; Flyvbjerg, 2011; Yin, 2003, 2009), the case is stud-
ied through a clearly defined initial protocol where all the characteristics of
the case are described and possibly even measured (e.g., McGrath, 2016).
Nevertheless, case studies are versatile in nature and entail collecting a vari-
ety of evidence, often over time, that results in a rich and complex collec-
tion of data about the case (Flyvbjerg, 2011).
In genre research, case studies have been used primarily to investigate
how individual or small clusters of students develop genre knowledge and
learn to write (e.g., Cheng, 2008; Guo, 2019; Kessler, 2021; Negretti
& Kuteeva, 2011; Wang, 2020), what impact different pedagogies and
teacher practices have (e.g., Huang, 2014; Troyan, 2016), and how spe-
cific genres are produced, including co-authoring processes and language
practices, both L1 and L2 (e.g., Flowerdew & Wan, 2010; McGrath, 2016;
Townley & Jones, 2016). Genre, in these studies, is usually defined fluidly
as a communicative event which a social group recognizes as typical for
a situation, with a typical purpose, and which therefore often presents a
set of discursive and linguistic features that are more or less conventional
(Miller, 1984).
In general, the definition of the case and the focus of the study – its
specific circumstances – are influenced by the tradition of genre-based
research that the researchers align themselves with. Studies with a focus on
genre learning, especially in L2, have fallen under the English for Specific
Purposes (ESP) umbrella (Swales, 1990), which places particular focus on
rhetorical characteristics of genres and the language that characterizes typi-
cal rhetorical moves and their variation (see for example Cheng, 2008).
Case studies of texts as examples of genre – which fall under genre analy-
sis – also adopt an ESP/LSP (Language for Specific Purposes) lens (e.g.,
Townley & Jones, 2016). Case studies with a focus on classroom practice
 Case studies 15

sometimes adopt Systemic Functional Linguistics (SFL) theories (Martin,

1992; see also Pessoa et al., 2017) where genres tend to be described based
on discourse functions. Case studies that explore literacy practices and
social identity are likely to be inspired by academic literacy theories – AcLits
(Lillis & Harrington, 2015; Lillis & Scott, 2007), which aim to study aca-
demic writing development in its sociohistorical context, with special atten-
tion to imbalances of power and narratives of dominance surrounding L2
writers (e.g., Casanave, 2003; Harwood & Petrić, 2017).

2.2 Goals
In terms of goals, in broad terms case studies in genre research aim to
understand the development of genre knowledge (Tardy et al., 2020) as
well as the social and individual circumstances that contribute to this devel-
opment (e.g., Casanave, 2010; Kobayashi & Rinnert, 2013). As such, case
studies are used to investigate research questions such as:

• How do different learners develop an increasingly sophisticated genre

knowledge over time, across genres, and/or languages?
• How do different learners develop knowledge of the same academic or
professional genre over time, given the same context?
• What is the impact of a genre-based instructional approach on a stu-
dent’s or a classroom’s development of writing skills?
• How do learners approach similar genres in different languages? Does
transfer of genre knowledge occur in these cases?
• How does a teacher use genre-based pedagogy in their practice and why?
(with possible extensions to the impact on students’ learning)
• How do sociohistorical and contextual circumstances affect the lived
experience of a writer(s) in the production of a genre?
• What are the social and collaborative activities that lead to the produc-
tion of a specific genre? (With a specific text/genre example as a case.)

2.3 Common research methods

Case study research is not confined to one type of data collection method
(e.g., interviews, observations, writing logs, think-alouds), but the
researcher has the flexibility to adopt various qualitative and quantitative
approaches, as long as there is richness in the data, (i.e., data is collected
either from various resources, for a period of time, and/or on various stages
in the process). As clearly put by Flyvbjerg (2011), “If you choose to do a
case study, you are therefore not so much making a methodological choice
as a choice of what it is to be studied [and why]” (p. 301).
16 Raffaella Negretti, Baraa Khuder

We make a distinction here between the use of case studies as a method

and as a methodology (drawing on Lillis’, 2008 distinction on the use of
ethnography as a method, methodology, and deep theorizing). Case stud-
ies are used as a method when a single method is used in data collection and
analysis (e.g., McGrath, 2016) and as a methodology when several methods
are used, such as interviews, text histories, questionnaires, and observations
to study a case.
A useful starting point to think about case study methods in genre
research, especially in relation to writing, is Casanave’s (2003) distinction
of cases as either a) written products (e.g., McGrath, 2016); b) writing pro-
cesses (De Oliveira & Lan, 2014), and c) writer (or teacher) identity (e.g.,
Tardy, 2005). This distinction can highlight the description and delimita-
tion of the case that is adopted in different studies which is often influenced
by the genre theories that researchers ascribe to, and may help in under-
standing the motivation behind specific data collection methods used in
the case study. Thus, what distinguishes case study research from other
methodological approaches, especially in genre research, is the focus of
the study on a specific case (which might be a person, a text, a process of
production, or a method of instruction), and the approach to the inter-
pretation of the evidence where, for example, researchers might consider
avoiding bias in data analysis by adopting triangulation.
Triangulation of methods of data collection and/or analysis is an impor-
tant procedure in case studies, since it provides confirmation of findings
which increases data trustworthiness and provides additional perspectives
about the case, resulting in a better understanding of complex phenomena
(e.g., De Oliveira & Lan, 2014). Denzin (1978) and Patton (1999) identi-
fied four types of triangulation: (a) method triangulation, where several
methods can be used to either ensure the trustworthiness of data or to
collect more information (e.g., using textual analysis of drafts as well as
interviews); (b) investigator triangulation, where more than one researcher
is involved in the research process as a whole (e.g., agreeing on theories
used, data collection methods, independent data analysis); (c) theory tri-
angulation, via using different theories to interpret the same data set (e.g.,
combing genre theory and identity theory, as in our example study); and
(d) data source triangulation, via using the same method to collect data
from various sources (e.g., interviewing all co-authors involved in one text).
As has been pointed out earlier, case studies do not fall under the umbrella
of a specific research paradigm. Rather, case studies can be considered as a
bridge between various paradigms (Luck et al., 2006) where one does not
have to use the traditional methodological, ontological, and epistemologi-
cal divides but can move freely on the paradigms’ continuum or use a mix
of paradigms.
 Case studies 17

Nevertheless, case studies, especially in genre studies, have typically been

associated with post-positivist, constructivist, critical, and interpretive meth-
odologies (Denzin & Lincoln, 2011) limiting the focus to understanding
social and contextual dynamics that surround a case, by adopting primarily
qualitative approaches to data collection, and missing on the opportunity to
investigate cases as larger entities. For example, instead of focusing mainly
on a person or a text in genre studies, researchers might seek to widen their
case and focus on the impact of learning a genre on the ranking of an edu-
cational institute by making the institute the case. Some of the data collec-
tion methods that are used in case study research, for instance, interviews of
different types and observations, are also common in other approaches such
as ethnography (see Tardy, this volume), especially when the focus of the
investigation is the experience of learners. Having said that, case studies are
versatile in nature and can include “a palette of methods” (Stake, 1995, pp.
xi–xii), or one single method of data collection used systematically.
Case studies can also be used as a part of a larger study, where a specific
case is studied in depth, opening the space for answering research questions
about the larger context (e.g., De Oliveira & Lan, 2014). Another possibil-
ity to approach a case study is by using a multiple case study design, which is
a newer trend to track issues of typicality and variation (Duff & Anderson,
2016). There are two ways to use the multiple case study: either by look-
ing for differences and/or similarities via cross-case analysis and aiming for
generalizability, thus, adopting a positivistic approach (Yin, 2003). This
can be done, for example, by investigating the impact of genre analysis
instruction on less and more successful writers. The second option to use
a multiple case study design is to understand the phenomenon through
the experiences of multiple cases – hence, similarities and differences are
not accounted for (Stake, 1995). It should be noted that in longitudinal
multiple case studies, not all the cases necessarily last the same length of
time (e.g., one and two years in the two cases in Tardy, 2005). Additionally,
participants do not even have to share the same profile: Tardy’s partici-
pants were one PhD and one master’s student in two different disciplines.
Regardless of how the case study is incorporated (e.g., single case study,
multiple case study, single case study as part of a larger study), data col-
lection might last for a varied amount of time depending on the focus of
the study, for example, at the end of the course (e.g., Huang, 2014) or be
retrospective in nature by collecting drafts of a text (e.g., McGrath, 2016).
As can be seen, designing a case study requires a careful design and well-
motivated choices. Hence, researchers’ reflexivity, where researchers acknowl-
edge their role in the research process and critically reflect on the various
phases of the research design and interpretation, is highly recommended in
case studies. We suggest the following exercise, which is presented in Table 2.1
18 Raffaella Negretti, Baraa Khuder

TABLE 2.1 
Case study design: reflective exercise

Theoretical Basis + RQs Study Design Methods of Data

Collection and Analysis

• What theories • Why is this specific • Which methods are

underpin my study? case worthy of best suited to collect
• Which specific studying? data that can answer
concepts will be • What will I focus my RQs?
tested/examined in on as a case? (e.g., a • Does the case
the case study? specific genre and its motivate unique types
• Why is a case study production; a process of data collection
the approach of choice of learning, teaching methods? Which and
in relation to these practices, social why?
theories/concepts? dimension of genre • Will data collection
• What can be learned such as situated or be continuous timed?
from this case individual practices) What motivates this
or cases? Is it an • What will delimit my timeline?
application or real- case, and what makes • What will be the
world example of a it typical/atypical role of the researcher
specific theory or to other cases in the in the case study?
concept? If so, what is same context? What advantages,
the specific question • Which relevant details disadvantages, and
that the case study do I need to include ethical implications
may answer? about my case in does this role entail?
order to describe it • Would a single or a
sufficiently in relation multiple case study
to the research design be more
question (e.g., age, suitable to answer the
first language etc.)? RQs?

2.4 Example studies
To better illustrate how case studies have been conducted in genre research,
in this section we present a selection of relevant studies that offer a vari-
ety of useful examples (see Table 2.2). The studies below show what we
think are useful examples in genre analysis on the use of case study as a
method, via using a single method (here textual analysis) (i.e., McGrath,
2016), multiple case study design (i.e., Tardy, 2005), reflexivity in conduct-
ing case studies (i.e., Casanave, 2010), the researcher as a teacher provid-
ing interventions (i.e., Huang, 2014), conducting trustworthiness checks in
case studies (i.e., De Oliveira & Lan, 2014), and various levels of reporting
a case study (i.e., Kessler, 2021). Importantly, these studies also highlight
a myriad of data that researchers might opt to collect, depending on the
focus of their study.
TABLE 2.2 
Example studies using case study design

Authors Context Research Questions (RQs) Data Collected Comments on Study

Design

Casanave (2010) • The cases are three • How did three PhD • Email interviews • Researcher was
PhD students writing students learn how to for two years; the participants’ supervisor
in a new genre write in a new genre in researcher’s reflective • Genre defined loosely
a context that does not memos as writing a qualitative
support this genre? dissertation
• Important reflections
on the positioning of
the researcher as the
teacher
De Oliveira and • The case is an • How did a science • Class observations; • The study uses SFL
Lan (2014) L2 student in an teacher use genre- drafts of texts • The case is
elementary school based materials in her produced before and representative of the
developing his teaching? after instruction class
science writing • How did the teacher • Explaining the
during a genre-based interact with the background of the
course students? researcher who
• How does the focal conducted inter-rater
case develop his writing reliability checks
during the course?
(Continued)
Case studies
19
TABLE 2.2 Continued
20

Authors Context Research Questions (RQs) Data Collected Comments on Study

Design
Huang (2014) • The case is a student • What is the impact of • Pre- and post-course • Capturing development
in a genre-based explicit teaching on interviews; class in genre knowledge
writing course for genre knowledge? interactions; text drafts over time via data
publication triangulation
• Presentation of rich
interview and textual
data throughout the
text
Raffaella Negretti, Baraa Khuder

Kessler (2021) • The cases are six L2 • To what extent do the • Background • Focusing on a
English learners in case study participants questionnaire; office professional genre
a Master of Laws develop genre memos; three semi- using metacognition
program learning to awareness of the office structured interviews theory
write a professional memo over a semester? and two modified
legal genre (the In which area(s) does stimulated recalls
office memo) over a that awareness develop?
15-week period • To what extent are
there variations in
students’ individual
trajectories in terms
of developing genre
awareness?
McGrath (2016) • The case is a • Does the study of a • 659 comments posted • Focusing on the
research blog in pure collaborative research in response to five blog process of producing a
mathematics blog shed light on entries genre
behind-the-scenes • The blog has
activities pertaining to comments from non-
writing for publication experts, whose input
in pure mathematics, are rarely investigated
and the research article in literature
genre?
Tardy (2005) • The cases are one • How did two students • Interviews; text drafts • Multiple case study
master’s and one develop their rhetorical of various genres; design
PhD student writing knowledge over time? supervisors’ feedback • Uses ACL as its
in various genres framework
Case studies
21
22 Raffaella Negretti, Baraa Khuder

One way to make use of case studies is to uncover behind-the-scenes

processes regarding how genres are written. For example, McGrath (2016),
in an exploratory case study, investigated how writing a research article was
negotiated via an open-access research blog in pure mathematics, making
the focal case a single research blog. The dataset consisted of 659 thread
comments in response to five blog posts by expert and non-expert blog
users who took part in the discussions. Tracking the comments over five
blog posts reflects part of the article writing process in pure mathemat-
ics where the creation of knowledge is often a public endeavor. The use of
a case study allowed McGrath to provide a detailed explanation regard-
ing how decisions about genre writing are made; these decisions are rarely
explicitly discussed. McGrath analyzed the data both inductively and
deductively, allowing for the use of previous research as well as for any
emerging themes and she ensured data trustworthiness by conducting
researcher triangulation where part of the data was analyzed independently
by another researcher and agreement was reached among both researchers.
These findings provide valuable insights into genre teaching. For exam-
ple, one of the comments on the collaborative blog was how students
rarely have the chance to peek at how negotiations happen. Thus, McGrath
encourages students to take part and observe such blogs. As for differentiat-
ing her study from other ethnographic and ethnographically oriented stud-
ies, McGrath duly reports how her study is not actually longitudinal and
does not track a text history over the whole publication period, but rather,
taps into part of the revision process using comments from a blog. This
shows the flexibility of the case study design which allowed the researcher
to give contextual insights but with limited data.
While McGrath (2016) used a single case study as a method, Tardy (2005)
used a multiple case study as a methodology by focusing on two students.
Tardy investigated rhetorical knowledge development in advanced literacy
by collecting interviews and text drafts of various writing genres: while
for one participant this included assignments, homework, job applications,
and reviews, the other participant was only involved in academic writing
genres such as academic proposals and conference papers. The excellent
use of multiple case study design allowed Tardy (2005) to draw on differ-
ences and similarities between the cases to answer a question about how
two students developed their rhetorical knowledge. Similar to the other
reviewed studies in this chapter, Tardy provided detailed information about
the participants. However, she uses the AcLits framework, which stresses
the importance of the sociocultural background of learners. She provides a
full picture not only of both participants’ rhetorical background knowledge
but also of their aspirations and perspectives on the importance of writing
to achieve their future goals. This essential information makes tracking
 Case studies 23

changes in both students’ conceptual as well as textual development more

feasible for the reader. Tardy reports how views on writing changed from
providing the reader with information (i.e., knowledge telling), to selling
ideas to the reader (i.e., knowledge transformation). The findings highlight
the importance of mentoring relationships in developing genre knowledge.
This finding is echoed in Casanave (2010), who examined how three
female PhD students wrote their qualitative-research dissertations in a
department that promotes quantitative research. Casanave investigated
learning a genre under specific circumstances and used the cases to high-
light the dimension of change in a multiple case study design. The study
provides a detailed description of the three cases, covering gender, age, and
similarities and differences across cases in relation to their academic writing
and English language experiences. Casanave also provides information on
her relationship with the participants, the amount of interaction they had,
and the reason for being part of the study, which makes this study nota-
ble in that it illustrates the importance of reflexivity in data collection and
analysis. This study also shows the wide range of data collection methods
that researchers can use when they have a specific relationship with the par-
ticipants. Data collection included email interviews and consultations with
the participants over two years. This was supplemented by the researcher’s
own reflective memos, written after reading and interacting with the three
participants.
Rather than following a specific framework or model of genre analy-
sis, this article provides perspectives on deductive methods for conduct-
ing genre analysis, which might be relevant for researchers working with
emerging genres. Casanave’s reflections toward the end on her positioning
as a mentor and how this impacted her teaching practices, enhances how we
think about case study design and the conclusions that could be drawn from
it. We also see some reflections on how taking part in the study changed
the participants’ perspectives on writing and sensitized them toward their
writing processes.
Similar to Casanave, Huang (2014) was also the participant’s teacher.
Huang investigated how explicit instruction during a 12-week genre-based
writing course impacted a Taiwanese PhD student’s genre knowledge. The
researcher provides a detailed explanation of the case: the participant's
motivation to learn about writing, discipline, and previous writing experi-
ences in terms of both process and product. The data included the student’s
oral interaction data and the multiple drafts of a text intended for publica-
tion. Data were analyzed using Tardy’s (2009) model of genre knowledge:
formal, process, rhetorical, and subject-matter knowledge. The researcher,
an outsider to the participant’s discipline, was unable to capture the sub-
ject-matter knowledge development.
24 Raffaella Negretti, Baraa Khuder

The findings show that the student’s genre knowledge developed in rela-
tion to formal (linguistic and structural features of a genre), process (the
composing process in which a genre is undertaken), and rhetorical (lan-
guage use that helps writers achieve their intended purposes) aspects of
writing for publication due to explicit instruction. The researcher reflects
on how the duration of the course impacted the quality of development.
This study sets an excellent example of data presentation in a case study,
where Huang successfully shows the richness of the data collected by pro-
viding clear and concise examples.
As explained earlier, case studies can be part of a larger study. De
Oliveira and Lan (2014) is an example of this. Their study aimed to identify
instructional practices for teaching upper elementary L2 English learners to
write school-based genres. The study’s research questions focused on three
issues: the implementation of genre-based pedagogy by a science teacher,
the teacher’s interactions with the students, and the writing development
of one student (the main case) as a result of this interaction. De Oliveira
and Lan used a case study design to answer the question related to how a
typical student in the classroom develops his writing in a course that uses
genre-based pedagogy informed by SFL, focusing on one L2 writer. This
L2 writer is reported to be representative of the classroom. Researchers
identified him as typical based on the science teacher’s views about the par-
ticipant’s struggles and challenges during L2 writing.
The researchers in this study report in detail how they ensured accounta-
bility, a trustworthiness criterion, of their findings (Lincoln & Guba, 1985)
by conducting intercoder reliability checks and asking another researcher
familiar with SFL to independently code parts of the interview data.
Findings show that the focal participant’s L2 writing developed in specific
areas relevant to the teaching he was exposed to, for example, the incor-
poration of more appropriate lexical-grammatical words such as discipline-
specific vocabulary.
The last example we provide is Kessler (2021). Kessler investigated how
six Masters of Law students developed awareness of a genre collectively
and individually by examining the development of two cases during a
15-week course that focused on teaching a professional legal genre (the
office memo). In this study, Kessler responded to an overall call for using
metacognition theory, which is learners’ conscious awareness of three types
of knowledge: declarative knowledge (what we know); procedural knowl-
edge (how to apply what we know); and conditional knowledge (the reason
the knowledge is relevant to the current learning situation) (see Negretti &
McGrath, 2018). Kessler used a background questionnaire, office memos,
three semi-structured interviews, and two modified stimulated recalls.
Thus, data were triangulated using several research methods.
 Case studies 25

The findings of this study show that the development of the various
types of metacognition knowledge varied across participants. A closer look
at two cases showed that these differences are highly related to previous
learning experiences; that is, although both cases did not have prior knowl-
edge of the office memo genre, one of them was mainly trained in writing
for tests, while the other participant’s work as a research assistant exposed
him to various law genres and he was able to transfer that knowledge to the
office memo genre.

2.5 Issues and challenges

In this section, we address three prevalent issues in the studies we reviewed:
alignment of citations with the study design, the relationship of the case
to its context, and the lack of or minimal discussion of reflexivity. When
conducting case studies, researchers often face several issues. The first issue
regards the fact that some of the case studies we described earlier cite case
study approaches that do not reflect the method taken in the study, specifi-
cally in relation to when in the research process the research question(s) are
determined. For example, McGrath (2016) cited Stake (2008), whereas her
research question is predetermined, thus, a reference to Yin (2003) would
have been more appropriate. This issue seems to emerge also in ethno-
graphic studies: for example, Lillis and Curry (2018) cite references to case
studies but refer to their methods as ethnography. Additionally, in some of
the articles we listed, case study citations are missing entirely, although the
authors quite clearly and explicitly adopt a case study approach. Several of
the studies presented above do not include any reference to case study (e.g.,
Tardy, 2005), and some researchers might go as far as not even dedicating a
methods section in their report of a case study (e.g., Casanave, 2010). Our
recommendation is thus for researchers aiming to embark on a case study to
clearly connect their approach to the methodology literature available and
clearly motivate their choices (see the “Study in-focus” in the next section).
Secondly, another less prevalent concern in some studies is the relation
of the case to its context. In other words, a detailed description of what is
common and what is particular about the case seems to be missing (e.g.,
Huang, 2014). To avoid that, a full description of the case and whether
it is an exemplar or outlier should be explicitly provided: this means for
example including details about what makes the selected case unique in any
way, or on the other hand if the case is a typical example of what could be
expected in the context of the study, for example, teachers are often able
to identify what a typical struggling student in their class would be (e.g.,
Huang, 2014). Clearly stating the reasons for choosing a case is important,
since contextual boundaries and additional relevant contextual information
26 Raffaella Negretti, Baraa Khuder

(physical, political, geographic) offer a comprehensive understanding of the

conditions surrounding the case.
Another potential issue is the lack of or minimal discussion of reflexivity
in case study publications. Explicitly articulating the rationale for the case
study design, including interpretation (see Table 2.1) and reflecting on the
researcher’s positioning in relation to the participants, would strengthen
the research. For example, Huang (2014) reflected on the fact that he was
an outsider to the focal discipline (as we reported earlier); however, we
know little about how being the participant’s teacher impacted data col-
lection and analysis, such as the power relation between the researcher and
the participant. Information on researchers’ positioning and relation to the
case improves our understanding of the data collection and analysis proce-
dures and how researchers obtained access to the case (e.g., Kobayashi &
Rinnert, 2013).
In relation to data collection, reflexivity means questioning and weigh-
ing carefully each methodological choice (Khuder & Petrić, 2021): is the
sampling conducted out of convenience or purpose? For example, Huang’s
sampling was a convenience sampling. Khuder and Petrić (2021) encourage
researchers to adopt methods that include close engagement with the par-
ticipants to conduct a high level of reflexivity that involves both reflexivity
on the research process as well as the product (i.e., the research text pro-
duced and its wording). This reflexivity work ensures the trustworthiness of
the research process. For instance, even though De Oliveira and Lan (2014)
conducted trustworthiness checks to some level, given the nature of their
data, a higher level of trustworthiness would have been achieved had the
authors explained why they chose a specific case and specific methods of
data collection. As an entry point to reflexivity, we recommend our exercise
in Table 2.1.

2.6 Study in-focus
As an exemplary study, we chose Kobayashi and Rinnert (2013), which
tracks the writing development of a multilingual writer (the case) longi-
tudinally. What makes this study stand out is the exemplary way in which
the case study methodology is used to illuminate existing theories on mul-
tilingual transfer, genre knowledge, and identity, as well as the original-
ity and variety of the methods used for data collection and analysis. The
development of linguistic competence across different languages in bi- and
multilingual speakers has become increasingly relevant in applied linguis-
tics after the bilingual turn (see Ortega & Carson, 2010) in the study of
second language acquisition, posing that a multicompetence perspective,
rather than a contrastive/deficit perspective, could better serve the study
 Case studies 27

of how people develop abilities in more than one language. Specifically in

writing, theories and research have strongly suggested that writing com-
petence comprises several components always in development, including
genre knowledge (Tardy, 2009), which cut across and are transferred across
languages. Multicompetence theory, as adopted in this study, thus describes
the development of writing skills in multilinguals as dynamic and fluid,
changing with experiences, with soft boundaries across languages (Cenoz
& Gorter, 2011). In addition to multicompetence theory, Kobayashi and
Rinnert (2013) use theories of genre as social practice and identity to pro-
vide a comprehensive view of how the development of writing competence
occurs over time in the three languages used by their multilingual partici-
pant, Natsu. Before proceeding further with our commentary, we urge the
reader to review Rinnert and Kobayashi’s study in the call-out box.

Citation
Kobayashi, H., & Rinnert, C. (2013). L1/L2/L3 writing development:
Longitudinal case study of a Japanese multicompetent writer. Journal of
Second Language Writing, 22(1), 4–33.

Research questions
• How do L1 and L2 writing by a multicompetent writer change over
two and a half years in terms of linguistic development (fluency, sen-
tence length, and lexical diversity) and text construction (choices of text
features)?
• What similarities and differences are there in the multicompetent writ-
er’s text construction and composing processes across L1, L2, and L3?
• How are individual and social factors, particularly attitude and identity,
related to the development of L1/L2/L3 writing?

Context and population

The “case” is Natsu, a multicompetent writer who can write in three lan-
guages: Japanese (L1), English (L2), and Chinese (L3). Natsu was chosen
because of her unique experiences of learning both an L2 and a foreign lan-
guage, having lived and studied high school for three years in an L2 con-
text in Australia, and having been an exchange student for a year to learn
Chinese. Although Natsu is the main case, an interesting feature of this study
is that her data is compared with data from five groups of experienced and
28 Raffaella Negretti, Baraa Khuder

inexperienced writers from a previous study. The genre under scrutiny is the
argumentative essay as used in the major English language proficiency tests
(e.g., IELTS).

Procedure
To trace Natsu’s development as a writer over time and across three lan-
guages, data collection occurred at two points in time with two and five
years in between them and comprised both quantitative and qualitative data.
These multiple data sources correspond to the three theories adopted in the
study, about text production and composing processes across languages,
genre knowledge, and author identity/attitude. In Period 1, data comprised
essays written in L1 and L2, a written questionnaire, and in-depth, semi-
structured interviews after the composition of each essay. In Period 2, Natsu
was asked to write essays in L1, L2, and L3, in-depth interviews after each
essay, as well as a final comprehensive interview. Additionally, retrospec-
tive stimulated recalls were used after each essay. Essay writing was time-
constrained (60 minutes, paper/pencil). For text production and composing
processes, the authors focused on language output measures, text features,
and pauses in the composing process, complementing their interpretation
of these measures through interview data. Natsu’s text production was com-
pared with data from other student groups in earlier studies.

Findings
Several themes and patterns of interference across the three languages were
detected in Natsu’s development. In terms of L1 and L2 linguistic develop-
ment over time, at Period 1, Natsu’s writing proficiency in L2 English far
exceeded that of the comparison student groups (novices and returnees),
while her proficiency in L1 Japanese writing was surprisingly low. In Period
2, Natsu’s L2 writing proficiency was still much higher than the comparison
student groups’ (expert writers), while her writing proficiency in L1 caught
up with that of the other undergraduate writers. Interviews revealed that her
high English proficiency was explained in part by the three years of L2 high-
school literacy, combined with her intense training to pass the proficiency
exams required to study abroad (TOEIC and IELTS). Interestingly, Natsu admit-
ted in interviews that she applied academic writing knowledge obtained in
L2 to writing in L1. Natsu also established a personal authorial identity in her
texts in L1, L2, and L3. Overall, findings revealed several overlaps in her L1, L2,
and L3 writing, suggesting that Natsu’s writing knowledge over time became
 Case studies 29

merged, but also included awareness of several unique linguistic and stylistic
differences related to the socio-rhetorical dimensions of the genre and the
readers.
Regarding composing processes in L1, L2, and L3, while Natsu’s overall
composing time was approximately the same for all three languages, findings
showed that different processes (planning, formulating, and re-formulating)
took different times during the process, due to consideration of genre and
identity attached to each language. These dimensions of genre knowledge
and identity emerged clearly from the interviews, which overall illuminated
how deliberate, stylistic choices across languages suggested a developing
meta-knowledge of genre across languages and an authorial self.

As mentioned, in addition to the solid theoretical grounding, a key

strength of this study is the originality of the data collection procedure
and the variety of the data collection methods employed, comprising both
qualitative and quantitative measures. An interesting feature to highlight
is that while this is a longitudinal study, it is not an ethnographic study:
the authors did not continuously follow the case over time (fieldwork)
nor provide a “thick, ethnographic description” (Kobayashi & Rinnert,
2013, p. 26) of the social context in which Natsu developed her skills,
but rather collected data at two points in time combining naturalistically
occurring data with interviews and quantitative data. Nevertheless, it suc-
ceeds in providing a comprehensive and complex picture of the partici-
pant’s writing development in three languages, by triangulating in a clever
way measures of text fluency and composing with data that provides the
participant’s own perspective, such as the stimulated recall sessions and
in-depth interviews.
Overall, a key characteristic that makes this study exemplary is the
care with which each data collection measure and procedure for analysis
was chosen, considering the specific case and the research questions. For
instance, measures of writing proficiency were adapted to be suitable for
English and Japanese. As mentioned, in order to provide a sense of com-
parison, this quantitative data from the case study participant was com-
pared with data obtained from other students, showing that a multitude of
data sources can be used in case studies if and when they are justified by
the research question. The adoption of stimulated recall sessions to discuss
pauses in composing was another creative method, providing both quanti-
tative, tangible data about the process (in the form of pauses) and an insider
explanation of the process that illuminated many interesting inter-language
30 Raffaella Negretti, Baraa Khuder

influences as well as unique reasons for the composing process associated

with each language. Finally, and consistently with Kobayashi and Rinnert’s
cognitive and sociocultural approach, the in-depth interviews right after
writing the essays revealed the participant’s unique history, attitudes, and
experiences of learning to write in the three languages, shedding light on
how motivation, goals, genre knowledge, personal authorial identity, and
experience interact in a multilingual writer’s thinking when composing in
either of their languages.
The case study methodology adopted by Kobayashi and Rinnert, while
limited to the argumentative time-constrained test essay genre, made a
strong contribution to multicompetence theory, by showing that the
development of writing competence is not only multifaceted but is still
ongoing both in L1 and L2 (and in part in L3) in multilingual writers,
with writing knowledge (genre knowledge) and authorial identity merg-
ing and refining as time and experiences accrue. Their study illustrates
how assumptions of L1 dominance are poorly suited for understanding
the writing competence of multilingual writers, where transfer may occur
from any language that the learner perceives as closer to their identity,
has recency, or provides some level of similarity with the language used.
Notably, this transfer is not necessarily unconscious but rather suggests
a merged, hybrid, and fluid meta-knowledge of writing as posed by
Gentil (2011), for instance of genres, their socio-rhetorical dimensions
across languages, and the organizational and discoursal patterns that are
shareable.

2.7 Future research directions

Here we present possible directions for research using case studies that
we think would be productive for genre research and pedagogy. First, it
would be interesting to investigate the value of case studies as a pedagogical
tool. While case studies are primarily a research approach, they have been
used successfully for teaching in various fields (Andersen & Schiano, 2014;
Christensen & Hansen, 1987). In genre research, published case studies
can be used to illustrate genre production and variation (e.g., Khuder &
Petrić, 2022b), and could thus be useful both in the genre classroom and
for teacher training. Some examples of application to genre pedagogy: nov-
ice writers could focus on a specific learner genre, typically assigned in
university courses at their institution, and use each other as cases, investi-
gating and discussing similarities and differences in task perceptions and
writing process. More advanced students, especially at the doctoral level,
could use case studies both as a methodology and as a way to learn about
 Case studies 31

behind-the-scenes processes in relation to high-stakes genres (e.g., grants)

or ­unfamiliar genres. Case studies provide the opportunity for learners
to develop context-dependent knowledge that over time builds expertise
(Flyvberg, 2011).
This focus on building concrete and contextually dependent knowl-
edge dovetails nicely with the aims of genre pedagogy: to help learners
develop an understanding of the contextually situated nature of genres, as
manifested in linguistic and rhetorical choices, as well as a broader meta-
knowledge of how to recontextualize and transfer genre knowledge across
situations (cf. Tardy et al., 2020). While rhetorical move-step analysis (see
Casal & Kessler, this volume) as a pedagogy typically requires the compari-
son of several examples of a genre (or genres), learners may also benefit from
the richness and depth that is to be gained by focusing more closely on a
case, to appreciate how the writing process interacts with the contextual
and socio-rhetorical dimension of genre production.
While our first suggestion for future research is pedagogical, the sec-
ond suggestion is methodological. The reviewed studies above focused
either on texts (e.g., McGrath, 2016) or on writers’ knowledge develop-
ment (e.g., Tardy, 2005) but rarely on both (e.g., De Oliveira & Lan,
2014). It would be interesting to investigate how the dialogic interaction
between teachers and students (e.g., in the form of feedback) or different
co-authors (comments to a text) leads to genre knowledge development
and genre production, for instance through text histories (see Khuder &
Petrić, 2022a). For example, if the case study focuses on a writer’s textual
and conceptual development of a specific genre, data can include inter-
views with all authors and literacy brokers involved in the text, including
the main case, in addition to the main case’s uptake of these comments
and perceptions of the changes made. Such a holistic overview of devel-
opment can give the reader a better idea of how interactions about texts
between various co-authors and/or language tutors impact both the text
and the writer.
Finally, we believe that genre research would be greatly enriched by
conducting case studies similar to our example study (i.e., Kobayashi
& Rinnert, 2013), examining the complex and dynamic relationship
between genre knowledge and learning experiences in multilingual writ-
ers. The need for this research has been advocated both by Gentil (2011)
and Tardy et al. (2020), as needed to move toward a clearer theoretical
articulation of how writing knowledge transfers and develops across lan-
guages and genres, or to borrow Gentil’s (2011) question: “How does
genre knowledge intersect with writing expertise and language knowl-
edge?” (p. 19).
 32 Raffaella Negretti, Baraa Khuder

References
Andersen, E., & Schiano, W. (2014). Teaching with cases: A practical guide. Harvard
Business Review Press.
Casanave, C. P. (2003). Looking ahead to more sociopolitically-oriented case study
research in L2 writing scholarship: (But should it be called “post-process”?).
Journal of Second Language Writing, 12(1), 85–102.
Casanave, C. P. (2010). Taking risks?: A case study of three doctoral students
writing qualitative dissertations at an American University in Japan. Journal of
Second Language Writing, 19(1), 1.
Cenoz, J., & Gorter, D. (2011). Focus on multilingualism: A study of trilingual
writing. Modern Language Journal, 95(3), 356–369.
Cheng, A. (2008). Analyzing genre exemplars in preparation for writing: The case
of an L2 graduate student in the ESP genre-based instructional framework of
academic literacy. Applied Linguistics, 29(1), 50–71.
Christensen, C. R., & Hansen, A. J. (Eds.). (1987). Teaching and the case method.
Harvard University Press.
De Oliveira, L. C., & Lan, S. W. (2014). Writing science in an upper elementary
classroom: A genre-based approach to teaching English language learners.
Journal of Second Language Writing, 25(1), 23–39. https://doi​.org​/10​.1016​/ J​
.JSLW​.2014​.05​.001
Denzin, N. K. (1978). Sociological methods: A sourcebook. McGraw-Hill.
Denzin, N. K., & Lincoln, Y. S. (2011). The SAGE handbook of qualitative research.
SAGE.
Duff, P., & Anderson, T. (2016). Case study research. In J. D. Brown & C. Coombs
(Eds.), Cambridge guide to language research (pp. 112–118). Cambridge
University Press.
Eisenhardt, K. M. (1989). Building theories from case study research. Academy of
Management Review, 14(4), 532. https://doi​.org​/10​.2307​/258557
Flowerdew, J., & Wan, A. (2010). The linguistic and the contextual in applied genre
analysis: The case of the company audit report. English for Specific Purposes,
29(2), 78–93. https://doi​.org​/10​.1016​/ J​.ESP​.2009​.07​.001
Flyvbjerg, B. (2011). Case study. In N. K. Denzin & Y. S. Lincoln (Eds.), The
SAGE handbook of qualitative research (pp. 301–316). SAGE.
Gentil, G. (2011). A biliteracy agenda for genre research. Journal of Second
Language Writing, 20(1), 6–23.
Guo, M. (2019). Intertextuality and nationalism discourse: A critical discourse
analysis of microblog posts in China. Asian Journal of Communication, 29(4),
328–345. https://doi​.org​/10​.1080​/01292986​.2019​.1628286
Harwood, N., & Petrić, B. (2017). Experiencing master’s supervision: Perspectives
of international students and their supervisors. Routledge. https://doi​.org​/10​
.4324​/9781315680934
Huang, J. C. (2014). Learning to write for publication in English through genre-
based pedagogy: A case in Taiwan. System, 45(1), 175–186. https://doi​.org​/10​
.1016​/ J​. SYSTEM​.2014​.05​.010
Kessler, M. (2021). The longitudinal development of second language writers’
metacognitive genre awareness. Journal of Second Language Writing, 53.
https://doi​.org​/10​.1016​/j​.jslw​.2021​.100832
 Case studies 33

Khuder, B., & Petrić, B. (2021). Walking on thin ice: Reflexivity in doing
ethnography. In I. Guillén-Galve & A. Bocanegra-Valle (Eds.), Ethnographies of
academic writing research: Theory, methods, and interpretation (pp. 106–123).
John Benjamins. https://doi​.org​/10​.1075​/ RMAL​.1​.06KHU
Khuder, B., & Petrić, B. (2022a). Academic texts in motion: A text history study
of co-authorship interactions in writing for publication. Journal of English for
Research Publication Purposes, 3(1), 51–77. https://doi​.org​/10​.1075​/ JERPP​
.22001​.KHU
Khuder, B., & Petrić, B. (2022b). Helping EAL academics navigate asymmetrical
power relations in co-authorship. Writing and Pedagogy, 14(1). https://doi​.org​
/10​.1558​/ WAP​.20495
Kobayashi, H., & Rinnert, C. (2013). L1/L2/L3 writing development:
Longitudinal case study of a Japanese multicompetent writer. Journal of Second
Language Writing, 22(1), 4–33.
Lillis, T. (2008). Ethnography as method, methodology, and “deep theorizing”:
Closing the gap between text and context in academic writing research.
Written Communication, 25(3), 353–388. https://doi​.org​/10​.1177​
/0741088308319229
Lillis, T., & Curry, M. J. (2018). Trajectories of knowledge and desire: Multilingual
women scholars researching and writing in academia. Journal of English for
Academic Purposes, 32, 53–66. https://doi​.org​/10​.1016​/ J​.JEAP​.2018​.03​.008
Lillis, T., & Harrington, K. (2015). Working with academic literacies: Case studies
towards transformative practice. Parlor Press.
Lillis, T., & Scott, M. (2007). Defining academic literacies research: Issues
of epistemology, ideology and strategy. Journal of Applied Linguistics and
Professional Practice, 4(1), 5–32. https://doi​.org​/10​.1558​/japl​.v4i1.5
Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. SAGE.
Luck, L., Jackson, D., & Usher, K. (2006). Case study: A bridge across the
paradigms. Nursing Inquiry, 13(2), 103–109. https://doi​.org​/10​.1111​/ J​.1440​
-1800​.2006​.00309.X
Martin, J. R. (1992). English text: System and structure. John Benjamins.
McGrath, L. (2016). Open-access writing: An investigation into the online drafting
and revision of a research article in pure mathematics. English for Specific
Purposes, 43, 25–36. https://doi​.org​/10​.1016​/ J​.ESP​.2016​.02​.003
Merriam, S. B. (2009). Qualitative research: A guide to design and implementation.
Jossey-Bass.
Miller, C. R. (1984). Genre as social action. Quarterly Journal of Speech, 70(2),
151–167.
Negretti, R., & Kuteeva, M. (2011). Fostering metacognitive genre awareness in
L2 academic reading and writing: A case study of pre-service English teachers.
Journal of Second Language Writing, 20(2), 95–110.
Negretti, R., & McGrath, L. (2018). Scaffolding genre knowledge and metacognition:
Insights from an L2 doctoral research writing course. Journal of Second Language
Writing, 40, 12–31. https://doi​.org​/10​.1016​/j​.jslw​.2017​.12​.002
Ortega, L., & Carson, J. (2010). Multicompetence, social context, and L2 writing
research praxis. In T. Silva & P. K. Matsuda (Eds.), Practicing theory in second
language writing (pp. 48–71). Parlor Press.
34 Raffaella Negretti, Baraa Khuder

Patton, M. Q. (1999). Enhancing the quality and credibility of qualitative analysis.

Health Sciences Research, 34(5 Pt 2), 1189–1208.
Pessoa, S., Mitchell, T. D., & Miller, R. T. (2017). Emergent arguments: A
functional approach to analyzing student challenges with the argument genre.
Journal of Second Language Writing, 38, 42–55. https://doi​.org​/10​.1016​/j​.jslw​
.2017​.10​.013
Polio, C., & Friedman, D. A. (2016). Understanding, evaluating, and conducting
second language writing research. Routledge.
Ragin, C., & Becker, H. (1992). What is a case? Exploring the foundations of social
inquiry. Cambridge University Press.
Stake, R. E. (1995). The art of case study research. Sage.
Stake, R. E. (2008). Qualitative case studies. In N. K. Denzin & Y. S. Lincoln
(Eds.), Strategies of qualitative enquiry (pp. 119–149). SAGE.
Swales, J. (1990). Genre analysis: English in academic and research settings.
Cambridge University Press.
Tardy, C. M. (2005). “It’s like a story”: Rhetorical knowledge development in
advanced academic literacy. Journal of English for Academic Purposes, 4(4), 325–
338. https://doi​.org​/10​.1016​/j​.jeap​.2005​.07​.005
Tardy, C. M. (2009). Building genre knowledge. Parlor Press.
Tardy, C. M., Sommer-Farias, B., & Gevers, J. (2020). Teaching and
researching genre knowledge: Toward an enhanced theoretical framework.
Written Communication, 37(3), 287–321. https://doi​.org​/10​.1177​
/0741088320916554
Townley, A., & Jones, A. (2016). The role of emails and covering letters in
negotiating a legal contract: A case study from Turkey. English for Specific
Purposes, 44, 68–81. https://doi​.org​/10​.1016​/ J​.ESP​.2016​.07​.001
Troyan, F. J. (2016). Learning to mean in Spanish writing: A case study of a genre-
based pedagogy for standards-based writing instruction. Foreign Language
Annals, 49(2), 317–335. https://doi​.org​/10​.1111​/ FLAN​.12192
Wang, W. (2020). Analysing Chinese language and discourse: Across layers and
genres. John Benjamins.
Yin, R. K. (2003). Case study research: Design and methods. SAGE.
Yin, R. K. (2009). Case study research: Design and methods. SAGE.
3
ETHNOGRAPHIC RESEARCH
Christine M. Tardy

3.1 Introduction to the approach and definition of genre

Ethnography is commonly associated with research in cultural anthropol-
ogy, in which researchers traditionally traveled to distant places, embed-
ding themselves in unfamiliar communities for months, even years, in order
to identify cultural patterns of meaning. Brodkey (1987) defined ethnog-
raphy as “the study of lived experience…[of] the meanings that groups
confer on individual actions, the selection and evaluation of certain actions
as meaningful or significant cultural acts” (pp. 25–26). For Li (2020), eth-
nography involves four processes: an interest in a social group’s practice or a
social phenomenon, immersion in that social practice or phenomenon over
substantial time to observe its variations, collection of data that show how
people in the social group interpret their practice or the phenomenon, and
sharing an account of the observation and interpretation.
Given ethnography’s emphasis on understanding social practices, one
might imagine that it is particularly well suited toward the study of literacy
and, more specifically, the study of genre. After all, genre is not simply a
linguistic form but rather has consistently been theorized as a social prac-
tice (e.g., Bazerman, 1988; Martin, 1992; Miller, 1984; Swales, 1990).
Drawing a distinction between genre-as-form and genre-as-social-practice
is critical in researching genres because seeing genre as a practice means
that to understand a genre we must attend to the social environment in
which it is (re)produced and used. A study that examines only linguistic
features of a text (that is, form) may suggest how such language patterns
result from (or even construct) particular social practices and values, but it

DOI: 10.4324/9781003300847-4
36 Christine M. Tardy

does not directly observe those practices. Put another way, a study that ana-
lyzes text alone can reveal insights into texts but is incomplete in revealing
insights into genre. Ethnography offers a mode of inquiry that offers great
potential for the study of genre as a social practice.
It is perhaps not a coincidence that the word genre entered applied lin-
guistics in part indirectly through the work of anthropologist Clifford
Geertz. John Swales, a pioneer in introducing and developing genre studies
within applied linguistics, recounts how he came to see genre as a key con-
cept for understanding academic writing:

I happened, also in 1980, to be casually reading a review of a recent work

about traditional Javanese customs by the great cultural anthropologist,
Clifford Geertz. The reviewer several times mentioned the word genre
in connection with cock-fighting rituals and the arrangements required
for princely audiences and the like. Suddenly, I realized that the concept
of genre was the one that I had sub-consciously been searching for over
the previous year or two.
(Swales, 2009, p. 141, italics in original)

Swales (1990) later defined genre as “a class of communicative events, the

members of which share some set of communicative purposes…recognized
by the expert members of the parent discourse community” (p. 58). He
further noted that “[t]he genre names inherited and produced by discourse
communities and imported by others constitute valuable ethnographic
communication, but typically need further validation” (p. 58). Around the
same time, in the related field of rhetoric, Carolyn Miller (1984) defined
genre as “a particular type of discourse classification…based in rhetorical
practice and consequently open rather than closed and organized around
situated actions (that is, pragmatic rather than syntactic or semantic)” (p.
155). Miller, like Swales, also emphasized the social aspect of genre, noting
that “[t]he classification I am advocating is, in effect, ethnomethodological;
it seeks to explicate the knowledge that practice creates” (p. 155).
At the heart of both of these definitions is the notion that genre is not
simply a text marked by patterned uses of language forms; rather, gen-
res are practices used by social groups or communities to carry out com-
mon actions. These groups develop preferred forms for these actions, and
these forms evolve over time as groups and surrounding social, political,
rhetorical, and technological contexts shift. Genres have therefore been
said to index a group’s systems of values, beliefs, and practices, while at
the same time reinforcing those. Given this highly social view of genre,
researchers have turned – to varying degrees – to research methodologies
that offer insight not just to texts but also to the influences on those texts,
 Ethnographic Research 37

that is, the practices and values that led them to develop in particular ways.
Ethnography is one such methodology. It has been adapted primarily to the
study of academic genres like dissertations (e.g., Paltridge et al., 2012) but
also to professional genres like nursing care plans (Parks, 2001) or public
genres like museum visitor books (Noy, 2015). Ethnographic research of
genres has explored the writing practices of novices (monolingual and mul-
tilingual) and experts, students, and professionals.

3.2 Goals
Broadly speaking, ethnography offers “a way of understanding the par-
ticularities of the social life of individuals and social groups” (Li, 2020,
p. 156). When used to study genres, ethnographic research explores the
values, beliefs, and practices of individuals and social groups who use a par-
ticular genre or set of genres, or to understand how a genre may influence a
group’s social practice. For example, ethnographic studies of genre may ask:

• How is the genre produced? Who is involved in its production, and

what are the related genres, activities, and technologies that influence
its production?
• How do newcomers learn to use the genre? What activities and/or peo-
ple support this learning, or make it more challenging?
• How are the genre’s production and form shaped by the community,
including its other genres, its values, and its practices?
• What barriers and scaffolds exist to accessing the genre and its production?
• How is the genre distributed and used by others? Who has access to the
genre, and who does not?
• How do users assess how a genre has been carried out? What makes a use
of the genre successful or unsuccessful?
• What is the importance of the genre within the community? Why? Has
that importance changed over time, or does it change in different con-
texts? What attitudes and behaviors toward the genre exist (possibly by
different members of the community)?
• What is the value or consequence of using or not using the genre (and
related genres) as expected?

Because of their focus on social context and situated practice, ethnographic

studies of genre are especially appropriate for studying links among various
genres – sometimes referred to as sets, systems, or networks – rather than
focusing on genres in isolation. Rather than zooming in on particular con-
ventions or features of a genre, in other words, an ethnographic approach
allows researchers a wide-angle lens at genres in use.
38 Christine M. Tardy

3.3 Common research methods

Scholars have outlined key features of ethnographic research in various
ways, but common among these are an emphasis on sustained immersion
in the research setting and on developing a productive tension between
emic (insider) and etic (outsider) perspectives (Lillis, 2008; Paltridge et al.,
2016) – that is, between the perspectives of those being studied and the
researcher. Drawing heavily on Watson-Gegeo’s (1988) definition of eth-
nography, Ramanathan and Atkinson (1999) offer a prototypical definition
of ethnographic research as:

A species of research which undertakes to give an emically oriented

description of the cultural practices of individuals, where ‘cultural’ is
extended as described just above. Additionally, ethnographic research
aims to bring a variety of different kinds of data to bear in such descrip-
tion, on the principle that multiple perspectives enable more valid
description of complex social realities than any single kind of data could
alone.
(p. 49)

The authors ultimately argue that such research has tremendous value in
the study of second or additional language (L2) writing because it can help
us understand complex social and cultural practices. More specifically, they
point to the potential of ethnographic research in highlighting the particu-
lars, complicating an understanding of culture, and bringing together an
array of research tools to gather multiple perspectives on a phenomenon.
Nearly a decade after Ramanathan and Atkinson’s (1999) exploration
of ethnographic research in L2 writing, Lillis (2008) offered a tripartite
framework for identifying the ways in which such research has been taken
up in the study of literacy. Most common, Lillis argues, has been the use of
ethnography-as-method, which includes the integration of some investiga-
tion of the contexts surrounding texts, typically through interviews with
writers or readers. This focus on “talk around texts” (p. 355) allows text-
based researchers to begin to understand writers’ and readers’ perspectives
on genre (or other literacy practices) but still offers a limited view of social
context or textual practices. Ethnography-as-methodology goes further by
integrating multiple data sources collected over a lengthy period of time,
allowing researchers “to explore and track the dynamic and complex situ-
ated meanings and practices that are constituted in and by academic writ-
ing” (Lillis, 2008, p. 355). Finally, and more rarely, ethnography can be
construed at the level of epistemology and ontology, what Lillis – draw-
ing on Blommaert (2007) – refers to as deep theorizing. At this level,
 Ethnographic Research 39

ethnographic research can help to narrow the gap between the exploration
of text and context and instead “help us explore the complex politics of
academic text production” (Lillis, 2008, p. 381). While all three approaches
to ethnographic research can offer insight into literacy practices in general
and genre in particular, the latter two go furthest in providing a rich, mul-
tilayered perspective into the social nature of texts.
As may be clear, ethnographic research – especially when viewed as a
continuum of approaches – can share many features of related methodolo-
gies, such as case study research, naturalistic inquiry, ethnography of com-
munication, or qualitative research more broadly (Ramanthan & Atkinson,
1999). Green and Bloome (2005) identify distinctions between doing eth-
nography, adopting ethnographic perspectives (less than a full-scale eth-
nography), and using ethnographic tools (adopting fieldwork methods and
techniques), though others use these and related terms somewhat inter-
changeably. Within applied linguistics, researchers are often reluctant to
label their research as ethnographic, likely because it often lacks the range
of research tools, length of time, and perhaps overall depth of data of a
comprehensive ethnography. Instead, it has become more common to refer
to related research as ethnograhically informed or ethnographically oriented,
gently implying that it draws on principles of traditional ethnography but
does not adopt the overall approach. Despite a lack of unified terminology,
however, researchers considering how to label their own work might reflect
on the extent to which their methodology implements common principles
of ethnographic research, such as the use of multiple data sources, partici-
pant observation, an interplay between emic and etic perspectives, sustained
involvement in a research setting, and researcher reflexivity (i.e., a critical
reflection on one’s own positionality and role in the research process and
data interpretation).
As already suggested, ethnographic research of genre draws on an array
of tools to gain insight into social practices. Interviews are extremely com-
mon for gathering the perspectives of genre users. Interviews may be single
informant-type events, or they may be repeated over a period of time, either
within (in situ) or outside of the natural setting of genre use. Repeated
interviews have the benefit of offering more depth in understanding an
individual’s use and interpretation of a genre over time and in varying cir-
cumstances; sustained use of interviews with specific genre users also helps
build a participant-researcher relationship of trust and one in which the
researcher may gradually gain a stronger emic perspective, contributing to
what Sarangi (2007) refers to as thick participation, a kind of researcher
involvement that moves beyond a simple fly-on-the-wall account. Given the
embedded nature of much ethnographic research, interviews can also be
less formal and can include the more casual interactions that researchers
40 Christine M. Tardy

have with members of the community (see, for example, Cimasko & Shin,
2017; Noy, 2015).
Two more related tools often used for building an emic perspective
into a culture or community are participant observation and fieldnotes.
By observing a community’s practices, possibly even as a participant in
those practices, and making notes of those observations, researchers gradu-
ally build a picture of a group’s practices, including common patterns and
variations and the meaning behind those practices. In my own study of
genre-based instruction, for example, I spent 15 weeks attending a graduate
student writing class, recording each class and taking notes about activi-
ties and student reactions (Tardy, 2009). As a graduate student myself, I
had much in common with the other students, and I was often invited by
the teacher or other students to participate in small group activities or to
contribute to group discussions. This sustained involvement made me a
familiar presence to the students in the class and also helped build trust
with my focal participants (4 of the 11 class members) and begin to gain
some insight into their own perspectives on the classroom.
Finally, the analysis of artifacts often plays an important role in ethno-
graphic research. In the study of genre, artifacts may include, for example,
various drafts of a text in the genre of focus with intervening feedback
or comments (e.g., Prior, 1998), a collection of guidelines or policies
that shape a text and its production (e.g., Berkenkotter, 2001; Ravotas &
Berkenkotter, 1998), or the network of texts that work together to carry
out a group’s (or individual’s) practices (e.g., Swales, 1998). Photographs
of research settings (e.g., Noy, 2015; Wickman, 2010) or visuals produced
by participants are also used fairly commonly in papers and monographs
sharing ethnographic research. Such visuals help to bring readers into the
setting and offer a glimpse at a community’s embodied use of a genre.
Because it studies social practices, ethnographic research is often carried
out over a lengthy period of time, giving researchers an opportunity to see
not just a snapshot of activity but rather the rich variation in social interac-
tion and practice that develops over a period of months or years. In studies
of genre, longitudinal research is rare but does exist. Focused more broadly
on literacy practices, Spack’s (1997) study of an undergraduate interna-
tional student writing across disciplines and genre followed the participant,
Yuko, for three years, and Leki’s (2007) study followed four undergraduate
multilingual writers for the entirety of their undergraduate careers. These
studies are more limited in their exploration of context (focusing more on
the individuals themselves) but offer insight into how much can change in
a writer’s literacy experiences over extended periods of time. In another
approach, focusing more firmly on context, Noy (2015) studied the use of
written genres in museum settings as a weekly visitor and data collector for
 Ethnographic Research 41

several years; his study highlights the value of immersing oneself in a set-
ting over a lengthy period, becoming a participant-observer.
One interesting adaptation of ethnography used in research of genre is
what Swales (1998) termed textography, which is “something more than
a disembodied textual or discoursal analysis, but something less than a
full ethnographic account” (p. 1). Textography is an approach to under-
standing texts and their social contexts through tools such as interviews,
observations, artifact analysis, with the goal of getting “an inside view of
the worlds in which the texts are written, why the texts are written as they
are, what guides the writing, and the values that underlie the texts that have
been written” (Paltridge et al., 2012, p. 334). The most extended example
of this approach is found in Swales (1998), which uses interviews, arti-
fact analysis, and participant observation to explore the cultures of writing
that inhabit three different floors of a small university building: a computer
resource center in the basement, a second-floor university herbarium, and a
third-floor English language institute. Swales’ approach differs most from
traditional ethnography in his emphasis on the analysis of written text, but
it still employs a rich description of the physical site (complete with numer-
ous pictures) and social communities built through participant observa-
tion, sustained involvement, and researcher reflexivity.
Whether their data include interviews, observational fieldnotes, artifacts,
or a combination of these, ethnographic studies tend to gather immense
amounts of data and can therefore be particularly daunting to researchers.
Approaches to ethnographic research, including data analysis, do vary, but
Pole and Morrison (2003) helpfully identify at least two fairly common fea-
tures: an interest in categorizing data into units or themes, and a belief in
grounded theory (Glaser & Strauss, 1967) or a desire to generate theories or
abstractions from the data. In fact, these two practices are linked, as theory
generation is part of a researcher’s process of making sense of their data.
Blommaert and Jie (2020) offer a nice analogy of the process of making
sense of a larger phenomenon from numerous “pieces” of data, compar-
ing it to walking around a table to describe a can of soda. Viewing the can
from different angles gives you a different description; by walking all the
way around and looking at the can from different perspectives, you gain a
more complete understanding of the whole can. By looking at data from
numerous perspectives, an ethnographic researcher can start to get a fuller
understanding of the phenomenon or activity they are studying. Gradually,
they can work toward a speculative theory about it as well.
One particular challenge in data analysis is simply contending with the
vast amount of data typically generated in an ethnographic study. Qualitative
coding programs, such as NVivo or Atlas​.t​i, can have a steep learning curve
but are nearly essential for storing and analyzing data in an ethnographic
42 Christine M. Tardy

project. These programs allow researchers to compile multiple modalities

of data, to establish and refine various levels of coding categories, and to
identify and even visually display connections across data sources, all aiding
in the generation of theory.
Several additional tools can facilitate ethnographic data analysis, while
attending to issues of trustworthiness, researcher positionality, and theory
generation. First, the triangulation of multiple sources and kinds of data
allows researchers to look at a phenomenon from different vantage points,
identifying patterns or even contradictions. For example, interview data
can help reveal how a writer might understand (or articulate an under-
standing of) a genre, but their actual engagement in a genre reveals how
they apply and adapt that understanding in practice. A participant’s inter-
view discussion of a genre and their textual instantiation of that genre can
therefore be compared and analyzed together for a more complex perspec-
tive. Another tool for data analysis and reflection is member checking, in
which researchers share preliminary analyses with participants, helping to
compare insiders’ and researcher perspectives and ascribed meanings. In
addition, creating researcher memos can be an essential part of contend-
ing with a great deal of data as well as tracking researcher interpretations
over a sustained period of time. Memos can be written throughout the
research process, taking notes on the data collection and recording initial
and ongoing thoughts on salient or important moments or even themes.
Memos also allow researchers to engage in reflective practice, considering
how their positionality and perspectives influence their participants and
also their own interpretations of the data.

3.4 Example studies
Depending on how ethnography is construed – as method, methodology, or
deep theorizing – studies adopting an ethnographic perspective to research
may look quite different. Table 3.1 highlights six studies that adopt, to var-
ying extents, principles and/or tools of ethnograhically informed research.
Some of these studies exemplify situated genre analyses that emphasize
textual analysis in the tradition of textography. Paltridge et al. (2012), for
example, adopt a textographic approach in studying practice-based doctoral
theses in visual and performing arts. Seeking to identify the forms of such
theses, the researchers analyze 36 doctoral texts through rhetorical move
analysis (also see Casal & Kessler in this volume, for more). Though their
primary focus is on understanding the texts’ macrostructure, the research-
ers also interview 15 students and 15 supervisors, observe related doctoral
activities, and examine relevant artifacts that influence this unique thesis
genre. Together, these additional methods gather the “talk around texts”
TABLE 3.1 
Example studies adopting ethnographic approaches

Authors Context Research Question(s) or Data Collected Comments on Study Design

Aims

Atkinson and • Writing in L2 pre- • What attitudes and • Participant observation of • Research goal was
Ramanthan university and L1 behaviors regarding four teacher orientation to understand how
(1995) first-year university academic writing and sessions, ethnographic the programs’ beliefs
writing programs its teaching pervade interviews with seven influenced teaching
in the US the organizations that program administrators norms
teach it? and six teachers, • The authors refer
• How are broader 47 hours of classroom to their study as
concepts of writing observations, program “ethnographically
manifested in teacher artifacts, researcher oriented”
behavior? memos

Flowerdew and • Company audit • The researchers • Interviews with four • Focuses on move
Wan (2010) reports in a Hong aimed to understand auditors and a manager, structure with additional
Kong accounting the communicative nine hours of observation contextual analysis
firm purpose of audit of an audit assignment, • The authors refer to
reports 25 audit reports the approach as using
“ethnographic methods”
as a way to understand
the contextual aspects of
genre
(Continued)
Ethnographic Research
43
TABLE 3.1 Continued

Authors Context Research Question(s) or Data Collected Comments on Study Design

Aims

Morton (2016) • The “desk-crit,” a • The researcher • Observations and video • The study is described
spoken genre in an explored how recordings of weekly as a “situated genre
architecture degree academic and design studio activities analysis” following
professional (12 weeks), interviews Dressen-Hammouda
44 Christine M. Tardy

architecture with disciplinary insiders, (2014), building on

communities shape documents related to textography (Swales,
the design studio the studio and academic 1998)
curriculum and degree
specifically the desk-
crit genre
Noy (2015) • The visitor book • What are the material • Sustained observations of • The author describes the
(and similar and discursive means writing in public spaces, study as an ethnography
writing platforms) through which heritage fieldnotes, interviews
in two Jewish institutions and with museum staff,
heritage museums museums rhetorically photographs of field sites,
in the US frame VBs so as to artifacts, interactions
serve as situated and with visitors, digital
symbolic surfaces of copies of the visitor books
and for writing?
• How, through their
addressivity structures
and the collaborative
efforts involved in
their composition, do
visitors’ texts perform
participation?
Paltridge et al. • Doctoral texts • The researchers aimed • Survey of doctoral • The article focuses
(2012) in the visual and to identify thesis supervisors, 36 doctoral primarily on the texts’
performing arts forms that are used by texts, interviews with macrostructure, but
at an Australian doctoral students in 15 doctoral students other data points are
university visual and performing and 15 supervisors, used to understand the
arts university guidelines, meaning of different
in-house art publications, forms
observations of doctoral • The study is described
roundtable discussions as taking a textographic
and exhibitions approach
Schryer (1993) • A medical record- • The researcher aimed • 80 interviews (students, • Emphasizes triangulation
keeping system at a to understand medical faculty, practitioners), of methods (between and
veterinary college records as a complex over 200 hours of within) as essential
site for literacy participant observations • Defines genre as
practice in classes, labs, and “stabilized-for-now or
clinics, ten faculty reader stabilized enough sites
protocols of student of social and ideological
papers, and extensive action” (p. 204)
artifact collection over six • The study is described as
months an ethnography
Ethnographic Research
45
46 Christine M. Tardy

that provide insight into how the genre is carried out, sometimes adopting
and other times adapting a traditional dissertation macrostructure to vary-
ing degrees.
Flowerdew and Wan’s (2010) study bridges contextual and textual analy-
sis perhaps even more, examining company audit reports in a Hong Kong
accounting firm. They describe their methods as ethnographic but “given
their limited scope and intensity … not constitut[ing] a complete ethnog-
raphy” (Flowerdew & Wan, 2010, p. 82). To understand the audit report
genre as a contextualized practice, they carry out interviews with auditors
and a manager as well as multiple observations of audit activity. Their pri-
mary analysis, however, is a rhetorical move analysis of a collection of 25
audit reports, so the text form still remains central. The researchers identify
six common moves and study their frequency and variability. The research-
ers’ contextual data offer insight into why different moves are used and any
constraints on their implementation–insight that could not be gained from
text analysis alone. Interviews, for example, highlighted that the reports are
primarily template-driven but that auditors often struggled with writing
reports for more complex audit situations; this finding offered insight into
the auditors’ pedagogical needs.
A third example by Morton (2016) demonstrates how textography
has been used as a springboard for a situated genre analysis (see Dressen-
Hammouda, 2014) that examines the relationship between a unique genre
and its community of users. Here Morton studies the “desk-crit,” a spoken
genre used weekly in design studios as a part of architecture education,
defined as “the presentation of artifacts and the response to these” (p. 58).
To explore how academic and professional communities enact and negoti-
ate the genre in the educational setting, Morton integrates a sociohistori-
cal analysis of the genre’s origins and evolution and a close analysis of one
recorded performance of the genre. To do so, she draws on several data
sources and collection methods, including observations and video record-
ings of weekly design studio meetings over 12 weeks; interviews with the
studio teacher, his students, and four other teachers from industry; and an
analysis of related artifacts. A close analysis of two excerpts from one desk-
crit performance demonstrates how the teacher and student negotiate com-
munity values within the genre. Though the analysis reported in Morton’s
paper focuses on only one performance of the desk-crit genre, the larger
collection of data and observations allow for a more nuanced and layered
understanding of the genre in context.
These first three examples emphasize – to varying degrees – how
researchers can explore texts within a genre while also integrating atten-
tion to the complex contexts in which genres are created and used. Other
studies of genre that adopt ethnographic research approaches tilt their gaze
 Ethnographic Research 47

more toward the study of context, giving minimal attention to textual

forms themselves. One of the first and most notable contextual or genre-in-
practice studies was carried out by Schryer (1993). Responding to a request
from college faculty to help them understand perceived limitations in their
students’ writing, Schryer explains that “rather than simply examining stu-
dent texts in order to taxonomize errors…I wanted the opportunity to ask
the radical, contextual question – what did literacy mean here – and the
opportunity to observe the teaching and learning practices that constituted
literacy at the college” (p. 201). Schryer describes her research as an eth-
nography, and her data pool reflects the type and breadth of tools typically
employed in ethnographic studies: 80 interviews, over 200 hours of par-
ticipant observation, ten think-aloud protocols of faculty reading student
papers, and a large collection of relevant artifacts, all collected over a six-
month period.
Schryer focuses on the problem-oriented veterinary medical record
(POVMR) – a variation on the problem-oriented medical record (POMR)
genre – that was used in the college but with some controversy among fac-
ulty. Her data allow her to depict a rich picture of the genre in practice. She
describes in detail, for example, the “intense questioning” (p. 220) from
students to a clinician in a demonstration with a young horse, and she shares
faculty reactions to various student exam responses to illustrate “the socially
constructed nature of ‘problems’ and their embeddedness in the discourse
of veterinary medicine” (p. 223). As Schryer explains, the students’ exam
responses are a different genre from the POVMR, but they are intricately
related as the problem-solving structure of the POVMR underlies instruc-
tional approaches and the students’ discourse socialization. Ultimately, the
study offers insight into the highly intertextual, contested, and dynamic
nature of genre in situ, and it is able to do so because of its deeply layered
engagement with context. This study is, in fact, the source of Schryer’s now
well-known insight that genres should be considered “stabilized-for-now,”
demonstrating the potential power of ethnographic research to contribute
substantially to theoretical knowledge.
A second example of ethnographic research that emphasizes contextual
aspects of genre can be found in Noy’s (2015) research of visitor books
and related genres found in Jewish heritage museums. Noy explains his
approach to adopting an emic perspective on the genre and its associated
practices: “Rather than assuming what [visitor books] are, I examined them
bottom-up, so as to begin from scratch and ask, how, where, and when are
they employed, and to what effect” (Noy, 2015, p. 214). Noy’s goal is not
to understand the textual form of a genre, but rather, to understand the
“rich textual environments” of visitor books and the “writing activities and
the public texts that they produce” (p. 197) within these environments. To
48 Christine M. Tardy

do so, he spent multiple years observing literate activity in weekly visits to

two museums, watching visitors interact with visitor books and with other
visitors. In addition to observations and fieldnotes, Noy carried out infor-
mal interviews (conversations) with visitors and analyzed visitor book texts.
His article weaves these data sources together to paint a rich picture of the
visitor books as a site of literate activity, incorporating photographs of the
books and their environments to help readers visualize the spaces.
So far, I have described ethnographically oriented studies that focus pri-
marily on the textual aspects of genre and those that focus primarily on the
contextual or social aspects of genre as practice. Now I turn to a study that
narrows the gap between text and context, studying genre as product and
practice. Atkinson and Ramanthan’s (1995) study was one of the first to
adopt an ethnographic approach in the study of L2 writing. Their research
does not examine a single genre but rather “cultures of writing” in a uni-
versity writing program and university English language program. Over a
ten-month period, the authors (who had teaching experience in the pro-
grams) collected a wide range of data: participant observations of teacher
orientation sessions, interviews with seven program administrators and six
instructors, 57 hours of classroom observations, program artifacts (e.g.,
faculty handbooks, course objectives, assignments, teacher feedback), and
researcher notes. This layered approach to data allowed the researchers to
locate important divergences in the two programs in terms of how they rep-
resent and teach academic writing. They note, for example, that the writing
program tended to value writing that displays “sophisticated thought and
expression” (p. 560), while the language program promoted the use of
deductive forms with “workpersonlike prose” (p. 560) (characterized by,
for example, straightforward communication of information and use of a
formulaic deductive structure). In the end, their work highlights the chal-
lenges for students who must navigate “diverse cultural territories” (p. 564)
as they move from one program to the other.
The six studies reviewed here demonstrate the diversity of what might
fall under the umbrella label of ethnographic or ethnographically oriented
genre research. Despite differing attention to text and context, they share
an aim to situate genre use and practice. All bring together a range of quali-
tative data, and in doing so, they offer a view into how genres are produced
and negotiated in social environments.
It is notable that many of these ethnographic studies do not research a
genre per se, but rather explore broader literacy practices in which genres
are important tools. Schryer (1993), for example, did not begin her study as
one of the POVMR, but rather, as an exploration of literacy in one educa-
tional context. The medical report emerged as an influential and contested
genre within that environment. Similarly, Atkinson and Ramanathan’s
 Ethnographic Research 49

(1995) study did not investigate one particular type of writing in the two
university language and writing programs, but it did reveal distinctions in
how writing in general and “essays” in particular reflect and perpetuate
particular writing values. As the lens of ethnographic research widens, it is
understandably a challenge (and perhaps less desirable) to focus on a single
genre, as genres are inextricably connected to larger textual and social envi-
ronments. Ethnographic research enables a wider ecological study of genre
and literacy practices.

3.5 Issues and challenges

Despite its powerful tools for understanding genre in general, as well as
specific genres as social practice, ethnographic research of genre is not as
common as one might think. This is especially the case in applied linguis-
tics, where text understandably tends to gather more research attention
than context in genre-related scholarship. (In contrast, much research of
genre from a rhetorical perspective tends to focus more on context than
text.) Aside from a tendency to foreground textual data, however, a major
reason for more limited use of ethnographic research into genres is that it is
a relatively challenging research approach.
First, and perhaps most obviously, ethnographic research is extremely
time-consuming. An ethnographically oriented study typically involves a
rather large set of data, as illustrated in the sample studies described here,
and a larger dataset means more time spent collecting, organizing, and
analyzing data. Data in ethnographic studies are also often compiled over
a sustained period of time in order to understand the setting more deeply,
so the collection phase itself can be quite lengthy. Many of the sample
studies discussed in this chapter, for example, collected data for at least one
year, though in some cases ethnographic work is as short as a few months.
Additional time may be needed to apply for and gain approval from insti-
tutional ethics boards, a practice that is typical for any research involving
human participants and must generally be carried out before any data col-
lection begins. Researchers may also need to apply for grants to compensate
for the time taken on projects that require numerous hours of immersion
in a research setting.
Access can also present challenges to ethnographic researchers. Some
genres, for example, are carried out in communities that are closed to out-
siders, and research may need to use various social networks to gain entry
into the community. In many of the sample studies described in this chap-
ter, researchers already had an “in” to their context, for example as teachers
(e.g., Atkinson & Ramanathan, 1995) or as a consultant (e.g., Schryer,
1993). But in some cases, researchers may have to forge new connections
50 Christine M. Tardy

to open up research opportunities. A related consideration for researchers is

the extent to which they are comfortable asking for participants’ time and
willingness to share community artifacts, to participate in interviews, to be
observed, or to answer surveys. Incentives such as financial compensation
for participation help, but it can also be difficult to acquire funding for such
compensation.
Crucially, ethnographic studies often raise ethical challenges for
researchers as well. Given the level of contextual detail shared, it can be
difficult to maintain the kind of anonymity that respects and maintains
participant privacy. Researchers also need to consider the ethics of par-
ticipant recruitment. As researchers often study sites with which they have
some access, they may be familiar with the possible participants, potentially
exerting pressure to participate. As Fiona Copeland questions, “is it ethi-
cal to recruit participants who are so well known to the researcher? What
choice do such participants really have when it comes to taking part or
later on, withdrawing, given this close relationship?” (Copeland & Creese,
2015, p. 167). Building or sustaining a relationship with participants who
open themselves up to such research may also put researchers in challeng-
ing situations in deciding what to do with data that can put participants in
a negative light or which may be interpreted differently by the researcher
and the participants. Reflexivity plays an important role in ethnographic
research for these reasons. That is, it is imperative that researchers consider
their positionality within the research setting and its role in everything
from recruitment to interpretation (Li, 2020).

3.6 Study-in-focus
Because ethnographic studies of genre take many different approaches, it
is challenging to identify a single exemplary study in this line of inquiry.
Nevertheless, I share here a study that showcases a strong grounding in
ethnographic principles, an impressive array of data, and an approach that
leans toward the kind of deep theorizing described by Lillis (2008).
Wickman’s (2010) research centers around the laboratory notebook,
a genre used in tandem with lab activity. Lab notebooks have been the
object of inquiry from a genre perspective in other studies as well (e.g.,
Bazerman, 1988, 1999), offering a glimpse into research activity and the
communication of scientific knowledge. Wickman notes that several texts
were a part of the lab activity he observed, but that the lab notebook was
“the only one that is actually assembled in tandem with specific labo-
ratory tasks” (p. 269), directly linking research with “replicable textual
knowledge” (p. 269). The call-out box shares an overview of Wickman’s
study.
 Ethnographic Research 51

CITATION

Wickman, C. (2010). Writing material in chemical physics research: The

laboratory notebook as locus of technical and textual integration. Written
Communication, 27(3), 259–292.

Research questions
• How do notebooks function in the context of day-to-day scientific work?
• What types of writing and inscriptional practices inform notebook
production?
• How do scientists use notebooks to transform the material and techni-
cal dimensions of laboratory work into durable communal knowledge?

Context and population

The study explores research questions focused on genre-in-use through an
ethnographic case study of one postdoctoral research fellow working in a
research group at a liquid-crystal physics and materials sciences institute.
Dave was responsible for creating materials samples for the group, and the
laboratory notebook – as a documentation tool – was a critical genre in this
work.

Procedure
To understand the functions of the laboratory notebook, Wickman used
three common ethnographic research methods over a one-year period of
data collection: observation, visual documentation and inscription, and
interviews. Observations were extensive, encompassing the research group’s
daily meetings, lab activities, office activity, weekly group meetings, brain-
storming sessions, practice presentations, and weekly seminars with visiting
scientists. Wickman focuses specifically on Dave’s work, drawing on approxi-
mately 100 pages of observational fieldnotes generated over three phases of
the study. These fieldnotes revealed insight into the lab notebook’s role in
different laboratory activities.
Wickman also created visual representations of Dave’s lab work, includ-
ing digital photographs, the researcher’s graphic representations of Dave’s
work, and Dave’s own graphic representations of his work and digital pho-
tos. These visuals offered a glimpse into the symbiotic relationship between
the research activity and the notebook. Finally, Wickman’s study draws on
52 Christine M. Tardy

the transcript of a 45-minute interview with Dave, serving to build a more in-
depth understanding of his perspective on the notebook and its function in
his work. Wickman describes the interview as relatively open-ended, allow-
ing for broad reflection from Dave on the relationship between his research
and texts.
In his analysis, Wickman shares several valuable tools for analyzing a
complex dataset. For instance, a table summarizes the texts used in Dave’s
work, where they were produced and used, and the role they played in
his materials production. Excerpts from his fieldnotes, photographs of the
lab, interview extracts, visual representations, and notebook pages provide
detailed glimpses of the role of the notebook in Dave’s laboratory activity,
as Wickman narrates a story of the sophisticated semiotic network Dave
employs. Further, his analysis is informed by a constructivist approach to
texts and their production in which texts are more than inscriptions but are
rather means for transforming scientific practice into knowledge.

Findings
Wickman’s complex analysis of textual practice illuminates the precise ways
in which the laboratory notebook serves as much more than a textual mani-
festation of research. Rather, readers see the ways in which the notebook
mediates the scientist’s lab activity and the communication of that activity
outside of the lab. Put simply, the notebook “help[s] to construct rather than
simply report scientific research findings” (p. 284). Tracing how the scien-
tist’s inscriptions mediate knowledge construction, Wickman’s article offers
both a methodological approach for the study of scientific practice and com-
munication and a glimpse into the role that texts play in that process.

One exemplary aspect of Wickman’s study is its intertwining of various

forms of data to create a vivid picture of genre-in-practice. He shares
excerpts from his fieldnotes, photographs of the lab work, and summaries
of a typical series of tasks that his focal participant, Dave, carries out while
recording the work in his notebook. Wickman also uses visual depictions of
the lab space and of Dave’s processes of materials and textual production,
always contextualizing the genre as part of practice. It thus offers a strong
example of how multiple data sources can be brought together to build an
understanding of when, why, and how a genre is used.
Though he refers to the study as ethnographic, Wickman does not dis-
cuss the ethnographic principles that his work employs; nevertheless, these
 Ethnographic Research 53

are clearly visible. Gathering data over a one-year period allows him to build
a close familiarity with the research context and a trusting relationship with
the main participant. In sharing fieldnotes from the first week and the fifth
month of his study, we see development in detail as well as later moments
where the participant has begun inviting Wickman to ask questions during
his activity. Wickman’s study also illustrates how ethnographic researchers
can build a productive dialogue between emic and etic perspectives on data,
as he interweaves his interpretations with those offered and clarified by his
focal participant.
In the spirit of ethnographic research, Wickman’s study also engages
the tension between what Ramanathan and Atkinson (1999) refer to as
the “particularizability” and generalizability of practice. This interplay
between the particular contextual details and more general interpretations
of practice is carried out as Wickman moves from close descriptions of the
writing site to an interpretive discussion that engages the particulars with
the theoretical implications. He builds on Witte’s (1992) notion of inter-
text, for example, to elaborate on how viewing notebooks as practice (not
simply as text) allows us to see “the convergence of Dave’s technical work…
interpretive work…and rhetorical work” (p. 285). Wickman moves the
implications beyond simply understanding the roles and functions of the
lab notebook to the complex practice of scientific knowledge production
and the role of texts within that process. From this larger picture, the study
illuminates the variety of semiotic resources that contribute to production
of scientific texts and the ways that those texts ultimately render practice
into knowledge. And, as Wickman notes, this ethnographic lens sharpens
our understanding of genre as a social practice.

3.7 Future research directions

Given the rather significant constraints and challenges, ethnographically
oriented research of genre is not an approach that every researcher is inter-
ested in or even willing to adopt. Nevertheless, this type of research is cru-
cial in offering glimpses into genre as social practice. While textual lenses
create valuable pictures of the formal features of text that can perhaps most
directly inform pedagogy, ethnographic lenses offer essential reminders
that learning and using genres involves complex layers of rhetorical, subject-
matter, process, and formal knowledge (Tardy, 2009) that are developed in
social worlds. Simply generating more ethnographic research into genre is
an important future direction. A growing body of this type of inquiry can
build our theoretical understanding of genre more generally, and studies
in particular contexts – such as classrooms – can build a deeper knowledge
of the social environments in which common and valued genres are used,
54 Christine M. Tardy

assessed, learned, and taught. Ethnographic research, for example, is well

suited to building an understanding of how classroom environments and
instructional choices might facilitate (or complicate) genre learning and how
learners might adapt their genre knowledge built in one context to another.
In academic writing, growing attention is now being given to digital gen-
res, including emerging public-facing research genres that appear in blogs,
on the web, or through social media (e.g., Graham, 2020; Luzón, 2013;
Pérez-Llantada, 2013, 2021). Studies that bring together observational data
of these genres’ production and reception would be particularly useful in
developing a picture of not just how these genres differ from traditional
research genres in form but also in their social practice. Given the inter-
textual links among these research genres, studies that provide pictures of
when, why, and how writers recontextualize these genres would be especially
enlightening. Albero-Posac and Luzón (2021) offer key considerations and
guidelines for carrying out ethnographic research of digital communication.
Finally, a valuable path for future ethnographic studies into genre would
be those that specifically explore crosslingual genre learning and use (Tardy
et al., 2020). For example, ethnographic research has strong potential to
reveal insight into the kinds of environments that might support or con-
strain writers’ use of their full linguistic repertoire. Fully contextualized
studies can also provide better understandings of the sociolinguistic con-
texts in which genres are learned across languages and how surrounding
practices, values, and ideologies may shape writers’ multilingual genre
knowledge and practices.

References
Albero-Posac, S., & Luzón, M. J. (2021). Understanding academics online:
Ethnographic approaches to the analysis of online academic discourse and
practices. In E. Guillén-Galve & A. Bocanegra-Valle (Eds.), Ethnographies of
academic writing research: Theory, methods, and interpretation (pp. 61–81).
John Benjamins Publishing Company.
Atkinson, D., & Ramanathan, V. (1995). Cultures of writing: An ethnographic
comparison of L1 and L2 university writing/language programs. TESOL
Quarterly, 29(3), 539–568.
Bazerman, C. (1988). Shaping written knowledge. University of Wisconsin Press.
Bazerman, C. (1999). The languages of Edison’s light. MIT Press.
Berkenkotter, C. (2001). Genre systems at work: DSM-IV and rhetorical
recontextualization in psychotherapy paperwork. Written Communication,
18(3), 326–349.
Blommaert, J. (2007). On scope and depth in linguistic ethnography. Journal of
Sociolinguistics, 11(5), 682-688.
Blommaert, J., & Jie, D. (2020). Ethnographic fieldwork: A beginner’s guide (2nd
ed.). Multilingual Matters.
 Ethnographic Research 55

Brodkey, L. (1987). Writing ethnographic narratives. Written Communication,

4(1), 25–50.
Cimasko, T., & Shin, D. (2017). Multimodal resemiotization and authorial agency
in an L2 writing classroom. Written Communication, 34(4), 387–413.
Copeland, F., & Creese, A. (2015). Ethical issues in linguistic ethnography: Balancing
the micro and the macro. In P. I. De Costa (Ed.), Ethics in applied linguistics
research: Language researcher narratives (pp. 161–178). Taylor & Francis Group.
Dressen-Hammouda, D. (2014). Measuring the voice of disciplinarity in scientific
writing: A longitudinal exploration of experienced writers in geology. English
for Specific Purposes, 34, 14–25.
Flowerdew, J., & Wan, A. (2010). The linguistic and the contextual in applied genre
analysis: The case of the company audit report. English for Specific Purposes,
29(2), 78–93.
Glaser, B., & Strauss, A. L. (1967). The discovery of grounded theory. Aldine.
Graham, S. S. (2020). Misinformation inoculation and literacy support tweetorials
on COVID-19. Journal of Business and Technical Communication, 35(1), 7–14.
Green, J., & Bloome, D. (2005). Ethnography and ethnographers of an in
education: A situated perspective. In J. Flood, S. B. Heath, & D. Lapp (Eds.),
Handbook of research on teaching literacy through the communicative and visual
arts (pp. 181–202). Lawrence Erlbaum Associates.
Leki, I. (2007). Undergraduates in a second language: Challenges and complexities
of academic literacy development. Routledge.
Li, W. (2020). Ethnography: Origins, features, accountability, and criticality. In
J. McKinley & H. Rose (Eds.), The Routledge handbook of research methods in
applied linguistics (pp. 154–164). Routledge.
Lillis, T. (2008). Ethnography as method, methodology, and “deep theorizing”:
Closing the gap between text and context in academic writing research. Written
Communication, 25(3), 353–388.
Luzón, M. J. (2013). Public communication in science blogs: Recontextualizing
scientific discourse for a diversified audience. Written Communication, 30(4),
428–357.
Martin, J. R. (1992). Genre and literacy-modeling context in educational linguistics.
Annual Review of Applied Linguistics, 13, 141–172.
Miller, C. R. (1984). Genre as social action. Quarterly Journal of Speech, 70(2),
151–167.
Morton, J. (2016). ‘Adjacent worlds’: An analysis of genre at the intersection of
academic and professional communities. Journal of English for Academic
Purposes, 22, 54–63.
Noy, C. (2015). Writing in museums: Toward a rhetoric of participation. Written
Communication, 32(2), 195–219.
Paltridge, B., Starfield, S., Ravelli, L. J., & Tuckwell, K. (2012). Change and
stability: Examining the macrostructures of doctoral theses in the visual and
performing arts. Journal of English for Academic Purposes, 11(4), 332–344.
Paltridge, B., Starfield, S., & Tardy, C. M. (2016). Ethnographic perspective on
academic writing. Oxford University Press.
Parks, S. (2001). Moving from school to the workplace: Disciplinary innovation,
border crossings, and the reshaping of a written genre. Applied Linguistics,
22(4), 405–438.
56 Christine M. Tardy

Pérez-Llantada, C. (2013). The article of the future: Strategies for genre stability
and change. English for Specific Purposes, 32(4), 221–235.
Pérez-Llantada, C. (2021). Research genres across languages. Multilingual
communication online. Cambridge University Press.
Pole, C., & Morrison, M. (2003). Ethnography for education. Open University
Press.
Prior, P. A. (1998). Writing/disciplinarity: A sociohistoric account of literate activity
in the academy. Routledge.
Ramanathan, V., & Atkinson, D. (1999). Ethnographic approaches and methods
in L2 writing research: A critical guide and review. Applied Linguistics, 20(1),
44–70.
Ravotas, D., & Berkenkotter, C. (1998). Voices in the text: Varieties of reported
speech in a psychotherapist’s notes and initial assessments. Textile, 18, 211–239.
Sarangi, S. (2007). The anatomy of interpretation: Coming to terms with the
analyst’s paradox in professional discourse studies. Text and Talk, 27(5–6),
567–584.
Schryer, C. F. (1993). Records as genre. Written Communication, 10(2), 200–234.
Spack, R. (1997). The acquisition of academic literacy in a second language: A
longitudinal case study. Written communication, 14(1), 3-62.
Swales, J. M. (1990). Genre analysis: English in academic and research settings.
Cambridge University Press.
Swales, J. M. (1998). Other floors, other voices: A textography of a small university
building. Lawrence Erbaum Associates.
Swales, J. M. (2009). Incidents in an educational life: A memoir (of sorts). University
of Michigan Press.
Tardy, C. M. (2009). Building genre knowledge. Parlor Press.
Tardy, C. M., Sommer-Farias, B., & Gevers, J. (2020). Teaching and researching
genre knowledge: Toward an enhanced theoretical framework. Written
Communication, 37(3), 287–321.
Watson-Gegeo, K. A. (1988). Ethnography in ESL: Defining the essentials. TESOL
Quarterly, 22(4), 575–592.
Wickman, C. (2010). Writing material in chemical physics research: The laboratory
notebook as locus of technical and textual integration. Written Communication,
27(3), 259–292.
Witte, S. (1992). Text, context, intertext: Toward a constructivist semiotic of
writing. Written Communication, 9(2), 237–308.
SECTION II

Analyses, tools, and

topics


4
METADISCOURSE
Ken Hyland

4.1 Introduction to the approach and definition of genre

Metadiscourse is the commentary on a text made by its producer in the course
of speaking or writing. This commentary allows writers to guide, direct,
and interact with their readers and demonstrate a concern for their ability to
understand a text as it is intended. It assists them to recognize the writer’s
stance, see connections between ideas, and feel involved in what is being
said. Examples include frame markers, such as in conclusion, in the next sec-
tion and endophorics, like see table and as noted above. Metadiscourse, then,
reminds us of our responsibilities as writers to take readers into account. It
is a recipient design filter, which shapes a text out of consideration for an
audience based on an estimation of how it can be helped to process, com-
prehend, and agree with what is being discussed. By explicitly stepping into
a text in this way, we see the writer’s assessment of his or her audience and
therefore something of the context in which language is being used. As a
result, metadiscourse has become one of the leading approaches to studying
texts, so that a Google Scholar inquiry returns 29,500 articles containing
the term, 11,000 of these between 2017 and 2022.
The term metadiscourse was originally introduced by the structural lin-
guist Zelig Harris (1959) but not taken up in genre studies until Williams’
(1981) and Crismore’s (1983) attempts to refine and operationalize it by
creating functional subcategories. The work by Vande Kopple (1985) and
Crismore (1989) was instrumental in helping the concept gain traction by
making a firm distinction between propositional material, what is talked
about, and metadiscourse, which signals the presence of a text-organizing

DOI: 10.4324/9781003300847-6
60 Ken Hyland

and content-evaluating author rather than the subject matter. Their work
was foundational in identifying two broad categories of metadiscourse, bor-
rowing Halliday’s (1994) terms for language metafunctions: textual and
interpersonal. Halliday, however, saw these resources as not neatly separable
parts of a text, but as integral to every clause and intrinsic to all language.
They followed the labeling but not the concepts.
So metadiscourse analysts seek to describe the ways that writers (and
speakers) organize their texts to help readers (hearers) understand and
assess propositional information (Ädel & Mauranen 2010; Hyland, 2005a).
It is informed by a view of communication that sees language as not only
referring to the world, concerned with exchanging information, but also to
itself: with material which helps readers to organize, interpret, and evalu-
ate what is being said. It connects with Jacobson’s (1980) metalinguistic
function of language, which refers to language focusing on the text itself,
and with Halliday’s (1985) metaphenomena, which are “categories of the
language, not of the real world” (p. 271).
Despite this fairly straightforward characterization of metadiscourse,
there are several interpretations of the concept that we can see as spread
along a continuum ranging from a narrow text-centered view to a broad
interpersonal one (Hyland, 2017). At one end sits a view championed by
Mauranen (1993), which she refers to as metatext or text reflexivity. This
attempts to sharpen the concept by restricting it to expressions that refer
to the direction, purpose, and internal structure of the text itself (e.g., the
presentation is in four parts or we now turn to look at the disadvantages).
Further along the continuum, there are studies which extend this view to
include how writers refer to themselves, their readers, and their texts. Ädel
(2006), for example, adds features which refer to the writer or address the
reader (e.g., inclusive we and you will see) to those that comment on the text
itself (in addition, as a first step). Also positioned along the cline are alterna-
tive conceptions, such as Beauvais’ (1989) attempts to limit metadiscourse
to expositive illocutionary predicates (such as argue, illustrate, and deny) or
Ifantidou’s (2005) reformulation based on a relevance framework.
At the far end of the cline, analysts see a writer or speaker’s commentary
on his or her unfolding text as representing a set of interpersonal options.
Here metadiscourse comprises the resources writers employ to both organ-
ize textual material and make it comprehensible and persuasive by project-
ing themselves into their discourse. Hyland (2005a; Hyland & Tse, 2004),
for example, seeks to capture the interpersonal character of communication
in the concept, with a distinction between interactive and interactional
resources (Thompson, 2001). Interactive refers to the writer’s manage-
ment of the information flow to guide readers through a text, setting out
the structure, referring to sources, and linking parts of the discourse, and
 Metadiscourse 61

interactional concerns the writer/speaker’s interventions to express a stance

and build a relationship with readers.
The interpersonal perspective therefore collects together the linguistic
devices we use to shape our messages for particular listeners or readers: a
coherent set of interpersonal resources which organize a discourse or the
writer’s stance toward either its content or the reader. In this view both
organizational and evaluative features are necessary to create an inter-
personally effective text, although some researchers feel that this broad
interpretation weakens the term by trying to include too much (Ädel &
Mauranen, 2010). Supporters, however, argue that the interpersonal model
offers a more inclusive view of how we monitor our production by making
decisions about the effects we are having on our audience and that a fin-
ished text is an outcome of this awareness of the reader.
It should be clear from this description that metadiscourse research is
a text-oriented endeavor. Analysts study (mainly) written and (sometimes)
spoken discourse to identify salient features of interaction. The notion of
genre is therefore central to metadiscourse studies as the type of text informs
the choices writers make and how metadiscourse items build a coherent
pattern of interaction. Genre, then, is understood as how we typically use
language to respond to recurring situations, which means that it is both a
way of categorizing texts and the situations in which they occur. It points to
the fact that writing is a practice based on expectations: the reader’s chances
of interpreting the writer’s purpose are increased if the writer takes the
trouble to anticipate what the reader might be expecting based on previous
texts they have read of the same kind. In other words, we possess a schema
of prior knowledge which we share with others and can bring to the situa-
tions in which we read and write to express ourselves effectively.
The genre conventions associated with particular schema are, of course,
specific to particular contexts and communities, acknowledging the impor-
tance of local practices. Metadiscourse researchers therefore tend toward a
view of genre associated with English for Specific Purposes (ESP). This is
distinguished from other genre models by focusing on the texts created to
fulfill group purposes in specific contexts of use rather than ways of writing
available in the language as a whole. It sees genre, then, as a class of struc-
tured communicative events employed by specific discourse communities
whose members share broad social purposes (Swales, 1990). These pur-
poses are the rationale of a genre and help to shape the ways it is structured
and the choices of content and style it makes available (Johns, 1997). At
the heart of ESP perspectives, then, are the concepts of community and
social purpose. Neither of these ideas has proved to be as straightforward
as first imagined, but they enable ESP to maintain a narrow concept of
genre which allows researchers to distinguish similar texts in terms of the
62 Ken Hyland

purposes recognized by members of a relevant community. In terms of

metadiscourse research, this particularity is essential.
This is not to say that metadiscourse research has been confined to writ-
ten academic genres. While this has been the mainstay of research, a recent
paper using bibliometric methods to track metadiscourse in the core collec-
tion of the Web of Science between 1983 and 2020, found a range of top-
ics, genres, and submitting countries (Hyland & Jiang, 2022). The results
show the importance of academic and business writing, the growth of stud-
ies comparing disciplines, languages, and genres, as well as the increasing
predominance of the interpersonal model. Academic discourse, academic
writing, and research articles have been consistent themes of metadiscourse
research since the beginning. Metadiscourse research therefore reflects, or
perhaps drives, the huge interest in interaction in academic texts. These
papers cover a wide area of interests with expert and student texts com-
monly researched. Studies include those addressing:

• A particular domain, such as student writing, business presentations, or

journalism.
• Written and spoken genres (e.g., research articles, essays, and TED
Talks).
• Comparative studies exploring variations across genres, modes, lan-
guages, first language writers, student proficiencies, and time.

In addition, while a great deal of research seeks to unpack and describe

texts using an overall model of metadiscourse, a considerable number of
studies focus on a particular aspect of the model, such as hedging or stance.
Examples of these areas will be discussed below.

4.2 Goals
The study of metadiscourse, informed by these different rhetorical and
social contexts, has been hugely influential in informing applied linguists
of the ways interaction works in (mainly) written and (occasionally) spoken
discourse. In addition, it has also been instrumental in revealing something
of the different communicative purposes of genres and the epistemological
norms and social practices of different linguistic and professional commu-
nities. For many researchers, moreover, pedagogical goals underlie these
descriptive and explanatory objectives.
Metadiscourse sees a writer/speaker’s ability to make statements about
the world coherent, intelligible, and persuasive to a particular audience as
key to effective communication. As such, researchers are careful to con-
fine their studies to a particular sphere of interaction and restrict their
 Metadiscourse 63

conclusions to that sphere. As noted above, one such sphere is the pat-
terning of metadiscourse in a particular domain or genre, such as in stu-
dents’ writing, Twitter posts, newspaper articles, blogs, undergraduate
textbooks, and social media discussion groups. Metadiscourse research has
also sought to describe how spoken interactants work to make sense to each
other and achieve successful understandings, ranging from financial earn-
ings calls (where companies declare their results) to classroom discourse.
The vast majority of metadiscourse research, however, focuses on an aca-
demic register and addresses research articles, and often their introductions
and abstracts. Many of these studies have a contrastive goal, to determine
genre or disciplinary differences in how writers/speakers construct argu-
ments and make their claims persuasive to particular audiences in specific
communicative contexts.
Contrastive studies have not only sought to describe, and often explain,
variations across genres and communities, but also modes, languages, first
language writers, student proficiencies, and time. The goal of this research
is to provide a map of professional practices, cultural and linguistic differ-
ences, traditional and digital distinctions, spoken and written variations,
or discoursal change. A productive area, for instance, has been to explore
disciplinary variation in metadiscourse use and to relate these differences
to the epistemological and social practices of the more discursive social
sciences and humanities compared to the physical sciences. Similarly,
many researchers are interested in the metadiscourse use of various first
language (L1) writers, either writing in their own language or in English,
to determine characteristics of both linguistic and user differences.
It should also be mentioned that not all studies address metadiscourse in
its entirety and often focus on specific aspects of metalinguistic meanings.
Interactional metadiscourse, for example, and particularly hedges, boost-
ers, and self-mention, have received considerable attention. Researchers
have been keen to understand how stance and engagement are managed in
particular genres and by particular writers, while research into the structur-
ing of academic discourse has also been popular.
To summarize, metadiscourse research tends to ask questions such as:

• What metadiscourse features distinguish specific genres or parts of a

genre?
• How do the same writers vary their use of metadiscourse in different
contexts?
• How does stance/engagement/interactional metadiscourse help con-
struct persuasive texts?
• How do disciplines or other communities vary in their use of
metadiscourse?
64 Ken Hyland

• What interactional differences characterize related spoken and written

genres?
• How does metadiscourse distinguish the texts of experts and novices?
• How does metadiscourse use distinguish learners of different
proficiencies?
• What challenges do novice writers experience when using/understand-
ing metadiscourse?
• How does gender influence metadiscourse use?
• How has metadiscourse changed over time in particular genres and how
can we explain this?

4.3 Common research methods

It is difficult to imagine metadiscourse research without the use of corpora.
Essentially a corpus is a collection of naturally occurring, computer-read-
able texts, often comprising many millions of words, which is considered
more or less representative of a particular domain of language use. While
a corpus does not contain any new theories about language, it shows fre-
quencies of words and strings which can indicate the relative importance
given to them by users. As a result, corpora offer fresh insights on familiar,
but perhaps unnoticed features of language use, replacing intuition with
evidence. Using text retrieval and concordancing software such as AntConc
(Anthony, 2022), analysts (or students) can isolate, count, and see patterns
of metadiscourse in large numbers of texts to better understand the ways
it is used in a particular genre, register, or community. The results suggest
that insiders, because of their regular encounters with a genre, develop
a certain familiarity with the ways that language is typically used in that
genre. They come to use and understand a repertoire of metadiscourse
choices almost automatically as being the most effective ways of achieving
their goals.
There are various ways of conducting corpus research, but Tognini-
Bonelli (2001) makes a distinction between corpus-based and corpus-driven
studies. The former uses a corpus as a source of examples to check the
researcher’s intuitions, while a corpus-driven analysis is more inductive:
the corpus is the data and the basis for any discovery of patterns of use.
Some metadiscourse studies adopt the latter approach, delegating the task
of identifying features to the computer to generate lists of high-frequency
items that can be whittled down and checked for their potential metadis-
course function. In most cases, however, researchers start by focusing on
potentially productive items based on previous metadiscourse studies, from
interviews with writers, or from the wider literature. These features can be
studied for how they link text participants as interactants in ways which are
 Metadiscourse 65

consistent with the norms of the community, genre, or other dimensions

of context.
It is important to point out here that the devices searched for are only
potential candidates for metadiscourse status. They are a starting point of
high-frequency items that commonly work as metadiscourse in a particular
register. Items need to be examined in their sentential contexts using con-
cordance software to ensure they are performing metadiscourse functions.
Additionally, it is also common to conduct manual sweeps of the corpus,
reading long representative samples to discover unexpected realizations,
such as metadiscourse nouns (e.g., interpretation, shortcoming, reason) (Jiang
& Hyland, 2016) in abstracts, hypertext links in commercial webpages
(Gonzalez, 2005), gestural silence in classroom talk (e.g., a teacher’s wrist
grip to indicate attention or scanning the class for a response) (Fogarty-
Bourget et al., 2019), or surprise markers in research articles (unexpectedly,
amazingly) (Chen & Hu, 2020).
Because of the polypragmatic nature of language, where the same forms
can perform more than one function, checks on the reliability of analysts’
coding are essential when identifying metadiscourse features. Inter-coder
reliability measures are frequently used with a second trained coder or co-
author independently examining a sample of data, usually between 5% and
30% of cases, comparing their coding decisions and resolving disagreements
through discussion. This process is often iterative so that the two coders
gradually refine their judgments and improve their agreement with a final
statistical measure such as Cohen’s kappa values to indicate the extent of
inter-coder reliability (e.g., Chen & Hu, 2020; Hyland & Jiang, 2020). An
alternative method focuses on a more limited set of words, lexico-grammat-
ical patterns, or phrases, that are known to reliably express metadiscoursal
constructs (see Chapter 7 of this volume on multidimensional analysis).
Statistical measures are also often employed to determine the significance
of differences in results, either of individual features or across corpora.
Typically, frequencies of target features are normalized to 1,000 words to
control for the varying lengths of texts and to facilitate comparison. These
are then typically analyzed using inferential statistics such as ANOVA and
multiple regression, or less commonly, binominal measures, such as binary
logistical regression, to capture distribution patterns.
While corpus approaches dominate metadiscourse research, analysts
also occasionally seek the views of insider informants, either writers or
readers of the examined texts. This involves recording and transcribing
semi-structured, discourse-based interviews. The key advantage of this
approach is that both quantitative and qualitative data are collected, ena-
bling the analyst to support their insights and hypotheses about user moti-
vations for language choices. Thus, Hyland (1998) interviewed scientists
66 Ken Hyland

concerning their use and views of hedging in biology research articles

while McGrath and Kuteeva (2012) conducted interviews with the five
authors in their corpus of mathematical research articles. In both cases
the researchers gained insights into the intended effects of the linguistic
resources used and how they responded to stance and engagement markers
as readers.

4.4 Example studies
In this section I briefly discuss a number of sample studies to illustrate
some of the issues raised earlier. The examples are selected to highlight dif-
ferent aspects of metadiscourse research, exploring some of the main ques-
tions and genres studied. The studies are summarized in Table 4.1.
The first study is Hyland’s (2004) early paper developing his model
of metadiscourse (Hyland, 2005a) and elaborating on the idea of writ-
ing as a social engagement in which writers project themselves into their
discourse to signal their attitudes and commitments. The paper explores
how advanced second language (L2) writers deploy these resources in a
high-stakes research genre. It employs both qualitative and quantitative
approaches, comprising frequency counts and text analyses of 240 doctoral
and masters dissertations by Hong Kong students, together with interviews
with the writers themselves.
The corpus comprised 20 masters and 20 doctoral dissertations from
each of six disciplines across the academic spectrum. The dissertations were
scanned from hard copies in the libraries of five Hong Kong universities to
produce an electronic corpus of four million words, 2.6 million in the PhDs
and 1.4 million in the masters texts. Because some forms can serve more
than one function, a sample of two texts from each discipline was initially
coded manually by the author and a research assistant. The entire corpus
was then searched electronically for all these expressions using a text analy-
sis program. To limit the laborious work of coding very high-frequency
devices such as modals and conjunctions, a sample of 50 from each disci-
pline was analyzed to ensure they were functioning as metadiscourse and
results extrapolated. All frequencies were then normed to occurrences per
10,000 words to facilitate comparison across corpora of different sizes. In
addition, two masters students and two PhD students from each discipline
were interviewed (24 students) as a way of both gaining insights into the
text data and of discovering something about their own preferences and
thoughts on disciplinary practices. Interviews used a semi-structured for-
mat in English or Cantonese as the interviewee preferred.
Results showed that the texts were dominated by hedges, comprising 41%
of interactional uses, and transitions, making up a similar figure in the
TABLE 4.1 
Example metadiscourse studies

Authors Context Research Questions (RQs) Data Collected Comments on Study

Design

Bax et al. (2019) • Exam essays by • Is there a difference in • 281 metadiscourse • A carefully designed
various L1 students the overall quantity of markers from comparative study
at three levels of metadiscourse markers used by 900 exam scripts using internationally
English proficiency lower- and higher-level writers? at CEFR B2-C2 recognized
• Is there a difference in the levels using the Text proficiency scales
quantity of interpersonal Inspector tool and
markers used by higher- and human analysts
lower-level writers?
• Is there a difference in the
variety of metadiscourse
markers used by higher- and
lower-level writers?
• Is there a difference in the
quantities of individual
markers in the following
categories used by higher-
and lower-level writers:
emphatics, endophorics,
logical connectives, and frame
markers?
(Continued)
Metadiscourse
67
TABLE 4.1 Continued

Authors Context Research Questions (RQs) Data Collected Comments on Study

Design

Dafouz (2008) • Newspaper articles • Which metadiscourse • 20 opinion articles • Small corpus size
68 Ken Hyland

in two prestigious categories predominate in in each of The (48,500 words), but

British and Spanish this genre, and how they are Times and El País the qualitative aspect
broadsheets distributed according to cross- newspapers is interesting and
linguistic preferences? • Informant original
questionnaires
to determine
persuasiveness
Hyland (2004) • Dissertations • How do writers perceive and • 120 doctoral • An exploratory study
written by doctoral engage with their disciplines and 120 masters testing the model
and masters through their deployment of dissertations written using a substantial
students in Hong interpersonal features in their by Hong Kong self-compiled corpus;
Kong dissertations? students of four it is a good example
million words of corpus analysis of
metadiscourse
Hyland and • Papers from same • To what extent are there • Six papers from • Innovative
Jiang five journals in four changes in the use of the top five SSCI study tracking
(2018) disciplines at three metadiscourse over the journals in four metadiscourse
periods over the past 50 years across four disciplines at three change and linking
past 50 years disciplines? points in time: 1965, these to changes in
1985, and 2015 academic practices
Hyland and Zou • Student three- • How do 3MT presenters • 120 3MT • The study is
(2022) minute thesis connect with their audiences? presentations from well-designed to
(3MT) presentations • Are there disciplinary the physical and collect highly rated
from several differences in these methods? social sciences presentations and to
disciplines and • How can we account for these downloaded from compare broad fields
university websites differences in terms of 3MT university websites
in various countries genre features? and transcribed
Kawase (2015) • PhD theses • How do L2 English research • Introductions from • The study shows
subsequently writers use metadiscourse in eight thesis/article the importance of
published as the introductions of their PhD pairs carefully selecting
research articles by theses and in subsequently comparative texts for
Japanese students in published research articles? contrastive analysis
Australia
Metadiscourse
69
70 Ken Hyland

interactive category. More importantly, there were considerable variations

across the corpora from the two degrees. Masters students used slightly
more interactional metadiscourse and the doctoral writers substantially
(10%) more interactive forms. The PhD dissertations also contained far
more metadiscourse overall, with 73% of all cases in the study in terms of
raw numbers and 35% more when normed for text length. There were also
considerable variations across disciplines, with the more discursive social
science disciplines employing more metadiscourse overall (56% of the nor-
med count) and over 60% of the interactional features.
The second focus study illustrates the cross-linguistic strain of metadis-
course research and features Dafouz’s (2008) paper on metadiscourse in
English and Spanish. Such comparative studies have been a productive area
of metadiscourse research with studies examining Chinese (Wang, 2019),
Iranian (Crismore & Abdollehzadeh, 2010), Finnish (Mauranen, 1993),
Slovene (Peterlin, 2005), Swedish (Ädel, 2006), and other languages.
Dafouz’s study is interesting as it moves away from an academic register
to examine the role that metadiscourse plays in constructing persuasion in
two influential newspapers, the British The Times and the Spanish El País.
She created a small corpus of 40 opinion columns, 20 from each newspaper,
matched for the topic to assist comparability. Both manual and automatic
searches for metadiscourse items were conducted and subjected to nonpara-
metric statistical measures (Mann–Whitney U test) as a conservative meas-
ure of differences between the English and Spanish data based on their
ranks below and above the median. Results were standardized to 1,000
words to compare frequencies.
In addition, the researcher selected six texts from each corpus containing
varying amounts of metadiscourse and devised a questionnaire concern-
ing informants’ sense of the persuasiveness of the texts. The questionnaire
was based on Connor and Lauer’s (1988) model for persuasive writing and
asked respondents to judge rational, credibility, and affective appeals. These
were sent to native speakers of the target languages who taught English
or Spanish as a foreign language and 67 were returned. Respondents were
asked to rank the texts for persuasiveness and offer comments regarding
their views.
Findings suggest that both textual and interpersonal metadiscourse
markers are present in English and Spanish newspaper columns, but the
Spanish texts used more textual (interactive) markers, and the English texts
used more interpersonal (interactional) markers, although the differences
were not significant. Hedges were the most frequent device overall and log-
ical markers and code glosses showed the greatest differences. Regarding
the persuasive effect, informants agreed that a balanced number of both
 Metadiscourse 71

textual and interpersonal markers was necessary to make the text persuasive
and reader-oriented.
The third example looks at Hyland and Zou’s (2022) study of one cat-
egory of metadiscourse in a spoken genre: engagement in 3 Minute Thesis
(3MT) presentations. Engagement refers to features such as second-person
pronouns, directives, and questions which seek to involve readers or hear-
ers. The audience has to be hooked and led to a desired conclusion, and
this is no more urgent than in the competitive environment of the 3MT
presentation. Here doctoral students present their research to a heteroge-
neous, non-specialist, audience using only one static slide in just 180 sec-
onds. The researchers applied Hyland’s (2005b) engagement framework to
a corpus of 120 3MT presentations, taken equally from the hard and social
sciences, and transcribed from videos posted on public domain sites such
as YouTube, threeminutethesis​.or​g, and university websites. All the pres-
entations were given by PhD students and were the top three finishers of
university-sponsored competitions.
This corpus of 54,000 words was separated into hard and soft sci-
ences and searched for 320 common engagement features using AntConc
(Anthony, 2019) with additional items added after a manual trawl of the
data for items not on the original list. All examples were then concordanced
and manually checked to ensure that they performed the engagement func-
tion assigned to them. Each author independently coded a 25% sample of
each corpus and achieved an inter-coder agreement of 95%. Intra-reliability
tests were also conducted by the second author recategorizing 20% of the
cases two weeks after the initial coding with full agreement. Finally, the
frequencies were normed to 1,000 words to allow for cross-corpora com-
parison and significance determined using a student t-test in SPSS.
The results show that all speakers are conscious of the need to engage
with their audiences, but normed results revealed 33.3 cases per 1,000
words in the social science talks and 53.7 in the hard fields with a statisti-
cally significant difference (Log Likelihood = 45.19, p < .001). In fact, speak-
ers from the physical and life sciences used more of every feature except
questions. This result is in contrast with studies of other academic genres,
which show that social scientists take more explicitly involved and personal
positions than those in science and engineering fields. This is because writ-
ers are less able to rely on the explanatory value of accepted procedures and
have to work harder to involve and persuade their readers with the force
of their discourse (Hyland, 2005a). So, while there are disciplinary prefer-
ences in the use of engagement features in this genre, speakers’ decisions
are based less on their considerations of discipline than their assumptions
about the knowledge and interests of their audiences. The study underlines
72 Ken Hyland

the view that communication is only successful to the extent that we are
able to create an appropriate relationship with our interlocutors.
Kawase’s (2015) study compares metadiscourse used in texts with the
same content but in different genres, examining metadiscourse in PhD the-
ses and subsequently published research articles. The researcher collected
eight theses in applied linguistics written by Japanese students at Australian
universities along with articles from them published in international jour-
nals. The selected research articles were the PhD authors’ initial major
publications, and the subject matter similarity between the genres allowed
analysis of how writers at a similar stage of their career and academic lit-
eracy development use metadiscourse in response to genre expectations.
After identifying forms based on Hyland’s (2005a) taxonomy, the analy-
sis examined the total number of metadiscourse items employed in each
text and occurrences of interactive/interpersonal items as well as individual
items. Texts were normalized to 100 sentences but no inter-coder reliability
or statistical tests seem to have been conducted.
The analysis showed that most writers used more metadiscourse in their
article introductions. The greatest differences were in phrases referring to
previous research, less reference to other parts of the text, and still less
use of phrases signaling authorial presence. The author explains the dif-
ferences in terms of genre-specific requirements, particularly that writers
of PhD thesis introductions present previews of the subsequent chapters.
He further suggests that the variations can be ascribed to the nature of the
PhD thesis as an educational genre and that of research articles as a profes-
sional genre in which writers need to endure severe competition to get their
manuscripts published.
The fifth example study is also a comparative analysis but explores
changes in the use of metadiscourse in published articles both by discipline
and over time. Hyland and Jiang (2018) took six articles from each of the
same five journals in four disciplines at three periods over the past 50 years:
1965, 1985, and 2015. The journals were chosen according to their five-
year impact factor and disciplines were selected to obtain a cross-section
of academic practices, representing both soft and hard sciences. Overall,
the corpus comprised 30 articles from each discipline from each year, 360
papers of 2.2 million words. The 12 sub-corpora were each searched for
metadiscourse items on Hyland’s (2005a) list and others were added by
sweeps through the corpus, making 500 items. The authors decided to
omit both and and or from the counts of transitions as these are default
options of marking conjunctive relations of addition and alternation rather
than rhetorical strategies. To avoid counting forms rather than the func-
tions they were performing, each author then independently checked a
sample of 15% of cases with an inter-coder agreement of 95%.
 Metadiscourse 73

The results showed writers are now using 11.5% more metadiscourse
than 50 years ago, but while there had been a statistically significant
increase in interactive forms, interactional types had significantly decreased.
Surprisingly, perhaps, interactional metadiscourse showed a marked decline
in the discursive soft knowledge fields and a substantial increase in the
science subjects. Given the attention stance has received from analysts
in recent years, this is rather surprising. Boosters and attitude markers
showed the greatest falls over the period, indicating changing preferences
for strong authorial standpoints, although self-mention increased substan-
tially. Interactive resources, in contrast, saw a significant rise, particularly
in evidentials (up 97%) and code glosses (59%). The former are references
to external sources (e.g., see X, refer to Y) while the latter clarify or rephrase
statements or words (for instance, in other words). Their greater use perhaps
reflects the growing complexity of scientific research or its increasing dis-
semination to less specialized audiences. Overall, then, the study found that
writers are now using more features to guide readers through more explic-
itly cohesive texts and fewer features to take a personal stance and engage
directly with readers. Moreover, there seems to be a growing convergence
between the hard and soft knowledge fields in the use of metadiscourse.
My final example, by Bax et al. (2019), is another comparison study, but
this one looks at the use of metadiscourse by L2 students across three levels
of English proficiency. The Council of Europe indicates that one of the key
areas of development in English at more advanced levels of proficient writ-
ing is an increasing variety of discourse markers and acknowledgment of
the intended audience. This study represents the first large-scale project to
investigate metadiscourse use in L2 learner writing. The researchers pared
down a large sample of 1,200 pass level scripts from the Cambridge English
General English exams to 300 papers at each of three levels: B2 First, C1
Advanced, and C2 Proficiency (formerly called FCE, CAE, and CPE). To
provide more reliable comparisons, only descriptive, expository, and argu-
mentative scripts were selected.
Following a review of the literature on the overuse or underuse of meta-
discourse forms by L2 students, the authors set out to discover if there is a
difference in the overall quantity and variety of these markers in the work
of writers of different proficiencies. The study followed Hyland’s (2005a)
model but distinguished the subcategories of frame markers (i.e., sequenc-
ing, label stages, announce goals, and topic shifts) giving a total of 13 cat-
egories. The study was carefully planned by using a source of data, which
offered reliable distinctions of proficiency levels, and focused on established
linguistic features. Only topics and text length were variable. Care was also
taken to ensure a balanced mix of L1s in each of the three levels. Analysis
was conducted using the Text Inspector tool (Bax, 2012) with trained
74 Ken Hyland

human analysts checking six scripts in groups and then coding another
set of six individually. Following discussions to reach agreement, the three
researchers then independently reviewed the computer outputs 100 texts
for each level, proposed changes, and agreed on coding. Results were then
normed to 100 words and statistically compared across the three levels.
Nonparametric tests were used since the sample was not normally distrib-
uted, with Kruskal-Wallis tests used to identify overall differences across
the three levels and Mann–Whitney U tests to identify differences in post-
hoc comparisons.
The findings showed that higher-level writers used fewer metadiscourse
markers than lower-level writers, especially emphatics, hedges, label stages,
person markers, relational markers, and topic shifts. But they used more
endophorics and evidentials and a significantly wider range of 8 of the 13
classes. The study also points to the value of analyzing not only the distri-
bution of metadiscourse categories but also the items themselves. Focusing
only on emphatics, endophorics, logical connectives, and four frame mark-
ers, the study found some evidence of more proficient writers using fewer
of the highest-frequency items in favor of a more varied range of markers.

4.5 Issues and challenges

Critics of metadiscourse argue that the diverse perspectives on it point to
fundamental conceptual difficulties with the term and an inability to pin
down what it is. One source of concern is, for example, that the traditional
distinction between propositional and metadiscourse elements of a text
cannot be sustained in practice. Items often identified as metadiscoursal,
such as therefore, in contrast, and as a result of, can act as metadiscourse
by connecting steps in an argument or work propositionally to connect
events in the world outside the text (Hyland, 2005a). Propositional mate-
rial is what is talked about – that is, what can be affirmed, denied, doubted,
insisted upon, qualified, regretted, and so on. Metadiscourse, on the other
hand, is what signals the presence of a text-organizing and content-evalu-
ating author rather than the subject matter. The meaning of a text is the
result of these two elements working together. Metadiscourse researchers
therefore tend to sidestep a rigid distinction and instead look for rhetorical
functions which writers and speakers use to talk about their own talk and
seek consistency in coding.
It is also worth noting that separate analyses of metadiscourse items with
single functions could result in missing the possible interactions between
them (Crismore et al., 1993; Hyland & Tse, 2004). In other words, while
metadiscourse theorists tend to see textual, interpersonal, and proposi-
tional (ideational) elements of the texts as discrete and separable, Halliday
 Metadiscourse 75

(1994) reminds us that each sentence is likely to combine elements of all

three. Discourse is a process in which writers are simultaneously creating
propositional content, interpersonal engagement, and the flow of text as
they write.
A second analytical difficulty is deciding what constitutes an instance of
metadiscourse. Metadiscourse can be realized in a variety of ways and by
units of varied length, from individual words to whole clauses or sentences.
The size of the linguistic unit is important as longer units might encompass
smaller units so that our conclusion could be categorized as an example of
a frame marker signaling an upcoming text segment, or as two units with
our also coded as self-mention. Identifying individual cases is therefore dif-
ficult, and indeed, can vary from one analyst to another.
A third issue is that the use of predefined sets of lexical items for corpus
searches can imply that the function of each form is unvarying (e.g., Ädel
& Mauranen 2010). The formal heterogeneity of metadiscourse means that
functions may be performed in different ways or items perform more than
one function simultaneously. As I mentioned in the section on research
methods, the same forms can convey different categories of metadiscourse,
so that quite can be a hedge (quite good) or a booster (quite extraordinary),
while particular functions, such as concessive connectives, can be expressed
in numerous ways (even if, of course, admittedly, although). This kind of
category overlap is well-known in discourse analysis and is a consequence
of the multifunctionality of language itself. It points to the need to see lists
of items as potentially performing metadiscourse functions, as a starting
point of high-frequency items that commonly work as metadiscourse in a
particular register. Examination of individual uses is crucial. Metadiscourse
studies therefore underline, rather than resolve, the problem of polyprag-
matic meanings.
A final concern is that frequency counts do not convey the overall
amount of metadiscourse in a corpus, but simply compare different patterns
of occurrence of metadiscourse in corpora of unequal sizes. As we can see
in the examples in the previous section, researchers tend to norm frequen-
cies to a given word length to facilitate comparison. We still need to be wary
of overall frequencies, however, as some metadiscourse items may be entire
clauses rather than words and that longer texts are likely to include more
text-organizing interactive metadiscourse than shorter texts.
Despite these issues, metadiscourse is seen as a useful way of viewing
how writers understand the context in which they are working by using
language to take readers’ needs, understandings, and existing knowledge
into account. Analysts regard writers’ successful management of these local
rhetorical resources as a key means of achieving their immediate commu-
nicative objectives.
76 Ken Hyland

4.6 Study-in-focus
In this section, I discuss a metadiscourse study in more detail, focusing
on Zou and Hyland’s (2019) study exploring differences in published peer
research and corresponding blogs for a general audience without special-
ized knowledge. Because we had examples of the two genres written by the
same authors and on the same topic, we were able to focus on genre differ-
ences. A summary of the study is provided in the call-out box, and you are
advised to read this and to consider the procedure followed.

Citation
Zou, H., & Hyland, K. (2019). Reworking research: Interactions in academic
articles and blogs. Discourse Studies, 21(6), 713–733.

Research questions
• How is the content of journal articles recontextualized in blogs?
• What differences are there in the use of stance and engagement features
in articles and blogs by the same authors?

Context and population

This study traced how writers recontextualize the contents of their published
journal articles as academic blogs by compiling two corpora comprising 30
blog posts corresponding to 30 previously published research articles writ-
ten by the same authors on the same topics. The articles were published in
SSCI journals in the social sciences and the blogs published on the LSE site,
which is one of the world’s most influential and prestigious hubs for the
exchange of research findings in the social sciences.

Procedure
The two corpora were searched for Hyland’s (2005a) stance and engagement
features using AntConc (Anthony, 2019) with additional items added after a
thorough reading of the data. All retrieved items were manually checked
to ensure that each performed the function to which it was assigned. A
10% sample was independently coded by both authors with an inter-coder
agreement of 96%. Intra-coder reliability tests were also conducted by each
author recategorizing 20% of the cases two weeks after the initial coding
with full agreement between the first and second categorizations. Finally,
 Metadiscourse 77

the frequencies of features in each category were calculated after normal-

izing the results to 1,000 words to allow for cross-corpora comparison. The
frequencies for each feature were analyzed using SPSS to determine log-like-
lihood statistical significances.

Findings
The study found 68.7 stance items per 1,000 words in the blogs compared
with 50.1 in the articles, 10.8 engagement features per 1,000 words in the
blogs, and 6.6 in the articles. There are significantly more interactive fea-
tures in the blog posts than in the articles (Log Likelihood = 31.15, p < .01 for
stance, and Log Likelihood = 9.51, p < .01 for engagement), with engage-
ment features being particularly heavily represented in the blogs with 61%
more examples than in the articles. The results also indicate that academics
draw on the full range of interactional devices. These are determined by both
the target audiences and rhetorical purposes of each genre, broadly speak-
ing, to inform general readers of scientific information in one and to gain
credit for their research claims in the other.

The rationale for the study was twofold: first, to explore the concept
of interactional recontextualization and how writers reconfigure their
research for a very different audience; and second, to describe blog features
for writers new to the genre. The reformulation of academic propositions
not only involves reframing knowledge, but reconstructing writer-reader
interactions to suit new conditions. It requires bloggers to adopt a, perhaps
unfamiliar, informality in their writing by the use of self-reference, openly
evaluative and affective commentary, and imagining the co-presence of het-
erogeneous readers. Responses are quick to appear, sometimes critical and
often blunt, encouraging circumspection tempered with the need to convey
ideas with conviction and authority. These features mean that blogs are like
no other academic genre writers may be familiar with, such as peer reviews
and book reviews, and they require a very different orchestration of rhetori-
cal resources.
The strength of the paper, however, lies not only in its originality and
potential usefulness to academics who are increasingly encouraged to raise
their visibility by reaching wider audiences. It is also an example of a care-
fully designed study which illustrates something of the challenges of con-
ducting metadiscourse research. The researchers explain that selecting
texts from two contexts written by the same authors helps to eliminate any
78 Ken Hyland

possible idiosyncratic preferences of individual writers by aggregating fea-

tures of the texts. It allows us to see how writers tailor their discourse for
audiences with different knowledge and processing needs. This approach is
therefore less concerned with the on-line processing of individual writers
and more with the different purposes for which language is being used,
allowing us to see the typical characteristics of the genres themselves rather
than of the authors.
The texts, both blogs and articles, are from prestigious peer-reviewed
sources and are directly comparable. The authors also decided to focus on
interactional features which, unlike text-organizing interactive resources
(such as we now turn to or in the next section), do not vary according to the
length of the text. Other tests to determine the integrity of the subsequent
results conducted by the authors were manual sweeps to include unex-
pected items and inter-coder reliability checks on samples of cases to ensure
accuracy of the coding. A 10% sample was independently coded by both
authors with 96% agreement and intra-coder tests also conducted by each
author recategorizing 20% of the cases two weeks after the initial coding.
While the authors do not describe specific classroom strategies, there
are clear pedagogical spin-offs from the study. The blog offers academ-
ics considerable advantages. This is an opportunity to present themselves
and promote their work to wider audiences beyond the confines of their
discipline. But in addition to the advantages of sharing academic knowl-
edge, blogs help academics promote a different and increasingly valued side
to themselves: as publicly minded intellectuals contributing to the wider
understanding of science. By explicating the linguistic features of blogs,
which are more personal and evaluative than we find in articles and which,
in combination, create a sense of proximity and dialogic exchange, novice
writers may be assisted to present their work more effectively.

4.7 Future research directions

There are clearly opportunities for further research in identifying new
metadiscourse items, describing metadiscourse patterns in other domains
and genres, comparing particular kinds of texts and, importantly, patterns
of metadiscourse in other languages or by speakers of other languages writ-
ing in English. Chinese, Iranian, and Spanish are the languages most fre-
quently analyzed (e.g., Gong et al., 2021; Khabbazi-Oskouei, 2016; Salas,
2015), but there is room for research into other languages and speakers.
Currently, published literature shows that research heavily privileges
written academic texts, particularly research articles, with a strong focus
on comparing metadiscourse across disciplines, genres, and languages. It
is also the case that a great deal of the most cited work actually focuses on
 Metadiscourse 79

a specific feature of metadiscourse, particularly stance, perhaps reflecting

the widespread interest in the individual. This is the age of “me” and the
author/speaker’s perspective is what matters: what he or she thinks or feels
about something, or decides to tell us, ranging from everything between
one’s favorite color to assessments about the status of knowledge. There
is a risk that the approach might remain too closely associated with the
description of a limited range of text types and fail to realize its potential
as a systematic means of gaining insights into interaction more generally.
Methodologically, studies are understandably dominated by discourse-
analytic procedures, particularly using corpora, although these are occasion-
ally supplemented with the views of text users. Further study might explore the
language choices writers make in their texts through discourse-based inter-
views and how these choices are understood by text users. Another important
area for further research is the pedagogical impact of metadiscourse, aiming
to develop more systematic and identifiable connections between scholarship
and classroom practice. Specifically, this involves investigating the ways in
which metadiscourse findings are being used in EAP and ESP teaching, the
effect this has on learning, and refining best practices for teachers.

References
Ädel, A. (2006). Metadiscourse in L1 and L2 English. John Benjamins.
Ädel, A., & Mauranen, A. (2010). Metadiscourse: Diverse and divided perspectives.
Nordic Journal of English Studies, 9(2), 1–11.
Anthony, L. 2019. AntConc (Version 3.5.8). Tokyo: Waseda University. http://
www​.antlab​.sci​.waseda​.ac​.jp/
Anthony, L. (2022). AntConc (Version 4.1.0). Waseda University. http://www​
.antlab​.sci​.waseda​.ac​.jp/
Bax, S. (2012). Text inspector. Online text analysis tool. https://textinspector​.com/
Bax, S., Nakatsuhara, F., & Waller, D. (2019). Researching L2 writers’ use of
metadiscourse markers at intermediate and advanced levels. System, 83, 79–95.
Beauvais, P. J. (1989). A speech-act theory of metadiscourse. Written
Communication, 6(1), 11–30.
Chen, L., & Hu, G. (2020). Surprise markers in applied linguistics research articles:
A diachronic perspective. Lingua, 248, 102992.
Connor, U., & Lauer, J. M. (1988). Cross-cultural variation in persuasive student
writing. In A. C. Purves (Ed.), Contrastive rhetoric (pp. 138–159). Sage.
Crismore, A. (1983). Metadiscourse: What is it and how is it used in school and non-
school social science texts. University of Illinois.
Crismore, A. (1989). Talking with readers: Metadiscourse as rhetorical act. Peter
Lang.
Crismore, A., Markkanen, R., & Steffensen, M. S. (1993). Metadiscourse in
persuasive writing. Written Communication, 10(1), 39–71.
 80 Ken Hyland

Crismore, A., & Abdollehzadeh, E. (2010). A review of recent metadiscourse

studies: The Iranian context. Nordic Journal of English Studies, 9(2), 195–219.
Dafouz, E. (2008). The pragmatic role of textual and interpersonal metadiscourse
markers in the construction and attainment of persuasion: A cross-linguistic
study of newspaper discourse. Journal of Pragmatics, 40(1), 95–113.
Fogarty-Bourget, C. G., Artemeva, N., & Fox, J. (2019). Gestural silence. In C. S.
Guinda (Ed.), Engagement in professional genres (pp. 277–295). John Benjamins.
Gong, H., Liu, L., & Cao, F. (2021). A cross-linguistic study of interactional
metadiscourse in English and Chinese research articles by the same Chinese
scholars. Journal of Language, Identity, and Education. https://doi​.org​/10​
.1080​/15348458​.2021​.1932504
Gonzalez, R. A. (2005). Metadiscourse in commercial websites. Iberica, 9, 33–52.
Halliday, M. A. K. (1985). Spoken and written language. Oxford University Press.
Halliday, M. A. K. (1994). An introduction to functional grammar. Edward Arnold.
Harris, Z. S. (1959). Computable syntactic analysis: Transformations and discourse
analysis papers. University of Pennsylvania.
Hyland, K. (1998). Hedging in scientific research articles. John Benjamins.
Hyland, K. (2004). Disciplinary interactions: Metadiscourse in L2 postgraduate
writing. Journal of Second Language Writing, 13(2), 133–151.
Hyland, K. (2005a). Metadiscourse. Continuum.
Hyland, K. (2005b). Stance and engagement: A model of interaction in academic
discourse. Discourse Studies, 7(2), 173–191.
Hyland, K. (2017). Metadiscourse: What is it and where is it going? Journal of
Pragmatics, 113, 16–29.
Hyland, K., & Jiang, F. (2018). “In this paper we suggest”: Changing patterns of
disciplinary metadiscourse. English for Specific Purposes, 51, 18–30.
Hyland, K., & Jiang, K. (2020). Text-organizing metadiscourse: Tracking changes
in rhetorical persuasion. Journal of Historical Pragmatics, 21(1), 138–165.
Hyland, K., & Jiang, K. (2022). Metadiscourse: The evolution of an approach to
texts. Text and Talk. https://doi​.org​/10​.1515​/text​-2021​- 0156
Hyland, K., & Tse, P. (2004). Metadiscourse in academic writing: A reappraisal.
Applied Linguistics, 25(2), 156–177.
Hyland, K., & Zou, H. (2022). Pithy persuasion: Engagement in 3-minute theses.
Applied Linguistics, 43(1), 21–44.
Ifantidou, E. (2005). The semantics and pragmatics of metadiscourse. Journal of
Pragmatics, 37(9), 1325–1353.
Jakobson, R. (1980). The framework of language. Michigan Studies in the
Humanities.
Jiang, K., & Hyland, K. (2016). Nouns and academic interactions: A neglected
feature of metadiscourse. Applied Linguistics. https://doi​.org​/10​.1093​/applin​
/amw023
Johns, A. (1997). Text, role and context: Developing academic literacies. Cambridge
University Press.
Kawase, T. (2015). Metadiscourse in the introductions of PhD theses and research
articles. Journal of English for Academic Purposes, 20, 114–124.
Khabbazi-Oskouei, L. (2016). Orality in Persian argumentative discourse: A case
study of editorials. Iranian Studies, 49(4), 677–691.
 Metadiscourse 81

Mauranen, A. (1993). Cultural differences in academic rhetoric: A textlinguistic

study. Peter Lang.
McGrath, L., & Kuteeva, M. (2012). Stance and engagement in pure mathematics
research articles: Linking discourse features to disciplinary practices. English for
Specific Purposes, 31(3), 161–173.
Peterlin,AP 2005 Text-organising metatext in research articles: an English–Slovene
contrastive analysis, English for Specific Purposes, 24, 3, Pages 307-319,
Salas, M. D. (2015). Reflexive metadiscourse in research articles in Spanish:
Variation across three disciplines (linguistics, economics and medicine). Journal
of Pragmatics, 77, 20–40.
Swales, J. (1990). Genre analysis: English in academic and research settings.
Cambridge University Press.
Thompson, G. (2001). Interaction in academic writing: Learning to argue with the
reader. Applied Linguistics, 22(1), 58–78.
Tognini-Bonelli, E. (2001). Corpus linguistics at work. John Benjamins.
Vande Kopple, W. J. (1985). Some exploratory discourse on metadiscourse. College
Composition and Communication, 26, 82–93.
Wang, W. B. (2019). The temporality of English and the spatiality of Chinese: A
contrastive analysis. Beijing: Foreign Language Teaching and Research Press.
Williams, J. M. (1981). Style: Ten lessons in clarity and grace. Scott Foresman.
Zou, H., & Hyland, K. (2019). Reworking research: Interactions in academic
articles and blogs. Discourse Studies, 21(6), 713–733.
5
RHETORICAL MOVE-STEP ANALYSIS
J. Elliott Casal, Matt Kessler

5.1 Introduction to the approach and definition of genre

Rhetorical move-step analysis (often referred to as move analysis, genre
analysis, or ESP genre analysis) is a text analytical approach that is used to
investigate the organizational and rhetorical structures of community-spe-
cific genre practices (Moreno & Swales, 2018; Swales, 1990, 2004). Using
a dance metaphor of broader moves which are performed through more
micro-level steps, a rhetorical move analysis frames the decision-making
of writers around the rhetorical and linguistic expectations of community
members as they engage in genre practices which are understood as “struc-
tured communicative events” (Swales, 1990, p. 45). In general, moves have
been defined as “discoursal or rhetorical units performing coherent com-
municative functions in texts” (Swales, 2004, pp. 228–229). Importantly, a
rhetorical move is genre-dependent, defined as a “functional, not a formal,
unit” (Swales, 2004, p. 229), and is characterized both by carrying out a
local purpose and contributing to the recurring communicative purpose
at the center of the genre practice itself. As a functional unit, a move can
be realized in a paragraph, a sentence, or even a phrase. Steps are more
narrowly defined functional fragments of a text which contribute to the
realization of rhetorical moves. In this way, steps may be regarded as the
building blocks that constitute a move.
Rhetorical move-step analysis emerged from Swales (1981), a pedagogi-
cally driven investigation of the recurring rhetorical structures employed by
authors in the introduction sections of published academic research articles.
In this groundbreaking study, Swales originally identified four moves that
authors routinely performed in their introduction sections. Later in Swales
DOI: 10.4324/9781003300847-7
 Rhetorical move-step analysis 83

(1990), the Create-a-Research-Space (CARS) model was introduced, which

included only the following three moves: establishing a territory, establish-
ing a niche, and occupying a niche. Although Swales is credited with estab-
lishing rhetorical move-step analysis, other researchers are also integral in
its development, including Bhatia (1993), who targeted professional genres
such as sales promotion letters. Indeed, although academic English genres
have been the object of the bulk of move analyses, the method can be
applied to any written (or spoken) genre.
To further illustrate these concepts with regards to academic writing, let
us take the first move of Swales’ CARS model, establishing a territory, as an
example. With this move, writers broadly attempt to establish their topic for
the reader by showing that it is worthy of investigation. Within an academic
research article, there are multiple ways that writers can accomplish this. For
instance, writers may realize this move by employing a step called claiming
importance (e.g., providing evidence the topic is critical to address). However,
writers may also use other steps, such as making topic generalizations (e.g.,
describing the current state of knowledge on the topic) or reviewing items
of previous research (e.g., synthesizing previous studies to help situate how
the current topic fills a gap in the existing literature). Thus, while writers of
academic research articles have been shown to be likely to perform the estab-
lishing a territory move in their introduction sections, they go about it in dif-
ferent ways and by using one, two, or all three of the aforementioned steps.
Rhetorical move-step analysis is strongly associated with the English for
Specific Purposes (ESP) school of genre theory, and as such, genre is con-
ceptualized according to definitions put forward by Swales. Swales (1990)
defines genre as “a class of communicative events, the members of which
share some set of communicative purposes. These purposes are recognized
by the expert members of the parent discourse community, and thereby
constitute the rationale for the genre” (p. 58). This approach to genre also
adopts a strongly linguistic approach to defining genre events, which is
reflected in the emphasis that move analysts place not only on identify-
ing recurring rhetorical patterns in genre practices, but also in examin-
ing the linguistic realization of writers’ (or speakers’) rhetorical goals (e.g.,
reporting verbs, lexical bundles). As described in the introduction to this
edited volume, Swales’ definition informed the ESP approach to genre-
based research and pedagogy (also see Hyon, 1996, 2017), and therefore, a
majority of move-step research to date has been published in ESP-oriented
academic journals such as English for Specific Purposes and the Journal of
English for Academic Purposes.
Of these published studies, most have targeted discipline-specific aca-
demic genre practices for broad genres that are common across academia
(e.g., published research articles, accepted conference abstracts, doctoral
84 J. Elliott Casal, Matt Kessler

dissertations), but attention has also been directed toward pedagogical

genres (e.g., class lab reports), along with non-academic genres (e.g., busi-
ness emails). Therefore, it is a methodology that is suitable for studying
both learner and expert texts. As will be discussed throughout this chap-
ter, points of comparison are often cross-disciplinary, cross-linguistic, or
expert-learner, and in recent years, move analysis has often been combined
with other analytical methods of analyzing genre.

5.2 Goals
The overarching goals of rhetorical move analysis are to identify and profile
the recurring rhetorical structures of community-specific genre practices
and the linguistic features which play prominent roles in the realization
of such rhetorical intentions. An overt focus on the pedagogical applica-
tions of such analyses has been integral to the analytical approach from the
beginning. The concept of discourse community is central to understanding
the analytical aims of a rhetorical move-step analysis. Swales (1990) defines
discourse communities as having a shared set of goals, means of commu-
nicating to exchange knowledge and feedback, genres which they use (and
thus own), specialized linguistic repertoires, and mechanisms of member-
ship. Scholarship has placed considerable focus on academic discourse com-
munities such as academic disciplines (e.g., applied linguistics), or journal
communities, but discourse communities exist in professional contexts such
as office spaces and social settings such as book clubs. Taken together, these
features of the discourse community drive the analytical aims of rhetorical
move-step studies, bringing a focus on the situated, discourse-community-
specific purposes and manners of interaction. As Hyland (2008) has noted,
many genre-based studies are conducted with the goal of understand-
ing the commonly employed rhetorical strategies and linguistic features
used by expert members of a discourse community. Tardy (2016) has also
explained that such research often represents attempts to deconstruct and
demystify the discourse practices of academic or professional communities.
By doing so, researchers can inform pedagogical approaches which aim to
provide novices with an entry point for engaging in new or unfamiliar gen-
res practices.
As such, many rhetorical move-step studies have targeted academic and
pedagogical genres that both first and second language (L1/L2) students
are likely to encounter within undergraduate university settings (e.g.,
persuasive essays, academic research articles, dissertations) or the genre
practices of academic discipline communities (with a strong emphasis on
writing). Another related body of scholarship has focused on exploring
the rhetorical practices of discourse communities beyond the classroom
 Rhetorical move-step analysis 85

in professional settings (e.g., business email requests, sales letters). Within

both academic and professional settings, considerable interest lies in investi-
gating the discourse practices adopted in occluded genres (see Swales, 1996),
which are genres that may be unpublished or private in nature (e.g., grant
applications, personal statements written for graduate school).
Irrespective of the genre and its occluded status, the primary goals of
many rhetorical move-step studies have generally remained the same. These
goals pertain to understanding the conventionalized rhetorical stages of
a given genre, identifying the linguistic means through which rhetorical
aims are realized, understanding the communities who engage in a given
genre practice (and why), and understanding how genres may vary depend-
ing on certain factors such as the community (often discipline, sometimes
language), the author (e.g., expert versus novice, cultural-specific practices),
or other relevant variables. Thus, when adopting a rhetorical move-step
analysis, researchers typically investigate questions such as:

• What rhetorical moves-steps do writers employ in the target genre?

• What linguistic features distinguish or contribute to the realization of
the moves-steps (e.g., lexical items, complexity measures, formulaic lexi-
cal bundles, or phrase-frames)?
• What are expert community members’ perceptions of the rhetorical and
linguistic construction of texts? (e.g., Why do they adopt these moves-
steps? What are their perceptions of novice writers’ practices)?
• What challenges do novice writers face when attempting to engage with
the genre?
• How do novice writers’ use of moves-steps differ from expert writers?
• How do rhetorical moves-steps and their realizations vary across dis-
course community contexts of related genres (e.g., mathematics versus
applied linguistics research articles; cross-linguistic comparisons)?
• What variations exist in how community members rhetorically construct
their texts, and what factors may influence these decisions (e.g., based
on their L1)?

5.3 Common research methods

Rhetorical move analysis is an approach to text analysis that examines the
rhetorical conventions of community-specific genre practices through dis-
course-based analysis of the semantic, structural, and linguistic choices that
writers (or speakers) make in signaling their rhetorical intentions. Analysis
entails the segmentation of texts into rhetorical moves and steps through
close (qualitative) analysis, with rhetorical moves representing broad, rec-
ognizable, recurring communicative goals or stages within a text that are
86 J. Elliott Casal, Matt Kessler

variable in length. Steps represent more specialized and narrow parts of

a move’s realization. Due to the iterative, close analysis that is required,
datasets for rhetorical move analysis studies are often smaller collections
of specialized genres, with discipline, discourse community, or language
boundaries serving as points of comparison.
We frame rhetorical move analysis procedures as various processes that
occur within three broad stages as researchers develop, apply, and assess
their rhetorical move-step frameworks. However, it is important to note
that researchers are likely to move back and forth between stages through-
out the analysis, as reliability checks, conversations with experts, or other
procedures may result in refining one’s framework. In this way, we frame
move analysis as both the iterative nature of developing and applying move-
step frameworks, and the high variability in researchers’ use and report-
ing of the move analysis approach since not all move analysts include all
the steps in the same order. This broad scheme aligns with Cotos et al.’s
(2017) categorizations of development, testing, and validation, although it
is intended to be a more fluid account of practices, rather than a report on
a particular application.
The first stage targets the development of a cohesive rhetorical move-
step model that concisely accounts for the patterns of rhetorical activity
present in the dataset and flexibly handles variability. An important starting
point is developing an understanding of the genre’s overarching purpose
within the community context to which it belongs and familiarizing oneself
with the texts linguistically and rhetorically. This is accomplished through
careful reading of all or many texts in the dataset, as well as consideration of
other contextual factors and related genre practices. To develop the rhetori-
cal move-step model, researchers identify recurring rhetorical actions that
may coalesce into moves and steps through close examination of content,
structural, and linguistic cues. Importantly, some scholars argue that the
identification of rhetorical move-step categories should rely primarily (or
exclusively) on semantic matter (e.g., Paltridge, 1994), while others recog-
nize and allow for the strong signals of rhetorical intent that can exist in lin-
guistic markers. Special attention is paid to rhetorical and topical shifts that
occur in the text (Swales, 1990). Particularly when examining well-studied
genres, many researchers begin with existing rhetorical move frameworks
as starting points and adapt them to account for the rhetorical activity of
the community practice in focus. Descriptions of the moves and steps are
prepared as well.
The second stage involves applying the move-step framework devel-
oped through the first stage to code the texts into rhetorical segments and
refine the framework and coding procedures. Multiple analysts generally
participate in annotation, with researchers often coding independently;
 Rhetorical move-step analysis 87

in some cases, training is offered (e.g., Parkinson, 2017). In segment-

ing the texts into rhetorical units, Moreno and Swales (2018) argue that
rhetorical steps should be given preeminence over moves in order to cap-
ture more locally relevant rhetorical activity, but many scholars segment
the texts into rhetorical moves first, which are then deconstructed into
rhetorical steps. To this point, Swales (2004) notes that “although [the
rhetorical move] has sometimes been aligned with a grammatical unit
such as a sentence, utterance, or paragraph…it is better seen as flexible
in terms of its linguistic realization” (pp. 228–229). Crucially, when it
comes to this stage, there is no consensus among researchers regarding
the correct way to identify/code moves-steps in a text. When identifying
moves, some researchers have based their decisions solely on linguistic
cues (e.g., present tense, certain nouns). This has been referred to as a
bottom-up approach; meanwhile, other researchers have adopted a top-
down approach in which they identify moves on the basis of content (see
Pho, 2008). Oftentimes, many researchers will adopt a combination of
bottom-up and top-down approaches when identifying/coding rhetorical
moves.
The aim of the processes we have grouped into the third stage is to assess
and refine the move-step framework and to ensure and enhance the reliability
and validity of the analysis. Rhetorical move analysis is an iterative process,
and some scholars explicitly involve a pilot coding procedure (e.g., Cotos et
al., 2017; Kwan, 2006), allowing for systematic testing of the framework,
coding procedures, and coders, as well as adjusting of the framework. Due
to the subjective (and time-consuming) nature of move analysis, multiple
researchers and/or coders are often involved in coding, and thus inter-coder
agreement and reliability are important considerations for increasing the
transparency of the procedures. Kanoksilpatham (2005) represents an early
study to include inter-coder reliability, which was conducted to illustrate that
“different individuals can demarcate the boundary of units at a sufficiently
high level of agreement” (p. 272). While this study attempted to identify
move boundaries in texts, other studies have sought inter-coder agreement
to identify the number of words devoted to realizing a move (e.g., Parkinson,
2017), the sequence of certain moves (e.g., Yoon & Casal, 2020a), or even
more generally as to whether a text includes a move-step at all (e.g., Kessler,
2020). Many authors assess inter-coder agreement by calculating a simple
percentage agreement or by reporting Cohen’s Kappa coefficient. However,
we refer readers to Rau and Shih (2021), which provides a detailed discus-
sion of the topic. Assessing agreement is also an essential means of refining
the framework. Another important means of validation in rhetorical move
analysis is expert (or member) checking, where researchers involve discourse
community experts directly to review coding decisions.
88 J. Elliott Casal, Matt Kessler

Beyond consulting members for validation, many rhetorical move analy-

ses include interview-based data sources to better understand the decisions
of experts in the target genre (e.g., Hyland, 2012; Kuteeva & McGrath,
2015) or audience members and gatekeepers (e.g., Kessler, 2020). Corpus
linguistics has also interfaced with move analysis as a means of investigating
the linguistic features that writers use in the realization of their rhetorical
goals (e.g., Biber et al., 2007; Lu et al., 2021). Particularly in recent years,
researchers have integrated corpus-based investigations into phraseologi-
cal features such as lexical bundles (e.g., Cortes, 2013; Li et al., 2020) and
phrase-frames (e.g., Casal & Kessler, 2020; Yoon & Casal, 2020b), lexical
and syntactic complexity measures (e.g., Lu et al., 2020; Tankó, 2017),
and multi-dimensional analysis (e.g., Gray et al., 2020) with large-scale
move analysis to more directly target the linguistic realization of rhetorical
moves.

5.4 Example studies
To better illustrate how rhetorical move-step analysis has been adopted and
applied, in this section we highlight several recent studies (see Table 5.1)
which have been selected to provide an overview of the breadth of move
analysis research over recent years. Beginning with Yoon and Casal
(2020a), in their study, the authors applied rhetorical move analysis to
one English written academic genre, conference abstracts accepted by the
American Association of Applied Linguistics. They framed the importance
of their study by emphasizing how conference presentations afford impor-
tant opportunities for junior scholars, the role that conference presenta-
tions play in the production of academic knowledge, and the semi-occluded
nature of the genre. They further noted that conference abstracts do not
generally receive considerable feedback like some academic genres (e.g.,
research articles), and their imposed brevity leads to propositional den-
sity, thus potentially creating difficulties for novice writers and L2 English
writers who aim to produce this promotional genre. The dataset included
625 accepted conference abstracts, and the analysis used a combination
of rhetorical move frameworks from various research article subsections
as a starting point (i.e., Introduction, Methods, Results, Discussion, also
referred to as part-genres).
The researchers independently analyzed the texts and collaboratively
modified the framework in an iterative fashion, until the existing frame-
works were consolidated into a “single, structurally coherent framework”
(p. 6) that concisely accounted for the rhetorical activity of conference
abstract writers. They annotated the texts with inter-coder reliability of 95%
at the move level and 88% at the step level. Findings revealed notable levels
TABLE 5.1 
Example studies adopting move-step analysis

Authors Context Research Questions (RQs) Data Collected Comments on Study

Design

Kanoksilpatham • Research • What is the textual • 180 research articles • Chi-square tests
(2015) articles in three organization of individual R A across civil, software, were used to
subdisciplines of sections in corpora from three and biomedical compare sections
engineering engineering subdisciplines? engineering across engineering
• What are the significant disciplines
statistical variations in textual
organization that distinguish
one engineering subdiscipline
from another?
Kessler (2020) • Grant statements • What are the rhetorical moves- • 50 Fulbright grantees • The grant statements
from applicants steps employed by successful contributed both a were taken from
to the Fulbright writers of the Fulbright ETA personal statement “successful” writers
English Teaching Personal Statement (PS) and and a statement of only, which was
Assistantship Statement of Grant Purpose grant purpose (100 operationalized as
program (SoGP)? total documents) “applicants who
• When reviewing Fulbright • Six experienced US received grants”
ETA PSs and SoGPs, what faculty/Fulbright (p. 184)
moves-steps do evaluators raters participated
react to and perceive as being in think-alouds and
effective and/or ineffective? semi-structured
• What are raters’ expectations interviews
for the rhetorical connections
and differences between PSs
and SoGPs?
Rhetorical move-step analysis 89

(Continued)
TABLE 5.1 Continued
90

Authors Context Research Questions (RQs) Data Collected Comments on Study

Design
Kuteeva and • Research • What is the rhetorical and • 22 mathematics • The authors
McGrath articles in pure organizational structure of research articles, collaborated
(2015) mathematics the pure mathematics research which were published with an expert
• The authors articles analyzed in the study? between 1998 and informant (i.e., a
were L1 English • Is there a consistent structure? 2011 mathematician) in
speakers • What connection can be traced • Five mathematicians the initial move
J. Elliott Casal, Matt Kessler

between the rhetorical and and study authors identification stage,

organizational structure of participated in who assisted in
the research articles and the interviews answering questions
knowledge making-practices of the authors had
the pure mathematics academic
community?
Martín and León • Research article • To what extent does cross- • 80 research article • This study is unique
Pérez (2014) introductions linguistic and cross-disciplinary introductions in in the emphasis on
in Spanish and variation exist in the writing of English and Spanish promotional rhetoric
English across English- and Spanish-speaking each in research writing
four disciplines scholars in the Introduction • Evenly distributed and the combination
sections of R As in two across four disciplines of cross-linguistic
disciplines? in social science and and cross-disciplinary
• To what extent do differences humanities comparisons
exist in writers’ use of different
steps to accomplish Move 3
(i.e., occupying the niche)?
Park et al. (2021) • Business email • What are the similarities and • 30 email requests • The authors used a
requests differences between Korean from L1 English discourse completion
and American employees’ speakers and 30 email task to elicit email
request emails in terms of (a) requests from L1 responses from their
move frequency, (b) move Korean speakers of participants
sequence, and (c) move length English
and associated lexical and
syntactic features?
Yoon and Casal • Conference • What are the rhetorical moves- • 625 conference • Authors followed up
(2020a) abstracts accepted steps used by authors of applied abstracts accepted by the move analysis by
to an applied linguistics conference abstracts? a major conference in examining recurring
linguistics • What are the dominant move applied linguistics phraseological
conference sequences that recur across the features (Yoon &
corpus? Casal, 2020b)
Rhetorical move-step analysis 91
92 J. Elliott Casal, Matt Kessler

of rhetorical consistency at the move level, in that four of the seven moves
in their framework were present in over 90% of conference abstracts, and six
of the seven moves were present in over 65% of the abstracts. Considerable
rhetorical variety was observed at the step level. The authors also discussed
the opening sequences, reporting that while over 75% of writers began with
the establishing a research territory move, approximately one-quarter of the
writers began by presenting the present work.
Many studies that employ rhetorical move analysis are limited in scope
by the time-consuming nature of manual coding, so Kanoksilpatham’s
(2015) analysis is therefore noteworthy for undertaking a move analysis of
180 research articles. In this study, the author analyzed the rhetorical move
structures of three subdisciplines of engineering (civil, software, and bio-
medical engineering), rather than comparing disciplinary practices to those
documented within applied linguistics as many studies do. Kanoksilpatham
targeted the major part-genres of research articles (i.e., Introduction,
Methods, Results, Discussion [IMRD]). The researcher did not discuss
the principles on which the coding was based in depth, but other meth-
odological decisions are discussed clearly. For example, Kanoksilpatham
outlines that inter-coder reliability was assessed by involving two experts
from each engineering field, training them, and conducting independent
annotation of four complete research articles. Agreement was above 80% in
all subdisciplines.
The analysis presents the percent of texts in each engineering field which
were analyzed and included each rhetorical move and step, and move-steps
were labeled as obligatory, conventional, or optional using percent-based
thresholds. Findings highlight both areas of expected consistency across
engineering fields and areas of community-specific variation. For exam-
ple, while between 70% and 87% of writers in all disciplines included the
establishing a niche move in their introductions, biomedical engineers were
four times more likely to present positive justification for filling the gap.
Similarly, software engineers were twice as likely to report outcomes and
claim the value of their research contributions in introductions. Overall,
this study provides a strong basis for many methodological aspects of rhe-
torical move analysis, and it reinforces the community-specific nature of
genre practices which emerge both socially as conventions and in relation
to the activities of the discourse community.
While many move-step analyses have targeted community-based differ-
ences in the rhetorical practices of academic research genres in English,
Martín and León Pérez (2014) further expand the scope of discourse com-
munity comparisons by examining both cross-linguistic and cross-discipli-
nary research article practices. The primary interest in how writers present
the value of their own study is through attention to “those steps that clearly
 Rhetorical move-step analysis 93

add promotional value” (p. 2), namely announcing principal outcomes and
stating the value of the present research (Swales, 2004). The data consisted
of 160 research article introductions in total, with 80 in English and 80
in Spanish. Each language sub-corpus is comprised of 20 articles each
from four disciplines (clinical and health psychology, dermatology, politi-
cal philosophy, and political science). The authors independently analyzed
the texts with an emphasis on “semantic or functional criteria rather than
on formal criteria” (p. 5), following Paltridge’s (1994) recommendation to
rely on the content rather than on linguistic cues. They also report high
inter-coder agreement, although the agreement was reported after differ-
ences were reconciled. In cases of uncertainty, disciplinary specialists were
involved in resolutions.
The authors reaffirm the importance of signaling the paper’s aims
through the widespread inclusion of the announcing present research step,
and they also uncover a number of differences between the health sci-
ence disciplines and humanities/social science disciplines. One example
is that the humanities/social sciences writers were more likely to outline
the structure of the paper for their readers, which the authors attribute
to the demands that the increased variation in textual organization may
place on readers. From a cross-linguistic perspective, the findings indicate
that English writers engage in markedly more promotional activity, from
a rhetorical steps perspective, than Spanish authors overall. However, the
authors also note an interplay between the language and discipline, under-
scoring that genre practices are informed by cultural identities in addition
to professional ones. Overall, this study contributes to our understand-
ing of the inextricable connections between genre practices and the com-
munity contexts in which they emerge, the complex interconnectivity of
discourse communities, and the challenges of conducting cross-linguistic
move analyses.
Similarly, Park et al. (2021) conducted a cross-cultural study involving
the non-academic genre of business email requests. The researchers were
specifically interested in international business exchanges in which English
often serves as a lingua franca. Therefore, they compared how L1 English
speakers and L1 Korean speakers of English approached composing requests.
Interestingly, while most move-step studies tend to collect texts that have
been produced from discourse community members’ everyday practices,
Park et al. took a different approach. The authors noted that because prag-
matics are contextually bound, they needed to control for the situation.
They did so by creating a discourse completion task (DCT; rarely associated
with move analysis but common in research on spoken pragmatics), which
provided writers with a scenario, requiring them to ask a colleague to deliver
a marketing presentation in their place. The authors specified information
94 J. Elliott Casal, Matt Kessler

such as why the request was being made (i.e., a scheduling conflict), the
relationship between the writer and email recipient, and more. Then, the
DCT was sent to 60 participants (30 American and 30 Korean business
professionals), who wrote email requests in response to the prompt.
When coding for moves, the authors relied on both a bottom-up and
top-down approach. Multiple coders were involved in the identification
stage, and any inconsistencies were discussed; however, no inter-coder reli-
ability was provided. In their findings, the authors reported 1) the fre-
quency of moves used by the writers, 2) the common sequences of those
moves, and 3) the number of words constituting each move. In terms of
frequency, interestingly, despite the writer and email recipient knowing
each another in the context of the prompt, the L1 Korean writers tended
to provide a self-introduction (73%) much more frequently than their L1
English counterparts (27%). The L1 English speakers tended to begin their
requests by engaging in small talk, whereas the L1 Korean writers preferred
to start with the self-introduction move. Finally, although the frequency of
move use was relatively similar between the groups, the L1 English speak-
ers tended to produce more words per move on average (20.7 words versus
13.2 words). Ultimately, Park et al. noted that understanding culturally
specific practices is important, especially when successful business transac-
tions may depend on them.
As discussed, many studies have attempted to understand the construc-
tion of rhetoric by integrating the perspectives of the texts’ writers or read-
ers. For example, Kuteeva and McGrath (2015) examined the rhetorical and
organizational structure of research articles within pure mathematics (e.g.,
algebraic geometry). In terms of motivation, the authors noted that most
research investigating research articles has tended to focus on scholarly writ-
ing conducted within the social sciences. However, little scholarship has
explored theoretical articles where experimentation is not the norm, and thus,
an IMRD organizational paradigm might not apply. To address this gap,
Kuteeva and McGrath collected 22 peer-reviewed research articles within
theoretical mathematics. They provided information about the focal articles
such as the impact factors of the journals in which they were published. They
also indicated the research articles shared a similar purpose, which was “to
establish a mathematical theorem evidenced by proof resulting from logical
mathematical reasoning” (p. 218). When conducting the move-step analysis,
the authors first investigated two articles by collaborating with an expert
informant, who helped answer questions about the texts. After developing
their move-step framework, Kuteeva and McGrath independently coded the
remaining 20 articles. No inter-coder reliability was reported.
In their findings, the authors reported discovering three compulsory
moves in the overall organizational structure (i.e., openings, proofs, and
 Rhetorical move-step analysis 95

endings). There were also a number of optional moves (e.g., a post-introduc-

tion providing notational conventions), along with considerable variation
among the articles’ steps. To better understand authors’ rhetorical choices,
Kuteeva and McGrath conducted interviews with five mathematicians, who
were the lead authors of some of the focal texts. When discussing their
writing, the interviewees provided insights into the idiosyncratic and com-
plex nature of composing a theoretical mathematics research article. For
instance, one informant noted how the mathematical argument itself often
influenced different moves being made as opposed to creating “a dense
piece of text describing what I did in the lab and what I observed” (p.
232). Thus, the integration of the writers’ voices provided rich insights
into discipline-specific practices, along with how those practices differed
substantially from research articles in other fields.
While Kuteeva and McGrath’s study included the perspectives of the
target genre’s authors, Kessler (2020) integrated readers’ perceptions. In
the study, Kessler investigated the rhetorical strategies of undergraduate
writers who applied for a Fulbright English Teaching Assistantship (ETA)
grant, which enables recipients to serve as English teachers in different
countries. As part of the application, prospective grantees were required
to write a personal statement and a statement of grant purpose. Therefore,
the study focused on investigating the moves-steps used in the statements
of successful applicants (i.e., those who received an ETA grant), along with
understanding what Fulbright reviewers felt constituted effective rhetoric.
To investigate this, Kessler first collected 50 of each statement type (100
texts in total) from individuals who were awarded grants between 2012 and
2016. The move-step analysis involved the author reading through 20 of
each statement and developing an initial coding scheme, which began with
step identification. After the scheme was developed, a second coder was
recruited and trained before independently coding 10% of each statement
type. A percentage agreement for inter-coder reliability was reported (0.88).
In the findings, Kessler reported a striking amount of shared rhetoric
across the two statements, with three moves that were shared and only one
that was specific to the personal statement (called framing childhood and
family history). Across the statements, writers also employed a number of
shared steps, with only a handful being specific to each statement. Due to
the large amount of overlapping rhetoric, Kessler was interested in under-
standing readers’ (i.e., Fulbright raters’) perspectives of the texts. Six expe-
rienced Fulbright ETA readers were recruited, and these readers engaged in
think-alouds in which they read sample statements while commenting on
anything that came to mind while reading (e.g., strengths and weaknesses).
Once finished, the Fulbright raters answered follow-up questions in a semi-
structured interview about why they made certain comments. Interestingly,
96 J. Elliott Casal, Matt Kessler

despite writers adopting similar rhetoric across both statement types, the
experienced readers unanimously stated this was not a positive feature and
instead reflected the novice writers’ lack of experience with the genre. The
readers also made recommendations about what they felt constituted an
ideal balance of rhetoric between the statements. Thus, the inclusion of the
readers’ perspectives provided unique insights into the mismatch between
writers’ practices and readers’ expectations.

5.5 Issues and challenges

Rhetorical move analysts are likely to face many practical analytical issues
when conducting a move analysis. One persistent difficulty lies in deter-
mining the move-step boundaries, as rhetorical units will not necessarily
constitute linguistically complete constructions. Moreno and Swales (2018)
illustrate this issue with a number of examples and articulate their ration-
ale for determining a shift to a new rhetorical step, placing an emphasis
on “new propositional meaning,” which is functionally interpretable by
a “competent reader of the genre” and advances the text to “achieve the
purpose(s) of the (part-)genre” (p. 49). They provide clear examples of what
was and was not coded separately, and while space for these detailed discus-
sions is rare outside of such methodologically oriented texts, the tangible
benefits of clarity in principle and exemplification of ambiguous spaces are
informative. It is important for researchers to clearly articulate the princi-
ples on which segmentation will be based to other researchers, coders, and
readers who the work is shared with, and there are methodological benefits
to the difficulties in segmentation and labeling which arise. After all, rhe-
torical move analysis is a time-consuming manual procedure, and insights
into the decisions that are made by genre analysts after reading numerous
genre exemplars are insightful in their own right.
On the other hand, one of the most persistent issues facing rhetorical move
analysis is that of methodological transparency (and subsequently, repro-
ducibility). Move-step identification and demarcation of rhetorical bounda-
ries are inherently complex and subjective. In many studies, the authors
are carefully transparent in articulating their procedures and the principles
on which the move-step segmentation were conducted, but Moreno and
Swales (2018) rightly lament that “little information is typically provided
about the identification processes followed by researchers” (pp. 41–42).
Methodological issues such as the number of annotators, the procedures
for developing/modifying a framework, the unit of analysis (linguistic or
rhetorical), inter-coder reliability, and the procedures for resolving disagree-
ments are not uniformly reported. Thus, we encourage authors to provide
as many details as possible. In addition, transparency and reproducibility
 Rhetorical move-step analysis 97

can be enhanced by maximizing the number of examples of moves and steps

provided in the text (including those which may be atypical) and consider-
ing making move-step annotated portions of the dataset available.
Other practical analytical issues with methodological design also arise.
For example, in comparative analyses it is important to consider the nature
of the comparison. In expert-learner comparisons, in many cases it is rea-
sonable to consider the populations as engaging in approximately the same
genre practice. However, when conducting cross-disciplinary or cross-lin-
guistic comparisons, researchers must consider to what extent the texts rep-
resent distinct genre practices that necessitate divergent move frameworks.
Another important methodological consideration is how to report rhetori-
cal move results. Move analysts have long since reported inclusion rates of
moves and steps (i.e., what percent of texts included the rhetorical activ-
ity) to provide insights into what has been called the obligatory or optional
nature of a particular rhetorical action. As mentioned, some researchers also
provide insights into the amount of text dedicated to a given move-step (e.g.,
Parkinson, 2017) or common move sequences (e.g., Yoon & Casal, 2020a).
As a final issue, it is important to note that move analysis has had a
strong pedagogical orientation from the beginning (e.g., Swales, 1981)
and is closely associated with ESP, or more broadly Languages for Specific
Purposes. While the impacts of move analysis on L2 writing pedagogy are
clear to those who teach genre-based writing courses, there is space for
more direct pedagogical relevance in move analysis research. Cheng (2019)
critically examined how authors of 36 move analyses of research articles or
research article part-genres motivated their work as addressing pedagogi-
cal needs or leverage their work to propose pedagogically relevant implica-
tions. However, Cheng concluded that many do not directly respond to a
pedagogical need, reaffirm well-established findings within the field with-
out adding new insights, or do not offer pedagogical implications. Clearly
this represents a consideration for those who intend to adopt move analysis
in the future, namely the question: what is the pedagogical point of this
study? We recommend that novice researchers develop their move analysis
projects with careful consideration of the pedagogical needs their project
may address, and we also recommend that they articulate clear, actionable
pedagogical implications of their work.

5.6 Study-in-focus
In this section, we discuss an exemplary study that adopts move-step analy-
sis. In particular, we have selected Parkinson (2017), which examines the
rhetorical structure of scientific laboratory reports written by undergradu-
ate students. The laboratory (lab) report genre is a commonly used written
98 J. Elliott Casal, Matt Kessler

pedagogical genre for students in both the physical sciences and engineer-
ing fields during their undergraduate studies. Unlike academic research
articles which report on new research, the lab report’s function is more
pedagogical in nature, as the students’ instructor is the target audience
and the purpose of the report is to assess the extent to which students
understood and followed predefined methodological procedures in a lab-
based experiment. Importantly, although students are frequently asked to
produce the genre, studies have suggested that students typically have little-
to-no experience with lab reports prior to beginning their undergraduate
studies. Thus, its unfamiliarity, frequent use, and pedagogical utility make
it a genre worthy of investigation. Before proceeding further with our com-
mentary on this article, we encourage the reader to review our summary of
Parkinson’s study shown in the call-out box.

Citation
Parkinson, J. (2017). The student laboratory report genre: A genre analysis.
English for Specific Purposes, 45, 1–13.

Research questions
• To what extent do laboratory reports’ macro-structure compare to
academic research article’s (research articles) macro-structure (i.e., the
IMRD sequence)?
• What rhetorical moves are made in the student laboratory report genre?

Context and population

This study examined the lab reports of 60 students in the United Kingdom
in engineering and the physical sciences (e.g., chemistry, food science,
biology). The lab reports were extracted from the British Academic Written
English (BAWE) corpus. All 60 writers were L1 English speakers. Additionally,
the selected reports from the BAWE corpus all received high grades (“A” or
“B” equivalent). The decision to include highly rated reports was made to
increase the chance the assignments would “conform to disciplinary con-
ventions” (p. 4).

Procedure
To analyze the moves-steps, Parkinson followed a seven-stage process that
involved: 1) reading the texts to understand their overarching rhetorical
 Rhetorical move-step analysis 99

purpose; 2) analyzing text segments and assigning codes (i.e., steps); 3)

grouping the steps into functional themes (i.e., moves); 4) piloting the cod-
ing scheme; 5) developing a protocol for an inter-coder reliability check; 6)
obtaining inter-coder reliability; and 7) making any necessary revisions to
the coding scheme.
For inter-coder reliability, Parkinson recruited one coder, a PhD candi-
date in linguistics, who was trained. This coder applied the codes to four
texts (6.7% of the dataset) in a trial run, and differences in coding were dis-
cussed. Then, the second coder independently coded 12 more texts (20% of
the dataset). Reliability was reported using Cohen’s Kappa (0.87), indicat-
ing high agreement. Lastly, Parkinson consulted with an expert disciplinary
informant, who was a university physics tutor. The tutor reviewed and veri-
fied the moves-steps.

Findings
Parkinson’s analysis revealed all 60 texts shared the same macro-structure
with published research articles, following an IMRD sequence. However,
there was variation with other sections among disciplines. For instance,
biology and engineering reports frequently contained an abstract, but only
engineering reports contained a conclusion section. Among the six iden-
tified macro-sections, there were 22 rhetorical moves. No move appeared
across all 60 texts. Therefore, Parkinson treated moves that appeared 80%+
of the reports as obligatory. This included five moves, which were establish-
ing topic (in the intro), introducing experiment (intro), describing experimen-
tal procedures (method), announcing results (results), and interpreting results
(discussion). Three other moves were frequently used, appearing in 50–79%
of the texts.
For each macro-section, Parkinson listed the moves and steps. For exam-
ple, in the lab report’s intro section, she listed three typical moves (e.g.,
establishing topic). The establishing topic move consisted of three possible
steps (i.e., claiming importance, referring to known information, and referring
to literature). Parkinson also provided excerpts of each step. Finally, she pro-
vided quantitative data about the number of words devoted to each move,
each move’s mean word length, and how many reports had at least one
occurrence. These data were further segmented by discipline. For instance,
the move contextualizing discussion appeared to be obligatory in the discus-
sion section of food science lab reports (93% of texts) but was less frequent
in chemistry reports (47% of texts).
100 J. Elliott Casal, Matt Kessler

As highlighted, apart from Parkinson’s solid rationales for investigating

the genre, another major strength of her study is the thoroughness with
which she describes how she conducted the move-step analysis. Specifically,
Parkinson clearly describes the procedure that she followed when develop-
ing and verifying her coding scheme, including a pilot phase and inter-
coder reliability checks. Importantly, Parkinson included inter-coder
reliability information and reports high reliability with her second coder
(0.87). She also engaged in extra effort to enhance the validity of her analy-
sis by seeking the insights of an expert disciplinary informant (i.e., a phys-
ics tutor). Unfortunately, many move-step studies often do not report any
measure of reliability. Conversely, in Parkinson’s study, the combination of
her efforts helps establish a considerable level of confidence that her results
are both trustworthy and reliable. The clarity with which the procedures
were explained also contributes to transparency.
As mentioned, beyond conducting a move-step analysis of a specific
genre, researchers today are increasingly providing analyses that attempt
to understand additional aspects of the moves-steps being utilized. This
may involve investigating different linguistic elements that contribute to
the realization of rhetorical goals, integrating readers’ or writers’ perspec-
tives, or engaging in cross-disciplinary comparisons. Parkinson’s study falls
into this final category, in that she makes multiple disciplinary comparisons
and compares the lab report to another academic genre. Her first com-
parison involves analyzing the extent to which scientific lab reports follow
an IMRD macro-structure, which is common among published research
articles in applied linguistics and other fields. The second comparison she
makes is more nuanced, which involves comparing and contrasting the use
of moves within different scientific disciplines. In this analysis, she illus-
trates the variation that may occur among different sciences themselves
(e.g., between food sciences and chemistry). These comparisons provide
important pedagogical implications for how writing instructors may sub-
sequently approach the teaching of the lab report in different contexts.
Importantly, it also provides insights into how lab reports are assessed
across disciplinary contexts, as only reports which were highly rated were
included in the analysis. As such, Parkinson’s study is an exemplary model
showing how one might conduct a move-step analysis.

5.7 Future research directions

In terms of future research directions, the first major opportunity for
future move analysts is to develop more systematic and identifiable
connections between analytical scholarship and pedagogical practice.
Specifically, an important area for further study involves investigating the
 Rhetorical move-step analysis 101

ways in which move analysis findings make their way into Language for
Specific Purposes classrooms and the impact they have on L2 instruc-
tion. Increasing attention to how rhetorical move analysis may directly
impact L2 instruction is demonstrated by two recent approaches to imple-
menting move analysis into genre-based writing pedagogies, Negretti and
McGrath (2020) and Casal (2020). Both of these studies target gradu-
ate student writers in English and assess the impacts of a pedagogy that
includes rhetorical move analysis on learners’ principled, agentive, and
deliberate decision-making. However, more pedagogically oriented stud-
ies and approaches (e.g., action research) are needed that investigate the
actual implementation of move-step pedagogy in the L2 (and/or L1)
classroom. This can include assessment of changes in learner production,
perceptions of students or instructors who use move analysis activities or
materials, or changes in language users’ metacognition, decision-making,
and confidence.
Apart from developing more identifiable connections between research
and practice, another research direction involves expanding the target of
move-step studies. In particular, readers of this chapter may notice that
most rhetorical move-step analysis studies to date have explored genres
involving academic English. Very few researchers have examined academic
discourse beyond the conventions of English language research communi-
ties (e.g., Martín & León Pérez, 2014). Likewise, apart from non-English
genres, studies are needed that explore non-academic genres in English or
other languages. There is a small body of literature that uses move-step
analysis to examine professional genres such as fundraising letters (e.g.,
Connor & Gladkov, 2004) and request emails (e.g., Park et al., 2021).
However, future researchers may wish to explore commonly used profes-
sional genres in addition to multimodal genres (e.g., slideshow presenta-
tions, advertisements), which adopt different meaning-making resources to
convey intended messages.

References
Bhatia, V. (1993). Analyzing genre: Language use in professional settings.
Longman.
Biber, D., Connor, U., & Upton, T. A. (2007). Discourse on the move: Using corpus
analysis to describe discourse structure. John Benjamins.
Casal, J. E. (2020). An integrated corpus and genre analysis approach to writing
research and pedagogy: Development of graduate student genre knowledge
[Unpublished doctoral dissertation]. The Pennsylvania State University.
Casal, J. E., & Kessler, M. (2020). Form and rhetorical function of phrase-frames
in promotional writing: A corpus- and genre-based analysis. System, 95, 102370.
https://doi​.org​/10​.1016​/j​.system​.2020​.102370
 102 J. Elliott Casal, Matt Kessler

Cheng, A. (2019). Examining the “applied aspirations” in the ESP genre analysis of
published journal articles. Journal of English for Academic Purposes, 38, 36–47.
https://doi​.org​/10​.1016​.j​.jeap​/2018​.12​.005
Connor, U., & Gladkov, K. (2004). Rhetorical appeals in fundraising direct mail
letters. In U. Connot & T. Upton (Eds.), Discourse in the professions: Perspectives
from corpus linguistics (pp. 257–286). John Benjamins.
Cortes, V. (2013). The purpose of this study is to: Connecting lexical bundles
and moves in research article introductions. Journal of English for Academic
Purposes, 12(1), 33–43. https://doi​.org​/10​.1016​/j​.jeap​.2012​.11​.002
Cotos, E., Huffman, S., & Link, S. (2017). A move/step model for methods
sections: Demonstrating rigor and credibility. English for Specific Purposes, 46,
90–106. https://doi​.org​/10​.1016​.j​.esp​.2017​.01​/001
Gray, B., Cotos, E., & Smith, J. (2020). Combining rhetorical move analysis with
multi-dimensional analysis: Research writing across disciplines. In U. Römer,
V. Cortes, & E. Friginal (Eds.), Advances in corpus-based research on academic
writing: Effects of discipline, register, and writer expertise (pp. 138–168). John
Benjamins.
Hyland, K. (2008). Genre and academic writing in the disciplines. Language
Teaching, 41(4), 543–562. https://doi​.org​/10​.1017​/S0261444808005235
Hyland, K. (2012). Disciplinary identities: Individuality and community in
academic discourse. Cambridge University Press.
Hyon, S. (1996). Genre in three traditions: Implications for ESL. TESOL Quarterly,
30(4), 693–722. https://doi​.org​/10​.2307​/3587930
Hyon, S. (2017). Introducing genre and English for specific purposes. Routledge.
Kanoksilapatham, B. (2005). Rhetorical structure of biochemistry research articles.
English for Specific Purposes, 24(3), 269–292. https://doi​.org​/10​.1016​/j​.esp​
.2004​.08​.003
Kanoksilapatham, B. (2015). Distinguishing textual features characterizing
structural variation in research articles across three engineering sub-discipline
corpora. English for Specific Purposes, 37, 74–86. https://doi​.org​/10​.1016​/j​.esp​
.2014​.06​.008
Kessler, M. (2020). A text analysis and gatekeepers’ perspectives of a promotional
genre: Understanding the rhetoric of Fulbright grant statements. English for
Specific Purposes, 60, 182–192. https://doi​.org​/10​.1016​/j​.esp​.2020​.07​.003
Kuteeva, M., & McGrath, L. (2015). The theoretical research article as a reflection
of disciplinary practices: The case of pure mathematics. Applied Linguistics,
36(2), 215–235. https://doi​.org​/10​.1093​/applin​/amt042
Kwan, B. (2006). The schematic structure of literature reviews in doctoral theses
of applied linguistics. English for Specific Purposes, 25(1), 30–55. https://doi​.org​
/10​.1016​/j​.esp​.2005​.06​.001
Li, L., Franken, M., & Wu, S. (2020). Bundle-driven move analysis: Sentence
initial lexical bundles in PhD abstracts. English for Specific Purposes, 60, 85–97.
https://doi​.org​/10​.1016​/j​.esp​.2020​.04​.006
Lu, X., Casal, J. E., & Liu, Y. (2020). The rhetorical functions of syntactically
complex sentences in social science research article introductions. Journal of
English for Academic Purposes, 44, 1–16. https://doi​.org​/10​.1016​/j​.jeap​.2019​
.100832
 Rhetorical move-step analysis 103

Lu, X., Casal, J. E., & Liu, Y. (2021). Towards the synergy of genre-and corpus-
based approaches to academic writing research and pedagogy. International
Journal of Computer – Assisted Language Learning and Teaching, 11(1), 59–71.
https://doi​.org​/10​.4018​/ IJCALLT​.2021010104
Martín, P., & León Pérez, I. K. (2014). Convincing peers of the value of one’s
research: A genre analysis of rhetorical promotion in academic texts. English for
Specific Purposes, 34, 1–13. https://doi​.org​/10​.1016​/j​.esp​.2013​.09​.002
Moreno, A. I., & Swales, J. M. (2018). Strengthening move analysis methodology
towards bridging the function-form gap. English for Specific Purposes, 50, 40–
63. https://doi​.org​/10​.1016​/j​.esp​.2017​.11​.006
Negretti, R., & McGrath, L. (2020). English for specific playfulness? How doctoral
students in science, technology, engineering and mathematics manipulate genre.
English for Specific Purposes, 60, 26–39. https://doi​.org​/10​/1016​/ k​.esp​.2020​
.04​.004
Paltridge, B. (1994). Genre analysis and identification of textual boundaries.
Applied Linguistics, 15(3), 288–299. https://doi​.org​/10​.1093​/applin​/15​.3​.288
Park, S., Jeon, J., & Shim, E. (2021). Exploring request emails in English for
business purposes: A move analysis. English for Specific Purposes, 63, 137–150.
https://doi​.org​/10​.1016​/j​.esp​.2021​.03​.006
Parkinson, J. (2017). The student laboratory report genre: A genre analysis.
English for Specific Purposes, 45, 1–13. https://doi​.org​/10​.1016​/j​.esp​.2021​.03​
.006
Pho, P. D. (2008). Research article abstracts in applied linguistics and educational
technology: A study of linguistic realizations of rhetorical structure and
authorial stance. Discourse Studies, 10(2), 231–250. https://doi​.org​/10​.1177​
/1461445607087010
Rau, G., & Shih, Y. (2021). Evaluation of Cohen’s kappa and other measures of
inter-rater agreement for genre analysis and other nominal data. Journal of
English for Academic Purposes, 53, 101026. https://doi​.org​/10​.1016​.j​.jeap​.2021​
.101026
Swales, J. (1981 [2011]). Aspects of article introductions. University of Michigan Press.
Swales, J. (1990). Genre analysis: English in academic and research settings.
Cambridge University Press.
Swales, J. M. (1996). Occluded genres in the academy: The case of the submission
letter. In E. Ventola & A. Mauranen (Eds.), Academic writing: Intercultural
and textual issues (pp. 45–58). John Benjamins.
Swales, J. M. (2004). Research genres: Exploration and applications. Cambridge
University Press.
Tankó, G. (2017). Literary research article abstracts: An analysis of rhetorical
moves and their linguistic realizations. Journal of English for Academic Purposes,
27, 42–55. https://doi​.org​/10​.1016​/j​.jeap​.2017​.04​.003
Tardy, C. M. (2016). Beyond convention: Genre innovation in academic writing.
University of Michigan Press.
Yoon, J., & Casal, J. E. (2020a). Rhetorical structure, sequence, and variation:
A step-driven move analysis of applied linguistics conference abstracts.
International Journal of Applied Linguistics, 1–17. https://doi​.org​/10​.1111​/
ijal​.12300
104 J. Elliott Casal, Matt Kessler

Yoon, J., & Casal, J. E. (2020b). P-frames and rhetorical moves in applied linguistics
conference abstracts. In U. Römer, V. Cortes, & E. Friginal (Eds.), Advances
in corpus-based research on academic writing: Effects of discipline, register, and
writer expertise (pp. 282–305). John Benjamins.
6
LEXICAL BUNDLES AND
PHRASE FRAMES
Viviana Cortes

6.1 Introduction to the approach and definition of genre

The study of expressions that frequently occur in particular genres and are
identified following a corpus-driven approach has taken center stage in the
spectrum of formulaic language analyses in the past 30 years. Constructs
such as lexical bundles (or bundles) and phrase frames (or p-frames) have
become the focus of studies conducted in a variety of genres. Many studies
of bundles and p-frames take a bottom-up approach, focusing first on these
linguistic features, instead of a top-down approach that examines commu-
nicative functions as a starting point. The study of bundles and p-frames
in discourse, however, often involves “various interrelated levels of analysis
at the same time: identification of communicative purpose(s), schematic
structure, grammatical features, lexical features, etc.” (Flowerdew, 2002, p.
95). These multidirectional analyses help explore the relationship between
formulaic expressions and communicative purposes represented by rhetori-
cal moves and steps (Swales, 1990, 2004). The analysis of moves and steps
is associated with the English for Specific Purposes (ESP) school of genre
theory (see Casal & Kessler, this volume), in which genre is described as
a group of communicative purposes recognized by the members of a lin-
guistic community. Moves and steps are “textual segments that make up a
genre’s organizational structure and help the genre achieve its purposes”
(Hyon, 2018, p. 27).
As moves and steps have been considered building blocks at the dis-
course level (Miller, 1993), and bundles and p-frames can be considered
building blocks at the sentence level (Biber & Conrad, 1999), analyzing

DOI: 10.4324/9781003300847-8
106 Viviana Cortes

the relationship between moves/steps and frequent formulaic expressions

like bundles and p-frames can open new venues to explore how writers
realize their rhetorical intents using specific organizations and linguistic
conventions.
Bundles and p-frames are respectively continuous and discontinuous
groups of words that frequently occur in a particular genre. They are sim-
ple expressions and structures whose only quality is that they recur fre-
quently. The approach used to identify n-grams (groups of n number of
words: bigrams, trigrams, 4-grams, etc.) and bundles started in the late
1970s and early 1980s with the development of faster computer process-
ing that could be used for linguistic analysis in large collections of texts or
language corpora. Studies by Allen (1975) and Altenberg and Egg-Olofson
(1990) were some of the first to introduce a corpus-driven approach to
the identification of continuous word combinations or n-grams. It was not
until the end of the 1990s, however, that Biber and his collaborators coined
the term lexical bundles (Biber et al., 1999) to label groups of three or more
words that frequently occur in a register. In conversation, some examples of
bundles are I don’t know, she said to me, do you want some more, and do you
know what I mean. In academic prose, frequent bundles are expressions like
in order to, on the other hand, the purpose of this paper is, and the rest of the
paper is organized as follows.
P-frames are discontinuous word combinations, “n-grams with an
internal variable slot” (Römer, 2010, p. 103), although some studies have
considered expressions with external slots. These slots can be filled with
different variants or fillers. For decades, general linguists and corpus lin-
guists used their intuition, perception, or general formulaic knowledge
to identify and analyze frames (e.g., Renouf & Sinclair, 1991). Lately,
the development of computer programs that can identify frequent frames
automatically has allowed the use of a corpus-driven approach to identify
p-frames and their fillers. Even though some researchers use their own
programs for this task (e.g., Gray & Biber, 2013), there is now freeware
that can automatically identify p-frames in a corpus. The identification of
p-frames can focus on structures with more than one variable slot (the * of
the *), but most studies in the literature center on p-frames with only one
open slot. In addition, some studies identify p-frames with internal and
external slots (e.g., Yoon & Casal, 2020), while others focus exclusively
on p-frames with internal slots (e.g., Ren, 2022; Gray & Biber, 2013).
Frequent p-frames of different sizes (number of words) and with different
types of open slots in written academic prose are structures like the * of, the
* to the, one of the most *, from the * of the, is the most common * of, while
in conversation, we can find structures such as I don’t * if, the * and the,
and I don’t * if you.
 Lexical bundles and phrase frames 107

6.2 Goals
Studies that analyze bundles and p-frames often intend to go beyond lexis
or grammar, as these two constructs are considered lexicogrammatical,
merging both perspectives. Most bundle and p-frame investigations try to
describe a particular genre, compare two or more text types, compare the
same type of texts across different contexts or different languages, or com-
pare expert-novice or first language (L1)-second language (L2) production
from a formulaic language perspective. The ultimate objective of the iden-
tification and classification of both bundles and p-frames is to establish the
formulaic profile of particular genres.
Due to the availability and easy accessibility of written corpora, the
bulk of bundle and p-frame studies can be found in the analysis of writ-
ing, particularly in research that examines written academic writing. There
have been some studies of bundles and p-frames in spoken registers, such
as everyday conversation or academic lectures, but due to the difficulties
often encountered in the collection of spoken corpora, these investigations
are generally much less frequent. The most frequently studied academic
texts from a bundle and p-frame use perspective have undoubtedly been
published research articles. Even before corpora and computers helped
automatize the identification of these expressions, research articles had
already been found to be highly formulaic and linguistically pervasive
(Flowerdew & Li, 2007), making them the perfect candidates for stud-
ies of recurrent expressions. Recent studies have turned their attention to
other academic genres such as conference abstracts, statements of purpose,
and medical case reports, as well as non-academic genres like governmen-
tal documents.
Genre analysis studies that take a corpus-driven approach to the analysis
of bundles and p-frames are often guided by a variety of research questions
including these examples:

• What are the most frequent lexical bundles/p-frames in a given genre?

(This question often specifies the size of the bundles or p-frames and the
position of the open slot/s in frames).
• How are the bundles/p-frames structurally classified (e.g., noun phrase
fragments, prepositional phrase fragments, verb phrase fragments)?
• What are the functions most frequently performed by the bundles?
• Are there any connections between the bundles identified and the com-
municative purposes of the authors (as represented by rhetorical moves/
steps)?
• What are the fillers of a particular p-frame? (What are the most frequent
fillers of a given frame? How do these fillers group semantically? Are
108 Viviana Cortes

there any infrequent fillers that can be grouped semantically with fre-
quent fillers)?
• What are the functions of particular p-frames and fillers in context?
• What is the distribution of p-frames and fillers across rhetorical moves
and steps?
• What is the perceived pedagogical potential of bundles/p-frames identi-
fied in a given genre?

6.3 Common research methods

The identification of bundles and p-frames uses a corpus-driven approach
instead of a corpus-based approach. A corpus-based approach relies on
established linguistic forms, studying whether and how these forms are
used in a given language collection, while a corpus-based approach relies
on corpus data that delineates the features and scope of a linguistic study
(Biber, 2009). In the case of bundles and p-frames, this process uses com-
puter programs especially designed for this type of analysis. Earlier bundle
studies, for example, used computer software designed by the researchers
conducting those studies (Biber et al., 1999; Cortes, 2004). Nowadays,
researchers use commercially available software like Collocate (Barlow,
2004), or freeware like Antconc and Antgram (Anthony, 2021; 2022) or
Kf Ngram (Fletcher, 2012) in order to identify and analyze bundles and
p-frames. In general, a program runs through a collection of texts looking
at every n-gram, that is, expressions with n number of words in a text: three
words for three-word bundles or p-frames; four words for four-word bun-
dles or p-frames, and so on. The program keeps track of recurring n-grams
and flags the number of texts in which the same expression occurs (called
the range). Bundle and p-frame identification should respect punctuation
to avoid crossing clause or sentence boundaries. Identified n-grams must
comply with conventionalized but strict thresholds for frequency and range
to be considered lexical bundles.
In the case of p-frames, the program first generates a list of n-grams
(of the pre-established number of words), which are later used to create
p-frames combining identical n-grams with the selected variable slot/s
(often represented by an asterisk *). The program user has the option of
selecting the number of slots, and whether these are inner or outer slots
(inside the expression like the * of; or at the beginning or at the end, as in
* this study or most of the *). The identification of these expressions is based
only on frequency of occurrence and not on whether or not the expressions
make semantic sense. This may result in bundles like and in the or which is
the, as opposed to expressions that form an intuitive chunk like in the next
section or previous research has shown.
 Lexical bundles and phrase frames 109

Conventionalized thresholds for frequency and range in lexical bundle

identification may vary with the size of the bundle. Biber et al. (1999)
set a cut-off point of ten times per million words (pmw) for frequency for
four-, five-, and six-word bundles, but later studies used more conservative
thresholds. Cortes (2004) used 20 times pmw and Biber et al. (2004) used
40 times pmw for four-word lexical bundles. As longer bundles become less
frequent, Cortes (2013) used different thresholds for the bundles she iden-
tified in her study: 20 times pmw for four-word bundles, ten times pmw for
five-word bundles, eight times pmw for six- and seven-word bundles, and
six times pmw for any bundle longer than seven words. Regarding range,
traditional studies use a threshold of five or more (5+) texts in order to
avoid any idiosyncratic use by only one or just a few authors. More recent
studies tend to use higher dispersion rates reflected in a percentage of the
texts in the corpus.
In the case of p-frames, the frequency and range thresholds established
by different studies in the literature show great variability. Gray and Biber
(2013) worked with both bundles and p-frames in their study in order to
draw some comparisons between the two constructs. They started examin-
ing structures that occurred 100, 40, 5, 1 time(s) pmw, but they compro-
mised on a threshold of 40 times pmw and a range of 5+ texts. While some
p-frame studies follow the lexical bundle trends in terms of frequency and
range cut-off points, studies that analyze smaller corpora tend to choose
lower frequency and range thresholds for frames, such as a minimum fre-
quency of five and a range of 5+ texts (e.g., Yoon & Casal, 2020).
Once bundles have been identified, they are frequently classified gram-
matically (e.g., as prepositional phrases) and functionally (e.g., as discourse
organizers or stance markers). Even though these expressions are not always
complete grammatical units, they can be grouped according to their struc-
tural types. Biber et al. (2004) established three main grammatical catego-
ries for lexical bundle classification: 1) bundles that incorporate verb phrase
fragments (e.g., is based on the); 2) bundles that incorporate dependent
clause fragments (that there is a); and 3) bundles that incorporate phrasal
fragments (the rest of the). Studies that have analyzed lexical bundles of dif-
ferent sizes have found bundles that contain complex prepositions or subor-
dinators such as according to the and in order to, and longer bundles that are
full clauses or sentences as in the rest of the paper will be organized as follows.
The grammatical classification is often followed by a basic analysis of
the functions that these bundles perform in context. For this categoriza-
tion, analysts examine the text that surrounds the bundle, studying collo-
cations and common meanings in the discourse that precedes and follows
the bundle to identify semantic tendencies. Most current studies of lexical
bundles use previously designed taxonomies to conduct this classification.
110 Viviana Cortes

The functions in Biber et al.’s (2004) taxonomy are based on bundles iden-
tified in two corpora of everyday conversation and university textbooks,
and they comprise four main categories: stance expressions (e.g., are more
likely to, and the fact that the), discourse organizers (on the other hand or as
well as the), referential expressions (in the case of and in the presence of ), and
special conversational functions (thank you very much). Each of these func-
tions has multiple subcategories that these authors illustrate with examples
from the corpus (pp. 384–388). The subcategories in Hyland’s (2008) tax-
onomy were based on Biber et al.’s (2004), but the three major catego-
ries, research-oriented (e.g., in the present study), text-oriented (in the next
section), and participant-oriented bundles (as can be seen), emerged from
bundles frequently found in published research articles. Analysts gener-
ally examine bundles in context using concordancing software and analyze
their functions, classifying the bundles under the previously established
categories.
In the case of p-frames, once all the structures that meet the pre-estab-
lished thresholds and slot selections are identified by the software, two
important characteristics come into play: predictability and variability. The
predictability of a p-frame is measured as “the percentage of occurrences
of a frame in which the most frequent filler word is used” (Gray & Biber,
2013, p. 120). According to this predictability, p-frames can be classified
into highly predictable, predictable, less predictable, and unpredictable.
The variability of a frame is calculated as the ratio between the number of
distinct fillers (types) and the frequency of the frame (tokens) (Ren, 2022).
According to this type/token ratio, p-frames can be classified as highly vari-
able, variable, or fixed.
Next, p-frames are often also classified structurally. Gray and Biber
(2013) suggested three major structural categories of p-frames: verb-based
(e.g., was * in the), other lexical word-based (on the * hand), and function
word-based (in the * that). An important step in the analysis of p-frames is
the semantic grouping of their fillers. The analyst generally goes over the
list of fillers, regardless of their frequency, to identify their meanings and
then groups them accordingly. Analyzing the fillers and their collocates,
for example, can help identify specific semantic domains and create groups
of fillers with strong semantic relationships with pedagogical application
potential. For instance, in her analysis of published medical discourse,
Mbodj (2021) showed the semantic groupings of a wide variety of p-frame
fillers. In the case of the p-frame at a * of, some fillers were used to indicate
quantification, with fillers such as dose, concentration, rate, and ratio, and
some others were used to indicate location in time, with fillers like time,
end, start, beginning, midpoint, day, and date. Finally, p-frames are ana-
lyzed together with their fillers to identify the functions they perform in
 Lexical bundles and phrase frames 111

discourse. The taxonomies and procedures used for this step in the analysis
are similar to those used for bundle functional classification.
Recent studies of bundles and p-frames have been extending the func-
tional analysis of expressions in context. These studies often connect bun-
dles or frames to communicative purposes. For this analysis, bundles and
p-frames are examined in context, identifying the communicative purposes
that the authors tried to convey. This can be done by identifying moves and
steps in that particular discourse, and then finding connections between
the bundles or frames and those moves/steps (see the next section for
example studies). This type of analysis can have direct applications to the
genre-based academic writing class, where instructors can introduce moves
and steps in particular genres together with the bundles and/or p-frames
that most frequently occur in those moves/steps.

6.4 Example studies
In this section, I highlight some studies that focus on lexical bundles and
p-frames (see Table 6.1 for a comprehensive list). They were selected because
they contributed to the advancement of the investigation of these lexico-
grammatical features from a methodological perspective. Some of them
make strong connections between the formulaic expressions and rhetorical
functions represented by moves and steps in different genres.
I first discuss Gray and Biber (2013). These researchers take a register
approach rather than a genre approach to the study of linguistic features
such as continued and discontinued frequent word combinations, consider-
ing genre a macro-structure (Biber & Conrad, 2019), and register a vari-
ety of language identified through its situational context (see also Casal &
Kessler and Goulart & Staples, both in this volume). Even though Gray and
Biber did not make explicit connections to rhetorical functions, the authors
analyzed both bundles and p-frames, contributing many methodologi-
cal advances to the identification and analysis of these two constructs and
established several classifications of p-frames that can be used in the analy-
sis of these expressions in relation to the rhetorical functions they help con-
vey. Gray and Biber explained the approaches used to identify bundles and
p-frames and explain the differences between identifying frames from a list
of lexical bundles, which is what most available software does, and identify-
ing a full set of discontinuous sequences in a corpus. Using database tech-
niques, they identified frequent discontinuous sequences in which the open
slots had a wide variety of fillers, which showed no association between
these discontinuous sequences (p-frames) and continuous sequences (lexi-
cal bundles). They used two sub-corpora from the Longman Corpus of
Spoken and Written English: American conversation and academic prose
TABLE 6.1 
Example studies of lexical bundles and p-frames

Authors Context Research Questions (RQs) Data Collected Comments on Study Design

Casal and • Analysis of the • What are the frequent • 148 personal • The move scheme used
Kessler relationship between p-frames that occur in the statements/ had been previously
112 Viviana Cortes

(2020) moves and p-frames PS/SoGP corpus? statements of grant designed for the
in promotional genres • What is the distribution purpose (94,749 identification of
(personal statements/ of these p-frames across words) rhetorical functions in
statements of grant the rhetorical stages of the this particular genre
purpose) examined promotional • Instructors’ and
writing? students’ evaluation
• To what extent do of the pedagogical
instructors and L2 usefulness of the
English learners rate these frames could result
p-frames in terms of their in direct applications
pedagogical usefulness for of the p-frame-
teaching and/or learning move connection to
promotional writing? genre-based writing
instruction
Cortes (2013) • Lexical bundles • Which are the most • 1,372 introductions • The study set staggered
analyzed in academic frequent 4+-word lexical (1,002,748 words) frequency thresholds
writing represented bundles in a corpus for different bundles
by published R A of research article sizes and identified long
introductions introductions from a bundles (7+ words)
variety of disciplines?
• What are the structural
and functional connections
between bundles and
rhetorical moves and steps?
Gray and • Everyday American • Is the identification of • Conversation: 3,436 • The authors used
Biber English conversation p-frames with no relation texts (3,929,500 database techniques to
(2013) and academic texts from to lexical bundles possible? words) identify the full set of
textbooks and R As from • If so, how do these • Academic four-word sequences
the Longman Spoken p-frames and their register Prose: 408 texts and could test the
and Written Corpus distribution compare to (5,331,800 words) possibility that not all
examined for bundle p-frames identified in highly frequent frames
and p-frame analyses previous analyses? are associated with a
common lexical bundle

Lu et al. • P-frame analysis • What are the rhetorical • 600 introductions • The researchers used
(2021) in a corpus of R A functions of p-frames in (513,688 words) interrater reliability
introductions from a large corpus of social measures to ensure
the social sciences science research article agreement in move/
(anthropology, applied introductions? steps classification
linguistics, economics, • They worked with
political science, conservative thresholds
psychology, and for frequency and
sociology) variability of fillers
(Continued)
Lexical bundles and phrase frames 113
TABLE 6.1 Continued

Authors Context Research Questions (RQs) Data Collected Comments on Study Design

Pérez- • Lexical bundles • Which are the high- • L1 English: • Statistical analysis was
Llantada analyzed in published frequency lexical bundles 360 texts used to determine
114 Viviana Cortes

(2014) R As written by L1 in each language variable? (2,146,347 words) convergent and

English authors, L1 What are the defining • L2 English: divergent use of bundles
Spanish authors writing features of these bundles? 336 texts among the three groups
in English, and L1 • Is the choice of bundles (1,771,727 words) of authors represented
Spanish authors writing determined by register/ • L1 Spanish: in the corpus
in Spanish genre? 360 texts
• Which are core bundles (1,811,071 words)
shared by the three
language variables? Which
bundles are shared only by
L1–L2 English and only by
L2 English–L1 Spanish?
Are these bundles similar
or different structurally and
functionally? How do these
bundles build discourse
meanings?
• Which bundles are
distinctive to L1 English,
L2 English, and L1
Spanish? Do these
bundles involve distinctive
structures and functions?
 Lexical bundles and phrase frames 115

from textbooks and research articles (Biber et al., 1999). In the first analytic
stage, they identified four-word lexical bundles, with a frequency threshold
of 40 times pmw and a range of 5+ texts. For p-frames, they focused on
expressions that had only internal variable slots (A*BC, AB*C), and they
used variability and predictability measures adapted from the literature.
They started their analysis of p-frames using the same threshold they had
established for bundles, but they soon discovered that doing this, frames
that are not associated with the continuous sequences would go unnoticed.
They then lowered the frequency threshold for p-frames to identify struc-
tures that were not associated to bundles and also p-frames that presented a
wide variety of fillers. The most frequent p-frame in both registers was the *
of the. Frequent p-frames in conversation were structures such as I was * to,
I * want to, and do you * to, while academic prose showed structures like in
the * of, of the * of, a * of the, and and the * of, among others.
The authors emphasized the disadvantages of the bundle-to-frame
approach, explaining how the bundle-frame connection is stronger in
p-frames with lower type-token ratios, that is, p-frames that are less vari-
able, as many p-frames of this kind are often lexical bundles but p-frames
with a highly variable number of fillers in their open slots are not. The
authors used a structural classification method that has since then been
widely used in the field, comparing frames with different levels of variabil-
ity in their fillers (highly variable, variable, and fixed) across conversation
and academic prose. Their findings need to be taken into consideration
when analyzing bundles and p-frames in relation to writers’ communicative
intents, as it is important to clearly define and differentiate these constructs
at the identification and structural analyses before going deeper into their
functions in discourse.
The next exemplar study is Pérez-Llantada (2014), which clearly shows
the potential of lexical bundles to compare language production among
different genres, different languages, and different groups of language
users. While various studies have centered on the use of lexical bundles
with respect to disciplinary variation, Pérez-Llantada used the now well-
known corpus-driven approach to the identification and analysis of lexical
bundles in three corpora of research articles of around one million words:
research articles by L1 English writers published in international journals,
research articles by English L2 Spanish writers from the same international
journals, and research articles by L1 Spanish writers written in Spanish and
published in Spanish journals carefully selected through an analysis of the
situational characteristics of the texts in order to draw reliable comparisons.
The researcher set the bundle frequency threshold at 20 times pmw and a
dispersion of 10% of the texts in each corpus for the identification of four-
word lexical bundles. After the list of lexical bundles was automatically
116 Viviana Cortes

extracted from each of the three corpora using Kf Ngram (Fletcher, 2012),
bundles were classified structurally and functionally using existing taxono-
mies. A further stage in the analysis compared bundle production across
language groups, identifying convergent and divergent uses.
In line with previous studies of lexical bundles across languages, the
findings show very different numbers of lexical bundle types and tokens
across the three groups of writers, with the L1 English corpus yielding
the shortest list and the L1 Spanish corpus yielding the largest number of
bundles. The study was able to identify a group of 12 bundles used by the
three groups of writers. Among these expressions, we can find some time
referential bundles such as at the same time/a la vez que; at the time of/en el
momento de, and referential bundles used to mark textual reference like in
the present study/en el presente trabajo. The study also identified divergent
bundle use in expressions that were distinctive to only one group. The most
distinctive bundles in the L1 English group were stance markers expressing
probability/possibility or evaluation as in it is likely that, it is possible that,
and it is important to. L2 English writers preferred prepositional phrase
fragments like by the fact that and for the first time, and verb phrases, mostly
fragments of passive constructions as in be taken into account, can be found
in, is based on the, and related to the, often used to introduce evidential
data. Finally, the L1 Spanish group of writers showed their divergent use
in many bundles with the quasi-passive “se” in expressions such as a lo que
se (to which) or por lo que se (for which), which are expressions bound to
the nature of Spanish, as well as the frequent use of noun phrase bundles
with post modification of the noun such as la influencia de la (the influence
of the); la importancia de la (the importance of the); and la posibilidad de
que (the possibility that). An important conclusion is that the L2 English
writers’ use of lexical bundles showed a group of hybrid expressions, with
some bundles that overlapped with those produced by L1 English writ-
ers and some others translated from Spanish into English. Although this
study did not directly analyze bundles in relation to rhetorical moves/steps,
the results show strong connections between certain bundles used by one
group of writers and particular rhetorical functions within the research
article genre. These connections can be seen in the bundles used by L1 and
L2 English writers, who used discourse organizing bundles used for identi-
fication/focus purposes (e.g., be taken into account/que tener en cuenta, the
analysis of the/el análisis de la) to introduce discussion points.
The next example studies described are Cortes (2013) and Lu et al.
(2021), which both go beyond the grammatical and functional classification
of bundles and p-frames, tying these frequent expressions to communica-
tive purposes as represented by rhetorical moves and steps (Swales, 1990,
2004) in the research article genre. In her study, Cortes (2013) introduced
 Lexical bundles and phrase frames 117

a new direction to the analysis of lexical bundles. After having identified

and classified lexical bundles in different types of academic registers and
languages (Cortes, 2004, 2008), the author focused on research article
(R A) introductions and took the functional analysis of bundles a step fur-
ther. The study analyzed a one-million-word corpus of R A introductions in
search of bundles of different sizes, from four-words to the longest bundle
that could be identified in her corpus. For this purpose, different frequency
thresholds were established considering that the longer the bundle, the
less frequently it occurs. Using an updated version of her Lexical Bundle
Program, Cortes identified some nine-word bundles that occurred more
than ten times pmw, something that seemed impossible in earlier studies
that have not found bundles longer than six words (see Biber et al., 1999).
After she analyzed these bundles structurally and functionally using previ-
ously designed taxonomies, she studied each bundle in context to identify
the move/step in which the bundle was used, using Swales’ (2004) scheme
for R A introductions.
In her analysis, Cortes discovered that some lexical bundles were exclu-
sively used in one move/step and others were used across different moves/
steps. In addition, she noticed that some bundles initiated the move/step
(triggers), while other expressions were used later in the discourse (com-
ments) with other words and expressions initiating the move/step. For
example, the bundle little is known about was tied to Move 2 (indicating
the gap), and it was used to trigger that move. The bundle in the sense
that, on the other hand, was tied to Move 3 Step 1 (announcing present
research descriptively and/or purposefully), but was used as a comment
bundle, adding information to other expressions that triggered that move.
In spite of having 13 different disciplines in her corpus, Cortes’s study did
not intend to show lexical bundle variation across disciplines. However, she
did find some bundles that appeared in only one or two disciplines, as in
the case of the rest of the paper will be organized as follows, which was used
in business article introductions.
The study conducted by Lu et al. (2021) had a purpose similar to that of
Cortes (2013) but focused on p-frame use in the introduction sections of
R As from the social sciences, analyzing the relationship between p-frames
and rhetorical moves and steps. The methodology employed was different
because the corpus they used had previously been entirely annotated for
moves and steps using a scheme adapted from Swales (1990, 2004). For
this study, the authors used a corpus of 600 introductions and about half
a million words. The identification of five and six-word p-frames with only
one internal open slot was done using Kf Ngram (Fletcher, 2012) and in a
pilot analysis, the authors explored different frequency cut-off points con-
sidering not only the number of p-frames generated by the software but
118 Viviana Cortes

also p-frames that could have potential pedagogical applications. They also
paid special attention to p-frames that had two or more variants or fillers
and that occurred across several texts in the corpora and across two or more
disciplines.
In their analyses, the authors identified p-frames that were associated
with only one move/step (specialized p-frames), p-frames that were asso-
ciated with more than one move/step but had a strong link to only one
move (semi-specialized p-frames), and p-frames that occurred in multiple
moves/steps (non-specialized p-frames). For example, the p-frame the * of
this study is, with fillers such as aim, focus, goal, and purpose, was only
found in Move 3 Step 1 (announcing present research). The results of this
study suggest certain associations between p-frames and their fillers and
particular moves and steps that deserve further analysis in order to explore
how these p-frames could become an important element in genre-based
academic writing classes.
In the last study selected, Casal and Kessler (2020) examined the forms
and rhetorical functions of p-frames in a corpus of the promotional writing
genre, represented by personal statements and statements of grant purpose
from accepted grant applications. The study had three phases: identification
of frequent p-frames; distribution of p-frames across the rhetorical stages
in the statements; and assessment of the potential pedagogical usefulness
of the p-frames for teaching and learning statement writing. They used a
move scheme previously designed for this genre (see Kessler, 2020), fol-
lowing a process that “involved consideration of rhetorical cues including
lexical, phraseological, and structural indications of rhetorical intent and
shift” (p. 4). Their analysis showed high levels of interrater reliability. The
resulting scheme comprised four moves: competence claims, motivation
for pursuing grant, motivation for applying to target country, and fram-
ing childhood and family history. The authors used Kf Ngram (Fletcher,
2012) to automatically extract five-word p-frames with internal and exter-
nal slots that occurred in at least five texts. Five-word structures were tar-
geted because they seem to be more “linguistically cohesive and complete”
than four-word frames (p. 5). The final list had 69 p-frames after some
p-frames were eliminated that were not linguistically complete, or crossed
clause boundaries. The next step was to analyze the p-frames in the moves
in which they occurred and calculate the association that each p-frame had
with each move.
The results of this frame-move identification showed some frames that
were bound to one or two moves. For example, Move 1, showed a strong
association to the p-frame I was able to *, reflecting the authors’ intentions
to show their competencies. For the last stage in the study, the authors
summoned a group of ten writing instructors and ten L2 English graduate
 Lexical bundles and phrase frames 119

students, who evaluated the pedagogical potential of the p-frames using a

Likert scale “ranging from 1 (no pattern, not useful), to 4 (pattern is rec-
ognizable and very useful)” (p. 7). Participants also provided comments on
the p-frame evaluations. The authors then focused on the analysis of groups
of p-frames that had received very high and very low ratings. Regarding the
comments received, instructors highly regarded p-frames that showed var-
ied useful fillers, had direct functional associations, and could result in easy
transferability to other genres. Casal and Kessler emphasize the potential
of their findings to inform pedagogical approaches that combine p-frames,
moves, and other forms that can connect linguistic exponents and rhetori-
cal functions.

6.5 Issues and challenges

Several problems may arise in the identification of lexical bundles and
p-frames. The main problem with lexical bundle identification comes from
the corpora used and the thresholds established for frequency and range.
It is important to explain that normalization, a procedure used to adjust
raw frequencies from texts or corpora of different sizes so that they can be
compared accurately, often results in problematic lexical bundles studies
(Bestgen, 2019). After normalization, smaller corpora tend to yield many
more lexical bundles than larger corpora because this process inflates fre-
quency counts. In addition, word combinations that occur only a few times
could be considered lexical bundles when normalization is used. It is possi-
ble to study lexical bundles in small corpora, but the frequency cut-off point
needs to remain high (e.g., ten times in a corpus of any size) to ensure that
the expressions identified are still relatively frequent. Comparisons among
corpora of different sizes may also be problematic in the study of p-frames.
Another important issue is that of overlapping bundles, which were first
discussed in Chen and Baker (2010). This issue is related to the independ-
ence of observation in the study of both bundles and p-frames because to
ensure that independence, each observation must be counted and analyzed
only once, particularly when further analyses focus on the relationship
between these expressions and communicative functions. An example of
this issue can be found in the bundle as a result of the, which is very frequent
in written academic genres. The expressions as a result of and as a result may
be embedded in the longer expression, but in order to be considered three
different lexical bundles, each of these word combinations should meet
the pre-established frequency and range thresholds independently. If the
expression as a result of the occurs 20 times pmw on its own, that expres-
sion would be considered a bundle. For as a result of and as a result to be
considered bundles, their individual frequencies, occurrences not linked
120 Viviana Cortes

to as a result of the, should meet their corresponding frequency thresholds

individually. This issue is particularly problematic when studies look at lexi-
cal bundles of only one size (e.g., four words or five words) without taking
care of the potential overlaps. This identification is important because some
of these expressions may have different functions in discourse, and these
functions need to be accurately classified and accounted for. While some
studies have paid attention to this issue in detail and completed a manual
analysis of any possible overlap (Cortes, 2013), Cortes and Lake (2023)
have created a computer program, the LBiaP (Lexical Bundle Identification
and Analysis Program), which yields lists of bundles of three to nine words,
eliminating all types of overlap. The issue of overlapping expressions may
also affect p-frames because in many studies, p-frames originate on a list of
n-grams or lexical bundles. In this case, an initial step that takes care of this
overlap at the n-gram identification stage, prior to obtaining frames, could
eliminate any of these inconsistencies.
With a marked rise in studies that analyze p-frames with outer slots, it
will be necessary to investigate the boundaries between n-grams/bundles
and p-frames. It can be argued that a p-frame with an outer slot at the
end of the expression could be a frequent n-gram or even a lexical bun-
dle which frequently collocates with certain words or expressions. Future
studies need to make methodological decisions in terms of an “either/or”
approach to each of these formulaic types, or an approach that may consider
these expressions as part of a continuum that goes from continuous to dis-
continuous expressions.
Finally, there may be problems that arise from comparing lexical bundle
or p-frame use across L1 and L2 language production. In the case of L2 use
of bundles, if the texts produced by learners contain grammatical errors,
frequent expressions that are incomplete or use words that do not follow
the tendency would not be identified as lexical bundles by the software that
uses a corpus-driven methodology. Thus, a preliminary analysis of some
core expressions (e.g., Shin et al., 2018), structures embedded in the most
frequent bundles under analysis, or the correction of those misuses in an
initial stage (e.g., Kim & Kessler, 2022), could help identify any problem
and make decisions before starting any type of comparison.

6.6 Study-in-focus
For this section, I highlight a study by Omidian et al. (2018). In their
study of lexical bundles in R A abstracts, the authors investigated discipli-
nary variation, analyzing the bundles used to convey different rhetorical
purposes in this genre. Before continuing, readers are encouraged to read
the description of the study in the call-out box.
 Lexical bundles and phrase frames 121

Citation
Omidian, T., Shahriari, H., & Siyanova-Chanturia, A. (2018). A cross-discipli-
nary investigation of multi-word expressions in the moves of research article
abstracts. Journal of English for Academic Purposes, 34, 1–14.

Research questions
• How do writers from hard and soft fields differ in terms of the distri-
bution of the sequences they frequently use in different moves of RA
abstracts?
• How do writers from hard and soft knowledge fields differ in their use
of the sequences employed to fulfill the rhetorical functions of different
moves in RA abstracts?

Context and population

The study analyzed lexical bundles of four words or longer in a corpus of RA
abstracts from top journals in six disciplines grouped into soft (applied lin-
guistics, sociology, and marketing) and hard (physics, biology, and mechani-
cal engineering) scientific areas. The corpus was about one million words
equally distributed across the six disciplines, with each sub-corpus having
around 980 abstracts and around 167,000 words.

Procedure
These are the steps the researchers followed in their analysis: 1) identify
4+word lexical bundles with staggered frequencies and a constant range; 2)
classify those bundles structurally and functionally using previously designed
taxonomies; 3) study each bundle in relation to the moves it helped realize;
4) calculate interrater reliability of move-bundle analysis; 6) perform statisti-
cal analysis to establish any significant difference of use across soft and hard
scientific fields; 7) compare bundle functions in moves across hard and soft
knowledge areas.

Findings
The results of the study showed a marked distinction between bundles used
by authors in hard and soft scientific areas in relation to particular moves. For
example, writers in hard knowledge fields used significantly more bundles
122 Viviana Cortes

in Move 3 (describing the methodology), reflecting the conventionality of

this section of the abstract in these disciplines. In soft scientific areas, writers
used more bundles in Move 2 (presenting the research and its purpose),
possibly to ensure the purpose of their study is clearly understood by their
readers.
The study also reported more research-oriented bundle use by hard science
writers while writers in soft fields made use of more text- and participant-
oriented bundles, showing that writers in each of these areas have different
priorities when describing their studies and promoting their findings.

In particular, Omidian et al. (2018) were selected because their research

methodology was sound and carefully executed, from the collection of a
well-balanced corpus to the thresholds used for bundle frequency and range.
They established various frequency thresholds for their study: twenty times
pmw for four words, ten times pmw for five words, seven times pmw for six
and seven words, and five times pmw for any longer bundles, and a range
of 5+ texts. For their rhetorical analysis of bundles, they used a scheme
adapted from Dos Santos (1996) and Pho (2008) and focused on the dif-
ferences in the use of bundles in moves across writers from hard and soft
scientific areas. The moves in their scheme were 1) situating the research,
2) presenting the research, 3) describing the methodology, 4) summarizing
the results, and 5) discussing the research. Each of these moves relates to
a section of IMRD (Intr​oduct​ion-M​ethod​s-Res​ults-​Discu​ssion​) in R As.
The study shows marked differences in the way the writers in the two
groups express their rhetorical purposes. Writers in the soft fields used
hedging stance bundles like are more likely to in Move 4 of their abstracts
(summarizing the results). The authors interpret this use as the writers’
intention to avoid overgeneralizations and present their results in a more
tentative way. Hard-field writers, on the other hand, use bundles like in
good agreement with the when summarizing their results in the same move,
in order to provide validity to those results. The authors conclude that the
difference in the tendencies of bundle use across the two groups of writers
can help “understand disciplinary variation and the construction of knowl-
edge in different sections of the R A” (p. 12).

6.7 Future research directions

Although it may seem lexical bundle and p-frame studies are numer-
ous, many genres have not been studied from a corpus-driven formulaic
 Lexical bundles and phrase frames 123

language perspective. As spoken corpora are relatively scarce, corpus-based

and corpus-driven studies of spoken registers are not as common as those
of written discourse, but there are still many written registers that need to
be explored through their use of bundles and p-frames, particularly specific
registers that are not academic, such as websites, blogs, popular science
reports, or text messages.
It is necessary to point out that there are very few studies that have
looked at both lexical bundles and p-frames in discourse. Analysts tend to
prefer only one of these two constructs. An exception seems to be Mbodj
(2021), whose study analyzed two-word lexical collocations, multi-word
collocations, lexical bundles, and p-frames in medical discourse from R As
and case reports in an attempt to describe the complete formulaic profile of
these registers.
It would also be important to revisit the functional taxonomies that have
been used to classify bundles and p-frames. Many studies that use these
taxonomies employ a top-down approach. That is, they use the existing
categories and classify new bundles comparing them to similar expressions
already classified. These taxonomies have proved to be sound, as they have
been very useful in the categorizations of bundles and p-frames from a
wide variety of registers and genres. However, the taxonomies only show
the functions that certain bundles performed most frequently. That means
that some bundles may be performing secondary functions that also need
to be investigated.
Finally, lexical bundles and p-frames seem to have a strong potential
for direct application to the teaching of specific genres, particularly for
the teaching of academic writing in different contexts. However, although
most studies of bundles and p-frames seem to have a pedagogical impetus,
few investigations explored real pedagogical applications for these expres-
sions and structures. Among these studies, we can find Cortes (2006),
who taught bundles in context to a group of students in a writing-inten-
sive history class. The results of her study show inconsistent gains in the
participants’ written production quality in general and in their use of lexi-
cal bundles in their writing tasks in particular. On the other hand, Kazemi
et al. (2014), who replicated Cortes’s study with a group of EFL university
students from Iran, claimed the teaching of lexical bundles had a strong
impact on the writing of the participants, attested by the results of pre-
and post-intervention statistical analysis of their writing samples. Shin and
Kim (2017) explored how the teaching of lexical bundles could be used
to help L2 writers improve some grammatical problem areas, such as the
appropriate use of definite and indefinite articles. Regarding p-frames,
even though they have been used to identify differences between L1 and
L2 language users and among L2 writers at different levels of proficiency,
124 Viviana Cortes

studies that show practical pedagogical applications of p-frames are rare.

Therefore, there is a real need for more investigations that could help
translate lexical bundle and p-frame research findings to the language
classroom.

References
Allen, S. (1975). Frequency dictionary of present-day Swedish based on newspaper, 3:
Collocations. Stockholm: Almqvist & Wiksell.
Altenberg, B., & Eeg-Olofsson, M. (1990). Phraseology in spoken English:
Presentation of a project. In J. Aarts & W. Meijs (Eds.), Theory and practice in
corpus linguistics (pp. 1–26). Amsterdam: Rodopi.
Anthony, L. (2021). AntGram (version 1.3.0) [Computer software]. Tokyo,
Japan: Waseda University. Retrieved from https://www​.laurenceanthony​.net​/
software
Anthony, L. (2022). AntConc (version 4.1.0) [Computer software]. Tokyo,
Japan: Waseda University. Retrieved from https://www​.laurenceanthony​.net​
/software
Barlow, M. (2004). Collocate (version 1.0) [Computer software]. Houston, TX:
Athelstan. Retrieved from https://collocationary​.com/
Bestgen, Y. (2019). Comparing lexical bundles across corpora of different sizes:
The Zipfian problem. Journal of Quantitative Linguistics, 27(3), 272–290.
Retrieved from https://www​.tandfonline​.com​/doi​/full​/10​.1080​/09296174​
.2019​.1566975
Biber, D. (2009). A corpus driven approach to formulaic language in English.
International Journal of Corpus Linguistics, 14, 275–311. https://doi​.org​/10​
.1075​/ijcl​.14​.3​.08bib
Biber, D., & Conrad, S. (1999). Lexical bundles in conversation and academic
prose. In H. Hasselgard & S. Oksefjell (Eds.), Out of corpora: Studies in honor of
Stig Johansson (pp. 181–189). Amsterdam: Rodopi.
Biber, D., & Conrad, S. (2019). Register, genre, and style. Cambridge: Cambridge
University Press
Biber, D., Conrad, S., & Cortes, V. (2004). ‘If you look at …’: Lexical bundles in
university teaching and textbooks. Applied Linguistics, 25(3), 371–405. https://
doi​.org​/10​.1093​/applin​/25​.3​.371
Biber, D., Johansson, S., Leech, G., Conrad, S., & Finegan, E. (1999). Longman
grammar of spoken and written English. New York: Longman. https://doi​.org​
/10​.1075​/ijcl​.18​.1​.08gra
Casal, J. E., & Kessler, M. (2020). Form and rhetorical function of phrase-frames
in promotional writing: A corpus- and genre-based analysis. System, 95, 102370.
https://doi​.org​/10​.1016​/j​.system​.2020​.102370
Casal, J. E., & Kessler, M. (2023). Rhetorical move-step analysis. In M. Kessler
& C. Polio (Eds.), Conducting genre-based research in applied linguistics: A
methodological guide. Routledge.
Chen, Y. H., & Baker, P. (2010). Lexical bundles in L1 and L2 academic writing.
Language Learning and Technology, 14(2), 30–49. Retrieved from http://hdl​
.handle​.net​/10125​/44213
 Lexical bundles and phrase frames 125

Cortes, V. (2004). Lexical bundles in published and student disciplinary writing:

Examples from history and biology. English for Specific Purposes, 23(4), 397–
423. https://doi​.org​/10​.1016​/j​.esp​.2003​.12​.001
Cortes, V. (2006). Teaching lexical bundles in the disciplines: An example from
a writing intensive history class. Linguistics and Education, 17(4), 391–406.
https://doi​.org​/10​.1016​/j​.linged​.2007​.02​.001
Cortes, V. (2008). A comparative analysis of lexical bundles in academic history
writing in English and Spanish. Corpora, 3(1), 43–58. https://doi​.org​/10​.3366​
/E1749503208000063
Cortes, V. (2013). The purpose of this study is to: Connecting lexical bundles
and moves in research article introductions. Journal of English for Academic
Purposes, 12(1), 33–43. https://doi​.org​/10​.1016​/j​.jeap​.2012​.11​.002
Cortes, V., & Lake, W. (2023). LBiaP: A solution to the problem of attaining
observation independence in lexical bundle studies. International Journal of
Corpus Linguistics, 28, 263–277. https://doi​.org​/10​.1075​/ijcl​.21100​.cor
Dos Santos, M. B. (1996). The textual organization of research paper abstracts in
applied linguistics. Text-Interdisciplinary Journal for the Study of Discourse, 16,
481e500. https://doi​.org​/10​.1515​/text​.1​.1996​.16​.4​.481.
Fletcher, W. H. (2012). KfNgram [Computer software]. Retrieved from http://
kwicfinder​.com​/ kf Ngram​/ kf NgramHelp​.html
Flowerdew, J. (2002). Genre in the classroom: A linguistic approach. In A. Johns
(Ed.), Genre in the classroom: Multiple perspectives (pp. 91–103). Mahwah, New
Jersey: Lawrence Erlbaum Associates.
Flowerdew, J., & Li, Y. (2007). Language re-use among Chinese apprentice
scientists writing for publication. Applied Linguistics, 28(3), 440–465. https://
doi​.org​/10​.1093​/applin​/amm031
Goulart, L., & Staples, S. (2023). Multidimensional analysis. In M. Kessler &
C. Polio (Eds.), Conducting genre-based research in applied linguistics: A
methodological guide. Routledge.
Gray, B., & Biber, D. (2013). Lexical frames in academic prose and conversation.
International Journal of Corpus Linguistics, 18(1), 109–135. https://doi​.org​
/10​.1075​/ijcl​.18​.1​.08gra
Hyland, K. (2008). As can be seen: Lexical bundles and disciplinary variation.
English for Specific Purposes, 27(1), 4–21. https://doi​.org​/10​.1016​/j​.esp​.2007​
.06​.001
Hyon, S. (2018). Introducing genre and English for specific purposes. Routledge.
Kazemi, M., Katiraeib, S., & Raseckho, A. (2014). The impact of teaching lexical
bundles on improving Iranian EFL students’ writing skills. Procedia: Social and
Behavioral Sciences, 98, 864–869. https://doi​.org​/10​.1016​/j​.sbspro​.2014​.03​.493
Kessler, M. (2020). A text analysis and gatekeepers’ perspectives of a promotional
genre: Understanding the rhetoric of Fulbright grant statements. English for
Specific Purposes, 60, 182–192. https://doi​.org​/10​.1016​/j​.esp​.2020​.07​.003
Kim, S., & Kessler, M. (2022). Examining L2 English university students’ uses of
lexical bundles and their relationship to writing quality. Assessing Writing, 51:
100589
Lu, X., Yoon, J., & Kisselev, O. (2021). Matching phrase-frames to rhetorical moves
in social science research article introductions. Journal of English for Academic
Purposes, 61, 66–83. https://doi​.org​/10​.1016​/j​.esp​.2020​.10​.001
 126 Viviana Cortes

Mbodj, N. B. (2021). Writing in the disciplines and within-discipline variations: A

comparison of the formulaic profiles of the medical research article and the medical
case report. Retrieved from https://scholarworks​.gsu​.edu​/alesl​_diss​/61/.
[[Doctoral Dissertation]. Georgia State University]
Miller, T. (1993). The introduction to the research article from a discourse
perspective. ASp, 2(2), 55–68.
Omidian, T., Shahriari, H., & Siyanova-Chanturia, A. (2018). A cross-disciplinary
investigation of multi-word expression in the moves of research article abstracts.
Journal of English for Academic Purposes, 36, 1–14. https://doi​.org​/10​.1016​/j​
.jeap​.2018​.08​.002
Perez Llantada, C. (2014). Formulaic language in L1 and L2 expert academic
writing: Convergent and divergent usage. Journal of English for Academic
Purposes, 14, 84–94.
Pho, P. D. (2008). Research article abstracts in applied linguistics and educational
technology: A study of linguistic realizations of rhetorical structure and
authorial stance. Discourse Studies, 10(2), 231e250. https://doi​.org​/10​.1177​
/1461445607087010.
Ren, J. (2022). A comparative study of the phrase frames used in the essays of
native and nonnative English students. Lingua, 274, 103376. https://doi​.org​
/10​.1016​/j​.lingua​.2022​.103376
Renouf, A. J., & Sinclair, J. (1991). Collocational frameworks in English. In K.
Ajimer & B. Altenberg (Eds.), English corpus linguistics: Studies in honour of Jan
Svartvik (pp. 128–143). New York: Longman.
Römer, U. (2010). Establishing the phraseological profile of a text type: The
construction of meaning in academic book reviews. English Text Construction,
3(1), 95–11. https://doi​.org​/10​.1075​/etc​.3​.1​.06rom
Shin, Y., Cortes, V., & Yoo, I. W. H. (2018). Using lexical bundles as a tool to
analyze definite article use in L2 academic writing. Journal of Second Language
Writing, 39, 29–41. https://doi​.org​/10​.1016​/j​.jslw​.2017​.09​.004
Shin, Y., & Kim, Y. (2017). Using lexical bundles to teach articles to L2 English
learners of different proficiencies. System, 69, 79–91. https://doi​.org​/10​.1016​/j​
.system​.2017​.08​.002
Swales, J. (1990). Genre analysis: English in academic and research settings.
Cambridge: Cambridge University Press.
Swales, J. (2004). Research genres: Explorations and applications. Cambridge:
Cambridge University Press.
Yoon, J., & Casal, J. E. (2020). P-frames and rhetorical moves in applied linguistics
conference abstracts. In U. Römer, V. Cortes & E. Friginal (Eds.), Advances
in corpus-based research on academic writing: Effects of discipline, register, and
writer expertise (pp. 282–305). Amsterdam: John Benjamins.
7
MULTIDIMENSIONAL ANALYSIS
Larissa Goulart, Shelley Staples

7.1 Introduction to the approach and definition of genre

Multidimensional analysis (MDA) is a research method initiated by Biber
(1988) which typically uses the statistical procedure of exploratory factor
analysis (EFA) (and more recently principal component analysis, PCA), but
also includes methodological processes that are particular to using EFA
with linguistic data. Data for MDA consists of a corpus of texts that rep-
resent individual writing or speaking events. While EFA and its corollary
PCA are used a great deal within applied linguistics, these statistical proce-
dures are usually applied to studies that investigate questionnaire data (e.g.,
examining beliefs, attitudes, and/or individual differences; see Loewen &
Gonulal, 2015). Thus, MDA can be considered a distinct method that can
be used to investigate spoken and written genres.
Genre has been defined differently within MDA studies and has changed
over time. In the first MDA published, Biber (1988) used genre to catego-
rize the texts included in his corpus. In this study, Biber (1988) defined
genre as “text categorizations made on the basis of external criteria relat-
ing to author/speaker purpose” (p. 68). For example, the genre of aca-
demic expository prose is defined primarily by its ideational (informational)
purpose (Biber, 1988, p. 39). Thus, his definition aligns with the shared
characteristic of all three genre schools, Rhetorical Genre Studies (RGS),
Systemic Functional Linguistics (SFL), and English for Academic/Specific
Purposes (EAP/ESP), with genre guided by external, social categoriza-
tions, primarily communicative purpose. In this seminal MDA study,
genre is seen as an outgrowth of a speech situation, which encompasses

DOI: 10.4324/9781003300847-9
128 Larissa Goulart, Shelley Staples

the “configurations of cultural, physical, temporal, and psychological fea-

tures that define the situational context of texts” (p. 25). In defining genre,
Biber (1988) draws heavily on sociolinguistic antecedents, particularly Dell
Hymes’ (1974) SPEAKING model, which includes message form, con-
tent, speaker, hearer, purposes, key, channels, and norms of interaction.
Importantly, the concept of genre was used in early MDAs to explore dif-
ferences between speech and writing, and to some extent this tradition
continues in recent MDAs, now under the aegis of register.1 Regardless of
whether MDA researchers use genre or register as their framework for clas-
sifying texts, they typically see three elements as important to the analysis:
the situation of use (i.e., speech situation; for example, academic prose or
face-to-face conversation); the linguistic features used within that situation
of use (e.g., technical vocabulary in academic prose vs. first-person pro-
nouns in face-to-face conversation); and the functional interpretation of the
linguistic features within the situation of use (e.g., first-person pronouns
are used to perform the interpersonal function in face-to-face conversation
while technical vocabulary is used to perform the ideational function in
academic prose). These three elements are usually referred to as the “reg-
ister triangle.”
In later work, Biber and colleagues (e.g., Biber & Conrad, 2019) iden-
tify communicative purpose as an aspect of register and shift to discussing
genre as a particular methodological perspective on complete texts, distin-
guishing genre markers (linguistic devices that might occur once within
a genre, such as a salutation in a letter) from register markers (pervasive
linguistic features within multiple texts from the same situation of use).
Two key MDAs that are particularly useful for genre scholars are Nesi
and Gardner (2012) and Gardner et al. (2019), who use a modified version
of SFL to identify genres in the British Academic Written English (BAWE)
corpus (Nesi et al., 2004-2007). To categorize texts, they focus not only
on communicative purpose but also the stages within the texts (Halliday &
Matthiessen, 2004), which can be seen as somewhat akin to Swales’ (1990)
moves. The genre classification of BAWE also was informed by interviews
with faculty and students as well as other artifacts (e.g., pedagogical mate-
rials). Thus, the BAWE team focused on the context of instruction and
members of the speech communities more than many other MDAs. They
also use the notion of genre families as a larger categorization, focusing
on communicative purpose and general (rather than specific) stages. For
example, essays have the purpose “to demonstrate/develop the ability to
construct a coherent argument and employ critical thinking skills” (Nesi
& Gardner, 2012, p. 38), and their stages generally involve “introduction,
series of arguments, conclusion” (p. 38). However, there are many kinds
of essays within this larger genre family, often distinguished by their more
 Multidimensional analysis 129

specific stages. For example, an expository essay might have the stages “the-
sis, evidence, restate thesis” (p. 99), but a challenge essay might have the
stages “challenge, evidence, thesis” (p. 101), thus distinguishing the two
genres.
Admittedly, the category of genre is not well defined in MDA as a
whole. However, MDA offers genre scholars another way to identify
“genre conventions,” including formal, specific genre knowledge (Tardy
et al., 2020). In this sense, the constellations of pervasive linguistic fea-
tures identified through MDA provide insight into the language typically
used to enact socially recognized genres, providing a complement to the
often more qualitative work of genre analysis. In addition, the co-occur-
rence of linguistic features allows scholars to focus on higher-level func-
tions within texts (rather than simply identifying forms that frequently
occur). Particularly when combined with move analysis (e.g., Connor &
Upton, 2004; Gray et al., 2020), MDA also provides possible inroads to
understanding the rhetorical purposes of the co-occurring lexico-gram-
matical features. Recent work using MDA also emphasizes the importance
of variation within registers, which aligns with genre scholars’ interest in
looking “beyond conventions” (see Tardy, 2016) to understand how writ-
ers vary their linguistic and rhetorical choices within a socially accepted
genre.

7.2 Goals
The understanding that linguistic variation occurs along multiple func-
tional layers (or dimensions) was the initial motivation to use a multivari-
ate statistical technique to investigate language variation (Biber, 2019).
Dimensions represent groups of linguistic features that tend to co-occur in
a text, and this co-occurrence is associated with the communicative func-
tions that these features convey. For example, researchers that have applied
MDA to university writing found that proposals and laboratory reports are
similar in their use of linguistic features that have the function of report-
ing a procedure (e.g., passive voice, technical nouns), but these two genres
differ in their use of linguistic features that have the function of expressing
possibility (e.g., modal verbs, if-conditionals, and present tense). That is,
proposals and laboratory reports are similar in the dimension of procedural
reporting, but different in the dimension of expression of possibility.
With the use of EFA, MDA allows researchers to go beyond the analy-
sis of a single linguistic feature (e.g., past tense or hedging strategies) and
consider the use of a variety of linguistic features. In Biber’s (1988) seminal
study, for example, the author accounted for the use of 67 linguistic fea-
tures in spoken and written texts.
130 Larissa Goulart, Shelley Staples

Traditionally, researchers have used this approach to describe the extent

to which a group of texts varies along multiple dimensions. Many MDAs
have explored language variation broadly, comparing written and spoken
discourse (e.g., Berber-Sardinha et al., 2014; Katinskaya & Sharoff, 2015).
These MDAs usually explore questions such as the extent to which gen-
res differ from each other in a specific language. More recently, however,
researchers have conducted MDAs in more specialized genres, such as fic-
tion (Egbert & Mahlberg, 2020), or TV shows (Quaglio, 2009). These
MDAs provide a more detailed description of how language varies within
specialized genres, taking into account variables such as discipline, level of
proficiency, or type of TV broadcast.
MDAs have described the language characteristics of a broad range of
genres. Nevertheless, as shown by a recent survey of MDA studies (i.e.,
Goulart & Wood, 2021), academic writing and student/learner genres
have been the focus of the majority of MDAs until now. This focus on
academic genres indicates that there are a number of other genres (e.g.,
professional genres, spoken discourse) that have yet to be described tak-
ing an MDA approach. Researchers conducting an MDA might investigate
research questions such as:

• What are the dimensions of linguistic co-occurrence in the genre(s)

investigated?
• To what extent is there linguistic variation across genres?
• What are the pervasive linguistic characteristics of a genre?
• How do the linguistic characteristics of a genre compare to another
genre(s)?
• How do the linguistic characteristics of novice writers (or language
learners) differ from that of expert writers?

7.3 Common research methods

As mentioned above, MDA uses a dimension-reduction technique (EFA
or PCA) to analyze the co-occurrence of a number of linguistic features in
a specific discourse domain. Although both of these statistical techniques
can be used to reduce a number of linguistic variables into factors of co-
occurrence, they differ in many aspects including how they account for the
total variance between features. While PCA accounts for the total variance
between features, including the error variance, EFA takes into considera-
tion only covariance. This makes PCAs a better approach to dealing with
data that includes zeros.
The factors extracted by the statistical analysis are functionally inter-
preted in relation to the linguistic features that load in each dimension.
 Multidimensional analysis 131

These factors have a positive and a negative side, but this does not mean
that the features on the positive side are better than the ones on the nega-
tive side. Rather, features on the positive and negative sides of a dimension
occur in complementary distribution. That is, a text that has a negative
loading will have a higher frequency of the features that occur on the nega-
tive side and a lower frequency of the features that occur on the positive
side than a text that has a high positive loading. In Gray (2015), for exam-
ple, Dimension 1 contains, among other features, pronouns (bold), gen-
eral adverbs (italicized), modal verbs (double underlined), if-conditionals
(underlined), and verb complement clauses (dotted underlined). These fea-
tures co-occur in certain texts as illustrated in the excerpt below, where
positive features are bolded:
Quantitative Biology

Due to the large number of results described here, we make a stepwise

and systematic approach, proving each step carefully even if they might
have been proved previously in other publications (always referenced).
This means also systematically, that we do not use the disproved models
to attempt to interpret later steps in the development, to avoid confusion.

The negative side of the same dimension contains, among other features,
nouns (underlined), prepositions (double underlined), past tense verbs
(bold), passives (italicized), and passive postnominal modifiers (between
brackets). These features tend to co-occur together in specific texts, as the
excerpt below illustrates (also from Gray, 2015, p. 147):
Quantitative Biology

Stable isotopes of carbon and nitrogen were used to estimate food-chain

length and identify patterns of material flow through dominant trophic
pathways for each food web. The aquatic food webs [analyzed] in this
study were delimited by fishes as consumers plus their aquatic and riparian
prey and production sources [consumed] throughout the web leading to
those consumers.

Based on a qualitative analysis of how these features function in the corpus,

the author interpreted this dimension as academic involvement and elabora-
tion on the positive side versus informational density on the negative side.
The results of Gray (2015) show that research articles in theoretical phi-
losophy have a high positive factor loading, meaning that research articles
written for this discipline have a great deal of features found on the positive
side (e.g., modal verbs, conditionals, and pronouns), and few features found
132 Larissa Goulart, Shelley Staples

on the negative side (e.g., nouns, prepositions, and past tense). Conversely,
research articles in quantitative biology have a negative factor loading for
this dimension, meaning that these texts use a great deal of features found
on the negative side of this dimension, and few features found on the posi-
tive side of this dimension.
Traditionally, MDAs are conducted using EFA, a statistical proce-
dure that considers the correlation between a wide number of variables.
Recently, however, a number of studies have adopted other dimension-
reduction techniques. Gray et al. (2020) and Wright (2020), for example,
adopted PCA rather than EFA due to the nature of their data (e.g., shorter
texts, low frequency of features). A number of step-by-step guides detail
how to conduct MDAs (e.g., Egbert & Staples, 2019; Friginal & Hardy,
2014), but there are key methodological decisions researchers need to make
beyond the statistical analysis. Goulart and Wood (2021) summarize these
decisions as follows: (1) identify or compile a corpus of interest, (2) choose
the linguistic variables for analysis, (3) annotate the linguistic variables, (4)
check the accuracy of the annotation, (5) obtain normed frequencies of the
linguistic variables in each text in the corpus, (6) check the assumptions for
conducting the statistical analysis, (7) run the statistical analysis, (8) inter-
pret the factors extracted, (9) calculate dimension scores for each text, and
(10) calculate mean dimension scores for the genres investigated. In the
following paragraphs, we will focus on the methodological decisions that
do not involve statistical analysis, specifically on points (1), (2), (3), (4), (8),
and (9); for more details on how to conduct EFA, the reader is referred to
Egbert and Staples (2019) and Friginal and Hardy (2014).
Similar to many of the other approaches introduced in this volume,
MDA starts with identifying (or compiling) an appropriate corpus of study.
Once a corpus has been selected or compiled, researchers consider whether
an additive or a novel MDA is best suited for the study’s research goals.
An additive MDA is ​​a way to add a text variety (or varieties) to dimen-
sions extracted in previous studies. This is usually done with Biber’s (1988)
dimensions: (1) involved versus informational production, (2) narrative
discourse, (3) situation-dependent versus elaborated reference, (4) overt
expression of persuasion, and (5) impersonal style. Through this approach,
the researcher calculates mean dimension scores for a group of texts based
on the linguistic features on each side of Biber’s (1988) original dimen-
sions (see Berber-Sardinha et al., 2019 for a description of how to conduct
an additive MDA). Additive MDAs are less statistically demanding than
novel MDAs since the researcher does not have to run a new EFA. The
advantage of running an additive MDA is that it allows for comparisons
against a range of studies that have also applied the same dimensions to dif-
ferent genres. For example, Çandarli (2022) conducted an additive MDA
 Multidimensional analysis 133

of online academic forums written by undergraduate and master’s students.

Using the results of her additive MDA, the researcher was able to compare
the linguistic patterns of these forum posts to those of university-written
genres (e.g., Nesi & Gardner, 2012) and other genres investigated in Biber
(1988).
In contrast, a novel MDA requires the researcher to conduct a full EFA,
thus extracting dimensions based only on the genres represented in the
corpus of study. Wright (2020), for example, identifies the patterns of lan-
guage variation in stand-alone literature reviews. The researcher extracted
six dimensions of variation in this study, which are detailed below in the
Example Studies section. Her results allow for comparison of linguistic
characteristics in literature reviews across years of publication or discipline
within the same genre.
If a novel MDA is selected, researchers need to follow the steps of a full
EFA, starting from selecting linguistic features that are relevant to the gen-
res investigated and annotating the corpus for these features. Most MDAs
have used the Biber Tagger2 (Biber, 1988) to annotate the linguistic vari-
ables included in the analysis. In Goulart and Wood’s (2021) methodologi-
cal synthesis, the authors show that 50 out of 95 novel MDAs published to
date have adopted the Biber Tagger as their tagger of choice. According to
Gray (2019), the prevalence of this tagger in MDA analyses can be explained
by two factors. First, the tagger has three different algorithms and prob-
abilities that can be adjusted depending on the text varieties in the corpus
(exposition, fiction, and spoken). In addition, the Biber Tagger not only
identifies part-of-speech information, but it also adds semantic and syn-
tactic information, adding up to more than 250 different tags. Although
researchers conducting MDAs have favored the use of the Biber Tagger,
other lexico-grammatical taggers (e.g., CLAWS, MAT Tagger, TreeTagger)
can also be used in the process of automatic annotation of linguistic fea-
tures (see Gray, 2019 for a review of different taggers). Regardless of the
tagger (or previously reported accuracy), researchers should evaluate the
accuracy of the tags in a sample of texts from their corpus.
The following stage involves running the statistical analysis. At this
point, researchers need to consider a number of subjective statistical deci-
sions, such as the cut-off point for communalities, the rotation method,
or the most appropriate number of factors to be extracted (see Egbert &
Staples, 2019; Friginal & Hardy, 2014). When it comes to factor extrac-
tion, researchers need to consider both the eigenvalue for each factor, the
variance explained, and the interpretability of the factors. That is, the num-
ber of factors extracted is not simply determined by a statistical threshold.
The researcher also needs to take into consideration the possible interpre-
tation of the factors. In Novikov (2021), for example, the author found
134 Larissa Goulart, Shelley Staples

that extracting three factors resulted in more variance explained, but the
third factor was not readily interpretable. Therefore, the author decided on
a two-factor solution, which statistically explains less of the variation, but
resulted in two factors that can be interpreted in the context investigated
(learners of Russian as a second language).
Once factors are extracted, researchers then interpret the dimensions
based on the shared function of the linguistic variables that appear together.
The final step is to calculate dimension scores for each text in the corpus
and mean dimension scores for each genre (or group of texts) represented
in the corpus. With these mean dimension scores, the researcher can then
compare the language profile of different genres or conduct further statisti-
cal analyses comparing the mean dimension scores across genres.
Although MDAs can be more statistically complex than other approaches
to genre analysis, MDAs can add to these other approaches by identifying
which linguistic features are pervasive in a genre. In fact, several approaches
to genre analysis can be used in tandem with an MDA. Gray et al. (2020),
for example, conducted an MDA in a corpus of research articles segmented
into rhetorical moves (see Casal & Kessler, this volume), while Biber (2006)
included lexical bundles (see Cortes, this volume) as one of the linguistic
variables in his MDA of university language.

7.4 Example studies
Previous MDAs have investigated various discourse domains, describing
the linguistic characteristics of a number of spoken and written genres (see
Goulart & Wood, 2021). In this section, we highlight five studies that illus-
trate how MDAs have been used to investigate genre variation. Table 7.1
provides an overview of the contexts, research questions, and data collected
for these studies.
Our review starts with Hardy and Friginal (2016). In this study, the
authors conducted an MDA on a corpus of undergraduate and graduate
written assignments from the Michigan Corpus of Upper-Level Student
Papers (MICUSP). The goal of this study was to describe the linguistic
characteristics of assignments students need to write as part of their course-
work. The authors argue that written assignments are central to students’
success in their undergraduate and graduate courses. Nevertheless, little is
known about the linguistic characteristics of these assignments that can
help prepare students in EAP and writing classes. The dataset used in Hardy
and Friginal included 825 assignments from 16 disciplines. These assign-
ments were classified into genre categories following a systematic analysis
of all texts in the corpus (Römer & O’Donnell, 2011). This categorization
resulted in seven genres (argumentative essay, creative writing, critique,
TABLE 7.1 
Example studies adopting an MDA

Authors Context Research Questions (RQs) Data Collected Comments on Study

Design

Connor and • Grant proposals • What are the linguistic • 68 grant proposals • The authors report on
Upton (2004) • Most of the proposals and rhetorical features submitted to the the overall dimension
(60 of 68) came of promotion and Indiana Center scores of the grants
from the discipline persuasion in grant of Intercultural as well as the way
of health and human proposals written by non- Communication each move loads
services profit organizations? fundraising corpus along Biber’s (1988)
• What are the dimensions
distinctive patterns of • This study is interesting
language variation in in its analysis of the
grant proposals and differences between
independent moves of persuasion and
grant proposals? argumentation
Goulart (2021) • Undergraduate • To what extent is there • 379 student papers in • The author compares
assignments written linguistic variation the L2 corpus divided the linguistic
by L1 and L2 English across registers in L2 into 11 genres of characteristics of the
speakers in British university writing? university writing same register across
universities • To what extent is there • 395 student papers in language background
linguistic variation the L1 corpus divided
across registers in L1 into 13 genres of
university writing? university writing
• Do L1 and L2 students
use different linguistic
features when writing
the same register?
Multidimensional analysis 135

(Continued)
TABLE 7.1 Continued

Authors Context Research Questions (RQs) Data Collected Comments on Study

Design
Gray et al. (2020) • High-quality, empirical • What are the linguistic • 900 research articles • This study is
research articles that characteristics of divided in 16 possible innovative in its
follow the IMRD/C research articles moves? moves analyses of linguistic
structure in social characteristics across
sciences, engineering, different moves and
and natural and applied how it deals with
sciences shorter texts
Hardy and • Senior year • What are the • 825 student papers • The authors report the
136 Larissa Goulart, Shelley Staples

Friginal (2016) undergraduate and quantitative linguistic from MICUSP divided methodological decisions
graduate assignments patterns in the into seven genres in detail (e.g., features,
written for four university genres (argumentative essay, rotation)
disciplinary groups represented in the creative writing, • Four dimensions were
(humanities, social MICUSP corpus? critique, proposal, extracted
sciences, biological report, research paper,
and health sciences, and response paper)
and physical sciences)
Wright (2020) • Published stand-alone • Which dimensions • 417 stand-alone reviews • This study is unique in
literature reviews that define stand-alone from 11 journals across the genre analyzed
synthesize the results literature reviews the disciplines of • The results show
of research articles on and how do those medicine, education, the extent to which
a carefully delineated dimensions compare to and psychology linguistic characteristics
topic dimensions found in vary within the same
other MDAs? genre across disciplines
and methodological
approach
 Multidimensional analysis 137

proposal, report, research paper, and response paper). For the MDA, the
researchers started with frequency counts for 110 linguistic features, which
were automatically annotated using the Biber Tagger (Biber, 1988). In the
final factor solution, 54 of these 110 features were retained. After consid-
ering different factor solutions, the authors found that a four-factor solu-
tion was optimal. These four factors were functionally interpreted as (1)
involved, academic narrative vs. descriptive, informational discourse; (2)
expression of opinion and mental processes; (3) situation-dependent, non-
procedural evaluation vs. procedural discourse; (4) production of possibility
statement and argumentation.
The patterns of genre variation observed in these four dimensions
show that university genres are distributed along a continuum. That is,
the authors find that there is a continuum between more involved genres
(e.g., creative writing, response papers, critique, argumentative essay) and
informational genres (e.g., proposal, report, research papers) in the four
dimensions identified. The results of this study show that creative writing,
response papers, argumentative essays, and critiques loaded on the positive
side of all four dimensions, indicating that these four genres are character-
ized by linguistic features associated with expression of opinion, situation-
dependent discourse, and production of possibility.
In a similar study, Goulart (2021) examined the patterns of linguistic
co-occurrence in a corpus of undergraduate written assignments across first
(L1) and second language (L2) speakers of English. The author sought to
identify the extent to which there is linguistic variation across the same
genres when written by L1 and L2 students. The data included a corpus
of 379 L2 undergraduate papers and a corpus of 395 L1 undergraduate
papers. In order to conduct the MDA, the author used the Biber Tagger to
annotate the linguistic features. The author reported selecting features that
were used in previous MDAs of academic writing and university writing.
In the final factor analysis, 34 features were retained. After several rounds
of piloting, the author found that a five-factor solution was optimal for the
study. These five dimensions were interpreted as (1) expression of personal
opinion vs. compressed procedural information; (2) expression of possibil-
ity vs. account of completed actions; (3) informational density vs. engaging
presentation; (4) involved academic narrative vs. elaborate description; (5)
stance towards the work of others.
Overall, Goulart’s results show an opposition between genres that are
usually associated with showing an understanding of content knowledge
(e.g., essays, critiques) and genres that simulate professional practice (e.g.,
laboratory reports, case studies). Additionally, this study shows that most
genres have the same linguistic characteristics when written by both L1 and
L2 students, the main difference being in genres such as laboratory reports,
138 Larissa Goulart, Shelley Staples

case studies, and design specifications where L2 students tend to use more
concrete features (e.g., concrete nouns, quantity nouns, passive voice) than
L1 students.
Connor and Upton (2004) also investigate the linguistic characteristics
of academic English, but different from these previous studies, they exam-
ine the extent to which the different moves of grant proposals have distinct
linguistic characteristics. Connor and Upton conducted an additive MDA
using Biber’s (1988) dimensions. That is, the authors analyzed how the
different moves of grant proposals (territory, gap, goal, means, competence
claim, importance claim, and benefits) interacted with Biber’s dimensions.
As mentioned above, one of the main advantages of using an additive MDA
is that it allows for comparisons between the genre investigated (grant pro-
posals) and other genres described in previous studies that used the same
dimensions. The first step of the study was to identify the rhetorical moves
of grant proposals. The authors drew on moves identified in published aca-
demic writing and piloted several move schemes to identify seven distinct
moves of grant proposals. The authors then present the results of two addi-
tive MDAs. The first examines the linguistic characteristics of grant pro-
posals as a whole and the second takes an in-depth look into the linguistic
characteristics of each move. The results of the additive MDA for the whole
genre of grant proposals show that these texts are characterized by features
of informational discourse (e.g., nouns, nominalizations, adjectives), non-
narrative concerns (e.g., present tense), and explicitness (e.g., wh- relative
clause, phrasal coordination).
The results of the MDA for each move show that there is considerable
variation within the genre of grant proposals. The move of benefits, for
example, has considerably more features of explicit discourse than the gap
move. In contrast, the move of territory has considerably fewer features of
overt expression of argumentation (e.g., infinitive verb forms like to live, to
walk, etc.; suasive verbs like agree, allow, beg, etc.; modals of prediction like
would, will, shall, etc.) than all other moves in grants. The results of this
study are particularly interesting as they show (1) the extent to which the
linguistic characteristics of grant proposals differ from other genres, and
(2) the extent to which there is variation across rhetorical moves that occur
within the same genre.
Gray et al. (2020) also explore the linguistic characteristics of rhetori-
cal moves, examining the lexico-grammatical features that are pervasive in
research article moves. The primary goal of this research was to map the
communicative function of the research article moves onto the patterns
of linguistic variation identified in a new MDA. In order to achieve this
goal, a corpus of 900 research articles was segmented according to the
move/step framework described in Cotos et al. (2015). This framework
 Multidimensional analysis 139

consists of three possible moves for the introduction and methods sections
of research articles and four possible moves for the results and discussion
sections of a paper. The final corpus consists of 11,043 text segments of
different moves in the corpus. For the MDA, the authors annotated the
linguistic features using the Biber Tagger and additional customized scripts
that counted the frequency of linking adverbials, individual modal verbs,
and common nouns + preposition combinations. The authors discuss the
process of automatic annotation in-depth, including the accuracy of tags in
academic writing.
This study introduces important methodological considerations for
researchers conducting MDAs on shorter texts. Shorter texts have a lower
frequency for certain linguistic features (with many features having a zero
count in shorter text segments), which might affect the results of a tra-
ditional EFA. In order to circumvent this issue, the authors adapted the
MDA in several ways. First, they excluded move segments with less than
a hundred words. In addition, they automatically excluded linguistic fea-
tures that did not occur in more than 90% of the texts. Third, they prior-
itized more general features (e.g., all passives) over more specific features
(e.g., by-passives and agentless passives). Aside from these inclusion and
exclusion criteria, they also decided to run a PCA, rather than EFA. The
PCA resulted in four components, subsequently interpreted as dimensions,
which provide a fine-grained account of language variation in research arti-
cles, looking at how each move can have different linguistic characteristics
even within the same section of a paper (i.e., introduction, results, meth-
ods). This study also provides a detailed report of the methodological con-
siderations in applying an MDA to rhetorical moves, from how to segment
texts to how to deal with shorter text segments.
While Gray et al. described the linguistic characteristics of published
research articles, Wright (2020) used MDA to examine the linguistic
characteristics of another published academic genre: stand-alone litera-
ture reviews. Considering how ubiquitous literature reviews have become
in recent years, Wright (2020) argues that there is a need to account for
how language varies within this genre. With this in mind, the author con-
ducts a new MDA in a corpus of 417 texts divided into three disciplines
(education, psychology, and medicine), three periods in time (1950, 1980,
and 2010), and two research approaches (quantitative and qualitative).
Similar to the other studies summarized in this section, the author started
by automatically annotating the linguistic features with the Biber Tagger.
Wright selected the same 70 features as used in Gray’s (2015) MDA of
research articles. The author then followed Gray et al.’s (2020) approach
and conducted a PCA in order to account for the fact that several variables
included in Gray’s study had zero frequencies in many texts included in the
140 Larissa Goulart, Shelley Staples

corpus. In addition, different from most MDAs, the author uses a Varimax
rotation, which assumes that there is no correlation between variables, to
rotate the factors. Wright justifies this choice based on the fact that adopt-
ing a Promax rotation, which assumes that there is a correlation between
variables as is customary with lexico-grammatical features, would result in
extremely low loadings for 11 features (p. 311).
The final MDA resulted in a six-factor solution that accounted for 36.8%
of the variance. These dimensions were interpreted as (1) human vs. tech-
nical/academic focus, (2) questioning/interpreting vs. knowledge-confer-
ring, (3) expression of stance, (4) author/discourse community vs. topic
focus, (5) abstract vs. concrete focus, and (6) methodological concerns vs.
description. This study is particularly relevant as it shows that new MDAs
can be useful when describing the language of genre-restricted sets of data,
examining variation within the genre of stand-alone literature reviews. In
addition, the author discusses the methodological considerations in detail
(i.e., adopting PCA rather than EFA). This methodological discussion
highlights ways that researchers might need to adapt the original method
in order to obtain interpretable dimensions.

7.5 Issues and challenges

There are a number of issues that researchers might encounter when con-
ducting an MDA. The first issue to consider is whether an MDA is the
most appropriate approach to answer the research question proposed by
the researcher. While MDAs provide comprehensive accounts of the lin-
guistic characteristics of a specific genre, the linguistic description pro-
vided by an MDA focuses on a number of linguistic features at the same
time. Therefore, if the researcher is interested in the use of a single linguis-
tic structure in a genre or text variety, MDA is not the most appropriate
method for the study.
A more practical hurdle is whether the researcher has the statistical
knowledge or access to statistical guidance to conduct an MDA. Essentially,
the researcher should be familiar with EFA and/or PCA. If the researcher
does not wish to conduct a new MDA, then it is sufficient to be familiar
with factor analysis conceptually. A good conceptual overview of MDA can
be found in Biber and Conrad (2019, Chapter 9). There are several research
tools available that will allow researchers to use the dimensions from Biber
(1988) and thus conduct an additive MDA. First, although not publicly
available, the Biber Tagger (Biber, 1988) provides results for the first five
dimensions of Biber (1988). A publicly available option is Nini’s (2019)
Multidimensional Analysis Tagger (MAT), which provides output based
on the Stanford Parser rather than the Biber Tagger, but the variable set is
 Multidimensional analysis 141

aligned with that of Biber (1988). Finally, the Lancaster Stats Tool Online
(Brezina, 2018)3 uses Biber’s (1988) dimensions as an option in its MDA
tool. These tools allow a researcher to investigate the dimensions in genres
not covered in Biber’s (1988) study or to compare, for example, similar
genres to those represented in Biber (1988) from more recent time periods.
However, to add variables not included in Biber (1988), researchers need
to understand the statistical principles of factor analysis or gain access to
statisticians that can help them with the various decisions needed to suc-
cessfully conduct a new MDA. Two resources are suggested for gaining
this knowledge: a good multivariate statistics textbook (e.g., Tabachnick
& Fidell, 2021) and a reference text on MDA (e.g., Berber-Sardinha &
Pinto, 2019).
Even with knowledge of factor analysis, additional considerations include
access to and selection of appropriately large corpora (MDA ideally is only
conducted on samples > 300 cases or at a ratio of 5/1 cases per variable).
Following the selection of a corpus, researchers must identify appropriate
variables for the analysis, which are contingent on the genre under investi-
gation as well as the research questions. In addition, the selection of vari-
ables to omit based on conceptual or statistical overlap is as important as
the initial section of variables. Another challenge includes identifying the
variables, either through existing automated programs, development of
computer programs, and/or manual coding along with accuracy checking
of the variables, which may be identified initially through automated means
(e.g., a part-of-speech tagger), but need to be checked manually on at least
a portion of the data. If the researcher wants to combine MDA with other
types of analysis, such as move analysis, additional coding and accuracy
checking are needed (see Casal & Kessler, this volume). A final challenge is
an interpretation of the factors, which requires knowledge of the variables
as well as the genres included in the analysis.

7.6 Study-in-focus
In this section, we discuss a model study that uses MDA to explore the
linguistic characteristics of university writing genres. Specifically, we have
selected Gardner et al.’s (2019) MDA of the BAWE corpus. The BAWE
corpus is divided into 13 genre families (case studies, critiques, empathy
writings, essays, exercises, explanations, designs, methodology recounts,
narrative recounts, literature surveys, proposals, problem-questions, and
research reports). In this study, the authors conducted a novel MDA
accounting for language variation across genres, disciplines, and students’
year of study. The authors explain that this broad linguistic description
of university writing can help inform the development of EAP teaching
142 Larissa Goulart, Shelley Staples

materials and lesson plans. Before proceeding with our discussion of this
article, please read the call-out box, which provides a summary of the
study’s procedures and findings.

Citation
Gardner, S., Nesi, H., & Biber, D. (2019). Discipline, level, genre: Integrating
situational perspectives in a new MD analysis of university student writing.
Applied Linguistics, 40(4), 646–674.

Research questions
• To what extent do linguistic features relate to situational perspectives
on student academic writing?
• What are the linguistic characteristics of university writing across genre,
level of study, and discipline?

Context and population

This study examined the language profile of the 13 genres in the BAWE cor-
pus. BAWE contains texts written by undergraduate and master’s students
in four universities in the United Kingdom. Texts in this corpus represent
successful undergraduate and graduate writing, regardless of students’ L1.
These texts received a merit or a distinction by the instructors of the courses
that required these assignments. All 2,760 texts in the corpus were used in
this study. In addition, BAWE texts are divided into four disciplinary groups
(arts and humanities, social sciences, life sciences, and physical sciences).

Procedure
To analyze the linguistic characteristics of university writing, Gardner et
al. (2019) conducted an EFA, which involved several steps: (1) texts in the
BAWE corpus were annotated for lexico-grammatical features using the Biber
Tagger; (2) an EFA was run; (3) dimension scores were computed for each
text in the corpus; (4) mean dimension scores were computed for each text
category (discipline, genre, and level); and (5) dimensions were interpreted
based on their underlying functions.
To select linguistic variables to include in the EFA, the authors started
with all features tagged by the Biber Tagger and excluded variables that had
low communalities in the initial EFA runs. They also eliminated features with
 Multidimensional analysis 143

considerable overlap (e.g., all pronouns vs. 1st, 2nd, 3rd person pronouns),
maintaining more specific features whenever possible. Additionally, some
variables were combined to avoid redundancy. The EFA analysis retained 39
linguistic features distributed in four factors (or dimensions).

Findings
The four dimensions identified were (1) compressed procedural information
vs. stance toward the work of others, (2) personal stance, (3) possible events
vs. completed events, and (4) informational density. These four dimensions
accounted for 39.3% of the cumulative shared variance. For each dimension,
Gardner et al. also accounted for the patterns of variation across disciplinary
groups and level of study. In terms of discipline, this MDA reveals a marked
opposition between arts and humanities and physical sciences in the first
three dimensions.

The first dimension identified in this study, compressed procedural infor-

mation versus stance towards the work of others, contains on its positive
pole features associated with concise report of a procedure (e.g., premodi-
fying nouns, common nouns, passives, action verbs), and on its negative
pole features associated with elaboration (e.g., communication verbs, stance
adverbials, proper nouns, stance nouns + that clauses).4 This dimension
distinguishes between genres that report a procedure (e.g., methodology
recounts, designs, exercises) on the positive pole and genres that include
references to the work of others (e.g., essays, narrative recounts, problem-
question). The positive side of Dimension 2 is characterized by features usu-
ally used to express personal opinion (mental verbs, stance verbs + that and
to- clauses, that- deletion, first-person pronouns, and past tense). The genre
of narrative recount loads extremely high on this dimension, followed by
problem-question, and empathy writing. All other university genres load on
the negative side of this dimension (lack of personal stance). Dimension 3
distinguishes between features used to express possibility (e.g., if- condition-
als, present tense, modal verbs, and verb to be) in its positive side and features
used to recount completed events (e.g., past tense and perfect aspect) in its
negative side. Genres that account for descriptions of past events (method-
ology recounts, research reports, narrative recounts, essays, and literature
surveys) load on the negative side of this dimension, while all other gen-
res load on the positive side. Dimension 4 contains features usually associ-
ated with academic writing (e.g., longer words, nominalizations, attributive
144 Larissa Goulart, Shelley Staples

adjectives, abstract nouns) on its positive side. Proposals, literature surveys,

case studies, and critiques load on the positive side of this dimension.
Gardner et al. (2019) is considered an exemplary study not only because
of its elucidating results but also for its level of detail when reporting the
methodological decisions involved in conducting the MDA. In this study,
the authors provide an extensive account of their criteria for the inclusion
and exclusion of different linguistic features. For example, the authors
exclude features that contain consistent overlap, such as all modal verbs in
favor of fine-grained counts like possibility, necessity, and modals of pre-
diction. Unfortunately, many MDAs do not describe their feature selection
criteria, which affects the dimensions extracted in the MDA. In addition,
Gardner et al. also report on a number of decisions that researchers have
to make when running the statistical analysis, such as the type of rota-
tion (Promax), the factor loading cut-off (±.35), and the criteria for factor
selection (scree plot and interpretability). The fact that the authors report
on these methodological decisions and provide the factor loadings of all
linguistic features in the supplemental materials allows other researchers to
add new genres to this MDA, as has been done in Staples et al. (2018). That
is, the level of methodological detail allows other researchers to replicate
this MDA in other corpora.
Finally, the results of Gardner et al. also provide a rich discussion of the
results, analyzing variation across genres, level of study, and discipline. The
authors find that texts in arts and humanities are considerably different
from those written for physical sciences, across three out of the four dimen-
sions extracted. It is worth noting that this study is also a model in terms
of how to discuss the results of an MDA, using text excerpts. The authors
illustrate the linguistic characteristics obtained through the statistical anal-
ysis, with plenty of examples from different genres. In sum, Gardner et al.
is an exemplary study in two ways, first in providing a clear methodological
outline of the MDA presented in the study, and second, by exemplifying
the results of quantitative analysis with clear text excerpts.

7.7 Future research directions

As discussed in the introduction, MDA generally has not been used exten-
sively with genre frameworks, with the important exception of Gardner et
al. (2019). Thus, there is a need for more theoretical grounding of genre in
MDA studies. Gardner et al.’s work provides an example of how MDA can
be usefully combined with a genre-based approach through their qualita-
tive identification of textual categories using SFL principles such as stages,
followed by quantitative identification of linguistic forms which are then
interpreted based on the qualitatively determined categories.
 Multidimensional analysis 145

Researchers interested in functional categories often included in SFL

research, such as appraisal (how writers or speakers express approval or dis-
approval), may either want to expand the formal categories under analysis
in order to more comprehensively capture those functions or tag the cor-
pus manually for those functions using a more traditional SFL approach.
For example, the Biber Tagger includes a limited set of features associated
with stance, which is related to appraisal. However, the focus of the tag-
ger has been on linguistic features that are reliably associated with stance
(such as complement clauses or modal verbs), and thus, numerous lexico-
grammatical resources (not to mention prosodic resources) for express-
ing stance are not included. Researchers interested in a more fine-grained
approach to appraisal might wish to expand the lexico-grammatical items
included in the analysis by creating their own computational programs
(which should in turn be followed by analysis of whether appraisal is being
used in all cases of the identified lexico-grammatical items). They alterna-
tively could consider using a manual annotation tool, such as the UAM
Corpus Tool (http://www​.corpustool​.com/), which allows users to create
an annotation scheme. In addition, researchers from all genre traditions
can add to MDA by (1) combining MDA with other methods commonly
used in genre analysis, such as move analysis, or by (2) increasing the qual-
itative, contextual analysis that takes place in any MDA but many times is
minimally covered by researchers from a more quantitative tradition.
Second, although MDA researchers have identified a need to explore
variation within registers, less work has been done to explore variation
within genres using MDA. Goulart et al. (2022) show how five differ-
ent genre families (essays, critiques, case studies, methodology recounts,
and explanations) in BAWE vary in terms of communicative purpose: the
majority of texts classified as essays, critiques, and case studies combine
two or three communicative purposes. Methodology recounts and explana-
tions, on the other hand, show much less variation in communicative pur-
pose. Such qualitative characterization of individual texts combined with
a quantitative MDA approach can provide further insight into variation
within genres.
Finally, while this book is primarily focused on written genres, we would
like to argue for the expansion of a genre-based perspective on MDA to
spoken and multimodal genres. While numerous researchers have used
MDA to explore variation in spoken texts, these researchers have not taken
a genre-based perspective in their analysis. Interestingly, researchers from
the Sydney School of genre early on used oral narratives and conversation
as a way to apply their genre theories (Rose, 2012). It would be fruitful to
add MDA to further understand distinctions between story genres, both
spoken and written. In addition, many genre researchers have begun to
 146 Larissa Goulart, Shelley Staples

incorporate multimodal genres into their analyses, but these genres have so
far not been included in MDAs.

Notes
1 In recent publications, however, Biber and colleagues distinguish between reg-
ister and genres. In Biber and Conrad (2019) the authors state that “the register
perspective combines an analysis of linguistic characteristics that are common
in a text variety with analysis of the situation of use of the variety,” while a
genre perspective focuses on “conventional structures used to construct a com-
plete text within the variety.” (p.2)
2 One of the reasons why researchers decide to adopt other linguistic annotators
in their MDA is that the original Biber Tagger is not freely available and can
only be used upon request.
3 A new MDA can also be conducted with the Lancaster Stats Tool, but
researchers are still advised to be familiar with the steps of EFA/PCA before
doing so.
4 The reader can find examples of these linguistic features and all the others men-
tioned in this section in the original study by Gardner et al. (2019)

References
Berber-Sardinha, T. B., & Pinto, M. V. (Eds.). (2019). Multi-dimensional analysis:
Research methods and current issues. Bloomsbury.
Berber-Sardinha, T., Kauffman, C., & Acunzo, C. M. (2014). Dimensions of
register variation in Brazilian Portuguese. In T. Berber-Sardinha & M. Pinto
(Eds.), Multidimensional analysis, 25 years on: A tribute to Douglas Biber (pp.
35–79). John Benjamins.
Berber-Sardinha, T., Pinto, M. V., Mayer, C., Zuppardi, M. C., & Kauffmann,
C. H. (2019). Adding registers to a previous multi-dimensional analysis. In
T. Berber-Sardinha & M. Pinto (Eds.), Multi-dimensional analysis: Research
methods and current issues (pp. 165–186). Bloomsbury.
Biber, D. (1988). Variation across speech and writing. Cambridge University Press.
Biber, D. (2006). University language. John Benjamins.
Biber, D. (2019). Multi-dimensional analysis: A historical synopsis. In T. Berber-
Sardinha & M. Pinto (Eds.), Multi-dimensional analysis: Research methods and
current issues (pp. 11–26). Bloomsbury.
Biber, D., & Conrad, S. (2019). Register, genre, and style. Cambridge University
Press.
Brezina, V. (2018). Lancaster stats tools. http://corpora​.lancs​.ac​.uk​/stats​/toolbox​
.php
Çandarli, D. (2022). Linguistic characteristics of online academic forum posts
across subregisters, L1 backgrounds, and grades. Lingua, 267, 103190.
Connor, U., & Upton, T. (2004). The genre of grant proposals: A corpus linguistic
analysis. In U. Connor & T. Upton (Eds.), Discourse in the professions: Perspectives
from corpus linguistics (pp. 235–256). John Benjamins.
 Multidimensional analysis 147

Cotos, E., Huffman, S., & Link, S. (2015). Furthering and applying move/step
constructs: Technology-driven marshalling of Swalesian genre theory for EAP
pedagogy. Journal of English for Academic Purposes, 19, 52–72.
Egbert, J., & Mahlberg, M. (2020). Fiction–one register or two?: Speech and
narration in novels. Register Studies, 2(1), 72–101.
Egbert, J., & Staples, S. (2019). Doing multi-dimensional analysis in SPSS, SAS,
and R. In T. Berber-Sardinha & M. Pinto (Eds.), Multi-dimensional analysis:
Research methods and current issues (pp. 125–144). Bloomsbury.
Friginal, E., & Hardy, J. A. (2014). Conducting multi-dimensional analysis using
SPSS. In T. Berber-Sardinha & M. Pinto (Eds.), Multi-dimensional analysis, 25
years on: A tribute to Douglas Biber (pp. 297–316). John Benjamins.
Gardner, S., Nesi, H., & Biber, D. (2019). Discipline, level, genre: Integrating
situational perspectives in a new MD analysis of university student writing.
Applied Linguistics, 40(4), 646–674.
Goulart, L. (2021). Register variation in L1 and L2 student writing: A
multidimensional analysis. Register Studies, 3(1), 115–143.
Goulart, L., Biber, D., & Reppen, R. (2022). In this essay, I will…: Examining
variation of communicative purpose student written genres. Journal of English
for Academic Purposes, 59, 101159.
Goulart, L., & Wood, M. (2021). Methodological synthesis of research using multi-
dimensional analysis. Journal of Research Design and Statistics in Linguistics
and Communication Science, 6(2), 107–137.
Gray, B. (2015). Linguistic variation in research articles: When discipline tells only
part of the story. John Benjamins.
Gray, B. (2019). Tagging and counting linguistic features for multi-dimensional
analysis. In T. Berber Sardinha & M. Pinto (Eds.), Multi-dimensional analysis:
Research methods and current issues (pp. 43–66). Bloomsbury.
Gray, B., Cotos, E., & Smith, J. (2020). Combining rhetorical move analysis with
multi-dimensional analysis: Research writing across disciplines. In U. Römer,
V. Cortes, & E. Friginal (Eds.), Advances in corpus-based research on academic
writing: Effects of discipline, register, and writer expertise (pp. 138–168). John
Benjamins.
Halliday, M. A. K., & Matthiessen, C. M. (2004). An introduction to functional
grammar. Routledge.
Hardy, J. A., & Friginal, E. (2016). Genre variation in student writing: A multi-
dimensional analysis. Journal of English for Academic Purposes, 22, 119–131.
Hymes, D. (1974). Foundations in sociolinguistics: An ethnographic approach.
University of Pennsylvania Press.
Katinskaya, A., & Sharoff, S. (2015). Applying multi-dimensional analysis to
a Russian webcorpus: Searching for evidence of genres. Proceedings of the
5th Workshop on Balto-Slavic Natural Language Processing 2015 (pp. 65–4).
INCOMA Ltd. https://aclanthology​.org​/ W15​-5311​.pdf.
Loewen, S., & Gonulal, T. (2015). Exploratory factor analysis and principal
components analysis. In L. Plonsky (Ed.), Advancing quantitative methods in
second language research (pp. 182–212). Routledge.
Nesi, H., & Gardner, S. (2012). Genres across the disciplines: Student writing in
higher education. Cambridge University Press.
148 Larissa Goulart, Shelley Staples

Nesi, H., Gardner, S., Thompson, P., & Wickens, P. (2004–2007). The British
academic written English (BAWE) corpus. www​.coventry​.ac​.uk​/ bawe
Nini, A. (2019). The multi-dimensional analysis tagger. In T. Berber-Sardinha &
M. Pinto (Eds.), Multi-dimensional analysis: Research methods and current issues
(pp. 67–94). Bloomsbury.
Novikov, A. (2021). Syntactic and morphological complexity measures as markers of
L2 development in Russian [Unpublished doctoral dissertation]. University of
Arizona.
Quaglio, P. (2009). Television dialogue: The sitcom Friends versus natural
conversation. John Benjamins.
Römer, U., & O’Donnell, M. B. (2011). From student hard drive to web corpus
(part 1): The design, compilation and genre classification of the Michigan
Corpus of Upper-level Student Papers (MICUSP). Corpora, 6(2), 159–177.
Rose, D. (2012). Genre in the Sydney school. In J. P. Gee & M. Handford (Eds.),
The Routledge handbook of discourse analysis (pp. 209–225). Routledge.
Staples, S., Biber, D., & Reppen, R. (2018). Using corpus-based register analysis to
explore the authenticity of high-stakes language exams: A register comparison of
TOEFL iBT and disciplinary writing tasks. Modern Language Journal, 102(2),
310–332.
Swales, J. (1990). Genre analysis: English in academic and research settings.
Cambridge University Press.
Tabachnick, B. G., & Fidell, L. S. (2021). Using multivariate statistics (7th ed.).
Pearson.
Tardy, C. M. (2016). Beyond convention: Genre innovation in academic writing.
University of Michigan Press.
Tardy, C. M., Sommer-Farias, B., & Gevers, J. (2020). Teaching and researching
genre knowledge: Toward an enhanced theoretical framework. Written
Communication, 37(3), 287–321.
Wright, H. R. (2020). Stand-alone literature reviews. In U. Römer, V. Cortes, & E.
Friginal (Eds.), Advances in corpus-based research on academic writing: Effects of
discipline, register, and writer expertise (pp. 307–332). John Benjamins.
8
COMPLEXITY, ACCURACY, AND
FLUENCY (CALF) MEASURES
Charlene Polio, Hyung-Jo Yoon

8.1 Introduction to the approach and definition of genre

Decades ago, education researchers began using complexity measures to
study the writing of children and adolescents. This research was rooted in
Hunt’s (1965) work on developmental measures, which had as its goal to
identify “trends in the frequency of various grammatical structures” (p. 17).
With a small sample size of 18 students each in grades 4, 8, and 12, Hunt
noted students’ progress across a range of complexity measures, including
those that incorporated the now well-known T-unit (i.e., an independent
clause and its associated dependent clauses). Although Hunt’s (1965) study
was not primarily focused on genre (or what he referred to as mode in his
study), many first language (L1) researchers drew on his work to compare
writing across genres.
A few earlier studies of L1 writers included measures of accuracy and
fluency, but a much wider range of syntactic and lexical complexity, accu-
racy, and fluency (CALF) measures came into use when researchers started
applying them to the writing of multilingual writers (henceforth, L2 writ-
ers). CALF measures were brought to the forefront by Wolfe-Quintero
et al. (1998), who were not concerned with studying genre but rather with
understanding how CALF measures related to writing proficiency. They
detailed a variety of CALF measures along with how well they correlated
with writing test scores or class placements. There was no discussion of
how CALF measures might vary across genres or how genre might affect
measures of writing proficiency.

DOI: 10.4324/9781003300847-10
150 Charlene Polio, Hyung-Jo Yoon

Around this time, researchers began looking at how measures of CALF

varied across tasks in oral language. This early research was motivated by
Skehan (1998), who claimed that more complex tasks would result in a
trade-off among complexity, accuracy, and fluency, and by Robinson (2001),
who claimed that certain types of task complexity could increase speak-
ers’ attention to various language components and result in more complex,
accurate, or fluent performance. Definitions of task abound, but there is no
direct link between genre and task in studies of spoken language.
Research into oral language variation on tasks proliferated and, for bet-
ter or worse, writing scholars adopted this line of task complexity research
by studying the effects of task variation on language in written tasks. For
example, an early study by Kuiken and Vedder (2008) varied the complex-
ity of a letter-writing task by manipulating the number of elements that
needed to be considered. Tavokoli (2014) later examined picture narratives
and manipulated the complexity of the storyline. Neither of these stud-
ies addressed genre, and, in fact, we might say that the genres (i.e., let-
ter to a friend and picture narrative) were held constant. However, some
research on task differences did not hold genre constant and included genre
as a feature of task complexity as distinct from others that have been stud-
ied (e.g., number of elements, time for planning). Ruiz-Funes (2015), for
example, examined differences between what she called narrative/analytic
and expository/argumentative essays arguing that the narrative genre was
less complex than the argumentative. Frear and Bitchener’s (2015) study
also was motivated by the same claim, namely that the argumentative genre
was more complex because it involved a requirement of reasoning, a feature
of task complexity noted by Robinson (2001). Ultimately, however, Frear
and Bitchener noted that the task differences found in their study may not
have been due to the cognitive complexity of the task but to the functional
or communicative linguistic needs of the two genres. As they stated:

Other prospective causes for these findings include the pragmatic

requirements of the tasks…Ryshina-Pankova and Byrnes (2013) noted
that increases in complexity, operationalized as the occurrence of nomi-
nalization, were associated with increased mastery of the academic reg-
ister. It is possible, therefore, that the differences in findings between
different subordinate clauses in our study were also a by-product of the
task type in that particular dependent clauses were not frequently needed
as a means of fulfilling the pragmatic requirements of the task.
(p. 52)

Historically, studies of L2 writers focusing on CALF measures have viewed

genre as a feature of tasks that could be manipulated along with other task
 Complexity, accuracy, and fluency (CALF) measures 151

features such as planning time or topic familiarity. For example, an experi-

mental study might examine learner language across different prompts
that varied according to genre. Such a study might have implications for
theories of task complexity, such as those by Robinson (2001) and Skehan
(1998) mentioned earlier. However, Yoon and Polio (2017) examined the
differences in CALF measures in narrative and argumentative essays and,
like Frear and Bitchener (2015), claimed that the differences (in complexity
measures) were related to the language needed to communicate in the two
genres. They drew on the work of Biber and Conrad (2009) and claimed,
for example, that longer clauses and sentences (a measure of syntactic com-
plexity) were used in argumentative writing because it included more nomi-
nal post-modifiers (e.g., that- relative clauses and prepositional phrases).
In other words, differences in language across genres are not related to
task complexity but rather to the functional requirements of the genre.
Studies such as Yoon and Polio (2017) used a very broad definition of genre
that includes classifications such as narrative, argumentative, descriptive, or
expository. These classifications overlap with what might be called mode
(e.g., Caplan, 2019) or text type in the Sydney School/Systemic Functional
Linguistics. For example, de Oliveira and Smith (2019) included narratives
and informational writing as examples of text types, their broadest level
of classification. Within narratives, they included stories and recounts as
examples of genre.
While Yoon and Polio (2017) drew on the work of Biber and Conrad
(2009) to explain differences across genres, we note that Biber and Conrad
see genre as a macro structure, such as the methods section of a research
article, similar to what might be analyzed as in an English for Specific
Purposes (ESP) approach. Biber and Conrad refer to the analysis of linguis-
tic features as register analysis. The focus is on the communicative function
of the language and not on social conventions (see Goulart & Staples, this
volume, for a more complete discussion).
This chapter focuses on studies that use CALF measures, most of which
see genre in broad terms and do not consider the role of social conventions,
in part, because of their origins in language development, namely, how
writers’ language changes over time. Although the goals of research using
CALF measures vary, as detailed in the next section, this research has often
been framed in an assessment context where the genre of the elicited writ-
ing is specified but never defined. Nor is there a discussion of what genre as
a concept means. For example, Beers and Nagy (2009), who examined dif-
ferences in L1 expository versus narratives, despite having genre in the title
of their study, never defined the term. Similarly, Lu (2011), who studied L2
writers, and examined narrative and argumentative writing, was concerned
with developmental measures and did not define genre. It is possible, and
152 Charlene Polio, Hyung-Jo Yoon

even likely that writers aimed to produce the five-paragraph essay (i.e., an
essay with an introduction, body paragraphs, and conclusion, among other
features) in these contexts. Whether or not the five-paragraph essay consti-
tutes a genre has been debated (e.g., Tardy, 2019), but it is worth noting
that much of the research reviewed here includes timed essays reminiscent
of an assessment context.
Finally, as noted, unlike other ways to study genre variation, CALF
studies consider mostly, but not exclusively, developmental aspects of writ-
ing, whether researching children and adolescents or L2 writers because
accuracy and fluency are typically not relevant to the study of target or
published texts. As will be discussed throughout the chapter, non-learner
text studies focus instead on the syntactic and lexical complexity of genre
variations.

8.2 Goals
Studies that use CALF measures in the context of genre differences have
quite varied goals. In the studies of children and adolescents, even when
genre was included as a variable, the goals of the studies may not have been
genre-focused. For example, Hunt (1965) wanted to find the best devel-
opmental measures to describe student writing, yet he only tangentially
discussed how topic and genre affected language use with regard to various
aspects of complexity.
Some L1 studies conducted after Hunt (1965) had the explicit goal of
comparing language across genres. For example, Crowhurst and Piche
(1979) examined narrative, description, and argumentative pieces in a
cross-sectional study to determine if development of complexity measures
was the same across genres, called modes in their study. The study was
framed mostly as an assessment or research methods study in that there is
little mention of pedagogy or theory. As an interesting side note, Crowhurst
and Piche also examined the effect of varying the audience (i.e., teacher vs.
friend) on writers’ syntactic complexity. One might say then that they were
also examining how register and genre interacted, but they never men-
tioned register in the study.
Jumping ahead to a more recent study, Beers and Nagy (2011) also con-
ducted a mixed cross-sectional/longitudinal study of genre differences
across grade levels. This study was clearly framed as a genre study with
an investigation into syntactic complexity measures and one fluency meas-
ure. One goal was to understand if the writers could vary their language
among genres, as evidenced by the complexity measures, across grades. In
fact, the students generally could. This study was not framed as an assess-
ment study, but rather as an extension and clarification of previous work on
 Complexity, accuracy, and fluency (CALF) measures 153

developmental indices across levels. The authors hint at pedagogical impli-

cations but do not provide explicit recommendations.
With regard to L2 research, as discussed earlier, studies that regard genre
as a task feature tend to have as their goal describing the cognitive processes
involved in writing. For example, Ruiz-Funes’s (2015) experimental study
manipulated genre as a task complexity feature to compare the influence of
Robinson’s (2001) and Skehan’s (1998) models. She compared less complex
tasks (analytical and personal essays) with more complex tasks (argumen-
tative and expository essays) while keeping the topic constant. Similarly,
Abdi Tabari (2019) used CALF measures to compare three types of writ-
ing (personal, narrative, and decision-making), also to test so-called task
complexity features.
Other studies seek to describe genre differences for methodological or
assessment reasons. In other words, if learner language varies with regard to
CALF measures across genres, writing assessments for research or placement
and gate-keeping purposes need to consider genre as a variable. Yoon and
Polio’s (2017) semester-long longitudinal study of changes in CALF meas-
ures across two genres was motivated by methodological issues, namely con-
cerns about developmental studies that did not control for topic or genre. Lu
(2011), in his corpus-based study of Chinese learners of L2 English, com-
pared syntactic complexity measures across genres because of the concern
that varying genres added additional errors to assessments of writing. In
other words, genre can be simply one feature of writing test specifications.
Finally, a smaller set of research has a very different set of goals, namely
to describe target or published texts. These studies sometimes use meas-
ures of syntactic complexity to describe different parts of a genre (e.g., Lu
et al., 2020) or a specific genre (Tankó, 2017). Both Lu et al. and Tankó
combined CALF measures with a rhetorical move analysis (see Casal &
Kessler, this volume), and to some extent, both had the goal of informing
pedagogy. Lu et al. also mentioned how describing a genre via complexity
and move analyses could inform automated scoring systems.
In sum, we can identify a range of goals across studies of L1 writers,
L2 writers, and published texts that use CALF measures. Studies have
addressed (or could address in the future) the following research ques-
tions that have the potential to inform teaching, assessment, and models
of writing:

• To what extent does writers’ language differ across genres in terms

of complexity, accuracy, and fluency? Is there any interaction with
development?
• Should assessments of writing take genre into consideration in prompt
specifications?
154 Charlene Polio, Hyung-Jo Yoon

• Should teachers focus on syntactic and lexical complexity when teaching

different genres, or will students use appropriate language because of the
communicative functions required by the different genres? What is the
cause of any differences?
• How do published genres, target genres, or parts of genres differ with
regard to complexity? What are the implications for teaching students to
write these genres?

8.3 Common research methods

Research concerned with genre and CALF measures is either experimental
or corpus-based. In experimental studies, researchers collect data specifi-
cally manipulating the genre or the conditions under which a genre is writ-
ten (e.g., Yoon & Polio, 2017, as shown in Table 8.1). Other studies rely
on existing corpora (e.g., Alexpoulou et al., 2017, as shown in Table 8.1)
and then identify genres, or other prompt features, in the dataset. In both
types of research, often (but not always), participants write within a time
limit and without access to outside sources. While this type of data col-
lection does not reflect real-world writing conditions, it allows researchers
to control the features of the prompt, including genre, that they want to
study. In either case, the population should be clearly described with regard
to background and proficiency level, but writing samples from learner cor-
pora may or may not include detailed information about the participants.
Furthermore, studies of development across genres may be longitudinal
or cross-sectional, and these designs correspond somewhat to experimen-
tal and corpus-based research, respectively, because many of the available
learner corpora are not longitudinal. Studies of target texts may rely on
existing corpora or researchers may construct their own, depending on
their area of inquiry.
Once the data are collected or the corpus is identified, researchers choose
CALF measures to assess the writing. Syntactic and lexical complexity have
been studied extensively, and each has been shown to be multidimensional.
For example, syntactic complexity can be assessed at the sentence, clause, or
noun phrase level. Sentence complexity may be measured by the number of
dependent clauses, and clause and noun phrase complexity by the number
of words in the constituent, among other ways. Lexical complexity consists
of several constructs including diversity, the number of different words, and
sophistication. In the next section, we provide sources for choosing among
the many measures, but we note that the studies reviewed often include
several measures to capture the multidimensionality of the construct. Both
syntactic and lexical complexity are studied using automated tools such
as those found at https://www​.lin​guis​t ica​naly​sistools​.org/. Accuracy and
TABLE 8.1 
Example studies adopting CALF measures

Authors Context Research Questions (RQs) Data Collected Comments on Study Design

Alexopoulou • The study • How does task complexity, • Essays from EFCAMDAT, • Referred to genre as a
et al. used essays task type, and/or an open-access corpus task type
(2017) submitted to instructional focus impact • 1,180,543 pieces of • Measures included
Englishtown, the complexity and accuracy writing with 0–40 words error frequency of
an online of the language use in global at lower levels to 150–180 two features, three
school as well as specific features or words at higher levels complexity measure, and
structures? • Writing came from one measure of lexical
174,771 learners with diversity
56.5% having written at
least three scripts
Lee (2021) • Students at • Do college-level Korean EFL • 481 untimed descriptive • Used 17 measures of
two levels of learners exhibit syntactic and argumentative essays syntactic and lexical
proficiency complexity and lexical of at least 100 words complexity
studying complexity differences across having an average of
English in two genres (argumentative 130 words
Korea and who and descriptive)? • Essays came from the
had Korean as • Does L2 proficiency level Ganchon Learner Corpus,
their L1 affect the measures of which includes 16,113
syntactic complexity and essays produced by 1,607
lexical complexity? undergraduate participants
(Continued)
Complexity, accuracy, and fluency (CALF) measures 155
TABLE 8.1 Continued

Authors Context Research Questions (RQs) Data Collected Comments on Study Design
Lu et al. • Published • Are there differences in • The introductions and • Sentence was the unit
(2020) research articles the syntactic complexity literature reviews came of analysis and each was
from six social of sentences that realize from the Corpus of Social coded for rhetorical
sciences different rhetorical functions Sciences Research function
in social science research • They used 600 published • Expanded the CARS
article introductions? If yes, research articles published model to account for
what are the differences? between 2012 and disciplinary variation
• What rhetorical functions 2016 (100 each in from • Used five measures of
are realized through anthropology, applied syntactic complexity
156 Charlene Polio, Hyung-Jo Yoon

the most syntactically linguistics, economics,

complex sentences in social political science,
science research article psychology, and sociology)
introductions?
Tankó (2017) • Abstracts • What rhetorical moves occur • 135 abstracts from five • Focus was on rhetorical
from literature in literature research article high-impact journals in moves but also examined
research articles abstracts? literature sentence complexity
• What are the characteristic with the abstracts and
features of the moves? in comparison to other
• What are the typical lexico- fields
grammatical features of the • Examined lexical
moves in literature research density, lexical
article abstracts? sophistication, and
• What characterizes the lexical variation across
syntactic complexity and moves
lexical richness of the
abstracts?
Xu and Li • Four genres • How does the syntactic • A corpus of translated • 14 measures of
(2021) (fiction, news, complexity of translated English (from 60 source complexity
general prose, English texts compare with languages) and a corpus • They compared the
academic prose) non-translational English of non-translated English; measures across
of texts written ones in quantitative terms? both included four genres translated and non-
in English and • What are the connections and 15 subgenres translated texts for
texts translated between various forms • The corpora were balanced each genre, but did not
into English and degrees of syntactic so that they were the same compare genres
complexity measures and size and included the same
explication across genres? sampling of subgenres and
publication dates
Yoon and • Students in • How do measures of • Argumentative and • Used 12 measures of
Polio ESL classes at a complexity, accuracy, lexical narrative essays from syntactic complexity,
(2017) US university complexity, and fluency 37 students written at six three of lexical
develop over the course of points in the semester complexity, one of
one semester in ESL learners’ fluency, and four of
writing? accuracy
• Is there any effect of
genre on the measures of
complexity, accuracy, lexical
complexity, and fluency?
• Is there an interaction effect
between time and genre on
any of the measures?
Complexity, accuracy, and fluency (CALF) measures 157
158 Charlene Polio, Hyung-Jo Yoon

fluency are less commonly assessed. As discussed later, accuracy is challeng-

ing and time-consuming to assess and, similar to fluency, is less applicable
to the study of L1 writers and target texts.
Studies that use CALF measures to study genre are comparative in some
way because CALF measures are relative. There is no set standard for what
makes a learner or target text more complex. Thus, the studies described
in the next section compare different genres, sections of a genre, the same
genre across disciplines, and translated versus non-translated genres.
CALF measures may be combined with other approaches; two studies
of target texts discussed in the next section combine CALF measures with
rhetorical move analysis (see Casal & Kessler, this volume) but in different
ways. Lu et al. (2020) first identified rhetorical moves and then calculated
the complexity of the different moves. Tankó (2017), on the other hand,
described a genre with respect to moves and then with respect to syntac-
tic and lexical complexity (as compared to other genres), but she did not
attempt to integrate the two analyses.
Multidimensional analysis (see Goulart & Staples, this volume) draws on
linguistic features that may be included in some complexity measures. For
example, a feature of the literate versus oral dimension can include finite
adverbial clauses or that clauses controlled by a likelihood verb (Biber et al.,
2016), both of which would be figured into dependent clause/total clauses
ratio. Yoon and Polio (2017) argued that such specific features could be
used to explain the differences in complexity measures across genres, but in
general, multidimensional analysis may be considered a separate approach
to analyzing texts for genre or register differences. Biber et al. (2016) did
not combine approaches but compared them. They examined writing from
the TOEFL independent and integrated writing tasks (as well as speak-
ing tasks). They found differences across mode and task on the complex-
ity measures, but they also found that a multidimensional analysis, which
analyzes groups of features, better predicted scores, and task differences.
Overall, when using CALF measures, the biggest decision researchers
have to make is which measures to use. We cannot describe all of the meas-
ures for each of the constructs in this section, but we hope that the example
studies in the next section will illustrate some of the myriad of measures
available.

8.4 Example studies
Although studies using CALF measures generally focus on L2 writers, we
have included studies of both language learners and target texts here to
show a wider range of research with the qualification that the target-text
studies do not analyze accuracy or fluency. These studies are summarized
 Complexity, accuracy, and fluency (CALF) measures 159

in Table 8.1. We begin with three examples of research on learner texts that
examined genre as one among other variables. The first two use a rhetori-
cal mode definition of genre, although it is important to note that mode
is defined quite differently by other researchers who investigate aspects of
genre (see D’Angelo & Marino, this volume). The third study uses more of
a task-based definition.
Yoon and Polio (2017) conducted a study of ESL students’ linguistic
development in two genres. Their study was not motivated by issues of
genre but rather by research methods. They were concerned that prior
studies of language development did not control for prompt variables, one
of which was genre. They collected timed essays from 37 US university
students six times during a semester, with topics in each genre (i.e., argu-
mentative and narrative) counterbalanced. Students were enrolled in an
ESL class, but the authors did not consider or observe instruction. Yoon
and Polio assessed the essays on 12 measures of complexity, four of accu-
racy, three of lexical complexity, and one of fluency. They obtained syntac-
tic complexity measures from the L2 Syntactic Complexity Analyzer (Lu,
2010), as the aim of the study was to conceptually replicate Lu (2011), in
which the two genres were analyzed using the same tool. The two dimen-
sions of lexical complexity targeted in Yoon and Polio were lexical diversity
and lexical sophistication. Specifically, one lexical diversity (vocd-D) and
two lexical sophistication (mean word length and mean frequency for all
words) measures that had been studied extensively in previous studies on
lexical development were computed using Coh-Metrix (McNamara et al.,
2014). Accuracy measures that counted the number of specific error types
(e.g., syntactic errors per 100 words or morphological errors per 100 words)
were hand-coded.
Yoon and Polio (2017) found that genre had a greater effect on language
than did development (i.e., the time at which the essay was written). With
regard to genre, they found that 8 of the 12 complexity measures were
greater for the argumentative essays. Lexical sophistication measures were
higher for the argumentative essays, but lexical diversity was higher for the
narratives (likely because of the repetition of prompt words in the argumen-
tative essays). Genre had no effect on accuracy or fluency, which were exam-
ined to try to understand if either genre was more cognitively complex.
Only fluency, spelling, and one complexity measure improved over time,
and the effects were weak. The authors argued that the reason for the genre
differences was the different communicative needs of the two genres and
not the cognitive complexity of the genres. They showed how the various
complexity features aligned with Biber and Conrad’s (2009) description of
register differences. For example, argumentative academic writing contains
more nominalizations and fewer personal pronouns, thus affecting word
160 Charlene Polio, Hyung-Jo Yoon

length, one of the lexical sophistication measures. Biber and Conrad also
noted that narratives are likely to contain more temporal adverbial clauses
and argumentative writing, more conditional adverbial clauses. Thus, it was
likely that these different types of subordination canceled each other out in
Yoon and Polio’s study.
Lee (2021) conducted a conceptual replication of Yoon and Polio (2017)
using the Ganchon Learner Corpus, version 2.1 (Carlstrom & Price, 2012).
His study is an illustration of how corpus studies can complement experi-
mental studies. Lee collected descriptive and argumentative essays from
students at two levels of proficiency, as determined by self-reported TOEIC
scores. Because he used an available corpus, he was able to analyze a some-
what large set of essays (N = 481), but the corpus did not include the two
different genres from each participant. His study differed from Yoon and
Polio’s (2017) in that the essays were untimed and in that he examined only
syntactic and lexical complexity (the 14 syntactic complexity measures from
the L2 Syntactic Complexity Analyzer, and the same three lexical complex-
ity measures from Coh-Metrix). In addition, the essays were shorter.
Overall, despite differences in population and essay length, the results
were strikingly similar to, but not as strong as Yoon and Polio’s, with argu-
mentative writing showing greater complexity in many syntactic measures
and in lexical sophistication. The reasons for the lack of identical results
could be due to a number of factors, but Lee showed how a corpus can be
used to conceptually replicate experimental studies that can control vari-
ables and collect (by design) longitudinal or repeated measures data.
Alexopoulou et al. (2017) also conducted a corpus-based study of
prompt variables, one of which was genre. They analyzed writing from
the EF-Cambridge Open Language Database (EFCAMDAT) containing
1,180,543 texts. The pieces were short with 20–40 words at the lower
levels to 150–180 words at the higher levels. They were collected from
174,771 learners, with 56.5% of them contributing at least three pieces.
Because 66.7% included error labels, Alexpoulou et al. were able to assess
accuracy as well. Prompts were classified by task type corresponding to the
genres of narration (writing a movie plot) and description (describing a bad
meal). In addition, they included professional tasks (writing a resume and
writing a job advertisement), which is a departure from most studies com-
paring learner language across genres. Following the suggestion of Norris
and Ortega (2009) that L2 researchers need to measure multidimensional
CALF constructs, Alexpoulou et al. used some global (e.g., mean sen-
tence length and mean clause length) and fine-grained (e.g., number of
wh-phrases per sentence and number of past tense verb forms) syntactic
complexity measures. They used one lexical diversity measure (Measure of
Textual Lexical Diversity, MTLD) for lexical complexity. The syntactic and
 Complexity, accuracy, and fluency (CALF) measures 161

lexical measures were calculated using the freely available code, but the
validity and reliability of the individual measures were not tested. Last, for
accuracy, based on the error labels in the corpus, they calculated relative
error frequency and correct use of prepositions and past tense verbs.
Their findings were complex, but they were able to compare their
results to Yoon and Polio (2017) confirming that what they called task-
type affected complexity. They also noted, “Error rate generally did not
discriminate between tasks (again mirroring Yoon & Polio, 2017[7]),
but accuracy of specific features did—although not in a predictable way”
(p. 203). This study, like Lee’s (2021) complemented Yoon and Polio’s
smaller longitudinal study. Because of the large sample size, Alexopoulou
et al. could use a range of measures, including more accuracy measures.
However, they had to use the prompts that were included in the corpus,
which may not have corresponded exactly to other studies comparing
genres.
Turning to studies of target texts, Lu et al. (2020) analyzed a corpus of
600 published articles from the Corpus of Social Sciences Research Article
Introductions. The authors drew on the introduction and literature reviews
of 100 articles from each of six social science journals. Five “non-niche”
(p. 4) journals were selected based on impact factor and consultation with
experts in their respective fields. The authors wanted to determine if syntac-
tically more complex language was associated with specific rhetorical moves
of the Create a Research Space (CARS) model (Swales, 1990). Thus, their
view of genre is firmly rooted in the ESP approach. And like most studies
using the CARS model, their approach was to inform pedagogy by show-
ing how rhetorical functions are realized linguistically. Furthermore, they
note that previous studies of rhetorical functions have been linked mostly
to lexical bundles (see Cortes, this volume) and not syntax.
Based on a modified version of the CARS model, Lu et al. (2020) first
coded the introductions by noting the rhetorical move/step for each sen-
tence. Also, acknowledging the need for a multidimensional exploration of
syntactic complexity (Norris & Ortega, 2009), the authors adapted the code
in the Syntactic Complexity Analyzer and calculated five measures of syn-
tactic complexity for each sentence: sentence length, nominalizations, finite
dependent clauses, nonfinite dependent clauses, and left-embeddedness.
This approach to syntactic complexity is unique in that it uses the sentence
as its unit of analysis and in that it focuses on the function of each sentence
as opposed to an entire section of a text. Lu et al. found that all five meas-
ures showed between-step differences to some extent. For example, using
significance testing, they found that sentences with the functions question
raising, and outlining the structure of the paper were shorter. Sentences
with the steps real-world contextualization and question raising contained
162 Charlene Polio, Hyung-Jo Yoon

fewer nominalizations whereas those stating the value of the research con-
tained more. They also conducted a second analysis by charting the rhetori-
cal functions of the most complex sentences for each of the five measures.
For example, they found four steps associated with longer sentences. One
step, outlining the structure of the paper, used shorter sentences and “was
often realized through a series of short and direct sentences that connected
a label with its contents” (p. 7).
Tankó (2017) is another example of a study that combined rhetorical
moves with an analysis of lexical and syntactic complexity. The focus of
his study is the moves found in the abstracts of journal articles in the field
of literature, thus following an ESP approach to genre. Tankó collected
135 abstracts and investigated the prevalence of the eight moves that he
identified, such as purpose, outcome, and methods, finding that some
moves, such as conclusion and implications, were relatively infrequent. For
software-driven analyses, the abstracts were lemmatized and POS-tagged
(part-of-speech), and then they were analyzed for linguistic complexity
using Lu’s syntactic and lexical analysis tools (2010, 2012). The specific
subconstructs targeted in this study include production unit length, lexical
sophistication, lexical variation, and lexical density.
With regard to complexity, Tankó found that the sentences were on aver-
age longer than in linguistics and natural sciences, based on cited studies,
and that the longest clauses were found in the moves describing the meth-
ods. As for lexical measures, he compared the lexical density, sophistica-
tion, and diversity to highly rated student writing (Tankó, 2016). Although
a comparison to other genres by expert writers would have been more
appropriate, he concluded that the abstracts were more lexically dense and
sophisticated, but not more lexically diverse.
The final example study, Xu and Li (2021), compared English texts to
texts translated into English across four genres (news, fiction, non-fiction
prose, and academic works). Except for academic writing, the genres were
further subdivided (e.g., fiction included six subgenres). This study of non-
learner texts is situated in translation theory, and as a result, the discussion
of genre is minimal, but we included this study because it examined the
influence of genre and translation together and because it used syntactic
complexity measures common in L2 research to examine non-learner lan-
guage. With regard to genre, the authors appeared to be using a definition
of genre that overlaps with Biber and Conrad’s (2009) analysis of regis-
ters. Indeed, they state in an endnote, “Varieties of language are termed
differently as registers, genres, text types, domains, and styles, which are
delineated in fine detail by some discourse analysts (Leech, 2001; Biber and
Conrad, 2009)” (p. 229), further illustrating the variety with which many
of the studies in this chapter view genre.
 Complexity, accuracy, and fluency (CALF) measures 163

According to Xu and Li (2021), prior studies of complexity in translated

works focused mostly on single structures with the perspective that trans-
lated sentences need to unpack dense sentences and that translation favored
explicit encodings. For example, one study found that the complementizer
that was more common following reporting verbs (e.g., say) in translated
works. With regard to genre, the authors did not speculate on how genre
and translation interact, but they did present some findings with regard to
genre and translation. They modified the code built for the L2 Syntactic
Complexity Analyzer (Lu, 2010) and calculated the 14 syntactic complex-
ity measures that were arguably based on more accurate syntactic parsing.
Overall, as expected, the translations were more complex on 13 of the
14 measures. All 14 of the measures were different in the fiction transla-
tions and non-translations, and 13 were different in news texts. Yet only
three measures differed across academic prose and only one across non-
fiction prose. The reasons given for the various differences across genres
are speculative, including that academic writing is influenced by the fact
that most writers of translated works also write in English, minimizing the
differences. In addition, it seems that if translators across all genres need
to unpack dense sentences, complex nominals would be less common in
translations. In fact, complex nominals were the one measure that did not
show a difference across translated and non-translated texts. The authors
suggest that studies such as theirs can raise awareness of complexity in the
training of translators.

8.5 Issues and challenges

Investigations into genre variation using CALF measures generally com-
pare writers’ language on different prompts that are either manipulated by
the researcher or that are taken from a corpus. In such studies, as with any
type of experimental study, there are two main challenges: measuring the
outcome variables and ensuring that the study is internally valid.
With regard to the outcome variables (i.e., complexity, accuracy, and
fluency), choosing the most reliable and valid measure for each of these is
not a straightforward decision. There has been much written about each
of these constructs, so we will only briefly discuss them here and instead
refer readers to key works. Fluency, perhaps the least complex, has generally
been defined as the number of words written in a given time. Abdel Latif
(2013), however, noted that such a measure did not account for differences
between writers who wrote consistently at a slow pace and those who wrote
quickly between pauses. Thus, van Waes and Leijten (2015) examined the
construct of fluency using Inputlog, free key-stroke logging software, and
showed that fluency was actually a multidimensional construct, reinforcing
164 Charlene Polio, Hyung-Jo Yoon

the notion that words per minute was only one part of fluency. Using
Inputlog to collect data can be time-consuming and the data can be chal-
lenging to interpret. A larger question to consider is what the implications
are of finding writers exhibit different fluency behaviors across genres. Yoon
and Polio (2017) found that in terms of words per minute, genre had no
effect on fluency, whereas Lee (2018) did find an effect using a wider range
of measures and a different population. Assuming that there is a difference
in how learners write different genres, another challenge is determining the
theoretical and pedagogical implications of this finding.
Genre differences in accuracy and complexity likely have clearer implica-
tions for assessment and research contexts, namely, because writers’ lan-
guage will vary across genres, at least in terms of complexity measures,
genres should be kept consistent. Yet choosing the most valid and reliable
method is not straightforward. Polio and Shea (2014) examined several
measures of accuracy and determined that no one measure was more valid,
but that they did differ in terms of reliability. While there are many auto-
mated measures for complexity (as described in the previous section), there
are no viable option for accuracy (but see Crossley et al., 2019, for an exam-
ple of a tool that has not been rigorously tested on a range of error types).
Rather, accuracy has to be hand-coded and must involve multiple raters for
inter-rater reliability purposes because there can be multiple interpretations
of error presence and error type. Furthermore, no measure of accuracy has
been validated as being the most appropriate measure (see Polio & Shea,
2014), and it is likely that accuracy, too, is multidimensional.
Both syntactic and lexical complexity have been studied extensively. For
syntactic complexity, Polio and Yoon (2018) found that automated meas-
ures have been found to be reliable (in comparison to human coding) and
that the two most common automated programs Syntactic Complexity
Analyzer and Coh-Metrix yielded similar results when comparing gen-
res, but that the Syntactic Complexity Analyzer was more transparent in
terms of how measures were calculated. Choosing the most valid meas-
ures of complexity, a multidimensional construct, is challenging. A recent
discussion by Kyle et al. (2021) suggests that in addition to traditional
phrasal and clausal measures, usage-based measures related to verb-argu-
ment constructions should be used as a better indicator of proficiency
(but they did not examine genre variation). Another issue, particularly for
measuring syntactic complexity, is that researchers may need to select dis-
tinct measures, given the possible redundancy of including multiple co-
linear measures in one study (e.g., clauses per T-unit, dependent clauses
per clause, and dependent clauses per T-unit). Or they may use statistical
techniques such as principal component analysis to establish composite
scores, which represent target constructs, instead of individual measures.
 Complexity, accuracy, and fluency (CALF) measures 165

Lexical complexity has been conceived of as a variety of more specific

constructs, most commonly, lexical sophistication and lexical diversity.
Both can be measured using a range of automated tools, as detailed by
Kyle (2022).
Finally, most of the studies assessing CALF measures study English, and
those researching other languages do not have information on the reliabil-
ity and validity of the same variety of automated and manual measures, (but
there are some exceptions such as Hu et al., 2022 for Chinese). As we note
in the final section of this chapter, CALF genre-based research is, as far as
we know, exclusively English-based, and finding valid and reliable measures
for other language continues to be a limiting factor.
At the beginning of the chapter, we noted that genre and task at times
have been conflated, but an additional issue is that among these studies,
most have employed timed, five-paragraph type tasks that elicit short essays
one might find in assessment contexts. While the use of timed tasks is a
proper methodological decision that enables greater control, we note that it
might be more pedagogically meaningful to use writing tasks that allow for
more extended and contextualized discourse. Finding the balance between
experimental research and data collection that reflects real-world condi-
tions is a challenge in all educational research, but perhaps more so with
genre research because if we want to move beyond the five-paragraph essay,
we want to approximate real-word writing conditions. As discussed earlier,
combining experimental research with corpus data collected in more realis-
tic educational settings could help solve this problem.

8.6 Study-in-focus
In this section, we discuss Yoon (2021), a study that employs CALF meas-
ures for genre comparisons. Previous genre research has used narrative
and non-narrative tasks (e.g., expository or argumentative) in exploring
genre-specific language use, and the findings of these studies have gen-
erally indicated that writers use more syntactically complex structures for
non-narrative than narrative tasks. As noted earlier, based on a hypothesis
associating language structures and cognitive demands, these findings have
been interpreted as evidence of different cognitive demands imposed by
different genres. Despite a considerable amount of cross-genre research on
linguistic complexity, however, little research has been conducted on how
L2 learners actually perceive writing genres and how genre interacts with
other task factors that are known to affect cognitive task complexity. Given
these, Yoon’s study attempted to clarify the cognitive demands of distinct
genres and the meaning of genre effects on language, leading to a clearer
explanation of the motivation for linguistic changes in writing. The callout
166 Charlene Polio, Hyung-Jo Yoon

box below provides a summary of his study. We encourage readers to review

it first before proceeding with our commentary.

Citation
Yoon, H. (2021). Challenging the connection between task perceptions and
language use in L2 writing: Genre, cognitive task complexity, and linguistic
complexity. Journal of Second Language Writing, 54, 100857.

Research questions
• How do the manipulations of genre and idea support influence L2
learners’ perceptions of writing tasks?
• How do the manipulations of genre and idea support influence meas-
ures of linguistic complexity in L2 learners’ writing?

Context and population

This study examined 304 essays written by 76 high-intermediate ESL stu-
dents at a US university. At the time of data collection, they were enrolled
in L2 writing courses, receiving 4–6 hours of instruction per week. Fifty
participants were L1 speakers of Mandarin Chinese, and seven were Arabic
speakers. The remaining participants had different L1s. Their mean length of
English study was approximately nine years, and their mean length of stay
in the US was 15 months.

Procedure
Four timed writing prompts (two narrative and two argumentative) were
designed, with the level of conceptual demands manipulated within each
genre (idea support). Yoon collected data at one-week intervals, with the
order of the tasks counterbalanced. Participants composed timed essays (30
minutes) without referring to any external resources. A self-rating question-
naire was given to the participants immediately after they completed each
writing task. After cleaning the essay data (e.g., removing spelling errors),
Yoon employed three natural language processing tools (i.e., Syntactic
Complexity Analyzer, Coh-Metrix, and the Multidimensional Analysis
Tagger) to explore several constructs of linguistic complexity. The study
included 12 measures, which tapped into length of production unit, sub-
ordination, phrasal complexity, and lexical complexity. While tapping into
multidimensional complexity aspects, these measures were selected based
 Complexity, accuracy, and fluency (CALF) measures 167

on the criteria of validity, construct distinctiveness, and redundancy. To

examine the effects of genre and idea support manipulations on perceptions
and linguistic complexity, multilevel analyses were computed with genre
(argumentative and narrative) and idea support (no support and support)
as fixed and subject as random effects.

Findings
Yoon found that the manipulation of idea support resulted in significant
changes in the students’ perceptions of task complexity and difficulty (i.e.,
writing prompts with supporting ideas regarded as less cognitively demand-
ing), while they did not perceive different genres as imposing different levels
of cognitive complexity and difficulty. However, the analyses of linguistic
complexity showed contrasting patterns. Genre effects were prevalent for
many linguistic measures (i.e., greater unit length and phrasal complexity in
argumentative than in narrative essays), while idea support played a limited
role in eliciting more complex language.
Given the statistical analyses of task perceptions and linguistic features,
Yoon argued that textual changes across task types might have little to do
with the learners’ perceptions of the tasks. Previous studies have tested the
validity of task complexity (i.e., task manipulation effects on cognitive task
complexity) with regard to significant changes in linguistic features, mostly
tapping into the construct of linguistic complexity. However, Yoon found
that the provision of idea support actually resulted in a significant decrease
in L2 learners’ task complexity and difficulty, but this valid manipulation of
task complexity did not push the learners to use different linguistic resources
for task completion, challenging a widely accepted hypothesis in task-based
research.

Yoon’s study confirmed the role of supporting ideas as a task variable affect-
ing cognitive complexity in writing and refuted the general assumption
that argumentative tasks would be cognitively more demanding to adult L2
writers than narrative tasks. The strengths of his study include its rigorous
methodological procedures. As a quasi-experimental study, the participants
were randomly assigned into one of the four groups, and the sequence of
the writing tasks was counterbalanced. The four tasks were on very similar
topics of language learning and use, which was an attempt to remove poten-
tial topic effects on language use, and the learners’ task perceptions were
examined immediately after their writing performance.
168 Charlene Polio, Hyung-Jo Yoon

Nevertheless, it should be noted that only one prompt was employed

for the provision of each of the four task conditions (e.g., narrative with
supporting ideas), and the use of similar topics that aimed to remove topic
effects might have potentially led to task repetition effects. It is also plau-
sible that the idea support prompts that included some additional input
with complex phrases might have affected the language use of the learners.
Other limitations, common to many reviewed in this chapter, are addressed
in the next section.

8.7 Future research directions

We address here three possible areas for future research: the examination of
a wider range of genres, including those written under real-life conditions,
and a wider range of languages; studies more focused on pedagogical impli-
cations; and combining CALF measures with other methods.
Studies of genre using CALF measures have been limited to broad
categories of what we might call modes written under timed conditions.
Another option would be to compare novice writers’ texts (both L1 and
L2) in research articles (perhaps published and rejected), to determine
where novice writers may need more assistance. This would be a within-
genre study. A possible between-genre study might be to examine CALF
differences, for example, between a lab report and email, or an academic
article and academic blog. In addition, we know of no studies replicat-
ing genre difference studies with other languages. For example, the dif-
ferences between narratives and argumentative essays in English has been
well documented, but if these differences are related to the communicative
functions of the genres, it is not a given that language will vary in the same
across all languages.
Second, it is clear from the research on CALF and genres that L2 writ-
ers adjust their writing accordingly, but the pedagogical implications of
this finding are not obvious other than, as Yoon and Polio (2017) argue,
we should not limit assignments to what is believed to be a so-called easier
genre. Thus, embarking on research that can better inform pedagogy and
that can link these linguistic differences to instruction would be helpful.
Given that most of the research has found differences in complexity and
not accuracy or fluency, we focus on syntactic complexity. Specifically, we
should investigate the effects on different genre production on develop-
ment. For example, in cases where students are limited to writing narra-
tives, we could challenge them with argumentative essays early on to see if
such writing speeds up development overall.
Finally, we end with a suggestion that CALF measures be combined
with other approaches discussed in this volume. CALF measure studies
 Complexity, accuracy, and fluency (CALF) measures 169

can be reductionist and do not provide full picture of the complexities of

writing proficiency. For example, we do not know if writers who can vary
their language actually understand the differences across genres or how
they came to know about language difference. Case studies that include
interviews, stimulated recall, or classroom observations, might provide a
fuller picture.

References
Abdel Latif, M. M. (2013). What do we mean by writing fluency and how can it be
validly measured? Applied Linguistics, 34(1), 99–105. https://doi​.org​/10​.1093​
/applin​/ams073
Abdi Tabari, M. (2019). Differential effects of strategic planning and task structure
on L2 writing outcomes. Reading and Writing Quarterly, 36(4), 320–338.
https://doi​.org​/10​.1080​/10573569​.2019​.1637310
Alexpoulou, T., Michel, M., Murkami, A., & Meurers, D. (2017). Task effects
on linguistic complexity and accuracy: A large-scale learner corpus analysis
employing natural language processing techniques. Language Learning, 67(S1),
180–208. https://doi​.org​/10​.1111​/ lang​.12232
Beers, S., & Nagy, W. (2009). Syntactic complexity as a predictor of adolescent
writing quality: Which measures? Which genre? Reading and Writing, 22(2),
185–200. https://doi​.org​/10​.1007​/s11145​- 007​-9107-5
Beers, S., & Nagy, W. (2011). Writing development in four genres from grades
three to seven: Syntactic complexity and genre differentiation. Reading and
Writing, 24(2), 183–202. https://doi​.org​/10​.1007​/s11145​- 010​-9264-9
Biber, D., & Conrad, S. (2009). Register, genre, and style. Cambridge University
Press.
Biber, D., Gray, B., & Staples, S. (2016). Predicting patterns of grammatical
complexity across language exam task types and proficiency levels. Applied
Linguistics, 37(5), 639–668. https://doi​.org​/10​.1093​/applin​/amu059
Caplan, N. (2019). Have we always taught the five-paragraph essay. In N. Caplan
& A. M. Johns (Eds.), Changing practices for the L2 writing classroom: Moving
beyond the five-paragraph essay (pp. 2–23) University of Michigan Press.
Carlstrom, B., & Price, N. (2012). The Gachon learner corpus. http://kor​eanl​earn​
erco​r pusblog​.blogspot​.kr​/p​/corpus​.html
Crossley, S. A., Bradfield, F., & Bustamante, A. (2019). Using human judgements
to examine the validity of automated grammar, syntax, and mechanical errors
in writing. Journal of Writing Research, 11(2), 251–270. https://doi​.org​/10​
.17239​/jowr​-2019​.11​.02​.01
Crowhurst, M., & Piche, G. L. (1979). Audience and mode of discourse effects on
syntactic complexity in writing at two grade levels. Research in the Teaching of
English, 13, 101–109.
de Oliveira, L., & Smith, S. L. (2019). Interactions with and around texts:
Writing in elementary schools. In N. Caplan & A. M. Johns (Eds.), Changing
practices for the L2 writing classroom: Moving beyond the five-paragraph essay (pp.
65–88).University of Michigan Press.
 170 Charlene Polio, Hyung-Jo Yoon

Frear, M. W., & Bitchener, J. (2015). The effects of cognitive task complexity on
writing complexity. Journal of Second Language Writing, 30, 45–57. https://doi​
.org​/10​.1016​/j​.jslw​.2015​.08​.009
Hu, R., Wu, J., & Lu, X. (2022). Word-combination-based measures of
phraseological diversity, sophistication and complexity and their relationship to
L2 Chinese proficiency and writing quality. Language Learning, 72(4). https://
doi​.org​/10​.1111​/ lang​.12511
Hunt, K. W. (1965). Grammatical structures written at three grade levels. National
Council of Teachers of English.
Kuiken, F., & Vedder, I. (2008). Cognitive task complexity and written output in
Italian and French as a foreign language. Journal of Second Language Writing,
17(1), 48–60. https://doi​.org​/10​.1016​/j​.jslw​.2007​.08​.003
Kyle, K. (2022). Writing and vocabulary learning. In R. Manchón & C. Polio
(Eds.), The Routledge handbook of second language acquisition and writing (pp.
183–198) Routledge.
Kyle, K., Crossley, S., & Verspoor, M. (2021). Measuring longitudinal writing
development using indices of syntactic complexity and sophistication. Studies
in Second Language Acquisition, 43(4), 781–812. https://doi​.org​/10​.1017​/
S0272263120000546
Lee, J. (2018). The effects of time constraints, genre, and proficiency on l2 writing
fluency behaviors and linguistic outcomes. [Doctoral dissertation, Michigan
State University].
Lee, J. (2021). Using corpus analysis to extend experimental research: Genre effects
in L2 writing. System, 100, 102563. https://doi​.org​/10​.1016​/j​.system​.2021​
.102563
Lu, X. (2010). Automatic measurement of syntactic complexity in child language
acquisition. International Journal of Corpus Linguistics, 14(1), 3–28. https://
doi​.org​/10​.1075​/ijcl​.14​.1​.02lu
Lu, X. (2011). A corpus-based evaluation of syntactic complexity measures as
indices of college-level ESL writers’ language development. TESOL Quarterly,
45(1), 36–62. https://doi​.org​/10​.5054​/tq​.2011​.240859
Lu, X. (2012). The relationship of lexical richness to the quality of ESL learners’
oral narratives. Modern Language Journal, 96(2), 190–208. https://doi​.org​/10​
.1111​/j​.1540​- 4781​.2011​.01232.x
Lu, X., Casal, J. E., & Liu, Y. (2020). The rhetorical functions of syntactically
complex sentences in social science research article introductions. Journal of
English for Academic Purposes, 44, 1–16. https://doi​.org​/10​.1016​/j​.jeap​.2019​
.100832
McNamara, D.S., Graesser, A.C., McCarthy, P., & Cai, Z. (2014). Automated
evaluation of text and discourse with Coh-Metrix. Cambridge University Press.
Norris, J., & Ortega, L. (2009). Towards an organic approach to investigating CAF
in instructed SLA: The case of complexity. Applied Linguistics, 30(4), 555–578.
https://doi​.org​/10​.1093​/applin​/amp044
Polio, C., & Shea, M. (2014). An investigation into current measures of linguistic
accuracy in second language writing research. Journal of Second Language
Writing, 26, 10–27. https://doi​.org​/10​.1016​/j​.jslw​.2014​.09​.003
 Complexity, accuracy, and fluency (CALF) measures 171

Polio, C. &. Yoon, H.J. (2018). The reliability and validity of automated tools
for examining variation in syntactic complexity across genres. International
Journal of Applied Linguistics, 28, 165–188.
Robinson, P. (2001). Task complexity, task difficulty, and task production:
Exploring interactions in a componential framework. Applied Linguistics, 22(1),
27–57. https://doi​.org​/10​.1093​/applin​/22​.1​.27
Ruiz-Funes, M. (2015). Exploring the potential of second/foreign language
writing for language learning: The effects of task factors and learner variables.
Journal of Second Language Writing, 28, 1–19. https://doi​.org​/10​.1016​/j​.jslw​
.2015​.02​.001
Ryshina-Pankova, M.,& Byrnes,H.(2013). Writing as learning to know: Tracing
knowledge construction in L2 German compositions. Journal of Second
Language Writing, 22, 179–197.
Skehan, P. (1998). A cognitive approach to language learning. Oxford University
Press.
Swales, J. (1990). Genre analysis: English in academic and research settings.
Cambridge University Press.
Tankó, G. (2016). Written summarisation for academic writing skills development:
A corpus based contrastive investigation of EFL student writing. In N. Dobric,
E.-M. Graf, & A. Onysko (Eds.), Corpora in applied linguistics: Current
approaches (pp. 53–78). Cambridge Scholars Publishing.
Tankó, G. (2017). Literary research article abstracts: An analysis of rhetorical
moves and their linguistic realizations. Journal of English for Academic Purposes,
27, 42–55. https://doi​.org​/10​.1016​/j​.jeap​.2017​.04​.003
Tardy, C. (2019). Is the five-paragraph essay a genre? In N. Caplan & A. M. Johns
(Eds.), Changing practices for the L2 writing classroom: Moving beyond the five-
paragraph essay (pp. 24–41). University of Michigan Press.
Tavakoli, P. (2014). Storyline complexity and syntactic complexity in writing and
speaking tasks. In H. Byrnes & R. M. Manchón (Eds.), Task-based language
learning: Insights from and for L2 writing (pp. 217–236). John Benjamins.
Van Waes, L., & Leijten, M. (2015). Fluency in writing: A multidimensional
perspective on writing fluency applied to L1 and L2. Computers and Composition,
38, 79–95. https://doi​.org​/10​.1016​/j​.compcom​.2015​.09​.012
Wolfe-Quintero, K., Inagaki, S., & Kim, H. Y. (1998). Second language development
in writing: Measures of fluency, accuracy, and complexity. University of Hawaii
Press.
Xu, J., & Li, J. (2021). A syntactic complexity analysis of translational English
cross genres. Across Languages and Cultures, 22(2), 214–232. https://doi​.org​
/10​.1556​/084​.2021​.00015
Yoon, H. (2021). Challenging the connection between task perceptions and
language use in L2 writing: Genre, cognitive task complexity, and linguistic
complexity. Journal of Second Language Writing, 54, 100857. https://doi​.org​
/10​.1016​/j​.jslw​.2021​.100857
Yoon, H., & Polio, C. (2017). ESL students’ linguistic development in two written
genres. TESOL Quarterly, 51(2), 275–301. https://doi​.org​/10​.1002​/tesq​.296
9
MULTIMODAL ANALYSIS
Larissa D’Angelo,1 Francesca Marino

9.1 Introduction to the approach and definition of genre

Genre-based research, whether involving academic or non-academic
texts, has captured the interest of applied linguists since the 1980s.
Genre is, in fact, an intuitively attractive concept that has, over the years,
helped researchers in different areas and disciplines categorize texts and
relate these to the context and cultures in which they are shaped. Genre
analysis entails the study of the structure, organization, and language
style of different text types in order to better understand the purpose,
conventions, and expectations of different target audiences for a variety
of community-specific academic and non-academic genres (e.g., Swales,
1990a, 1990b, 2004).
Traditionally, genre-based research in applied linguistics has revolved
around the production and reception of monomodal texts, either written
or spoken. However, the notion of genre has been dramatically altered by
the proliferation of multimodal texts resulting from rapid developments
in digital technologies. Multimodal products and texts are countless (e.g.,
websites, TV programs, YouTube, PowerPoint presentations). The emer-
gence of multiple multimodal genres has urged scholars to adopt new mul-
timodal frameworks and methodologies, allowing researchers to address
unexplored genre-specific features. This multimodal turn in genre analysis
has also been fueled by a gradual awareness among scholars of the ubiquity
of multimodality in our daily lives (Kress & van Leeuwen, 2006).
First coined in the mid-1990s, the term multimodality is nowadays
widely used in numerous academic contexts across different fields, clearly
demonstrating a shared interest in the phenomenon. However, within the
DOI: 10.4324/9781003300847-11
 Multimodal analysis 173

frameworks that have been developed to conduct multimodal analyses, it

is still problematic to pinpoint a common object of study since the con-
cept of multimodality is articulated and operationalized differently across
disciplines and research traditions (Jewitt et al., 2016). Therefore, we find
it necessary to first define the meaning of the term by unpacking the two
complementary notions of mode and multimodal products.
The term mode, sometimes referred to as semiotic resource (e.g.,
O’Halloran, 2011; van Leeuwen, 2005), has taken slightly different mean-
ings within the multiple multimodal approaches to applied linguistics
research. Typically, modes have been seen as both linguistic (i.e., written
or spoken language) and non-linguistic (e.g., visual, aural, gestural) com-
municative resources (Jewitt & Kress, 2003; Kress & van Leeuwen, 2001).
Examples of modes include written text, speech, images, videos, layout,
gestures, and soundtracks (Kress, 2010). To this list, van Leeuwen (2005)
adds tools like pen and ink and computer software and hardware. As Jewitt
and Kress (2003) have pointed out, modes are “broadly understood to be
the effect of the work of culture in shaping material into resources for rep-
resentation” (Jewitt & Kress, 2003, p. 1). As a result, they are socially and
culturally shaped in that their use, and interpretation heavily depends on
the contexts in which they are produced (Kress & van Leeuwen, 2001).
As suggested, what a mode or semiotic resource is depends on the aca-
demic discipline in which we operate. As Jewitt et al. (2016) indicated,
the meanings ascribed to either mode and semiotic resource have varied in
applied linguistics, with scholars advocating, for instance, for the inclusion
of actions and movement into modes, or arguing for a distinction between
visual modes, such as illustrations and photographs. However, regardless of
how we decide to categorize them, most scholars would agree that modes
are combined into multimodal products to generate meaning, and it is pre-
cisely this interplay of modes through which we communicate in our eve-
ryday life that goes under the name of multimodality. Within this chapter,
we will consider a multimodal product as any text simultaneously displaying
different modes/semiotic resources to convey meaning.
Although genre analysis has been approached differently by distinc-
tive theoretical schools, two of them, English for Specific Purposes (ESP)
and Systemic Functional Linguistics (SFL), have been particularly instru-
mental in developing the flourishing domain of multimodal analysis. ESP
genre analysis studies rhetorical patterns and moves in specialized genres
(see Casal & Kessler, this volume), which have been defined as discour-
sal or rhetorical units performing a coherent communicative function in
written (or spoken) discourse (Swales, 1990a). Within this context, genres
are defined and identified through a number of characteristics that makes
them unique from the point of view of their purposes, structures, and the
174 Larissa D’Angelo, Francesca Marino

linguistic elements they present (e.g., Hyland, 2004, 2006, 2009; Swales,
2004). These characteristics are embedded within given discourse com-
munities, and writers tend to tap into repertoires of linguistic resources,
conventions, and patterns specific to the genre in which they are writing to
meet the expectations of their readers, who draw on the same genre-specific
knowledge to understand texts and make assumptions about the writer’s
aim (Hyland, 2009).
Adding to this conception of genre, SFL conceives genre as a step-
by-step pursuit to reach a certain communicative goal. Importantly, this
frame emphasizes the role played by culture and context in the construc-
tion of a text within a certain genre. In SFL, communication is not seen
as occurring in a vacuum, but in meaningful units called texts which
are produced and influenced by context. Within SFL, the construct of
genre proposed by Halliday and Matthiessen (2004) is perceived as only
part of a comprehensive social semiotic theory of language and context.
Specifically, SFL posits genre as a construct which sits at the highest level
of abstraction, followed by context, which entails the variables of field (the
topic being talked about), tenor (the relationship of participants), and
mode (the channel of communication). These variables help to explain
how the use of language predominantly depends upon different choices
in language. For SFL, then, genre represents systems of social processes
that constitute a culture or, in Martin and Rose’s (2008) words, “recur-
rent configurations of meanings […] that […] enact the social practices of
a given culture” (p. 6).
In early genre research adopting multimodal theoretical and methodo-
logical approaches, scholars have explored text-visual interrelations within
academic genres. For instance, within ESP research, particular attention
has been focused on the increasing role played by visual elements in science
textbooks and papers (e.g., Bateman, 2008; Bazerman, 1988; Berkenkotter
& Huckin, 1995; Gunel et al., 2006; Jones, 2007; Miller, 1998) as well as
academic-related multimodal texts created by students (e.g., Tardy, 2005;
Pacheco & Smith, 2015). Recently, interest not only in written but also
oral texts have increased, and a shift from monomodal to multimodal stud-
ies has brought a rich amount of research conducted on academic genres
such as conference presentations (e.g., Querol-Julian & Fortanet-Gomez,
2012; Rowley-Jolivet, 2004), lectures (e.g., Bernard-Mecho, 2017; Fox &
Artemeva, 2013), posters (e.g., D’Angelo, 2016; Rowe & Ilic, 2011), and
science communication blogs (e.g., Tessuto, 2021; Zou & Hyland, 2022).
Within SFL, a flourishing set of studies has concentrated on corporate com-
munication, political business discourse, and related non-academic genres,
such as websites in educational contexts (e.g., Djonov, 2008), corporate
websites (e.g., Coupland & Brown, 2004; Denti, 2015), tourist brochures
 Multimodal analysis 175

(e.g., Francesconi, 2011), print advertisements (e.g., O’Halloran, 2008),

and corporate podcasts and blogs (e.g., Archibald, 2007; Corona, 2021).
Exactly how multimodal academic and non-academic genres have been
explored within the field of applied linguistics will be the focus of the fol-
lowing sections.

9.2 Goals
When multimodal analysis began to gain momentum, researchers started
to move away from the traditional opposition of verbal and non-verbal com-
munication, which saw the verbal mode as primary and all other means as
secondary. Multiple linguists such as Saussure, Jakobson, and Benveniste
perceived language as the highest, most complex, and thus, most impor-
tant form of communication. Semioticians and linguists such as Goodwin
(2001), Kress and van Leeuwen (2006), and O’Halloran (2004, 2011)
introduced the idea that although different semiotic resources offer dif-
ferent possibilities for meaning-making, one resource does not have more
or less potential than another. Rather, these different modes (e.g., writing,
speech, sound, gesture) should not be perceived as separate but combined
into an “integrated, multimodal whole” (Jewitt et al., 2016, p. 2). As such,
a study relying solely on the analysis of written or spoken text will inevitably
provide partial results because “the meanings of the maker of a text as a
whole reside in the meanings made jointly by all the modes in a text” (Kress,
2012, p. 37; italics in original).
What makes a study multimodal then? This depends not so much on the
fact that the genre considered is multimodal, but on the approach selected
by the researcher and the methodological framework developed so that it
accounts for all (or at least some) of the modes simultaneously at play in
the genre. Multimodality, in fact, calls for researchers to move beyond the
empirical boundary of their discipline and develop theories and methods
that interpret how we use different modes together to produce meanings
(Jewitt et al., 2016). For instance, studies using multimodal approaches
have flourished within genre research to address the semiotic complexity of
multiple genres, including research papers (e.g., O’Halloran, 2015), confer-
ence presentations (e.g., Morell, 2015), and school textbooks (e.g., Bezemer
& Kress, 2016), as we illustrate in the following sections.
Although there are numerous methodologies available, three approaches,
in particular, provide useful conceptual tools for applied linguists interested
in written multimodal genres, including (1) Systemic Functional Multimodal
Discourse Analysis (SF-MDA); (2) Social Semiotics Multimodal Analysis;
(3) Multimodal Corpus Analysis (MCA). All of these can be applied (to a
certain extent) to multimodal research, generating research questions that
176 Larissa D’Angelo, Francesca Marino

span different research areas. The abovementioned approaches will be tack-

led in the following paragraphs, and examples of questions deriving from
each approach follow.
Systemic Functional Multimodal Discourse Analysis (SF-MDA) is an
interdisciplinary approach to the investigation of multimodal texts that
originally stems from Halliday’s (1978, 1985) theory of SFL, but was fur-
ther developed by O’Toole (2011), Kress and van Leeuwen (2006), Martin
and Rose (2007), and O’Halloran (1999). SF-MDA can be understood
as an approach that tests the application of the key principles of SFL to
the analysis of semiotic systems other than language, along with their
interactions with each other and with language. Specifically, in SF-MDA,
Halliday’s (1978, 1985) metafunctions are expanded and applied to the
study of multimodal products. In a nutshell, Halliday (1978, 1985) identi-
fied three metafunctions of language simultaneously operating in texts to
create meaning. These are defined as (1) ideational, concerning the ability
of language to categorize human experiences and further configure them
into complex patterns; (2) interpersonal, dealing with how social roles and
attitudes are negotiated through language; and (3) textual, concerning how
language creates discourse. In SF-MDA, the abovementioned metafunc-
tions are respectively extended to (1) how participants (i.e., who/what is
depicted in the multimodal text) are represented, (2) what relationship
exists between the text, the producer, and the audience, and (3) how differ-
ent elements are placed and orchestrated to achieve cohesion and coherence
in multimodal texts.
Although SF-MDA is one of the most popular approaches for the analy-
sis of multimodal texts and genres in discourse studies, MDAs can also
be conducted by taking a variety of different approaches (see O’Halloran,
2011). To provide an example in this chapter, we illustrate a comparative
approach to the analysis of the online review genre (see Section 9.4). Within
this approach, digital or print multimodal texts such as videos, advertise-
ments, and websites, as well as educational material displaying text and
images, are examples of materials collected and analyzed. Examples of
emerging research questions are:

• How are text-image relations established in various genres (e.g., in

textbooks, picture books, and/or informative panels in museum
exhibitions)?
• What is the nature of these relationships?
• To what extent are the text-image relationships compatible with the
communicative purpose of the genre?
• Can text-image relations pose challenges for certain groups of readers
(e.g., special needs readers)?
 Multimodal analysis 177

The second approach we highlight is Social Semiotics, which was first

introduced by Halliday (1978) and then further developed in the work of
Kress and Hodge (1979, 1988) and van Leeuwen (2005). Social Semiotics
is another approach to multimodality that argues against the traditional
semiotic separation between language as a formal system and its use in the
context of social relations and processes including power and ideology. It
is centered on the assumption that meanings derive from social action and
interaction using semiotic resources as tools. Not only it is not possible to
separate a sign system from society and the people that use that sign system
in context, but individuals and societies are shaped by different processes of
meaning-making. With this in mind, Social Semiotics sets out to under-
stand the social dimension of multimodal artifacts, how they are produced,
circulated, and interpreted within society.
As such, a study that adopts a multimodal Social Semiotic approach
might pose research questions such as:

• What multimodal text-making patterns can be found through the anal-

ysis of writing, image, layout, and typography in genres (e.g., school
textbooks)?
• What educational implications arise when these semiotic modes are
utilized?

The third and final approach we describe is Multimodal Corpus Analysis

(MCA). MCA is used to systematically analyze a collection of texts so that
strong empirical grounds can be set within a multimodal analysis. Simply
put, the word corpus is used to refer to a collection of texts which is rep-
resentative of a particular variety or use of language(s). Different types of
corpora exist (see McEnery, 2019), but the general idea within corpus lin-
guistics is that the analyses carried out on larger corpora tend to lead to
more representative data and sounder generalizations. The technological
tools now available allow researchers to process large amounts of digital
data in search of significant patterns.
A key issue in MCA is deciding on the size of the corpus one wants
to investigate. Large machine-readable corpora can be found in corpus
linguistics studies (e.g., the Lancaster-Oslo/Bergen Corpus of British
English, and the Brown Corpus of American English both contain one
million words). However, an MCA study may decide to utilize a smaller
corpus because the amount of data deriving from a multimodal corpus
(capable of collecting a range of printed and digital texts such as web-
sites, newspapers, brochures, and technical manuals, and simultaneously
considering different modes) can easily overwhelm the researcher. Even
small-scale MCA studies (collecting five to ten artifacts) allow researchers
178 Larissa D’Angelo, Francesca Marino

to conduct a fine-grained analysis and pose numerous research questions,

such as:

• Which visual and textual elements can be found in the target multi-
modal genre(s)?
• What factors influence the multimodal structure of these texts-types?
• How do the patterns found change over time?

9.3 Common research methods

As discussed, there are multiple approaches available to applied linguists
to explore multimodal texts, and each approach relies on specific meth-
odologies and tools. Beginning with SF-MDA, this approach utilizes a
type of multimodal text analysis based on descriptions of textual and visual
resources organized according to three metafunctions (experiential, inter-
personal, and textual meaning). Specific system choices are also selected to
subsequently code the multimodal text analyzed and highlight language
and image relations. To annotate and code text, images, and videos in a
variety of genres, the Multimodal Analysis Image software can be used
(see O’Halloran et al., 2012), which proposes a ready-made set of systems
but allows users to edit these systems and add new ones if necessary. The
software ultimately allows the multimodal analyst to understand how semi-
otic choices combine with language choices to create meaning and identify
significant semiotic combinations or hotspots (O’Toole, 2011) such as gaze,
stance, perspective, and framing. Importantly, taking a pedagogical per-
spective, SF-MDA analysis can also be used to explore the meaning-making
process in which L2 learners are engaged when working on multimodal
composing tasks, and most importantly, how L2 learners develop the nec-
essary genre-related metalanguage. For instance, research has investigated
how L2 students employ linguistic and visual modes in genres such as mul-
timodal PowerPoint presentations and Glogster-based multimodal argu-
mentative essays (e.g., Shin et al., 2020).
For Social Semiotics, researchers adopting this approach typically ana-
lyze multimodal examples in context to understand how the social world
is represented through them. To understand why and how multimodal
texts are produced and designed to represent meaning, Social Semiotics
typically requires the collection of a small or large number of multimodal
materials and the observation of the amount and type of semiotic resources
and modal affordances present in the text. Researchers must always con-
sider the social, cultural, or professional context within which the writ-
ten multimodal genre operates. It is often the case that multimodal works
are compared so that similarities and differences emerge depending on
 Multimodal analysis 179

the creator of the text, the cultural and social context, and the audience
addressed.
For example, a Social Semiotic analysis might explore two websites dedi-
cated to the same topic (e.g., contemporary art) but with different audi-
ences in mind, the first website designed for child readers, and the second
website for adult readers. Numerous distinctions would probably be dis-
cernible in the different mode choices made by the two design teams. On
the adult website, the amount of written text will presumably vary greatly
so that writing becomes dominant and foregrounded as well as traditionally
organized linearly in paragraphs, headings, and subheadings. Conversely,
the children’s website might foreground images and audio files and display
writing only in the form of captions for images. The menu bar might be
traditionally positioned on the right side of the screen and images never
challenge the centrality of the text in the adult website, whereas the navi-
gation (as a point of entry and subsequent movements through the website
content) might be decided by the child user, to allow maximum freedom in
the reception of the material displayed. Thus, affordances of modes become
central in this type of analysis, as well as the importance of modularity (i.e.,
each design team knows how to use the different modes and how to arrange
them to address audiences). Although no specific software is necessary to
work with this approach, researchers typically collect multimodal material
containing images and text and use video and audio recordings as well as
transcriptions of interviews.
Finally, turning to MCA, corpus linguists have traditionally used soft-
ware such as AntConc and WordSmith Tools. MCA relies on software
capable of handling linear and non-linear data (Bateman, 2014). The
former consists of written or oral texts that unfold linearly, such as the
sections of an academic article or a recorded speech. The latter involves
text types that have multiple entry points and which can easily be read
non-linearly (e.g., websites, advertisements, academic posters, or museum
informative panels). Importantly, MCA may also be used to address L2
pedagogy. For instance, MCA could be implemented in the teaching of the
meaning-making processes and recurrent patterns found in multimodal
genres. Studies have been conducted on the integration of multimodal cor-
pora into university syllabi and their use as tools for teaching (e.g., Ackerly
& Coccetta, 2007; Baldry & Thibault, 2006; Fortanet-Gómez & Querol-
Julián, 2010).
Linear and non-linear data can be collected, processed, and annotated
using software with different annotating capabilities for multimodal cor-
pus linguistics analyses. For linear data, Jewitt et al. (2016) suggest using
ELAN (see Wittenburg et al., 2006), the Multimodal Corpus Authoring
system (see Baldry & Thibault, 2006), or Multimodal Analysis Video (see
180 Larissa D’Angelo, Francesca Marino

O’Halloran, 2008; O’Halloran et al., 2017), which are also used in other
approaches to multimodal research (e.g., SF-MDA). For non-linear data,
other tools are suggested, such as the Genre and Multimodality (GeM)
annotation schema (see Bateman, 2008) and Multimodal Analysis Image
(see O’Halloran et al., 2017). Both are capable of storing, handling, and
visualizing media files, creating frameworks that are used to annotate
the material collected, and subsequently exporting data into Excel files.
In detail, the GeM annotation schema uses Extensible Markup Language
(XML) for storing annotations, and it is particularly suitable for genre
analysis since it enables researchers to search for as well as quantify struc-
tural patterns within the multimodal corpus. In detail, the GeM annota-
tion schema consists of four analytical layers, including (1) the base layer,
in which the content displayed in the multimodal text is broken down into
semiotic components (e.g., text, images, videos, etc.); (2) the layout layer,
which can be used to analyze content organization (e.g., spatial arrange-
ment, graphic characteristics); (3) the rhetorical layer, which allows explor-
ing the relationship between elements in the multimodal text; and (4) the
navigation layers, specifically designed to navigate the corpus and connect/
compare texts.

9.4 Example studies
To better comprehend the range of multimodal analyses that can be carried
out in genre-based research, we highlight sample studies, each having writ-
ten text as one of the modes under scrutiny (see Table 9.1). These studies
not only utilize the approaches described in the previous paragraph, but
also offer useful examples of how multimodal analyses have been conducted
to explore various academic and non-academic genres.
Academic genres traditionally have been investigated by examining
written texts. However, in the last decade, multimodal analyses have been
growing in importance within genre research. An example is offered by
O’Halloran’s (2015) study on mathematical research papers. Drawing on
SF-MDA, O’Halloran investigates the register used in a research article
published in 2011 in a major journal of theoretical physics, which the author
uses as an example to support her claim of expanding the language-based
conceptualization of the mathematical register to non-linguistic semi-
otic resources. After annotating all modes by using Multimodal Analysis
Image software (http://multimodal​-analysis​.com/), the researcher started
by analyzing excerpts from the article by looking for linguistic features
contributing to the inaccessibility of the language used in mathematics to
a non-expert audience (e.g., dense noun phrases, specialized word mean-
ings, syntactic ambiguity). Importantly, the abovementioned features were
TABLE 9.1 
Example studies using multimodal analysis approaches

Authors Context Research Questions (RQs) Data Collected Comments on Study Design

Adami (2009) • Interaction practices • What video-interaction • 835 videos from • Drawing on Social Semiotics
and meaning- practices are enacted in a video-thread • Important reflections on
making in YouTube video-threads? generated on the interactional exchanges
video-threads YouTube in 2007 occurring between each
response and the initial video
Chik and • Online user- • What are the similarities • 120 reviews (60 • Drawing on MDA
Vásquez generated consumer and differences in from each site) of • Contributing to the body of
(2017) reviews in two restaurant reviews posted one-star Michelin research in the online review
different websites on two different sites in restaurants serving genre from a multimodal
and cultural contexts terms of review format, Asian-style cuisine perspective
(one in the US, one content discussed, • Graphic representation of
in Hong Kong) and use of semiotic the two websites is used to
resources? compare the composition
path for reviewers
Hiippala (2017) • Long-form • How does the • 12 online long- • Drawing on MCA
journalism genre interaction of multiple form articles • Using the GeM annotation
modes make meaning in published in schema
the long-form genre? 2012–2013 from • Annotation tools are stored
US newspapers in a digital corpus that is
available online
(Continued)
Multimodal analysis 181
TABLE 9.1 Continued
182

Authors Context Research Questions (RQs) Data Collected Comments on Study Design
Morell (2015) • International • How are varied modes • Four audio- • Drawing on SF-MDA
conference of communication recorded oral • Developing an SFL and
presentations used in conference presentations (two multimodal framework for
paper presentations, from the social conference presentations
and how are they sciences, two and informing academic oral
combined to ensure from the technical communication pedagogy
effective and successful sciences)
communication?
O’Halloran • Registers in • How is the grammatical • One research • Drawing on SF-MDA
(2015) mathematical difficulty of the article from a • Developing a multimodal
research papers mathematics register major journal in approach to the study and
Larissa D’Angelo, Francesca Marino

enacted through theoretical physics teaching of the mathematics

linguistic features register
of scientific English,
mathematical
symbolism, and
mathematical images?
Rowley-Jolivet • International • How do geology, • Three videos • Drawing on Social Semiotics
(2004) conference medicine, and physics of scientific • Sheds light on similarities
presentations use modes to make conference and differences between
meaning and create presentations each disciplines
coherence in conference from a different
presentations? discipline (i.e.,
• Are there discipline- medicine, physics,
specific meaning-making and geology)
strategies?
 Multimodal analysis 183

derived from the linguistic structures identified by Halliday (1993) in sci-

entific English.
To explore how mathematical symbolic notations and images create
meanings in mathematical texts, O’Halloran first analyzed the semantic
density of mathematical symbols used in equations and showed how the
grammatical strategies required to process symbolic notation (e.g., rule of
order, use of spatial notion) differ substantially from those employed to
decode natural language. Then, the author explored how mathematical vis-
ual reasoning works by closely analyzing a geometrical diagram displaying
a tetrahedron. Varied features contributing to grammatical difficulty were
identified in this mathematical image, such as the use of special visual con-
ventions to display patterns of relations (e.g., Cartesian Coordinate Plane)
and the abstractness of the processes being depicted. Finally, the author
looked at how symbols and images related to one another through lan-
guage (e.g., technical names are used to call the parts of a tetrahedron).
Overall, this study illustrates how meaning is achieved in mathematical
papers through the interaction of linguistic features of scientific English,
mathematical symbolism, and visual representations, and how these com-
ponents result in the difficulty of interpreting mathematical texts.
The SF-MDA framework has also been expanded to the study of the
conference paper presentation genre. For instance, in her small-scale quali-
tative study, Morell (2015) sheds light on the varied meaning-making
resources, from textual material to body language mode, used by English
as a Lingua Franca (ELF) speakers to deliver successful talks. Taking a
holistic view of conference paper presentations, Morell emphasizes the
affordances of multiple modes of communication and the role played by
the combination of varied multimodal strategies in effective presentation
delivery. The dataset included a total of four audio-recorded oral presenta-
tions, two from the social sciences and two from the technical sciences.
Notably, Morell triangulated the findings from the multimodal analysis by
conducting semi-structured interviews with each presenter to gain insights
into the presenters’ English proficiency levels, previous presentation experi-
ences, and beliefs in terms of what features matter in conference presenta-
tions in their fields.
The researcher separately analyzed four modes, including spoken
English, written English, non-verbal material, and body language. For each
mode, the author respectively annotated: (1) the linguistic and paralinguis-
tic features of the talks; (2) the text found on the accompanying slides; (3)
the visual material displayed on the slides and the purposes they served
(i.e., expository, illustrative, decorative); and (4) the presenters’ body lan-
guage. The combination of the modes was then analyzed by looking at the
simultaneous or sequential use of modes. For example, the author observed
184 Larissa D’Angelo, Francesca Marino

how the verbal mode or non-verbal resources were used to facilitate the
understanding of the concise text on the slides. Finally, the presentations
were compared in order to determine what characteristics rendered the
talks effective. As a result of her analysis, Morell developed an SFL and
multimodal framework that ELF presentation trainers and instructors may
use to prepare novice presenters on what/how to present at a conference.
Among the characteristics identified by the author are ideational, organi-
zational, and interactional tools, including conciseness, use of discourse
markers, and inclusive interpersonal devices. No annotating tool was used
in Morell’s study to conduct the analysis, an issue that was noted by the
author as a limitation (e.g., annotating tools such as ELAN could have been
used to obtain more reliable findings).
Another study applying a multimodal approach to the study of confer-
ence presentations is Rowley-Jolivet (2004). In this paper, the researcher
illustrates how conference presentations adhere to specific community
norms, conventions, and practices by using field-specific resources to create
coherence and texture. Using a Social Semiotic analysis and placing particu-
lar emphasis on the slides used by each presenter, Rowley-Jolivet analyzed
three videos of scientific conference presentations from international con-
ferences, each from a different discipline (i.e., medicine, physics, and geol-
ogy). The researcher first conducted a content analysis (Bell, 2001) of the
slide presentations, which led to the identification of four visual categories
that the author defined as (1) scriptural (i.e., textual moves); (2) numerical
(i.e., tables and mathematical formulas); (3) figurative (i.e., photographs);
and (4) graphical (e.g., graphs, models, spectra). These categories were
based on two criteria: (1) semiotic resources used (i.e., linguistic, math-
ematical, or visual); and (2) number of modes displayed (i.e., monosemic
vs. polysemic). Then, the researcher examined the overall number of visuals
per conference presentation, the average of visuals per minute, and in what
proportions each visual was used in each presentation. For instance, find-
ings revealed that presentations in the field of medicine tended to include
more numerical visuals than presentations from the other two fields, while
geology relied more heavily on figurative and graphical visuals.
To further her analysis, the author explored how each meaning-making
resource was used in the three fields by focusing on the spatial and tempo-
ral compositional dimensions of visuals, respectively referring to how visu-
als are integrated into the slides, and the presentation order. For example,
the author analyzed the cohesive bonds between visuals appearing on the
slides by looking at the mutually supportive relations between: (a) high–low
iconic visual pairs (e.g., microscope photos are matched with computer-
produced diagrams of the objects depicted in the photo); and (b) overview-
detail visual pairs (e.g., list of controlled trials in medicine are matched
 Multimodal analysis 185

with an example from one specific trial). As for the temporal dimension of
the presentations used to bound discourse segments, the author explored
the sequence in which slides are shown and identified field-specific formats
which, interestingly, resonates with the moves identified by previous genre
analyses of research papers from each field.
Multimodal analysis approaches have also found application in the study
of multiple non-academic genres, among which are journalistic genres. For
instance, using an MCA approach, Hiippala (2017) analyzed digital long-
form journalism, an emerging genre typically used for storytelling, which
combines written text and visual resources. After compiling 12 articles pub-
lished in 2012–2013 gathered from US newspapers, the author started by
annotating semiotic modes and transitions between the semiotic modes by
using the XML markup language annotation schema provided via GeM.
The latter enabled the researcher to calculate the distribution of modes and
transition patterns in the corpus and search for particular potential mode
combinations. The nine semiotic modes annotated included: text-flow (i.e.,
written narratives), photos, static illustrations (i.e., comics and other images
organized in sequence), dynamic image-flow (i.e., looping videos), page-
flow (i.e., page layouts), still animations (i.e., digital images organized into
shots), and animated illustrations. Additionally, the author identified six
transitions between semiotic modes used to join the modes together (e.g.,
click, dissolve, scroll, wipe, zoom, or none).
The tools used for conducting the analysis were written in the coding
language Python (www​.python​.org). After determining the most promi-
nent modes in the articles, Hiippala evaluated what functions they served
and how they were combined. For example, exploring the integration of
text-flow and dynamic image-flow, the researcher illustrated how a written
third-person narrative of an avalanche survivor had been paired with a brief
first-person video in which the survivor narrated her experience, a multi-
modal strategy used to emphasize the interpersonal meanings associated
with the personal testimony reported in the article. As the author points
out, similar processes of resemiotization in which the same topic is dis-
played in a different form are widely used in digital long-form journalism.
In terms of compositional features, findings illustrated that articles tend
to adopt a linear structure in which complex and distracting audiovisual
narratives are avoided and short looping videos are preferred. Overall, this
study contributes to our understanding of the multimodal nature of digital
long-form journalism and shows how multimodal corpora can be used in
genre analysis.
Customer reviews represent another genre that has witnessed a remark-
able evolution in recent years. Although review articles are not new, the
appearance of platforms and apps where customers can evaluate products
186 Larissa D’Angelo, Francesca Marino

and services has led to the emergence of a new genre in which independ-
ent users can publicly share their experiences: online user-generated con-
sumer reviews. Taking a comparative approach to the investigation of this
genre, Chik and Vásquez (2017) conducted an MDA of restaurant reviews
produced in two different linguistic contexts and websites, one in the US
(Yelp) and one in Hong Kong (OpenRice). The dataset consisted of 120
reviews (60 from each site) of one-star Michelin restaurants serving Asian-
style cuisine. After downloading and storing the reviews in a database,
the researchers translated the reviews from the Hong Kong website from
Chinese into English to compare the texts. Each review was coded in terms
of procedural sections, concerned with the rating system and connoisseu-
rial sections, where the reviewers express subjective opinions. Each author
manually coded half of the data independently, and inter-coder reliability
was established by comparing the authors’ interpretations on a coding sam-
ple. As for the procedural sections, the authors annotated all types of rating
categories, whereas, for the connoisseurial sections, multiple features were
annotated, including word counts, references to specific content catego-
ries related to the dining experience, pictures, emojis, capitalization, and
emphatic punctuation marks, among others.
A fine-grained qualitative analysis was undertaken to determine the sim-
ilarities and differences between reviews from the two websites in terms of
features such as review length and distribution of content categories, and
use and functions of other semiotic resources. For instance, the research-
ers analyzed the subject of the pictures, and if the pictures came with a
caption or evaluation or not. Importantly, a graphical representation (see
Martinec & Van Leeuwen, 2020) of the interfaces of the two websites was
also created to facilitate the comparison of the composition paths prompted
by each platform. Paying attention to how websites or apps are designed is
essential when planning to compare how genres are deployed in different
digital environments since platform- or app-specific features may influence
users’ behaviors and, in turn, the texts they produce. Similarities between
the reviews posted on the two sites were found with regard to the range
length and distribution of some content categories. They differed in terms
of the degree of detail included in the descriptions of the dining experience,
the content being prioritized, and the use of multimodal resources. As the
authors pointed out, intra-genre differences were constrained by both the
cultural contexts and the affordances of the sites where the reviews had
been produced.
For the final example study, while some genres have been reshaped or re-
contextualized, some new genres have also emerged. One of these is a video-
thread, namely users’ video responses given to an initial video uploaded by
another user on a platform. By adopting a Social Semiotics approach to
 Multimodal analysis 187

the study of this genre on YouTube, Adami (2009) conducted a multi-

modal analysis on a video-thread that started from a video that went viral
on YouTube in 2007. The video, titled “@–– Where Do You Tube?––@”,
had been uploaded by a popular YouTuber known as ChangeDaChannel.
The data, compiled into a corpus, comprised approximately 800 videos,
including the initial video, responses provided by other users, and a video-
summary posted by ChangeDaChannel.
After conducting an analysis of the video by ChangeDaChannel to deter-
mine how the YouTuber solicited his potential viewers’ responses, Adami
examined how other users responded to the YouTuber’s content-specific
prompt, namely providing their geographical location. After counting
the number of visual and auditory modes used by responders to represent
their location, the researcher explored the extent to which the original
prompt had been addressed. The last step of the analysis entailed inves-
tigating the video-summary posted by ChangeDaChannel. Specifically,
the researcher examined how ChangeDaChannel creatively resumed the
whole video-thread by analyzing the YouTuber’s remix practices to trans-
form content generated from the video responses. The analysis allowed
Adami to explore how interactions take place in the video-thread genre.
As the author pointed out, users’ responses tended to address only the
initial video, and in numerous video responses, unexpected content was
provided to produce humor. Interestingly, this content playfully reinter-
preted the prompt by rendering the video-thread genre a creative chal-
lenge, mainly relying on the participants’ interest-driven manipulation of
the initial prompts.

9.5 Issues and challenges

Approaching multimodal research from an applied linguist’s perspective
is not an easy task because the first action the researcher has to make is to
position oneself within this diverse field. As Jewitt et al. (2016) suggest,
there is no “right” or “best” approach to multimodality. One has to con-
sider instead the appropriateness and coherence of the approach selected for
the multimodal study one intends to do, all the while shaping the research
topic and modus operandi depending on one’s disciplinary area, theoretical
background, and academic training. Also, the need to have a solid under-
standing of different methodologies and the frequent use of a plurality of
technological tools is what sometimes renders working with multimodal
material problematic. In many of the approaches considered, technological
advances render the empirical research work easier and faster, but they also
call for continuous software and hardware updates and a good command of
these instruments on the part of the researcher.
188 Larissa D’Angelo, Francesca Marino

Notably, multimodality also challenges the notion, traditionally dear to

linguists, that language is the most important element to research and that
it can be studied separately from other present modes (for further discus-
sion, see Adami & Kress, 2014). To name an example, we argue that a
textbook cannot be analyzed by taking into consideration the text only, as
visuals play an important role in communicating information. Therefore,
within this field of research, applied linguists need to engage with all modes
simultaneously, recognizing that the meaningful units that people pro-
duce are nowadays almost always multimodal and that each mode plays an
equally important role in meaning-making processes, as demonstrated by
the example studies.
Finally, numerous studies have been carried out using every approach
mentioned so far, but instances of analyses integrating more than one mul-
timodal approach are rare. The reason can probably be found if we con-
sider a number of limitations. Although these approaches are undoubtedly
useful for multimodal analyses, these are laborious and time-consuming
and lead the researcher to limit the amount of material to be investigated,
because only small fragments can realistically be analyzed. As Bernad-
Mechó (2021) rightly states: “These analyses […] are fine-grained and
produce complex qualitative results that open up new possibilities for
expanding research” (p. 195). Therefore, even if integrating more than
one approach leads to solid qualitative and quantitative analyses, these
necessarily need to be based on small fragments of data. This means that
careful planning is needed before conducting multimodal analyses to
make sure that the examples chosen are truly representative of the target
genre.

9.6 Study-in-focus
In this section, we illustrate an exemplary study by Bezemer and Kress
(2016), which adopts a Social Semiotics approach to analyzing textbooks
for secondary school students and contemporary digital learning materi-
als. As the researchers argue, learning processes have always been shaped
by a wide range of educational resources (and related modes) available to
instructors and learners, among which are textbooks. With this in mind,
they explored the semiotic shifts in textbooks over the last century by con-
sidering how modes, such as written text, image, layout, typography, and
speech, have changed in school textbooks over three periods of time (i.e.,
1930–1940, 1980–1990, and 1995–2005). The results from this study
shed light on the gains and losses in multimodal text-making of textbooks
over the century. Before proceeding further with our commentary on this
article, we urge readers to review the summary provided in the call-out box.
 Multimodal analysis 189

Citation
Bezemer, J., & Kress, G. (2016). The textbook in a changing multimodal land-
scape. In N. M. Klug & H. Stöckl (Eds.), Handbuch Sprache im multimodalen
Kontext (pp. 476–498). Walter de Gruyter.

Research questions
• What has changed in the multimodal text-making of textbooks for sec-
ondary school students over the last century? Have these changes been
influenced by the changing position of textbooks from the only learning
medium to one of the available media over time?
• What may have been gained and what may have been lost in the poten-
tial for learning as a result of these changes?

Context
The authors analyzed 92 excerpts from 59 textbooks as well as 16 digital
resources randomly chosen from English, Science, and Math textbooks
retrieved from the library of the Institute of Education, University of London.
Importantly, the texts had been gathered from three different periods in
the 1930s, 1980s, and 2000s, each of which differs with respect to the role
played by books as a medium of communication and mode of writing.

Procedure
Drawing on Social Semiotics, the researchers developed an analytical frame-
work to analyse four modes in the books, including: writing, typography,
image, layout, and speech. To code the writing, the researchers looked at
mood and clause relations, whereas for typography, they focused on resources
such as spacing, orientation, and space alignment. For the images, the authors
separately coded photos and drawings by considering aspects such as the
provenance of the images and the relationship between images and text, and
by annotating elements such as contextualization, color, pictorial detail, illumi-
nation, depth, and movement (see Kress & van Leeuwen, 2006). The analysis
of the layout of pages focused on various features, including page format,
grid, page orientation, alignment, number of columns per page, and column
width. Speech and moving and dynamic images were also considered when
found in digital learning resources, and special attention was given to mood,
clause relations, and the use of intonation (see Halliday, 2015).
190 Larissa D’Angelo, Francesca Marino

Findings
The analysis of the textbooks revealed changes in terms of multiple modal
aspects over time. For instance, compared to textbooks published in the
1930s, recent textbooks contained fewer complex sentences and passive
sentence forms, and overall, the genre has seen a shift from the formulation
of a scientific hypothesis via the report to a mix of instruction and report. In
addition, modern textbooks are bigger in size and less rigid with page layout,
and the variety of typefaces and fonts being used has increased over time, a
change that is in line with the increasing freedom of choice given to design-
ers. Notably, they also display many more images (from almost no images in
the 1930s to three in every four pages in the 2000s), and their function has
changed from illustrative to complementary to the text.
Finally, the findings show an astonishing change in the production and
use of additional learning materials from the 1930s to the 2000s. In the
1930s, the only resources used in combination with textbooks were “refer-
ence books” and “information books” specifically designed for the school
curriculum. Starting from the 2000s, a wide array of learning resources has
been developed and made available for students, including videos, apps,
and wikis. These changes and the adoption of technology in education have
shifted the position of the school textbook from a central and dominant site
of learning to a more peripheral site.

Drawing on Social Semiotics, this study highlights the interconnection

between semiotic practices and the social world as well as the changing
relationship between production and consumption. For instance, as the
researchers point out, writing sophistication and syntactic complexity have
recently been replaced by an extensive use of images, which may reduce
cognitive demands involved in learning.
Bezemer and Kress (2016) gathered their data from an existing and
easily accessible electronic catalog. Using a cross-sectional corpus of
textbooks produced at three different time points enabled the research-
ers to capture the evolution of the genre over time, something which
is not common in genre analysis. Furthermore, the fine-grained com-
parative analysis of multiple modes in textbooks provided in this study
shows that the evolution of genres not only encompasses changes in the
writing style but can also be marked by changes in other modes (and
mode-specific affordances), and as a result, in the multimodal meaning-
making of texts.
 Multimodal analysis 191

9.7 Future research directions

Multimodal analyses can undoubtedly become exciting endeavors that pro-
vide voluminous amounts of data and lead to important paths forward.
They also entail complex and time-consuming efforts on the part of the
researchers, who have to situate themselves in a continuously evolving field
of study where multiple approaches and tools are available. The frontier
of multimodal research, we believe, lies in the purpose-built software and
hardware that is continuously being developed, such as Multimodal Analysis
Image, Multimodal Analysis Video, and the variety of eye-tracking tools
and bio-sensors now available. These applications not only allow the crea-
tion of online and offline databases that store great amounts of multimodal
material and annotations for further data processing, but they also simplify
collaborations between disciplinary and interdisciplinary research groups.
If the combination of approaches contributes to a universal understanding
of multimodal genres that yields reflections on the contexts, the systems
of choices, and the social actors, it is the use of state-of-the-art technology
that will open a new door to future research in this academic field.
In particular, we hope that the coming years will see a growing body
of literature utilizing this new technology to study the writing, design,
and implementation of multimodal genres. More specific research could
be carried out, for example, on the design and writing of digital material
for educational purposes, paying special attention to its inclusiveness. Also,
social media platforms represent still relatively uncharted territory in the
sense that few studies have applied a well-established theoretical framework
to the analysis of online social interactions. Social media usage changes
continuously, and it has undoubtedly infiltrated numerous aspects of peo-
ple’s everyday lives. It does so by aptly utilizing an array of multimodal
tools and modes, which subsequently calls for thorough studies that are
well-grounded in theory.

Note
1 Although this paper has been jointly planned and developed, Larissa D’Angelo
is responsible for Sections 9.1–9.3, and Francesca Marino is responsible for
Sections 9.4–9.6. Section 9.7 is a joint effort.

References
Ackerley, K., & Coccetta, F. (2007). Enriching language learning through a
multimedia corpus. ReCALL, 19(3), 351–370.
Adami, E. (2009). ‘We/YouTube’: Exploring sign-making in video-interaction.
Visual Communication, 8(4), 379–399.
 192 Larissa D’Angelo, Francesca Marino

Adami, E., & Kress, G. (2014). Introduction: Multimodality, meaning-making,

and the issue of “text”. Text and Talk, 34(3), 231–237.
Archibald, J. (2007). Freedom of expression, opinion and the press in the Gulf
states: Weblogs and emergent international and regional media policy. In
G. Garzone, G. Poncini, & P. Catenaccio (Eds.), Multimodality in corporate
communication: Web genres and discursive identity (pp. 167–179). Franco
Angeli.
Baldry, A., & Thibault, P. J. (2006). Multimodal transcription and text analysis: A
multimedia toolkit and coursebook. Equinox.
Bateman, J. (2008). Multimodality and genre: A foundation for the systematic
analysis of multimodal documents. Palgrave Macmillan.
Bateman, J. A. (2014). Using multimodal corpora for empirical research. In C. Jewitt
(Ed.), The Routledge handbook of multimodal analysis (pp. 238–252). Routledge.
Bazerman, C. (1988). Shaping written knowledge: The genre and activity of the
experimental article in science. The University of Wisconsin Press.
Bell, P. (2001). Content analysis of visual images. In T. van Leeuwen & C. Jewitt
(Eds.), Handbook of visual analysis (pp. 10–34). Sage.
Berkenkotter, C., & Huckin, T. N. (1995). Genre knowledge in disciplinary
communities. Lawrence Erlbaum.
Bernad-Mechó, E. (2017). Metadiscourse and topic introductions in an academic
lecture: A multimodal insight. Multimodal Communication, 6(1), 39–60.
Bernad-Mechó, E. (2021). Combining multimodal techniques to approach
the study of academic lectures: A methodological reflection. Atlantis, 43(1),
178–198.
Bezemer, J., & Kress, G. (2016). The textbook in a changing multimodal landscape.
In N. M. Klug & H. Stöckl (Eds.). Handbuch Sprache im multimodalen Kontext
(pp. 476–498). Walter de Gruyter.
Chik, A., & Vásquez, C. (2017). A comparative multimodal analysis of restaurant
reviews from two geographical contexts. Visual Communication, 16(1), 3–26.
Corona, I. (2021). A window to the world: Visual design and research visibility of
European research projects’ homepages. European Journal of English Studies,
25(3), 352–368.
Coupland, C., & Brown, A. D. (2004). Constructing organizational identities on
the web: A case study of Royal Dutch/Shell. Journal of Management Studies,
41(8), 1325–1347.
D’Angelo, L. (2016). The academic poster genre: Friend or foe? In K. Hayland &
P. Shaw (Eds.), The Routledge handbook of English for academic purposes (pp.
392–402). Routledge.
Denti, O. (2015). Gazing at Italy from the East: A multimodal analysis of Malaysian
tourist blogs. Lingue, Culture, Mediazioni, 2(15), 47–68.
Djonov, E. (2008). Children’s website structure and navigation. In L. Unsworth
(Ed.), Multimodal semiotics: Functional analysis in contexts of education (pp.
216–236). Continuum.
Fortanet-Gómez, I., & Querol-Julián, M. (2010). The video corpus as a multimodal
tool for teaching. In M. C. Campoy-Cubillo, B. Bellés-Fortuño, & M. L. Geo-
Valor (Eds.), Corpus-based approaches to English language teaching (pp. 261–
270). Continuum.
 Multimodal analysis 193

Fox, J., & Artemeva, N. (2013). The cinematic art of teaching university
mathematics: Chalk talk as embodied practice. Multimodal Communication,
1(1), 83–103.
Francesconi, S. (2011). Images and writing in tourist brochures. Journal of Tourism
and Cultural Change, 9(4), 341–356.
Goodwin, C. (2001). Practices of seeing, visual analysis: An ethnomethodological
approach. In T. van Leeuwen & C. Jewitt (Eds.), The handbook of visual analysis
(pp. 157–182). Sage.
Gunel, M., Hand, B., & Gunduz, S. (2006). Comparing student understanding
of quantum physics when embedding multimodal representations into two
different writing formats: Presentation format versus summary report format.
Science Education, 90(6), 1092–1112.
Halliday, M. A. K. (1978). Language as social semiotic: The social interpretation of
language and meaning. Edward Arnold.
Halliday, M. A. K. (1985). An introduction to functional grammar. Edward Arnold.
Halliday, M. A. K. (1993). On the language of physical science. In M. A. K.
Halliday & J. R. Martin (Eds.), Writing science: Literacy and discursive power
(pp. 54–68). Falmer.
Halliday, M. A. K. (2015). Intonation and grammar in British English. De Gruyter
Mouton.
Halliday, M. A. K., & Matthiessen, C. (2004). An introduction to functional
grammar. Oxford University Press.
Hiippala, T. (2017). The multimodality of digital longform journalism. Digital
Journalism, 5(4), 420–442.
Hodge, R., & Kress, G. (1988). Social semiotics. Cornell University Press.
Hyland, K. (2004). Disciplinary discourses: Social interactions in academic writing.
University of Michigan Press.
Hyland, K. (2006). English for academic purposes: An advanced resource book.
Routledge.
Hyland, K. (2009). Academic discourse. Continuum.
Jewitt, C., Bezemer, J., & O'Halloran, K. (2016). Introducing multimodality.
Routledge.
Jewitt, C., & Kress, G. (2003). Multimodal literacy. Peter Lang.
Jones, J. (2007). Multiliteracies for academic purposes: Multimodality in textbook
and computer-based learning materials in science at university. In A. McCabe,
M. O’Donnell, & R. Whittaker (Eds.), Advances in language and education
(pp. 103–121). Bloomsbury.
Kress, G. (2010). Multimodality: A social semiotic approach to contemporary
communication. Routledge.
Kress, G. (2012). Multimodal discourse analysis. In J. P. Gee & M. Handford
(Eds.), The Routledge handbook of discourse analysis (pp. 35–50). Routledge.
Kress, G., & Hodge, R. (1979). Language as ideology. Routledge.
Kress, G., & van Leeuwen, T. (2006). Reading images: The grammar of visual
design. Routledge.
Kress, G. & van Leeuwen, T. (2001). Multimodal discourse: The modes and media
of contemporary communication. Arnold.
 194 Larissa D’Angelo, Francesca Marino

Martin, J. R., & Rose, D. (2007). Working with discourse: Meaning beyond the
clause (2nd ed.). Continuum.
Martin, J. R., & Rose D. (2008). Genre relations: Mapping culture. Equinox.
Martinec, R., & Van Leeuwen, T. (2020). The language of new media design: Theory
and practice. Routledge.
McEnery, T. (2019). Corpus linguistics. Edinburgh University Press.
Miller, T. (1998). Visual persuasion: A comparison of visuals in academic texts and
the popular press. English for Specific Purposes, 17(1), 29–46.
Morell, T. (2015). International conference paper presentations: A multimodal
analysis to determine effectiveness. English for Specific Purposes, 37, 137–150.
O’Halloran, L.K., Podlasov, A., Chua, A., & Marissa, K. L. E. (2012). Interactive
software for multimodal analysis. Visual Communication, 11(3), 363–381.
O’Halloran, L. K. (1999). Towards a systemic functional analysis of multisemiotic
mathematics texts. Semiotica, 124(1/2), 1–29.
O’Halloran, L. K. (2004). Multimodal discourse analysis. Continuum.
O’Halloran, K. L. (2008). Systemic functional-multimodal discourse analysis
(SF-MDA): Constructing ideational meaning using language and visual
imagery. Visual Communication, 7(4), 443–475.
O’Halloran, L. K. (2011). Multimodal discourse analysis. In K. Hyland & B.
Paltridge (Eds.), Companion to discourse (pp. 120–137). Continuum.
O’Halloran, L. K. (2015). The language of learning mathematics: A multimodal
perspective. Journal of Mathematical Behaviour, 40, 63–74.
O’Halloran, K. L., Tan, S., & Marissa, K. L. E. (2017). Multimodal analysis for
critical thinking. Learning, Media and Technology, 42(2), 147–170.
O’Toole, M. (2011). The language of displayed art (2nd ed.). Routledge.
Pacheco, M. B., & Smith, B. E. (2015). Across languages, modes, and identities:
Bilingual adolescents’ multimodal codemeshing in the literacy classroom.
Bilingual Research Journal, 38(3), 292–312.
Querol-Julian, M., & Fortanet-Gómez, I. (2012). Multimodal evaluation in
academic discussion sessions: How do presenters act and react? English for
Specific Purposes, 31, 271–283.
Rowe, N., & Ilic, D. (2011). Poster presentation–a visual medium for academic and
scientific meetings. Paediatric Respiratory Reviews, 12(3), 208–213.
Rowley-Jolivet, E. (2004). Different visions, different visuals: A social semiotic
analysis of field-specific visual composition in scientific conference presentations.
Visual Communication, 3(2), 145–175.
Shin, D., Cimasko, T., & Yi, Y. (2020). Development of metalanguage for
multimodal composing: A case study of an L2 writer’s design of multimedia
texts. Journal of Second Language Writing, 47, 100714.
Swales, J. (1990a). Genre analysis: English in academic and research settings.
Cambridge University Press.
Swales, J. (1990b). Discourse analysis in professional contexts. Annual Review of
Applied Linguistics, 11, 103–114.
Swales, J. (2004). Research genres: Exploration and applications. Cambridge
University Press.
Tardy, C. M. (2005). Expressions of disciplinarity and individuality in a multimodal
genre. Computers and Composition, 22(3), 319–336.
 Multimodal analysis 195

Tessuto, G. (2021). Medicine and biology science communication blogs:

Investigating stance patterns for gender identity construction. In L. D’Angelo
& C. Consonni (Eds.), CERLIS series vol. 10: New explorations in digital
metadiscourse (pp. 143–166). CELSB.
van Leeuwen, T. (2005). Introducing social semiotics: An introductory textbook.
Routledge.
Wittenburg, P., Brugman, H., Russel, A., Klassmann, A., & Sloetjes, H. (2006).
ELAN: A professional framework for multimodality research. In Proceedings of
the 5th international conference on language resources and evaluation (LREC
2006) (pp. 1556–1559). European Language Resources Association (ELR A).
Zou, H., & Hyland, K. (2022). Stance in academic blogs and three-minute theses.
International Journal of Applied Linguistics, 32(2), 225–240.
SECTION III

Focus on the learner and

learning


10
MULTILINGUALISM AND
MULTICOMPETENCE
Bruna Sommer-Farias

10.1 Introduction to the approach and definition of genre

Multicompetence is a paradigm that encourages researchers to move away
from the monolingual norm. More than a theory or model to be disproved
(Murahata et al., 2016), multicompetence as defined by Cook (1992, 2008)
posits that the bilingual mind is not composed of two monolinguals in one
mind but is uniquely and qualitatively different with the presence of more
than one language in one mind (Cook, 2016; Grosjean, 1982). Since mul-
ticompetence postulates hybrid and complex sets of knowledge, genre stud-
ies within this framework are interested in how genre knowledge is built
across languages and in relationship to other types of knowledge.
Studies investigating genres written by users of two or more languages
(i.e., multilingual writing) and the composing processes involved do not
commonly define genre explicitly or claim to adopt a multicompetence par-
adigm. The same happens with other multilingual approaches in the second
language (L2) writing field at large (Gentil, 2022; Kobayashi & Rinnert,
2016), such as the focus on multilingualism (Cenoz & Gorter, 2011) and
translanguaging (Otheguy et al., 2015). For this reason, studies included
in this chapter examine multilingual writing through a genre perspective,
although they do not always explicitly state a specific genre school or mul-
ticompetence as the leading paradigm. These studies include writing pro-
duced by users in a range of proficiency levels and learning stages.
Multilinguals become multicompetent language users (Cook, 2008)
through exposure to and engagement in communities that use more than
one language, thus definitions of genre as social action (Miller, 1984) and
genres as social practices of discourse communities (Swales, 1990) can
DOI: 10.4324/9781003300847-13
200 Bruna Sommer-Farias

provide compatible frameworks to consider the role played by the con-

text in which users speak and (learn how to) write in multiple languages.
Qualitative methods are frequently adopted to investigate how multilin-
guals build and expand their genre repertoires dynamically. One example is
Canagarajah’s (2011) analysis of a student’s (Sivatamby) shuttling between
languages, identities, and cultures in English and Tamil. Though not named
as such, Sivatamby demonstrates multicompetent abilities learned through
multilingual practices in Tamil communities and English-speaking Western
academic practices to skillfully write for academic and non-academic audi-
ences in both languages. In fact, the community’s influence on individual
language practices and policies is addressed by the latest definition of mul-
ticompetence, which has been defined as both “the compound state of a
mind” and “the overall system of a mind or a community that uses more
than one language” (Cook, 2016, p. 3).
Genre research has investigated writing produced by emergent
bilinguals or multilingual users, but examining the role previous lan-
guages play in L2 writing development and, more specifically, in genre
knowledge development, is a more recent endeavor (Gentil, 2011).
For example, more than half of the 60 empirical studies reviewed by
Tardy (2006) had multilingual writers as participants, but information
on linguistic background, when obtained, was not often part of the
analysis. Nonetheless, the influence of previous languages and literacy
practices was frequently acknowledged in discussion sections in arti-
cles across Systemic Functional Linguistics (SFL), English for Specific
Purposes (ESP), and Rhetorical Genre Studies (RGS) genre schools
(e.g., Artemeva & Myles, 2015; Yasuda, 2011), though very few studies
have examined genre knowledge across languages specifically (Sommer-
Farias, 2020).
Metacognitive knowledge has been considered key for multilingual
genre knowledge (Gentil, 2011) based on findings such as the strong asso-
ciation between first and second language writing proficiency mediated by
metacognitive knowledge and linguistic fluency, and the non-language-
dependent nature of metacognition (Schoonen et al., 2011). Recent studies
have argued that genre awareness may be metacognitive and non-language-
dependent (Negretti & Kuteeva, 2011; Negretti & McGrath, 2018), and
therefore, available for multilingual users across languages (Tardy et al.,
2020). Conversely, genre-specific knowledge, or “the knowledge that writ-
ers hold of a particular genre or group of genres” (Tardy et al., 2020, p.
294), would be at least partially language-dependent.
Most genre-based studies on multilingual writing have been conducted
in English-medium academic settings, though advocacy for more studies
on understudied populations (e.g., refugees, immigrants, heritage learners)
 Multilingualism and multicompetence 201

has increased (Ortega & Carson, 2010; Ortega, 2019). One of the conse-
quences of the bilingual and social turns in SLA is the need to advance the-
oretical frameworks to account for multicompetent users’ abilities (Ortega
& Carson, 2010), which is extendable to genre studies. To better account
for the sociocultural, cognitive, and linguistic aspects of composing gen-
res across languages, transdisciplinary efforts should be implemented to
inform more mixed-method approaches to inquiry. For instance, a bilit-
eracy perspective on genre aims to understand writing produced in two or
more languages combining “the insights from research in biliteracy and
bilingualism in order to shed light on how multilingual writers develop and
use genre expertise in more than one language” (Gentil, 2011, p. 7).
A range of genres written by multilinguals have been analyzed with a
multicompetence lens, but academic texts are still the most frequent. The
use of Catalan, Spanish, and English in research group blogs (Luzón,
2017) is one example that represents the growing body of research on
the use of multiple languages across digital platforms for academic and
non-academic audiences (Pérez-Llantada, 2021). The influence of writing
instruction, language proficiency, and antecedent genre knowledge used
across languages is another relationship explored. Séror and Gentil (2020),
for instance, show how multilingual university students became aware of
insights gained in L1 and L2 writing classes while critically examining con-
text-appropriate cross-linguistic versus compartmentalization of language
choices to write genres. Traditionally, SFL-oriented studies analyze novice
learners’ production, so a few studies have used SFL principles to examine
features in secondary school writing across languages (e.g., Lindgren et al.,
2017), whereas studies following ESP and RGS principles focus more on
advanced learners and include adapting knowledge and using metacogni-
tive genre awareness from instructional to workplace settings (e.g., Kessler,
2021, Rounsaville et al., 2008).

10.2 Goals
Questions surrounding multilingual writing seek to understand the rela-
tionship between the languages known by the writer and other types
of knowledge at play during the composing process. More specifically,
research is interested in how languages interact or influence each other and
how possible overlaps across languages and types of knowledge characterize
how writers compose specific genres across languages. Drawing from bilin-
gual cognition and education research (e.g., Jessner, 2006), possible cogni-
tive benefits from learning genres multilingually would potentially include
“genre awareness, rhetorical flexibility, and audience sensitivity” (Gentil,
2011, p. 20). These characteristics were observed in Spanish-English
202 Bruna Sommer-Farias

bilinguals using genre-specific knowledge in L3 Portuguese to write cur-

ricula vitae (CVs) in English. They identified appropriate formal conven-
tions for each language and justified the rhetorical framing to the pertinent
audience instead of resorting to direct translation (Sommer-Farias, 2020).
Theoretical hypotheses such as whether there is a common underlying
proficiency across languages (Cummins, 2000), whether genre knowledge
dimensions are dependent or independent of one specific language (Gentil,
2011; Tardy et al., 2020), and how proficiency level plays a role in genre
knowledge across languages have inspired various research designs. For exam-
ple, one line of research analyzes multilingual writing by looking at genre
knowledge as part of a broader model of repertoire of knowledge combining
genre, topic, disciplinary, and L1/L2/L3 writing knowledge (Rinnert &
Kobayashi, 2016). Another line of research includes biliteracy genre studies
(Gentil, 2011), which draw from the continua of biliteracy (Hornberger,
2003), among other theories, to analyze the role of social context in com-
mitments and developments of literacy in more than one language. Finally,
there is research that uses Tardy’s (2009) genre knowledge framework to
examine the dependence and overlap of rhetorical, process, formal, and con-
tent knowledge across languages (e.g., Tardy et al., 2020), along with the
role of genre awareness when writing across languages. All these approaches
investigate the social and linguistic aspects of writing specific genres. To
do so, questions about multicompetent users writing specific genres require
eliciting data about familiarity with similar genres written in other languages
or contexts (e.g., professional and instructional settings). The adaptation of
knowledge from one context to another, which could potentially describe
language use across genres, can be found in the literature via notions such as
high and low road transfer (e.g., Rounsaville et al., 2008), adaptive transfer
(e.g., De Palma & Ringer, 2011), conditional knowledge (e.g., Negretti &
McGrath, 2018), and recontextualization (e.g., Cheng, 2007).
Thus, when investigating genres from a multicompetence lens, research-
ers often pose questions such as:

• How do languages interact or influence each other when writing genres

in an additional language?
• What is the influence of L1 writing instruction and previous L1 writing
experience in developing L2 genre knowledge?
• How does genre knowledge overlap with writing expertise and language
knowledge?
• What are the similarities and differences in genre-specific knowledge (i.e.,
formal, rhetorical, process, subject-matter knowledge) across languages?
• How do multilingual genre repertoires contribute to understanding
genres in additional languages?
 Multilingualism and multicompetence 203

• What is the role of (metacognitive) genre awareness in developing genre

knowledge across languages?
• How do writers build genre knowledge in another language when in a
new context (e.g., study abroad, academic setting, professional setting)?
• How do students build genre knowledge across languages over time?
• How can teachers use students’ genre knowledge in one language to
leverage the learning of specific genres in additional languages?

10.3 Common research methods

A clear definition of theories elected to frame questions and methods is
paramount to operationalizing sociocultural, cognitive, or social dimen-
sions of genres composed by multilingual writers. Once the constructs have
been defined, the researcher is able to select methods suitable to elicit data
specific to that construct. Researchers interested in writers’ performance
of one or more specific genres across languages, also defined as “a proce-
dural knowledge of the genre” (Tardy et al., 2020, p. 295, emphasis in the
original), should implement within-writer designs to analyze production
by the same writers across languages. This form of analysis aims to iden-
tify similarities and differences across languages (Cenoz & Gorter, 2011)
but also rhetorical patterns (Abasi, 2012) and meaning-making processes
(e.g., Lindgren et al., 2017). To facilitate comparability, genres may be pro-
duced in response to similar prompts under similar conditions in differ-
ent languages or collected from instructional or professional settings in
naturally occurring conditions. Interviews and questionnaires about previ-
ous writing experiences can help situate antecedent genre knowledge, and
stimulated recalls can provide a glance at writing strategies shared across
languages. These analyses have been reported as case studies varying from
one semester to two or more years of data collection (e.g., Kim & Belcher,
2018; Kobayashi & Rinnert, 2013), and less frequently as large-scale exper-
imental studies (e.g., Rinnert & Kobayashi, 2012).
Through self-report data, researchers can elicit writers’ awareness
of genre, or the declarative knowledge of a specific genre (Tardy et al.,
2020). Writers are asked to describe genres or their previous experience
with genres via tools such as diagnostic assessments (e.g., Artemeva & Fox,
2010), prior genre use questionnaires (e.g., Sommer-Farias, 2020; Reiff &
Bawarshi, 2011), genre knowledge surveys (e.g., Chsherbakov et al., 2022),
genre analysis tasks (e.g., Cheng, 2007), and autobiographical and text-
based interviews (e.g., Gentil, 2005; Tardy, 2009). Participants can also
be asked to analyze their own texts and composing processes via written
reflections (e.g., Negretti & Kuteeva, 2011), written analyses (e.g., Kuteeva,
2013), and genre analysis of writers’ own texts (e.g., Sommer-Farias, 2020),
204 Bruna Sommer-Farias

a variation of consciousness-raising tasks where writers analyze and justify

their own choices instead of collected genre samples (e.g., Cheng, 2018;
Swales, 1990). The analysis may be accompanied by questions about bilit-
eracy practices or encounters with genres across languages in familial, pro-
fessional, and academic settings. Activation of previous knowledge through
conceptual mind maps can reveal how writers select and organize content
in different languages (Sommer-Farias, 2020).
Due to the tacit nature of genre knowledge, self-report data may be
insufficient because participants may report prior writing experiences inac-
curately or overgeneralize genre conventions (Artemeva & Fox, 2010;
Artemeva & Myles, 2015). To achieve a more nuanced account of writ-
ers’ genre knowledge, the previous methods along with stimulated recalls,
classroom observations, and field notes should be triangulated with writing
assignments. The combination of text-based with ethnographic-oriented
procedures (see Tardy, this volume) opens a window to both social and
linguistic aspects to interpret multilingual writing. In addition, language
policy documents provide information on community policies and possible
influence on levels of literacy and languages of choice (e.g., Gentil, 2005),
especially minority and heritage languages.
Finally, mixed methods have the potential to identify correlations among
quantitatively and qualitatively measured factors. For example, Cenoz and
Gorter’s (2011) empirical study on Basque, English, and Spanish school-age
writers, though grounded in the “Focus on Multilingualism” approach,
can shed light on future genre-based studies. Results showed that students
tended to write the same text type (i.e., story or picture description) on the
same themes in the three languages when writing about pictures in any text
type. Following multicompetence principles, correlations were calculated
across the three languages to verify multidirectionality in language transfer,
composition scores, writing skills, and language mixing in non-instruc-
tional contexts based on extracts from social media posts and a background
questionnaire on their knowledge and use of Basque in social networks.
Correlational methods can be implemented to analyze influences between
proficiency and academic writing skills when composing similar or different
genres across languages (see Rinnert & Kobayashi, 2013).

10.4 Example studies
This section presents an overview of selected studies examining multilin-
gual writing. Table 10.1 displays a summary of each study highlighting
how researchers accounted for participants’ multilingual practices in study
design. As mentioned earlier, studies capturing multicompetence under a
genre lens are scarce, so the selected studies operationalized genre as one
 Multilingualism and multicompetence 205

of the foci and treated participants as multicompetent users of multiple

languages.
The first study by Gentil (2005) draws from biliteracy, critical social
theory, and philosophical hermeneutics’ principles to analyze how three L1
French-speaking graduate students developed L2 disciplinary genre knowl-
edge while attempting to advance L1 writing knowledge for research pur-
poses in an English-speaking university. A longitudinal case study approach
was adopted to describe the multiplicity of practices through which Katia,
Peter, and François built multilingual genre knowledge. For three years,
the researcher collected academic and non-academic writing in English and
French, observed classes, analyzed field notes, context documentation, and
interviews. Among other practices, findings showed how participants used
resources from one language to write in another, discussed institutional
power to influence individual linguistic choices (i.e., English-dominant
academic discourse over other languages), and revealed participants’ pro-
ficiency and confidence growth supported by individuals acting as brokers
including in informal contexts.
The next study highlighted is by Parks (2001), in which the author also
adopted a longitudinal case study design to examine the use of English
and French by 11 nurses who completed their studies in an L1 French
university and had to demonstrate expertise writing care plans in an
Anglophone hospital. The researcher collaborated with an expert member
of the nursing community in both Francophone and Anglophone con-
texts to evaluate the expertise level with which the care plans were per-
ceived. For approximately 22 months the researcher collected work-related
documents, and care plans written at three different points in time, con-
ducted interviews with the nurses and the clinical educator, recorded ses-
sions with the clinical educator providing feedback to the nurses’ writing,
and observed interactions in the workplace. The discussion was centered
around how nurses adapted their L1 genre-specific knowledge to conform
to genre expectations in another language. The study highlighted bound-
aries between school and the workplace marked by cultural differences,
revealing misaligned expectations when educating students to write for
professions. Parks (2001) drew from the notion of genre as social action
from RGS and activity theory to emphasize the role of context and media-
tion in how genres are shaped and reshaped to respond to social expecta-
tions and, ultimately, negotiate taken-for-granted social expectations in
multilingual communities. In this sense, it is important to notice that both
Gentil (2005) and Parks (2001) included data that elicited evidence from
the contexts to account for the role of the discourse community in shaping
writers’ exigencies, language policies, and genre-specific features seen as
expert writing in each language.
206 Bruna Sommer-Farias

The last example of a longitudinal study in the table is Kobayashi and

Rinnert (2013), who investigated one multilingual writer, Natsu, writing
in Japanese (L1), English (L2), and Chinese (L3) for two-and-a-half years.
The researchers explicitly refer to multicompetence and the biliteracy genre
perspective (Gentil, 2011) as analytical principles, focusing on 1) second
language acquisition and languages in interaction during composing and
revising processes, and 2) process, social, and textual dimensions as inte-
grated and indissociable from social practices. Data collection included nat-
urally occurring and elicited data, such as argumentation essays produced
in the three languages at two points in time, retrospective simulated recalls,
interviews, and classroom observations. The researchers sustained interac-
tion with the participant via emails and class observations. For the elicited
data, essay types and writing conditions were the same for both periods for
consistency and comparability. The methods and data analysis performed
allowed a multidimensional description of how multicompetent writers
compose genres across languages, considering the multidirectionality of
interactions between languages, and the influence of perceptions of profi-
ciency level and identity on text construction and composition processes.
In an earlier article, Kobayashi and Rinnert (2012) analyzed the acquisi-
tion, expansion, and interrelationship of language systems through writing
knowledge using a model of text construction and knowledge repertoire.
In this study, genre was considered a situational factor along with audi-
ence, task, topic, and setting. Essays composed in English and Japanese by
64 first-year university students in Japan were compiled and grouped into
four categories corresponding to expertise levels and writing instruction
experience (Japanese writers in an EFL context, North-American writers in
a Japanese as a foreign language context, and L1 Japanese writers in an L1
context). Analyses were run for cross-linguistic, cross-sectional, and longi-
tudinal comparisons of text features. Writers in the first group composed
the genre of their choice in response to two prompts to test high school
training effects, while the second group selected one of two prompts for
an argumentative essay to verify argument construction across languages.
Essays were written in their L1 first, then in their L2 with the help of
dictionaries (if needed). Retrospective interviews were conducted to shed
light on the writers’ composing processes, and background questionnaires
provided information on linguistic repertoire and academic experience. The
texts were coded by the two researchers for “types of discourse, subtypes of
argumentation, and introduction/conclusion components” (p. 111), as they
were areas with the potential to evidence cross-rhetorical differences and
language transfer. Inter-coder reliability was calculated post-hoc on 10% of
the English essays selected randomly, and differences across groups were
reported via descriptive statistics. The analyses allowed the observation of
 Multilingualism and multicompetence 207

1) features acquired from early schooling and higher education experiences,

stressing the importance of the literacy experiences in multiple contexts, 2)
progressive overlap of L1 and L2 knowledge as L2 input increases, and 3)
individual differences in language transfer.
The articles by Lindgren et al. (2017) and Kim and Belcher (2018) also
report on case studies. Lindgren et al. (2017) conducted their study in
12 schools in a multilingual indigenous language context, but they selected
texts written by three 15-year-old learners to share. The analysis focused on
ways of making meaning on ideational, interpersonal, and textual levels from
an SFL perspective across three languages: North Sami, Swedish/Finish/
Norwegian, or English. The study does not define genre per se, but explicitly
uses SFL as an analytical framework and subscribes to a multicompetence
perspective to examine how writing develops across languages. Six writing
topics were individually assigned for each student along with what language
to compose descriptive and argumentative texts. Students wrote one or two
texts in studios equipped with computers and a keyboard with Sami special
characters. The writing was completed in three or six sessions depending on
how fast they wrote. The study exemplifies how genre studies can shed light
on the impact of language policies in communities of minorities for language
revitalization purposes. Future studies could benefit from following students
using a longitudinal design to examine the impact of writing in minority
languages for identity and voice construction across languages in later years.
Adding interviews and texts produced outside of classrooms would enable
a more nuanced view of community versus school-based language policies.
Finally, Kim and Belcher (2018) reported on the multilingual academic
writing of four South Korean exchange students in the US. Subscribing to
RGS, the research questions targeted L2 genre knowledge development
during study abroad and how antecedent genre knowledge, including gen-
res across languages, was adapted to English. The knowledge of Korean and
other languages was obtained from questionnaires targeting previous and
current writing experiences in both Korean and English, general writing strat-
egies, and overall changes at the end of the program. Interviews conducted
in Korean were coded for the themes “(a) L2 learning trajectories; (b) prior
genre knowledge in L1 and L2 that had been accumulated in Korea; and (c)
genre practices that occurred at the study abroad” (Kim & Belcher, 2018, p.
60), followed by recurrent themes for L1 and L2 across time and space. These
tools elicited data on perceived prior genre knowledge, yielding the limitation
that “all of the participants were unable to point out specific instruction [on
Korean writing], though they knew how to write in school-related genres,
such as reports and essays in Korean” (p. 66). This example sheds light on the
distinction between genre awareness and genre-specific knowledge and how
procedural and declarative knowledge of genre writing differ.
208 Bruna Sommer-Farias

10.5 Issues and challenges

Researchers interested in genre research through multilingualism and
multicompetence lenses should be aware of a few issues on data elicitation
and analysis. First, defining genre and related constructs (genre-specific
knowledge, genre awareness, etc.) is key, especially for quantitative studies,
because clear operational definitions enable more reliable data collection
and analytic measures for replication (Kobayashi & Rinnert, 2016). For
example, coding compositions in two languages for genre-specific knowl-
edge dimensions will differ from coding survey responses about language
choices used to write the genre (metacognitive understanding of genre-spe-
cific knowledge) or coding possible genre-specific knowledge dimensions
manifested in the texts. Additionally, in qualitative studies, even when tri-
angulating various data sources, an individual’s experience and proficiency
is still situated and often limited. Therefore, researchers should be cautious
not to generalize writers’ genre knowledge.
The elicitation potential of current research tools for studying genre
knowledge is still to be tested and replicated, urging us to address ethi-
cal concerns and report findings cautiously. For instance, when coding for
process, content, formal, and rhetorical knowledge, the heuristic nature of
genre knowledge (Tardy, 2009) makes it difficult to recognize boundaries
between each type of knowledge as they naturally overlap, as shown by
Chsherbakov et al. (2022). For qualitative analyses, commitment to multi-
ple coders, data triangulation, and member-checking are paramount. For
quantitative analyses, obtaining inter-coder reliability and reporting sta-
tistical measures are encouraged best practices. Both designs can benefit
from reflective practices through journals and positionality statements to
support evolving notions of genre and how they shape research tools and
inquiry-based practices. For instance, Chsherbakov et al. (2022) pointed
out that the academic expertise legitimized in Western universities guided
their Likert scale survey design. Following other academic traditions may
have resulted in a different design.
The third issue, as argued by multiple researchers (Canagarajah, 2006;
Kobayashi & Rinnert, 2016), is that multilingual research should be con-
ducted by researchers who are able to interpret writing in the languages
involved and, similarly, discuss unequal power relations across languages.
While this chapter drew mostly from studies circulating in North America,
research conducted in other regions, such as the Global South, can contrib-
ute to a more comprehensive agenda of languages and skills developed in
other contexts. The power forces exerted by English in writers’ biliteracy
development as well as when negotiating choices in translingual practices
 Multilingualism and multicompetence 209

should also be addressed (Canagarajah, 2013, also see Corcoran et al.,

2019; Pérez-Llantada, 2021 for an extended discussion).
A fourth issue is the need to be critical of our role in deciding what
counts as literacy to interrogate and oppose the class, patriarchal, and
cultural biases that oppress minorities and hinder diversity (Luke, 1996).
Specifically for heritage and minority-language speakers, attention should
be paid to how literacy is framed and measured (Ortega, 2019), and how
research is informed by “the sociopolitical realities of systemic margin-
alization and linguistic insecurity that shape the inequitable multilingual
learning experiences of heritage speakers and their minoritized communi-
ties” (Ortega, 2019, p. 27). To better integrate ethnographic approaches,
establishing relationships with the participants before gathering reliable
data is important for literacy research, especially with learners from diverse
cultural backgrounds (Leung & Brice, 2013). Sustained engagement with
learners will therefore assist researchers in cultivating “a socially situated
and sensitive approach to working with underrepresented groups to protect
their interest” (De Costa et al., 2020, p. 125). In the case of both L2 and
heritage speakers, this can be done by selecting culturally and linguisti-
cally appropriate genres and tasks for research design. For heritage learners,
avoiding the divide between primary versus secondary genres, or everyday
versus academic genres, validates genre repertoires built both at home and
at school (Samaniego & Warner, 2016).

10.6 Study-in-focus
Chsherbakov et al. (2022) was selected as an in-focus study because it is
one of the first studies to examine the overlap of language and dimensions
of genre-specific knowledge (Tardy, 2009; Tardy et al., 2020). Readers are
encouraged to review the summary below to learn more about the research
design adopted.

Citation
Chsherbakov, A., Goodman, B., & Kapashev, D. (2022). Multilingual aca-
demic genre knowledge: Insights from a mixed-method study of post-gradu-
ate students in Kazakhstan. Journal of Second Language Writing, 55, 100872.

Research questions
• To what extent do students perceive their genre-specific knowledge
development across Russian, English, and Kazakh?
210 Bruna Sommer-Farias

• Does self-reported genre-specific knowledge in each language correlate

with their overall self-reported proficiency in that language?
• Can Tardy’s (2009) four-dimensional model be effectively used to evalu-
ate academic genre-specific knowledge in multiple languages through
a Likert scale questionnaire?
• How can the survey design help us understand the structure and
development of academic genre-specific knowledge in the Kazakhstani
context?

Context and population

The participants were faculty, administrators, and masters and doctoral stu-
dents from six universities in Kazakhstan offering English-medium instruc-
tion and multilingual programs where students write in Russian, English,
and Kazakh. The data included a survey to measure levels of genre-specific
knowledge in three languages (n= 283), focus groups, and interviews about
genre-based teaching with students (n= 43), teachers (n= 34) and adminis-
trators (n= 30). To account for varying levels of language proficiency, writers
could answer the surveys in any of the three languages.

Procedure
A survey was created based on Tardy’s (2009) genre knowledge framework
and the latest terminology (Tardy et al., 2020). The survey was submit-
ted for expert content validation (Marsden & Wright, 2010), that is, genre
theory experts selected the items that better aligned with each dimension
of the framework and the survey was piloted by students for accessibil-
ity. The final version included 25 seven-point Likert scale questions with
seven items on process, formal, and rhetorical knowledge and four ques-
tions for content knowledge repeated for each language. Profile questions
(age, courses, language of instruction, academic achievements) and self-
reported proficiency in the three languages based on CEFR and IELTS were
also included. SPSS was used to sum up item values and calculate aver-
ages for each language and genre-specific knowledge dimension, resulting
in 12 aggregated variables. Besides descriptive, reliability, and inferen-
tial statistics, maximum likelihood factor analysis (MLFA) was run to test
the model and create subsets grouped by languages and genre-specific
knowledge dimensions. Interview data were analyzed via open coding and
grouped into themes using NVivo.
 Multilingualism and multicompetence 211

Findings
A one-way ANOVA reported no significant differences between genre-
specific knowledge dimensions within each language (RQ1) but showed
significant differences across languages with a strong correlation between
overall proficiency and genre-specific knowledge, especially for Kazakh
and English (RQ2), showing that “genre-specific knowledge (or expertise)
strongly depends on the knowledge of the language” (p. 9). An MLFA on
non-aggregated items identified irregular items that did not group by lan-
guage within genre-specific knowledge dimensions (p. 6). A second MLFA
performed in subsets of dimensions by language (e.g., formal, content, pro-
cess, and rhetorical knowledge for English, Kazakh, and Russian) excluding
irregular items revealed more dimension overlaps in Kazakh and Russian
than English, with formal knowledge distinctly separated from others in all
languages (RQ3 and 4). Based on qualitative data, patterns could be due
to varied instruction and academic practice (e.g., absence of work on cita-
tions and oral presentation skills in either language could explain irregular
items), whereas English instruction focused on structure more than other
languages, and resource shortage in Kazakh could explain overlaps of other
dimensions except for formal knowledge.

This article proposes an innovative way to study multilingual genre knowl-

edge. Unlike the case studies illustrated earlier in Table 10.1, Chsherbakov
et al. (2022) empirically test a theoretical framework’s potential to describe
how languages interact when writing genres. The authors created a Likert
scale can-do statement survey informed by the genre knowledge framework
(Tardy, 2009; Tardy et al., 2020) to measure genre-specific knowledge in
English, Kazakh, and Russian. The first characteristic that calls attention
in this study is the instrument creation procedure. To transform theoretical
concepts into measurable items, the researchers consulted experts in genre
theory to list items associated with rhetorical, content, process, or formal
knowledge. Statistics tested the consistency and validity of the survey, and
students piloted a final list of items for accessibility. The article clearly speci-
fied the quantitative methods used, reporting the results of all statistical
measures performed. The data visualization through diagrams demon-
strated the possible separation and overlap among dimensions, facilitating
association with the original theoretical model by Tardy (2009). When
discussing the results, the authors argued that genre-specific knowledge
is dependent on language, but proficiency level does not guarantee genre
TABLE 10.1 
Example studies focused on multilingualism and multicompetence
212

Authors Context Research Questions (RQs) Data Collected Comments on Study Design

Gentil • Longitudinal case • How are participants’ • Autobiographical and text- • Oral and written
(2005) studies of three commitment to English based interviews, inventories, reports were shared
Francophone and French academic and academic and non- with participants and
university students literacies’ development? academic writing in English comments on their
in an English- • How do they negotiate and French, classroom-based performance were added
Bruna Sommer-Farias

medium university academic discourses observations, field notes to the data analysis
in Quebec across peers and and context documentation
languages? What is collected over two-and-a-half
the context’s role in years
facilitating or hindering
sustained development
of biliteracy?
Kim and • Case study of four • What is L2 students’ • 12 semi-structured interviews • Participants were
Belcher South Korean prior genre knowledge about 1) previous writing surveyed in regard to
(2018) exchange students when studying abroad? experiences, 2) writing genre network potential
enrolled in a US • How do L2 students experiences in both Korean (Nesi & Gardner, 2012)
university use prior genre and English and general in both languages and
knowledge to write new writing strategies, 3) writing contexts
assignments during experiences of assignments • Interviews were
study abroad? written in study abroad, and conducted in Korean
• How is genre overall perceived changes at the
knowledge built during of the program
study abroad?
Kobayashi • Four groups of • How is L1 and L2 • L1 and L2 essays written • The study compiled
and first-year Japanese writing knowledge by 64 writers with varied and reanalyzed data
Rinnert university acquired and expanded? backgrounds and writing from previous studies
(2012) students’ L1 and • How do L1 and L2 expertise; retrospective via cross-linguistic,
L2 essays writing knowledge interviews, and background cross-sectional, and
interrelate in the questionnaires longitudinal comparisons
knowledge repertoire? of text features
• How do L1, L2, and • The study comprised
L3 writing knowledge four stages according to
interrelate in the writing levels of expertise and
knowledge repertoire? experience with prior
essay-writing instruction
Kobayashi • Longitudinal • How does L1 and L2 • Elicited and naturally occurring • Interaction was sustained
and case study of writing (linguistic data collected for two- using emails and class
Rinnert one Japanese development and text and-a-half years, including observations during
(2013) multilingual writer construction) change argumentation essays, the years; essay data
(Japanese L1, over time? What are simulated recall, interviews, collected in two periods
English L2, and the similarities and and observations (2006 and 2009)
Chinese L3) in a differences across the • Essay types and writing
Japanese academic three languages? conditions were the
context • How do attitude same for both periods
and identity relate to for consistency and
multilingual writing comparability
development?
(Continued)
Multilingualism and multicompetence 213
TABLE 10.1 Continued
214

Authors Context Research Questions (RQs) Data Collected Comments on Study Design
Lindgren • Case study of • What are the meaning- • One argumentative and one • The study was conducted
et al. three 15-year-old making strategies descriptive text written in each in 12 schools, but the
(2017) trilingual students’ and rhetorical devices one of the three languages study reports results for
writing in the employed by three • Questionnaires about language three students
languages North trilingual writers in use, literacy practices, and • Each student was
Sami (indigenous), argumentative and language attitudes toward each assigned six topics to
Bruna Sommer-Farias

either Finnish, descriptive texts? language were completed on be written in a specific

Norwegian or • How are they employed the language of their choice language
Swedish (majority), between and across with parent input
and English languages?
Parks • Longitudinal • How expert is the • Interviews with the nurses • An expert member
(2001) case studies of writing produced by and the clinical educator, tape of the community
11 nurses trained nurses trained in an recordings of feedback sessions (a Francophone
in Francophone Anglophone setting with the clinical educator, nurse experienced
universities and perceived when writing observations, work-related in Francophone and
employed in an in a Francophone documents, and nursing care Anglophone contexts)
English-medium setting? plans written in three different was the evaluator for the
hospital in Quebec points in time nurses’ care plans
• Data was collected over
approximately 22 months
 Multilingualism and multicompetence 215

expertise since writers demonstrated lower English proficiency but more

distinguished genre knowledge dimensions. Qualitative data from inter-
views were used to interpret results, attributing more explicit teaching of
formal features, and writing skills done in IELTS preparation classes than
in other academic environments in Russian and Kazakh. The use of mixed
methods has thus proven valuable to discuss the effect of socially and cul-
turally situated literacies at play, which could explain “the mismatch of the
genre-specific knowledge items and the theoretical model in one language
compared to another” (p. 10).
The study also demonstrates important steps to create and test new
tools to research genre knowledge across languages. While the survey was
validated through pilot testing and demonstrated high internal consist-
ency after the exclusion of irregular items, it is important to emphasize
that responses were based on students’ genre awareness or their declarative
knowledge of genres. To test students’ genre performance (or their pro-
cedural knowledge of the genre), their writing assignments, ideally in all
languages, should have been analyzed and possibly compared to the survey
results. In this sense, the authors acknowledge that the results reveal how
well conceptualized that genre-specific knowledge dimension is rather than
the writers’ expertise level in that dimension. In their words,

While this finding may be interpreted as challenging the validity of the

instrument or the underlying theoretical model (and, to a certain extent
they are), we propose that they are useful indications of the situatedness
of genre, and thus genre-specific knowledge
(Chsherbakov et al., 2022, p. 9)

These results can shed light on the overlap between genre awareness and
genre-specific knowledge (Tardy et al., 2020), as well as the elicitation
potential of this information through a Likert scale survey. Results also
showed that self-report survey data is contingent upon the awareness writ-
ers have developed, which to some extent works as evidence of the overlap
of metacognitive genre awareness and genre-specific knowledge. Finally,
the fact that the situatedness may hinder genre-specific knowledge adapta-
tion to other contexts should be further empirically tested because the fact
that the survey responses did not reveal transfer is insufficient to affirm
that writers are not necessarily applying genre-specific knowledge from one
language when writing in another (i.e., when using procedural knowledge
to compose the genre).
216 Bruna Sommer-Farias

10.7 Future research directions

Research on underrepresented populations is a crucial need for genre stud-
ies to move beyond the focus on university students writing L1/L2 argu-
mentative or expository essays in English. There are studies on Japanese,
Chinese, Portuguese, and Spanish from multicompetence perspectives
as described in this chapter. There is also an increasing number of peda-
gogically oriented pieces on genres in languages such as French, Turkish,
Basque, and German, among others (e.g., Crane, 2016; Manterola, 2019;
Yiğitoğlu & Reichelt, 2014), though the role of L1s in L2 production has
not been their major focus.
Writing produced by multicompetent users in more than two or three
languages is a productive research avenue (Rinnert & Kobayashi, 2016) as
well as writing produced by younger learners in K-12 contexts, especially
with the emergence of dual-language schools (in the US) and English as a
medium of instruction across the globe. Similarly, more attention can be
paid to multimodal genres in both academic and non-academic realms (see
Lim & Kessler, 2022). There has been encouragement for academics to pub-
lish in languages other than English (Luzón, 2017, Pérez-Llantada, 2021),
but research on genres across languages may show that, while the transfer
may be facilitated by multilingual genre knowledge, the specificity of gen-
res (genre-specific knowledge) in each culture and language will require
more preparation from multilingual scholars (Perez-Llantada, 2021). An
anecdote in Tardy et al. (2020) gave a glimpse of what recontextualization
of genre knowledge across languages may look like, but more empirical
research is needed to clarify when and how prior genre knowledge can
facilitate the production of familiar and unfamiliar genres across languages.
Finally, more research can be conducted on the role of multiple languages
in text construction, including the role of each language in composing pro-
cesses and how linguistics, cognitive, and social factors affect multilingual
writing (Manchón, 2013; Rinnert & Kobayashi, 2016).
In the realm of genre theory, empirical research such as Chsherbakov
et al. (2022) can provide evidence to prove, disprove, or refine genre
frameworks. For example, investigating how genre-specific knowledge can
contribute to metacognitive genre awareness and vice-versa is yet to be
observed. Since multicompetence is concerned with the use of a diverse
repertoire of knowledge, genre studies can use the notion of recontextual-
ization to answer questions about using conditional knowledge from one
language to write in another and defining it as either a process, an aware-
ness, or another type of knowledge. For that, more effective tools need to
be developed and tested for looking at genre-specific knowledge and aware-
ness multilingually, as well as how genre-specific knowledge works across
 Multilingualism and multicompetence 217

languages. The use of tools such as think-aloud protocols, retrospective

interviews, keystroke logging, and eye-tracking may enable observation of
more nuanced aspects of each genre-specific knowledge dimension. Other
pedagogical-based strategies can be used to research multilingual students’
composing process, such as screen recordings during students’ composing
process (Gentil, 2018).
In sum, benefits for future studies previously voiced by Gentil (2022)
and Tardy (2006) remain in vogue, especially for multilingual and multi-
competence studies. Using mixed-methods design informed by cognitive
and ethnographic research can help empirically identify how multicompe-
tent writers use their “greater sensitivity, mental flexibility, communicative
sensitivity, selective attention and metalinguistic awareness” (Gentil, 2011,
p. 20) to write genres across languages.

References
Abasi, A. (2012). The pedagogical value of intercultural rhetoric: A report from a
Persian-as-a-foreign-language classroom. Journal of Second Language Writing,
21(3), 195–220.
Artemeva, N., & Fox, J. (2010). Awareness versus production: Probing students’
antecedent genre knowledge. Journal of Business and Technical Communication,
24(4), 476–515.
Artemeva, N., & Myles, D. (2015). Perceptions of prior genre knowledge: A case
of incipient biliterate writers in the EAP classroom. In G. Dowd & N. Rulyova
(Eds.), Genre trajectories: Identifying, mapping, projecting (pp. 225–245).
Palgrave Macmillan.
Canagarajah, A. S. (2006). Toward a writing pedagogy of shuttling between
languages: Learning from multilingual writers. College English, 68(6),
589–604.
Canagarajah, A. S. (2011). Writing to learn and learning to write by shuttling
between languages. In R. Manchón (Ed.), Learning-to-write and writing-to-
learn in an additional language (pp. 111–132). Johns Benjamins.
Canagarajah, A. S. (Ed.). (2013). Literacy as translingual practice: Between
communities and classrooms. Routledge.
Cenoz, J., & Gorter, D. (2011). A holistic approach to multilingual education:
Introduction. Modern Language Journal, 95(3), 339–343.
Cheng, A. (2007). Transferring generic features and recontextualizing genre
awareness: Understanding writing performance in the ESP genre-based literacy
framework. English for Specific Purposes, 26(3), 287–307.
Cheng, A. (2018). Genre and graduate-level research writing. University of
Michigan Press.
Chsherbakov, A., Goodman, B., & Kapashev, D. (2022). Multilingual academic
genre knowledge: Insights from a mixed-method study of post-graduate students
in Kazakhstan. Journal of Second Language Writing, 55, 100872.
Cook, V. (1992). Evidence for multicompetence. Language Learning, 42(4),
557–591.
 218 Bruna Sommer-Farias

Cook, V. (2008). Multi-competence: Black Hole or wormhole for second language

acquisition research. In Z. Han (Ed.), Understanding second language process
(pp. 16–26). Multilingual Matters.
Cook, V. (2016). Premises of multi-competence. In V. Cook & L. Wei (Eds.), The
Cambridge handbook of linguistic multi-competence (pp. 1–25). Cambridge
University Press.
Corcoran, J., Englander, K., & Muresan, L. (2019). Diverse global perspectives
on scholarly writing for publication. In J. N. Corcoran, K. Englander, & L.-M.
Muresan (Eds.), Pedagogies and policies for publishing research in English: Local
initiatives supporting international scholars (pp. 1–16). Routledge.
Crane, C. (2016). Understanding and evaluating L2 personal letter writing:
A systemic functional linguistics analysis of student texts in German. Die
Unterrichtspraxis, 49(2), 122–139.
Cummins, J. (2000). Language, power and pedagogy: Bilingual children in the
crossfire. Multilingual Matters.
De Costa, P., Lee, J., Rawal, H., & Li, W. (2020). Ethics in applied linguistics
research. In H. Rose & J. McKinley (Eds.), The Routledge handbook of research
methods in applied linguistics (pp. 122–130). Routledge.
De Palma, M.-J., & Ringer, J. M. (2011). Toward a theory of adaptive transfer:
Expanding disciplinary discussions of “transfer” in second-language writing
and composition studies. Journal of Second Language Writing, 20(2), 134–147.
Gentil, G. (2005). Commitments to academic biliteracy: Case studies of
francophone university writers. Written Communication, 22(4), 421–471.
Gentil, G. (2011). A biliteracy agenda for genre research. Journal of Second
Language Writing, 20(1), 6–23.
Gentil, G. (2018). Modern languages, bilingual education, and translation studies:
The next frontiers in WAC/WID research and instruction? Across the Disciplines,
15(3), 114–129.
Gentil, G. (2022). Multicompetence and L2 writing. In R. Manchón & C. Polio
(Eds.), The Routledge handbook of second language acquisition and writing (pp.
109–122). Routledge.
Grosjean, F. (1982). Life with two languages: An introduction to bilingualism.
Harvard University Press.
Hornberger, N. (2003). Continua of biliteracy. Multilingual Matters.
Jessner, U. (2006). Linguistic awareness in multilinguals: English as a third
language: English as a third language. Edinburgh University Press.
Kessler, M. (2021). The longitudinal development of second language writers’
metacognitive genre awareness. Journal of Second Language Writing, 53,
100832. https://doi​.org​/10​.1016​/j​.jslw​.2021​.100832
Kim, M., & Belcher, D. (2018). Building genre knowledge in second language
writers during study abroad in higher education. Journal of English for Academic
Purposes, 35, 56–69.
Kobayashi, H., & Rinnert, C. (2012). Understanding L2 writing development
from a multicompetence perspective: Dynamic repertoires of knowledge and
text construction. In R. Manchón (Ed.), L2 writing development: Multiple
perspectives (pp. 101–134). De Gruyter Mouton.
 Multilingualism and multicompetence 219

Kobayashi, H., & Rinnert, C. (2013). L1/L2/L3 writing development:

Longitudinal case study of a Japanese multicompetent writer. Journal of Second
Language Writing, 22(1), 4–31.
Kuteeva, M. (2013). Graduate learners’ approaches to genre-analysis tasks:
Variations across and within four disciplines. English for Specific Purposes, 32(2),
84–96.
Leung, C., & Brice, A. (2013). Ethical issues in conducting research with bilingual/
dual language learners. In C. S. Rhodes & K. J. Weiss (Eds.), Ethical issues in
literacy research (pp. 63–75). Routledge.
Lim, J., & Kessler, M. (2022). Directions for future research on SLA, L2 writing,
and multimodality. In R. M. Manchón & C. Polio (Eds.), The Routledge
handbook of second language acquisition and writing (pp. 325–338). Routledge.
Lindgren, E., Westum, A., Outakoski, H., & Sullivan, K. (2017). Meaning-
making across languages: A case study of three multilingual writers in Sápmi.
International Journal of Multilingualism, 14(2), 124–143.
Luke, A. (1996). Genres of power? Literacy education and the production of
capital. In R. Hassan & G. Williams (Eds.), Literacy in society (pp. 308–338).
Longman.
Luzón, M. J. (2017). Connecting genres and languages in online scholarly
communication: An analysis of research group blogs. Written Communication,
34(4), 441–471.
Manchón, R. M. (2013). Writing. In F. Grosjean & P. Li (Eds.), The psycholinguistics
of bilingualism (pp. 100–115). Blackwell Publishing.
Manterola, I. (2019). A study of recipes written by Basque L2 immersion students:
Any evidence for language revitalization. In N. Yiğitoğlu & M. Reichelt (Eds.),
L2 writing beyond English (pp. 117–134). Multilingual Matters.
Marsden, P., & Wright, J. (2010). Handbook of survey research. Emerald Group
Publishing.
Murahata, G., Murahata, Y., & Cook, V. J. (2016). Research questions and
methodology of multi-competence. In V. Cook & L. Wei (Eds.), The Cambridge
handbook of linguistic multicompetence (pp. 26–49). Cambridge University
Press.
Miller, C. R. (1984). Genre as social action. Quarterly Journal of Speech, 70(2),
151–167.
Negretti, R., & Kuteeva, M. (2011). Fostering metacognitive genre awareness in
L2 academic reading and writing: A case study of pre-service English teachers.
Journal of Second Language Writing, 20(2), 95–110.
Negretti, R., & McGrath, L. (2018). Scaffolding genre knowledge and
metacognition: Insights from an L2 doctoral research writing course. Journal of
Second Language Writing, 40, 12–31.
Nesi, H., & Gardner, S. (2012). Genres across the disciplines: Student writing in
higher education. Cambridge University Press.
Ortega, L. (2019). The study of heritage language development from a bilingualism
and social justice perspective. Language Learning, 70(S1), 15–53.
Ortega, L., & Carson, J. (2010). Multicompetence, social context, and L2 writing
research praxis. In T. Silva & P. K. Matsuda (Eds.), Practicing theory in second
language writing (pp. 48–71). Parlor Press.
 220 Bruna Sommer-Farias

Otheguy, R., García, O., & Reid, W. (2015). Clarifying translanguaging and
deconstructing named languages: A perspective from linguistics. Applied
Linguistics Review, 6(3), 281–307.
Parks, S. (2001). Moving from school to the workplace: Disciplinary innovation,
border crossings, and the reshaping of a written genre. Applied Linguistics,
22(4), 405–438.
Pérez-Llantada, C. (2021). Research genres across languages: Multilingual
communication online. Cambridge University Press.
Reiff, M. J. & Bawarshi, A. (2011). Tracing discursive resources: How students
use prior genre knowledge to negotiate new writing contexts in first year
composition. Written Communication, 28(3), 312–337.
Rinnert, C., & Kobayashi, H. (2016). Multicompetence and multilingual writing.
In R. M. Manchón & P. K. Matsuda (Eds.), Handbook of second and foreign
language writing (pp. 365–385). De Gruyter Mouton.
Rounsaville, A., Goldberg, R., & Bawarshi, A. (2008). From incomes to outcomes:
FYW students’ prior genre knowledge, meta-cognition, and the question of
transfer. WPA: Writing Program Administration, 32(1), 97–112.
Samaniego, M., & Warner, C. (2016). Designing meaning in inherited languages. In
M. Fairclough & S. Beaudrie (Eds.), Innovative strategies for heritage language
teaching: A practical guide for the classroom (pp. 191–213). Georgetown
University Press.
Schoonen, R., van Gelderen, A., Stoel, R., Hulstijn, J., & de Glopper, K. (2011).
Modeling the development of L1 and EFL writing proficiency of secondary
school students. Language Learning, 61(1), 31–79.
Séror, J., & Gentil, G. (2020). Cross-linguistic pedagogy and biliteracy in a
bilingual university: Students’ stances, practices, and ideologies. Canadian
Modern Language Review, 76(4), 356–374.
Sommer-Farias, B. (2020). “This is helping me with writing in all languages”:
Developing genre knowledge across languages in a foreign language course
[Unpublished doctoral dissertation]. University of Arizona.
Swales, J. (1990). Genre analysis: English in academic and research settings.
Cambridge University Press.
Tardy, C. M. (2006). Researching first and second language genre learning: A
comparative review and a look ahead. Journal of Second Language Writing,
15(2), 79–101.
Tardy, C. M. (2009). Building genre knowledge. Parlor Press.
Tardy, C. M., Sommer-Farias, B., & Gevers, J. (2020). Teaching and researching
genre knowledge: Toward an enhanced theoretical framework. Written
Communication, 37(3), 287–321.
Yasuda, S. (2011). Genre-based tasks in foreign language writing: Developing
writers’ genre awareness, linguistic knowledge, and writing competence.
Journal of Second Language Writing, 20(2), 111–133.
Yiğitoğlu, N., & Reichelt, M. (2014). Using a genre-based approach for writing
instruction in a less-commonly-taught language. Language Awareness, 23(3),
187–202.
11
SYSTEMIC FUNCTIONAL LINGUISTICS
AND THE (EXPANDED) TEACHING
AND LEARNING CYCLE
Kathryn Accurso, Jennifer Walsh Marr

11.1 Introduction to the approach and definition of genre

Genre analysis using systemic functional linguistics (SFL) is an empirical
method of identifying and functionally labeling the structure of texts in
different cultures and situations (Halliday & Hasan, 1985; Martin, 1992).
It is commonly used for purposes of description, classification, teacher pro-
fessional development, and/or the design of pedagogy and curriculum.
Following a “top-down” method of analysis (Martin, 1992, p. 167), ana-
lysts organize texts into genre stages, or phases of meaning that account for
the overall coherence of a type of text, and identify the lexicogrammatical
features of each stage. This method has been well-developed for mapping,
teaching, and critiquing dominating school literacy practices, reflecting the
approach’s theoretical commitments to equity, justice, and “the dialogical-
ity of student interests, power dynamics and socio-political issues” (Troyan
et al., 2021, p. 386). Genre analyses often support teachers in developing
curriculum, instruction, and assessments that focus students’ attention on
the social functions of texts, the linguistic and multimodal choices that
comprise them, and how stages interact with one another within larger
texts.
SFL genre theory and genre analysis emerged from the linguistic the-
ory of M.A.K. Halliday as a socially accountable linguistics influenced by
Marxist aims of class emancipation. Halliday (1978) conceptualized social
contexts as being semiotically structured along three dimensions, such that
any instance of text includes linguistic choices that construe happenings
(ideational meanings), roles and relationships (interpersonal meanings),
and the channel of communication (textual meanings).
DOI: 10.4324/9781003300847-14
222 Kathryn Accurso, Jennifer Walsh Marr

As illustrated in Figure 11.1, this relationship between text and con-

text is dialogical, meaning people write, read, speak, view, and listen in
ways that reflect their situational and cultural context. In turn, people’s
meaning-making choices within a particular context dynamically shape
that context and which semiotic choices become associated with and antici-
pated within that context over time. Moreover, language is a series of lay-
ers that inform one another, from small details such as particular word
choices and grammatical structures (combined and termed lexicogrammar)
to the genre stages in which they are situated (and that in turn inform those
very choices), and the situational and cultural contexts in which these texts
occur. None of these layers are independent of one another, as shifts in one
can have consequences for those above and below.
Halliday theorized that ideational, interpersonal, and textual meanings
are best analyzed at the level of register, as lexicogrammatically realized
within and across clauses, the central grammatical unit in SFL (Halliday &
Hasan, 1985). Consider the following clause from a Grade 6 science text-
book: The habitat is the home of a plant or animal (Accurso, 2017, p. 9).
The cultural context is Western science, which values definition, taxonomy,
and neutral observation. The situational context is classroom instruction.

FIGURE 11.1 
SFL model of text/context dynamics
 Systemic functional linguistics and the (expanded) teaching and learning cycle 223

As illustrated in Table 11.1, these aspects of context are construed at the

clause level through particular lexicogrammatical choices.
Martin (1992) proposed that lexicogrammatical choices are further
coordinated beyond the clause by an aspect of context he called genre (see
Figure 11.1). Genres are “staged, goal-oriented social process[es]” (p. 505)
realized in recurrent discourse semantic (i.e., text structure) patterns that
have evolved over time.1 Genres are goal-oriented because they serve social
functions, such as defining a concept (an information genre), telling a story
(a narrative genre), or giving instructions (a procedural genre). They are
social processes because members of a culture interact with one another to
produce them. They are staged because it usually takes more than one step
to accomplish a goal. And they are evolved systems because no individual
sits down to design a genre; rather genres arise as members of a culture
negotiate meaning to live their daily lives. While genres and their discourse
semantic stages have some stability as recognizable patterns for making and
understanding meanings within a culture, they are also flexible, dynamic,
and evolving. Further, genres are ideologically and politically inflected, and
their perceived status is influenced by the unequal distribution of semiotic
resources in society.
According to SFL genre theory, when a person wants to accomplish
something using language, they (consciously or unconsciously) select a
genre and register, cueing a text structure they expect to be effective for
accomplishing that goal, as well as lexicogrammatical resources members
of their culture typically associated with their situation type.2 As Coffin
(2006a) points out, “different genres have distinct ‘beginnings’, ‘middles’,
and ‘ends’, and these structural elements or stages can be identified on the
basis of shifts in lexical and grammatical patterning that correlate with dif-
ferent micro-functions operating at different points in the text” (p. 416). In
sum, SFL genre analysis is a process of identifying discourse stages on the
basis of shifts in authors’ lexicogrammatical choices.

TABLE 11.1 
Sample clause analysis of a Grade 6 science textbook excerpt

Text Register: Lexicogrammatical Choices

The habitat is Ideational: Definition of a specific concept constructed

the home of through specialized terminology (habitat), being verb
a plant or (is), and generalized nouns (home, plant, animal)
animal. Textual: Key concept is made important by putting it in
italics and theme position (i.e., starting the clause)
Interpersonal: Neutral presentation of information
through declarative mood (statement that constructs a
fact) and lack of first-person voice
224 Kathryn Accurso, Jennifer Walsh Marr

Returning to the example of Grade 6 science, when the clause about habi-
tats is analyzed as part of a larger text, we can see that the text’s purpose is not
just to define a term, but to organize and present information about a phe-
nomenon for a distant audience who may or may not be presently observing
the phenomenon. As shown in Table 11.2, to accomplish this purpose, the
text has two obligatory genre stages: identification/definition of the phenom-
enon; and description of its properties (Brisk, 2015; Schleppegrell, 2004).
The purpose of each stage is accomplished by finer-grained register choices
at the level of lexicogrammar. Moreover, the boundary between one genre
stage and the next is marked by a shift in register choices, such as the shift
from generalized to more specific nouns to construct ideational meanings.
Pedagogically, this kind of genre analysis has been incorporated into an
SFL-inspired Teaching and Learning Cycle (TLC) for disciplinary literacy

TABLE 11.2 
Sample genre analysis of a Grade 6 science textbook excerpt (factual
information report genre)

Text Genre: Discourse Register: Lexicogrammatical

Semantic Stages Choices

The habitat is the home Identification/ Ideational & Textual: Italics

of a plant or animal – definition of the and starting theme position
the place where it lives. phenomenon identify the overarching
A large number of things (habitat) phenomenon (habitat);
can share the same being verb (is), generalized
habitat. For example, in nouns (a plant or animal,
Europe, deer, Building information the place where it lives) used
woodpeckers, oak trees, about the to categorize and define the
and bluebells share the phenomenon and phenomenon broadly.
same habitat – the its properties Shift toward more specific
woodland. All the plants nouns (deer, woodpeckers,
and animals living in a oak trees, bluebells, the
habitat form a woodland), more specific
community of living circumstances (in Europe),
things. and more abstract being
verbs (share, live in) to
illustrate and provide more
specific information about
the phenomenon.
Interpersonal: No shift;
maintains neutral
presentation of information
through declarative mood
and lack of first-person
voice throughout
 Systemic functional linguistics and the (expanded) teaching and learning cycle 225

instruction in which teachers scaffold students’ critical orientation toward

and ability to control genres by guiding them through analysis of genre
stages and register choices (Rose & Martin, 2012). Key scaffolding phases
include (1) joint deconstruction of texts in a particular genre, (2) joint con-
struction of a new text in the target genre, and then (3) independent stu-
dent writing. Impressive results in early applications of the TLC have led
to pedagogical uses of genre analysis being widely adopted in Australian
elementary and secondary departments of education, migrant education
programs, university EAP programs, and museum education programs
(Martin, 2012). Subsequently, the TLC spread globally. More recently,
Gebhard (2019) proposed an Expanded Teaching and Learning Cycle
(ETLC) that adds teacher action research elements. These elements include
identifying a problem of practice related to equity and literacy, planning a
curriculum that addresses this problem, and reflecting on the impacts of
instruction to share findings with students, families, researchers, or other
teachers.
Studies that undertake SFL genre analysis within the context of the
TLC often have two phases: First, a “linguistic phase” (Coffin, 2006a, p.
414) that analyzes model texts students are typically assigned in content-
area studies. Second is a teaching and learning phase that involves develop-
ing and implementing curriculum, instruction, and assessment practices to
support students’ writing development alongside content learning. Those
using the ETLC undertake a third phase, analyzing students’ writing to
understand student repertoires and the impact of instruction on students’
literacy development. Section 4 of this chapter highlights a number of stud-
ies of these types.
Because SFL genre theory aims to encompass meaning potential across
a culture, this method can be used to analyze any kind of text produced
in a culture. Past analyses have explored disciplinary genres such as scien-
tific explanations (e.g., Halliday & Martin, 2003), essays in the humanities
(e.g., Nesi & Gardner, 2012), and legal arguments (e.g., Bartley, 2018);
K-12 genres of schooling such as reports, narratives, and procedures (e.g.,
Brisk, 2015); and non-academic genres such as newspaper editorials (Ansary
& Babaii, 2005) and Truth and Reconciliation testimonies (i.e., a genre
found in restorative justice practices in post-apartheid South Africa; see
Hattingh, 2011).

11.2 Goals
As the introduction to this volume explains, SFL genre analysis shares
similar aims with New Rhetoric and English for Specific Purposes (ESP)
approaches. All three approaches attempt to describe and explain patterns
226 Kathryn Accurso, Jennifer Walsh Marr

in purpose, form, and situated social action across texts. Moreover, like
ESP, SFL analysts are typically committed to making findings from genre
analyses usable for teachers (see Casal & Kessler, this volume). However, as
Figure 11.1 illustrates, SFL provides a more detailed framework for analysis
across levels of language and context. In addition, it is more often applied to
K-12 contexts than ESP and New Rhetoric approaches, with a focus on sup-
porting students’ ability to control, critique, and remix so-called “genres
of power” (Rose & Martin, 2012, p. 67; see also Harman & Khote, 2018).
In line with these overarching interests, researchers adopting SFL genre
analysis typically explore questions like:

• What types of purposes do writers use text to accomplish in different

contexts?
• How are texts structured in stages to accomplish these purposes?
• What stages are obligatory for accomplishing these social purposes
within a particular cultural context? What genre stages are optional?
How does stage placement vary and to what effect?
• What patterns of register choices distinguish or contribute to the realiza-
tion of different genre stages (i.e., lexicogrammatical choices that con-
strue ideational, interpersonal, and textual meanings)?
• How do different texts within a cultural context relate to one another in
genre families?
• How do genre and register patterns with a genre family compare (e.g.,
literary argument versus legal argument)?
• How can we describe and/or classify new types of texts as contexts
evolve (e.g., academic blogs)?

When adopted within the context of the TLC or ETLC, teachers and
researchers also explore questions like:

• What is the impact of genre-based instruction on student writing over

time?
• How do students’ use of genre stages compare to more experienced users
of the genre in a particular cultural context?
• What challenges do student writers face when engaging with a particular
genre or genre family?
• What is the relationship between students’ genre knowledge, sociopo-
litical consciousness, and their ability to navigate high-stakes schooling
contexts?
• How do students play with genre conventions or use language in novel/
unexpected ways to accomplish social purposes that are important to
them?
 Systemic functional linguistics and the (expanded) teaching and learning cycle 227

Through these kinds of questions, SFL genre analysis aims to bring to

consciousness the ways texts are structured and lexicogrammatically shaped
to accomplish different social purposes, not to prescribe assimilation to
particular structures, but to “reflect systematically and critically on the
social role, purpose, and textual strategies of texts” within a given context
(Coffin, 2006a, p. 414).

11.3 Common research methods

Broadly speaking, SFL genre analysis involves collecting a corpus of texts
and analyzing them at the level of discourse (genre) and clause (register).
We characterize this type of analysis as having five steps, which are listed in
Table 11.3, though not all analysts report engaging all five steps or follow-
ing them in the same order. Moreover, though we list the steps sequentially,
in practice, they are often quite iterative as analysts move back and forth
between text samples, analyzing at the levels of discourse and clause, and
consulting experts or previous analyses.
First, analysts must collect a corpus of authentic texts of analytical inter-
est. Some select texts they believe to be of a similar purpose or type with
the intention of analyzing them for obligatory and optional structural ele-
ments. Others collect a corpus of different kinds of texts from a shared con-
text (e.g., student writing in higher education). Either way, analysts must
have some understanding of the context of use, as what proceeds will be
a text-in-context analysis. Corpus sizes vary widely depending on analysts’
goals, from numbers of texts in the single digits to several thousand. Among
analysts interested in genre description and classification, some undertake
broad analyses of a greater number of texts (e.g., Nesi & Gardner, 2012),
while others undertake an exhaustive analysis of fewer texts (e.g., Salmaso,

TABLE 11.3 
Five steps for SFL genre analysis

Steps Description

Step 1 Collect a corpus of authentic texts for which you have some
understanding of the context of use.
Step 2 Read each text in the corpus, laid out as it was written or published,
to get a sense of each text’s meaning and note your initial reactions.
Step 3 Review each text a second time to identify purpose, audience, and
authorship.
Step 4 Conduct linguistic analysis, identifying discourse semantic stages in
each text and the lexicogrammatical features that distinguish one
stage from another.
Step 5 Look across texts in the corpus to describe stabilities and variations
within the genre, or obligatory and optional genre stages.
228 Kathryn Accurso, Jennifer Walsh Marr

2022). However, quantity and depth are not mutually exclusive. Teachers
employing SFL genre analysis with students as part of the TLC or to ana-
lyze the impacts of their instruction as part of the ETLC tend to work with
smaller numbers of texts.
Second, analysts conduct a preliminary reading of each text in the cor-
pus, laid out as it was written or published. The goal of this step is to get a
sense of each text’s meaning and note initial reactions. There is no manip-
ulation of the text yet during this step, such as breaking it into clauses.
Painter (2005) explains that this is because in SFL analysis “meaning rather
than form should be the ‘way in’ to the text” (p. 73).
Third, analysts review each text a second time to identify purpose, audi-
ence, and authorship. Because SFL is a semiotic theory of society, “analysts
approaching genre from the perspective of SFL attend to more than just the
language involved” (Martin, 2020, p. 230). Across genre analyses reviewed
for this chapter, we found that analysts always identified the purpose of
texts in their corpus, but did not always explicitly attend to audience or
authorship. However, Melissourgou and Frantzi (2017) argue that identi-
fying “who to whom relations” is a critical part of genre identification (p.
382). Moreover, Accurso and Mizell (2020) argue that without attention to
these interpersonal aspects of a text, there is little evidence to suggest that
genre analysis will lead teachers or students to develop a critical orientation
toward text.
Fourth, analysts begin linguistic analysis, identifying discourse seman-
tic stages in each text and the lexicogrammatical features that distinguish
one stage from another. As analysts segment each text into goal-oriented
stages, they typically give each stage a functional name regarding its pur-
pose (e.g., define, orient, explain). Functional labels are analytically use-
ful as they provide researchers, teachers, and students with some terms
for talking about genres and their social functions. They are also peda-
gogically useful, as plentiful research has now shown that talking about
language scaffolds language development, and functional labels for genre
stages facilitate such talk (e.g., Painter, 2005). Analysts may consult exist-
ing analyses of genre stages as they segment and label the structure of
their texts (see Martin, 2012, p. 56). Clause-level linguistic analysis serves
to confirm or refine identified genre stages, as shifts in lexicogrammar
indicate movement from one stage to another.3 Lexicogrammatical shifts
might include changes in types of participants or processes that indicate
who is doing what to whom (ideational meaning resources; see Halliday
& Hasan, 1985), changes in mood or modality (interpersonal meaning
resources; see Arancón, 2013), or changes in theme that signal a new
point of departure in the text (textual meaning resources; see Ravelli,
2000). Detailed guides for clause-level analysis include Derewianka and
 Systemic functional linguistics and the (expanded) teaching and learning cycle 229

Jones (2016) and Humphrey et al. (2010) for teachers, Eggins (2004) for
a range of beginning analysts, and Halliday and Matthiessen (2014) for
comprehensive connections to theory. Though some argue that this level
of analysis can quickly become too detailed or delicate, clause-level analy-
sis is critical because it provides linguistic justification for the boundaries
analysts draw between text structure stages, protecting their analysis from
critiques of being simply “intuitive” or “ad hoc” (Christie & Unsworth,
2005, p. 14). Moreover, it allows analysts to better classify genres by com-
paring the frequency of particular lexicogrammatical features (e.g., Nesi &
Gardner, 2012).
Fifth and finally, analysts look across texts in the corpus to describe sta-
bilities and variations within the genre. They identify obligatory elements
whose presence is necessary for texts to be seen as complete instances of the
genre and optional elements that may affect the overall structure of a text,
but whose presence does not affect a text’s recognition as an instance of the
genre (Christie & Unsworth, 2005). Analysts may explore how obligatory
and optional elements tend to be ordered across texts in their corpus and/
or address the possibility of iteration (e.g., Salmaso, 2022). Overall, this
part of the analytical process allows the analyst to determine whether texts
in their corpus are structured differently because they are different genres
or because they are different versions of the same genre.
It is worth noting that SFL genre analysis involves a certain amount of
“play” in interpretation. As Martin (2019) points out:

At a given point in a text there may be more than one analysis available;
there may be a need to revise analysis as the text unfolds; and there may
in the end be multiple analyses available that cannot be resolved. Some
analysts experience this difference between discourse interpretation and
grammar analysis as a source of frustration; others experience it with a
sense of liberation.
(p. 359)

Given this element of subjectivity, analysts can engage a number of strate-

gies to enhance reliability and validity in their genre study. For example,
depending on their goals, analysts might enhance quantitative reliability
by comparing their findings to texts in a larger corpus of their target genre
and/or related genres (e.g., Rothery & Stenglin, 2000). They may take a
constant comparative approach to qualitative analysis, working collabora-
tively with other analysts aiming to enhance validity and achieve a certain
amount of inter-analyst or inter-coder reliability. Or they may consult pro-
ducers of the genre to confirm that their analysis resonates as valid with
other members of the cultural context.
230 Kathryn Accurso, Jennifer Walsh Marr

11.4 Example studies
This section presents five studies conducted in K-12 contexts that illustrate
the methodological steps outlined in the previous section as they unfold for
more descriptive or more pedagogical aims (see Table 11.4). In general, we
selected studies that show how researchers, teachers, and teams connect genre
analysis to broader social action, such as the pursuit of more equitable edu-
cational experiences for marginalized learners. We focus on K-12 studies as
there is a particularly rich history of these in SFL scholarship. Readers inter-
ested in post-secondary genre analyses might consult Abdel-Malek (2020),
Kawamitsu (2015), Melissourgou and Frantzi (2017), Nesi and Gardner
(2012), or Pessoa et al. (2020) as recent example studies. Readers interested
in SFL genre analyses that are not connected to pedagogy might consult
Ansary and Babaii (2005), Martin (2020), or Thompson et al. (2017).
We begin with Coffin (2006a, 2006b), who exemplified the breadth of
layers SFL genre analysis encompasses and how to work across these layers
in a large corpus for purposes of genre description and classification. The
study focused on genres in the Australian history curriculum. Here and in
Table 11.4 we make reference to both an article-length (2006a) and book-
length (2006b) account of this research for readers interested in different
degrees of analytical detail. Over the course of two years, Coffin collected
approximately 1,000 samples of student writing from Grades 7–12 his-
tory classrooms in Sydney. Ethnographic data such as national curriculum
standards, textbooks, lesson plans, and teacher assessments of the writing
supported Coffin’s understanding of the context of use and the meanings
students were making in these texts. Coffin also conducted semi-structured
interviews with history educators and experts, using their perspectives to
identify a narrower “mini-corpus” of approximately 100 “successful” texts,
which she subjected to detailed linguistic analysis.
Coffin’s analysis revealed three main purposes for student writing across
grade levels: recording, explaining, and arguing. These overarching pur-
poses were accomplished through students’ use of nine distinct genres,
which Coffin arranged along a grade-level continuum, noting that mean-
ings became more abstract as students progressed. Through more detailed
linguistic analysis, Coffin demonstrated that “grammatical and lexical
shifts correlate[d] with shifts in genre” (2006a, p. 422). She substantiated
this claim by plotting several patterns along a continuum of development.
For instance, as grade-level expectations shifted from talking about what
individual historical actors did to historical phenomena and their impacts,
Coffin found increased use of nominalization in student texts (i.e., the
making of nouns from verbs, a type of grammatical abstraction found at
the clause level).
TABLE 11.4 
Example studies adopting SFL genre analysis and the (E)TLC

Authors Context Research Questions (RQs) Data Collected Comments on Study Design
Coffin (2006a, • Secondary history • RQ1. What are the • RQ 1 and 2: Corpus • “Successful” texts were
2006b) (Grades 7–12) in main types of texts of 1,000 student texts identified through
Australia students produce within • RQ2: 100 interviews with teachers
the secondary history “successful” student and historians (i.e.,
curriculum? texts disciplinary experts)
• RQ2. What are • RQ3: Student writing
linguistic features of from 17 teachers’
“successful” texts? classrooms before
What are similarities and after genre-based
and differences in texts instruction
as students progress
through grade levels?
• RQ3. How is student
writing influenced by
genre-based instruction?
Gebhard et al. • Elementary social • RQ1. How did a • RQ1: Six model texts, • Children learned
(2014) studies, English Grade 3 teacher and teaching materials, metalanguage for genre
language arts, science her students use SFL video recordings of analysis a little bit at
(Grade 3) in the U.S., metalanguage to instruction a time, returning to
focus on multilingual analyze explanation • RQ2: 42 samples a text multiple times
learners genres across disciplines? of student writing; with different analytical
• RQ2. What was the role standardized lenses, a recommended
of SFL metalanguage assessment scores approach for beginning
in ELLs’ literacy gains analysts and teachers
over the course of an
academic year?
Systemic functional linguistics and the (expanded) teaching and learning cycle 231

(Continued)
TABLE 11.4 Continued
232

Authors Context Research Questions (RQs) Data Collected Comments on Study Design
Gebhard et al. • Secondary English • RQ1. How can genre • RQ1: 18 model texts • Teachers/researchers
(2019) language arts, analysis support (across five genres), did genre analysis
science, math, social refugee-background teaching materials, on model texts to
studies (Grade 9–10) learners in drawing video recordings of inform instruction,
in the U.S., focus on on multilingual and instruction then involved students
multilingual learners multimodal resources • RQ2: 72 post- in genre analysis as
to read, write, and instruction texts part of instruction, then
discuss disciplinary written by 19 refugee- analyzed student texts to
texts? background students evaluate the impacts of
• RQ2. How did instruction
students’ literacy
Kathryn Accurso, Jennifer Walsh Marr

practices change
over time as they
participated in genre-
based instruction?
Graham (2023) • Secondary English • RQ1. What genre stages • RQ1: Corpus of • Reflects all five
language arts (Grade and linguistic features 10 model texts methodological steps:
8) in the U.S. do Grade 8 students • RQ2: Post-instruction compile corpus; read
identify through a texts written by four for meaning; identify
genre analysis of author focal students purpose, audience,
bio texts? author; linguistic
• RQ2. To what degree analysis; identify
do students use these obligatory/optional
stages and linguistic features
features in their own • Exemplifies the ETLC
bio writing?
Moore (2019) • Elementary English • RQ1. What are the • RQ1 and 2: Corpus • Clearly explicates
language arts (Grades features of literary of 45 student texts reliability and validity
4–5) in the U.S. arguments written by strategies (e.g., multiple
primary grade students analysts, constant
who participated in comparative coding, use
genre-based lessons? of agreement measures,
• RQ2: In what ways seeking “deviant
were their texts similar cases” to test emerging
and varied? patterns, transparency in
methods)
Systemic functional linguistics and the (expanded) teaching and learning cycle 233
234 Kathryn Accurso, Jennifer Walsh Marr

Graham (2023) used a smaller corpus for explicitly pedagogical aims,

engaging with the ETLC by doing genre analysis with her secondary stu-
dents. We selected this study because it reflects all five methodological steps
described in the previous section and is aimed at novice scholars and/or
teachers contemplating an SFL-informed approach to pedagogy and teacher
action research. Graham first guided her diverse Grade 8 students through
a genre analysis of author bios, such as one might find on a book jacket.
Then, she independently conducted a genre analysis of author bios written
by her students to analyze the impact of this instruction. This approach
was motivated by what is referred to within the ETLC as a “problem of
practice” (Gebhard & Accurso, 2023, p. 8). Graham wanted to introduce
students to authors they would be reading in her class that year (a content
goal), build their genre knowledge and a shared language for text analysis
(a language/literacy goal), and affirm and showcase students’ identities as
authors themselves (an equity goal). Toward these aims, Graham compiled
a corpus of ten author bio texts. Then, Graham and her students identi-
fied purposes and audiences for these texts; read each one for meaning;
iteratively identified and named discourse semantic stages (first as a whole
class, then in small groups and individually); determined which stages were
obligatory and optional; and explored lexicogrammatical choices that were
functional for constructing different genre stages.
Graham introduced some SFL metalanguage for students to recognize
and label text structure and ideational, interpersonal, and textual mean-
ing-making resources (e.g., genre, genre move, register, field, tenor, mode,
appraisal, time tracker). She also encouraged students to contribute their
own metalanguage (e.g., SUPER EXCITING language, bragging lan-
guage). Classroom transcripts showed that across five lessons, Graham’s
students, including multilingual learners, used SFL and class metalanguage
with ease to analyze texts. For example, students identified four obligatory
genre stages: author background, types of writing they engage in, most recent
piece of writing, and information about the author not related to writing.
They incorporated these stages into a new text, working together to write
Graham’s bio before working on their own individual bios. Graham then
used genre analysis on her students’ texts. One multilingual student’s bio
is highlighted in the study with genre moves and register choices labeled
to suggest the pedagogy was effective for meeting Graham’s content, lan-
guage/literacy, and equity goals.
Gebhard et al. (2019) also employed SFL genre analysis in a secondary
classroom to make language features more visible for multilingual students
who were expected to read, write, discuss, and critique progressively more
analytical genres. We included this study because it analyzed the cumulative
effects of this practice longitudinally and across content areas. The authors
 Systemic functional linguistics and the (expanded) teaching and learning cycle 235

identified five commonly assigned genres at a “high risk” and “underper-

forming” high school (autobiography, poetry, scientific description, math
report, social studies argument) and curated a mini-corpus of three to eight
sample texts for each genre with the help of other content teachers and dis-
ciplinary experts. Then, they designed text analysis guides based on existing
genre descriptions (e.g., Derewianka & Jones, 2016; Gebhard et al., 2019;
Schleppegrell, 2004). Students used these guides to analyze corpus texts
for purpose and audience, obligatory and optional text structure stages,
and relevant lexicogrammatical features. Then, they practiced joint and
independent writing, using learning from their analyses to “make strategic
decisions in their writing” (p. 258). This approach sought to acknowledge,
honor, and expand students’ multilingual meaning-making resources, but
also support them in developing varieties of English needed to meet gradu-
ation requirements and access a wider range of post-secondary opportuni-
ties in the city where they lived.
Case study data showcased students’ valuation of learning how authors
write in expected genres through the process of joint analysis and text con-
struction. Longitudinal data demonstrated a focal student’s significantly
improved performance in spite of socio-economic hindrances beyond her
control. These findings support calls for increased collaboration in which
teachers and researchers use genre analysis on model texts to inform instruc-
tion, involve students in genre analysis as part of instruction, and then
analyze student texts to evaluate the impacts of instruction and address
destructive discourses about students’ potential and the role of their full
linguistic repertoires in the development of disciplinary literacies.
Gebhard et al. (2014) exemplified this type of collaboration around
genre analysis in the context of a bilingual elementary school. We selected
this study for inclusion because the authors analyzed a single genre fam-
ily (i.e., explanations) across three content areas. After a review of existing
genre studies, the authors analyzed Grade 3 historical explanations, bio-
graphic explanations, and scientific explanations with students, richly illus-
trating the connection between language strata as they identified stabilities,
differences, and lexicogrammatical features across the three genres. The
authors also detailed which metalanguage students used for their analysis,
and the ways that children approached the task of learning functional meta-
language a little bit at a time, returning to a text multiple times with dif-
ferent analytical lenses, an approach beginning analysts and teachers would
do well to emulate.
Regarding the impacts of incorporating this kind of genre analysis into
instruction, the authors analyzed 42 student texts produced over the course
of one academic year as well as 85 hours of recorded instruction. Three
focal cases illustrated the types of literacy development captured in this
236 Kathryn Accurso, Jennifer Walsh Marr

corpus, including student texts which were typically longer, more coherent,
and written more independently and with enhanced accuracy, logic, and
overall genre and register knowledge.
Moore (2019) also analyzed student writing but delved into an under-
explored genre: literary arguments written in the elementary grades. We
included this study because Moore outlined his specific strategies for
achieving reliability and validity, which other studies rarely explicated.
Following the pilot of a genre-based curriculum, Moore collected a cor-
pus of 45 student-written responses to literature from two classrooms, one
Grade 4 and one Grade 5. Like other studies highlighted here, Moore’s
analytical process involved labeling each text’s genre stages and use of
lexicogrammatical choices to construe attitude (an interpersonal meaning
resource), theme (a textual meaning resource), and participants and processes
(ideational meaning resources). However, Moore also “generated notes
about the perceived successes and struggles of students’ writing to advance
through these stages and write in the target genre” (p. 441) and subjected
these notes to a constant comparative thematic coding process. Moreover,
he scored each text using a rubric that defined the qualities of successful,
partially successful, and unsuccessful literary arguments. Regarding reli-
ability and validity, each step was undertaken by multiple researchers who
compared and corroborated findings as Moore also searched for “deviant
cases” to test emerging patterns. Rubric scores among researchers reached
a statistical threshold of agreement.
These steps allowed Moore to describe variation and stabilities in the
corpus with some degree of confidence, ultimately providing substantial
evidence that a majority of writers in the two classrooms produced logi-
cally organized texts with the following stages: establishing an interpretive
stance, providing a reason for that stance, specific textual evidence to support
claims, interpretation of character attitudes, and evaluation of characters.
Some texts also included a successful orientation to evidence stage that con-
textualized the evidence to be presented. Finer-grained register analyses
richly described the ways students more and less effectively accomplished
these stages.

11.5 Issues and challenges

A persistent challenge facing SFL genre analysts is that this approach is
time-consuming and requires one to manage various, interacting levels of
detailed analysis (i.e., discourse semantic and lexicogrammatical systems).
SFL is admittedly a “comprehensive” and “extravagant” theory (Halliday,
1997, pp. 6–7) that offers many lenses for looking at texts. Yet Halliday
contends that this is so because language is “very big,” and does the work
 Systemic functional linguistics and the (expanded) teaching and learning cycle 237

of construing “all of our experience” and enacting “all of our interper-

sonal processes” (p. 6, emphasis original). Therefore, he maintains, “the
extravagance of modeling the same domain of experience in more than one
way leads to a richer and more life-supporting account” (p. 8), which is, of
course, a core goal of SFL genre analysis. Nevertheless, beginning analysts
may question the depth of systemic functional grammar knowledge needed
to undertake a genre analysis in this tradition.
Macken-Horarik (2008) addressed this tension between the technical-
ity and practicality of SFL genre analysis by arguing that an exhaustive
knowledge of SFL is not needed, but rather a “good enough grammatics”
and a “mind alive to changes of form (and mode) and able to see common-
alities across forms of textuality” (p. 43). This “good enough” approach
means that beginning analysts need not know all of the functional labels
SFL provides across levels of language to get started (Ravelli, 2000). Nor
should they see their work as “a proliferation of labeling for the sake of
having labels”; rather, analysis should be a form of “grammatical reason-
ing” (p. 37) making distinctions between genre stages based on semantic
and lexicogrammatical grounds. Our review of studies for this chapter sug-
gests this type of reasoning may be most productively undertaken collabo-
ratively, whether the collaboration is between researchers, teachers, and/or
students, as illustrated in Graham’s (2023) analysis of author bios with her
students.
A second issue facing analysts is the need to distinguish the deep descrip-
tion characteristic of SFL genre analysis from normative approaches to lin-
guistic analysis, which serve to objectify or naturalize the (often dominating)
language practices being analyzed while drawing attention away from con-
ditions of inequality in their contexts of use (Rosa, 2021). Gebhard et al.
(2014, 2019), Graham (2023), and others included in our exemplar studies
provide models of how to articulate critical commitments clearly and resist
reporting results of analyses and the design of genre-based pedagogy in
ways that make language seem fixed, objective, decontextualized, or that
position one variety as inherently more valuable than another. Given the
critical origins and aims of SFL as a theory of text-in-context, we recom-
mend analysts avoid unintentional decontextualization by investing time in
understanding the history, not simply the structure, of genres, and situat-
ing their research, texts, participants, purpose, and context(s) accordingly.
A third, related issue is the lack of genre analyses of translanguaged texts
(i.e., texts that reflect multilingual repertoires in their use of multiple lan-
guages/language varieties) available for reference. As there are currently
many efforts underway to understand, affirm, and teach with translan-
guaged texts (e.g., García et al., 2021; Khote & Tian, 2019; Ramírez,
2020), having more studies of this nature would reflect SFL’s underlying
238 Kathryn Accurso, Jennifer Walsh Marr

disposition toward inclusion and support analysts interested in this line of

inquiry.

11.6 Study-in-focus
In this section, we discuss a final example study. Salmaso (2022) exam-
ined an overlooked narrative genre: the short story. As a university English
professor in Argentina, Salmaso found that she often worked with short
stories, yet there did not seem to be an existing description or classification
in previous genre studies. Therefore, she aimed to better understand the
genre potential of short stories and what, if anything other than length,
makes them distinct from other narrative genres. Salmaso collected a cor-
pus of six texts that she regularly used with her students, acknowledging
the small size of this corpus, but asserting that this exploratory study privi-
leged an “exhaustive analysis” (p. 42) of the texts over a larger data set.
Salmaso’s findings are detailed in the call-out box, which we encourage
readers to review before continuing with our commentary.

Citation
Salmaso, G. S. (2022). Modelling of the genre short story for teaching pur-
poses: An approach from a systemic functional perspective. International
Journal of TESOL Studies, 4(2), 39–55.

Research questions
• What is the structure of the literary genre “short story”? What purpose
and contribution does each genre stage make to the text as a whole? Do
genres occur within the genre of the short story (e.g., narratives within
narratives, descriptive reports within narratives)?
• How can genre knowledge support university students’ reading
comprehension?

Context and population

This study examined six short stories of similar length from upper under-
graduate courses (Language III and Language IV) in the English Department
of a university in Argentina. Observing that instructors tended to guide stu-
dents’ analysis of short stories “intuitively” rather than from a linguistic per-
spective, the researcher aimed to precisely model aspects of the short story
to support the teaching and learning process.
Systemic functional linguistics and the (expanded) teaching and learning cycle 239

Procedure
Salmaso’s SFL genre analysis followed six steps: (1) read the texts thoroughly
with previous work on narrative genre families in mind to identify each text’s
overarching purpose and typical structure (i.e., recount, anecdote, narra-
tive, exemplum, or observation); (2) within each text, identify genre stages
and each stage’s function toward the text’s larger social purpose; (3) note
examples of genres inserted within stages of the larger text; (4) if/where
inserted genres are present, identify their genre stages and the purpose of
each stage; (5) compare the genre structure of each text in the corpus to
identify similarities and differences; (6) based on the comparison, identify a
generalized macro-structure of short story texts in the corpus and the range
of possible genre-inside-genre insertions.

Findings
Short stories in the corpus shared the following macro-structure: an abstract/
orientation that foreshadowed the main event of the story and introduced
characters and setting; a record-of-events that constructed a general narrative;
a disruptive event that interrupted the general narrative; either a reaction to
and interpretation of the disruption or a resolution and evaluation of it; and
an optional reorientation to time and place and/or coda that offered a final
reflection on the events of the short story.
Within this macro-structure, Salmaso found other genres inserted to
help the short story achieve its purpose. For example, a recount genre may
be inserted into the record-of-events stage, and in this corpus, always was.
Narrative genres were sometimes inserted into the disruption stage to con-
struct disruptive events and reactions to them. However, narrative genres
were not inserted into descriptive stages, such as the orientation or resolu-
tion, because these stages present and describe participants in the short
story, not events.
Regarding pedagogy, Salmaso found that attention to short story macro-
structure and possible insertions enabled better comprehension when stu-
dents revisited stories they had already read. Productive scaffolding activities
included: cutting a short story into its stages without labeling them (e.g.,
orientation, events, disruption) and having students analyze the purpose of
each stage following a series of guided questions; teacher and students jointly
constructing a new or expanded stage of the short story based on the macro-
structure; and students independently constructing a new short story based on
a familiar character and general narrative but leading up to a new disruption.
240 Kathryn Accurso, Jennifer Walsh Marr

As the call-out box highlights, two strengths of this study are Salmaso’s
clear outline of the steps she followed for genre analysis and her immedi-
ate application of findings to practice. Published SFL genre analyses are
generally quite good about including samples of analysis, but not always as
good at clearly detailing the analytical steps followed. Perhaps this is due
to word-limit constraints in journals, but Salmaso managed to do both
quite clearly. In the second half of this study, Salmaso detailed the kinds
of pedagogical activities she designed to support her students’ reading of
these texts based on her analysis. Together, Salmaso’s clarity of analysis and
practical applications make the findings both understandable and usable to
other educators.

11.7 Future research directions

A major strength of the SFL approach is the systematic connection between
descriptive analysis and pedagogical practice. We have highlighted just a
few such examples, but readers will find many more if they trace the his-
tory of implementation in Australia and around the world. However, as
pedagogical horizons rapidly expand in terms of multimodality and multi-
lingualism, more studies are needed that use genre theory to analyze these
aspects of text production (e.g., Cocetta, 2018; Harman & Burke, 2020;
Hood & Maggiora, 2016). These types of inquiries will make important
contributions, extending existing discussions of semiotic features across
disciplines (e.g., Walsh Marr & Mahmood, 2021; Walsh Marr & Martin,
2021), and ongoing efforts to map curriculum genres in both K-12 con-
texts and higher education (e.g., Dreyfus et al., 2016; Ferreira & Zappa-
Hollman, 2019).
In addition, critical SFL scholars have called for new lines of work that
reinvigorate the concept of genre as a tool for social change (e.g., Mahboob
& Szenes, 2010). For example, future genre analyses might look beyond
traditional genres of schooling to explore knowledge-rich community gen-
res that have been historically marginalized for reasons related to race, class,
gender, and so on (e.g., Black and Indigenous oral histories). Mizell (2022)
argues that analyzing community genres can play a key role in the pursuit
of pedagogies that “center, value, and sustain the literacies and languaging
practices of racialized youth and their communities” (p. 3). Accurso and
Mizell (2020) further advocate for studies that engage side-by-side analysis
of dominating texts and community countertexts. However, remember-
ing SFL’s origins as both a “socially accountable” and “appliable” theory
(Troyan et al., 2021, p. 294), we urge researchers interested in pursuing
this work to orient to it not as research on a particular community, but as
research with a particular community.
 Systemic functional linguistics and the (expanded) teaching and learning cycle 241

Notes
1 Some SFL scholars differ in the ways they understand genre and thus, how they
model text/context dynamics. Halliday saw genre as part of textual meaning-
making (a mode resource), while Hasan argued that obligatory elements in
text structure are a matter of ideation (a field resource) (Halliday & Hasan,
1985). Martin (1992) maintained that genre is a “pattern of register patterns”
(p. 506) corresponding to culture, as shown in Figure 1. This chapter draws
on a Martinian conception of genre, not intending to downplay theoretical
differences or promote the common misconception that Halliday, Hasan, and
Martin’s models are the same. Lukin et al. (2011) review the meaningful differ-
ences in the positioning of genre in different communities of SFL scholarship.
2 Because this is a guidebook to conducting research in applied linguistics involv-
ing written genres, text in this chapter largely refers to written texts. However,
SFL scholarship defines text much more broadly, including written, oral, and
multimodal compositions.
3 Examples of detailed analyses relating genre stages to lexicogrammatical pat-
terns can be found in Martin’s (2020) analysis of legal exemplum, Hasan’s
(2015) analysis of fairy tales, or Eggins’ (2004, pp. 23–53) analysis of stories.

References
Abdel-Malek, M. (2020). Empowering Arabic learners to make meaning: A genre-
based approach. System, 94, 102329.
Accurso, K. (2017). Exploiting the language of science: Teaching informational
texts across grade levels. SPELT Quarterly Journal, 32(4), 2–16.
Accurso, K., & Mizell, J. D. (2020). Toward an antiracist genre pedagogy:
Considerations for a North American context. TESOL Journal, 11(4), e554.
Ansary, H., & Babaii, E. (2005). The generic integrity of newspaper editorials: A
systemic functional perspective. RELC Journal, 36(3), 271–295.
Arancón, P. R. (2013). The use of SFL genre theory for the analysis of students’
writing skills in ESP. Revista Española de Lingüística Aplicada, 1, 245–262.
Bartley, L. V. (2018). “Justice demands that you find this man not guilty”: A
transitivity analysis of the closing arguments of a rape case that resulted in a
wrongful conviction. International Journal of Applied Linguistics, 28(3),
480–495.
Brisk, M. E. (2015). Engaging students in academic literacies: Genre-based pedagogy
for K-5 classrooms. Routledge.
Christie, F., & Unsworth, L. (Eds.). (2005). Researching language in schools and
communities: Functional linguistic perspectives. Bloomsbury.
Cocetta, F. (2018). Developing university students’ multimodal communicative
competence: Field research into multimodal text studies in English. System, 77,
19–27.
Coffin, C. (2006a). Learning the language of school history: The role of linguistics
in mapping the writing demands of the secondary school curriculum. Journal of
Curriculum Studies, 38(4), 413–429.
Coffin, C. (2006b). Historical discourse: The language of time, cause and evaluation.
Bloomsbury.
 242 Kathryn Accurso, Jennifer Walsh Marr

Derewianka, B., & Jones, P. (2016). Teaching language in context (2nd ed.). Oxford
University Press.
Dreyfus, S., Humphrey, S., Mahboob, A., & Martin, J. R. (2016). Genre pedagogy
in higher education: The SLATE project. Palgrave Macmillan.
Eggins, S. (2004). Introduction to systemic functional linguistics. A&C Black.
Ferreira, A. A., & Zappa-Hollman, S. (2019). Disciplinary registers in a first-year
program: A view from the context of curriculum. Language, Context and Text,
1(1), 148–193.
García, O., Flores, N., Seltzer, K., Wei, L., Otheguy, R., & Rosa, J. (2021). Rejecting
abyssal thinking in the language and education of racialized bilinguals: A
manifesto. Critical Inquiry in Language Studies, 18(3), 203–228.
Gebhard, M. (2019). Teaching and researching ELLs’ disciplinary literacies: Systemic
functional linguistics in action in the context of U.S. school reform. Routledge.
Gebhard, M., & Accurso, K. (Eds.). (2023). In pursuit of a multilingual equity
agenda: SFL teacher action research. Routledge.
Gebhard, M., Accurso, K., & Harris, G. (2019). Putting it all together: SFL in
action. In M. Gebhard (Ed.), Teaching and researching ELLs’ disciplinary
literacies (pp. 239–271). Routledge.
Gebhard, M., Chen, I. A., & Britton, L. (2014). “Miss, nominalization is a
nominalization:” English language learners’ use of SFL metalanguage and their
literacy practices. Linguistics and Education, 26, 106–125.
Graham, H. (2023). “How do you do this?”: Getting started with SFL and genre
pedagogy with multilingual middle schoolers. In M. Gebhard & K. Accurso
(Eds.), In pursuit of a multilingual equity agenda: SFL teacher action research
(pp. 31–60). Routledge.
Halliday, M. A. K. (1978). Language as social semiotic. Arnold.
Halliday, M. A. K. (1997). Linguistics as metaphor. In A. Simon-Vandenbergen, K.
Davidse, & D. Noël (Eds.), Reconnecting language (pp. 3–28). John Benjamins.
Halliday, M. A. K., & Hasan, R. (1985). Language, context, and text. Deakin
University Press.
Halliday, M. A. K., & Martin, J. R. (2003). Writing science: Literacy and discursive
power. Falmer Press.
Halliday, M. A. K., & Matthiessen, C. M. (2014). An introduction to functional
grammar (4th ed.). Routledge.
Harman, R. M., & Burke, K. J. (2020). Culturally sustaining systemic functional
linguistics praxis: Embodied inquiry with multilingual youth. Routledge.
Harman, R. M., & Khote, N. (2018). Critical SFL praxis with bilingual youth:
Disciplinary instruction in a third space. Critical Inquiry in Language Studies,
15(1), 63–83.
Hasan, R. (2015). The nursery tale as genre. In C. Cloran, D. Butt, & G. Williams
(Eds.), Ways of saying, ways of meaning (2nd ed., pp. 51–72). Bloomsbury.
Hattingh, N. (2011). A systemic functional analysis of two truth and reconciliation
commission testimonies [Doctoral dissertation]. University of the Western Cape,
ProQuest.
Hood, S., & Maggiora, P. (2016). A lecturer at work: Language, the body and
space in the structuring of disciplinary knowledge in law. In H. de Silva Joyce
(Ed.), Language at work in social contexts (pp. 108–128). Cambridge Scholars.
 Systemic functional linguistics and the (expanded) teaching and learning cycle 243

Humphrey, S., Martin, J. R., Dreyfus, S., & Mahboob, A. (2010). The 3x3: Setting
up a linguistic toolkit. In A. Mahboob & N. Knight (Eds.), Appliable linguistics
(pp. 185–195). Bloomsbury.
Kawamitsu, S. (2015). Introducing genre into Japanese-as-a-foreign-language. L2
Journal, 7(4), 63–90.
Khote, N., & Tian, Z. (2019). Translanguaging in culturally sustaining systemic
functional linguistics: Developing a heteroglossic space with multilingual
learners. Translation and Translanguaging in Multilingual Contexts, 5(1), 5–28.
Lukin, A., Moore, A. R., Herke, M., Wegener, R., & Wu, C. (2011). Halliday’s
model of register revisited and explored. Linguistics and the Human Sciences,
4(2), 187–213.
Macken-Horarik, M. (2008). A ‘good enough’ grammatics: Developing an
effective metalanguage for school English in an era of multiliteracies. In Wu,
C., Matthiessen, C., & Herke, M. (Eds.), Proceedings of ISFC 35: Voices Around
the World (pp. 43–48). 35th ISFC Organizing Committee.
Mahboob, A., & Szenes, E. (2010). Linguicism and racism in assessment practices
in higher education. Linguistics and Human Sciences, 3(3), 325–354.
Martin, J. R. (1992). English text: System and structure. Benjamins.
Martin, J. R. (2012). Grammar meets genre: Reflections on the ‘Sydney school’.
Arts: The Journal of the Sydney University Arts Association, 22, 47–95.
Martin, J. R. (2019). Discourse semantics. In G. Thompson, W. Bowcher, L.
Fontaine, & D. Schöntal (Eds.), Cambridge handbook of systemic functional
linguistics (pp. 358–381). Cambridge University Press.
Martin, J. R. (2020). Genre and activity: A potential site for dialogue between
Systemic Functional Linguistics (SFL) and Cultural Historical Activity Theory
(CHAT). Mind, Culture, and Activity, 27(3), 216–232.
Melissourgou, M. N., & Frantzi, K. T. (2017). Genre identification based on SFL
principles: The representation of text types and genres in English language
teaching material. Corpus Pragmatics, 1(4), 373–392.
Mizell, J. D. (2022). Culturally sustaining systemic functional linguistics: Towards
an explicitly anti-racist and anti-colonial languaging and literacy pedagogy.
Linguistics and Education, 72(22), 101108.
Moore, J. (2019). Choice and constraint: Using SFL genre theory to teach primary-
grade students to write arguments about literature. Journal of Writing Research,
10(3), 429–464.
Nesi, H., & Gardner, S. (2012). Genres across the disciplines: Student writing in
higher education. Cambridge University Press.
Painter, C. (2005). Learning through language in early childhood. A&C Black.
Pessoa, S., Gomez-Laich, M. P., & Mitchell, T. D. (2020). Mapping the case
analysis genre continuum in an Information Systems program. Journal of
Writing Research, 12(2), 291–320.
Ramírez, A. (2020). The case for culturally and linguistically relevant pedagogy:
Bilingual reading to learn for Spanish-speaking immigrant mothers. System, 95,
102379.
Ravelli, L. (2000). Getting started with functional analysis of texts. In L. Unsworth
(Ed.), Researching language in schools and communities: Functional linguistic
perspectives (pp. 27–64). Continuum.
 244 Kathryn Accurso, Jennifer Walsh Marr

Rosa, J. (2021, March 20). Never enough language/language is never enough:

Raciolinguistic perspectives on applied linguistic theories of change [plenary talk].
American Association of Applied Linguistics, virtual conference.
Rose, D., & Martin, J. R. (2012). Learning to write, reading to learn: Genre,
knowledge, and pedagogy in the Sydney school. Equinox.
Rothery, J., & Stenglin, M. (2000). Interpreting literature: The role of appraisal.
In L. Unsworth (Ed.), Researching language in schools and communities (pp.
222–244). Cassell.
Salmaso, G. S. (2022). Modelling of the genre short story for teaching purposes:
An approach from a systemic functional perspective. International Journal of
TESOL Studies, 4(2), 39–55.
Schleppegrell, M. J. (2004). The language of schooling: A functional linguistics
perspective. Routledge.
Thomson, E. A., Sano, M., & de Silva Joyce, H. (Eds.). (2017). Mapping genres,
mapping culture: Japanese texts in context. John Benjamins.
Troyan, F. J., Harman, R., & Zhang, X. (2021). Critical SFL praxis in teacher
education: Insights from Australian SFL scholars. Language and Education,
35(5), 383–401.
Walsh Marr, J., & Mahmood, F. (2021). Looking past limiting conditions:
Prioritizing meaning in EAP. Journal of English for Academic Purposes, 51,
100979.
Walsh Marr, J., & Martin, J. L. (2021). Pomp and circumstances: From research, in
practice, for students. Íkala, 26(1), 227–242.
INDEX

abstracts 1, 63, 65, 83, 88, 91–92, corpus-driven approach 106–108, 115
107, 120–122, 156, 162 Create-a-Research-Space (CARS) 83,
academic discourse 62–63, 84, 156, 161
205, 212 critical orientation 225, 228
academic literacy (AcLits) 15, 22, 72
accuracy (as a measure) 149, 150, deconstruction 225
152–161, 163–164, 168, 236 delimitation (of a case) 14, 16, 18
action research 101, 225, 234 discourse community 2, 36, 83–86,
adaptive transfer 202 92–93, 140, 205
automated tools/measures 6, 141, dispersion 109, 115
153–154, 164–165
email 84–85, 91, 93–94, 101, 168,
Biber Tagger 133, 137, 139, 140, 142, 206, 213
145–146 emic 38–40, 47, 53
bilingualism 201 engagement 63, 66, 71, 75–77
bilingual mind 199 engineering 71, 89, 92, 98–99,
British Academic Written 121, 136
English (BAWE) Corpus 98, 128, English for Specific Purposes (ESP) 1,
141–142, 145 61, 83, 105, 173, 200, 225
business 62, 84–85, 91, 93–94, equity/equity goal 221, 225, 234
117, 174 ethics 49–50, 208
etic 38–39, 53
cognitive complexity 150, 159, 167
communicative purpose 36, 62, factor analysis 127, 129–130, 132–133,
82–83, 105, 107, 111, 116, 137, 139–141, 143, 146, 210
127–128, 145, 176 fillers 107–108, 110–111, 113, 115,
concordance 65, 71 118–119
conditional knowledge 24, 202, 216 fluency 149–150, 152–153, 157–159,
contrastive studies 63 163–164, 168, 200
core expression 120 Fulbright 89, 95


246 Index

gatekeepers 88 New Rhetoric 1, 4, 225–226

Genre and Multimodality (GeM) nominalization 138, 143, 150, 159,
180–181, 185 161–162, 230
genre-specific knowledge 200, 202, non-language dependent 200
205, 207–211, 215–217 novice 28, 30, 37, 64, 78, 84–85, 88,
genre stages 221–222, 224, 226–228, 96–97, 107, 130, 168, 184, 201, 234
232, 234, 236–239, 241
occluded genres 85, 88
ideational meaning 3, 74, 127–128,
176, 184, 207, 221–222, 224, 226, paradigm 14, 16, 94, 200
228, 234, 236 part-genres 89, 92, 97
informational density 131 participant observation 39–41, 43, 45,
interactional metadiscourse 63, 70, 73 47–48
inter-coder/inter-rater agreement and persuasive writing/texts 63, 68,
reliability 19, 65, 71–72, 76, 78, 70–71, 84
87–88, 92–96, 99–100, 164, 186, Principal Component Analysis (PCA)
206, 208, 229 127, 130, 132, 139–140, 146
interpersonal (meaning) 3, 5, 60–62, promotional genre 88, 93, 112
67–68, 70–72, 74–75, 128, 176,
178, 184–185, 207, 221–224, 226, range (when identifying n-grams)
228, 234, 236–237 108–109, 115, 119–122
recontextualization 77, 202, 216
language development 153, 159, 228 reflexivity 17–18, 23, 25–26, 39, 41,
lexical complexity 150, 152, 154, 155, 50, 60
158–160, 162, 165–166 register 17, 22
lexicogrammar/lexicogrammatical rhetorical functions 111–112, 116,
features 107, 112, 145, 156, 118–119, 121, 156, 161–162
221, 223–224, 226–229, Rhetorical Genre Studies (RGS) 1–2,
234–237, 241 4–7, 127, 200–201, 205, 207
linguistic phase 225
single case 17, 22
mathematics 22, 66, 85, 90, 94–95, socially accountable 221, 240
180, 182–184 social action 177, 199, 205–206, 230
member checking 42, 87, 208 social practice 27, 35–37, 39–40, 49,
metacognitive knowledge 200–201, 53–54, 62–63, 174, 199, 206
203, 208, 215–216 social semiotics 174–175, 177–179,
metafunctions 60, 176, 178 181–182, 184, 186, 188–190
Michigan Corpus of Upper-Level stance 60–63, 66, 73, 76–77, 79,
Student Papers (MICUSP) 134, 136 109–110, 116, 122, 137, 140, 143,
mode 149, 158–159, 173–175, 179, 145, 178, 236
184–185, 188, 241 strata 235
multilingualism 199, 204, 208, 240 syntactic complexity 88, 151–157,
multilingual writing 200, 202, 159–164, 168, 190
204, 213 Systemic Functional Linguistics (SFL)
Multimodal Corpus Analysis (MCA) 2–7, 15, 19, 24, 127, 140, 144–145,
175, 177, 179, 181, 185 173–174, 176, 182, 184, 200–201,
multimodality 172–173, 175, 177, 207, 221–224, 226–231, 234,
180, 187–188, 240 236–241
multimodal product 172–173, 176 Systemic Functional Multimodal
multimodal text-making 177, 188–189 Discourse Analysis (SF-MDA)
multiple case 13, 18, 21–23 175, 177
 Index 247

task-based 159, 167 translanguaging 199

Teaching and Learning Cycle (TLC)/ triangulate/triangulation 16, 20,
Expanded Teaching and Learning 22, 24, 29, 42, 45, 183,
Cycle (ETLC) 3–5, 7, 224–226, 204, 208
228, 234 trigger 17
text-in-context analysis 227, 237 trustworthy/trustworthiness 16, 18,
textography 41, 42, 44–46 22, 24, 26, 42, 100
textual meaning 178, 221–222, 226,
228, 234, 236, 241 university writing 48, 129, 135, 137,
translanguaged texts 237 140, 142